FOREST RESOURCE CONSERVATION AND RURAL LIVELIHOOD IMPROVEMENT IN GHANA: THE PROMISE OF A MODIFIED TAUNGYA SYSTEM By Doe Adovor A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Forestry – Doctor of Philosophy 2021 ABSTRACT FOREST RESOURCE CONSERVATION AND RURAL LIVELIHOOD IMPROVEMENT IN GHANA: THE PROMISE OF A MODIFIED TAUNGYA SYSTEM By Doe Adovor This dissertation addressed four questions: 1) What landcover changes have occurred in Yaya, Nsemre, and Sawsaw forest reserves in Ghana’s Brong Ahafo region before and after Ghana’s MTS program was launched in 2002? 2) What factors influenced the community and household selection into the MTS program? 3) What changes in livelihood assets have occurred among MTS participants and non-participant households since the launching of the program in 2002? And 4) To what extent are the changes in household livelihood assets attributable to the MTS program? An unsupervised landcover classification of January 1990 and 2000 LANSAT images and January 2012 DMC satellite images of the three reserves suggest an initial growth of 0.26 Km2/year in forest cover in Yaya between 1990 and 2000 and then a doubling of growth to about 0.52 Km2/year between 2000 and 2012. Unlike Yaya, forests cover in Nsemre declined by -0.37 Km2/year during the pre-MTS period and then increased sharply at an annual rate of 0.45 Km2 during the post-MTS era. Similar to the trends in Nsemre, Sawsaw likewise experienced an initial decline of -0.73 Km2 in forest cover before increasing annually at a rate of 0.31 Km2 post-MTS. To address research questions two and three, 878 households were surveyed in 19 communities in Yaya, Nsemre, and Sawsaw. Included in the Yaya survey were 406 MTS participants and 240 non-participant households. Also surveyed were another 232 non- participating households in nine communities around Nsemre (4) and Sawsaw (5) where the MTS program did not exist. A Binomial Probit Model (BPM) generated from the survey data, suggests that cash, farm inputs, and labor assistance from local religious organizations, migrant work, type of roof, and proximity to paved roads significantly (p<0.001) increased the expected probability of MTS community selection. Average land ownership exceeding two acres had a significantly (p<0.001) negative influence on MTS community selection. A GIS community mapping exercise conducted in the 19 research communities produced 1,446 household data points used in generating another BPM used to determine the extent to which proximity of households to communal physical assets influenced MTS selection. According to BPM, seven factors significantly increased the predicted probability of MTS community selection, while nine significantly reduced the chance of selection. Also, five factors positively influenced selection while two negatively impacted selection. In order to track changes in livelihood assets post- MTS, five livelihood indexes were constructed and compared for two time periods before and after MTS. The result suggests that aggregate household physical capital index increased by 17% while social and financial capital both increased by 2% between 1999 and 2009. Also, the human capital index increased by 3% while the natural capital index decreased by 5%. To address question four, a Difference in Difference (DID) model was used to isolate changes in the livelihood indexes between 1999 and 2009 that may be a directly attributed to the MTS project or spillover. The DID results indicate that on average, MTS resulted in a 6.8% increase in annual household income and income sufficiency index, a 2.3% increase in the combined household natural capital index, and a 4.8% increase in cropland ownership index. The DID also suggests that MTS participation may be directly responsible for the 2.5% increase in combined household physical capital index and the 1.5% aggregate household social capital index among project participants between 1999 to 2009. Of the DID livelihood indexes, the combined household human capital index (CHCI) was the least impacted by the MTS (0.4%). Copyright by DOE ADOVOR 2021 This work is dedicated to the people of Ghana’s Yaya, Nsemre and Sawsaw Forest Reserve Communities. It is my fervent hope and prayer that this work draws attention to your daily struggles to achieve descent and sustainable livelihoods from your forest resources. v ACKNOWLEDGEMENTS The field work for this research was conducted in 2009 and the dissertation successfully defended in a public forum in 2013. Several individuals and institutions contributed immensely to this research throughout the period of data collection leading to defense. This this section acknowledges all those who played a role in making this work a success. Michigan State University (MSU): I am truly grateful to Professor Yin my academic advisor and chair of my dissertation committee for all the support and encouragement he gave me throughout my doctoral study at MSU. Professor Yin (one of the leading authorities in Forest Resource Economics) helped shape my dissertation research right from the very conception, through predissertation fieldwork, data collection, and eventual defense and no words can describe the depth of gratitude I owe Professor Yin. I also owe a debt of gratitude to Professor John Kerr for his immerse interest in my graduate work and the guidance he provided throughout my graduate study and dissertation research leading up to my defense. Professor Kerr supported me with a lot of relevant literature, provided constructive feedback on my research, and took a special interest in preparing me for my final defense. Professor Kerr is the one person every student must have on their committee, and I count myself extremely lucky to have had him on mine. I wish to thank the rest of my committee members, Professor Maureen McDonough and Professor Karen Potter-Witter for their encouragement and insightful comments on my thesis. I would like to thank my very good friend Dr. Andrew Kizito Muganga for helping me with my econometrics/data analysis. I also owe a special thanks to my friend Mr. Felix Yeboah a Ph.D. student in Community Agriculture and Resource Studies for his constructive feedback on some of my chapters. vi MSU-USAID Partnerships for Food Industry Development- Fruits and Vegetables (PFID): I would like to thank the MSU-USAID PFID-F&V Southern Africa and Central America projects for financially supporting my doctoral study at MSU through a USAID funded Graduate Research Assistantship. Specifically, I would like to thank Dr. Russell Freed PFID- F&V Acting Director, the Deputy Director Dr. Andrea Allen, and the PFID-F&V Central America Project Manager Dr. Luis Flores for supporting my graduate work and also allowing me to grow as a student researcher with the project. Forest Research Institute of Ghana (FORIG): No words can describe the immense contributions of researchers at FORIG towards my dissertation research. Foremost, I would like to thank Dr. Victor Agyeman the Director of Forest Research Institute of Ghana (FORIG) for taking the time to meet with me during my predissertation fieldwork to discuss forestry in Ghana and possible areas of dissertation research. I am particularly grateful to Dr. Agyeman for introducing me to Dr. Earnest Foli a Senior Research Scientists who later served as my primary liaison and support person at FORIG. Besides connecting me to other vital resource persons in FORIG, Dr. Foli together with Dr. Beatrice Obiri-Darko reviewed my entire dissertation proposal and survey instruments for face and content validity, and for this, I am truly grateful. I would also like to thank Dr. Emmanuel Opuni-Frimpong also a Research Scientist at FORIG for designating three of his research assistants to help me during my field data collection and also loaning my dissertation project a vehicle to be used in transporting enumerators to and from my research communities. Together, Dr. Foli and Dr. Opuni-Frimpong sent five research assistants to help with my field data collection. To Dr. Foli and Dr. Opuni-Frimpong and their FORIG research assistants: Mr. Lord Kofi Ameyaw, Ms. Sandra Acheampong Owusu, Ms. Esther Osei vii Agyeman, Mr. Samuel Boakye Baafi, and Mr. Adade Nana Yaw I say a big thank you for your invaluable contributions to this work. Ghana Forestry Services Department (FSD) – Sunyani District: I am very grateful to Mr. Dickson Sakyi Adjei (Sunyani Forest District Manager) for taking his time to meet and discuss forestry activities in his district and the implementation of Ghana’s pilot Modified Taungya System (MTS) within the district. During my predissertation work, Mr. Dickson designated one of his Technical Officers and Range Supervisor Mr. Joseph Aggrey to take me around to each of the ten MTS pilot communities surrounding the Yaya Forest Reserve. I owe Mr. Aggrey a debt of gratitude for the adventure ride to each of the ten Yaya communities and for introducing me to the community leaders and chiefs whose support would become invaluable to my community entry and the successful interactions with community members during data collection. I am also grateful to Mr. Sakyi Adjei for granting permission for two National Service Personnel under his supervision; Mr. Pius Babuna and Ms. Anabertha Owusu-Bempah (both recent graduates of KNUST) to assist me with my fieldwork. Both Mr. Babuna and Ms. Owusu- Bempah participated diligently in all aspects of my fieldwork including household surveys/interviews, GIS community mapping exercises, and data entry and for this, I am truly grateful. I am also grateful to the following Sunyani District FSD Plantation/Range Supervisors: Madam Ruth N. Gyapong, Mr. Mohammed Issaka, Mr. Paul Sowah, and Mr. Holiday for taking their time to review my data collection instruments and also provide inputs into plantation development activities within the Sunyani District. Ghana Forestry Services Department (FSD) - Kumasi: I owe a debt of gratitude to Mr. Lawrence Akpalu the GIS Specialist at FSD Kumasi District office for sharing with me several base maps of forests in Ghana. viii University of Ghana: I am highly indebted to Mr. Prosper Evadzi, a GIS Specialist and Teaching Assistant at the University of Ghana’s Geography Department for taking me through a one-month crash course on GIS and Remote Sensing. Mr. Evadzi thought me how to use ArcMap to analyze spatial data and create maps and also how to use ENVI 4.7 to analyze remotely sensed imagery. All the maps included in this dissertation were the result of training and careful guidance provided by Mr. Evadzi and for this, I am truly grateful. Kwame Nkrumah University of Science and Technology (KNUST) Department of Forest Resources - Sunyani: I would also like to extend my sincere appreciation to Mr. Eric Gyebi a final year KNUST Forest Resource student who printed all my field survey instruments. On several occasions, Mr. Gyebi had to work through the night to make copies of revised survey instruments ready for pick up as early as 6 AM and for this, I am truly grateful. I would also like to thank Mr. Samuel Sky for his help with data collection and entry. Participatory Forest Resource Management Project in the Transition Zone (PAFORM): My dissertation fieldwork benefited immensely from the field expertise of nine PAFORM Community Facilitators who until March 2009 supported the PAFORM project in Nsemre Forest Reserve Communities. I owe a debt of gratitude to Mr. Samuel Akurugu, Mr. Mohammed Salam, Mr. Prince Henneh, Mr. Alfred Sangpour, Ms. Veronica Twumwaa, Mr. John Laar, Mr. Ebenezer Kumih, Mr. Dapilah Tarcisius, and Mr. Philip Kwaku Neigbija who assisted my dissertation fieldwork as enumerators. While all nine Community Facilitators assisted with household surveys, I would like to single out Mr. Akurugu, Mr. Salam, and Mr. Henneh for special recognition because these three gentlemen in addition to conducting focus group interviews also had the task of working with community leaders to mobilize households ix the evening before each day’s field exercise. No words can describe the debt of gratitude I owe to Mr. Akurugu, Mr. Salam, and Mr. Henneh. Other enumerators: I would like to thank Mr. Tingbani Abdul-Razak a final year student of Tamale Polytechnic for assisting with field data collection. Forest Community Leaders: I would like to extend my sincere gratitude to Yaya MTS Community leaders. These MTS leaders not only mobilized both MTS and Non-MTS members for briefing before the arrival of the survey team to each community but also encouraged households to participate in the study. Specifically, I would like to thank Mr. J.K Tawiah (Asuakwaa community), Mr. Fourjour and Mr. Yaw Sumana (Sawiah community), Mr. Adjei (Ahyiem), Mr. Ati Mohamed and Mr. Antoni Kusi (Ayigbekrom), Mr. Papa Ekow (Konsua), Mr. Yaw Annor and Mr. Ati Boateng (Amangoase), Mr. Earnest Darko and Mr. Bob (Mallamkrom), Mr. Leonard and Mr. Nana Say (Abrefakrom), Mr. Joseph Marfo and Mr. David Sarfo (Buoku), Mr. Emmanuel Kyeremeh and Mr. Kofi Appiah (Amoakrom). I also owe a debt of gratitude to all the Chiefs and Queen Mothers as well the Elders in all the 19 communities surveyed as part of this dissertation research. My dissertation would not have been successful if not for the wholehearted support of all the above-mentioned community leaders. Friends: I am extremely grateful to all my friends including Andrew Muganga, Felix Yeboah, Rohit Jindhal, Joshua Ariga, Luis Flores, Keith Oberg, and others who besides providing an avenue to escape the stress of grueling course work but also took particular interest in my research and personal development. Family: No amount of words can describe the depth of gratitude I owe my parents Mr. Samuel Adovor (a.k.a Charley) and Mrs. Grace Adovor (a.k.a Sister) for the solid upbringing x and foundation they provided for me to build my life on. Besides their moral and financial support throughout my education, I am particularly lucky to be born to parents who understood the fundamental principles that all hands were not created equal and that my siblings and I were endowed with different abilities that required customized nurturing. In matters of education, my parents’ doctrine has always been to create the enabling environment for individual growth and success through perseverance. I am a beneficiary of second, third, and sometimes fourth attempts at success and parents who refuse to throw in the towel too soon. My parents have always thought me that while success may be rewarding, there are equally important life-lessons to be gained from failures. I endeavor to honor my parents by always recognizing the disparate strengths in humans and treat others with similar compassion. I would like to thank all my siblings (Sister Woetsa, Ehui, Fo Yaw, Fo Dzidzinyor, Atsu, Elenu, Ekeke, Nyuiewodze, Dan, Amenyedu, Albert, and Alex) for their encouragements and support throughout my doctoral study. I am particularly grateful to my younger brother Albert for his absolute dedication to my doctoral fieldwork. Albert who had recently completed his undergraduate study at KNUST, joined my dissertation field crew from the start of my fieldwork and stayed with me until the end, assisting with household surveys, GIS community mapping, and data entry, and for this, I am truly grateful. I am highly indebted to my wife Clarice for her unwavering devotion to me and our children Volta and Qwekqem. To my daughters Volta and Qwekqem, and all my siblings' children some of whom would call me Torga (Big Father), Tordey (Small Father), or Wofa, I want you to look upon this work as your father’s attempt to help improve the livelihoods of a people living in dire circumstances largely through no fault of their own. Changing the plight of the world’s poor often time requires a force similar to that needed to move mountains. However, xi I want you to always remember that moving a mountain requires intergenerational effort and all I have done with this work is to move a tiny pebble. The world is counting on you and future generations to play your part. xii TABLE OF CONTENTS LIST OF TABLES ...................................................................................................................... xx LIST OF FIGURES ................................................................................................................ xxvii CHAPTER 1: INTRODUCTION ................................................................................................ 1 1.1 Introduction to the Problem................................................................................................... 2 1.2 Background of the Study ....................................................................................................... 3 1.3 Problem Statement ................................................................................................................ 8 1.4 Purpose of the Study ........................................................................................................... 11 1.5 Method used in the Study .................................................................................................... 12 1.6 Significance of the Research ............................................................................................... 14 1.7 Layout of the Dissertation ................................................................................................... 15 CHAPTER 2: FORESTRY IN GHANA................................................................................... 17 2.1 Introduction ......................................................................................................................... 18 2.2 Literature ............................................................................................................................. 18 2.3 The Crisis of Tropical Deforestation ................................................................................... 19 2.4 Historical Background on Ghana ........................................................................................ 23 2.5 Agro Climatic Conditions in Ghana .................................................................................... 28 2.6 Ghana’s Agro-ecological Zone (AEZ) and High Forest Zones (HFZ) ............................... 29 2.7 Traditional System of Governance and Land Tenure ......................................................... 34 2.8 Historical Background on Ghana’s Forest Policy ............................................................... 36 2.8.1 Crown Land Ordinance of 1894 and the Public Land Bill of 1897 .............................. 39 2.8.2 The Concession Ordinance of 1900 .............................................................................. 40 2.8.3 Timber Protection Ordinance of 1907 .......................................................................... 42 2.8.4 Forest Ordinance of 1911 ............................................................................................. 43 2.8.5 The Forest Ordinance of 1927 ...................................................................................... 48 2.8.6 The 1948 Forest Policy ................................................................................................. 50 2.8.7 Aftermath of the 1948 Forest Policy ............................................................................ 55 2.9 Ghana’s Economy after Independence (1957) .................................................................... 57 xiii 2.9.1 Ghana’s Cocoa Economy (1885-2008) ........................................................................ 59 2.10 Structural Adjustment Programs (1983-2000) .................................................................. 61 2.10.1 Impact of SAP Loans on Adjusting Countries ........................................................... 63 2.10.2 Ghana’s Economic Recovery Program (ERP) and SAP Adoption in 1983 ............... 65 2.10.3 Changes in Ghana’s forest cover under SAP era (1983-2003) ................................... 68 2.10.4 Ghana’s forest-cover after SAP .................................................................................. 70 2.11 Secure Tenure Rights and Community involvement in Forest Management ................... 70 2.12 Tax Reform and Timber Certification ............................................................................... 72 2.13 Role of the Media in Forest Management ......................................................................... 72 APPENDICES ........................................................................................................................... 74 APPENDIX A: Agro Climatic Conditions in Ghana ............................................................ 75 APPENDIX B: Logging Activity in Yaya Forest Reserve.................................................... 81 CHAPTER 3: HOUSEHOLD LIVELIHOODS APPROACH ............................................... 82 3.1 Introduction ......................................................................................................................... 83 3.2 Literature ............................................................................................................................. 84 3.3 Livelihood/Asset-based Terminology ................................................................................. 85 3.3.1 Livelihood ..................................................................................................................... 85 3.3.2 Livelihood Assets ......................................................................................................... 86 3.3.3 Vulnerability Context ................................................................................................... 87 3.3.4 Poverty and Sustainable Livelihood ............................................................................. 90 3.4 Livelihood Capital/Assets ................................................................................................... 91 3.4.1 Financial Capital ........................................................................................................... 91 3.4.2 Social Capital ................................................................................................................ 93 3.4.3 Physical Capital ............................................................................................................ 94 3.4.4 Natural Capital .............................................................................................................. 96 3.4.5 Human Capital .............................................................................................................. 96 3.5 Asset-Based Approaches ..................................................................................................... 98 3.5.1 Historical Background .................................................................................................. 98 3.5.2 Asset-Based Methods ................................................................................................. 102 CHAPTER 4: METHODOLOGY .......................................................................................... 111 4.1 Introduction ....................................................................................................................... 112 4.2 Literature Review .............................................................................................................. 112 xiv 4.3 Pre-dissertation Field Reconnaissance .............................................................................. 114 4.4 Research Community and GIS Community Mapping ...................................................... 115 4.4.1 Forest Reserves in Brong Ahafo Region .................................................................... 115 4.5 Research Communities and GIS Community Mapping Exercise ..................................... 117 4.5.1 GIS Data Collection Protocol ..................................................................................... 117 4.5.2 Data Retrieval from GPS Unit .................................................................................... 118 4.5.3 Image Processing ........................................................................................................ 119 4.6 Background Information on Yaya Reserve Communities ................................................ 120 4.6.1 Household Information from Yaya Community Maps ............................................... 123 4.7 Background Information on Nsemre Forest Reserve ........................................................ 124 4.7.1 Household Information from Nsemre Community Maps ........................................... 127 4.8 Background Information on Sawsaw Forest Reserve communities ................................. 128 4.8.1 Household Information from Sawsaw Community Maps .......................................... 131 4.9 Survey Instrument Design and UCHRIS Approval .......................................................... 132 4.9.1 Instrument Design....................................................................................................... 132 4.9.2 Obtaining IRB UCHRIS Approval ............................................................................. 135 4.9.3 Content Validity of Survey Instruments ..................................................................... 136 4.9.4 Yaya Community and Household Selection ............................................................... 137 4.9.5 Selection of Nsemre (Non-MTS) Communities and Households .............................. 142 4.9.6 Selection of Sawsaw (Non-MTS) communities and households ............................... 146 4.9.7 Focus Group Interviews .............................................................................................. 147 4.9.8 Managing the Survey and Field Data Collection........................................................ 148 4.9.9 Data Analysis .............................................................................................................. 150 APPENDICES ......................................................................................................................... 154 APPENDIX A: Initial Application to MSU Internal Review Board (IRB) ........................ 155 APPENDIX B: IRB Approval Letter .................................................................................. 163 APPENDIX C: Letter of Collaboration to the Ghana Forestry Research Institute ............. 164 APPENDIX D: Livelihoods Monitoring Tool (LMT) ........................................................ 165 APPENDIX E: Community Profile Instrument................................................................... 182 APPENDIX F: Household Livelihood Strategies ............................................................... 184 APPENDIX G: Institutional Profiles ................................................................................... 188 APPENDIX H: Schedule of MTS Benefit Sharing Agreement .......................................... 193 APPENDIX I: Demographic Information of MTS Communities ....................................... 196 APPENDIX J: Descriptive Statistics from Community Mapping Exercise ....................... 198 xv CHAPTER 5: LANDCOVER CHANGE DETECTION IN YAYA, SAWSAW AND NSEMRE FOREST RESERVES ............................................................................................ 204 Abstract ................................................................................................................................... 205 5.1 Landcover change in Yaya, Nsemre and Sawsaw Forest Reserves .................................. 205 5.1.1 Image selection protocol ............................................................................................. 208 5.1.2 Image acquisition ........................................................................................................ 208 5.1.3 Image processing protocol .......................................................................................... 209 5.1.4 Chapter organization ................................................................................................... 209 5.2 Landcover changes in all three reserves ............................................................................ 210 5.2.1 Landcover changes in all three reserves from 1990 to 2000 ...................................... 210 5.2.2 Landcover changes in all three reserves from 2000 to 2012 ...................................... 215 5.3 Land-cover changes in Yaya Reserve ............................................................................... 219 5.3.1 Landcover change in Yaya Reserve between 1990 and 2000 .................................... 220 5.3.2 Landcover Change in Yaya Reserve between 2000 and 2012 ................................... 223 5.4 Land-cover changes in Nsemre Reserve ........................................................................... 226 5.4.1 Land-cover changes in Nsemre Reserve between 1990 and 2000 ............................. 226 5.4.2 Landcover changes in Nsemre Reserve between 2000 and 2012 ............................... 229 5.5 Landcover changes in Sawsaw Reserve ............................................................................ 232 5.5.1 Land-cover changes in Sawsaw Reserve between 1990 and 2000 ............................. 232 5.5.2 Land-cover changes in Sawsaw Reserve between 2000 and 2012 ............................. 234 5.6 Summary of findings ......................................................................................................... 236 5.6.1 Annual changes in forest cover in Yaya before and after MTS ................................. 236 5.6.2 Annual changes in forest cover in Nsemre before and after MTS ............................. 238 5.6.3 Annual changes in forest cover in Sawsaw before and after MTS ............................. 241 5.7 Conclusions and Implications for forest policy................................................................. 243 5.7.1 Monoculture Forest Plantations under MTS .............................................................. 244 5.7.2 Polyculture Forest Plantations under MTS ................................................................. 247 5.8 Recommended areas for further research .......................................................................... 248 CHAPTER 6: COMMUNITY AND HOUSEHOLD PLACEMENT INTO GHANA’S MTS PROGRAM ............................................................................................................................... 251 Part I: Impact of Socio-Economic Factors on Community and Household Placement into Ghana’s Modified Taungya System ........................................................................................ 252 Abstract ................................................................................................................................... 252 6.1 Binary Probit Analysis ...................................................................................................... 253 xvi 6.1.1 Functional form for Probit Model .............................................................................. 254 6.1.2 Probit results from household livelihood survey ........................................................ 256 6.1.3 Hypothesized effects of BPM variables on MTS community selection ..................... 267 6.1.4 Factors that significantly influence MTS community selection ................................. 271 6.1.5 Other factors that influence the probability of MTS community selection ................ 279 6.1.6 Power of BPM to predict MTS community placement .............................................. 289 6.1.7 Probit Results for MTS household placement ............................................................ 291 6.1.8 Power of BPM to Predict MTS Community Placement ............................................. 296 6.1.9 Summary, conclusion, and policy implications .......................................................... 299 Part II: Impact of Spatiotemporal Factors on Community and Household Placement into Ghana’s Modified Taungya System ........................................................................................ 301 Abstract ................................................................................................................................... 301 6.2 Binary Probit Models Generated from Community Maps ................................................ 302 6.2.1 Binomial Probit Analysis Results- Community Participation in MTS....................... 310 6.2.2 Binomial Probit Analysis Results- Household Participation in MTS ........................ 321 6.2.3 Summary, conclusions, and policy implications ........................................................ 330 Part III: Patterns of Local Spatial Autocorrelation that Influenced Household Placement into Ghana’s Modified Taungya System – Insight from Ayigbekrom Community ....................... 333 Abstract ................................................................................................................................... 333 6.3 Data Generation and Analysis ........................................................................................... 334 6.3.1 Housing Infrastructure Index (HII) ............................................................................. 337 6.3.2 Household Livestock Production Index (HLI) ........................................................... 339 6.3.3 Household Proximity Index (HPI) .............................................................................. 341 6.3.4 Household Male Index (HMI) .................................................................................... 344 6.3.5 Results ........................................................................................................................ 344 6.3.6 Summary, conclusions, and Policy Implications ........................................................ 361 APPENDICES ......................................................................................................................... 363 APPENDIX A: Yaya Forest Reserve Community Maps .................................................... 364 APPENDIX B: Nsemre Forest Reserve Community Maps ................................................ 491 APPENDIX C: Sawsaw Forest Reserve Community Maps ................................................ 529 APPENDIX D: Population, Religion and Poverty in the Northern Ghana ......................... 576 CHAPTER 7: BETWEEN AND WITHIN GROUP CHANGES IN HOUSEHOLD LIVELIHOOD ASSETS IN YAYA, NSEMRE AND SAWSAW FOREST RESERVE COMMUNITIES....................................................................................................................... 583 xvii Abstract ................................................................................................................................... 584 Introduction ............................................................................................................................. 585 7.1 Financial Capital ............................................................................................................... 588 7.1.1 Two Sample T-test of between and within group differences in 1999 and 2009 ........ 595 7.1.2 Summary Results ........................................................................................................ 602 7.2 Human Capital................................................................................................................... 607 7.2.1 Two Sample T-test of between and within group differences in 1999 and 2009 ........ 618 7.2.2 Summary Results ........................................................................................................ 632 7.3 Physical Capital ................................................................................................................. 634 7.3.1 Two Sample T-test of between and within group differences in 1999 and 2009 ........ 641 7.3.2 Summary Results ........................................................................................................ 650 7.4 Natural Capital .................................................................................................................. 653 7.4.1 Two Sample T-test of between and within group differences in 1999 and 2009 ........ 663 7.4.2 Summary Results ........................................................................................................ 677 7.5 Social Capital .................................................................................................................... 680 7.5.1 Two Sample T-test of between and within group differences in 1999 and 2009 ........ 685 7.5.2 Summary Results ........................................................................................................ 695 7.6 Livelihood Asset Pentagons .............................................................................................. 697 7.6.1 Comparative Analysis of Household Livelihood Indexes/Asset (HLIs) .................... 697 7.6.2 Interpretation of Livelihood Pentagons ...................................................................... 698 7.6.3 Changes in Household Livelihood Indexes for all Groups Between 1999 and 2009 . 699 7.6.4 Within Group Changes in MTS and non-MTS Asset Indexes between 1999 and 2009 ............................................................................................................................................. 702 7.6.5 Between Group Changes in MTS, non-MTS and NsemSaw HLIs for 1999 and 2009 ............................................................................................................................................. 704 7.6.6 Within Group Changes in MTS and NsemSaw Asset Indexes between 1999 and 2009 ............................................................................................................................................. 706 7.6.7 Between Group Changes in MTS and NsemSaw HLIs for 1999 and 2009 ............... 708 APPENDICES ......................................................................................................................... 711 APPENDIX A: Descriptive Analysis of Financial Capital Indexes .................................... 712 APPENDIX B: Descriptive Analysis of Human Capital Indexes ....................................... 747 APPENDIX C: Descriptive Analysis of Physical Capital Indexes ..................................... 780 APPENDIX D: Descriptive Analysis of Natural Capital Indexes ....................................... 806 APPENDIX E: Descriptive Analysis of Social Capital Indexes ......................................... 834 xviii CHAPTER 8: USING DIFFERENCE-IN-DIFFERENCE (DID) METHODOLOGY TO ISOLATE DIRECT AND SPILLOVER EFFECTS OF GHANA’S MTS PROJECT ON PARTICIPANT AND NON-PARTICIPANT PROJECT COMMUNITIES ..................... 862 Abstract ................................................................................................................................... 863 8.1 Difference in Difference (DID) ......................................................................................... 864 8.1.1 Background information on data used for DID .......................................................... 865 8.1.2 Definition of DID Estimators for Household Livelihood Indexes ............................. 870 8.1.3 Difference in Difference Estimates of Combined Livelihoods Indexes ..................... 872 8.2 Direct and Spillover MTS Project Effect on CHHFCI ..................................................... 874 8.2.1 Direct and Spillover MTS Project Effect on CHCI1 and CHCI2 ............................... 877 8.2.2 Direct and Spillover MTS Project Effect on CHPCI1 and CHPCI2 .......................... 880 8.2.3 Direct and Spillover MTS Project Effect on AHHSCI ............................................... 883 8.2.4 Direct and Spillover MTS Project Effect on CHHNCI1 and CHHNCI2 ................... 885 8.3 Summary, conclusions, and policy implications ............................................................... 889 BIBLIOGRAPHY ..................................................................................................................... 890 xix LIST OF TABLES Table 2.8.1: Timber (Log) Exports in Ft3 (x1000): 1937-1948 .................................................... 54 Table 4.6.1: Yaya Reserve Household Demographic Information ............................................. 123 Table 4.6.2: Yaya Reserve Housing Infrastructure Per Capita ................................................... 123 Table 4.6.3: Yaya Reserve Community Livestock Ownership................................................... 124 Table 4.7.1: Nsemre Reserve Household Demographic Information ......................................... 127 Table 4.7.2: Nsemre Reserve Housing Infrastructure Per Capita ............................................... 127 Table 4.7.3: Nsemre Reserve Community Livestock Ownership ............................................... 128 Table 4.8.1: Sawsaw Reserve Household Demographic Information ........................................ 131 Table 4.8.2: Sawsaw Reserve Housing Infrastructure Per Capita .............................................. 131 Table 4.8.3: Sawsaw Reserve Community Livestock Ownership .............................................. 132 Table 4.9.1: Sample sizes of MTS and non-MTS households in Yaya ...................................... 141 Table 4.9.2: Sample sizes of Nsemre households ....................................................................... 146 Table 4.9.3: Sample sizes of Sawsaw households ...................................................................... 147 Table 4.9.4: Number of households selected in MTS communities ........................................... 196 Table 4.9.5: Gender, Marital, Residential Status and education of selected MTS Participants . 197 Table 4.9.6: Average Age, Household size and Year’s resident in the Brong Ahafo Region .... 197 Table 4.9.7: Population Information from Field GPS Exercise .................................................. 198 Table 4.9.8: Population Densities ............................................................................................... 199 Table 4.9.9: Household Livestock Ownership Information........................................................ 200 Table 4.9.10: Type of Housing Infrastructure ............................................................................ 201 Table 4.9.11: Proximity of Household to Communal Asset ....................................................... 202 Table 5.2.1: Land-cover change statistics in pixels and Km2 for all reserves from 1990 to 2000 ..................................................................................................................................................... 213 Table 5.2.2: Land-cover change statistics in pixels and Km2 for all reserves from 2000 to 2012 ..................................................................................................................................................... 218 Table 5.3.1: Yaya Land Cover Change Statistics in pixels and Km2 from 1990 to 2000........... 222 Table 5.3.2: Yaya Land Cover Change Statistics in pixels and Km2 from 2000 to 2012........... 224 Table 5.4.1: Nsemre Land Cover Change Statistics in pixels and Km2 from 1990 to 2000....... 228 Table 5.4.2: Nsemre Land Cover Change Statistics in pixels and Km2 from 2000 to 2012....... 230 Table 5.5.1: Sawsaw Land Cover Change Statistics in pixels and Km2 from 1990 to 2000 ...... 233 Table 5.5.2: Sawsaw Land Cover Change Statistics in pixels and Km2 from 2000 to 2012 ...... 235 xx Table 6.1.1: Variable Definitions ................................................................................................ 257 Table 6.1.2: Hypothesized Effects of Explanatory Variables on Selection ................................ 259 Table 6.1.3: BPM Analysis of Factors that Influence MTS Community Selection -1999 ......... 269 Table 6.1.4: Classification of Predicted and Actual MTS Community Placement .................... 290 Table 6.1.5: BPM Analysis of Factors that Influence Household Selection into MTS-1999 ..... 293 Table 6.1.6: Classification of Predicted and Actual MTS Household Placement ...................... 298 Table 6.2.1: Household Variable Definitions ............................................................................. 304 Table 6.2.2: Hypothesized Effects of Explanatory Variables on Participation .......................... 306 Table 6.2.3: Impact of 21 Selected Livelihood Assets on MTS Community Participation........ 311 Table 6.2.4: Classification of Predicted and Actual MTS Community Participation ................. 320 Table 6.2.5: Impact of 21 Selected Livelihood Assets on MTS Household Participation ......... 323 Table 6.2.6: Classification of Predicted and Actual MTS Household Participation .................. 329 Table 6.2.7: Impact of 21 Selected Livelihood Assets on MTS Community Participation........ 330 Table 6.2.8: Impact of 21 Selected Livelihood Assets on MTS Household Participation ......... 331 Table 6.3.1: Housing Infrastructure Index .................................................................................. 339 Table 6.3.2: Household Variable Definitions ............................................................................. 341 Table 6.3.3: Frequency Distribution of HII ................................................................................ 345 Table 6.3.4: Frequency Distribution of HLI ............................................................................... 348 Table 6.3.5: Frequency Distribution of HPI (Meters)................................................................. 351 Table 6.3.6: Frequency Distribution of HMI .............................................................................. 354 Table 6.3.7: Variables included in BPM ..................................................................................... 357 Table 6.3.8: Livelihood Indexes used in BPM ........................................................................... 358 Table 6.3.9: Hypothesized Effects of Livelihood Indexes .......................................................... 359 Table 6.3.10: Effects of Livelihood Indexes on Household Selection ....................................... 361 Table 7.1.1: Definition of Household Financial Capital Indexes ............................................... 589 Table 7.1.2: T-Test for Estimated Differences Between and Within Groups ............................. 603 Table 7.1.3: Summary of T-Test for Estimated Differences Between and Within Groups ........ 606 Table 7.2.1: Definition of Household Human Capital Indexes .................................................. 608 Table 7.2.2: T-Test for Estimated Differences in Human Capital Indexes Between and Within Groups ......................................................................................................................................... 627 Table 7.2.3: Summary of T-Test for Estimated Differences Between and Within Groups ........ 633 Table 7.3.1: Definition of Household Physical Capital Indexes ................................................. 635 Table 7.3.2: T-Test for Estimated Differences in Physical Capital Indexes Between and Within Groups ......................................................................................................................................... 647 Table 7.3.3: Summary of T-Test for Estimated Differences Between and Within Groups ........ 652 xxi Table 7.4.1: Definition of Household Natural Capital Indexes .................................................. 654 Table 7.4.2: T-Test for Estimated Differences in Natural Capital Indexes Between and Within Groups ......................................................................................................................................... 672 Table 7.4.3: Summary of T-Test for Estimated Differences Between and Within Groups ........ 679 Table 7.5.1: Definition of Household Social Capital Indexes .................................................... 681 Table 7.5.2: T-Test for Estimated Differences in Natural Capital Indexes Between and Within Groups ......................................................................................................................................... 691 Table 7.5.3: T-Test for Estimated Differences Between and Within Groups ............................. 696 Table 7.5.4: Primary Sources of Household Income - Frequency .............................................. 713 Table 7.5.5: Primary Sources of Household Income - Descriptive ............................................ 714 Table 7.5.6: Secondary Sources of Household Income - Frequency .......................................... 715 Table 7.5.7: Secondary Sources of Household Income - Descriptive ........................................ 717 Table 7.5.8: Primary Sources of Household Expenditure - Frequency ...................................... 719 Table 7.5.9: Primary Sources of Household Expenditure - Descriptive ..................................... 720 Table 7.5.10: Secondary Household Expenditure - Frequency .................................................. 721 Table 7.5.11: Secondary Household Expenditure – Descriptive ................................................ 722 Table 7.5.12: Household Savings and Loan Account Ownership - Frequency .......................... 723 Table 7.5.13: Household Savings and Loan Account Ownership - Descriptive ........................ 724 Table 7.5.14: Frequency of Household Savings and Loan Activities - Frequency .................... 726 Table 7.5.15: Frequency of Household Savings and Loan Activities - Descriptive ................... 727 Table 7.5.16: Total Household Savings and Loan Amount Taken - Frequency......................... 728 Table 7.5.17: Total Household Savings and Loan Amount Taken - Descriptive ....................... 729 Table 7.5.18: Annual Household Income - Frequency ............................................................... 730 Table 7.5.19: Annual Household Income - Descriptive ............................................................. 730 Table 7.5.20: Annual Household s Income Sufficiency - Frequency ......................................... 731 Table 7.5.21: Annual Household Income and Sufficiency - Descriptive ................................... 732 Table 7.5.22: Number of Village Stores or Kiosks that Sell Consumer Goods - Frequency ..... 734 Table 7.5.23: Number of Village Stores or Kiosks that Sell Consumer Goods - Descriptive .... 734 Table 7.5.24: Average Price of Items Sold in Village Store- Frequency ................................... 735 Table 7.5.25: Average Price of Items Sold in Village Store- Frequency - Descriptive .............. 735 Table 7.5.26: Household Expenditure in Village Store - Frequency .......................................... 737 Table 7.5.27: Household Expenditure in Village Store - Descriptive ........................................ 738 Table 7.5.28: Household Expenditure in Village Store – Frequency ......................................... 739 Table 7.5.29: Household Expenditure in Village Store - Descriptive ........................................ 739 Table 7.5.30: Sale of Household Possessions - Frequency ......................................................... 741 xxii Table 7.5.31: Sale of Household Possessions - Descriptive ....................................................... 742 Table 7.5.32: General Trend in Accumulation of HH Possessions - Frequency ........................ 743 Table 7.5.33: General Trend in Accumulation of HH Possessions - Descriptive ....................... 744 Table 7.5.34: English and Local Language Literacy .................................................................. 747 Table 7.5.35: Primary, Secondary and College Level Education - Frequency ........................... 748 Table 7.5.36: Education and Literacy - Descriptive ................................................................... 749 Table 7.5.37: Household Migration out of the Village - Frequency ........................................... 751 Table 7.5.38: Household Migration out of the Village - Descriptive ......................................... 752 Table 7.5.39: HDD-Frequency of Consumption of Major Staples - Frequency ......................... 753 Table 7.5.40: HDD-Frequency of Consumption of Major Staples - Descriptive ....................... 754 Table 7.5.41: HDD- Frequency of Consumption of Vegetables and Fruits - Frequency ........... 755 Table 7.5.42: HDD- Frequency of Consumption of Vegetables and Fruits - Descriptive .......... 756 Table 7.5.43: HDD- Frequency of Consumption of Animal and Plant Proteins - Frequency .... 758 Table 7.5.44: HDD- Frequency of Consumption of Animal and Plant Proteins - Descriptive .. 759 Table 7.5.45: HDD- Frequency of Consumption of Vegetables and Fruits - Frequency ........... 761 Table 7.5.46: HDD- Frequency of Consumption of Other Food Items - Descriptive ................ 762 Table 7.5.47: HDD-Trend of Consumption of Major Staple Foods - Frequency ....................... 764 Table 7.5.48: HDD-Trend of Consumption of Major Staple Foods - Descriptive ..................... 765 Table 7.5.49: HDD- Trend of Consumption of Vegetables and Fruits- Frequency ................... 766 Table 7.5.50: HDD- Trend of Consumption of Vegetables and Fruits- Descriptive .................. 767 Table 7.5.51: HDD- Trend of Consumption of Plant and Animal Proteins - Frequency ........... 769 Table 7.5.52: HDD- Trend of Consumption of Plant and Animal Proteins - Descriptive .......... 770 Table 7.5.53: HDD- Trend of Consumption of Other Foods - Frequency ................................. 772 Table 7.5.54: HDD- Trend of Consumption of Other Foods – Descriptive ............................... 773 Table 7.5.55: Frequency of Disease and Illnesses within the Household - Frequency .............. 775 Table 7.5.56: Frequency of Disease and Illnesses within the Household - Descriptive ............. 776 Table 7.5.57: Frequency of Mortality cases within the Household ............................................ 777 Table 7.5.58: Frequency of Mortality cases within the Household - Descriptive ...................... 778 Table 7.5.59: Home Ownership Status – Frequency .................................................................. 780 Table 7.5.60: Home Ownership Status – Descriptive ................................................................. 780 Table 7.5.61: Average Number of Person per Room - Frequency ............................................. 781 Table 7.5.62: Average Number of Person per Room - Descriptive ............................................ 781 Table 7.5.63: Physical Construction of the House - Frequency ................................................. 782 Table 7.5.64: Physical Construction of the House - Descriptive ................................................ 783 xxiii Table 7.5.65: Type of Roofing Material/Construction - Frequency ........................................... 783 Table 7.5.66: Type of Roofing Material/Construction - Descriptive .......................................... 784 Table 7.5.67: Source of Energy for Lighting the House - Frequency ......................................... 785 Table 7.5.68: Source of Energy for Lighting the House - Descriptive ....................................... 785 Table 7.5.69: Source of Drinking Water in the House - Frequency ........................................... 786 Table 7.5.70: Source of Drinking Water in the House - Descriptive .......................................... 786 Table 7.5.71: Type of Kitchen in the House - Frequency ........................................................... 787 Table 7.5.72: Type of Kitchen in the House - Descriptive ......................................................... 788 Table 7.5.73: Type of Bathroom in the House - Frequency ....................................................... 788 Table 7.5.74: Type of Bathroom in the House - Descriptive ...................................................... 789 Table 7.5.75: Type of Toilet Facility in the House - Frequency ................................................. 789 Table 7.5.76: Type of Toilet Facility in the House - Descriptive ............................................... 790 Table 7.5.77: Disposal Site for Household Liquid Waste - Frequency ...................................... 791 Table 7.5.78: Disposal Site for Household Liquid Waste - Descriptive ..................................... 791 Table 7.5.79: Disposal Site for Household Solid Waste - Frequency ........................................ 792 Table 7.5.80: Disposal Site for Household Solid Waste - Descriptive ....................................... 792 Table 7.5.81: Proximity of Household to Different Markets- Frequency................................... 793 Table 7.5.82: Proximity of Household to Different Markets - Descriptive ................................ 794 Table 7.5.83: Proximity of Household to Different Health Facilities - Frequency .................... 795 Table 7.5.84: Proximity of Household to Different Health Facilities - Descriptive ................... 795 Table 7.5.85: Proximity of Household to any Postal Service - Frequency ................................. 796 Table 7.5.86: Proximity of Household to any Postal Service - Descriptive ............................... 796 Table 7.5.87: Proximity of Household to Transportation Services - Frequency ........................ 797 Table 7.5.88: Proximity of Household to Transportation Services - Descriptive ....................... 797 Table 7.5.89: Cell Phone Ownership .......................................................................................... 798 Table 7.5.90: Cell Phone Network/Services ............................................................................... 799 Table 7.5.91: Cell Phone Network Reliability ............................................................................ 800 Table 7.5.92: Basic Household Possessions One - Frequency ................................................... 801 Table 7.5.93: Basic Household Possessions Two - Frequency ................................................... 801 Table 7.5.94: Basic Household Possessions - Descriptive ......................................................... 802 Table 7.5.95: Luxury Household Possessions - Frequency ........................................................ 803 Table 7.5.96: Luxury Household Possessions - Descriptive ....................................................... 804 Table 7.5.97: Cropland Ownership ............................................................................................. 806 Table 7.5.98: Cropland Ownership ............................................................................................. 807 xxiv Table 7.5.99: Crop Production .................................................................................................... 808 Table 7.5.100: Crop Production .................................................................................................. 809 Table 7.5.101: Trends in Quantity of Specific Crops Produced ................................................. 810 Table 7.5.102: Trends in Quantity of Specific Crops Produced ................................................. 811 Table 7.5.103: Subsistence Oriented Crop Production ............................................................... 812 Table 7.5.104: Subsistence Oriented Crop Production ............................................................... 813 Table 7.5.105: Market Oriented Crop Production ...................................................................... 814 Table 7.5.106: Market Oriented Crop Production ...................................................................... 815 Table 7.5.107: Livestock Production (Ownership) .................................................................... 816 Table 7.5.108: Livestock Production (Ownership) ..................................................................... 817 Table 7.5.109: Livestock Production (Quantity) ........................................................................ 818 Table 7.5.110: Livestock Production (Quantity) ........................................................................ 819 Table 7.5.111: Trend in Livestock Production (Quantity) .......................................................... 820 Table 7.5.112: Trend in Livestock Production (Quantity) .......................................................... 821 Table 7.5.113: Subsistent Oriented Livestock Production.......................................................... 822 Table 7.5.114: Subsistent Oriented Livestock Production.......................................................... 823 Table 7.5.115: Subsistent Oriented Livestock Production.......................................................... 824 Table 7.5.116: Subsistent Oriented Livestock Production.......................................................... 825 Table 7.5.117: Soil Fertility and Fertilizer Application .............................................................. 826 Table 7.5.118: Soil Fertility and Fertilizer Application .............................................................. 826 Table 7.5.119: Trend in Fertilizer Application ........................................................................... 827 Table 7.5.120: Trend in Fertilizer Application ........................................................................... 828 Table 7.5.121: Harvest of NTFPs ............................................................................................... 829 Table 7.5.122: Harvest of NTFPs ............................................................................................... 829 Table 7.5.123: Trends in NTFP Harvest ..................................................................................... 830 Table 7.5.124: Trends in NTFP Harvest ..................................................................................... 831 Table 7.5.125: Time Spent Harvesting NTFPs ........................................................................... 832 Table 7.5.126: Time Spent Harvesting NTFPs ........................................................................... 833 Table 7.5.127: Number of Close Relatives in the Village - Frequency ...................................... 834 Table 7.5.128: Number of Close Relatives in the Village - Descriptive .................................... 835 Table 7.5.129: Help from Relatives in the Village - Frequency ................................................. 836 Table 7.5.130: Help from Relatives in the Village - Descriptive ............................................... 837 Table 7.5.131: Help from Relatives Outside the Village but in Brong Ahafo - Frequency ....... 838 Table 7.5.132: Help from Relatives Outside the Village but in Brong Ahafo - Descriptive ...... 839 xxv Table 7.5.133: Help from Relatives Outside Brong Ahafo but in Ghana - Frequency ............... 840 Table 7.5.134: Help from Relatives Outside Brong Ahafo but in Ghana - Descriptive ............. 841 Table 7.5.135: Help from Relatives Outside Ghana - Frequency ............................................... 842 Table 7.5.136: Help from Relatives Outside Ghana - Descriptive ............................................. 843 Table 7.5.137: Support from Government and Various NGOs - Frequency .............................. 844 Table 7.5.138: Support from Government and Various NGOs - Descriptive ............................ 845 Table 7.5.139: Support from Friends Inside or Outside the Village - Frequency ...................... 846 Table 7.5.140: Support from Friends Inside or Outside the Village - Descriptive ..................... 847 Table 7.5.141: Membership in Community Organizations/Associations - Frequency ............... 848 Table 7.5.142: Membership in Community Organizations/Associations - Descriptive ............. 849 Table 7.5.143: Attendance of Community Organizations/Associations Meetings - Frequency . 850 Table 7.5.144: Attendance of Community Organizations/Associations Meetings - Descriptive 851 Table 7.5.145: Membership in Religious Groups - Frequency ................................................... 852 Table 7.5.146: Membership in Religious Groups - Descriptive ................................................. 852 Table 7.5.147: Attendance of Religious Group Meetings - Frequency ...................................... 853 Table 7.5.148: Attendance of Religious Group Meetings - Descriptive ..................................... 853 Table 7.5.149: Support from Religious Groups - Frequency...................................................... 854 Table 7.5.150: Support from Religious Groups - Descriptive .................................................... 855 Table 7.5.151: Size of Household - Frequency........................................................................... 856 Table 7.5.152: Size of Household - Descriptive ......................................................................... 856 Table 7.5.153: Joint Household Activities - Frequency ............................................................. 857 Table 7.5.154: Joint Household Activities - Descriptive ............................................................ 858 Table 7.5.155: Short-term Seasonal Migration - Frequency ....................................................... 859 Table 7.5.156: Short-term Migration - Descriptive .................................................................... 859 Table 7.5.157: Long-term Seasonal Migration - Frequency ....................................................... 860 Table 7.5.158: Long-term Seasonal Migration - Frequency ....................................................... 861 Table 8.1.1: Definition of Household Aggregate Livelihood Indexes........................................ 868 Table 8.1.2: DID Estimates of Combined Livelihoods Indexes ................................................. 872 xxvi LIST OF FIGURES Figure 2.1.1: Changes in Forest Cover in Selected West-African Countries (1900-1990) .......... 21 Figure 2.1.2: Administrative Map of Ghana Highlighting Research Region ............................... 27 Figure 2.1.3: Agro-ecological Zones of Ghana ............................................................................ 31 Figure 2.1.4: Land Cover Situation Map of Ghana...................................................................... 32 Figure 2.1.5: Land Cover Situation Map of Brong Ahafo Region................................................ 33 Figure 2.1.6: Map of Ghana’s Reserves and National Parks....................................................... 46 Figure 2.1.7: Map of Brong Ahafo’s Forest Reserves and National Parks ................................. 47 Figure 2.1.8: Ghana’s Timber Exports (1938-1948) .................................................................... 51 Figure 2.1.9: Timber (Log) Exports in Ft3 (x1000): 1937-1948 .................................................. 55 Figure 2.1.10: Change in Ghana’s Forest Cover between 1948 and 1990 .................................. 57 Figure 2.1.11: Ghana’s Ranking and Percentage Shares of World Exports in 1950 .................. 58 Figure 2.1.12: Ghana’s cocoa outputs (1970-1983) .................................................................... 60 Figure 2.1.13: Ghana’s Mineral Outputs under SAP ................................................................... 66 Figure 2.1.14: Changes in Ghana’s Unreserved Forest Area (1948-1990) ................................. 68 Figure 2.2.1: January and February Average (mm) Rainfall Patterns in Ghana ........................ 75 Figure 2.2.2: March and April Average (mm) Rainfall Patterns in Ghana ................................. 76 Figure 2.2.3: May and June Average (mm) Rainfall Patterns in Ghana ..................................... 77 Figure 2.2.4: July and August Average (mm) Rainfall Patterns in Ghana .................................. 78 Figure 2.2.5: September and October Average (mm) Rainfall Patterns in Ghana ...................... 79 Figure 2.2.6: November and December Average (mm) Rainfall Patterns in Ghana ................... 80 Figure 2.3.1: Illegal chainsaw logging on a local farm in Yaya near Sewiah Community.......... 81 Figure 3.1.1: Sustainable Livelihood Approach (SLA) .............................................................. 105 Figure 3.1.2: Sustainability Livelihoods Framework ................................................................. 108 Figure 4.1.1: Maps of Brong Ahafo Showing the locations of Yaya, Nsemre and Sawsaw Forest Reserves ...................................................................................................................................... 116 Figure 4.1.2: Maps of Yaya Forest Reserve and Yaya Research Communities ......................... 122 Figure 4.1.3: Maps of Nsemre Forest Reserve and Nsemre Research Communities ................. 126 Figure 4.1.4: Maps of Sawsaw Forest Reserve and Sawsaw Research Communities ............... 130 Figure 4.1.5: Brong Ahafo Region- Districts and Forest Reserve Maps ................................... 144 Figure 4.1.6: Forest Reserves in Brong Ahafo’s Wenchi District .............................................. 145 Figure 5.0: Brong-Ahafo land-cover situations between 1990 and 2000 .................................. 207 xxvii Figure 5.1: Land-cover changes in Yaya, Nsemre and Sawsaw from 1990 to 2000 .................. 212 Figure 5.2: Land-cover changes in Km2 in Yaya, Nsemre and Sawsaw from 1990 to 2000 ...... 214 Figure 5.3: Land-cover changes in Yaya, Nsemre and Sawsaw from 2000 to 2012 .................. 216 Figure 5.4: Land-cover change in Km2 Yaya, Nsemre and Sawsaw from 2000 to 2012 ........... 219 Figure 5.5: Yaya Reserve Land Cover Change from 1990 to 2012............................................ 220 Figure 5.6: Yaya Reserve Land Cover Change in Km2 from 1990 to 2000................................ 223 Figure 5.7: Yaya Reserve Land Cover Change in Km2 from 2000 to 2012................................ 225 Figure 5.8: Nsemre Reserve Land Cover Change from 1990 to 2012 ....................................... 226 Figure 5.9: Nsemre Reserve Landcover Change in Km2 from 1990 to 2000 ............................. 229 Figure 5.10: Nsemre Reserve Land Cover Change in Km2 from 2000 to 2012 ......................... 231 Figure 5.11: Sawsaw Reserve Land Cover Change from 1990 to 2012 ..................................... 232 Figure 5.12: Sawsaw Reserve Land-Cover Change in Km2 from 1990 to 2000 ........................ 234 Figure 5.13: Sawsaw Reserve Land Cover Change in Km2 from 2000 to 2012 ......................... 236 Figure 5.14: Yaya Forest Reserve and Communities in 2012 .................................................... 238 Figure 5.15: Nsemre Forest Reserve and Communities in 2012 ................................................ 240 Figure 5.16: Sawsaw Forest Reserve and Communities in 2012 ............................................... 242 Figure 6.3.1: Spatial Distribution of Households Relative to Physical Assets in Ayigbekrom .. 336 Figure 6.3.2: Physical Features used in Constructing HHI ....................................................... 338 Figure 6.3.3: Livestock Categories used in Constructing HLI ................................................... 340 Figure 6.3.4: Communal Assets used in Constructing HPI ........................................................ 343 Figure 6.3.5: HII Distribution and Cluster Analysis Maps ........................................................ 347 Figure 6.3.6: HLI Distribution and Cluster Analysis Maps ....................................................... 350 Figure 6.3.7: HPI Distribution and Cluster Analysis Maps ....................................................... 353 Figure 6.3.8: HMI Distribution and Cluster Analysis Maps ...................................................... 355 Figure 6.3.9: Yaya Reserve Map Showing the Location of Abrefakrom .................................... 364 Figure 6.3.10: Abrefakrom Community Map ............................................................................. 367 Figure 6.3.11: Abrefakrom Community – Buffer on Public Toilets ........................................... 368 Figure 6.3.12: Abrefakrom Community – Buffer on Primary Schools ....................................... 369 Figure 6.3.13: Abrefakrom Community – Buffer on Corn Mill .................................................. 370 Figure 6.3.14: Abrefakrom Community – Buffer on Mosque ..................................................... 371 Figure 6.3.15: Abrefakrom Community – Buffer on Weekly Market .......................................... 372 Figure 6.3.16: Abrefakrom Community – Buffer on Major Roads ............................................. 373 Figure 6.3.17: Abrefakrom Community – Buffer on Provision Kiosks....................................... 374 Figure 6.3.18: Abrefakrom Community – Buffer on Junior Secondary School.......................... 375 xxviii Figure 6.3.19: Abrefakrom Community – Buffer on Dumpster .................................................. 376 Figure 6.3.20: Abrefakrom Community – Buffer on Borehole ................................................... 377 Figure 6.3.21: Yaya Reserve Map Showing the Location of Ahyiem ......................................... 378 Figure 6.3.22: Ahyiem Community ............................................................................................. 380 Figure 6.3.23: Ahyiem Community – Buffer on Borehole .......................................................... 381 Figure 6.3.24: Ahyiem Community – Buffer on Market ............................................................. 382 Figure 6.3.25: Ahyiem Community – Buffer on Road ................................................................ 383 Figure 6.3.26: Ahyiem Community – Buffer on Mosque ............................................................ 384 Figure 6.3.27: Ahyiem Community- Buffer on Corn Mill ........................................................... 385 Figure 6.3.28: Ahyiem Community – Buffer on Nursery and Kindergarten .............................. 386 Figure 6.3.29: Ahyiem Community – Buffer on Primary School ................................................ 387 Figure 6.3.30: Ahyiem Community – Buffer on Provision Kiosk ............................................... 388 Figure 6.3.31: Ahyiem Community – Buffer on Public Toilet .................................................... 389 Figure 6.3.32: Yaya Reserve Map Showing the Location of Amangoase ................................... 390 Figure 6.3.33: Amangoase Community ...................................................................................... 392 Figure 6.3.34: Amangoase Community – Buffer on Borehole.................................................... 393 Figure 6.3.35: Amangoase Community – Buffer on Church ...................................................... 394 Figure 6.3.36: Amangoase Community – Buffer on Clinic ........................................................ 395 Figure 6.3.37: Amangoase Community – Buffer on Public Dumpster ....................................... 396 Figure 6.3.38: Amangoase Community – Buffer on Provision Kiosks ....................................... 397 Figure 6.3.39: Amangoase Community – Households with Livestock ....................................... 398 Figure 6.3.40: Amangoase Community – Buffer on Corn Mill .................................................. 399 Figure 6.3.41: Amangoase Community – Buffer on Primary School ......................................... 400 Figure 6.3.42: Amangoase Community – Buffer on Major Road ............................................... 401 Figure 6.3.43: Amangoase Community – Buffer on Public Toilet ............................................. 402 Figure 6.3.44: Yaya Reserve Map Showing the Location of Amoahkrom .................................. 403 Figure 6.3.45: Amoahkrom Community ..................................................................................... 405 Figure 6.3.46: Amoahkrom Community – Buffer on Borehole ................................................... 406 Figure 6.3.47: Amoahkrom Community – Buffer on Preschool and Kindergarten .................... 407 Figure 6.3.48: Amoahkrom Community – Buffer on Provision Kiosks ...................................... 408 Figure 6.3.49: Amoahkrom Community – Households with Livestock ...................................... 409 Figure 6.3.50: Amoahkrom Community – Buffer on Market ...................................................... 410 Figure 6.3.51: Amoahkrom Community – Buffer on MTS Livestock Program .......................... 411 Figure 6.3.52: Amoahkrom Community – Buffer on Corn Mill ................................................. 412 xxix Figure 6.3.53: Amoahkrom Community – Buffer on Primary School ........................................ 413 Figure 6.3.54: Amoahkrom Community – Buffer on Major Road .............................................. 414 Figure 6.3.55: Amoahkrom Community – Buffer on Public Toilet ............................................. 415 Figure 6.3.56: Yaya Reserve Map Showing the Location of Asuakwa ....................................... 416 Figure 6.3.57: Asuakwa Community .......................................................................................... 418 Figure 6.3.58: Asuakwa Community – Buffer on Borehole ........................................................ 419 Figure 6.3.59: Asuakwa Community – Buffer on Church .......................................................... 420 Figure 6.3.60: Asuakwa Community – Buffer on Market ........................................................... 421 Figure 6.3.61: Asuakwa Community – Buffer on Junior Secondary School .............................. 422 Figure 6.3.62: Asuakwa Community – Buffer on Major Road ................................................... 423 Figure 6.3.63: Asuakwa Community – Buffer on Corn Mill....................................................... 424 Figure 6.3.64: Asuakwa Community – Buffer on Preschool and Kindergarten ......................... 425 Figure 6.3.65: Asuakwa Community – Buffer on Primary School ............................................. 426 Figure 6.3.66: Asuakwa Community – Provision Kiosks ........................................................... 427 Figure 6.3.67: Asuakwa Community – Buffer on Public Toilet .................................................. 428 Figure 6.3.68: Yaya Reserve Map Showing the Location of Ayigbekrom .................................. 429 Figure 6.3.69: Ayigbekrom Community ...................................................................................... 431 Figure 6.3.70: Ayigbekrom Community – Buffer on Boreholes ................................................. 432 Figure 6.3.71: Ayigbekrom Community – Buffer on MTS Livestock Program .......................... 433 Figure 6.3.72: Ayigbekrom Community – Buffer on Church ...................................................... 434 Figure 6.3.73: Ayigbekrom Community – Buffer on Dumpster .................................................. 435 Figure 6.3.74: Ayigbekrom Community – Buffer on Preschool and Kindergarten .................... 436 Figure 6.3.75: Ayigbekrom Community – Buffer on Provision Kiosk ........................................ 437 Figure 6.3.76: Ayigbekrom Community – Livestock Ownership ................................................ 438 Figure 6.3.77: Ayigbekrom Community – Buffer on Mosque ..................................................... 439 Figure 6.3.78: Ayigbekrom Community – Buffer on Corn Mill .................................................. 440 Figure 6.3.79: Ayigbekrom Community – Buffer on Major Road .............................................. 441 Figure 6.3.80: Ayigbekrom Community – Buffer on Public Toilet ............................................. 442 Figure 6.3.81: Yaya Reserve Map Showing the Location of Buoku Community ........................ 443 Figure 6.3.82: Buoku Community Map ...................................................................................... 445 Figure 6.3.83: Buoku Community – Buffer on Boreholes .......................................................... 446 Figure 6.3.84: Buoku Community – Buffer on Church ............................................................... 447 Figure 6.3.85: Buoku Community – Buffer on Public Dumpster ............................................... 448 Figure 6.3.86: Buoku Community – Buffer on Junior Secondary School .................................. 449 xxx Figure 6.3.87: Buoku Community – Buffer on Provision Kiosks ............................................... 450 Figure 6.3.88: Buoku Community – Buffer on Market ............................................................... 451 Figure 6.3.89: Buoku Community – Buffer on Mosque .............................................................. 452 Figure 6.3.90: Buoku Community – Buffer on Corn Mill ........................................................... 453 Figure 6.3.91: Buoku Community – Buffer on Preschool and Kindergarten ............................. 454 Figure 6.3.92: Buoku Community – Buffer on Primary School ................................................. 455 Figure 6.3.93: Buoku Community – Buffer on Public Toilet ...................................................... 456 Figure 6.3.94: Buoku Community – Buffer on Major Roads ...................................................... 457 Figure 6.3.95: Yaya Reserve Map Showing the Location of Konsua ......................................... 458 Figure 6.3.96: Konsua Community Map .................................................................................... 460 Figure 6.3.97: Konsua Community – Buffer on Borehole .......................................................... 461 Figure 6.3.98: Konsua Community – Buffer on Church ............................................................. 462 Figure 6.3.99: Konsua Community – Buffer on Provision Kiosks ............................................. 463 Figure 6.3.100: Konsua Community – Buffer on Mosque .......................................................... 464 Figure 6.3.101: Konsua Community – Buffer on Corn Mill ....................................................... 465 Figure 6.3.102: Konsua Community – Buffer on Primary School ............................................. 466 Figure 6.3.103: Konsua Community – Buffer on Major Roads .................................................. 467 Figure 6.3.104: Konsua Community – Livestock Ownership ..................................................... 468 Figure 6.3.105: Konsua Community – Buffer on Public Toilet .................................................. 469 Figure 6.3.106: Yaya Reserve Map Showing the Location of Malamkrom ................................ 470 Figure 6.3.107: Malamkrom Community Map ........................................................................... 472 Figure 6.3.108: Malamkrom Community – Buffer on Borehole ................................................. 473 Figure 6.3.109: Malamkrom Community – Buffer on Church ................................................... 474 Figure 6.3.110: Malamkrom Community – Buffer on Provision Kiosks .................................... 475 Figure 6.3.111: Malamkrom Community – Livestock Ownership .............................................. 476 Figure 6.3.112: Malamkrom Community – Buffer on Mosque ................................................... 477 Figure 6.3.113: Malamkrom Community – Buffer on Corn Mill ............................................... 478 Figure 6.3.114: Malamkrom Community – Buffer on Primary School ...................................... 479 Figure 6.3.115: Malamkrom Community – Buffer on Major Roads .......................................... 480 Figure 6.3.116: Malamkrom Community – Buffer on Public Toilet ........................................... 481 Figure 6.3.117: Yaya Reserve Map Showing the Location of Sewiah ........................................ 482 Figure 6.3.118: Sewiah Community Map ................................................................................... 484 Figure 6.3.119: Sewiah Community – Magnified ....................................................................... 485 Figure 6.3.120: Sewiah Community – Buffer on Major Roads .................................................. 486 xxxi Figure 6.3.121: Sewiah Community – Buffer on Borehole ......................................................... 487 Figure 6.3.122: Sewiah Community – Buffer on Church ........................................................... 488 Figure 6.3.123: Sewiah Community – Buffer on Corn Mill ....................................................... 489 Figure 6.3.124: Sewiah Community – Buffer on Primary School .............................................. 490 Figure 6.3.125: Nsemre Reserve Map Showing the Location of Ahwene .................................. 491 Figure 6.3.126: Ahwene Community Map .................................................................................. 493 Figure 6.3.127: Ahyiem Community – Buffer on Borehole ........................................................ 494 Figure 6.3.128: Ahyiem Community – Buffer on Church ........................................................... 495 Figure 6.3.129: Ahyiem Community – Buffer on Preschool and Kindergarten ......................... 496 Figure 6.3.130: Ahwene Community – Buffer on Mosque.......................................................... 497 Figure 6.3.131: Ahwene Community – Buffer on Corn Mill ...................................................... 498 Figure 6.3.132: Ahwene Community – Buffer on PAFORM Alternative Livelihood Program .. 499 Figure 6.3.133: Ahwene Community – Buffer on Major Road ................................................... 500 Figure 6.3.134: Ahwene Community – Buffer on Public Toilet ................................................. 501 Figure 6.3.135: Nsemre Reserve Map Showing the Location of Asuofre ................................... 502 Figure 6.3.136: Asuofre Community Map .................................................................................. 504 Figure 6.3.137: Asuofre Community – Buffer on Borehole ........................................................ 505 Figure 6.3.138: Asuofre Community – Livestock Ownership ..................................................... 506 Figure 6.3.139: Ahyiem Community – Buffer on Corn Mill ....................................................... 507 Figure 6.3.140: Ahyiem Community – Buffer on PAFORM Alternative Livelihood Program ... 508 Figure 6.3.141: Ahyiem Community – Buffer on Primary School .............................................. 509 Figure 6.3.142: Ahyiem Community – Buffer on Major Roads .................................................. 510 Figure 6.3.143: Asuofre Community – Buffer on Public Toilet .................................................. 511 Figure 6.3.144: Nsemre Reserve Map Showing the Location of Kofitsum ................................. 512 Figure 6.3.145: Kofitsum Community Map ................................................................................ 514 Figure 6.3.146: Kofitsum Community – Buffer on Borehole ...................................................... 515 Figure 6.3.147: Kofitsum Community – Buffer on Church ........................................................ 516 Figure 6.3.148: Kofitsum Community – Livestock Ownership ................................................... 517 Figure 6.3.149: Kofitsum Community – PAFORM Alternative Livelihood Program ................ 518 Figure 6.3.150: Kofitsum Community – Buffer on Primary School ........................................... 519 Figure 6.3.151: Kofitsum Community – Major Road ................................................................. 520 Figure 6.3.152: Nsemre Reserve Map Showing the Location of Pepewase ............................... 521 Figure 6.3.153: Pepewase Community – Map............................................................................ 523 Figure 6.3.154: Pepewase Community – Buffer on Borehole .................................................... 524 xxxii Figure 6.3.155: Pepewase Community – Buffer on Church ....................................................... 525 Figure 6.3.156: Pepewase Community – Livestock Ownership ................................................. 526 Figure 6.3.157: Pepewase Community – PAFORM Alternative Livelihood Program ............... 527 Figure 6.3.158: Pepewase Community – Buffer on Major Road ............................................... 528 Figure 6.3.159: Yaya Sawsaw Map Showing the Location of Ayaayo ....................................... 529 Figure 6.3.160: Ayaayo Community Map ................................................................................... 531 Figure 6.3.161: Ayaayo Community – Buffer on Boreholes ....................................................... 532 Figure 6.3.162: Ayaayo Community – Buffer on Church ........................................................... 533 Figure 6.3.163: Ayaayo Community – Buffer on Clinic ............................................................. 534 Figure 6.3.164: Ayaayo Community – Buffer on Dumpster ....................................................... 535 Figure 6.3.165: Ayaayo Community – Buffer on Preschool and Kindergarten ......................... 536 Figure 6.3.166: Ayaayo Community – Buffer on Provision Kiosks ............................................ 537 Figure 6.3.167: Ayaayo Community – Buffer on Market ........................................................... 538 Figure 6.3.168: Ayaayo Community – Buffer on Mosque .......................................................... 539 Figure 6.3.169: Ayaayo Community – Buffer on Corn Mills ..................................................... 540 Figure 6.3.170: Ayaayo Community – Buffer on Primary School .............................................. 541 Figure 6.3.171: Ayaayo Community – Buffer on Major Road ................................................... 542 Figure 6.3.172: Sawsaw Reserve Map Showing the Location of Domeabra ............................. 543 Figure 6.3.173: Domeabra Community Map ............................................................................. 545 Figure 6.3.174: Domeabra Community – Buffer on Boreholes.................................................. 546 Figure 6.3.175: Domeabra Community – Buffer on Dumpster .................................................. 547 Figure 6.3.176: Domeabra Community – Buffer on Provision Kiosks....................................... 548 Figure 6.3.177: Domeabra Community – Livestock Ownership ................................................ 549 Figure 6.3.178: Domeabra Community – Buffer on Corn Mill .................................................. 550 Figure 6.3.179: Domeabra Community – Buffer on Major Road .............................................. 551 Figure 6.3.180: Sawsaw Reserve Map Showing the Location of Ntema .................................... 552 Figure 6.3.181: Ntema Community – Buffer on Public Toilet .................................................... 554 Figure 6.3.182: Ntema Community – Buffer on Public Toilet .................................................... 555 Figure 6.3.183: Ntema Community – Buffer on Public Toilet .................................................... 556 Figure 6.3.184: Ntema Community – Buffer on Public Toilet .................................................... 557 Figure 6.3.185: Ntema Community – Buffer on Public Toilet .................................................... 558 Figure 6.3.186: Sawsaw Reserve Map Showing the Location of Papasu .................................. 559 Figure 6.3.187: Papasu Community Map................................................................................... 561 Figure 6.3.188: Papasu Community – Livestock Ownership ..................................................... 562 xxxiii Figure 6.3.189: Papasu Community – Buffer on Major Road ................................................... 563 Figure 6.3.190: Sawsaw Reserve Map Showing the Location of Pipotrim ................................ 564 Figure 6.3.191: Pipotrim Community Map ................................................................................ 566 Figure 6.3.192: Pipotrim Community – Buffer on Church ......................................................... 567 Figure 6.3.193: Pipotrim Community – Buffer on Dumpster ..................................................... 568 Figure 6.3.194: Pipotrim Community – Buffer on Preschool and Kindergarten ....................... 569 Figure 6.3.195: Pipotrim Community – Buffer on Provision Kiosks ......................................... 570 Figure 6.3.196: Pipotrim Community – Livestock Ownership ................................................... 571 Figure 6.3.197: Pipotrim Community – Buffer on Mosque ........................................................ 572 Figure 6.3.198: Pipotrim Community – Buffer on Corn Mill ..................................................... 573 Figure 6.3.199: Pipotrim Community – Buffer on Primary School ........................................... 574 Figure 6.3.200: Pipotrim Community – Buffer on Major Road ................................................. 575 Figure 6.3.201: Ghana’s Population Distribution by Region in 2010 ...................................... 577 Figure 6.3.202: Dominant Religions in Ghana’s three Northmost Region in 2010 ................... 578 Figure 6.3.203: Poverty Rates in Ghana’s three Northern-most Region in 2010 ...................... 580 Figure 6.3.204: Extreme Poverty Rates in Ghana’s three Northmost Region in 2010 .............. 581 Figure 7.6.1: Changes in Aggregate Asset Indexes between 1999 and 2009 ............................. 700 Figure 7.6.2: Within Group Changes in MTS and non-MTS Asset Indexes between 1999 and 2009............................................................................................................................................. 701 Figure 7.6.3: Between Group Changes in MTS and non-MTS Asset Indexes for 1999 and 2009 ..................................................................................................................................................... 703 Figure 7.6.4: Within Group Changes in MTS and NsemSaw Asset Indexes between 1999 and 2009............................................................................................................................................. 705 Figure 7.6.5: Between Group Changes in MTS and NsemSaw Asset Indexes for 1999 and 2009 ..................................................................................................................................................... 707 Figure 7.6.6: Within Group Changes in non-MTS and NsemSaw Asset Indexes between 1999 and 2009............................................................................................................................................. 709 Figure 7.6.7: Between Group Changes in non-MTS and NsemSaw Asset Indexes for 1999 and 2009............................................................................................................................................. 710 Figure 8.1.1: Illustration of Normal and Treatment Effect in a DID Approach ........................ 865 Figure 8.1.2: DID Estimates of Direct and Spillover MTS Project Effects on CHHFCI ........... 876 Figure 8.1.3: DID Estimates of Direct and Spillover MTS Project Effects on CHCI1 .............. 878 Figure 8.1.4: DID Estimates of Direct and Spillover MTS Project Effects on CHCI2 .............. 879 Figure 8.1.5: DID Estimates of Direct and Spillover MTS Project Effects on CHPCI1 ............ 881 Figure 8.1.6: DID Estimates of Direct and Spillover MTS Project Effects on CHPCI2 ............ 882 Figure 8.1.7: DID Estimates of Direct and Spillover MTS Project Effects on AHHSCI ........... 884 xxxiv Figure 8.1.8: DID Estimates of Direct and Spillover MTS Project Effects on CHHNCI1......... 886 Figure 8.9: DID Estimates of Direct and Spillover MTS Project Effects on CHHNCI2............ 888 xxxv CHAPTER 1: INTRODUCTION 1 1.1 Introduction to the Problem Tropical forests unarguably represent one of humanity’s greatest natural assets. Occupying approximately 7% of the entire earth’s surface (Pearce and Brown, 1994) and spread across more than 60 countries (Grainger, 1993), tropical forests account for more than fifty percent of the world’s biodiversity (Grainger, 1993, Visseren-Hamakers and Glasbergen, 2007). Sadly, across all topical countries, natural forests are declining at alarming rates with no one particular theory able to adequately explain the reasons for the decline. While some may be quick to link deforestation in the tropics to rapid population growth and demand for farmlands, Guppy (1984, p. 932) warn that superficial and sometimes obvious factors such as population growth, demand for agriculture land and settlements, upon close examination are indecisive and not necessarily important causes of deforestation. For example, in a recent study, it was revealed that “an important channel through which trade policy affects forests in Brazil, Indonesia, Malaysia and the Philippines is through agricultural expansion (López and Galinato 2005, p. 145).” López and Galinato (2005, p. 145) further assert that “economic growth has a negative and relatively large impact on forest cover in Brazil, Indonesia, Malaysia and the Philippines.” Ali, Benjaminsen and Hammad et al. (2005) also found that in the Western Himalayas of Pakistan, construction of access roads through the natural forest in the 1960s attracted both legal and illegal commercial harvesting leading to more than 50% decline in natural forest cover by end of the twentieth century. Like in many developing countries, rural communities in Ghana depend on forest resources for their livelihoods, however the decline of forests in Ghana have continued unabated, raising serious questions regarding the sustainability of rural livelihoods. Since the early 1900s, several attempts at preserving Ghana’s forests, for example through the establishment of 2 protected forest reserves and a combination of different Community-Based Forest Management Programs (CBFMP) have had mixed results. A few reasons cited for the rapid and sustained forest loss in Ghana include, rising population and a correspondent increase in the demand for agricultural land (Agyarko, n.d., p.17), and fuel wood (Adam, 1999), ineffective forest concession policies (Kufuor, 2000, p.53), implementation of structural adjustment policies (Hilson, G. M., 2004, Tockman 2002 and Owusu, 2006) and the increase in legal and illegal chainsaw operations (Kotey et. al., 1998, p.43, Lawson and McFul, 2010, p.12). Perennial bushfire (Agyarko, n.d.), has also been cited as one of the culprits of deforestation in Ghana. For example, several researcher attribute some of the loses in Ghana’s vegetative cover to the forest fires of 1983 and 1984 where wildfires were blamed for destroying at least 50% of Ghana’s vegetative cover in 1983 alone (Adam, 1999). It is also estimated that in 1984 and 1985 alone, as many as 1005 wildfires were recorded in Ghana’s tropical rainforest and savannah regions with at least 307 destroying portions of the semi-deciduous forests zone (Ampadu-Adjei, 1987 cited in Adam, 1999). Other important contributing factors to forest lost in Ghana particularly in the late 1960s leading to the early 1990s are loss of communal control (Dei, 1992) and loss of state power, legitimacy and authority (Ewusi, 1984b) characteristic of political unstable governments. Among the possible solutions to Ghana’s forest loss, some researchers have argued for radical shifts in Ghana’s forest policy from one of purely resource extraction and enforcement (Asante, 2005), to policies that promote local ownership and participatory management of forests resources (Agyeman, Kansanga, Danso et. al, n.d). 1.2 Background of the Study In response to tropical deforestation, there has been a variety of participatory programs worldwide focusing on reforestation. One such approach is Participatory Forestry (PF) adopted 3 in India in the 1970s and later modified into the infamous Joint Forest Management (JFM) in the 1990s (see Damodaran, and Engel, 2003 and Basu, 2012). The entire premise of JFM is to promote participatory management and joint ownership of forest resource between local communities and national forestry departments (see also Datta and Sarkar, 2010, and Vemuri, 2008). Another reforestation program that was practiced in most tropical countries since the mid- nineteen century is Taungya forestry. Taungya is a forest plantation system in which timber tree seedlings are intercropped with annual food crops for a period ranging between 2-3years. Unlike JFM that emphasizes local ownership of forests, multiple cropping regimes, forest resource development and equitable sharing of benefits (Dasgupta and Debnath, 2008) in traditional Taungya, program participants have no ownership rights in forest resources and the system promotes commercial tree monocultures with a heavy revenue and profit orientation. Critics of traditional community forestry regimes (Sanwal, 1988) particularly traditional Taungya (Agyeman, Kansanga, Danso et. al, n.d) argue that these program fail to prevent forest loss and because of a lack of equitable benefit distribution system. The critics also claim that apart from a daily wage compensations paid for plantation establishment and occasional gains from annual intercrops, peasants in traditional Taungya systems have little or no stake in the profits from timber (Milton, 1994, cited in Agyeman et. al., 2003). Though traditional Taungya may be traced as far back to16th century China (Xu Guangqi, 1639, p. 36a cited in Menzies, 1988, p. 363) the system has been ineffective in rescuing tropical forests. A primary reason for the failure of traditional Taungya systems is because they focus rather narrowly on commercial tree production and revenue generation for national forestry departments without ensuring that the system as a whole improved the livelihoods of the local people. The failure of traditional Taungya as an effective reforestation 4 scheme has spawned several modified taungya systems (MTS) in Kenya (Takeda, 1992), Nigeria (Lowe, 1987 and Evans, 1982), Thailand (Gajaseni 1992) Tanzania (Mgeni, 1992) and Ghana (Agyeman, Kansanga, Danso et. al, n.d). Ghana’s new MTS was launched under a National Forest Plantation Development Program (NFPDP) established in 2001 under a special presidential initiative (NFPDP, 2004). Like the traditional Taungya system, the new MTS’ mandate is to allocate degraded portions of forest reserves to mostly resource/asset poor farmers in forest fringe communities. While both the traditional Taungya and the new MTS allow participating farmers to intercrop tree plantations with annual food crops for periods ranging between 2-3 years, the two programs differ significantly in their treatment of benefits accruing to timber harvested under Taungya. In the traditional system, farmers have no stake in the profits from harvested timber while the new MTS recognizes farmers’ contributions in the form of sweat equity built into established forest plantations. The new MTS thus attempts to correct previous inequities embedded in the traditional system by guaranteeing current MTS participants 40% shares in the profits from harvested timber (Forestry Commission, 2002). The Forestry Department’s share of profit is also 40% while 15% is expected to go to the landowners (Forestry Commission, 2002). Landowners in this context refer to the traditional authorities and chiefs entrusted with all communal lands in a township including forest reserves within the town (Forestry Commission, 2002). The remaining 5% of Taungya profit is earmarked for various forest community support programs (Forestry Commission, 2002). Another major shift from previous Taungya systems is reliance of the new MTS on a formal binding contract between the program implementers (i.e. Forest Service Division- FSD) and local communities and farmers. A memorandum of understanding (MOU) between the FSD 5 and local communities supposedly makes it binding on the MTS program implementers and participants to honor their part of the contract. The initial and continued success of the MTS in improving forest vegetation and the socio-economic status of participants is predicated on the degree to which the MOU is upheld by both parties. Hence to pave the way for the successful implementation of the new MTS, the MOU mandates the Ghana Forest Services Division (FSD) to partner with local communities and farmers to establish short rotation timber plantations in degraded forest reserves. Under the NFPDP initiative, the Plantations Department (PD) of the FSD is responsible for the implementation, coordination and management of the NFPDP (Forestry Commission, 2002). Also, the FSD and Forest Research Institute of Ghana (FORIG) on one hand are expected to provide technical direction, field surveys and demarcation of degraded forest reserves as well as supply pegs and seedlings for plantation development (CSIR-FORIG, 2003). Local farmers on the other hand are responsible for providing all the labor inputs needed for site clearing, pegging, planting and maintenance, and fire protection. In the present MTS, a greater proportion of degraded forest reserves are been reforested with Tectona grandis (Teak) due to its relatively short rotation period and high economic value. While Ghana’s current MTS sounds promising on multiple fronts, one major drawback/threat to the program is the failure of the rotation periods to fit well into farmers’ economic horizon (Gajaseni and Watanabe, 1992 also made this observation in Thailand). Currently three rotation periods of eight, fifteen and twenty-five years have been proposed for Teak plantations established under MTS. Assuming intercropping is allowed for a maximum of 3 years, the 8, 15 and 25 year rotation periods translates into a wait period of 5 to 22 years before benefits from MTS teak plantations may be realized. Hence in order to lend additional support to the MTS program, particularly during the 5 to 22 year lag periods for which no other benefits of 6 Taungya may be forthcoming, the Government of Ghana, the Global Environment Facility– Small Grants Program of the United Nations Development Program (UNDP-GEF, 2008)) and the African Development Bank (AfDB) have since 2002 implemented a variety of alternative livelihood support programs dubbed the Community Forest Management Program (CFMP). The CFMP program, currently under pilot test in the 10 MTS villages/communities around the Yaya Forest Reserve, in addition to allocating land to famers CFMP also provides MTS participants in Yaya forest communities with start-up capital for small-scale livestock production and tree seedling establishment. Ghana’s MTS thus appears to have a dual intent of improving the forests and the livelihoods of forest communities. One of the factors that might affect the success of the MTS program is community and participant selection. The statement below provides some insight into how one Yaya community might have been chosen for the MTS. “My parents founded this town sometime in 1942 or 1943 but the land we have been cultivating since then belongs to other chiefs. Recently, we the elders petitioned the government to allow our people to cultivate portions of the forest so that we the landless can also feed our families. Thankfully, the government granted us permission to farm portions of the Yaya forest (Amangoase, MTS focus group interview, 2009).” From the above statement, it may be logical to think that the poor and landless are more likely to encroach on forests for subsistence than the resource rich. If the preceding argument is valid then it is important that forest policy governing MTS implementation get MTS participant selection right in order for the program to achieve the two main objectives. At present, Ghana’s MTS in the Brong Ahafo Region has been running for nearly ten years if not more and yet several important questions still remain regarding selection of program participants and the impact of the program on the forest and people. If policy makers believe that involving local 7 people is essential to making reforestation work, then a conscious effort must be made to involve the right people, the people who are a real threat to the forest. It is against this background that this dissertation research was launched to answer the following pressing research questions:1 ) What land cover changes have occurred in Yaya, Nsemre and Sawsaw forest reserves in Ghana’s Brong Ahafo region before and after the MTS was launched in 2002, 2) What factors influence MTS community and participant household selection? 3) What changes in livelihood assets have occurred among MTS participant and non-participant household since launching of the program in 2002? and 4) To what extent are changes in household livelihood assets attributable to the MTS program implementation? 1.3 Problem Statement The MTS concept has been practiced in Thailand since the 1960’s. Thailand’s MTS, also known as the “Forest Village System,” constituted groups of 100 families into village units for the purpose of forest plantation establishment (Watanabe, Sahunalu and Khemnark, 1988). Unlike Ghana’s MTS that shares timber profits with participating farmers, the Forest Village allocates 0.16 hectares of land in the program area for program beneficiaries to build their homes and farm (Watanabe, Sahunalu and Khemnark, 1988, p. 170). Additionally, the system provides health care, education, and other social services to MTS families in return for farmers tending 1.6 ha of tree plantations each year (Watanabe, Sahunalu and Khemnark, 1988, p. 170). In general MTS has been around for nearly half a century however no studies have been done to determine how the program jointly impacts livelihood assets and the forest. At present majority of the research on both the traditional Taungya and MTS have focused primarily on understanding: tree physiology (Ross, 1959, Palupi and Owens, 1997), interactions between tree vegetation and soils (Bhatia, 1958, Egunjobi, 1974, Borggaard, Gafur, 8 and Petersen, 2003), tree herbicide interactions (Ross, 1958 and 1961), competition between trees and food crops (Watanabe, 1992, Schlönvoigt & J. Beer, 2001 and Imo, 2009), tree and yield improvement (Gajaseni and Jordan, 1990, Ennion, 1998, Rugmini and Jayaraman, 2008, Adesina, 1990), economic benefits of tree improvement (Kjær and Foster, 1996) and cost- benefits of Taungya (Agbeja, 2004, Current, Lutz, Scherr, 1995, Borggaard, Gafur, Petersen, 2003). Other studies of social dimensions have mainly served to enrich the historical context within which Taungya developed in Burma (Spate, 1945, Bryant, 1993, 1994 &1996), China (Menzies, 1988), India (Sivaramakrishnan, 1995, and Guha and Gadgil, 1989), Indonesia/Java (Peluso, 1991 & 1993, Boomgaard, 1992), Nigeria (Egunjobi, 1974) and Ghana (Bedele, 1988, England, 1993). Of the few studies that have attempted to apply socio-economic tools to investigate the livelihood impact of Taungya or other similar programs, Maung and Yamamoto (2008) used principal component methodology to cluster household in three Taungya villages in Burma for the purpose of tracking changes in socioeconomic and livelihood patterns. Maung and Yamamoto’s (2008) work essentially predicted common patterns in economic activities of Taungya participant and non-participants. Maleson’s et al. (2008) study of rural communities in Cameroon, Nigeria and Ghana paved the way for using community maps and a weighting scheme to rank household wealth status based on an aggregate score of household physical assets. Yao, Guo and Huo’s (2010) research on the “effects of China’s land conversion program on farmers’ income growth and labor transfer” demonstrate how cutting-edge econometric tools such as Difference in Difference (DID) may be used to isolate the effect of reforestation policy on household livelihood assets. Yao, Guo and Huo’s (2010) study demonstrated how DID may be used to isolate the impact of a reforestation project on different livelihood assets. In a recent 9 study to determine the livelihood impact of a carbon offset forestry project in Mozambique, Jindhal, Kerr and Carter (2012) used Logit and Tobit econometric models to determine factors that influence participant selection into the reforestation program. Jindhal, Kerr and Carter’s (2012) study also tracked changes in livelihood assets before and after the project implementation. While several studies including those cited above have made tremendous contributions to the body of literature on the impact of reforestation programs on household livelihood assets, there still remain gaps in the literature that needs to be filled. Some of the gaps identified in the literature include identifying the determinants of community or household selection into afforestation programs as well as the extent to which a household’s physical location, relative other communal assets influence selection into a livelihood support program. In this dissertation research, I attempt to fill the gaps identified in the literature by first comparing remotely sensed images taken of the study area reserves in 1990, 2000 and 2012. In order to detect and classify land cover changes Yaya, and two other non-MTS reserves (Nsemre and Sawsaw), I conducted a post classification analysis of LANSAT images from 1990 and 2000 and also 2012 DMC image of all three forest reserves. The two other reserves were used as a control to better explain changes in the MTS reserve (Yaya). To address the question of community and household selection into MTS, I used a Binary Probit Model (BPM) to determine which factors significantly influence community and household selection into the MTS. I also used data generated from geo-referenced community maps to model the effect of households’ relative location to communal physical assets on selection into MTS. (e.g. borehole, major roads). To quantify changes in household livelihood assets, I used descriptive statistics generated from a five-part survey instrument targeting all five major livelihood asset categories (Financial, Social, Physical, Natural and Human capital assets). From the mean values obtained for each 10 asset category, I generated a list of livelihood indexes for each livelihood asset category and then conducted a t-test to determine before and after MTS changes and the level of significance of the observed changes. In order to determine the changes in livelihood assets that may be attributed to the MTS, I used the DID model. The findings of the empirical analysis presented in this dissertation thus provides a basis for comparing changes in forest cover as well as livelihood changes that have occurred ten years after the MTS project implementation and to what extent the observed changes may be attributed the MTS. By quantifying changes in livelihood assets and establishing a relationship between observed changes and the MTS program or lack thereof, this research provides information that enhances forest policies in favor of resource sustainability and rural livelihoods improvement. 1.4 Purpose of the Study From the above discussion, one can safely conclude that selection of communities and participating households is critical for the success of Ghana’s MTS. Yet to date, very little is known about how communities and households were selected. It may be possible that failure of past national reforestation programs to effectively improve forests and household livelihood is because the right communities and household were not recruited to begin with. Rescuing Ghana’s forests will thus require researchers to understand not only how afforestation program affect forests and livelihoods but also factors that influence community and household selection into these programs. The MTS is past 10 years and now is a good time to evaluate its progress and also factors that influence recruitment. Currently there is not a single study on taungya be it traditional or modified that investigates how communities and program participants are recruited and what factors influence recruitment and participation. This dissertation attempts to bridge these literature gaps by first quantifying changes in forest cover and livelihood assets before and 11 after MTS, and then isolating the changes in livelihoods that result from the MTS implementation. Isolating the impacts of MTS is accomplished by the use of cutting-edge econometric techniques such as the DID. To determine factors that influence MTS community and household selection, two groups of Binary Probit Models are used. Knowing how different livelihood assets impact community and participant selection will help refine future MTS implementation policies towards achieving program goals. The purpose of this dissertation research is thus to investigate the factors that influence MTS community and household selection and also changes in the forest and livelihoods. 1.5 Method used in the Study This dissertation research uses an asset-based approach (Chambers and Conway, 1991, Siegel and Alwang, 1999, Scoones, 1998, Krantz, 2001, Carter and Barrett, 2006 and Moser, 2006) to guide survey design, data collection and analysis. Data was collected on five livelihood asset categories among MTS participant and non-participant households. A census approach was used to select all MTS participants while non-participants were selected randomly from the list of all households generated during community mapping exercises. In communities that had 30 or fewer households, a census approach was used in selecting households for the study. Both quantitative and qualitative household data was collected from all 10 communities along the Yaya reserve currently participating in the MTS program and 4 communities along the Nsemre reserve that do not participate in the program. A similar set of data was also collected from 5 communities fringing the Sawsaw reserve. The qualitative data consisted of 29 focus group interviews of 10 MTS and 10 non-MTS representatives in Yaya, and at least one group each in each of the Nsemre and Sawsaw communities. 12 For both MTS and non-MTS participant households and communities, the study generated data for the two time periods 1999 and 2009. Since communities along Sawsaw and Nsemre reserves do not have the MTS program, these two reserves together serve as the control group so that the results of the study may provide an opportunity to uncover differences in household assets if any between MTS (i.e. project) and non-MTS (i.e. non-project) participant households and communities. By comparing pre- and post-MTS livelihood asset levels it was possible to determine what changes if any had occurred during the two periods and the extent to which these changes may be attributed to the MTS program or the lack thereof. The individual household survey was designed so as to allow each respondent to select responses on a Likert-scale. All responses within different asset category were used to generate asset indexes for the purpose of comparison and econometric analysis. The method for assigning weights to different categories of livelihood assets (see Maleson et al., 2008) allowed the responses to be analyzed using descriptive statistics, simple multi-linear and probit regressions, Difference in Difference (DID) models. The quantitative data analysis was done in several steps. First, household livelihood indexes were computed from the descriptive statistics obtained for each of the five household livelihood categories (i.e. human, financial, natural, physical and social). Then using Microsoft excel, household livelihood pentagons were constructed (see Messer and Townsley, 2003) to allow graphical representation of between and within group differences in each asset category for the two time periods 1999 and 2009. Following construction of the livelihood pentagons, a two-sample t-test was conducted in STATA 12 to determine the extent to which within and between group differences are significant for the two time periods 1999 and 2009. In the second step, a probit model was used to determine factors that influence community and household selection into the MTS program. 13 The GIS maps were especially important in determining the effect of a household’s relative location to communal physical capital assets on selection into the MTS program. During the third step a series Binary Probit Models were used to determine factors that affect MTS community and household selection. The fourth step involved the use of remotely sensed LANSAT and DMC images to detect changes in forest cover within the three forest reserves before and after MTS program implementation. To detect changes in forest cover before and after the MTS program, LANSAT images taken of the study site in 1990 and 2000 and DMC images taken in 2012 were processed using image processing software ArcGis 10.0 and Envi 4.7. The study used the unsupervised classification method to track changed in forest cover, degraded forest areas, settlements/open spaces and road, and grass and shrubs within the reserves. By combing data from remotely sensed LANSAT and DMC images of all three forest reserves with the survey results on livelihood assets this dissertation was able to adequately investigate the impact of Ghana’s MTS program on forest and household livelihood assets. 1.6 Significance of the Research Till date no study on rural livelihoods has simultaneously tracked changes in all five major livelihood asset categories and shown how these asset categories interact to define a household’s livelihood status. This dissertation tracked changes in forest cover and household livelihood assets and show how a household’s asset endowments may be impacted by forest resource management regime such as the MTS. The dissertation research effort also provides a systematic approach to monitoring the impact of Ghana’s new MTS program and making changes to the MTS policy where necessary. The pre- and post-MTS information on household livelihoods first provides a baseline for continues tracking of livelihood assets and secondly 14 allows for monitoring of how well forest resources are conserved and rural livelihood assets improved. By developing measurable indicators and demonstrating how these can be used to continuously track livelihood impacts of Ghana’s MTS, this study provides an opportunity for researchers to adapt the study methodology to similar afforestation programs in Ghana and other countries. 1.7 Layout of the Dissertation The rest of the dissertation is divided into nine chapters. Chapter two presents literature on the following topics: a) tropical deforestation, b) historical background information on Ghana, c) Ghana’s land tenure system and the evolution of forest policy from pre-colonial, colonial and the post independent era d) review of Ghana’s economic conditions and its impact on forests with special attention to the three decades (1965-1983) leading to Ghana’s adoption of World Bank and IMF Structural Adjustment Programs (SAP); e) Ghana’s Structural Adjustment Program (SAP) and its impact on forest cover particularly in Ghana’s High Forest Zone (HFZ); f) review of Taungya literature and the role of a novel afforestation programs such as Ghana’s MTS in forest resource and rural livelihood improvement, and g) review of literature on the “Sustainability Livelihoods Framework (SLF)” and how it is used to guide instrument design, data collection and analysis. Chapter three presents the methodological approach (the household livelihoods approach) used in data collection and analysis, while chapter four presents the findings of the research. Chapter four provides background information into the research communities, the sampling procedure used in executing household surveys and community GIS mapping and data analysis. Chapter five tracks changes in three landcover classes within the Yaya, Nsemre and Sawsaw forest reserves that provide a source of livelihood for the 19 communities included in the study. 15 The chapter also describes the methodology uses in acquisition of satellite images as well as the processing procedure and interpretation of landcover change results and their implications for future reforestation programs in Ghana and similar countries confronting deforestation crisis. Chapter six used the household survey data as well as data from the GIS community maps to generate Binomial Probit Models (BPMs) that predicted the extent to which different socio- economic factors within a household and the household’s spatial situation influences communal and household selection into the MTS program. In chapter seven, household survey data generated in the 10 MTS and nine non-MTS communities was used to generate household livelihoods indices across five asset categories. The chapter then uses a two-sample t-test to determine between and within group differences in the five livelihood indexes and the extent to which these differences were significant for the two time periods before MTS (1999) and after MTS (2009). Chapter eight uses Difference-in-Difference methodology to determine how much of the observed changes in changes in each livelihood asset may be attributed directly to the MTS program and how much of the changes may be a result of a spillover from the project. All literature cited in this dissertation are presented in chapter nine (bibliography) while chapter 10 contains each of the previous chapter appendixes. Chapters five through eight each contain chapter conclusions and policy recommendations. 16 CHAPTER 2: FORESTRY IN GHANA 17 2.1 Introduction This dissertation examines how a novel participatory reforestation programs similar to India’s Joint Forest Management (JFM) and Thailand’s Forest Village System is being implemented in Ghana to improve both forests and livelihoods. Hence in an effort to explore what is already known about JFM/participatory and apply it specifically to the Ghanaian situation, this chapter begins first reviews a wide range of literature on the causes and consequences of global deforestation and the approaches adopted at the global, regional and local levels to manage the crisis. Following the global and regional overview on the crisis of tropical deforestation and forest management, the rest of the chapter focuses on forest policies under colonial and independent governance regimes in Ghana. 2.2 Literature Deforestation no matter where it occurs is like a melting pot in which no one particular cause or solution bubbles to the top. Hence in discussing the subject of deforestation, this chapter first attempts to uncover the diverse causes before proceeding to discuss possible solutions. The chapter first discusses deforestation within a global context and then directs attention towards contributing factors to forest loss in Ghana. In order to sets the stage for discussing the problem of deforestation in Ghana, chapter two also provides a brief historical background on Ghana and the land tenure system before discussing pre-colonial, colonial and post-colonial forest policies and their impact on the country’s forests. A brief review of Ghana’s economic situation leading to adoption of World Bank and International Monitory Fund’s (IMF) Structural Adjustment Programs (SAP) in the 1980s and the link between SAP and possible changes in forest cover is also discussed. Chapter two divided into seven sections covering: 1) The Crisis of Tropical Deforestation 18 2) Historical Background on Ghana 3) Agro Climatic Conditions in Ghana 4) Ghana’s Agro-ecological Zone (AEZ) and High Forest Zones (HFZ) 5) Traditional System of Governance and Land Tenure 6) Historical Background on Ghana’s Forest Policy 7) Ghana’s Economy after Independence (1957) 8) Structural Adjustment Programs (1983-2000) 9) Secure Tenure Rights and Community involvement in Forest Management 10) Tax Reform and Timber Certification 11) Role of the Media in forest Management 2.3 The Crisis of Tropical Deforestation The importance of tropical forests in supporting entire agro-ecological zones is well documented. For humans, tropical forests provide benefits in the form of timber and non-timber forest products (NTFPs) such as fruits, mushrooms, medicinal herbs, energy in the form of firewood, and charcoal (Culas, 2008). Besides livelihood support, tropical forests also help to regulate the climate by acting as sink for carbon (Culas, 2008). Global deforestation particularly in the tropics has gained a lot of attention due to the critical role forests play in providing important livelihood services (Deacon 1994) as well as regulating the climate (Guppy 1984, Perman, Ma and Common et al., 2003). According to Pearce and Brown (1994, p.3-4), “the Intergovernmental Panel on Climate Change estimated that in the 1980’s, tropical deforestation alone accounted for approximately 1.6 billion tons of all carbon emissions, representing one-fifth to a quarter of all greenhouse effects.” Amelung and Diehl, (1992) also claim that between 1981 and 1990, permanent agriculture (defined as land for pasture, permanent crops, and arable land) 19 accounted for 45% of deforestation in all tropical countries (Amelung and Diehl, 1992 in Woollery, 1997, p. 22).” Some researchers have argued that since total greenhouse gases from tropical deforestation alone surpasses those from all the planet’s transportation systems combined (Leonard, Kopp and Purvis, 2010), the problem of climate change cannot be fully resolved without finding a lasting solution to the problem of tropical deforestation (Boyd, 2010). A recent study by NASA revealed that tropical forests in 75 countries across three continents store over 245 billion tons of carbon (Saatchi, Harris and Brown et. al., 2011), hence policy actions towards reducing tropical deforestation will be more expedient and cost-effective in reducing annual GHG emissions (Bosetti and Lubowski, 2010) while mitigating climate change (Cattaneo, 2010). In effect, the large stocks of carbon trapped in tropical forests, means its destruction, if not controlled, will have serious repercussions for global carbon cycles, which in turn will influence future changes in the climate (Fearnside, 2004). The following statement sums up the urgent need for policy change in favor of tropical forest conservation: “….. controversies concerning the impact of present climate change on tropical forests remain unfinished. The possibilities of substantial impacts that damage the forest and introduce positive feedback effects into the climate system are sufficiently large that these impacts should be an important consideration in defining policies that affect both land- use change and global greenhouse gas emissions. The need for more research is obvious, but policy changes should not be held hostage to the results of further research (Fearnside, 2004, p.301).” Whether tropical forests are valued for what they truly represent remains an open question. A 2005 Global Forest Resource Assessment (GFRA) raised very serious questions regarding the future of tropical forests and their ability to continue to regulate climatic conditions as well as sustain livelihoods. According to the GFRA, the world’s baseline forest area in 2005 was approximately 4 billion hectares (Ha) (representing 30% of total global land area) (FAO, 2006). The same report also noted that 13 million hectares per year of forests was converted into 20 agricultural land between 2000 and 2005 (FAO, 2006). South America and Africa respectively recorded the most severe net forest losses of about 4.3 million and 4.0 million hectares between 2000 and 2005 (Lindquist, D’annunzio and Gerrand, et. al., 2012). Since the turn of the 20th century, the process of desertification, and accompanying drought and famine has threatened the livelihoods of several African countries. In 1985, a meeting of African Head of States at the Organization of African Unity summit in Addis Ababa Ethiopia revealed shocking details of the extent of desertification in Africa. According to the OAU report, the process of desertification in Africa is occurring at the rate of about 8-10 kilometers per annum and out of 36 countries facing serious food shortages, 24 are affected by desertification (Maloka, 2002, p.180). The graph below represents changes in forest cover in six West African countries between 1900 and 1990. Figure 2.1.1: Changes in Forest Cover in Selected West-African Countries (1900-1990) Change in Forest Cover- 1900-1990 ) 0 0 0 1 x ( a H n i a e r A 25000 20000 15000 10000 5000 0 1900 (Forest Cover) 1948 Forest Cover 1985 Forest Cover 1990 Forest Cover Benin Ghana Guinea Ivory Coast Liberia Sierra Leone Togo Source: FAO 1993: Annex 1. Tables 7a and 8a, Sayer, Harcourt and Collins (1992) Cited in Leach and Fairhead (2000) and Ghana Forestry Department 1949-194 Annual Report. Cited in Dadebo and Shinohara (1999) Most international conservation and forestry researchers acknowledge the extent of West Africa’s unprecedented forest loss since the beginning of the 1900s and particularly in the last 21 three decades (Leach and Fairhead, 2000). For example, the World Resources Institute’s (WRI) assessment of West Africa’s forest cover suggests that nearly 90% of the original forest in the region has disappeared and even what remains is heavily fragmented and degraded (Bryant, Nielsen, and Tangley, 1997). Figure 2.1.1 above is a stark reminder of changes that have occurred on the West African landscape in the 20th century (1900 and 1990) alone. At present, a host of reasons including tropical country’s struggle for economic development, weak forest policies, and rising demand for agricultural land, seem to broadly explain some of the trends presented in figure 2.1.1. For the most part, majority of research findings and publications on the subject of deforestation seem to attribute the high rates of tropical deforestation, to rising populations in forest fringe communities and subsequent increase in the demand for farmland and fuel wood. While increasing populations within forest fringe communities provide the necessary conditions, population growth alone does not sufficiently explain the high rates of deforestation in tropical forests since the beginning of the1900s. Leach and Fairhead (2000) warn that the neo-Malthusian population-deforestation narrative is overly simplistic and masks the true causes of the problem. The basic argument is that “neo-Malthusian narratives not only misrepresent the relationships between people and forests, but also grossly misstate the ways these relations have evolved for centuries (Leach and Fairhead, 2000, p.18).” While uncontrolled population growth in part may explain trends in tropical deforestation, over reliance on a neo- Malthusian theory particularly in framing forest policies obscure the mechanisms by which forest communities have preserved their forests for generations even in the face of rising population growth (Leach and Fairhead, 2000 also see Moran and Ostrom, 2005). A historical perspective of forest policies and their influence on traditional land tenure systems, as well as an in-depth analysis of national economic trends and associated government policies all of which affect land 22 use, offers a more balance explanation and understanding of past and current trends in deforestation. Chapter two of this dissertation attempts to provide a balance perspective of forest loss in Ghana. 2.4 Historical Background on Ghana Until the later parts of the 15th century, the nature and habitation of present-day Ghana was largely unknown to most Europeans. It was not until January 1471 that the first Portuguese expedition made landfall at Edina (present day Elmina in Cape Coast) and later in 1482 built the first European castle and forte at Elmina (Ward, 1948, p. 61-62). It is believed that European expeditions were driven largely by the rediscovery of gold along the shores of Cape Three Points and the quest to secure trade routes. The name Gold Coast was thus coined out of the country’s wealth in gold and the subsequent gold trade that ensued between natives of the coast and their new European visitors (Ward, 1948, p.17, 60, and 81). Notwithstanding all attempts by the Portuguese to monopolize trade in the Gold Coast, as is evidence in the building of several other fortes at Axim, Shama and Accra (Ward, 1948, p.164) their presence was constantly challenged by the Spaniards, French, English and Dutch. By the end of the 18th century, at least “thirty-five villages along the entire Gold Coast had fortes built, abandoned, sold, attacked, exchanged, captured and recaptured (Ward, 1948, p.85). For example, in June of 1637 the Dutch after a previously failed attempt, eventually defeated the Portuguese, captured the Elmina castle and later in 1642 expelled the Portuguese completely out of the Gold Coast (Ward, 1948, 72-73). During the 160 years of Portuguese presence on the Gold Coast, trade was centered primarily on gold and slaves, both purchased and transported to Europe and the Americas. While the Portuguese to a large extent bear some responsibility for the highly criticized slave trade on the Gold Coast, they are also credited for introducing high value crops such as “oranges, lemon, 23 lime, rice and sugar cane from the Far East (Ward, 1948, 74).” Nearly all of the major staple crops currently eaten in the country today including cassava and maize owe their origins to the Portuguese, having brought these together with guava, pineapples, and tobacco from the Americas (Ward, 1948, p.74). To maintain trade and political dominance within the Gold Coast, the British combined military aggression with diplomatic buyout of other European nations. First, they bought out all the major European competitors within the territory beginning with the Danes in March of 1850 and ending with the Dutch 1871 (Horton, 1868). The second strategy that helped consolidate British political power in the Gold was the total defeat of the Ashanti by the turn of the 20th century. Militarily, the first British and native allied victory over the Ashanti started with the Akatamanso (also called Dodowa) War of 1837. The Akatamanso war set in motion several others victory such as Wolsely’s expedition of 1874 (Beckett, 2009, p.365), and Sir Francis Scott’s expedition of 1895. In spite of previous victories, the final defeat of the Ashanti came during the Yaa Asantewaa War of 1900-1901 (Boahen, 2003), resulting in the annexation of Kumasi and declaration of the entire Ashanti territory as a British protectorate (Kimble, 1963, p.264-333, and Beckett, 2009, p.374). Victory in the Yaa Asantewaa War thus completed British conquest of the entire region of present-day Ghana. It is important to note that the period stretching from the mid-18th century leading to the late 19th century, was marked by Ashanti conquest and dominance of nearly all the major tribes along the Gold Coast and the hinterlands. The Ashanti Kingdom during the period of their conquest stretched over nearly all of present-day Ghana (Edgerton 1995, p.5). Hence the defeat of the Ashanti as will be discussed later in this chapter, had serious consequences for Ghana’s forest particular the portion later identified in the chapter as the High Forest Zone (HFZ). It may 24 be possible that the ban on slave trade in 1807 and the need to fill the trade vacuum with other natural resources drove the British to expand their territory beyond the coast thus bringing them into conflict with Ashanti and other powerful inland tribes. Another theory is that the scramble for Africa, possibly fueled by a ban on the trans-Atlantic slave trade, ignited a race among all the major European nations for territorial control within Africa. The result of that race was the partition of the continent into artificial boarders at the Berlin Conference of 1884-1885 (Currey, 1992). Hence by the first quarter of the 20th century, more than 90% of the African continent was partition into some 55 countries, 40 political units and seven territories controlled by Europeans (Currey, 1992, p22). The Gold Coast, Ashanti and Northern Territories were awarded to the British who subsequently governed by colonial rule until 1956. The British Togoland also known as the Trans Volta Togoland was annexed from the Germans and brought under British control in 1916 thus increasing the size of the British colony of the Gold Coast (Haskett, 1981). One year prior to Ghana’s independence in 1957, the passing of a referendum in May 1956 added the British Trans Volta Togoland permanently to independent Ghana (Haskett, 1981). The name Ghana was chosen to commemorate the once powerful Ghana Empire that ruled most of West Africa for nearly a thousand years before disappearing in the 13th century. According to historians, Kumbi Saleh the capital of ancient Ghana Empire is located some hundred miles north of modern Bamarko (Edgerton, 1995, p.2). Present day Ghana is located on the West Coast of Africa (Figure 2.1.2), has a population of about 24,658,823 million people (Ghana Statistical Service, 2013) and occupies a land area of about 238,538 square kilometers. In terms of size Ghana is often compared to Oregon in the United States or the United Kingdom in Europe. Administratively, the country is divided into 10 regions with each region having a regional capital which also serves as the seat of government 25 for a government appointed regional Minister and his/her or her cabinet. Accra serves both as the national capital and the regional capital for Greater Accra Region. 26 Figure 2.1.2: Administrative Map of Ghana Highlighting Research Region 27 2.5 Agro Climatic Conditions in Ghana In terms of the agro climate, Ghana is a tropical country and experiences two distinct seasons across most parts of the country, a wet and a dry season. The wet season is characterized by major rainfall events starting in April and ending in July and a minor raining season starting September and ending in late October early November. Most parts of the Northern, Upper East and Upper West regions experience only one raining season beginning in April and ending in early October. The minor raining season is experienced mostly in the south most parts of the country from the Brong Ahafo Region southwards. The dry season also known as the Harmattan season is characterized by dry Sahara winds that blow from the Northern parts of the country to the South beginning in November or December and lasting until March. Harmattan winds typically lowers atmospheric humidity, increases the daytime temperature while lowering nighttime temperatures. The annual average temperature ranges between 21-32oC (70-90oF) while annual rainfall ranges from about 33 inches in parts of the North and 87 inches in parts of the South. Approximately 40 percent of the country’s labor force is engaged in subsistent rain- fed agricultural production. Appendix A below represents Ghana’s rainfall patterns for the different months from January through December. 28 2.6 Ghana’s Agro-ecological Zone (AEZ) and High Forest Zones (HFZ) In the first quarter of the 20th century, Ghana’s Agro Ecological Zone (AEZ) was divided into five major zones (Chipp, 1921) including the Savannah and Savannah Forests, Transition Zone (TZ), Deciduous Forest, Rainforest and Coastal Savannah. By the end of the 20th century, a substantial portion of the Guinea Savannah in the Northern part of the country had been replaced by the Sudan Savannah associated with extremely dry desert environments. Also, the Rainforest had shrunk significantly by more than 70% by the end of the 20th century. The HFZ occupies all of the Rainforest and Deciduous forest and sizable portions of the Guinea Savannah south of latitude 8o zone. Both the HFZ and Savannah Zones (SZ) are partitioned into reserved and unreserved forests but the HFZ alone contains more than 75% of all gazetted forest reserves in the country. The total land mass of the HFZ is estimated at 8.2 million Ha, approximately one- third of the country’s total land surface (Agyarko, 2008). Currently there are some 266 gazetted forest reserves throughout the country out of which 204 (1.6million Ha) are located in the HFZ alone and the remaining 62 reserves (0.6million Ha) located in the Savanna zone (Opoku, 2006 and Odoom, 2002). From the graph in figure 2.4 below it is obvious that both reserved and unreserved forests in the HFZ and SZ have not been spared by the massive rate of deforestation that occurred in the Ghana since the beginning of the 18th century. For example, several reports estimate that the total area of Ghana’s High Forest (HF) decreased from a baseline level of 8.2 million Ha to about 4.3 million Ha by 1948 (Agyarko, 2008), and further decreased to about 1.71 million Ha by 1990 (Odoom, 2002 and Opoku, 2006). Intact unreserved areas of timber resources in the HFZ, also known as Protected Timber Land (PTL), estimated at about 2.8 million Ha in 1948, decreased by about 98% to only 31,000 Ha in 1990 (Dadebo and Shinohara, 1999, p.2). Out of the 1.68 29 million hectares of reserved forests in the HFZ managed by the Forest Services Department (FSD) since 1990, an estimated 1.49 million Ha (84%) is said to display signs of partial to extremely degradation (Hawthorne and Abu-Juam 1993 and Odoom, 2002). Ghana’s forest zone covers approximately “two-thirds each of the Ashanti and Eastern Regions and one-third of the Brong Ahafo Region and one-eighth of the Volta Region (Ewusi, 1984b, p.18).” Trends of deforestation experienced in Ghana over the last century have thus been remarkably similar to global trends described by the FAO-GFAR and WRI reports. Figure 2.1.3 below describe Ghana’s Agro Ecological Zones (AEZ) while figures 2.1.4 and 2.1.5 present Ghana and Brong Ahafo Region (study region) land cover situation maps, respectively. 30 Figure 2.1.3: Agro-ecological Zones of Ghana 31 Figure 2.1.4: Land Cover Situation Map of Ghana 32 Figure 2.1.5: Land Cover Situation Map of Brong Ahafo Region 33 2.7 Traditional System of Governance and Land Tenure Administratively, Ghana is divided into 10 regions and has a total of 170 districts. The country is made up of eight ethno-linguistic groups of between 45 and 50 distinct dialects/languages (Dakubu, 1988 cited in Fridy, 2007 p.66-68). In the traditional system of governance, each district is sub-divide into traditional councils with each traditional council headed by a paramount chief. Traditional councils are subdivided into towns which are themselves made up of several villages. Each town is administered by a chief under whose authority are several sub-chiefs who control several villages within the township. By virtue of being first occupiers or having historically fought for and conquered the land, the chief of a town wields ultimate power over land within a township. The structure of Ghana’s land tenure system and the role of the chieftaincy system in defining ownership rights have undergone very little evolutionary change since the days preceding colonialism. Practically all land in Ghana falls under three main categories: 1) stool land, 2) tribal or family land and 3) individual ownership/property (Mends and De Meijere, 2006, Dzokoto and Opoku, 2010 and Hilhorst and Porchet, 2012). The quote below summarizes the importance of the “stool” in Ghanaian culture. “As a symbol, the stool is loaded with multifaceted meanings. Stool in Asante is the bedrock of religious observances by way of venerating the ancestors. The stool is the origin of kinship and the social basis of Ashanti world. The stool is the foundation of “Oman” (the Nation) and the political order of chieftaincy. This will ensure a clearer understanding and the utility of stools as both natural and supernatural objects. So important was the stool, especially, those that symbolize political office that people accorded it with high esteem. Before British colonization denuded the authority of such stools, they were at the apex of the political order, and society elevated those who held office very high in the communities they reigned (Kwakye-Nuako, 1999: 57).” Approximately 80% of all land including forest reserve and unreserved land in Ghana is held under the customary tenure system (Mends and De Meijere, 2006) and all forest lands are stool lands governed by customary tenure arrangements (Dzokoto and Opoku, 2010, p9). The 34 stool thus represents the seat of government in a township with the chief of the town exercising jurisdiction over all land under the stool (i.e., stool land). The chief is thus responsible for demarcating land within a town and designating areas suitable for different activities (residential, farming, hunting, collection of forest products, graveyards, religious ceremonies, or sacred grooves, etc.). Since stool lands are communal lands vested in a chief, tribes and families under the chief are often given access typically for subsistence production however these privileges do not necessarily convert into absolute ownership (Blocher, 2006 p179-180). Family or tribal lands are administered in nearly the same manner as stool lands only that land under this tenure system is the property of specific tribes, clans or resident in the village. Like stool land, all individuals within a family have rights to portions of family land for farming purposes although this also does not automatically convert to private ownership. Land under private ownership/property is usually acquired as a gift from chiefs or through sale of stool or family lands. While land is rarely sold outright, there may be instances where land may be sold outright for example to help defray arbitration costs associated with land disputes, or to raise funds for community development projects. Land under private property is that which is owned by individuals by virtue of purchase from the stool or family. In order to avoid total or complete alienation of land, the traditional system in Ghana often prohibits the outright or complete sale of land. In the place of outright sale, stool lands are usually leased on long term basis for duration stretching anywhere from fifty to ninety-nine years. The statement below explains the rationale behind the customary tenure system of land ownership in Ghana: “…….land belongs to a vast family of which many are dead, few are living and countless numbers are still unborn….the living must use it so that the interests of future generations are not jeopardized… each member of the community has a right to occupy 35 and use part of the land….and no individual could alienate these rights to another (Ollenu, 1962 cited in Mends and De Meijere, 2006, p.5). The above statement from Olenu (1962) suggest that while it is typical for land in Ghana to have “multiple owners with the chief holding the highest title, and numerous other rights- holders claiming lesser rights of possession, use, or transfer (Blocher, 2006 p179),” the land tenure regime attempts to safeguard total alienation of land (for in depth discussion on the different land tenure regimes in Ghana, see Sarpong, 2006). Section 2.8 below discusses pre- and post-colonial forest policies and how these policies in most instances fueled conflicts over land ownership and use. It is possible that misunderstanding of traditional tenure systems and how these systems may have helped preserve forests over several generations explain the deforestation that occurred during the pre- and post-colonial as well as post independent era. 2.8 Historical Background on Ghana’s Forest Policy The quotation below from Chips’ 1927 dissertation research in the Gold Coast, provide a vivid description of Ghana’s forests nearly two decades after the creation of the colony’s Forestry Department in 1909. According to Chips: “…the traveler, who tries to make his way through the forest, constantly finds his way barred by massive lianes and by great trees whose crowns of thick foliage are so densely interlocked as to almost shut out the rays of the tropical sun overhead. It is like an immense, almost impenetrable, vault supported by gigantic pillars and spreading a mysterious shade; lianes like enormous cable stretch from tree to tree and bind the crown firmly into an inseparable canopy. This wonderful interwoven canopy is a source of perplexity for, it is impossible to see the shape and development of the crowns of these forest giants which run straight up without a branch for 50 feet or more and stand towering over the canopy surface at a height of nearly 200 feet above the ground. It is impossible to obtain flowers or leaves from any particular tree-crown, for this sea-like forest canopy stretches apparently unending, with a depth of thickness of over 50 feet, and it is a long time before flowers and leaves, and often fruits, shed by the giants, penetrate this mass and arrive on the ground. Thus, baffled in his attempt to see through the opaque roof of the forest, a collector may, in desperation, determine upon the total destruction of a tree by cutting it down, in order to obtain the materials, he so much desire. It is then that the full effect of the lianes, great and small, and of the strangling 36 Figs is realized. First, one tree must be cut, and then another, and then a third, and before long the astonishing spectacle of a dozen or more trees cut right through near the bases yet remaining in position, almost dangling by their crown from the canopy-mesh woven inextricably by these band-like creepers. It is only when many trunks have been cut through does the canopy sag, and crash to the ground with a mighty roar, leaving the giants standing gaunt and erect, shorn of their massive limbs, which have been torn off in the general subsidence of the surrounding and supporting canopy. Such is the nature of the forest which extends as a coastal belt from the east of Sierra Leone to the middle of the Gold Coast and Ashanti, reaching a varying distance from the coast up to a maximum of about 250 miles (Chips, 1927: p18-20).” The formulation of forest policy and subsequent creation of a Forestry Department may be said to have emerged out of concerns for excessive deforestation and rapid decline in valuable timber species within the territory’s High Forest Zone (HFZ). Long before any contact with Europeans the people of the Gold Coast (now present-day Ghana) have for generations depended on their forests for resources such as timber and none-timber forest (NTFPs) products. Some researchers, for example Parren (1994) suggests that, among all forest products including timber, NTFPs were of the most importance to indigenous West African tribes and thus dominated intra- regional trade during the pre-colonial era. One possible reason for the importance of NTFPs relative to timber may be explained by the construction of indigenous West African houses. For several generations, indigenous West African houses were constructed primarily from mud/earth, straw and bamboo thus requiring very little timber for construction. Chipp’s (1927, p75) extensive dissertation work on the Gold Coast Forest in 1927, for example documents the importance of Borassies Flabellifer (a fan-leaved palm tree also called Borasus) in building construction in most parts of the Ashanti Colony (present day Ashanti and Brong Ahafo Regions). Dickson (1969) and Parren (1994), suggests that commercial logging and export of timber was initiated by Europeans sometime in the late 1700s or early 1800s. Owusu, (1993, p.14), chronicled six distinct phases of development in Ghana’s formal timber industry between 1887 and 1983. It is believed that establishment of Ghana’s formal timber industry was fist 37 preceded by trial exports in 1887 of some 240 superficial cubic feet of Mahogany (Khava and Entandrophragma spp.) to Britain (Owusu, 1993). These initial successful trials ushered Ghanaian timber onto the world market and thus set the stage for commercial timber extraction from Ghana’s High Forest Zone (see Wealth in Wood, 1950). One major development that ultimately paved the way for increased timber harvests and the rampant deforestation that ensued was the end of major tribal conflicts and wars within the Ashanti territory by the late 1800s and early 1900s (see Beckett, 2009, for details on Wolseley’s 1874 expedition resulting in the defeat of the Ashanti’s and siege of Kumasi). The British victory over the Ashanti’s in February of 1874 (Beckett, 2009) gave the British firm control over the Southern parts of Ghana and the also Ghana’s High Forest Zone (HFZ) which houses the bulk of the country’s timber. According to Owusu, (1993), timber exports increased dramatically from the few trials in 1887 to 450,000ft3 by 1894 thus establishing the country’s timber industry. Logging of timber in the closed forests and the deforestation that ensued paved the way for new agrarian communities to be established with a primary focus on cocoa production (Chipp, 1927, Owusu, 1993, and Parren and de Graaf, 1995). The massive rate of deforestation taking place on the Gold Coast and its territories in the late 19th century and early 20th century drove the one time Legos Governor Cornelius Alfred Moloney to publish a book describing the crisis and its detrimental effect on both the environment and livelihoods. Parts of Maloney’s (1887) extensive work on commercial forest products from West Africa during the colonial era, highlights the problem of massive deforestation and its impact on the coastal areas of the Gold territory. Maloney used the word “detimbering” to describe the senseless shaving of the forests of timber in areas immediately surrounding the coastal areas of Accra (Maloney, 1887: p.236 & 244)”. Maloney (1887) further argued that the process of deforestation resulted in a scarcity of firewood 38 and water supply in Accra particularly during the dry seasons. The following statements summarize Maloney’s concern about deforestation and the call for a government strategy that supposedly ensures sustained timber supply and forest conservation along the West African Coast: “It has been truly said that the there is a tendency in newly-settled countries to regard timber as a mere encumbrance to the land, and, as it generally occupies the most fertile soil, the finest timber is usually that selected for destruction by fire, by ring barking and other crude and wasteful methods……It is not withstanding a difficult matter, and one that might with reason be viewed as arbitrary, to put in force legislative restraints upon the clearing of the land in young colonies, and, so as to obviate resort to any legal machinery in the matter, people interested should not look on supinely while timber is disappearing and a country is gradually markedly becoming treeless, and apart from their economic advantage, bereft of the shade so essential to health of tropical climate (Maloney’s, 1887, p231).” As far as forest policy in West Africa and in the Gold Coast in particular, Maloney’s 1887 publication “Sketch of the Forests of West Africa” in one way or another shaped the debate on deforestation and provided the road map for future forest policy in the Gold Coast. First Maloney acknowledged the problem of reckless deforestation and its implications for future timber resource availability as well for the environment; and secondly, he identified the primary causes of deforestation and then provided policy guidelines for forest conservation. By 1874, several events including the rise in gold-mining and the increasing number of concessions granted, and associated land disputes prompted the then Governor of the Gold Coast to side with Maloney in an attempt to craft a solution to the land problems (Kimble, 1963, p.332). 2.8.1 Crown Land Ordinance of 1894 and the Public Land Bill of 1897 In less than a decade after Maloney’s famous publication, the British Government of the Gold Coast attempted unsuccessfully in 1894 and 1897 to pass what was deemed a controversial Crown Land Ordinance and the Public Land Bill respectively (Ashurst, Morris and Crisp, 1911 in Belfield 1912). Both laws if passed would have vested all “wasteland” and forest lands as 39 well as all mineral resources in the Crown (Moris, 1912, Parren, 1994), essentially doing away with the traditional land tenure system, a good part of which exists till this day (for more details on the debates and proceedings in the Gold Coast Legislative Council against both Land Bills see Omosini, 1972, 464-463). Some researchers have concluded that both the 1894 and 1897 bills were designed by the Colonial Government to “control native land in order to tap its resources” (Parren and de Graaf, 1995, p.35). It is argued that the Colonial Government’s insistent on passing the Crown Ordinance and Public Land Bills was because “whoever owned the land or secured the rights of its administration essentially dictates the method of its exploitation (Omosini, 1972, p.454).” 2.8.2 The Concession Ordinance of 1900 Two major developments in the Ashanti colony gave the colonial government greater access to Ghana’s tropical forest region which it otherwise would not have had. In February of 1874, a British led expedition commanded by Major General Brevet and Colonel Sir Garnet Wolseley fought and defeated the Ashanti’s and took siege of Kumasi the Ashanti capital on February 4th (Beckett, 2009, p.365). Approximately two decades following Wolseley’s expedition, Colonel Sir Francis Scott led the 1895 expedition that ceased Kumasi and exiled more than 60 members of the Ashanti royal family including the King Prempeh (Edgerton, 1995, and Boahen, 2003). Final British victory in September of 1900 resulted in the annexation of Kumasi and declaration of the entire Ashanti territory as a British protectorate (Kimble, 1963, p.264-333, and Beckett, 2009, p.374). Following the Colonial Government’s victories over the Asante’s and subsequent annexation of the entire Ashanti territory, the British gained access to vast resources including gold within the territory. One of the major catalysts of the 1900-1901 Yaa Asantewaa war between the British and the Asante’s came as a result of granting excessive 40 gold mining concessions particularly the “Ashanti Goldfields at Obuasi between October 1895 and December 1900” (Boahen, 2003, p.33). It has been suggested that the influx of European prospectors seeking concessions for the development of gold mines within the Ashanti territory started in 1874, and reached its peak following the final British victory over the Ashanti’s in 1895 (Kimble 1963, p331). Other researchers suggest that the result of numerous disputes and litigations associated with timber and gold concessions prompted the Concession Ordinance of 1900 albeit as a regulatory measure (Omosini, 1972). In a rare ruling, then Secretary of State for the Gold Coast Colonies Joseph Chamberlain sided with the natives and declared that the soils of the Gold Coast should indeed remain in native hands, thus ordering a revised version of the 1897 Public Lands Bill to be passed but serve only as a regulative measure (Moris, 1912, p.100 and 104). The revised 1897 bill thus became known as the Concession Ordinance of 1900 which gave the government power to grant concessions to mining companies albeit with all native ownership rights to the land remaining intact. The Concession Ordinance of 1900 allowed the alienation to an individual or company of at least five square miles of land for gold mining and 20 square miles for timber production (Moris, 1912, 107). Some researchers have argued that the vagueness of the Concession Ordinance of 1900 which did not stipulate the exact allocations that could be granted an individual prospector, gave the government indirect access to excessive forest lands since all concessions had to be certified and approved by the government in order to be legally binding on both parties (Owusu1993, p76). The following statement from Belfield (1912) reveals a major flaw in the Concession Ordinance that possibly increased the rate of deforestation within Ghana’s High Forest Zone: “…..when inquiry is made concerning the areas granted, the Chiefs and their advisors have no idea, even at the present time, what extent of their country is contained in a given 41 number of square miles. They always imagine that what they are granting is only a fraction of the area which they actually dispose of. Perceptions of the actual state of affairs do not dawn on them until survey has been completed and the matter has gone too far for alteration. I have been told by almost every chief who has granted concessions, that if he had known the extent of the country which was been sought from him, he would have refused to concede all that was asked for (Belfield 1912, p.9).” Testimonies presented by the native chiefs and their representatives to Belfield between February and March of 1912 suggest that most chiefs in charge of granting concession were non- lettered (i.e. could neither read nor write English) and also did not understand the European units of measurement used in demarcating land under the concession contract. In addition to being illiterate as far as reading and writing the English language, Belfield (1912) contends that the clerks who read and interpreted the deeds to the chiefs as well as the lawyers who formalized the documents were all hired and paid by the concessionaire hence were more mindful of the interests of their clients/applicants than those of the chiefs (Belfield 1912, p.10). It may be argued that in the wake of it becoming law, the Concession Ordinance of 1900 predisposed even larger portions of Ghana’s tropical forests to deforestation. The deforestation that followed immediately after 1900 may be attributed to two factors: 1) inability of native chiefs to comprehend the details of concession contract leading to larger than intended areas being conceded to foreign timber and mining firms, and 2) the failure of government to effectively protect the natives from concessionaires who preyed on their ignorance of the English language and their lack of understanding of the European measuring scales. The government’s response to the continuing deforestation was to institute new laws. 2.8.3 Timber Protection Ordinance of 1907 After almost seven years of implementing the Concession Ordinance of 1900, the rate of deforestation within the Gold Coast, Ashanti and other territories only seemed to get worse each 42 year thus prompting the government to pass yet another bill dubbed the “Timber Protection Ordinance of 1907”. The goal of the Timber Protection Ordinance of 1907 was essentially to prevent indiscriminate cutting of immature trees by limiting timber harvest to only mature trees of specified girth and width (Brown, 1911, in Belfield 1912, p.111). To enforce the Timber Protection Ordinance the government recruited Forestry Officers who’s first task in 1908 was to conduct a detailed survey of the forests in the territories (Parren and de Graff, 1995, p39). Following the 1908 survey, a Forestry Department was created in 1909 to administer all matters related to forestry in the Gold Coast (Kimble, 1963, p363). 2.8.4 Forest Ordinance of 1911 Approximately two years after creating a national Forestry Department in 1909, a highly contested Forest Ordinance was passed in 1911 giving the government power to declare parts of Gold Coast territorial forests as national reserves to be placed under tight government control. Native protest against the Forest Ordinance of 1911 prompted the British Crown to dispatch Special Commissioner Conway Belfield to the Gold Coast in 1912 to listen to testimonies and report on the situation in the Gold Coast territories. The native chiefs and their representatives argued that the 1911 Forest Bill was a reincarnate version of the Crown Land Ordinance and the Public Lands Bill of1894 and 1897 respectively that failed to achieve the very objectives the government was intent on achieving with the 1911 Ordinance. According to Ashurst, Morris and Crisp (1911 in Belfield, 1912, p.103), the Forest Ordinance of 1911 gives the government unwarranted powers to administer stool, family as well as tribal lands, the proceeds of which would be inequitably distributed with three-fifths going to the government, and two-fifths to landowners. Yet in the face of national protest, the government still maintained its position on the native’s destruction of forests and the need to rescue the situation by passing and 43 implementing the Forest Ordinance of 1911. The Government’s reasons for the new legislation prompted Brown (1911) in a deposition before Commissioner Belfield in 1912 to make the following statements: “If the desire is to introduce foreign capital into the country, which seems to be the likely reason, then there is no necessity for the Forest Ordinance. For under the 1900 Concession Ordinance agricultural rights and implied forestry rights can be acquired subject to amendments ….. If on the other hand the sole objective is to conserve the forests because of the alleged reckless deforestation on the part of natives, then it is worth pointing out, that those who are really deforesting the country are the mining companies, who indiscriminately cut timber for timbering their mines……. Forest conservation is not unknown to our people. To effectually prevent deforestation, the reservation is invariably associated with prohibitions, the willful infraction of which is believed to bring upon the wrong-doers, the anger of the tribal family, or national gods or the displeasure of the spirits of our forebears when the sanctity of their sleeping place is disturbed, besides penalties in the way of fines (Brown, 1911 in Belfield 1912, p11).” The new Forest Ordinance of 1911 supposedly would have given the government exclusive power to alienate any “unoccupied lands” that have not been inhabited or cultivated for at least a decade. The critical question that remained unresolved since the demise of the previous 1894 and 1897 Land Bills was the meaning of the term “abandoned or wasteland” introduced into both previous bills as a pretext to alienate native lands. Yet again a similar term “unoccupied land” seemed to have aroused lots of anger and suspicion among most natives and their chiefs. The traditional shifting cultivation system of agriculture practiced by most African societies is often characterized by alternating lag periods in which previously cultivated farmlands are left to fallow in some cases for a decade to enable natural regeneration of the soil. To recent settlers unfamiliar with the traditional system of shifting cultivation, all fallow lands are simply ‘wasteland and/or abandoned land.’ The natives argued that previous land transactions with the British were conducted under customary laws which respected shifting cultivation as the 44 prevailing system of agriculture, hence any new laws which would otherwise confiscate land were in clear violation of the same laws which the British had previously recognized. In a little over a decade since the failed attempt to alienate native lands, the term “wasteland” had metamorphosed and reappeared in other modified bills taking on new names such as ‘abandoned land,’ ‘unoccupied land’ and ‘undeveloped land.’ In his1912 report presented to the British House of Parliament under the title: “Legislation Governing the Alienation of Native Lands in the Gold Coast Colony and Ashanti…,” Commissioner Belfield (1912), suggested that the indigenes’ objections to the term “unoccupied land” in the Forest Ordinance of the 1911, was on the ground that the term implied the existence of land over which no rights of ownership or occupancy is claimed. Belfield suggested that future bills adopt the term “undeveloped land” in place of “unoccupied land” so as to avoid or at least minimize any resistance from the native chiefs (Belfield, 1912, p.39). Native opposition and agitation over the Forest Ordinance of 1911 rendered it dormant for almost two decades only to re-emerge in a rather potent form in 1927 leading to the creation of some 266 National Forest Reserves and Parks within the boundaries of present-day Ghana (figures 2.1.6) and almost 30 in the Brong Ahafo region alone (figure 2.1.7 below). 45 Figure 2.1.6: Map of Ghana’s Reserves and National Parks 46 Figure 2.1.7: Map of Brong Ahafo’s Forest Reserves and National Parks 47 2.8.5 The Forest Ordinance of 1927 After two failed attempts first in 1894 and second in 1897 to pass the Crown Ordinance and the Public Land Bills respectively, the British colonial government of the Gold Coast finally passed a Forest Ordinance in 1911. As with most forest policies in the Gold Coast and present- day Ghana, passing the Forest Ordinance of 1911 was one thing and enforcement was another. Just like the 1894 and 1897 bills, native resistance against the Forest Ordinance of 1911 made it impossible for the British to enforce this law. Hence for the next 16 years following 1911 the British adopted new strategies meant to allay public fears and suspicions towards future forest and land related regulations (see Belfield’s 1912 report on the “Legislation Governing the Alienation of Native Lands in the Gold Coast and Ashanti.)” In 1927, another Forest Ordinance was passed containing specific language that appealed to the Natives Authorities. De Grassi (2003, p.4), asserts that by recognizing that reservations would in no way affect native land ownership rights, the Forest Ordinance of 1927 allowed the colonial government to regain the native’s trust in establishing forest reserves. This new strategy adopted by the government to preserve the status quo as far as land ownership rights seemed to have paid off, for Wardell (2005, p.18) points out that when the Forestry Ordinance of 1927 was read for the first time in 1926, it was met with very little to no resistance from the native chiefs and their representatives. The Forest Ordinance of 1927 gave Native Authorities certain powers some of which charged them with establishing and protecting forest reserves. What the Natives Authorities probably overlooked however was section 4(4) of the same 1927 Forest Ordinance that gave the government power to over-ride the powers of the Native Authorities should they fail to constitute forest reserves (for specific details of the Forestry Ordinance Section 4(4) see de Grassi, 2003, p4). Hence like its 1911 prototype, the Forestry Ordinance of 1927, allowed the government to 48 achieve its aim of establishing forest reserves throughout most of the Gold Coast and Ashanti colonies, only this time, with the local authorities leading the charge. Some researchers argue that the Native Authorities also had a vested interest in constituting reserves due to revenues generated from royalties (Amanor, 1999, p.51) while others speculate that cooperation was largely a pre-emptive strategy to avoid any new or future restrictive land legislations (de Grassi, 2003, p.7). By 1940, the Native Authorities in collaboration with the Forestry Department established 214 forest reserves covering an area of over 15,000 square kilometer in the Gold Coast and Ashanti territories and 160 square kilometers in the Northern Territories (Wardell, 2005, p.11). The fact that the Native Authorities through their own initiative established more than 50% (127) of all reserves within the colonies is a testament to the effectiveness of the 1927 Forest Ordinance (de Grassi, 2003, p.7). By 1945, strains in the relationship between the Native Authorities and Forestry Department had become visible. In a 1945 report on Forestry in the Northern Territories of the Gold Coast, the Chief Conservator of Forests R.C Marshall observed that the Native Authorities were either not “capable or willing to administer Forest Reserves under their jurisdiction” (Marshall, 1945, p.5). Marshall’s comments stem from previous reports suggesting that squatters had taken over and established farms on some reserves while a passive Native Authorities acquiesced. The following statements from Marshall echoed sentiments among some in the Colonial Forestry Departments regarding the effectiveness of the Native Authorities in protecting newly established forest reserves: “Negligence on the part of the Native Authority staff and apathy on the part of the Native Authority would appear to preclude the possibility of administering Forest Reserves properly. The remedy is an adequate and impartial staff. It is too much to expect a Native Authority Official to act contrary to the wishes of his employers and it must, in this connection, be borne in mind that the Forest Reserves are no more popular in the Northern Territories than elsewhere. If Reserves are to be Reserves in the acceptable 49 sense of the term, Government Rangers and Government Forest Guards should control them (Marshall, 1945, p.5).” 2.8.6 The 1948 Forest Policy Marshall’s concerns regarding the challenges involved in collaborating with the Native Authorities and his recommendations to “fixing the problem” appear to have resonated with the colonial administration who responded by passing into law the 1948 Forest Policy. The new forestry law essentially abandoned all local cooperation and stripped the Native Authorities of their administrative powers over forest reserves. Some researchers have suggested that the 1948 Forest Policy was driven in part by the colonial government’s desire to exploit timber to satisfy the ever-increasing export demand (Boakye and Baffoe, 2006). As figure 2.1.8 below suggest, timber exports begun to increase rapidly and reached a peak of some 7 million cubic feet soon after all 264 national forest reserves were created in the Gold Coast, Ashanti and Northern Territories. 50 Figure 2.1.8: Ghana’s Timber Exports (1938-1948) Ghana's Timber Exports (x1000) (1938-1948) Quantity in Cubic Feet (x1000) Value in British Pounds (x1000) 3000 2500 2000 1500 £ 1000 500 0 3 t F 8000 7000 6000 5000 4000 3000 2000 1000 0 1938 1940 1942 1944 1946 1948 Source of Data: Wealth in Wood, 1950. Some authorities on the subject have suggested that increase demand for timber especially during World War II accounts for the trend observed in figure 2.8. For example, Owusu claims that to support the Western war effort, the “Forestry Department organized the supply of forest products for use by British and American armies based in the Gold Coast (Owusu, 1993, p.20-21).” Also, by 1948, it appeared the Native Authorities’ involvement in Forest Management posed a threat to sustained timber exports hence the decision to revert to a more centralized management regime as stipulated by the 1948 Forest Policy. By 1948, several scientific studies, reports, manuals, and forest handbooks had been produced detailing the nature and extent of the Gold Coast’s forests (Chips, 1922 and 1927), tree composition (Chalk et al, 1933) and silvicultural practices (Kinlock, 1945) as well as threats to forests from various sources (Maloney, 1887 and Thompson, 1910, Marshal, 1945). 51 Increased scientific knowledge resulting from information gathered over the decades, gave the colonial government considerable leverage in matters concerning forest reserve management and exploitation. For example, part of the new 1948 Forest Policy also gave the Forestry Department the mandate to create Permanent Forest Estates for the purpose of protecting the environment. These Permanent Forest Estates it was argued will among other things, ensure the protection of water supplies, maintain agro-climatic condition, as well as serve as a resource for educating the public on forestry and resource conservation (Boakye and Baffoe, 2006). In spite of what many have come to suspect as a plot to exploit timber, part of the colonial government’s stated rational for centralized management of forest reserves in 1948 seemed well intentioned. However, table 2.8.1 appears to support critiques’ claim about the real motives behind the 1948 forest policy. Table 2.8.1 suggests that the 11 years preceding 1948 (i.e., 1937- 1947), only seven timber species were exported and even that only Mahogany made it onto the export roaster for all the 11years. Some species such as Odum had reached such low levels as to warrant a total ban on further exports by 1947 (Wealth in Wood, 1950). However, the very year the 1948 Forestry Ordinance was passed, all 16 commercial timber species with the exception of Odum, made it onto the export roaster (Table 2.8.1). Figure 2.1.9 below suggests that the 1948 Forest Policy drastically increased timber exports to an all-time high, exceeding seven million cubic feet and bringing in revenue of over 2.5 million pounds (Wealth in Wood, 1950). While timber extraction and export were focused primarily on sixteen species, Mahogany alone accounted for more than 95% of all log exports between 1937 and 1944 (figure 2.1.9). Mahogany’s contributions to total exports eventually declined to 65% not due to a decline in its exports volume but due to a tremendous increase in the export volumes of other species. For example, while Mahogany exports quadrupled (i.e. from 52 approximately 1 million cubic feet in 1937 to approximately 4 million cubic feet), other species such as Sapele and Wawa increased by more than 300 folds. It is worth noting also that until the 1948, only seven species (Mahogany, Odum, Walnut, Mansonia, Wawa, Sapele and Baku) were commercially extracted for exports. 53 Table 2.8.1: Timber (Log) Exports in Ft3 (x1000): 1937-1948 Year No Species 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 Mahogany 1 Odum 2 Walnut 3 Mansonia 4 Ofram 5 Avodire 6 Danta 7 Dahoma 8 9 Edinam 10 Guarea Emeri 11 12 Baku 13 Wawa 14 Kusia 15 Sapele 16 Utile Source: Wealth in Wood, (1950, p.16) 1034 652 502 7 23 18 - - - - - - - - 5 - - - 9 4 2 - - - - - - - - - - - 6 - - - - - - - - - - 5 - 2 - 944 2000 - - - - - - - - - - - - 1 - 1079 1403 2122 2487 2514 3287 3 - - - - - - - - - - - - 2 - 1 - - - - - - - - - - - - 59 - 2 - - - - - - - - - - - - 37 - 2 - - - - - - - - - - 3 - 11 - 3 - - - - - - - - - - 9 - 17 - 1 21 2 - - - - - - - 17 18 308 - 3468 3976 12 6 - - - - - - - 115 237 887 - 5 75 35 96 7 165 97 54 67 282 442 27 657 67 54 Figure 2.1.9: Timber (Log) Exports in Ft3 (x1000): 1937-1948 Timber Exports (x1000 Ft3) and Contribution from Mahogany 1937-1948 Mahogany (x1000) All Other Species (x1000) Mahogany (%) 100 90 80 70 60 50 40 30 20 10 0 % 3 t F 4500 4000 3500 3000 2500 2000 1500 1000 500 0 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 Source of Data: Wealth in Wood, 1950. 2.8.7 Aftermath of the 1948 Forest Policy By 1948, all 266 forest reserves had been established and managed by the Forestry Department. It is abundantly clear that the Forestry Department did not have adequate manpower to effectively protect valuable timber species within vast tracts of forests brought under the reserved system. It is possible that the traditional authorities out of fear of the repercussions of local encroachment may have voluntarily protected forest reserves under their jurisdiction from local encroachment. With some government and possibly local protection of reserved forests, areas designated unreserved forests became the primary source of timber for local communities, timber firms and foreign concession-holders. In the 1950 publication by the public relations department of the British colonial government, the forestry department portrayed a dooms day scenario for unreserved forests and thus ordered complete removal of all valuable timber species within those forests. For unreserved forests, the forestry department demarcated and categorized 55 areas likely to be farmed in the next decade and those not likely to be heavily farmed within the same period (Wealth in Wood, 1950, p.8). In areas suspected to be farmed in the next decade, a clear instruction was given to concession-holders to “cut and completely remove all valuable timber as quickly as possible (Wealth in Wood, 1950, p. 8).” It is clear how such an order may have orchestrated a “commons tragedy” resulting in a race among concessionaires and villagers to completely decimate unprotected forests. The result of this “rat race” is in part portrayed in figure 2.10 below which compares the rate of deforestation in Ghana’s reserved and unreserved forests between 1948 and 1990. The forgoing discussion and figure 2.10 appear to support Asante’s (2005, p.176) claim that pre- and post-colonial era forest policies are plagued with “conflicts and contradictions.” Asante (2005) asserts that these contradictions in forest policy implementation are to blame for the conflicts that have come to dominate Ghana’s forest policies even after independence. Hence in spite of several forest policies aimed at rescuing Ghana’s forests, deforestation continued and even spread to the protected reserves including Yaya, Nsemre and Sawsaw. 56 Figure 2.1.10: Change in Ghana’s Forest Cover between 1948 and 1990 Change in Ghana's High Forest Cover From 1948-1990 Reserved Forest Unreserved Forests Total Forest Cover 5000 4500 4000 3500 3000 2500 2000 1500 1000 500 0 ) 0 0 0 1 x ( a H n i a e r A 1948 1952 1955 1962 1965 1969 1972 1975 1978 1981 1984 1987 1989 1990 Source of Data: Forestry Department annual reports (1949 - 1994) cited in Dadebo and Shinohara (1999, p.2) The colonial era policy formulation and implementation according to Asante (2005, p.176) “have remained an enduring part of the forest policy process in post-independence Ghana.” To further strengthen Asante’s assertions, Owusu (2010) claims that a good part of the economic policies under independent Ghana only served to solidify and institutionalize state control over resources. Hence like her colonial master, independent Ghana instituted centralized forest policies that guaranteed cheap access to natural resources on behalf of the corporate sector and transnational agencies (Owusu, 2010, p. 263). In the sections that follow, an attempt is made to establish linkages between Ghana’s post-independent economic development and policies that favor unsustainable forest resource extraction. 2.9 Ghana’s Economy after Independence (1957) Ghana gained her independence from the British on March 6th, 1957. Being the first sub- Saharan African country to gain its independence, all seemed well for the newborn country as 57 per capita income increased steadily from $354 in 1950 to $427 in 1975 nearly doubling the average for Africa (Ewusi, 1989). Ghana thus became one of a few African countries to attain middle income status at the time of independence (UNIDO, 1986). According to May and Schoone (1989) Ghana’s Central Bank Reserve at independence was in excess of 17 months of imports and its human resource endowments as measured by the level of education was the highest of any sub-Saharan African country. The abundance of natural resources including gold, diamond, manganese, bauxite, timber, and cocoa undoubtedly made Ghana one of the richest countries in Africa and the world. Figure 2.1.11 below, illustrates Ghana’s share of world cocoa, manganese, diamond, and gold outputs in 1950. As shown in figure 2.1.11, Ghana was the largest exporter of cocoa (36%), second largest for manganese (24%), third largest for diamond (8%) and fifth for gold (3%) in 1950 making the country a very important player on the global market. Figure 2.1.11: Ghana’s Ranking and Percentage Shares of World Exports in 1950 Ghana's Ranking and % Share of World Exports in 1950 % 40 35 30 25 20 15 10 5 0 1 2 Share of World Exports in % Rank 3 5 Cocoa Manganese Diamond Gold Timber Source of Data: Ewusi, 1989 58 2.9.1 Ghana’s Cocoa Economy (1885-2008) Of all primary export commodities, cocoa however wielded the most influence in the Ghanaian economy, partly owing to its contributions to GDP and employment and also its impact on other productive sectors including agriculture and forestry. Throughout the 1950s, 1960s and the first half of the 1970s, Ghana’s cocoa industry alone accounted for approximately 60% of GDP and foreign exchange earnings, 40% of internal government revenues and employed at least 25% of the labor force (Nyanteng, 1980). Sarris and Sham (1991) describe Ghana’s foreign reserve situation to be very healthy as a result of cocoa exports. For a good part of the 20th century, cocoa remained the single most important cash crop in Ghana’s economy with production and exports playing a central role in the country’s economic development. Following trends in cocoa output, GDP per annum for example increased from 1.8% to 3.5% between the periods 1891-1901 and 1901-1910 respectively during the colonial era (Ewusi, 1989). Similarly, both GDP and agricultural outputs grew even faster during the post-colonial era reaching an annual growth rate of 4.1% between 1950 and 1960, thus stimulating a similar trend of 4.3% growth in agricultural output for the same period (Sarris and Sham, 1991). Sarris and Sham, (1991, p.1) claim that due to its high growth, cocoa exports alone provided the basis for rising investment. Hence most of the key large-scale infrastructural development projects commissioned since the 1920s particularly in the transportation, health and educational sectors were all financed primarily with export revenue from cocoa (see Ewusi, 1989 for a detailed account of specific projects). The rise of cocoa and favorable cocoa policies may be said to have had negative impacts on forest resources since large portion of the high forest zone had to be cleared to make way for cocoa plantations. As shown in figure 2.12 below, Ghana’s cocoa production started at a meager 59 2000 tons in 1900, climbed quickly to 200,000 tons in 1925 and then peaked at 585,000 tons in 1965 before plunging steeply to less than 160,000 tons in 1983. Figure 2.1.12: Ghana’s cocoa outputs (1970-1983) Ghana's Cocoa Production in Metric Tons (x1000) (1885-2008) 0 0 0 1 X s n o T c i r t e M 800 700 600 500 400 300 200 100 0 Source of Data: Gill and Duffus Group (n.d) cited in Kolavali and Vigneri, 2011, p.202 By 1983, Ghana’s cocoa outputs had declined by approximately 73% relative to 1965 levels (figure 2.1.12) thus setting into motion a series of downward trends characterized for example by a 14% decline in agricultural value added, 43% decline in manufacturing, 17% decline in mining, and 37% decline in construction (Ewusi, 1989). Several reasons have been cited for the downturn in Ghana’s economy leading to a near collapse by 1983. For cocoa, a combination of factors both domestic and international helps explain production and export trends in the late 1970s and early 1980’s (see Nyanteng, 1980, p72- 78 and UNIDO, 1986). Other researchers attribute part of the decline in Ghana’s cocoa production and associated decline in economic growth in the 1970s and 1980s to other world events (see May and Schoone, 60 1989). Due to its immense contributions to GDP per capita and growth, any decline in cocoa production and exports is registered in GDP figures. Hence declining cocoa outputs throughout most of the 1970 leading to 1983, partly explains the 32% decline in per capita income by 1982 (Ewusi, 1983), and a 30% fall in per capita GDP by 1983 (Ewusi, 1989). By 1983 literally all sectors of Ghana’s economy were in a free fall, thus prompting the call for an Economic Recovery Program (ERP) and adoption of International Monetary Fund (IMF) and World Bank Structural Adjustment Programs (SAP) with further devastating results for forest resources. 2.10 Structural Adjustment Programs (1983-2000) Structural Adjustment Programs have been described as a set of market-based economic policies implemented by the IMF and World Bank between the 1970s and 2000s to help revive ailing developing economies (for detailed historical perspective on SAP see Owusu, 1998, 2001 and Olukoshi, 1999). Between 1970 and 2000 most of the developing world particularly Sub- Saharan Africa experience grave economic crisis. According to Ibhawoh (1999) the global economic recession and collapse of world commodity prices in the late 1970s and early 1980s disproportionately affected Africa. Hence in spite of Africa’s vast natural resource endowments, the continent still experienced an average decline of 15% in GDP between 1977 and 1985 (Ibhawoh, 1999) and “20 of the 31 least developed countries world-wide resided in Africa (OAU, 2002, p. 31-32)”. The economic situation in Africa by 1980 was thus described as so desperate that by 1980 as many as 24 African countries adopted some form of SAP (Ibhawoh, 1999) and by the mid-1980s every Sub-Saharan African country with the exception of South- Africa, Botswana, and Namibia had adopted some form of SAP (Olukoshi, 1999, Hilson, 2004). To the IMF and the World Bank Sub-Saharan Africa’s economic crisis of the 1970s and 1980s was largely a result of excessive government intervention in domestic economic activities. 61 Hence SAPs were prescribed as a means to allow the invincible hand of the free market to restore any structural deformities or imbalances within these ailing economies (Owusu, 1998 and Olukoshi, 1999). The belief was that the currencies of adjusting countries were over-valued, and their governments interfered excessively and often times unnecessarily in economic processes (Owusu, 1998, p.193). Shah (2010) explains that in return for World Bank and IMF SAP loans, adjusting countries agreed to implement a set of neo-liberal economic agenda as a precondition for adjustment loans. Some of the preconditions of SAP loans listed by Shah include: “a) cutback to government subsidies to agriculture, education, health, as well as other social and environmental support programs, b) economic liberalization with a heavy focus on resource extraction, c) reduction in state interference in the domestic economy by scaling back protection of domestic industries (i.e. removal of tariffs), d) devaluation of local currencies, e) elimination of food subsidies and f) opening up the economy to foreign investors (Shah, 2010).” The World Bank and IMF policies such as currency devaluation were intended to help adjusting countries attract foreign investment (as goods and services become cheaper) and discourage the import of expensive foreign goods. By devaluing their currencies struggling economies in the long-run are able to stimulate growth of their local manufacturing sector, improve capital utilization and overall unemployment situation. Some researchers charge that a counterproductive policy to the World Bank and IMF currency devaluation mandate was making a large percentage of the Sector Adjustment Loans (SECAL) loan approval under SAP subject to the purchase of relatively expensive equipment and services from the West (Owusu 1998). Another downside of SAP policy was the emphasis on export of raw materials/commodities instead of semi-processed or finished products. By not engaging in value added processes, the process of technology transfer was delayed in adjusting countries thus further prolonging 62 continued dependence on finished products from the West (Robbins, 1999). Owusu, (2010, p.261) charge that Ghana’s quest to sustain timber export volumes at all cost during the SAP era adversely affected the traditional forward linkage between the sawmills and the downstream processors thus resulting in serious lumber shortages on the domestic market which were supplied by illicit chain saw operators. The very policies of unequal trade implemented under SAP resembles those implemented under Africa’s colonial administrations resulting in the creation of impoverished dependent economies all throughout the African continent (Robbins, 1999). 2.10.1 Impact of SAP Loans on Adjusting Countries The two popular IMF and World Bank SAP loan facilities were the Enhanced Structural Adjustment Facility (ESAF) and the Structural Adjustment Facility (SAF) both established in 1986 and 1987 respectively. Niaman and Wartkins (1999) claim countries that implemented ESAF, SAF and other SIMILAR SAP loan programs generally experienced lower economic growth than countries that did not subscribe to such loans. Niaman and Wartkins (1999) further observed that in general African countries that received ESAF and SAF loans fared even worse as their total external debt as a share of GNP increased from 58% in 1988 to 70% in 1996 as compared to other countries’ that increased from 71.1% to 87.8%, between 1985 and 1995. Other researchers and organizations are also critical of SAP mostly because full implementation meant adjusting countries literally were compelled to empty their stock of natural resources so rapidly as to cause widespread social and environmental crisis within the adjustment countries (Owusu, 2006). At the core of the critics’ arguments is that SAPs in addition to restricting funding for critical welfare programs such as education and health (Olukoshi, 1999) also inadvertently favored increased production and exports of selected unprocessed natural resources (cocoa, 63 coffee, tea, timber and gold) simultaneously among several adjusting countries. The IMF and the World Bank often ensure that their prescribed economic liberalization policies take precedence over the adjusting country’s social and environmental policies (Tockman, 2002) by making disbursement of critical SAP loans contingent on compliance to adjustment policies (Olukoshi, 2000). For example, by relaxing mining and forestry laws and encouraging mostly log (i.e. unprocessed timber) exports come at huge environmental and economic costs to adjusting countries. Tockman (2002), further explain that “deforestation among 41 of the world’s most heavily indebted countries greatly exceeded the global average between 1990 and 1995 and ironically, during that time frame 70% of those countries adopted some form of IMF’s SAP.” By simultaneously saturating export markets with similar unprocessed forest products (e.g. timber, minerals, cocoa) without a corresponding increase in demand eventually caused prices of such exports from SAP countries to crash (George, 1990). Over time as the prices of primary exports from SAP countries crashed adjusting countries were forced to increase the volume of subsequent exports in order to generate sufficient foreign capital to service SAP loans thus ultimately trapping themselves in a cycle of debt. Another downside of SAP attributed to the export of mostly unprocessed natural resources instead of semi-processed or finished products was that the process of technology transfer (commonly associated with research and development) in the adjusting countries was delayed thus further prolonging continued dependence on expensive finished products from relatively developed countries mostly in the West (Owusu, 2006). Robbins (1999) charge that by discouraging value addition and promoting unequal trade/exchange SAP, like colonial policies created impoverished dependent economies throughout the African continent (Robbins, 1999). While specializing in the production and export of selected commodities such as timber, cocoa, coffee, tea, and gold, generated much 64 needed revenue for rapid economic recovery; Robbins (1999) suggests that foreign export earnings from unprocessed raw materials were often lost in the import of processed goods valued more than unprocessed commodities. This situation over time thus resembled a form of visual cycle within which poor indebted countries become trapped. Buyers from rich countries in the West enjoy cheap abundant resource supply while SAP countries became stuck in degraded environments and poverty. 2.10.2 Ghana’s Economic Recovery Program (ERP) and SAP Adoption in 1983 Precious minerals played a prominent role in Ghana’s economy long before colonization and decades following colonial rule. For example, Ghana’s ranked among the top five producers for gold, diamond and manganese in the first three quarters of the 1900s (see figure 2.1.11 above) however between 1980 and 1984 precious mineral outputs dropped significantly (see figure 2.1.13 below) with the lowest average decline of 55% occurring in 1983 (Ewusi, 1983). According to Benneh (1984) in 1935 there were 35 vibrant gold mines in Ghana however by the first half of the 1960s most gold mines appeared to have been exhausted and by 1983 only four remained viable. Benneh (1984) further explains that a 1981 government report claimed Ghana had sufficient gold reserves to produce over two million ounces of gold annually for the next 740 year however a 20-year capital investment of some $3 billion to establish these new mines (Benneh, 1984, p14). Like cocoa precious minerals were not immune to the economic crisis of the 1970s and 1980s. While it is often suggested that trends in most sectors of the Ghanaian economy are tied in one way or another to trends in cocoa production and exports, the decline in mineral exports particularly gold by 1983 was probably the result of a much serious problem. 65 Figure 2.1.13: Ghana’s Mineral Outputs under SAP Ghana's Major Mineral Outputs x 10000 Ounces (1980-2002) Gold (Ounces) Diamond (Carats) 300 250 200 150 100 50 0 Source of Data: Hilson (2004, p.65) Some researchers suggest that Ghana’s economic crisis between 1970 and 1990 leading to the adoption of SAP in 1983 is somewhat rooted in the country’s adherence to post-colonial policies that guaranteed cheap access to natural resources on behalf of the corporate sector and transnational agencies (Owusu, 2010). Clearly, the recovery of the mining sector hinged on capital investments without which old mines could not be resuscitated and new one built. In a study that examined the impact of mining sector reforms under Ghana’s SAP on micro and macroeconomic activities, Hilson (2004) revealed that Ghana’s mining sector experienced a 500% increase in gold production, 300% increase in Diamond and 600% increase in bauxite production during the SAP era between 1983 and 2004 (see figure 2.13 above). On the surface Ghana’s ERP may be touted by SAP proponents as rescuing an otherwise ailing mining sector on the verge of collapse, however mining sector reforms under SAP were not without social and environmental costs. Owusu (2006) suggests that “Ghana’s ERP like similar programs in other countries’ resulted in excessive exploitation of the country’s natural resources with huge 66 environmental costs. The downside of Ghana’s ERP was that “Ghana’s quest to sustain timber export volumes at all cost adversely affected the traditional forward linkage between the sawmills and the downstream processors resulting in serious lumber shortages on the domestic market which were supplied by illicit chain saw operators (Owusu, 2010, p.261).” Other literature on the performance of Ghana’s mining sector under SAP also suggests that since the adoption of trade liberalization, Ghana’s mining sector witnessed increased foreign capital investment without a corresponding increase in environmental protection to save guard the forest resources and the livelihoods of forest communities. By promoting policies that privatized industry benefits while socializing the environmental costs associated with mining, Ghana’s SAP literally decimated natural forests and laid waste large tracts of potentially viable agricultural lands. For example, Hilson (2004) observed that during Ghana’s SAP era widespread environmental problems stemming from land degradation and contamination from chemicals was commonplace and associated with the rapidly expanding mining sector. The quote from Awudi (2002) succinctly describes some of the environmental costs associated with mining during Ghana’s SAP era: “The surface mining concessions have taken 70% of the total land in the Tarkwa area alone. Rich vegetation has been cleared; ridges have been targeted and mined from top to bottom through a series of benches. The huge scales of excavation have led to a complete change of landform unsuitable for agricultural and any other livelihood activity. Huge craters have been formed and hillsides and parts of mountains removed, affecting the sources of many rivers and streams, and also causing deforestation. On the other hand, heaps of mine waste have been dumped and often occupy large areas of land and disfigure the landscape with attendant massive rainfall-based water pollution, given the location of Ghana’s mining industry in the South West part of the country where annual rainfall exceeds 1500 mm. Surprisingly, mining companies only pay lip service to rehabilitation of lands they have degraded as it ought to be under Ghana’s EIA requirement Awudi (2002, p.7 cited in Peprah, 2008, p.3). 67 2.10.3 Changes in Ghana’s forest cover under SAP era (1983-2003) Many scholars trace the excessive losses in Ghana’s forest cover since the 1900s and particularly during the SAP era to ineffective forest policies that favored large-scale mining and timber concessionaires. Also, the inability of forest policy to guarantee secure tenure rights to forest communities and landowners undoubtedly created open access conditions that led to neglect and indiscriminate plunder of forest resources in both reserved and unreserved forests. Figure 2.14 below suggests that the area of Ghana’s forests outside of national forest reserves declined steadily between 1948 and 1990 reaching a minimum between 1984 and 1991 a period in which Ghana’s SAP was in full swing. Figure 2.1.14: Changes in Ghana’s Unreserved Forest Area (1948-1990) Changes in Ghana's Unreserved Forest Area (1948-1990) Unreserved Forests 3000 2500 2000 1500 1000 500 0 ) 0 0 0 1 x ( a H n i a e r A 1948 1952 1955 1962 1965 1969 1972 1975 1978 1981 1984 1987 1989 1990 Source of Data: Forestry Department annual reports (1949 to 1994) cited in Dadebo and Shinohara (1999, p.2) In a recent landcover change analysis using “Image Differencing of Landsat TM,” Kusimi (2008) identified population growth, logging, mining and socio-economic and cultural practices as the major drivers of landcover change in Ghana’s forests between 1986 and 2002. 68 Kusimi (2008) further explain that logging and mining were the major factors that influenced landcover change in primary forests while illegal logging and mining activities, farming and rapidly expanding settlements accounted for the reduction in the area of secondary forests during the same SAP period (1986-2002). Kusimi (2008) estimates that primary forest reduced by 11% (from 88% to 69%) while secondary forests reduced 6% (from 9 to 3%). The rapid rate of deforestation in Ghana during the SAP era may be attributed directly or indirectly to the program’s loose mining and concession laws. In another study that investigated the impact of SAP on forest and biodiversity loss in Ghana, Benhin and Barbier (2004), found that cocoa and mining but not maize or staple crop production accounted for significant losses in forest cover during the SAP period. According to the authors, trade liberalization and tax incentives for foreign investors under SAP created an attractive investment climate for multinationals thus bringing large portions Ghana’s mining sector under the control of large-scale mining companies most of whose activities generated huge environmental costs relative to their contributions to GDP (Benhin and Barbier, 2003). Tockman (2002) also suggests that relaxed mining laws in Ghana during the SAP era decimated the country’s tropical rainforests and forest communities between 1989 and 1990. Tockman (2002) claims that under Ghana’ SAP, new mining laws devoid of environmental protection measures ceded 58,167square kilometers of forest land to some 250 mining companies enabled the exploitation of already vulnerable forests ecosystems. Specifically, Tockman (2002) suggests during the SAP era approximately 60% of rainforests in Ghana’s Wassa District in the Western Region was destroyed by mining activities which contaminated both surface and ground water sources with cyanide and other chemical. 69 2.10.4 Ghana’s forest-cover after SAP Since the end of Ghana’s SAP around the mid-2000s several studies and analytical works have emerged that propose creative ways of managing and internalizing the social and environmental costs associated with timber and precious minerals extraction from the country’s forests. Some advocate giving forest communities secure tenure rights to their resources as a means of encouraging greater community involvement in managing forest resources while others suggest amendments to the tax code on exported timber and mineral resources. Promoting timber and mineral export certification programs as well as media campaigns against the ills of illegal mining and logging have also been cited as possible strategies for curbing excessive forest resource extraction and associated social and environmental costs. 2.11 Secure Tenure Rights and Community involvement in Forest Management Effective management of forests in Ghana to a large extent requires that local communities have secure tenure rights to their forests. Lack of secure tenure exacerbates the open access problem and leads communities to decimate their natural resource or simply acquiesce while others do. The statement below (from a farmer in Sewiah community one of the Yaya Reserve communities included in this research) paints a vivid picture of the realities of the open access problem within Ghana’s forests. “A few months ago, I caught someone on my cocoa farm trying to cut down the trees on the farm. This person admitted he was going to sell the trees. I had so many Odum trees on my farm but every now and then these illegal chainsaw operators would sneak onto my farm and cut down the trees and in the process destroy my cocoa trees and other crops. Have you gone to see my cocoa farm? If you did you would understand what I just described to you. Having trees on my farm always keeps me on edge. If I don’t cut them someone else would and, in the process, destroy my farm. So as is often the case I don’t benefit from the trees, yet I lose my crops because of these same trees. This time around I took matters into my own hands and contracted a chainsaw operator to cut down this tree (see Appendix B). I know I will lose some of my crops but at least I will also benefit from selling the wood from this tree. I did not obtain a permit from the Forestry Department which I know I should…for that I am sorry (Farmer from Sewiah, 2008).” 70 Until the reintroduction of Taungya forestry in Ghana in 2001 (albeit in a modified form known as the Modified Taungya System-MTS), previous forest policies starting with the 1927 Forest Ordinance and the 1948 Forest Policy (see sections 2.8.5 and 2.8.6) appear to have alienated local forest communities. Lack of secure tenure undoubtedly created some variant of open access situation that encouraged adverse practices such as illegal logging; destructive small-scale surface mining and deliberate brush/bushfires to trap wild game. Some researchers thus attribute excessive loses in Ghana’s vegetative cover in recent years to the forest fires of 1983 and 1984 which allegedly destroyed at least 50% of Ghana’s vegetative cover in 1983 alone (Ampadu-Adjei, 1987, in Adam, 1996). While destructive practices such as deliberate uncontrolled bushfires may be easily blamed as primary causes of forest destruction, the root cause of these fires may to a large extent be traced to a lack of secure tenure and the resultant open access conditions that prevented private investments in best practices and enforcement. Open access to forest resources it is argued “reduces the private benefits of forest management below those of social benefits thus creating a disincentive for incurring any additional costs of sustainable management (Afful-Koomson, 2000, p.20)”. Thus, while the estimates from 1984 and 1985 alone, suggest as many as 1005 cases of wildfires in Ghana’s tropical rainforest and savannah regions with at least 307 destroying portions of the semi-deciduous forests zone (Ampadu-Adjei, 1987, in Adam, 1996) majority of those fires may have been prevented by joint (public-private/local) investments in fire barriers and enforcement of good forestry practices. By ensuring that local communities and individuals have secure tenure rights a strong coalition of local communities, public sector (FSD, law enforcement-police and military) and other civil society groups may be formed that will help enforce good forest practices while discouraging illegal activities. 71 2.12 Tax Reform and Timber Certification In a recent study that examined gold mining operations by foreign mines in Ghana, Akpalu and Parks (2007) suggested that the environmental costs of mining can be effectively addressed by including those costs into the tax rate levied on mining firms. Using a dynamic model, Akpalu and Parks (2007), thus demonstrated that value-added tax/ad-valerom severance tax (VAT) paid by gold mines on gross gold revenue provides an avenue for internalizing environmental costs if the tax is equal to the ratio of marginal damage from gold extraction (Apkalu and Parks, 2007). The authors argue that the 3% SAP era tax rate levied on gold outputs fall far short of the external costs of forest resource extraction in Ghana (Apkalu and Parks, 2007). In another study that investigated the relationship between timber certification systems and sustainable forest management (SFM) in Ghana, Afful-Koomson (2000) concluded that international certification has a dual advantage of helping to promote SFM while simultaneously guaranteeing market access for certified timber products. According to Afful-Koomson (2000) shortage of certified timber on the world market holds immerse prospects for Ghana as the bulk the country’s timber outputs are exported to EU countries most of whom already signaled a preference for certified timber by switching to exporter countries with credible certification programs. 2.13 Role of the Media in Forest Management Using data from seven selected countries including Ghana, Lawson and McFul (2010) estimated that about 17 million hectares of tropical forests have been preserved as a result of the decline in illegal logging thus preventing the release of some 1.2 -14.6 billion tons of CO2 into the atmosphere. According to Lawson and McFul (2010, p.7), increasing global, regional and local media coverage of illegal logging activities and their effects on the local and global climate 72 in part stimulated massive enforcement directed at halting the practice. Lawson and McFul (2010) also suggested that developed country imports of illegal wood from the countries reviewed declined by as much as 30%. The study however revealed that in Ghana most of the media coverage on the subject of illegal logging sadly blames the practice on small and medium– scale artisanal chainsaw operators but not the large-scale commercial companies responsible for the bulk of illegal logging in the country (Lawson and McFul, 2010, p.12). In an effort aimed at curbing export of illegal timber from Africa, the EU and its African counterparts reached an agreement in Accra, Ghana on September 3rd, 2008, that proposed to establish boarder control mechanisms to prevent unlicensed wood shipments from entering any of the EU’s 27 blocks of countries. 73 APPENDICES 74 APPENDIX A: Agro Climatic Conditions in Ghana Figure 2.2.1: January and February Average (mm) Rainfall Patterns in Ghana (mm) (mm) 75 Figure 2.2.2: March and April Average (mm) Rainfall Patterns in Ghana (mm) (mm) 76 Figure 2.2.3: May and June Average (mm) Rainfall Patterns in Ghana (mm) (mm) 77 Figure 2.2.4: July and August Average (mm) Rainfall Patterns in Ghana (mm) (mm) 78 Figure 2.2.5: September and October Average (mm) Rainfall Patterns in Ghana (mm) (mm) 79 Figure 2.2.6: November and December Average (mm) Rainfall Patterns in Ghana (mm) (mm) 80 APPENDIX B: Logging Activity in Yaya Forest Reserve Figure 2.3.1: Illegal chainsaw logging on a local farm in Yaya near Sewiah Community 81 CHAPTER 3: HOUSEHOLD LIVELIHOODS APPROACH 82 3.1 Introduction Past, present and future fears of anthropogenic events that contributes to adverse changes in the climate leading to; rising sea levels, desertification, food insecurity, and poverty are often couched in images of increasing human population and unsustainable natural resource exploitation especially in developing countries. While the world’s population over the last century has increased rapidly, so to have the knowledge and methods for combating anthropogenic problems, the brunt of which have for generations disproportionately affected resource-poor communities in the developing world. The refinement of old research methods, emergence of new ones, and conscientious efforts to merge the strengths of new and old approaches is a vindication of the claims by many poverty experts that conventional methods alone are grossly inadequate in addressing the problems of poverty. This current study recognizes the multi-dimensional realities of resource-poor households in Ghana and thus employs an asset-based framework to investigate households’ decisions to participate in the country’s Joint Forest Management (JFM) program (dubbed the Modified Taungya System - MTS) and the livelihood impacts resulting from participation. The asset-based approach employed in this study ensured that poverty in the study communities and households is defined by the poor themselves whose voices undoubtedly will help steer public policy in a direction that effectively addresses their concerns and challenges. The rest of chapter three is divided into the following sections: 1. Literature 2. Livelihood/Asset-based Terminologies 3. Livelihood Capital/Asset 4. Asset-based Approaches 83 3.2 Literature Some researchers (see Sen, 1981, Rodríguez-Bilella, 2009) point out the weaknesses of conventional production, employment, and poverty-line approach to understanding rural poverty and deprivation in developing countries. Others also argue that in the social science discipline, “the context of accelerating change and uncertainty is often confronted by conservational conservatism in concepts, methods, values and behavior (Chambers and Conway, 1991. p2). Hence solutions to rural poverty and suffering in developing countries are often defined within the narrow confines of production, employment and the poverty-line (Chambers and Conway, 1991). By defining the problem of hunger and malnutrition as one of primarily low agricultural production, employment and income underscores the importance of entitlements and claims over resources (Sen, 1981), which in most cases is seriously lacking in poor and vulnerable rural households. Chambers and Conway (1991), charge that conventional concepts typically share the two fundamental flaws of been an industrial country imprint and reductionism for the ease of measurement and comparison of conditions among rural households, communities and even countries. The purpose of chapter three is to build a theoretical and conceptual foundation for this study based on past present and emerging literature on livelihoods and poverty. The underlying thesis guiding the literature review is the gradual shift in concept, from a more traditional approach to poverty research based primarily on income and expenditure to one that stresses the use of different livelihood assets of which income and expenditure are included. The quote below explains the importance of taking a broader approach to studying poverty that goes beyond income and expenditure parameters. “We must not make the mistake- common in some circles- of taking the growth rate of GNP to be the ultimate test of success, and of treating the removal of illiteracy, ill-health 84 and social deprivation as- at best- possible means to that hallowed end. The first and the most important aspects of what we have to do is to make clear that the elimination of illiteracy, ill-health and other avoidable deprivations are valuable for their own sake- they are “the tasks” that we face. The more conventional criteria of economic success (such as, high growth rate, a sound balance of payments, and so forth) are to be valued only as means to deeper ends. It will, therefore, be a mistake to see the development of education, healthcare, and other basic achievements only or primarily as expansions of “human resources”- just the means of production and not its ultimate end (Sen, 1994, p.3).” While the asset-based approach appears to still be in its infancy, its usefulness in focusing policy on the needs and problems identified by the poor themselves cannot be overemphasized. A body of literature detailing the historical perspectives of poverty and livelihoods research, income-expenditure based methods and emerging asset-based approaches is thus reviewed. By thoroughly discussing the theoretical and conceptual underpinnings of the asset-based method, this literature review helps shed light on the gaps in traditional income and consumption as well as asset-based methods and how these gaps can be bridged with new methods. Section 3.2 below describes major terminologies used in livelihood research and the asset-based approach. 3.3 Livelihood/Asset-based Terminology 3.3.1 Livelihood What is a livelihood, how households sustain their livelihoods and what are the appropriate indicators for measuring changes in livelihoods are three very important development questions that have eluded researchers for decades. The concept of livelihoods is increasingly used in the development literature to describe people’s capabilities, and social as well as material assets important to sustaining a means of living (Kanji and Barrientos, 2002). Messer and Townsley (2003) define livelihood as basically comprising the means a household uses to achieve and sustain a certain level of well-being. Just how sustainable a household’s 85 livelihood is depends on several factors some of which vary widely across space and time. Chamber and Conway’s (1991) comprehensive definition of livelihoods based in part on Sen’s (1981) capability and entitlement concept is that which has come to be widely adopted by several development agencies (e.g. DFID, USAID, IDS, FAO, UNDP, and CARE etc). Chambers and Conway (1991) define livelihood as comprising of capabilities, tangible (material) and intangible assets (claims and access) and incentives required for a means of living. Chambers and Conway (1991) assert that for a livelihood to be sustainable however, it should be capable of recovery from stress and shocks, while at the same time maintaining or enhancing its capabilities and asset base. According to Chambers and Conway (1991), the ability of a livelihood to contribute net benefits to other livelihoods while simultaneously providing opportunities for future generations is the hallmark of sustainability. Hence if in the process of providing a livelihood for example through daily economic activities results in a degraded resource base then that livelihood strategy in question is unsustainable in the long term (Messer and Townsley, 2003). On the contrary, if households’ livelihoods are built and sustained on a resource base (e.g., forests, farmlands, lakes, rivers and oceans) that remains uncompromised over the long-term then the livelihood in question is likely to be more sustainable (Messer and Townsley, 2003). 3.3.2 Livelihood Assets Livelihood assets may be defined as all the resources that go into creating a livelihood Messer and Townsley (2003) hence both tangible and intangible assets represent the two main categories of livelihood assets. Assets therefore take on tangible forms such as natural, physical, and financial assets (e.g. land, labor, capital, and savings) or intangible forms such as human, social, vulnerability, empowerment, political, and institutional assets (Siegel and Alwang, 1999). Other researchers view assets as stocks of different types of capital to be used either directly or 86 indirectly to create livelihoods (Elasha, Elhassan, Ahmed and Zakieldin, (2005). Hence the process of livelihood construction often involves both depletion and accumulation of assets (Elasha, Elhassan, Ahmed and Zakieldin, (2005). According to Jansen, Siegel and Francisco (2005), a household’s asset portfolio provides a roadmap for income generation and risk management strategies, and over time changes in well-being may be attributed to either a strong or weakening asset-base. It may therefore be argued that the stronger a household’s asset-base the higher the chances of minimizing its vulnerabilities while improving its well-being. Livelihood assets thus serve very important consumption smoothing, risk management and productive function as well as income generation (Jansen, Siegel and Francisco, 2005, Dorward, Anderson, and Clark et al., 2001). 3.3.3 Vulnerability Context Vulnerability “is a function of the risks to which people may be exposed, the sensitivity of their livelihood system to such risks, and their ability to adapt to, cope with, or recover from the impacts of external ‘shocks’ to their livelihood system (Allison and Horemans, 2006, p.757).” Vulnerability thus represents “insecurity in well-being of individuals, households, and communities in the face of changes in their external environment (Serrat, 2008, p.3).” The three major characteristics that define a household’s vulnerability context are shocks such as conflicts (e.g., civil/tribal wars), illnesses, natural disasters (e.g. floods, fires, storms, droughts, famine, pests, and diseases), seasonality (e.g. prices), and employment opportunities. Other factors that may affect a household’s vulnerability context are demographic, environmental, economic, governance, and technological trends (Serrat, 2008). Following the above explanations, vulnerability context may thus be described as the external environment within which household livelihoods are constructed Baumann (2002), the risks associated with their environment Elasha 87 et al (2005) and the household’s resilience to varying degrees of shocks (i.e. exposure to risks) embedded in that environment (Fouracre, 2001). Vulnerability context may this be grouped into two major categories depending on whether the shocks to a household are registered externally in tangible or internally as intangible livelihood outcomes. For example, external outcomes of shocks such as loss of physical, financial and natural assets often manifest themselves directly or indirectly in tangible forms as homelessness, physical disability or illness, hunger, or even death. Internal outcomes of shocks on the other hand are more subtle, intangible in nature, difficult to observe and quantify while adversely affecting household well-being by stripping the household of entitlement, capabilities (Sen, 81), as well as other form of defense mechanisms. Depending on the nature of the relationship between a household and other more powerful groups or individuals, Horemans (2006) argues that the household may either enrich its assets or become excluded from economic opportunities, social networks and political processes necessary to effectively withstand and recover from shocks. A household’s access and entitlements to tangible and intangible assets according to Baumann (2002), provides the capabilities to device livelihood strategies in response to shocks. The vulnerability context thus affects all aspect of livelihoods including both tangible and intangible assets as well as the livelihood strategies employed by households to minimize these risks. It is therefore logical that asset-based methods such as embodied by the SLF, SLED, and HLS begin with and have at their core the vulnerability context (see figures x, y, z below). In a recent effort to apply the Sustainable Livelihoods Framework (SLF) to water projects in Asia, Africa, and South America, Nicol (2000) found that understanding the vulnerability context within which households’ gain access to water rights represents the first and critical step to understanding access to water security. As rural households embark on a constant transition in and out of poverty, Serrat (2008) suggests 88 that research methodologies that incorporate the concept of vulnerability will help track or capture the process of change in well-being more effectively than other methods. Fouracre (2001), also found that understanding the VC, allows the researcher to uncover the structures and processes associated with trends, shocks as well as the cultural settings influencing household livelihood strategies. One of the advantages of an asset-based method is that it allows the researcher to understand how rural households sustain their livelihoods and what livelihood strategies they adopt to cope with and recover from shocks, trends and seasonality associated with their vulnerability context. Understanding a household’s vulnerability context, according to Hussein and Nelson (n.d) involves a thorough understanding of two pressing analytical and practical questions. Analytically, the researcher needs to understand the institutional arrangements that work in favor of or against secure sustainable livelihoods while practically uncovering policies that enhance or impede such livelihoods (Hussein and Nelson, n.d). Adger et al (2002) contends that shocks sometimes increase the vulnerability of some households while providing an opportunity for others to re-evaluate their livelihoods strategies and make necessary adjustments in favor of more sustainable outcomes. Hence by understand the vulnerability context, as well as the livelihood assets and how they are combined in different livelihood strategies, researchers are able to determine which assets are limiting, and which strategies are more effective against specific shocks. With a thorough understanding of the vulnerability context, researcher may be able to propose and design policies and interventions that help the poor escape poverty or cope with shocks. Some of the strategies may involve improving access to markets, improving transportation networks and roads, improving access to education, and health care services, land tenure systems and improving access to credits. Baumann (2002) also suggests that the outcomes 89 of shocks are often mutually reinforcing in the sense that the vulnerability of livelihood strategies to such shocks means households are often unable to cope with stresses when they occur, or even manipulate their environment to reduce stress in the future, and also unable to benefit from positive trends even when these do occur. The asset-based method and the vulnerability context which it embodies according to Baumann (2002) helps researchers to effectively study the various factors that directly or indirectly unleash untold hardships on individual households and yet are often outside of their immediate control. Shocks often draw attention to policies that sometimes result in unsustainable exploitation of resources and the need to make necessary adjustments in favor of more sustainable livelihoods strategies and outcomes. In a study aimed at uncovering changes and similarities in livelihood strategies within farm families in Bawku District of North-Eastern Ghana, Whitehead (2002) adopted an asset-based method to study changes in household’s vulnerability context and the livelihood strategies that were adopted to deal with such shock in two distinct time periods 1975 and 1989. According to Whitehead (2002), focus on the vulnerability context allows researchers to uncover the “complexity and interrelatedness of livelihoods, their links with wider institutions and processes, together with the focus on people as agents making decisions, albeit constrained.” Nicol (2000) lists some of the risk factors that usually affects households as the household’s physical environment such topographical factors, soil and water availability; the transparency of government, availability of local private sector companies, and the channels of communication through which claims to entitlements are expressed. 3.3.4 Poverty and Sustainable Livelihood In the last two decades, socio-economic research appeared to have favored rather heavily the use of quantitative livelihood indicators such as income and expenditure (that can be readily 90 measured) while largely ignoring other qualitative development indicators such as a household’s capabilities and entitlements (see Sen, 1981 and 1985). Sahn and Stifel (2000) suggest that moving poverty research more towards the use of an asset-based approach allow researchers to adequately study how household assets are accumulated in the bid to escape income inequalities and the poverty trap. Poverty used within the context of this study refers to an inadequate livelihood outcome resulting from a household having inadequate access to basic assets such as land, water, credit or social support (Messer and Townsley, 2003). Poverty may be caused or exacerbated by public policies, institutions and processes that fail to support adequate livelihood (Messer and Townsley, 2003) and ignores individual freedoms and capabilities (Sen, 1981). Messer and Townsley, (2003) further explain that the very strategy adopted by households in combining the available livelihood assets at their disposal in of itself may result in poverty or inadequate livelihoods. This study adopts Hussein and Nelson’s (n.d, p.3) definition of sustainable livelihood which states that “livelihood is sustainable if it can cope with and recover from stresses and shocks and maintain or enhance its capabilities and assets both now and in the future, while not undermining the natural resource base.” Sections 3.1 through 3.5 discuss the five key livelihood assets that constitute the cornerstone of a typical household’s livelihood. 3.4 Livelihood Capital/Assets 3.4.1 Financial Capital Financial capital/asset plays an important role in building and sustaining livelihoods in both rural and urban communities. Held in the form of cash or available stock (Carney, 1998, Frankenberger, Drinkwater and Maxwell, 2002, Kollmair and Gamper, 2002), financial assets may be derived from several off and on-farm activities. The most common means of generating cash in rural communities includes sale of farm produce and wages from both off and on-farm 91 activities. Remittances (Ellis-Jones and Mason, 1999) and inheritance also accounts for some of the financial capital flows into rural communities. In most urban economies, financial capital in the form of cash is the standard means of business transaction and exchange of goods and services. In rural economies however financial capital may be tied up in lumpy assets such as livestock, jewelry or stored grains which require elaborate negotiations and time to be converted into cash. Kollmair and Gamper (2002) assert that while financial capital is easily converted into other forms of capitals such as natural (e.g. livestock and crops) and physical capital (e.g. jewelry, housing), it remains the least available among rural households. The advantage of financial assets relative to other forms of capitals lies in their versatility. Sanders (2000) revealed that unlike rural communities, the urban poor live by arranging complex systems of loans and debt servicing, borrowing small amounts and calling in debts from others to pay bills as they arise. Sanders (2000) also argue that for the urban poor, accumulating financial assets is almost always a key activity for greater livelihood security. While households in rural communities with relatively higher formal education often convert their knowledge and skills into financial capital in the form of cash paid in salaries or wages, the primary source of financial capital in most rural households in Ghana is sale of farm produce or wage/farm labor. Formal and informal credits also provide a means of financial capital to rural households. In recent years production subsides in the form of input vouchers which may be redeemed as fertilizer, seeds and other farm inputs have served the purpose of cash flow. 92 3.4.2 Social Capital Over the past decades, the term social capital has dominated the development debate, however “little consensus exists about what it is, how it is observed and measured, which outcomes it supports, and more importantly, which outcomes it does not support (Krishna, 2004, p292)”. Social capital/assets represent the quantity and quality Frankenberger, Drinkwater and Maxwell (2002) of all social resources that households depend on or use to sustain their livelihoods (Kollmair and Gomper, 2002). Hence for households or communities to benefit from social assets requires some form of membership in a group or network Portes (1998) as well as investments in interpersonal relations within the network (Lin, 1999). The quote below shed more light on the nature of social capital and how cultural specificity influences indexes designed to measure it. “Social capital is not directly observable; people carry it inside their heads. What is observed and measured are some manifestations or behavioral consequences of social capital, including both structural and cognitive elements. Different cultures uphold different expressions of social capital; hence its observable aspects will vary contextually. Different measures of social capital are appropriate for different social and cultural contexts. Hence developing a locally appropriate index of social capital is the first step to examine its utility and to investigate how to add to its stock (Krishna 2004, p.296).” Traditionally, access to embedded resources within social networks has been mediated through birth right, age, gender, political affiliation, education, and location. In instances where group membership alone is not enough to guarantee access to resources within a network, strong relationships of trust (Ellas-Jones and Mason, 1999) and extended family ties and networks (Frankenberger, Drinkwater and Maxwell, 2002), have proven to be effective in negotiating demands for rights and claims to network resources (Carney, 1998 and Sanderson, 2000). Like other livelihood assets, the quantity and quality of social capital within a household significantly 93 affects livelihood outcomes and the effectiveness of development programs. Several indicators including: memberships in different groups (e.g., churches, mosques, JFM project, government and non-governmental associations, producer groups etc.), and number of close friends and family within one’s network provide simple and straight forward measures for estimating the quantity of social capital within a household. The challenge however lies in identifying appropriate quality indicators and measuring them. For example, within one’s social network, indicators such as the level of connectedness, trust, cooperation and influence (Frankenberger, Drinkwater and Maxwell, 2002) may provide clues about the quality of the network. In terms of group membership, quality indicators such as effectiveness of rules, sanction, norms, and security of transfer of network resources (Siegel and Franscisco, 2005) provide a hint into the quality/ability of the group to adequately support livelihoods. For the most part quality indicators at best may be described as blurry, unobservable, and subjective thus posing extremely difficult problems for researchers to identify and measure. The following quote from Krishna (2004) demonstrates the challenges in quantifying or measuring social capital indicators: 3.4.3 Physical Capital Physical capital is the basic infrastructure (Siegel and Alwang, 1999, Serrat, 2008) and producer goods (Kollmair and Gamper, 2002) required to maintain and enhance livelihoods (Fouracre, 2001, Frankenberger, Drinkwater and Maxwell, 2002). Though physical capital is often generated through economic production (Carney, 1998), the level of access to physical assets is often determined by their proximity to communities (Ellis-Johnes and Mason, 1999) and cost of production (Gordon and Craig, 2001). Affordable and secure access to critical assets such as shelter, clean/safe/portable drinking water, primary healthcare and basic education are the basic rights of all humans yet provision of physical assets that guarantee these rights, their 94 maintenance and enhancement is mediated through initial investments in other key assets such as transportation (e.g. roads), communication and market infrastructure. These initial investments thus make these basic assets cost prohibitive and thus lacking in most rural communities. Kollmair and Gamper (2002) use the concept of opportunity costs/trade-offs to explain how a lack of basic infrastructure often precludes access to critical livelihood services such as education, access to health care and income generation opportunities. According to the authors investments in irrigation facilities for example reduces the time spent on non-productive activities, such as the collection of water thus freeing up time that could be spent in school or other productive ventures. Similar arguments can be made for investments in mechanized production (plows/cultivators/planters, harvesters, and grain mills) that significantly increase food production and reduce the drudgery of manual labor particularly among females that are often burdened with these tasks (see Siegel and Alwang, 1999). Janvry and Sadoulet (2000) argue that physical capital such as electrical power plants/grid; roads and irrigation facilities are important public goods that affect the value of other assets held by households. For example, availability of affordable electric power often creates the opportunities for rural households (often with some assistance of public and civil society groups) to invest in small-scale enterprises (e.g. grain mills,) that would have otherwise been extremely to operate with fossil fuels such as diesel or petrol (see Gordon and Craig, 2001). Also, electricity allows community members both adults and children to partake in evening school thus improving on their human capital. From the preceding discussion it is clear how investments in essential physical assets help to enrich other assets such as financial (through income generating small-scale industry) and human capital (through provision of basic shelter, educational facilities, portable water and energy). Chapter seven explains how the physical capital indicators included in this study were generated. 95 3.4.4 Natural Capital Natural capital within the context of this study describes the natural resource stock (Fouracre, 2001, Kollmair and Gamper, 2002) from which resource flows useful for livelihoods may be derived (Frankenberger, Drinkwater and Maxwell, 2002). Hence land, water, forests, wildlife, non-timber forest products and other environmental resources may be collectively described as natural capital. In most rural communities in West Africa natural capital is of particular important as the bulk of rural livelihoods are derived from it. Anthropogenic activities including excessive deforestation, wildfires, unsustainable exploitation of land and aquatic flora and fauna that compromise natural capital resources also affects the quality of livelihoods derived from these resources. Some researchers argue that construction of road networks through tropical forests for example may result in improved access to natural assets and management of forest resources (Fouracre, 2001) however these also carries the risk of opening up natural forests to new settlements often with detrimental effect to the natural resource base on which livelihoods are constructed. Kollmair and Gamper (2002) further explain that the close relationship between natural capital and the vulnerability context of most rural livelihoods creates feedback effects where anthropogenic (fire, deforestation, erosion, water pollution) and natural processes (floods, earthquake, landslides) that destroy natural capital also have devastating effects on livelihoods. 3.4.5 Human Capital Some researchers define human capital as the sum of an individual or household’s skill sets and experiences (Sanders, 2000). Others define it as the knowledge, abilities (Frankenberger, Drinkwater and Maxwell, 2002), and health status (Kollmair and Gamper, 2002) of individuals or households. Mincer, (1996) suggests that differences in the levels of human capital stock held 96 within households may in part explain observed differences in wages across households. According to Mincer (1996) there is a direct relationship between human capital growth and economic growth and that most of the socio-economic transformation observed for example in advanced economies such as the United States in the 19th century was a result of increased human capital endowment within households. The key human capital indicators described by Mincer (1996) are increased literacy and education resulting from increased high school graduations. Siegel and Alwang (1999) also argue that the qualitative attributes of human capital indicators such as health and nutritional status, skills and knowledge are the most important components of any labor force in that returns on investments in these attributes helps vulnerable households escape adverse conditions. Like in most rural communities, Weatherly (2003) found that human capital can be the single most important asset and liability within a firm because it has the potential of been influenced, invested in wisely or wasted thoughtlessly but never completely controlled. Weatherly (2003) define human capital as the collective sum of a household’s size, age distribution, life experiences, knowledge, and education. According to Weatherly (2003) individuals often enhance their human capital by building strong bond with family, community and interest groups, following particular religions or belief systems, ensuring physical fitness, engaging in productive economic activities and investing in education (Weatherly, 2003). Evidence from the development literature also suggests a high intergenerational correlation between an individual or household’s economic status and persistent disparities in health between the rich and the poor (Chakraborty and Das, 2005). In a recent research to determine the interactions between health, wealth and human capital accumulation, Chakraborty and Das (2005) found a positive correlation between an individual’s health and socio-economic status. Private investment in health and the interplay between income 97 and mortality it is argued determines who falls into and stay inside the poverty trap (Chakraborty and Das, 2005). Evidence across both the developed and developing countries (USA, Brazil and Bangladesh) suggests that relative to higher income households, the risk of death from illness is two to three times greater for individuals or households at the bottom of the income distribution scale (Chakraborty and Das, 2005). Over the years, researchers have used school enrollments, average years of schooling and highest educational attainments as a proxy for gauging the human capital endowment within households (Mulligan and Sala-Martin, 2000). The preceding discussion suggests growing evidence in favor of a strong relationship between human capital endowments and economic growth. Poor endowments in human capital it is argued often leads to failure of these assets to provide safety nets under adverse conditions. 3.5 Asset-Based Approaches 3.5.1 Historical Background For decades, economists appear to have favored a more welfarist approach to accessing household poverty and well-being Sahn and Stifel, (2000) largely based on the income generating potential of the household’s tangible assets particularly land labor and physical capital (Siegel and Alwang, 1999). While an income and expenditure-based approach may be entirely relevant particularly within a developed country context, extending purely welfarist approach to developing countries has come under intense scrutiny in recent years. The poverty debate of the 1980s and 1990s called into question the merits of conventional research methods based on income, and expenditure (see Chambers 2007). The basic tenet of the critiques’ argument is that, in developing economies, the aggregates of a household’s consumption expenditure simply fail to adequately capture the true nature of poverty and its causes. By failing to capture the true causes of poverty, the income or consumption expenditure methods also fails to offer the right 98 policy prescriptions for poverty alleviation among poor rural households. Chief among the critiques are Sen (1981, 1985 and 1987 and Chambers and Cornway (1992). The critiques suggest the need for a new pro-poor approach to poverty research based both tangible and intangible household assets. The asset-based approach to studying poverty and a household’s well-being thus integrates both welfarist indicators such as income, expenditure and consumption as well as non-welfarist indicators such basic capabilities, malnutrition, poor health and entitlements advocated by Sen (1981), Chambers, Cornway (1992), and Sahn and Stifel, (2000). The quote below reveals the interconnectedness and feedback-loop between typical welfarist indicators such as income and expenditure and other qualitative human attributes that produces and sustains economic wellbeing. “The bettering of human life does not have to be justified by showing that the person with a better life is also a better producer. Basic education, good health and other human attainments are not only directly valuable as constituent elements of our basic capabilities, but these capabilities can also help in generating economic success of a more standard kind, which in turn can contribute to enhancing the quality of human life even more. Many of the ingredients of a good quality life- including education, health and elementary freedom- clearly do have instrumental roles in making us more productive and helping us to generate more outputs and incomes (Sen, 1994, p.3).” In a recent study of rural livelihoods strategies of forest communities in Córdoba (central Argentina), Rodríguez-Bilella (2009) claimed that conventional/traditional research methodologies based on simplistic and deterministic analysis often fail to adequately capture human causes of environmental degradation. Rodríguez-Bilella (2009), in an attempt to explain the shortfalls of deterministic methodologies, describe the concept of “Positivism and Instrumentalism” that has for generations dominated the natural sciences and until recently informed various socioeconomic research endeavors. According to Rodríguez-Bilella (2009, p.2), “Positivism and Instrumentalism” are theoretical and methodological constructs that “admit 99 precise predictions and rigorous methods of testing hypotheses, especially reproducible experiments involving quantifiable predictions and measurements.” Positivism he argued: “…takes statements about facts, including social events, as congeries of observable sense data to be further confirmed or disconfirmed. Therefore, when applied to human beings and society, it would be the observable behavior of individuals, including their verbal utterances, which must be given priority in arriving at conclusions about social phenomena. It was in this sense that the obsession with quantitative methods and statistics that works on collected data from the behaviors of respondents became the most reliable instrument to the positivist-minded researcher in social research (Rodríguez- Bilella, 2009, p.2-3). While it may be tempting for researchers to want to extend purely natural science methodologies broadly to other disciplines, the unobservable characteristics of human interactions within society and the effects such interactions have on the environment calls for additional research methodologies. Using survey data from the Living Standards Measurement Study (LSMS), Sahn and Stifel (2000), demonstrated the advantages of using an asset-based approach to define poverty in Côte d’Ivoire, Ghana and Vietnam. The decision to use an asset- based approach according to Sahn and Stifel (2000) was due to apparent “conceptual and technical problems associated with the traditional money-metrics of welfare based on income and consumption expenditures.” By constructing an asset index, Sahn and Stifel (2000), were able to profile poor households and then estimate a demand function for the nutritional status of children undoubtedly an important poverty outcome. Based on their findings, Sahn and Stifel (2000) confirmed the appropriateness of employing asset-based methods for poverty research in poor rural communities based on its conceptual soundness and ease of measurement relative to income and consumption expenditure. The following statements by Sahn and Stifel (2000) summarize the argument in favor of asset-based methods: “Since meaningful poverty alleviation is largely predicated on the individual’s ability to accumulate productive assets, and income inequality will be reduced by addressing the 100 unequal distribution of income generating assets, there is considerable merit in moving the process of poverty measurement away from expenditure-based measures, toward a more asset-based focus. Such a focus will, in turn, have implications for poverty reduction strategies. It implies more emphasis on economic and social forces that contribute to asset inequality, instead of anti-poverty measures that are targeted and evaluated based on expenditure levels (Sahn and Stifel 2000, p.3).” Some researchers point out that because a household’s assets represent the primary means of sustaining a livelihood, an approach that explores the relationships between assets and the context within which they are employed and the livelihood outcomes resulting from that combination offers the best means of studying poverty. Siegel (2005) suggests that by understanding the quantity, quality and productivity of available assets within a household places researchers in an advantageous position to determine the potential for long term growth and poverty reduction within the household. It thus follows that because assets represent a household’s engine of growth, efforts towards understanding poverty and how to improve a household’s well-being must first begin with the question of how to improve the household’s tangible and intangible asset base. In most poor rural communities where assets are likely to be inequitably distributed and household exposure to natural, economic and social risks likely to be high, Siegel (2005) contends that an asset-based approach offers the best chance of truly understanding the plight of households and crafting appropriate policies for improving well- being. Different variants of the asset-based approach have been adopted by major development agencies (e.g., IADB, USAID, FAO, IFAD, DFID, CARE, and the Ford Foundation) for delivering their development programs (Siegel, 2005). The asset-based approach adopted in this study owe its origins to the broader debate surrounding assets and capabilities Sen, (1981, 1984, 1987), hunger and entitlements (Sen, 1987, Borton and Shoham 1991), freedom of choice (Sen, 1987, 1997), risk and vulnerability 101 (Chambers and Cornway, 1991). To the proponents of asset-based methods the multi- dimensionality of poverty Chambers (1991) demand that a distinction be made between “poverty as a static concept and vulnerability as a dynamic one (Moser, 2006, p8).” Hence in order to adequately capture the multifaceted nature of poverty requires a methodology that captures households’ entitlements and asset accumulation strategies, as well as the various shocks that predispose either the household or individuals within it to different levels of vulnerability. According to Moser (2006), insecurity within a household often results from exposure to various forms of risks with the outcome vulnerability manifesting itself as declining well-being. Dorward, Anderson, and Clark et al (2001) contend that asset-based frameworks developed in response to the ensuing debate provides opportunities to examine in detail poor peoples’ access to different types of assets and the functions of those assets within changing livelihood strategies. Employing an asset-based method in this study helps shed light on the varied causes of household poverty as well as household’s transition in and out of poverty in my research communities. 3.5.2 Asset-Based Methods Asset-based methods such as the Sustainable Livelihood Approach (SLA), Sustainability Livelihood Framework (SLF), Sustainable Livelihood Enhancement and Diversification (SLED) and Household Livelihood Security (HLS) discussed below have been touted as providing excellent opportunities for livelihood researchers to go beyond mere identification of the causes of poverty to recommending policies that favor poverty reduction as well as environment and social sustainability (Jansen, Siegel and Francisco, 2005). Dorward, Anderson, and Clark et al., (2001) suggest that analysis of poor people’s livelihoods must include four critical steps: first the process must examine the functions of different assets within households’ asset portfolio, second 102 researchers must identify priorities for policy and other interventions that are likely to support expanded access to assets, third the process must relate poor people’s access to assets to the functions those assets play within the livelihood strategy and then finally the research effort must identify the most effective livelihood development pathway and how the changing roles of different assets and livelihood strategies influence that pathway (Dorward, Anderson, and Clark et al., 2001). Because the type of assets and the strategies required for obtaining a particular livelihood outcome depends on the local and global context (vulnerability, policy etc.), households require a range of asset/capital endowment for a desired livelihood outcome. Assets thus constitute the basic building blocks of a household’s livelihood in the sense that the household’s capabilities Sen (1981), as well as the power to effectively demand claims to entitlements (Sen and Foster 1997), are embedded in the household’s asset endowments. The effectiveness of an asset-based approach thus results from the opportunity it offers to view poverty from the lenses of the poor themselves. Hence by incorporating both tangible and intangible assets employed by poor households as well as the vulnerability context within which the household’s livelihood is constructed, researchers may be able to truly understand the nature of poverty and the household’s strategies to dealing with it. The Context within which a household is situated greatly influences the household’s livelihood outcomes by dictating the quantity and quality of both tangible and intangible assets available to the household. Context within asset-based frameworks represent policies or institutions and risk that results from either anthropogenic or natural causes (see Figure 3.1 and Figure 3.2). By influencing asset availability, the Context in effect influences how assets are combined and livelihood strategies adopted in the combination. Hence as an adaptation to or anticipation of changes within the livelihood Context, households generally develop continues feedback loops amongst livelihood asset, strategies and 103 well-being outcomes. While some aspects of livelihood Context may be influenced by households, most typically fall outside the household’s immediate sphere of influence thus predisposing the household to risks that often may be unpredictable. Siegel and Alwang (1999) contend that since both tangible and intangible assets of household play very important roles in managing risk and vulnerability it only makes sense that they are both considered in poverty analysis. 3.5.2.1 Sustainable Livelihood Approach (SLA) Asset-based approaches described under sections 3.4.1 through 3.4.5 above have at their core five broad categories (Natural, Physical, Human, Financial and Social assets). To build and sustain livelihoods, a household must combine intangible assets such as capabilities, skills, and knowledge with other tangible asset endowments such as land, labor, capital, and savings. Other factors such as policy and risk/vulnerability context operating beyond households’ immediate sphere of influence also affect livelihood strategies and ultimately the household’s wellbeing. Figure 3.1.1 below adapted from Jansen, Siegel and Francisco (2005), illustrates conceptually the interrelationships amongst livelihood assets, context (policy, risk/vulnerability) livelihood strategies and outcomes. 104 Figure 3.1.1: Sustainable Livelihood Approach (SLA) 105 3.5.2.2 The Sustainability Livelihood Framework (SLF) The “Sustainability Livelihood Framework” (SLF) is another example of an asset-based framework that emerged in the early 1990s. According to Kollmair and Gamper, (2002), SLF’s became popular because they allowed researchers to determine if households that escaped poverty started off with a particular combination of assets, and if such a combination is transferable to other livelihood settings. Kollmair and Gamper, (2002) believed that the SLF provides an opportunity to discover the substitutability potential of assets within households’ asset bundles. By employing an asset-based approach researchers are now able to account for important assets likely to influence a household’s well-being. For example, questions regarding how a lack of financial capital within poor rural households may be compensated by high social capital endowment may be addressed using asset-based frameworks such as SLF (see Messer and Townsley, 2003). Both SLA and SLF feature the five key livelihood assets and provide the pathway for all five assets to be quantified and represented graphically in the form of a pentagon. Each of the five corners of the pentagon represents the levels of a livelihood asset accumulated by a household or community (see chapter seven). According to Messer and Townsley (2003), households with relatively large, well-balanced and regular livelihood pentagons have a stronger asset base while those with small, distorted pentagons have fewer assets from which to build their livelihoods. Varying access or claims to assets within or outside a community also affects the shape of a household’s asset pentagon and ultimately the livelihood strategies and outcomes. Jansen, Siegel and Francisco (2005) observed that asset complementarities between different livelihood assets results in some assets been viable only when combined with others. Siegel and Francisco (2005) thus coined the term “Locational Capital” to explain asset complementarities 106 between the geographical location of a household and some livelihood assets. Possible examples of “locational capitals” include population density, road density, distance to markets, and access to public services. According to Jansen, Siegel and Francisco (2005), the influence of certain natural assets for example land on livelihood outcomes such as income and overall well-being also depends on the location of the land relative to other livelihood assets such as roads, markets, access to credits, access to production technologies and transportation. 107 Figure 3.1.2: Sustainability Livelihoods Framework 108 3.5.2.3 Sustainable Livelihood Enhancement and Diversification (SLED) One asset-based approach that seemed to have evolved out of the SLA is the Sustainable Livelihood Enhancement and Diversification (SLED) approach. SLED was designed specifically to ensure that local coastal communities obtain maximum benefits from natural resource conservation Cattermoul, Townsley and Campbell (2008). Like the SLF, SLED first focuses on human and social capital/assets that influence a household’s access to other livelihood assets. The goal is to begin the enquiry into livelihoods by first understanding the household or individual’s personal characteristics likely to influence access to other livelihood resources. Hence human capital indicators such as education, gender, age and ethnicity for example help lay the groundwork for understanding current livelihood strategies and outcomes and how these may be influenced by program participation and policy changes. SLED also requires that the researcher pay attention to the combination of assets (financial, natural and physical) employed by the household to achieve livelihood outcomes. Like the SLF, SLED also acknowledges the importance of contexts (enabling agencies and external factors) into its framework. The assumption is that enabling agencies such as government, non-governmental groups and service providers through their action also influence the household’s livelihood outcomes. While individuals, households and communities may influence enabling agencies, the effects of external factors such as natural disasters (flood, fires, drought, pests and disease outbreak), global trends seasonal changes may be out of the household’s immediate control, yet these factors have tremendous effects on livelihood outcomes. The main objective of SLED is to help researchers understand the sources of poverty within rural communities before launching into development programs. Currently SLED has been pilot tested in several countries in Asia (see Cattermoul, Townsley and Campbell (2008).). 109 3.5.2.4 Household Livelihood Security (HLS) Framework Another asset-based method that has gained popularity in recent years is the Household Livelihood Security (HLS). Frankenberger, Drinkwater, Maxwell (2000) claim that while the HLS approach owes its origin to the food security perspective of the 1970 and 1980s, HLS was developed on the assumption that food is only one basic need among several, and that in some instances adequate food consumption may be sacrificed for other more important livelihood assets. CARE International’s decision to adopt HLS in 1994 for its development work around the globe was on the grounds that “HLS provides a framework to analyze and understand the web of poverty and people’s mechanisms for dealing with it (Frankenberger, Drinkwater, Maxwell, 2000).” Currently HLS is used by CARE for most of its program analysis, design, monitoring and evaluation (Frankenberger, Drinkwater, Maxwell, 2000). Carney (2003) provides more details on the current thinking and evolution of different asset-based approaches, as well as the strengths and challenges of each method and how different agencies (DFID, OXFAM, ODI, CARE, USAID etc.) have come to adopt different versions for their development programs. 110 CHAPTER 4: METHODOLOGY 111 4.1 Introduction Chapter three of this dissertation discusses the methods used in conducting the study. The chapter first discusses literature reviewed and their sources. Following the literature review section, the chapter then describes the pre-dissertation field work leading to survey instrument design (Appendices E, F, and G), and MSU-IRB application (Appendix B) and University Committee on Research Involving Human Subjects (UCRIHS) approval (Appendix C). Instrument validity and data collection protocol and analysis procedures are also discussed under the instrument design section. Section 4.4 describes the research communities and the GIS community mapping process used to establish a basis for research sample selection. The rest of the chapter discussed the data collection/interview process (individual household, focus group and key informant) and data analysis methods used in addressing all four research questions. 4.2 Literature Review Since this research examines the effectiveness of Ghana’s MTS reforestation program in improving both forests and livelihoods, the study first started with review of a wide range of literature (see chapter two) on tropical deforestation and the range of reforestation programs worldwide. Literature review begun sometime in October 2008 and continued at throughout all stages of the dissertation process. Prior to pre-dissertation field work, literature focused broadly on the plight of tropical deforestation and the different afforestation program used in combating deforestation globally. The deforestation and reforestation literature review then narrowed down to the case of Africa, West Africa generally and then Ghana specifically. Since no single source adequately explains past and present-day deforestation in Ghana, the deforestation and reforestation literature started with Ghana’s history and land tenure system, the evolution of forest policy from pre-colonial, colonial and the post independent era; and Ghana’s economic 112 conditions and its impact on forests. Literature on Ghana’s Economic Recovery Program of the 1980s leading the adoption of World Bank and IMF Structural Adjustment Programs (SAP) is also reviewed in order to understand the extent to which SAPs may have impacted Ghana’s forests. Following the deforestation literature review, attention was turned to global afforestation programs particularly India’s JFM and MTS in Asia that tries to leave forest ownership and management in the hands of local people. The last section of the literature review focus on the “Sustainability Livelihoods Framework (SLF)” and how it was used to guide the entire literature review, instrument design, data collection and analysis process (see chapter three). The appeal for SLF in this research is because it allows the various facets of livelihoods to be examined within the context of Ghana’s MTS program. The MSU Main Library provided an excellent source of hard paper materials on the crisis of deforestation and reforestation programs around the globe including Africa and Ghana. Google and JSTOR were also useful in searching relevant electronic materials on all the five areas of literature described above. For specific information on research community and forest Ghana’s forest reserves the Forestry Research Institute of Ghana’s (FORIG) annual reports and research publications were consulted. Also, Ghana’s University of Development Study’s (UDS) Integrated Third Trimester Field Practical Program (ITTFPP) student reports and case studies provided useful background information on the history, culture and socio-economic status of some of the research communities surveyed as part of this dissertation. For literature on household livelihoods assets, the Institute of Development Studies’ (IDS, Brighton) research papers and publications provided useful information used in designing the household survey instrument and analytical framework for addressing question pertaining to livelihood asset 113 changes ten years after Ghana’s MTS implementation. The Pro-Quest dissertation archives was used to retrieve past dissertation research on global and Ghana specific research on deforestation, reforestation, livelihoods, SAPs, and other works deemed relevant for this current study. 4.3 Pre-dissertation Field Reconnaissance In order to understand the different reforestation programs ongoing in Ghana and also decide on a dissertation research direction, I embarked on a two-month pre-dissertation field exercise in November and December of 2008. As part of the pre-dissertation field reconnaissance, I met with the national director of FORIG Dr. Victor Agyemang to discuss ongoing research activities within FORIG and a possible fit with my dissertation research interests. It was after meeting and discussion with Dr. Agyemang that I later decided to focus my research attention on Ghana’s newly introduced MTS program which had been running for at least eight years and thus ripe for a livelihoods impact assessment. To help understand the intricacies of this new MTS program and also gather field data for my dissertation proposal, Dr. Agyemang introduced me to Dr. Earnest Foli who was to serve as my point person for all matter concerning my research in the Brong Ahafo Region. Exactly one week after been introduced to Dr. Foli, he personally accompanied me to the Sunyani where I was introduced to the Sunyani District Forest Service Department (FSD) Manager Mr. Dickson Adjei Sakyi and his Range Supervisor Mr. Joseph Aggrey. After my initial meeting with the Sunyani District FSD Manager and his Range Supervisor, I scheduled another visit the following week with Mr. Aggrey to be introduced to all 10 MTS community leaders. During my second visit to Sunyani, I spent approximately five days in the region during which time Mr. Aggey took me on his FSD motorbike around all ten communities to be introduced to the community chiefs, elders and their MTS leaders (also called Taungya heads). Visiting all ten Yaya MTS 114 communities, gave me good insight into the MTS activities, member selections and the benefit sharing arrangement proposed by the government as motivation for preserving trees until their final clear cut rotation periods of 25 years following initial teak plantations establishment (see Appendix H). After returning from my second field visit to Sunyani, Dr. Foli later introduced me to Dr. Emmanuel Opuni-Frimpong and Dr. Obiri-Darko Beatrice both of whom become vital resource persons for my research. Both Dr. Foli and Dr. Obiri-Darko later reviewed my entire dissertation proposal for face and content validity. During my research field exercise in March through July of 2009, both Dr. Foli and Opuni-Frimpong sent their research assistants with ample knowledge of the research area (six research assistants in total) to assist with field data collection. Also Dr. Opuni loaned the my research project one of his personal vehicles to be used for data collection for the entire period of my field work. 4.4 Research Community and GIS Community Mapping 4.4.1 Forest Reserves in Brong Ahafo Region Brong Ahafo region has twenty 20 national forest reserves covering a total land mass of 233,469 Ha or 2334.69 square Kilometers (Forestry Commission, 2002). Yaya, Nsemre and Sawsaw reserves are located in the Brong Ahafo region’s Dry Semi-deciduous Forest Zone (DSFZ). In 1994 and 2001 all three reserves were listed among the most degraded reserves in Ghana (see Global Environmental Facility Small Grant Program (GEF) report, 2008). Due to high levels of forest degradation in these reserves the MTS program was launched in 2001/2002 to help restore forest vegetation and simultaneously improve the livelihoods of reserve communities. Figure 4.1.1 below is a map of Brong Ahafo showing Yaya, Nsemre and Sawsaw. 115 Figure 4.1.1: Maps of Brong Ahafo Showing the locations of Yaya, Nsemre and Sawsaw Forest Reserves 116 4.5 Research Communities and GIS Community Mapping Exercise Initial preparations for community mapping began sometime in November and December of 2008 during my pre-dissertation field reconnaissance. During the pre-dissertation survey, I visited all ten Yaya reserve communities to understand the community’s physical structures and also plan for future mapping. Community mapping activities begun in March through June 2009 by which time GPS data for all 19 communities were collected. Three research assistants, one using a Garmin eTrex Venture HC GPS Receiver, another using Sony digital camera collected georeferenced GIS information on community physical assets while the third collected household demographic information. With the exception of Abrefakrom, Ahyiem, Amangoase, Ayigbekrom and Buoku communities that had more than 100 housing units, and thus took longer than two days of GPS field exercise the rest of the communities were mapped within two days. GPS data collection typically began during the early morning house between 6-7am in the morning or when the sky was clear and lasted till 5pm with a one-hour break in the morning and another hour break between 12 till 2pm. Though the Garmin eTrex usually picks up signals from multiple satellites even under cloudy conditions, data collection during cloudy conditions was avoided due to the potential for error readings. 4.5.1 GIS Data Collection Protocol The error margin on the Garmin eTrex used in this study is +/-0.74 to 0.9 meters hence to ensure the highest level of accuracy of the data, the field assistant handling the GPS units were trained to stands in the center of each housing unit while picking up the coordinate for the house. For standalone structures such as boreholes, latrines, small shops and kiosks, the coordinates were taken as physically close to the structure as possible while ensuring the strongest possible satellite signal. A housing unit used in the context of this research is a single house with a single 117 room or a large contiguous structure with multiple rooms often strung together in a circle or enclosure. Relatively large housing units often housed up to 10 or more rooms and were occupied by either the same family or different family units. While the GPS coordinate of a house is being taken, the second field assistant concurrently takes at least five pictures of the same housing unit focusing on the roof, kitchen, bathrooms and housing structure (cement or clay). The third field assistant then records the unique house identification number generated by the GPS unit for each housing unit as well as their picture numbers. Basic demographic information including the name of household head, number of adult males and females older than 12 years and all children below the age of 12 years are also recorded. The number of rooms and separate family units occupying each housing unit were also recorded. The data collection protocol involving picking coordinates, taking pictures and recording unique GPS coordinate identification was also repeated for all physical structures such as boreholes, churches, schools, toilet facilities, markets and shops within a community. All major roads passing through each research community or connecting one community to another were tracked using the Garmin’s tracking option. Since the Garmin eTrex came with a relatively small (24 mega bites) internal memory which allowed for only 500 maximum data points or coordinates, data for each community was completely downloaded from the GPS device and saved in labeled folders on a laptop computer. A Toshiba Satellite 305 laptop computer was always kept handy in the field and used to clear out the GPS device any time it filled up in the middle of data collection. 4.5.2 Data Retrieval from GPS Unit The data from the Garmin eTrex GPS device was retrieved using MapSource software version 6.16.3. To retrieve the data in excel format, the plotted data on MapSource was exported 118 into the desired format. In other cases, data was also be retrieved directly by connecting the device via a USB cable to a laptop and dragging the data files onto a designated folder. One advantage of using MapSource was that it allowed the research team to display the data on the Ghana map and inspect it before downloading and proceeding to the next community. After retrieving the data, all the data cleaning was done in excel before importing into ArcMap 9.3 and then later ArcMap10.0 for processing and mapping. 4.5.3 Image Processing In order to be able to map the data points from the research communities and create community maps, a base map of Ghana in an ArcMap shape file extension/format was obtained from the Ghana Survey Department and loaded into ArcMap 10.0. Once in ArcMap, all data points were projected onto the base map making sure that both the coordinate system of the base map and the data points from the GPS unit matched. Since the Garmin eTrex used for data collection came with the Global Coordinate System GCS_WGS_1984, all base maps including that for Ghana was transformed into GCS_WGS_1984 using the “Projection and Transformation” tool in the ArcMap 10.0 “Toolbox”. With all data points projected onto the Ghana map, household units within research communities were displayed and labeled with the appropriate legends. For the purpose of this study, the boundaries of each community are marked using the houses or communal structures furthermost from the center of the community. In order to determine households’ relative distance from communal assets, three separate buffer zones of 60, 120 and 180 meters were created around all communal physical assets of interest (for example boreholes, major roads, churches, schools and toilet facilities) using the ArcMap 10.0 buffer extension wizard. Once all housing units were projected and labeled, appropriate legends created and 119 physical assets buffered, a polygon representing an outline of each community was digitized using tools from ArcMap 10’s “Editor” menu. The final digitized polygons were saved in a shape file format and imported on the map. The maps of each community are represented below their reserves in the sections that follow. Also immediately following this section are community and household information results from the GIS mapping exercise. 4.6 Background Information on Yaya Reserve Communities In an effort to ensure sustained income from MTS particularly in years following canopy closure, a number of collaborative Community-Based Natural Resource Management projects were introduced in select forest reserve communities. For example, between 2006 and 2008, the UNDP Global Environmental Fund (GEF), through its Small Grants Program (SGP) provided financial assistance to the Yaya Taungya Farmers Integrated Community Groups (YTFICG). The two-year SGP provided seed capital for the establishment of sheep, goat, pig and cane rat/grass- cutter rearing operations for 50 households in Asuakwaa, Sawiah, Amangoase, Mallamkrom, Ayigbe, and Konsua communities fringing the Yaya forest reserve (UNDP-SEF, 2008). YTFICG within each community had the option of selecting one alternative livelihood program for which they wished to receive assistance. The SGP was later scaled-up in 2008 to include Ahyiem, Abrefakrom, Bouku, and Amoakrom communities all surrounding the Yaya reserve. The SGP’s end of project report blames lack of adequate incentives to attract forest communities into forest management for continued degradation of the Yaya reserve (this observation was also reported in the Joint FORIG-FAO, 2008 report). The SGP further recommended assisting forest communities with revolving fund programs in order to provide incentives for sustainable plantation development in the Yaya forest reserve. The current Community Forest Management Project (CFMP) funded jointly by the government of Ghana and 120 the African Development Bank is a scaled-up version of UNDP’s SGP. The new CFMP program provides alternative livelihoods assistance to 452 households in 10 Yaya communities participating in the MTS project (see figure 4.2 below and Appendix I). One of the long-term goals of the CFMP program was to help establish a revolving fund to serve as seed capital for alternative livelihood enterprises. The alternative livelihood programs in essence complement income from crop production and Taungya activities under the MTS within the Yaya reserve. The CFMP’s alternative livelihood program is what distinguishes Yaya’s MTS program from other similar Taungya systems around the country. Figure 4.1.2 below presents a map of the Yaya Forest Reserve highlighting the locations of all ten communities within which the MTS program is currently implemented. Tables 4.6.1 through 4..6.3 below presents information on the demographics, housing infrastructure and livestock ownership of Yaya reserve communities. Appendices A, B and C discussed under chapter six provides detailed community maps highlighting types of housing structure, livestock ownership and availability and distribution of physical capital assets in each of the ten Yaya communities as well as those of Nsemre and Sawsaw. 121 Figure 4.1.2: Maps of Yaya Forest Reserve and Yaya Research Communities 122 4.6.1 Household Information from Yaya Community Maps Table 4.6.1: Yaya Reserve Household Demographic Information No. Name of Community Yaya Reserve Abrefakrom Ahyiem Amangoase Ayigbekrom Buoku Konsua 1 2 3 4 5 6 7 Malamkrom Amoahkrom 8 Asuakwa 9 10 Sewiah No. of males older than 12 years No. of females older than 12 years No. of males less than 12 years No. of females less than 12 years Total Population 235 235 265 315 413 63 73 129 114 43 213 203 285 281 393 62 70 118 143 38 217 211 251 279 360 62 69 135 133 36 195 186 271 237 355 74 68 126 122 36 860 835 1,072 1,112 1,521 261 280 508 512 153 7,114 Pooled 1,885 1,806 1,753 1,670 Table 4.6.2: Yaya Reserve Housing Infrastructure Per Capita No. Name of Community No of housing units 1 Abrefakrom 2 Ahyiem 3 Amangoase 4 Ayigbekrom 5 Buoku 6 Konsua 7 Malamkrom 8 Amoahkrom 9 Asuakwa 10 Sewiah 136 133 158 171 190 48 53 82 50 20 Pooled 1041 No. of family or household s units 156 144 162 192 216 55 53 86 127 34 1225 123 No of No of No. of bedrooms housing to familiy bedroom to family ratio ratio 357 328 444 465 663 116 144 224 213 73 3027 0.87 0.92 0.98 0.89 0.88 0.87 1.00 0.95 0.39 0.59 0.85 2.29 2.28 2.74 2.42 3.07 2.11 2.72 2.60 1.68 2.15 2.47 Table 4.6.3: Yaya Reserve Community Livestock Ownership No. Name of Community CFMP/ MTS Households No of housi ng units 1 Abrefakrom 2 Ahyiem 3 Amangoase 4 Ayigbekrom 5 Buoku 6 Konsua 7 Malamkrom 8 Amoahkrom 9 Asuakwa 10 Sewiah 25 13 34 43 38 19 23 18 24 17 136 133 158 171 190 48 53 82 50 20 Poultry Birds (0= No 1=Yes) 0.4265 0.2105 0.6329 0.6316 0.3684 0.5625 0.8113 0.5610 0.6400 0.7000 Pooled 254 1,041 0.5053 Goat 0=No 1=Yes 0.3162 0.2632 0.6076 0.3509 0.1947 0.3958 0.7736 0.4878 0.4600 0.3500 0.3852 Pig (0=No 1=Yes) 0.0221 0.0977 0.0063 0.0058 0.0000 0.1042 0.0000 0.0000 0.0200 0.1000 0.0250 Cattle (0=No 1=Yes) 0.0147 0.0075 0.0000 0.0058 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0038 4.7 Background Information on Nsemre Forest Reserve Between 2004 and 2009 the Japan International Corporation Agency’s (JICA) established a Participatory Forest Resource Management (PAFORM) project, in five communities surrounding the Nsemre reserve. While PAFORM and the CFMP program in Yaya may be similar in some regards, major differenced exist between both programs. One such difference is that while the CFMP provides material resources and training in teak seedlings production as well as livestock (e.g. goat, sheep, and pig and cane rat/grass-cutter) production and distribution to participants the PAFORM project provided Nsemre communities with only training skills in the a number alternative livelihoods programs (including snail and mushroom production, soap making as well as skills on how to use citrus and other fruit trees for fire protection within the Nsemre reserve. Having mature fruit trees in forest reserves provide some form of safety net particularly during seasons when staple crops are out of season and grain stocks are low. It is assumed that the value placed on fruit trees within the forest reserves will warrant their 124 protection from indiscriminate bush fires and in so doing protect the forest indirectly. Hence the PAFORM program helped Nsemre communities construct fire belts along the reserve using mostly citrus fruits. The citrus seedlings were also provided free of charge to the communities by the PAFORM program. Both CFMP and PAFORM maintained trained technical staff that provided periodic training and program monitoring however after the end of PAFORM in 2008 (after having run for nearly five years) all forms of community support ended with communities were expected to pick up from where PAFORM left off. While the CFMP-MTS and PAFORM projects may sound similar, the two fundamental differences are benefit sharing under the CFMP-MTS and also material support to MTS program hence the decision to lump Nsemre- PAFORM communities with Sawsaw communities and treat these two as a single unit of non- MTS communities for the purpose of analysis under chapters five through eight. The Nsemre reserve and surrounding communities were included in this study because though both Yaya and Nsemre shared a common boarder they had different forest management programs hence the decision to include Nsemre as control. By comparing forest cover (chapter five) and livelihood changes (chapter six) in both reserves this research will be able to provide policy recommendation for improving the conditions within the reserves and communities. Hence all four Nsemre communities supported by the JICA PAFORM project were included in the study. Figure 4.3 below presents the map of Nsemre forest reserve and four research Nsemre communities included in this study. Tables 4.7.1 through 4.7.3 below present information on household demographics, physical assets and livestock ownership within Nsemre Research communities. 125 Figure 4.1.3: Maps of Nsemre Forest Reserve and Nsemre Research Communities 126 4.7.1 Household Information from Nsemre Community Maps Table 4.7.1: Nsemre Reserve Household Demographic Information No. Name of Community Pepewase 11 Asuofre 12 13 Kofitsumkrom 14 Ahwene Pooled No. of males older than 12 years 105 - 91 113 309 No. of females older than 12 years 81 - 88 108 277 No. of males less than 12 years No. of females less than 12 years Total Population 73 - 69 86 228 79 - 69 111 259 338 - 317 418 1,073 Table 4.7.2: Nsemre Reserve Housing Infrastructure Per Capita No. Name of Community No of housing units No. of family or household units No. of bedrooms Housing unit to family ratio Bedroom to family ratio 11 Asuofre 12 Pepewase 13 Kofitsumkrom 14 Ahwene 46 24 56 66 Pooled 192 90 24 60 70 244 168 144 202 514 0.51 1.00 0.93 0.94 0.79 1.87 2.40 2.89 2.39 127 Table 4.7.3: Nsemre Reserve Community Livestock Ownership No. Name of Community No of housin g units 11 Asuofre 12 Pepewase 13 Kofitsumkrom 14 Ahwene 46 24 56 66 Poultry Birds (0= No 1=Yes) 0.5000 1.0000 0.2321 0.2576 Goat 0=No 1=Yes 0.0870 0.0000 0.2857 0.2121 Pig (0=No 1=Yes) 0.0000 0.0000 0.0000 0.0000 Cattle (0=No 1=Yes) 0.0000 0.0000 0.0000 0.0000 Pooled 192 0.4010 0.1771 0.0000 0.0000 4.8 Background Information on Sawsaw Forest Reserve communities Unlike Yaya reserve communities that had an active reforestation and alternative livelihoods project (MTS) and Nsemre communities that recently received some alternative livelihoods support from the JICA PAFORM project, Sawsaw communities had no active reforestation or alternative livelihood programs in recent years leading to this study. Of the three reserves Sawsaw is by far the largest and the most degraded of all three reserves (see chapter five). In terms of size Sawsaw is more than twice the size of Nsemre and larger than both Yaya and Nsemre combined. This study assumed prior to the introduction of CFMP-MTS in Yaya communities or the JICA PAFORM in Nsemre all three forest reserve communities had identical livelihood assets and had similar forest management practices. Hence Sawsaw reserve communities were included in the study to serve as control for the analysis of both direct and spillover effects of the recent CFMP-MTS project on households in Yaya communities (see chapter eight). Figure 4.4 below presents the map of Sawsaw forest reserve and the five research communities included in this study. Tables 4.8.1 through 4.8.3 below present information on household demographics, physical assets, and livestock ownership within the five Sawsaw 128 communities. Community and household selection for this study is discussed under section 4.9 below. 129 Figure 4.1.4: Maps of Sawsaw Forest Reserve and Sawsaw Research Communities 130 4.8.1 Household Information from Sawsaw Community Maps Table 4.8.1: Sawsaw Reserve Household Demographic Information No. Name of Community 15 Ayayo 16 Domeabra 17 Ntema Papasu 18 19 Pipotrim No. of males older than 12 years 115 92 29 17 76 No. of females older than 12 years 101 95 32 11 66 Pooled 329 305 No. of males less than 12 years No. of females less than 12 years Total Population 129 152 37 16 52 386 118 177 24 11 66 396 463 516 122 55 260 1,416 Table 4.8.2: Sawsaw Reserve Housing Infrastructure Per Capita No. Name of Community No of housing units 15 Ayayo 16 Domeabra 17 Ntema 18 Papasu 19 Pipotrim Pooled 71 67 17 12 46 213 No. of family or household s units No. of bedrooms Housing unit to family ratio Bedroom to family ratio 192 248 61 25 165 691 0.96 1.00 1.00 1.00 0.94 0.97 2.59 3.70 3.59 2.08 3.37 3.16 74 67 17 12 49 219 131 Table 4.8.3: Sawsaw Reserve Community Livestock Ownership No. Name of Community No of housing units 15 Ayayo 16 Domeabra 17 Ntema 18 Papasu 19 Pipotrim 71 67 17 12 46 Pooled 213 Poultry Birds (0= No 1=Yes) 0.3662 0.5075 0.6471 0.5833 0.2391 0.4178 Goat 0=No 1=Yes 0.4648 0.7015 0.3529 0.5000 0.3261 0.5023 Pig (0=No 1=Yes) 0.0000 0.0299 0.0000 0.0000 0.0217 0.0141 Cattle (0=No 1=Yes) 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 4.9 Survey Instrument Design and UCHRIS Approval 4.9.1 Instrument Design Four sets of survey instruments (Appendices D through G) were used in this study to generate data on five critical livelihood assets among residents in 19 villages fringing Yaya, Nsemre and Sawsaw forest reserves in Ghana’s Brong Ahafo region. The first of the four instruments (Appendix E) titled the “Livelihoods Monitoring Tool,” (LMT) has five sections with each section focused on one of five livelihood assets (Human, Natural, Physical, Social or Financial). For each livelihood asset, households were first asked to describe the 2009 (present/MTS period) condition of the asset within specific boundaries provided in the questionnaire before describing through recall, the 1999 (before MTS) condition of those same assets within the same boundaries provided by the questionnaire. The 17-page LMT was designed and administered using primarily the World Bank’s year 2000 “Guidelines for designing household survey questionnaires for developing countries,” the Ghana 2000 Census report and the years 2000 and 2008 Ghana Living Standards Survey Reports (GLSS). The 2003 FAO guidelines for analyzing local institutions and livelihoods in developing 132 countries (Messer and Townsley, 2003), and the 2002 and 2006 USAID Food and Nutrition Technical Assistant (FANTA) reports (Hoddinott and Yohannes, 2002 and Swindale and Bilinsky, 2006) were also used in designing the LMT. While the World Bank guidelines provided a general framework for designing and implementing the LMT, the Ghana Census and GLSS documents provided much more tailored questions that fit the Ghanaian context. USAID’s FANTA reports provided a means to include nutritional status in the computation of household’s aggregate human capital assets (see chapter 7). The LMT was used to gather data for two time periods before the MTS (1999) and approximately ten years into the MTS implementation (2009). The instrument included the following four categories of household human capital assets: a) education and employment, b) nutrition and c) general health and fitness of the household as captured by frequency of illness and disease occurrences among household members. While the data on education and employment, as well as household health and fitness were easily collected for the two time periods before MTS (1999 recall) and after MTS (2009), collecting similar information on the nutritional status of households required some modifications to the USAID’s FANTA guidelines. Using Household Dietary Diversity Score (HDDS) based on consumption of 15 different food groups within a 24-hour period (Hoddinott and Yohannes, 2002 and Swindale and Bilinsky, 2006) without any modification would have restricted this study (much like others) to describing the diversity of foods consumed only within a 24-hour recall period without any reference to periods prior. Hence unless HDDS was collected on the research groups in previous years or time periods, the 24-hour recall approach precluded any meaningful comparisons to periods dating back beyond the 24-hour recall period. Another limitation of FANTA’s HDDS approach was that it does not capture the quantities, frequencies or trends in household’s consumption of 133 the different food groups. In order to correct for these deficiencies in the 2002 and 2006 FANTA guidelines, two additional questions on frequency and general trend in quantity of food items consumed were included following each questions on the 24-hour recall of a particular food group consumed by the household (see pages three and four of LMT). To capture the frequency with which food groups are consumed within a household, the LMT asked respondents to rate on a four point-scale how often each food was consumed at the time of the survey in 2009 (MTS period) and then recall similar trends in 1999 (before MTS). Similarly, the LMT asked households to rate on a five point-scale the general trends in the quantities and levels of sufficiency of the different food groups consumed in in 2009 (MTS period) and 1999 (before MTS). In order to ensure consistency in responses to the modified HDDS, households selected based on a well-defined four-point scale the frequency with which each of the 15 different food groups were consumed in the household. Thus, respondents that said they consumed a particular food groups at least once a week also knew that meant the food group was “always” consumed on the four-point rating scale. Similarly consuming a food group once a month meant “often”, once in six months “occasionally” and not at all or once in a year meant the food group in question was “never” consumed within the household. In order to gauge trends in consumption of different food groups within a household, respondents were asked to first describe consumption and sufficiency trends at the time of the survey in 2009 by stating whether consumption of a particular food group was: a) decreasing but not sufficient, b) increasing but not sufficient, c) decreasing but sufficient, d) stable but sufficient or e) increasing but sufficient. Following the description of consumption trends and level of sufficiency in 2009, respondents were asked to recall similar patterns of consumption in 1999 the period prior to the MTS project. This five-point consumption sufficiency scale is extremely 134 important in gauging not only general trends in quantities of particular food groups consumed within a household but also whether particular food group are sufficient for the household. In a study to monitor the impact of Joint Forest Management (JFM) in India, Pandey (2005, p.18) used a sufficient scale to gauge not only the changing trends in natural capital indicators (area of key NTFPs and their values) but also whether the perceived changes were sufficient for the household. Pandey’s (2005) research provided insight into the use of a “sufficiency scale” in this study to effectively gauge perceived trends in quantities of different food groups consumed by a household and levels of sufficiency for the household. The same sufficiency scale was used in the LMT to elicit responses on perceived trends in quantities and level of sufficiency of natural capital assets such as crop yields and NTFPs harvested, (see LMT page 6), and financial capital assets such as household possessions and cash (LMT page 15). But for the last page that asked survey respondents to provide comments and suggestions, all other questions in the questionnaire were closed ended and responses were recorded on 2-to-10-point point-Likert scales depending on the type of question (Appendix E). The FAO guidelines for analysis of local institutions and livelihoods (Messer and Townsley, 2003) were used to design the three survey instruments used in the focus group interviews (Appendices F and G). 4.9.2 Obtaining IRB UCHRIS Approval To ensure that the rights of all research subjects were protected, the study design together with survey instruments were submitted to Michigan State University’s Institutional Review Board’s (IRB) Committee on Research Involving Human Subjects (UCRIHS). The electronic IRB application together with the survey instruments were submitted on March 31st, 2009 (Appendix A) and approval letter received on June 19, 2009 (Appendix B). Recruiting and training of field enumerators took approximately two months between April and May while field 135 survey begun on June 22, 2009 following IRB approval. The field survey and geo-referenced exercise lasted approximately three months from June 22 to September 30th with occasional follow-ups/verifications exercises occurring up until December of 2009. 4.9.3 Content Validity of Survey Instruments Content validity is the degree to which a test or survey instrument’s content is tied to the instructional domain it intends to measure (OERL, 2008). Generally, a question’s wording is extremely important more so when the instrument is been administered by non-native speakers who have to translate questions to non-native respondents. Researchers have confirmed that slight changes in the way questions are worded can have a significant impact on how people respond to the extent that minor changes in a question’s wording can produce more than a 25 percent difference in people’s opinions (see StatPac, 2008 for more details on content validity of survey instrument). Thus, to ensure facial and content validity, all the survey instruments were reviewed by two Senior Research Scientists from the Forestry Institute of Ghana (FORIG), one senior Forest Extension Officer from the Ghana Forest Services Department (FSD) assigned to the research communities in Yaya. Additionally, six community facilitators previously assigned to the research communities in Nsemre also reviewed the study instruments prior to field testing. The final draft of the survey instruments was field tested among selected households in Asuakwa (one of the ten Yaya communities). Using feedback from the pre-tests, some questions particularly in LMT were revised, others scrapped, and entirely new questions included. By the time, all the corrections were made and the LMT for example was ready to be implemented in the field, the instrument had increased from 14 to 17 pages. All questions were carefully worded to reflect objectivity in the research process, and also to avoid biasing respondents. To help respondents rate their responses, several rating scales 136 mostly ranging from two to five (depending on type question) were provided in survey instrument. Nearly all questions in the LMT were closed ended with the “don’t know” or “neutral” option intentionally omitted to compel survey respondents or households to think critically through each question before providing answers. Also, since the long history of survey research demonstrates that questionnaires without the “don’t know” option are likely to produce a relatively greater volume of accurate data (see also StatPac, 2008). In order to understand the profile of a community in terms of livelihood strategies adopted and the various institutions governing livelihood activities, open ended instruments (Appendices F and G) were used in facilitate focus group discussions aimed. 4.9.4 Yaya Community and Household Selection Since this study investigates livelihood impacts of the MTS project on participating households in Yaya, all 10 MTS communities fringing the Yaya forest reserve were automatically included in the study. Hence in June 2009 completed MTS recruitment forms for Yaya communities were obtained from the Regional Forestry Services Department (FSD) office in Sunyani. Data from the completed recruitment forms suggested that, at the time of this research 456 eligible individuals were recruited from the 10 Yaya communities to participate in the MTS project. A community resource mapping exercise subsequently conducted as part of this research (see chapter 6) suggests that the 456 MTS beneficiaries were recruited from 254 households (see table 4.9.1). Unlike the MTS program that targeted eligible adults in a household (husband, wife and adult children), the community resource mapping exercise geo-referenced the household units, hence the disparity in the number of MTS members recruited by the FSD (456) and the number of MTS households geo-referenced as part of the community resource mapping exercise (254). 137 Since the emphasis of this research is the household (including husband, wife, children, and close relatives), the 254 MTS households became the target population for the study. By focusing sampling efforts on the 254 households and not FSD’s 456 MTS beneficiaries, the study essentially avoided a possible overestimation of the sample size as well as multiple counting of livelihood indicators (e.g., income, household size etc.). A census approach was used in targeting all 254 MTS households for interview. Prior to interviews in each community, several consultative meetings were held with the village Chiefs, Queen Mothers, village Elders and Headsmen as well as the Community Taungya Heads (MTS leader). At each meeting, the study objectives as well household survey protocol and timelines were discussed. Each meeting ended with a tentative agreement on the day and time for the survey exercise that works for the community. The Taungya Heads were responsible for mobilizing all their MTS group members while the Village Chief and Headsmen were responsible for mobilizing all other community members. While community members are often informed at least a week prior to the survey exercise, a member of the survey team visits the particular community the day before the survey to remind the Chiefs, their Headsmen and Community Taungya Heads to mobilize their community members for the exercise. To mobilize community members, most chiefs usually send their young men to each household to inform them about the pending exercise and the need to assemble at a designated spot in the village at the specific time agreed. The community Taungya Head and his Secretary also send reminders to their members to remind them of the day, time and location of the exercise. On each day of the survey exercise the survey team often arrived in the community between eight and nine o’clock in the morning or at a time previously agreed with the Village Chiefs and Taungya Heads. Before the interviews begin, households from the village gather in a 138 designated location for a 30-minute briefing of the study objectives and survey protocol followed by a brief section to address all questions and concerns about the study. After the briefing section, all MTS and non-MTS members are separated and later grouped into their family units (i.e., husband, wife, adult children and relatives) and assigned enumerators. Each enumerator makes a list of his/her assigned family heads and accompanies the first on the list into their home for the interview. The rest of the families are made to disperse and wait in their homes. After each interview, a member of the completed household leads the enumerator to the home of the next family/household on the list of households to be interviewed. Conducting interviews in respondents’ homes gave enumerators the chance to verify pertinent household physical characteristics such as type of materials used in constructing the main structure of the housing unit and roof (clay, cement or thatch), type of bathroom and kitchen construction and materials used in finishing the floors. Since there were only 254 MTS participant households, a census approach was adopted to recruit all 254 members for the study. For non-MTS households in Yaya however, 10% of the 971 (1225 less 254 MTS households) non-MTS household population was targeted (table 4.10 below). Though all 254 MTS households in Yaya were targeted, 203 (80%) were interviewed for the study. For non-MTS households in Yaya, 120 (12%) were interviewed and included in the study. A combination of census and random selection approaches were used in selecting non- MTS households in Yaya for interviews. In relatively smaller communities such as Sewiah, Konsua and Malamkrom with less than 40 non-MTS households the census approach was used in targeting households. Hence all non-MTS households assembled during the 30 minutes community briefing sections (to discuss the study objectives and survey protocol) were targeted. 139 No non-MTS members were interviewed in Konsua because the community leaders insisted that all households in Konsua were members of the MTS program at the time of this research. Though the GIS community resource mapping exercise revealed that 36 non-MTS households existed in Konsua at the time of the research, no attempt was made to interview them. The close proximity of Malamkrom to Konsua presented an opportunity to address the issues surrounding of non-MTS participants in Konsua. As shown by the Yaya Reserve Community Map in figure 4.2, Konsua and Malamkrom amalgamate into each other to the extent that both communities may be treated as one community. It was thus assumed that non-MTS households in both communities were similar in most respects hence the decision to focus on only non-MTS members in Malamkrom. 140 Total no. of households identified Name of Community during No. GIS mapping in Yaya 1 Abrefakrom 2 Ahyiem 3 Amangoase 4 Ayigbekrom 5 Buoku 6 Konsua 7 Malamkrom 8 Amoahkrom 9 Asuakwa 10 Sewiah 156 144 162 192 216 55 53 86 127 34 25 13 34 43 38 19 23 18 24 17 25 13 20 20 29 19 23 18 19 17 100 100 59 47 76 100 100 100 79 100 131 131 128 149 178 36 30 68 103 17 971 16 8 28 9 18 0 9 8 10 14 120 12 6 22 6 10 0 30 12 10 82 12.36 Table 4.9.1: Sample sizes of MTS and non-MTS households in Yaya CFMP/MTS households georeferenced during mapping CFMP/MTS households interviewed Percentage of CFMP/MTS households interviewed. (%) No of non- MTS households (Total no. less 254 MTS households) Percentage No of non- of MTS households interviewed non-MTS households interviewed. (%) Pooled 1,225 254 203 79.92 141 For relatively larger communities such as Abrefakrom, Amangoase and Buoku with more than 120 non-MTS households, a random selection approach was used to select at least 50% of households present during the community briefing sections that preceded the interviews. For the random selection, the Microsoft excel command “=RAND()” was used to generate random numbers between 0 and 1 approximated to the fifth decimal. Using the “Sort” command under the “Data” option in excel, the “=RAND()” column was sorted in ascending order. After sorting the data on “=RAND(),” the first 20 households were then selected for interview. The non-MTS sample sizes for Ahyiem and Ayigbekrom were below the targeted 10% because these were the first communities surveyed at a time when enumerator numbers were only four and below full strength. Furthermore, the four enumerators thought trained on how to administer the survey still needed some more time to master the 17-page instrument (LMT). Hence in addition to spending relatively longer hours on each survey, a few of the completed surveys were discarded due to excessive errors. At the peak of the survey exercise the number of enumerators increased to 15 thus allowing for more households to be interviewed in a relatively shorter time. Because the productivity of enumerators improved significantly as more households were interviewed increasingly more questionnaires were completed under an hour instead of two to three hours at the onset. Table 4.9.1 above presents the sample sizes of MTS and non-MTS households in Yaya. 4.9.5 Selection of Nsemre (Non-MTS) Communities and Households Due to constraints on available funding for this research, the selection criterion for control groups of non-MTS communities was based largely on their proximity to Yaya Reserve communities (the experimental group). Hence only other forest reserves and communities in close proximity with Yaya and specifically within the Wenchi District were considered for 142 inclusion in the study as control groups (see figures 4.4 and 4.5 below). Of the three other reserves in Wenchi District (Nsemre, Sawsaw and the Bui National Park), Nsemre Forest Reserve is the closest to Yaya and shares a common border with Yaya’s North-western portion thus was automatically included in the study. Once selected, the decision of which Nsemre communities to include in the study focused on communities that were situated in close proximity to Nsemre and thus likely to have the most impact on the reserve and vice-versa. During the selection process for Nsemre communities, it was discovered that the nearest communities to Nsemre Reserve were those that also recently received livelihood support from JICA’s five-year PAFORM project hence were also included in the study. 143 Figure 4.1.5: Brong Ahafo Region- Districts and Forest Reserve Maps 144 Figure 4.1.6: Forest Reserves in Brong Ahafo’s Wenchi District Selecting Nsemre Reserve and specifically PAFORM communities guaranteed to an extent that the research and control groups were similar in most regards except for the different approach to forest management implemented by both PAFORM and CFMP/MTS projects. Once the Nsemre communities were selected, households within these communities were randomly selected using the same random selection process used in selecting non-MTS households in Yaya. Like non-MTS households in Yaya, at least 10% of Nsemre Communities’ 244 households were targeted and 25% (61) were interviewed and included in the study. Table 4.9.2 below presents disaggregated statistics of the number of households interviewed in each community and corresponding percentage. 145 Table 4.9.2: Sample sizes of Nsemre households No. Name of Community Total no. of housing units georeferenced during GIS mapping in Nsemre Total no. of households counted during GIS mapping in Nsemre No of Nsemre Percentage of Nsemre Households interviewed households interviewed 11 Asuofre 12 Pepewase 13 Kofitsumkrom 14 Ahwene 46 24 56 66 Pooled 192 90 24 60 70 244 16 11 20 14 61 4.9.6 Selection of Sawsaw (Non-MTS) communities and households (%) 18 46 33 20 25.00 As described previously, one of the criteria for selecting forest reserves to serve as control groups is that they reside within Wenchi District. Aside from residing within Wenchi District, proximity of a reserve and its communities to Yaya was considered another important selection criterion. Hence of the three other forest reserves in Wenchi, Nsemre was selected because it is the closest to Yaya and also shares a physical boundary with Yaya’s Northeastern portion. Between Sawsaw Reserve and Bui National Park, Sawsaw is approximately 10 kilometers from Yaya while Bui National Park is about 80 kilometers (figure 4.6). Sawsaw was thus selected to serve as a second control group based on its relative proximity to Yaya and also the fact that unlike Yaya and Nsemre, Sawsaw communities had no recent history of livelihood support programs from either public or civil society groups. Once the decision was made to include Sawsaw in the study, the next step was deciding on a criterion for selecting Sawsaw communities. Like Yaya and Nsemre, the selection of Sawsaw communities was based on their relative proximity to the Sawsaw Reserve. By focusing on Sawsaw communities in close proximity to the reserve ensured that the study included communities most impacted by the 146 reserve and vice-versa. Unlike Yaya and Nsemre in which surveyed communities surrounded the reserves (figures 4.2 and 4.3), Sawsaw communities were selected along the Sunyani-Wenchi road (figure 4.4). Domeabra the furtherer’s of the five Sawsaw communities lay approximately four kilometers south of the reserve along the Sunyani-Wenchi road while Pipotrim is situated less than half a kilometer from the reserve. It is worth noting also that Papasu community is entirely situated within Sawsaw Reserve. By focusing research efforts on communities along a 20 kilometer stretch of road cutting across Sawsaw enabled the research to include Sawsaw within the limited research budget. Selection of Sawsaw households for interview followed the same protocol as those of Nsemre. Like Nsemre, at least 10% of the 219 households in Sawsaw were targeted for the study and 25% (55) were interviewed. Table 4.9.3 below presents disaggregated statistics of households interviewed in each of the five Sawsaw communities. Table 4.9.3: Sample sizes of Sawsaw households No. Name of Community Total no. of housing units georeferenced during GIS mapping in Sawsaw Total no. of households identified during GIS mapping in Sawsaw No of Sawsaw Percentage of Sawsaw households interviewed households interviewed 15 Ayayo 16 Domeabra 17 Ntema 18 Papasu 19 Pipotrim 71 67 17 12 46 Pooled 213 4.9.7 Focus Group Interviews 74 67 17 12 49 219 11 13 8 9 14 55 (%) 15 19 47 75 29 25.11 In order to capture pertinent information about household livelihoods that might have been missed by the LMT surveys, two focus groups interviews (one for MTS and one for non- 147 MTS members) were conducted in each of the 10 Yaya reserve community. Also, one focus group interview was conducted in each of the nine control communities in Nsemre (4) and Sawsaw (5). Focus group members were selected based on two criteria; a) they must be resident in the community for at least 10 years at the time of this study and b) they must have some stature or wield influence in the community. The selection criteria thus caved the space for the ordinary citizen of each community together with leaders with decision making authority (Chiefs, Queen Mothers, Village Headsmen, Taungya Heads, and Village Elders) to speak on behalf of their communities. Three open-ended survey instruments (Appendices F and G) were used to guide focus group discussions on household livelihood strategies adopted in coping with daily activities as well as the institutions (formal and informal) governing activities within the community. The focus group interviews used open ended questions to facilitate discussions touching on all five livelihood assets outlined in the LMTS. During discussions the group members were often asked to juxtapose actual and perceived changes in livelihood indicators at the time of the research in 2009 (the MTS period) and 1999 (10 years before MTS (1999). In all, a total of 29 tape-recorded focus group interviews were conducted in all 19 research communities. The information gathered from the focus group interviews were used primarily to triangulate some of the key findings from the household surveys. 4.9.8 Managing the Survey and Field Data Collection In order to facilitate timely data collection, I recruited, trained and supervised 15 enumerators. All 15 enumerators were trained on how to use the LMT to collect household level data while five received extra training on how to conduct focus group interviews using the three open-ended focus group instruments (Appendices F and G). In addition to the household level and focus group interviews, four of the 15 enumerators were also trained on using a handheld 148 Garmin GPS unit to capture geo-reference information for community resource mapping. Most importantly, these four enumerators were trained on how to read the satellite information on the Garmin unit to determine the appropriate time to capture waypoints with maximum accuracy. The four enumerators were also trained on how to photograph geo-referenced physical assets and record pertinent household information for further processing. Of the 15 enumerators recruited and trained, five were graduate research assistants from Forestry Research Institute of Ghana (FORIG), two were graduate assistants from the Forest Services Department (FSD) of Ghana, two were students from Kwame Nkrumah University of Science and Technology (KNUST) and six were previously community facilitators on the PAFORM project. All the enumerators were fluent in Twi (the Akan dialect) spoken or understood by most households in the research communities. Two of the enumerators were Dagaaba natives from Wa in Upper West Region and thus spoke fluent Dagaare in addition to Twi. The two Dagaare speakers were thus assigned to the few households that spoke only Dagaare. The front page of all survey instruments provided detailed consent information and a brief instruction on how to complete the questionnaire. Also, the front page of all three instruments used in the focus group interviews had a consent notice that was read aloud by the facilitator to all group members seeking their consent to audio tape the discussions. After reading out the consent information, the group was given the opportunity to select one member to sign the consent form on their behalf. In all the following data was collected from 19 research communities during a three- month data collection period: 1) 878 household level data on livelihood asset indicators, 2) 29 focus group interviews and 3) georeferenced data and photographs on housing units (1,446) and 149 communal physical asset (e.g., churches, schools and drinking water sources). The following equipment was used in field work: two vehicles (one double cabin Nissan Pick-up and one Nisan Pathfinder), one Sony digital camera, one Garmin GPS unit, two Sony voice recorders, six Laptop computers (five Toshiba Satellite computers and one Sony Vaio). After data collection, 10 of the enumerators were retained and paired to enter data from the LMT into excel. Also, three Social Science students from the University of Ghana fluent in Twi and Brong (major languages spoken in the research communities) and one of the Dagaare speaking enumerators transcribed all 29 focus group interviews. I frequently checked on enumerators during data collection in each community to monitor their progress and also address any questions and concerns. In large, dispersed communities such as Buoku and Ayigbekrom with strong phone signals, cellphones provided additional means of communication with enumerators. At the end of each survey section in a community, enumerators exchanged their questionnaire, and each reviewed the another’s completed surveys to check for missing information/responses and inconsistencies. The strategy of reviewing each completed instrument multiple times significantly reduced incidence of missing data and time spent cleaning up the data. 4.9.9 Data Analysis Data from Landsat images, household interviews (LMT surveys and focus groups) and geo-spatial community maps were used to address all four research questions. Research question one addressed changes in land cover in Yaya, Nsemre and Sawsaw forest reserves in Ghana’s Brong Ahafo region (from here-on referred to as region of interest-ROI) before and after the country’s Modified Taungya afforestation program was launched in 2002. To analyze changes in forest cover before and after the MTS, 1990 and 2000 Landsat images as well as 2012 DMC 150 image of the ROI were processed using image processing tools in ArcMap 10.0 and Envi 4.7. Using unsupervised classification tools in Envi, I tracked changed in three land classes (forest cover, settlements and farmlands/open spaces) within the ROI before and after the MTS project. Chapter five describes the methodology and results obtained from the analysis of all three images. The second research question addressed factors that influenced MTS placement in communities and households within the research area. Two sets of data were used in addressing research question two. The first dataset was generated from household level surveys and focus group interviews while the second was generated from geo-spatial community resource maps. To use the first dataset, all completed questionnaires were first numbered, grouped by communities and then entered into Microsoft Excel. Using tools in Excel the data was coded and imported into SPSS and STATA for further analysis. To address factors that influenced MTS placement into community and household (research question two), I used Binomial Probit Models (BPM) generated from household surveys (LMT) and geo-spatial data of household and communal physical assets. To run my BPMs, all “Yes” responses to MTS participation were coded 1 while “Nos” were coded 0. The BPMs were executed with the binary dependent variable (DV) MTS participation regressed against other independent variables (IV). Data generated from the geo-referenced community maps also formed the basis for generating another set of BPMs used in determining the extent to which a community or household’s location relative to major physical assets (e.g., major roads, market, hospitals, water source etc.) influenced MTS program placement. The geo-referenced data was first retrieved from the handheld Garmin GPS unit and then saved in Microsoft Excel following which the data was imported into ArcMap. Community resource maps were then generated in ArcMap 151 highlighting the location of all physical assets within and around a community. While taking geo-referenced points (waypoints) of physical assets, a still camera was also used to capture images of those same physical assets (e.g., houses, boreholes, kiosks, schools, and places of worship) within each research community. All still photographs were then coded in Microsoft Excel and linked to their respective waypoints (Latitudes and Longitudes) before importing into ArcMap for final mapping and analysis. Using different symbols and color codes in ArcMap, community resource maps were produced showing the concentrations of different types of housing units in a community and their relative distances from key communal assets (Appendices K, L, and M). Chapter six describes the methodology and BPM results obtained from the geo-reference data set. The third research question addressed changes in livelihood assets among MTS participant and non-participant household since launching of the program in 2002. Using tools in SPSS and STATA, descriptive statistics (frequencies, means, standard deviations, Chi-squares and Pearson’s p-value) were generated from the LMT survey data. Two sample T-tests were then computed in STATA to compare the differences in within and between group means for each of the five livelihood indexes for the two time periods (1999 and 2009) among MTS participants and non-participants. Essentially the T-tests results examined whether statistically significant differences exist in livelihood indicators among the research groups before and after the MTS program. Using the means from the descriptive analysis, I constructed Household Livelihood Pentagons (HLPs) in Microsoft Excel to compare the between and within group differences in means for each of the five livelihood indicators. House Livelihood pentagons illustrate changes within groups while also comparing differences between groups for the two time periods. 152 Chapter seven describes in greater detail the methodology used in analyzing data from the LMT data analysis and the results obtained from the T-tests and Livelihood pentagons. While the descriptive statistics and HLPs help describe differences in livelihood indexes within and between group before and after the MTS they do not explain how much of the observed changes in livelihood indexes may be attributed directly to the MTS or spillover from the project. Hence chapter eight used a Difference-in-Difference (DID) method to isolate the direct impact of the MTS project on MTS participants as well as the spillover effect on non- participants in project communities. To execute the DID I assumed that the in the absence of treatment, the unobserved differences in livelihood indexes between MTS (treatment) and non- MTS (control) groups will remain unchanged over time. With this assumption I used data from the mean livelihood indexes of MTS and Non-MTS households in the pre-treatment period (1999) to estimate a “normal difference” between MTS and non-MTS groups and then compared that with the difference after treatment in 2009. Since Nsemre and Sawsaw households did not benefit directly form the MTS project, I combined the data from these two non-MTS groups and used that for computing the normal difference for each livelihood index. Using the combined data for Nsemre and Sawsaw to generate normal differences between MTS and non-MTS groups in Yaya I computed the direct and spillover effects of the MTS project on households in Yaya. Chapter eight in greater detail the methodology used in executing the DID and the results obtained. This chapter provided a brief description of the methods used in generating and analyzing data to address all four research questions in this dissertation. Chapters five through eight details how each dataset was collected, analyzed and the results obtained from each analysis. 153 APPENDICES 154 APPENDIX A: Initial Application to MSU Internal Review Board (IRB) APPLICATION FOR INITIAL REVIEW APPROVAL OF A PROJECT INVOLVING HUMAN SUBJECTS Biomedical, Health Sciences Institutional Review Board (BIRB) Social Science, Behavioral, Education Institutional Review Board (SIRB) 202 Olds Hall, Michigan State University East Lansing, MI 48824-1047 Phone: (517) 355-2180 Fax: (517) 432-4503 E-mail: irb@msu.edu Office Hours: M-F (8:00 A.M.-5:00 P.M.) IRB#: ID# 155 156 157 158 159 160 161 162 APPENDIX B: IRB Approval Letter 163 APPENDIX C: Letter of Collaboration to the Ghana Forestry Research Institute 164 APPENDIX D: Livelihoods Monitoring Tool (LMT) The Livelihood Monitoring Tool (LMT) below will be used to quantify changes in households’ livelihood indicators for two time periods, 1999 and 2009 (i.e., before and after the implementation of MTS in the Yaya reserve). A similar yet less extensive version of this LMT was used by Pandey et. al., (2005) to assess stakeholders’ perceptions of changes in livelihoods as a result of the implementation of JFM in Rajasthan, India. In the current LMT, each livelihood asset is assigned a weight of either: 4, 3, 2, or 1 indicating the degree of improvement or deterioration (i.e., 4=best, 2=moderate and 1=worst) in livelihood indicators before (1999) and after (2009) the MTS implementation. Cumulative scores for each livelihood asset group provides a means of assessing changes in livelihood indicators based on households’ perceptions (i.e., whether HH are generally better-off, worst-off or simply remained the same for the two time periods in question). In order to determine the impact of MTS, each HH is asked to rate on a four-point Likert scale the level of change in the HH’s assets/indicators for the two time periods 1999 and 2009. Name of Enumerator: ____________________________________ Day and Time of Interview________________________________ 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 APPENDIX E: Community Profile Instrument Consent Form This Community Profile (CP) is been conducted as part of the broader study to determine Livelihood Strategies adopted by communities fringing the Yaya Forest Reserve. Conducting a CP will help the researchers determine the most effective strategy to improve livelihood assets in Yaya communities while protecting the Yaya Forest Reserve. Limited resources as well as time constraints necessitate the conducting of a community profile. The purpose for undertaking this community profile is to: 1) help the researchers focus on relevant household strategies, b) help the researchers determine which local institutions play a key role in shaping household livelihood strategies and hence worthy of further investigation, c) help the researchers identify the linkage between household livelihoods and local institutions as well as the context within which both operate. The checklist below is used to guide the discussions during the CP process. The interview will last for about 45 minutes to 1 hour. Please feel free to talk openly. You do not have to answer any question you don’t like. If you choose to participate, all information you provide shall remain confidential and the investigators will protect your privacy to the maximum extent allowed by law. This means that your name will not be associated with any information you provide. Aliases will be used in all reports arising from the study. Your participation in the study is completely voluntary. You may choose not to participate at all, or decline to participate at any point. It is expected that participation in the study is associated with minimal risk for you as your name will be known only to the investigator. There are no known benefits that will accrue to you individually as a result of your participation in the study. If you have any questions or concerns during the course of the study, please contact: Doe Adovor at adovordo@msu.edu or +233241572195 or 001-517-410- 9225 or 208 Natural Resource Building, East Lansing, MI 48824, USA or Dr. Ernest Foli at efoli@csir-forig.org.gh, +233-51-60123/60373. Your written consent indicates your voluntary agreement to participate in this study. Please select one person in the group to sign the consent form. Signature _________________________ Since I don’t want to miss anything we talk about, would it be alright if I tape-record our conversation? This will help me since it is hard for me to listen and write at the same time. Initial for tape recorder consent_______________ Would it be ok to contact you afterwards if I have any further questions or to review our findings? Initial to be contacted after this interview________________ ___________________ Date 182 • What formal organizations and associations are in the community? • What rules regulations and customs are in place? • Who is affected by them and how? 5. Community infrastructure • What services are available in the community (transport, communication, power and water supply, waste e.g. sewage and garbage disposal, health service e.g. hospitals/clinics, markets, agricultural extension, education e.g. primary, JSS, SS etc.? • Who has access to these services? • How expensive are the user fees for these services? 6. Community history • How long has the community been in existence and how was it founded? • When did different social, economic, ethnic and groups settle in the community? • How has the community changed over time and what has caused those changes? FAO, Checklist for community profiling (Messer and Towsley, 2003, p.29) 1. Resources • What is the principal or main natural resource available to the community? • Who uses them and how are they used? • Where are they located? 2. Livelihoods • What are the different activities that households in the community use to support their livelihoods? • Who is involved in those livelihood activities? (Men/women, young/old, different social and economic groups etc.) • When do those activities take place (time of day/month/seasons) and where? 3. Community structure • How many people and households live in the community? • What is the gender composition and age structure of the community? • What different social, economic, ethnic and cultural groups are in the community? • How are those groups defined? • Where do those different social, economic and ethnic and cultural groups live? 4. Local institutions 183 APPENDIX F: Household Livelihood Strategies Consent Form The purpose of the study is to determine Household Livelihood Strategies (HLS) adopted by households fringing the Yaya Forest Reserve. This study is been conducted as part of a broader study to determine the most effective strategy to improve livelihood assets in Yaya communities while protecting the Yaya Forest Reserve. In order to analyze in-depth the Livelihood Strategies employed by households living around the Yaya forest Reserve, the study will use a check list of 8 items to guide this focus group interview. You have been identified as individuals that have lived in this community for considerable length of time (e.g. >10 years) and have ample knowledge of the community, household and the Yaya Forest Reserve. You are requested to participate in an interview for this study. The interview will last for about 45 minutes to 1 hour. Please feel free to talk openly. You do not have to answer any question you don’t like. If you choose to participate, all information you provide shall remain confidential and the investigators will protect your privacy to the maximum extent allowed by law. This means that your name will not be associated with any information you provide. Aliases will be used in all reports arising from the study. Your participation in the study is completely voluntary. You may choose not to participate at all, or decline to participate at any point. It is expected that participation in the study is associated with minimal risk for you as your name will be known only to the investigator. There are no known benefits that will accrue to you individually as a result of your participation in the study. If you have any questions or concerns during the course of the study, please contact: Doe Adovor at adovordo@msu.edu or +233241572195 or 001-517-410- 9225 or 208 Natural Resource Building, East Lansing, MI 48824, USA or Dr. Ernest Foli at efoli@csir-forig.org.gh, +233-51-60123/60373. Your written consent indicates your voluntary agreement to participate in this study. Please select one person in the group to sign the consent form. Signature _________________________ Since I don’t want to miss anything we talk about, would it be alright if I tape-record our conversation? This will help me since it is hard for me to listen and write at the same time. Initial for tape recorder consent_______________ Would it be ok to contact you afterwards if I have any further questions or to review our findings? Initial to be contacted after this interview________________ ___________________ Date 184 HOUSEHOLD LIVELIHOOD (HH) STRATEGIES 1. Household information • Number of household members, sex, age, ethnic group, health status, (disabilities etc.), residency status, role in different livelihood activities 2. Human Capital • What is the educational status of household members? • Percentage of school age children attending school • What skills, capacity, knowledge and experience do different household members have (training, labor capacity etc.) • Number of people that work outside of the village on a daily basis per capita • Number of people that migrate from the village to work outside for a long period of time (e.g. 6- 12month at a time or more) • Infant mortality per capita • Physical health of HH members (are there any physical or mentally disabled HH members?) 3. Natural Capital • What land, water, livestock or forest resources do household members use? • What do they use them for? • What are the terms of access (ownership, rental, share arrangements, open access, leasing) • Area of key NTFP/capita • Area under MTS/capita • Average time spent collecting fuel wood per household per month? • Average time spent collecting water per household per month? • Average time spent collecting fodder/capita • Value of annual firewood production/capita • Value of annual NTFP production/capita • Annual food grain production/capita 4. Physical Capital • Types of housing: o Main structure: (Wood, metal, cement/concrete, earth/mud bricks etc) o Roofing: (Corrugated, slate or asbestos, cement, tiles, thatch, palm or raffia leaves etc) o Floor: Concrete, wood, earth/mud o Power supply (Electric, lantern, etc.) 185 • What infrastructure do household members have access to and use (transport, marketing facilities, health services, water supply, schools etc)? • What infrastructure do they not have access to and why? • What are the terms of access to different infrastructure (payment, open access, individual or pooled etc) • What tools or equipments do different household members use during different livelihood activities and what are the terms of access to them (ownership, hire, sharing, etc.)? • Means of transport (Does the household own the following: car, motorcycle, bicycle, donkey, horse, etc)? If so how many? • What is the average time traveled to nearest paved road? • What is the average time traveled to nearest market? • Area of own or communal land under irrigation (or irrigated land/capita) • Number of functioning tractors/capita? • What is the source of water supply and number of functioning sources per capita (tap, bore holes pumps, dug out wells etc.)? 5. Financial Capital • What are the earning of the household from different sources (e.g. income generating activities, remittances, etc) • Wages from forestry activities for example from MTS (3year average) • Savings account and total money deposited • What other sources of finances are available and how important are they (e.g. Bank credits, NGO support, etc.)? • Number of shops selling consumer goods • Average price of the 3 most expensive items in the shop 6. Social Capital • What links that the household have with other households or individuals in the community (kinship, social group, membership of organizations, political contacts, patronage etc.)? • In what situations do those links become important and how (mutual assistance, pooling, labor etc.) • Proportion of adults involved in MTS • Proportion of MTS participants that are women • Number of MTS meetings and attendance • • Is there a micro-credit or self-help group in the village or community? Is there any collective selling of agricultural or forest products? • Has collective selling resulted in improved prices? 7. Vulnerability Context 186 • What are the seasonal patterns of some of different activities that individuals are engaged in? • What seasonal patterns are there in food supply, income, expenditure, residence, etc.? • What crisis has the household faced in the past (health crisis, natural disasters, crop failures, civil unrest, legal problems, indebtedness, etc.) and how did they deal with them? • What longer-term changes have taken place in the household’s natural, economic and social environment and how have they dealt with these changes? 8. Policies, Institutions and Processes • What organizations, institutions, and associations (societies, cooperatives, political parties etc,) do household members participate in and what role do they play in them? • How are decisions reached within these organizations, institutions, and associations? • Who makes decisions about use of natural and physical resources in the community and how are those decisions reached (what are the centers of decision making)? • What laws, rules and regulations affect the household? 187 APPENDIX G: Institutional Profiles Consent Form The purpose of the study is to determine Household Livelihood Strategies (HLS) adopted by households fringing the Yaya Forest Reserve. This study is been conducted as part of a broader study to determine the most effective strategy to improve livelihood assets in Yaya communities while protecting the Yaya Forest Reserve. In order to analyze in-depth, the Livelihood Strategies employed by households living around the Yaya forest Reserve, the study will use a check list below to guide this focus group interview. You have been identified as individuals that have lived in this community for considerable length of time (e.g., >10 years) and have ample knowledge of the community, household, and various institutions within the Yaya communities. You are requested to participate in an interview to help profile the key institutions such as the MTS in your community. The interview will last for about 45 minutes to 1 hour. Please feel free to talk openly. You do not have to answer any question you don’t like. If you choose to participate, all information you provide shall remain confidential and the investigators will protect your privacy to the maximum extent allowed by law. This means that your name will not be associated with any information you provide. Aliases will be used in all reports arising from the study. Your participation in the study is completely voluntary. You may choose not to participate at all or decline to participate at any point. It is expected that participation in the study is associated with minimal risk for you as your name will be known only to the investigator. There are no known benefits that will accrue to you individually as a result of your participation in the study. If you have any questions or concerns during the course of the study, please contact: Doe Adovor at adovordo@msu.edu or +233241572195 or 001-517-410-9225 or 208 Natural Resource Building, East Lansing, MI 48824, USA or Dr. Ernest Foli at efoli@csir-forig.org.gh, +233-51-60123/60373. Your written consent indicates your voluntary agreement to participate in this study. Please select one person in the group to sign the consent form. Signature ___________________ Since I do not want to miss anything we talk about, would it be alright if I tape-record our conversation? This will help me since it is hard for me to listen and write at the same time. Please initial for tape recorder consent____________ Would it be ok to contact you afterwards if I have any further questions or to review our findings? Sign or place your initials in the line if it is okay to contact you after this interview________________ ___________________ Date 188 INSTITUTIONAL PROFILE This survey on institutional profile was developed by Messer and Townsley, (Rural Development Division of the FAO, 2003, p.16-63)) Legality ➢ What is the legal status of the institution? E.g. Legitimacy ➢ How and when did the institution or organization originate? • Does it have an official legal ➢ What sort of local support does the status? Is it registered? • ➢ How was that status determined? E.g. • By a policy decision? • By legislation • By registration ➢ Who was involved in establishing that legal status? Formality ➢ What procedures or formally established of behavior does the institution or organization have? ➢ What formal roles and tasks are established within the organization? ➢ How are meetings called? ➢ How often are meetings called? ➢ Are minutes recorded during meetings? ➢ Who decides procedures? ➢ Who calls the meetings? institution or organization command and why? ➢ Who initiated the creation of the institution or organization? E.g. • Local people? • Local leaders? • Outsiders (NGO, government etc)? ➢ Who regards the institution as legitimate? Informality ➢ What role is played by different informal rules or processes? E.g. • Gender? • Kinship? • Class? • Social status? • Ethnic group? ➢ How do these informal rules affect what the institution or organization does? E.g. • Do they influence who comes to meetings? • Do they influence who talks at meetings? • By registration ➢ Who establishes or influences informal rules or processes? Level Geographic coverage 189 ➢ At what level does the institution or organization operate? E.g. • Family? • Lineage, clan or tribe? • Professional group? • Community • • Women or men (Gender) Inter-community? ➢ Who or what determines the level at which the institution or organization operates? ➢ What area does the institution or organization cover? E.g. • Neighborhood? • Village? • Beyond the village? ➢ How (and whom) is the coverage of the institution or organization determined? “Objectives” and “Activities” Capabilities Willingness ➢ What are the stated objectives of the institution or organization? ➢ What is the capacity of the institution or organization to reach those objectives? ➢ Are the objectives realistic when compared to its capacity? ➢ Who is involved in establishing, changing or influencing the objectives of the institution and its capacity to achieve those objectives? ➢ Do leaders and community members sometimes disagree on the management of the institution or organization? ➢ What commitment is there on the part of the institution or organization and its members to achieve its objectives and to follow its rules? ➢ Are the names of members and their rights and duties posted on the village council door, or elsewhere? ➢ Does the institution or organization have a ‘vision’? If yes, is it stated or expressed anywhere? ➢ Who is involved in influencing the commitment of the institution to achieving its objectives? Mandated objectives and activities Ad-hoc objectives and activities 190 ➢ What objectives and activities does the ➢ What objectives or activities besides institution or organization have a mandate to achieve or carry out? ➢ How was the mandate established? E.g. • By government policy? • By local consensus? • By established practice? ➢ Who gave them that mandate? E.g. • Government? • Traditional authorities? • Local people? Actual activities ➢ How does the institution or organization achieve or try to achieve its objectives? ➢ What activities does the organization undertake now? ➢ Who participates in activities and who determines where, how and when activities are carried out? the stated ones have emerged over time? ➢ How have they been addressed? ➢ Does the institution or organization defend the interest of a particular group of people (Does it play an advocacy role)? For whom? ➢ Who is involved in establishing these ad-hoc objectives and activities? Future activities ➢ How does it plan to achieve objectives in the future? ➢ What activities are planned for the future? ➢ Who will participate in future activities? ➢ Who determines what future activities will be undertaken? “Membership and Participation” Conditions of participation Contributions ➢ What are the conditions for membership? ➢ Does membership in this institution or organization exclude membership elsewhere? ➢ Who participates in the institution or organization? ➢ What fees or other forms of contributions are expected from members? E.g. • Lump sum investment, food, time, charitable contributions, sharing of land, water, animals, labor etc? ➢ Who is excluded from the institution or inheritable? ➢ Is membership transferable or organization and why? 191 ➢ How are these contributions determined? Rules of the game ➢ How when and by whom were the rules established to determine who benefits (most/least)? ➢ How, if at all, do women participate in the institution or organization? ➢ What form of patronage and protection, if any, does the institution or organization provide? ➢ Who are the institutions or the organization’s main beneficiaries? ➢ Who decides on how the benefits from the institution are distributed? ➢ Who determines these contributions? ➢ Who collects the contributions? ➢ Who decides how contributions are used? Non-working rules and sanctions ➢ Are some rule applied differently to different people? ➢ Are there any rules that are no longer working or applied? ➢ What sanctions are there in place for not following the rules? ➢ How are they enforced? By whom? ➢ Are they applied to everyone in the same way? ➢ How often have they been applied in the past? ➢ Who decides on and enforces these rules and sanctions? ➢ Who had been subject to them now and in the past? 192 APPENDIX H: Schedule of MTS Benefit Sharing Agreement Schedule A A copy of the mapped area shall be attached to this agreement as Schedule A. As part of the MTS benefit sharing agreement signed between the MTS project and beneficiaries/farmers, a copy of the mapped area highlighting the specific forest allocated to the farmer is included as part of the agreement. Schedule A of the contract thus includes a copy of this map. Below is a generic copy of the Yaya Reserve map that is included under Schedule A. 193 Particulars of Individual Members of………………………………………………… Taungya Farmer Group Schedule B PICTURE TO BE AFFIXED HERE : ……………………………………………………………. : …………………………………………………………….. 1. Name of Farmer : ……………………………………………………………. 2. Age 3. Sex 4. Village 5. Address 6. Hometown 7. Total Area allocated: ……………………………………………………………. 8. Name of Next of Kin (Successor): …………………………………..………….. 9. Signature/Thumbprint of Farmer: …..…………………………………………. : ……………………………………………………………… ……………………………………………………………... : …………………………………………………………….. : …………………………………………………………….. Signature/Thumbprint of Witness (Taungya Farmer Group Executive Member): ……………………………………………… 194 This Schedule outlines the benefit flow for Tectona grandis (Teak) over the production cycle of the Modified Taungya plantation investment Schedule C Activity Farmer Benefits Investor/FC Benefits Landowner Benefits Local Community Agricultural cropping and Proceeds from agricultural tending of plantation cropping - - Benefits - 40% of thinning proceeds 40% of thinning 15% of thinning 5% of thinning (no benefits if plantation is proceeds proceeds proceeds abandoned before 1st thinning) - - - - 40% of thinning proceeds 40% of thinning 15% of thinning 5% of thinning - proceeds proceeds proceeds - - - Production cycle sequence (Years) 0-4 Year 5 6-10 Years 10-12 Year 18 1st Thinning Tending of Plantation 2nd Thinning 13-17 Tending of plantation 3rd Thinning 40% of thinning proceeds 40% of thinning 15% of thinning 5% of thinning 19-24 Tending of plantation Year 25 - proceeds proceeds proceeds - - - 5% of thinning (Final Harvest) Clear felling of trees 40% of final harvest proceeds 40% of final harvest 15% of final harvest proceeds proceeds proceeds - 195 APPENDIX I: Demographic Information of MTS Communities Table 4.9.4: Number of households selected in MTS communities Community Abrefakrom Ahyiem Amangoasekrom Amoakrom Asawakwaa Ayigbekrom Buoku Konsua Malamkrom Sewiah Total No. of MTS household Percent No. of MTS household Percent 34 16 49 35 40 73 58 43 52 52 452 7.5 3.5 10.8 7.7 8.8 16.2 12.8 9.5 11.5 11.5 100 34 16 49 35 40 73 58 43 52 - 402 8.5 4.0 12.3 8.8 10.0 18.3 14.5 10.8 13.0 - 100 196 Table 4.9.5: Gender, Marital, Residential Status and education of selected MTS Participants Gender (N=402) Frequency (n) Percent Male Female Total Missing System 249 151 400 2 61.9 37.6 99.5 0.5 Marital Status(N=402) Frequency (n) Percent Married Single Total Missing System 336 54 390 12 83.6 13.4 97.0 3.0 Resident Status(N=402) Frequency (n) Percent Migrant Native Total Missing System 217 173 390 12 54.0 43.0 97.0 3.0 Educational Status(N=402) Frequency (n) Percent No Formal and Informal Primary Middle JSS SSS or Higher Total Missing System 191 35 88 31 14 359 43 47.5 8.7 21.9 7.7 3.5 89.3 10.7 Table 4.9.6: Average Age, Household size and Year’s resident in the Brong Ahafo Region (N=402) Age Household Size Years in Brong Ahafo Frequenc y (n) 398 398 214 Minimum Maximum Mean Std. Deviation 18.00 1.00 1.00 77.00 15.00 75.00 43.5653 6.4523 15.6729 11.84724 2.83225 12.35680 197 APPENDIX J: Descriptive Statistics from Community Mapping Exercise Table 4.9.7: Population Information from Field GPS Exercise No. Name of Community No. of males No. of females older than 12 years older than 12 years No. of males less than 12 years No. of females less than 12 years Total Population 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Yaya Reserve 1,885 1,806 1,753 1,670 Abrefakrom Ahyiem Amangoase Ayigbekrom Buoku Konsua Malamkrom Amoahkrom Asuakwa Sewiah Sawsaw Reserve Ayayo Domeabra Ntema Papasu Pipotrim 235 235 265 315 413 63 73 129 114 43 329 115 92 29 17 76 Nsemre Reserve 309 Asuofre Pepewase Kofitsumkrom Ahwene 105 - 91 113 213 203 285 281 393 62 70 118 143 38 305 101 95 32 11 66 277 81 - 88 108 217 211 251 279 360 62 69 135 133 36 386 129 152 37 16 52 228 73 - 69 86 195 186 271 237 355 74 68 126 122 36 396 118 177 24 11 66 259 79 - 69 111 7,114 860 835 1,072 1,112 1,521 261 280 508 512 153 1,416 463 516 122 55 260 1,073 338 - 317 418 Pooled 2,523 2,388 2,367 2,325 9,603 198 Table 4.9.8: Population Densities No. of No. of No of family or households bedrooms in housing units 1225 units 3027 No. Name of Community No of housing units Yaya Reserve 1041 1 Abrefakrom 2 Ahyiem 3 Amangoase 4 Ayigbekrom 5 Buoku 6 Konsua 7 Malamkrom 8 Amoahkrom 9 Asuakwa 10 Sewiah 136 133 158 171 190 48 53 82 50 20 Sawsaw Reserve 213 11 Ayayo 12 Domeabra 13 Ntema 14 Papasu 15 Pipotrim 71 67 17 12 46 156 144 162 192 216 55 53 86 127 34 219 74 67 17 12 49 housing to families ratio 0.85 0.87 0.92 0.98 0.89 0.88 0.87 1.00 0.95 0.39 0.59 No of bedroom to family ratio 2.47 2.29 2.28 2.74 2.42 3.07 2.11 2.72 2.60 1.68 2.15 0.97 3.16 0.96 1.00 1.00 1.00 0.94 2.59 3.70 3.59 2.08 3.37 0.79 2.11 0.51 1.00 0.93 0.94 1.87 2.40 2.89 357 328 444 465 663 116 144 224 213 73 691 192 248 61 25 165 514 168 144 202 Nsemre Reserve 192 244 16 Asuofre 17 Pepewase 18 Kofitsumkrom 19 Ahwene 46 24 56 66 90 24 60 70 Pooled 1,446 1,688 4,232 0.86 2.51 199 Table 4.9.9: Household Livestock Ownership Information No. Name of Community No of housing units Poultry Birds (0= No 1=Yes) Goat 0=No 1=Yes Pig (0=No 1=Yes) Cattle (0=No 1=Yes) Yaya Reserve 1,041 0.5053 0.3852 0.0250 0.0038 1 Abrefakrom 2 Ahyiem 3 Amangoase 4 Ayigbekrom 5 Buoku 6 Konsua 7 Malamkrom 8 Amoahkrom 9 Asuakwa 10 Sewiah 136 133 158 171 190 48 53 82 50 20 0.4265 0.2105 0.6329 0.6316 0.3684 0.5625 0.8113 0.5610 0.6400 0.7000 0.3162 0.2632 0.6076 0.3509 0.1947 0.3958 0.7736 0.4878 0.4600 0.3500 0.0221 0.0977 0.0063 0.0058 0.0000 0.1042 0.0000 0.0000 0.0200 0.1000 0.0147 0.0075 0.0000 0.0058 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Sawsaw Reserve 213 0.4178 0.5023 0.0141 0.0000 11 Ayayo 12 Domeabra 13 Ntema 14 Papasu 15 Pipotrim 71 67 17 12 46 0.3662 0.5075 0.6471 0.5833 0.2391 0.4648 0.7015 0.3529 0.5000 0.3261 0.0000 0.0299 0.0000 0.0000 0.0217 0.0000 0.0000 0.0000 0.0000 0.0000 Nsemre Reserve 192 0.4010 0.1771 0.0000 0.0000 16 Asuofre 17 Pepewase 18 Kofitsumkrom 19 Ahwene 46 24 56 66 0.5000 1.0000 0.2321 0.2576 0.0870 0.0000 0.2857 0.2121 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Pooled 1,446 0.4786 0.3748 0.0201 0.0028 200 Table 4.9.10: Type of Housing Infrastructure No. Name of Community No of housin g units Type of Roof Housing structure (0= Raffia 1=Metal) 0=Mud 1=Cemen t House wired for electricity (0=Not wired 1=Wired) Type of bathroom floor (0=Mud/ gravel 1= Cemented) Type of kitchen 0= No kitchen 1= Secure roofed cooking space Yaya Reserve 1,041 0.5677 0.4966 0.2171 0.2133 0.4755 1 Abrefakrom 2 Ahyiem 3 Amangoase 4 Ayigbekrom 5 Buoku 6 Konsua 7 Malamkrom 8 Amoahkrom 9 Asuakwa 10 Sewiah 136 133 158 171 190 48 53 82 50 20 0.3676 0.3459 0.9051 0.5029 0.8684 0.4375 0.2453 0.1341 0.8600 0.6500 0.3235 0.1504 0.8354 0.4094 0.8368 0.3125 0.2264 0.2805 0.7400 0.2500 0.0000 0.0000 0.5823 0.0000 0.4947 0.0000 0.0000 0.0122 0.7800 0.0000 0.1765 0.0602 0.1772 0.1462 0.4211 0.2500 0.0755 0.0854 0.6400 0.1000 Sawsaw Reserve 213 0.1221 0.1033 0.0000 0.0376 11 Ayayo 12 Domeabra 13 Ntema 14 Papasu 15 Pipotrim 71 67 17 12 46 0.1127 0.1343 0.1176 0.0000 0.1522 0.0704 0.1194 0.1765 0.0000 0.1304 0.0000 0.0000 0.0000 0.0000 0.0000 0.0704 0.0149 0.0588 0.0000 0.0217 0.3456 0.2857 0.3734 0.4620 0.6053 0.6458 0.3208 0.6463 0.8000 0.8000 0.4789 0.4366 0.4925 0.6471 0.2500 0.5217 Nsemre Reserve 192 0.3073 0.2135 0.0000 0.0729 0.5677 16 Asuofre 17 Pepewase 18 Kofitsumkrom 19 Ahwene Pooled 46 24 56 66 0.4565 0.1667 0.0714 0.4545 0.2174 0.0833 0.1429 0.3182 1,446 0.4675 0.4011 0.0000 0.0000 0.0000 0.0000 0.1563 0.0217 0.1250 0.1071 0.0606 0.1687 0.6087 0.4583 0.6250 0.5303 0.4882 201 Table 4.9.11: Proximity of Household to Communal Asset No. Name of Community No of housing units Corn mill /Nika- Nika Proximity of household to communal assets 0=None, 1=>180meters, 2=120-180meters 3=60-120meters, 4=0-60meters Church Mosque Borehole JSS KG Primary Yaya Reserve 1,041 2.1758 1.2978 1.4573 2.0019 0.3698 0.9683 1.3391 1 Abrefakrom 2 Ahyiem 3 Amangoase 4 Ayigbekrom 5 Buoku 6 Konsua 7 Malamkrom 8 Amoahkrom 9 Asuakwa 10 Sewiah 136 133 158 171 190 48 53 82 50 20 1.6324 0.0000 1.7868 2.1765 1.0294 0.0000 1.1912 2.1128 0.0000 1.8872 1.8195 0.0000 1.4812 1.2105 2.5443 1.6519 1.8671 2.5380 0.0000 0.0000 1.8101 2.0292 2.2398 1.6550 2.1579 0.0000 1.4503 0.0000 2.6368 2.4632 1.6842 1.5263 1.0000 1.7421 1.9421 0.0000 2.1875 0.0000 1.9375 0.0000 0.0000 1.1667 2.5849 0.0000 2.3585 2.4906 0.0000 0.0000 2.6226 2.6585 0.0000 0.0000 1.2927 0.0000 1.9024 1.6463 2.6200 2.6800 0.0000 3.1000 1.1000 1.5200 1.7200 1.3000 0.0000 0.0000 1.0000 0.0000 0.0000 0.0000 Sawsaw Reserve 213 2.4038 1.1925 1.2441 1.5117 0.0000 1.1549 1.1274 11 Ayayo 12 Domeabra 13 Ntema 14 Papasu 15 Pipotrim 71 67 17 12 46 2.4789 1.7042 1.9155 2.0282 0.0000 1.9859 2.0286 2.3284 0.0000 0.0000 2.6567 0.0000 0.0000 0.0000 2.4706 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 3.0000 2.8913 2.8043 0.0000 0.0000 2.2826 2.1087 Nsemre Reserve 192 1.8438 2.2760 0.9375 2.4063 0.0000 0.8021 0.5677 16 Asuofre 17 Pepewase 18 Kofitsumkrom 19 Ahwene 46 24 56 66 2.7391 2.6739 0.0000 1.7391 0.0000 0.0000 1.1522 0.0000 1.4583 0.0000 2.4167 0.0000 0.0000 0.0000 0.0000 2.2857 0.0000 1.9643 0.0000 0.0000 1.0000 3.4545 2.2879 2.7273 3.2424 0.0000 2.3333 0.0000 Pooled 1,446 2.1653 1.4122 1.3568 1.9834 0.2663 0.9737 1.2055 202 No. Name of Community No of housing units Table 4.9.11 (cont’d) Proximity of household to communal assets 0=None, 1=>180meters, 2=120-180meters 3=60-120meters, 4=0-60meters Dumpste r Market Kiosk Toilet Major Road Business Center Pharmacy Yaya Reserve 1,041 1.0298 1.0471 2.9462 1.8934 2.5072 0.4265 0.6369 1 Abrefakrom 2 Ahyiem 3 Amangoase 4 Ayigbekrom 5 Buoku 6 Konsua 7 Malamkrom 8 Amoahkrom 9 Asuakwa 10 Sewiah 136 133 158 171 190 48 53 82 50 20 1.8382 1.2426 2.8897 1.5221 1.7426 0.0000 0.0000 1.6165 2.4511 1.1654 1.9624 0.0000 1.1899 0.0000 2.8987 2.0000 2.2911 0.0000 1.7778 0.0000 3.3567 2.2047 2.8363 0.0000 1.7368 2.0526 3.0316 2.0789 2.7158 2.3368 0.0000 0.0000 3.5833 2.5000 2.8125 0.0000 0.0000 0.0000 0.0000 2.2073 2.5660 2.9634 1.9245 2.2317 2.9245 2.7683 0.0000 0.0000 0.0000 2.7000 3.7800 2.3200 3.8000 0.0000 0.0000 0.0000 0.0000 0.0000 2.1000 0.0000 0.0000 0.0000 1.7089 0.0000 2.0684 0.0000 0.0000 0.0000 0.0000 0.0000 Sawsaw Reserve 213 1.5962 0.5775 2.0000 0.0000 2.7225 0.0000 0.6338 11 Ayayo 12 Domeabra 13 Ntema 14 Papasu 15 Pipotrim 71 67 17 12 46 1.6761 1.7324 1.9859 0.0000 3.4789 0.0000 1.6119 0.0000 1.9851 0.0000 1.8806 0.0000 0.0000 0.0000 0.0000 0.0000 2.6923 0.0000 0.0000 0.0000 0.0000 0.0000 3.9167 0.0000 2.4565 0.0000 3.3043 0.0000 2.4783 0.0000 1.9014 0.0000 0.0000 0.0000 0.0000 Nsemre Reserve 192 0.0000 0.0000 0.0000 0.9740 3.1771 0.0000 0.0000 16 Asuofre 17 Pepewase 18 Kofitsumkrom 19 Ahwene Pooled 46 24 56 66 0.0000 0.0000 0.0000 1.5870 3.4130 0.0000 0.0000 0.0000 0.0000 0.0000 2.1250 0.0000 0.0000 0.0000 0.0000 0.0000 3.7500 0.0000 0.0000 0.0000 0.0000 1.7273 2.9091 0.0000 0.0000 0.0000 0.0000 0.0000 1,446 0.9765 0.8389 2.4156 1.4924 2.6276 0.3071 0.5519 203 CHAPTER 5: LANDCOVER CHANGE DETECTION IN YAYA, SAWSAW AND NSEMRE FOREST RESERVES 204 Abstract To determine land-cover changes in Yaya, Nsemre and Sawsaw forest reserves, unsupervised land-cover classification was carried-out using ENVI 4.7. Remotely sensed Landsat images from January 1990 and 2000 and a Disaster Monitoring Constellation (DMC) image from January 2012 were processed. The aggregate post-classification results of all three reserves, indicates that a total of 5.60 Km2 and 2.15 Km2 of land classified as forest and farmlands respectively were lost to settlements and open areas/bare ground between 1990 and 2000 (i.e. before the MTS). Between 2000 and 2012 however (the period of MTS implementation), aggregate forest cover increased by 15.33 Km2. Of the total net gains in forest cover between 2000 and 2012, approximately 82% (12.64 Km2) came from settlements, open areas, shrub and grassland while 18% (2.69 Km2) came from farmland. The disaggregated classification results suggest that between 1990 and 2000, Yaya experienced a net gain of 2.56 Km2 (6.3%) in forest cover while both Nsemre and Sawsaw experienced loses of 3.67 Km2 (18.06%) and 7.33 Km2 (12.81%) respectively. Between 2000 and 2012 all three reserves experienced net gains in forest cover with Nsemre recording the largest percentage gain of 27% (5.28 Km2) compared to 15.52% (6.30 Km2) for Yaya and 6.47% (3.69 Km2) for Sawsaw. Knowing the extent of land-cover change for the different land-classes within Yaya, Nsemre and Sawsaw forest reserves before and after the MTS program is the first step to understand the impact of the program on natural capital and other livelihood assets held by the forest communities. 5.1 Landcover change in Yaya, Nsemre and Sawsaw Forest Reserves This dissertation investigated the impact of MTS not only on household livelihoods but also on land-cover change with particular attention on the forest. Chapter five thus use 1990 and 205 2000 remotely sensed Landsat images and a 2012 DMC image to track land-cover changes in Yaya forest reserve before and after MTS. For the purpose of the study, the period between 1990 and 2000 is considered the non-MTS period while between 2000 and 2012 is considered the MTS period. In order to adequately establish a link between MTS and possible changes in the Yaya forest cover, three similar sets of images were analyzed for Nsemre and Sawsaw reserves for the same time period. Since Nsemre and Sawsaw communities at the time of this research did not participate in MTS, they were included in the change analysis as a control against which the impact of MTS on observed land-cover changes are discussed. All three reserves (Yaya, Nsemre and Sawsaw) are located in the Brong Ahafo region’s Dry Semi-deciduous Forest Zone (DSFZ). The land-cover situation in all three reserves between 1990 and 2000 suggests heavy deforestation with the reserves dominated by patches with open forests and scattered trees (see figure 5.0 below). A 2008 Global Environmental Facility Small Grant Program (GEF) report claim that throughout the 1990s leading up to the early 2000s all three reserves were dominated by grass species mostly Chromolaena odoratum (Acheampong weed) (GEF, 2008). The land- cover change results discussed below appear to support the GEF report however the analysis in this chapter provides new information on the land-cover situation in all three reserves between 2000 and 2012. Before discussing the land-cover change results however, sections 5.1.1 through 5.1.3 below describes the image selection, acquisition and processing protocols. 206 Figure 5.0: Brong-Ahafo land-cover situations between 1990 and 2000 207 5.1.1 Image selection protocol Images for all three reserves were obtained for the month of January 1990, 2000 and 2012. Due to emphases on greenness (forest vegetation) within the reserves, the 4-3-2 Red- Green-Blue (RGB) false color composite was used. Images were selected for the month of January (i.e., typically the dries month of the year in most of Ghana) to minimize false representations from vegetation cover other than trees. Also, because January is a lean cropping season with reduced farming activity, the effect of annual crops particularly maize is also likely to be minimized during this period. Another advantage of using images from January is because during this period of dry weather, there are relatively fewer clouds and rains to affect image quality. 5.1.2 Image acquisition The 1990 and 2000 LANDSAT images were obtained from the USGS Global Visualization Viewer (GLOVIS) online areal and satellite image archives. While the GLOVIS archive has satellite images for the study area going as far back as November 1973, images prior to 1986 were found to be missing lower bands one and two and were also mostly of poor quality as a result of clouds. Also, images from 2003 to 2013 contained multiple strips of missing data and hence could not be processed. For the years of interest for which relatively good images were available (1990 and 2000), 22-meter resolution images were acquired by inputting latitude 7.2 and longitude -2.8 into the GLOVIS online search engine and specifying the year and month of interest. The images for 1990 and 2000 were downloaded from GLOVIS in compressed Tagged Image File Format (TIFF). TIFF generally allow for the original image to be transferred with very little to no loss in image quality. 208 5.1.3 Image processing protocol The entire area of interest including Yaya, Nsemre and Sawsaw Forest Reserves was initially processed as a single unit before disaggregated analysis was done on the individual reserves. After importing both Landsat and DMC TIFF images of the study area into ArcMap, a shape file of the three serves was used to delimit the study area boundaries prior to image classification and change detection. Initial cropping was done in ArcMap by utilizing the Raster processing tools to clip out the three reserves. The images were then exported from ArcMap in a TIFF format into a designated folder where they were imported into Envi 4.7 for further processing. Once imported into Envi, the 4-3-2 bang combination was selected and the image for each year saved in Envi Standard format. The Envi Standard images were later imported back into Envi and classified using the ISO-Data classification scheme under the unsupervised classification menu. After classifying the images and selecting the desired classes for change detection, the post-classification menu in Envi was used to compute the change statistics for each class for an initial and final time period. Post-classification results were saved in a “notepad” format in a designated folder and imported later into Microsoft excel for change detection analysis and graphical representation of changes for the different time periods. Prior to analyzing the changes, a field validation of the classified image was conducted to appropriately identify the various classes and their associated land-cover. All pixel counts were multiplied by a conversion factor of 0.0009 to obtain their Km2 equivalent. 5.1.4 Chapter organization The rest of chapter five discusses land cover changes in Yaya, Nsemre and Sawsaw reserves for the periods before MTS (1990) the period around which MTS was introduced (2000) and approximately 12 years after the MTS (2012). A combination of reserve maps and their 209 change statistics are used in the change detection analysis and discussion. For the sake of simplicity and also to focus attention on changes in forest vegetation, only three land-cover classes (forest, settlements and farms) were used in the analysis. The term “settlements” is used broadly to mean open areas, village settlements (houses), grass and shrub cover while forest refers to primary and secondary forest comprising mostly of recently planted teak and coppice. Farm or farmland generally refers to fallow land or areas recently farmed as either open space or intercrops in teak plantation. 5.2 Landcover changes in all three reserves 5.2.1 Landcover changes in all three reserves from 1990 to 2000 Figure 5.1 and table 5.2.1 below describe land-cover changes in all three forest reserves combined (Yaya, Nsemre and Sawsaw) before the MTS was launched. All three reserves combined cover a land mass of approximately 120 Km2 with the largest reserve being Sawsaw (57 Km2) and the smallest being Nsemre (20 Km2). Yaya reserve occupies a landmass of about 41 Km2. The years 1990 and 2000 were chosen for the analysis in order to investigate changes in land-cover situation in the research forest reserves that might have necessitated launching of the MTS sometime in 2001/2002. A close examination of figure 5.1 suggests extremely high rates of forest loss in all three reserves combined between 1990 and 2000 with the highest losses occurring in the Sawsaw reserve. The change statistics presented in table 5.2.1 below capture the net gains and losses in each of the three land classes in pixels and square kilometers. Thus, from table 5.2.1 below, a total of 5.6 Km2 and 2.15 Km2 of forest and farmlands respectively were lost to settlements, open areas, grass and shrubs between 1990 and 2000. Figure 5.2 below presents the 1990 baseline levels of each land class in square kilometers (Km2) and the changes that 210 occurred by 2000 (10 years from the baseline). Figure 5.2 suggests that out of the total land area of approximately 120 Km2 in 1990, 40% (44.84 Km2) was settlements, open areas and shrubs while 36% (43.66Km2) was forest, and 24% (29.35Km2) farmlands. 211 Figure 5.1: Land-cover changes in Yaya, Nsemre and Sawsaw from 1990 to 2000 212 Table 5.2.1: Land-cover change statistics in pixels and Km2 for all reserves from 1990 to 2000 2000 "to" Class Number of Pixels Percentage of total changed pixels 1990"from" class Number of Pixels Percentage of total changed pixels Net gain/loss pixels Net gain/loss pixels in percentage Net gain/loss Km2 Settlements 58440 50.98 Settlements 49824 45.37 8616 5.61 7.7544 Forest 42280 36.88 Forest 48507 44.17 -6227 -7.29 -5.6043 Forest Intercrop 30221 26.36 Forest Intercrop 32610 29.6935 -2389 -3.33 -2.1501 Total Pixels 130941 100 Total Pixels 130941 100 To and from class change statistics with net gain/loss expressed in pixels and square kilometers (Km2). Conversion factor used in calculating land area is Km2 = number of pixels x 0.0009. 213 m K q S ( ) a e r A Figure 5.2: Land-cover changes in Km2 in Yaya, Nsemre and Sawsaw from 1990 to 2000 Combined Yaya, Nsemre and Sawsaw Land Cover Change in Square Kilometers Between 1990 and 2000 60.00 50.00 40.00 30.00 20.00 10.00 0.00 -10.00 52.60 Settlements/Open Area/ Road Forest Farms/ Forest Intercrop 44.84 43.66 29.35 38.05 27.20 7.75 Change -2.15 -5.60 Year-1990 Year-2000 214 5.2.2 Landcover changes in all three reserves from 2000 to 2012 Figure 5.3 and table 5.2.2 below describe land-cover changes in all three forest reserves combined (Yaya, Nsemre and Sawsaw) 12 years after the MTS was launched in 2000. The years 2000 and 2012 were chosen for the analysis in order to investigate changes in land cover situation in the research forest reserves that may have resulted from 10 to 12 years of the MTS implementation within the Yaya Forest Reserve. Both Nsemre and Sawsaw were included to serve as controls against which the impact of the MTS on Yaya can be assessed. A visual appraisal of figure 5.3 reveals stark differences in land cover change within Yaya (the MTS reserve), Nsemre and Sawsaw (the two non-MTS reserves) between 2000 and 2012. From the analysis of the 1990 and 2000 images, it may be concluded that the forest cover situation in all three reserves went from bad to worse between this period (see figure 5.1 above). Figure 5.3 however reveals a relatively high rate of land conversion from settlement and farmlands into forest in both Yaya and Nsemre reserves. The forest cover situation for Sawsaw appears to have gotten even worse in most areas of the reserve during between 2000 and 2012 with a few dense patches of forest vegetation in the northeast and southeastern portions of the reserve. Since no other reforestation programs has been recorded within Yaya besides the MTS, it logical to assume that the increase in Yaya reserve forest cover is most likely due to the MTS. Even though Nsemre does not have the MTS, one logical explanation to increase forest cover between 2000 and 2012 may be due to its close proximity to Yaya but most importantly the implementation of JICA’s Participatory Forest Management Program (PAFORM) which was implemented sometime between 2000 and 2009. The PAFORM project encouraged planting of fruit (citrus and mango) trees within the Nsemre reserve to serve as a food safety net during the lean seasons as well as disincentivize deliberate forest fires used in trapping small game. 215 Figure 5.3: Land-cover changes in Yaya, Nsemre and Sawsaw from 2000 to 2012 216 The change statistics presented in table 5.2.2 below capture the net gains and losses in each of the three land classes in pixels and square kilometers. Thus, from table 5.2.2, a total gain of 15.33 Km2 was recorded for forests as against loses of -12.647 Km2 for settlements, open areas and shrubs and -2.69 for farmlands. Figure 5.4 below presents the 2000 baseline levels of each land class in Km2 and the changes that occurred by 2012 (10 years from the baseline). Figure 5.4 also suggests that out of the total land area of 15 Km2 converted into forests, approximately 82% (12.64 Km2) came from settlements and open areas while 18% (2.69 Km2) came from farmlands. Figure 5.4 below presents in Km2, changes in the three land-cover classes from 2000 to 2012. 217 Table 5.2.2: Land-cover change statistics in pixels and Km2 for all reserves from 2000 to 2012 2012 "to" Class Number of Pixels Percentage of total changed pixels 2000"from" class Number of Pixels Percentage of total changed pixels Net gain/loss pixels Net gain/loss pixels in percentage Net gain/loss Km2 Settlements 57043 49.76 Settlements 60033 54.66 -2990 -4.90 -2.6910 Forest 59562 51.96 Forest 42529 38.73 17033 13.23 15.3297 Forest Intercrop 16329 14.24 Forest Intercrop 30372 27.66 -14043 -13.41 -12.6387 Total Pixels 132934 100 Total Pixels 132934 100 To and from class change statistics with net gain/loss expressed in pixels and square kilometers (Km2). Conversion factor used in calculating land area is Km2 = number of pixels x 0.0009. 218 Figure 5.4: Land-cover change in Km2 Yaya, Nsemre and Sawsaw from 2000 to 2012 Combined Yaya, Nsemre and Sawsaw Land Cover Change in Square Kilometers Between 2000 and 2012 54.03 53.61 51.34 38.28 27.33 Settlements/Open Area/ Road Forest Farms/ Forest Intercrop 14.70 15.33 Year-2000 Year-2012 Change -2.69 -12.64 60.00 50.00 40.00 30.00 20.00 10.00 0.00 -10.00 -20.00 5.3 Land-cover changes in Yaya Reserve Figure 5.5 and table 5.3 below present changes in the three land-cover classes for only m K q S ( ) a e r A Yaya reserve for the time periods 1990 to 2000 and from 2000 to 2012. The land-cover situation in Yaya (with forests occupying less than 60%) in 1990 and 2000 appear to confirm earlier reports of excessive forest loss within the reserve. According to a UNDP-SEF, (2008) report, a 1994 Multi Resource Inventory (MRI) to assess forest conditions in Ghana, ranked Yaya reserve five on a six-point scale with six signifying the highest or worst level of resource degradation. Seven years following the 1994 national MRI, Yaya was not even considered for the 2001 MRI due to a lack of standing natural forests (UNDP-SEF, 2008 and Forestry Services Commission, 2002). The 1994 and 2001 MRI were a comprehensive evaluation of flora and fauna levels within the reserve hence the state of natural forest vegetation was only one part of the entire MRI. The land-cover situation in Yaya in the 1990s leading to the 2000s most likely prompted 219 the MTS program within the reserve in an attempt to rescue whatever natural forest was left and convert other open areas into tree plantations (with teak being species of choice under the MTS). 5.3.1 Landcover change in Yaya Reserve between 1990 and 2000 Figure 5.5: Yaya Reserve Land Cover Change from 1990 to 2012 The change statistics presented in table 5.3.1 and figure 5.6 below capture net gains and losses in each of the three land classes in pixels and Km2 for Yaya Reserve only. From table 5.3.1 below, a total gain of 2.56 Km2 and 4.702 Km2 was recorded for forests and settlements (including open areas and shrubs lands) respectively in Yaya as against loses of -2.56 Km2 for farmlands. Figure 5.6 presents the 1990 baseline levels of each land class in Km2 and the changes that occurred by the year 2000 (10 years from the baseline). Figure 5.6 thus suggests that between 1990 and 2000 farmlands within Yaya Reserve declined by -7.26 Km2 and of the total 220 losses in the area of farmlands (approximately 65% or 4.70 Km2) was taken up by settlements while 35% (2.56 Km2) was converted into forest cover. 221 Table 5.3.1: Yaya Land Cover Change Statistics in pixels and Km2 from 1990 to 2000 2000 "to" Class Number of Pixels Percentage of total changed pixels 1990"from" Number of class Pixels Settlements 15140 33.42 Settlements 9915 Percentage of total changed pixels 21.88 Net Net gain/loss Net gain/loss pixels pixels in percentage gain/loss Km2 5225 11.53 4.7025 Forest 27502 60.70 Forest 24662 54.43 2840 6.27 2.556 Forest Intercrop 2665 5.88 Forest Intercrop 10730 23.68 -8065 -17.80 -7.2585 Total Pixels 45307 100 Total Pixels 45307 100 To and from class change statistics with net gain/loss expressed in pixels and square kilometers (Km2). Conversion factor used in calculating land area is Km2 = number of pixels x 0.0009. 222 Figure 5.6: Yaya Reserve Land Cover Change in Km2 from 1990 to 2000 Yaya Reserve Land Cover Change in Square Kilometers Between 1990 and 2000 22.20 8.92 9.66 24.75 13.63 Settlements/Open Area/ Road Forest Farms/ Forest Intercrop 2.40 4.70 2.56 Year-1990 Year-2000 Change -7.26 30.00 25.00 20.00 15.00 10.00 5.00 0.00 -5.00 -10.00 m K q S ( ) a e r A 5.3.2 Landcover Change in Yaya Reserve between 2000 and 2012 In table 5.3.2 and figure 5.7 below, the number of pixels and Km2 change statistics in the three land cover classes for only Yaya reserve is presented for the time periods from 2000 to 2012. The year 2000 captures baseline forest levels around the time of MTS introduction into Yaya while 2012 capture land-cover changes that have occurred within the reserve 12 years after program implementation. Table 5.3.2 suggests that forest cover in Yaya reserve expanded by approximately 6.30 Km2 between the year 2000 and 2012 while settlements and farmlands respectively shrunk by -5.9103 Km2 and -0.3861 Km2. It is worth noting that of all the three land classes, only forest cover recorded an expansion in area for the period between 2000 and 2012. All the gain in forest cover was absorbed from the two other land classes (settlements and farmlands). Figure 5.7 below depicts graphically the land-cover changes in Yaya between 2000 and 2012. 223 Table 5.3.2: Yaya Land Cover Change Statistics in pixels and Km2 from 2000 to 2012 2012 "to" Class Number of Pixels Percentage of total changed pixels 2000"from" Class Number of Pixels Percentage of total changed pixels Net gain/loss pixels Net gain/loss pixels in percentage -6567 6996 Net gain/loss Km2 -5.9103 6.2964 -14.58 15.49 Settlements 8407 18.65 Settlements Forest 34438 76.38 Forest 14974 27442 Forest Intercrop Total Pixels 2222 45067 4.97 100 Forest Intercrop 2651 Total Pixels 45067 33.23 60.89 5.88 100 -0.91 -429 -0.3861 To and from class change statistics with net gain/loss expressed in pixels and square kilometers (Km2). Conversion factor used in calculating land area is Km2 = number of pixels x 0.0009. 224 Figure 5.7: Yaya Reserve Land Cover Change in Km2 from 2000 to 2012 35.00 30.00 25.00 20.00 15.00 10.00 5.00 0.00 -5.00 -10.00 Yaya Reserve Land Cover Change in Square Kilometers Between 2000 and 2012 30.99 24.70 13.48 Settlements/Open Area/ Road Forest Farms/ Forest Intercrop 7.57 6.30 2.39 2.00 Year-2000 Year-2012 Change -0.39 -5.91 225 m K q S ( ) a e r A 5.4 Land-cover changes in Nsemre Reserve 5.4.1 Land-cover changes in Nsemre Reserve between 1990 and 2000 Figure 5.8 and table 5.4.1 below presents the situation of the three land-cover classes in Nsemre in 1990, 2000 and 2012. In 1990, land classified as forest vegetation represented 44.43% (9.02 Km2) of the total land mass (20.3 Km2) in Nsemre. By 2000 however, forest vegetation declined by nearly about 41% (3.7 Km2) and occupied only 26.38% (5.35 Km2). Figure 5.9 below presents net changes in all three land-cover classes in Nsemre between 1990 and 2000. Figure 5.8: Nsemre Reserve Land Cover Change from 1990 to 2012 226 Of the total losses in forest cover between 1990 and 2000, approximately 90% was lost to settlements while 10% was lost to crop farms and forest intercrop. The alarming rate of forest loss within Nsemre between 1990 and 2000 possibly prompted the intervention of JICA’s Participatory Forest Management Project (PAFORM) in the early 2000s. While both ADB and GoG sponsored Community Forest Management Project (CFMP) (later modified into the MTS) and JICA’s PAFORM project are similar in their objective to rehabilitate forests, the two programs differ in their alternative livelihood approach. The primary difference between both programs is that for JICA, selected communities around Nsemre received training on a range of alternative livelihood programs (bee keeping, snail rearing, soap making, planting of fruit trees etc.) whereas for CFMP/MTS beneficiary groups received livestock, teak seedlings, pegs for transplanting teak seedlings and technical advice on teak management as part of the support. Table 5.4.1 and figure 5.9 illustrate changes in land-cover in Nsemre between 1990 and 2000. 227 Table 5.4.1: Nsemre Land Cover Change Statistics in pixels and Km2 from 1990 to 2000 2000 "to" Class Number of Pixels Percentage of total changed pixels 1990"from" class Number of Pixels Percentage of total changed pixels Net gain/loss pixels Net gain/loss pixels in percentage Net gain/loss Km2 Settlements 11189 49.61 Settlements 7558 33.51 3631 16.10 3.2679 Forest 5949 26.38 Forest 10020 44.43 -4071 -18.05 -3.6639 Forest Intercrop 5415 24.01 Forest Intercrop 4975 22.06 440 1.95 0.3960 22553 100 Total Pixels To and from class change statistics with net gain/loss expressed in pixels and square kilometers (Km2). Conversion factor used in calculating land area is Km2 = number of pixels x 0.0009. Total Pixels 22553 100 228 Figure 5.9: Nsemre Reserve Landcover Change in Km2 from 1990 to 2000 Nsemre Land Cover Change in Square Kilometers Between 1990 and 2000 10.07 9.02 6.80 4.48 5.35 4.87 Settlements/Open Area/ Road Forest Farms/ Forest Intercrop 3.27 0.40 Year-1990 Year-2000 Change -3.66 12.00 10.00 8.00 6.00 4.00 2.00 0.00 -2.00 -4.00 -6.00 m K q S ( ) a e r A 5.4.2 Landcover changes in Nsemre Reserve between 2000 and 2012 In table 5.4.2 and figure 5.10 below, the number of pixels and Km2 change statistics in the three land-cover classes for only Nsemre reserve is presented for 2000 and 2012. According to table 5.6, forest cover in Nsemre reserve expanded by approximately 28% (5.28 Km2) between 2000 and 2012 while settlements and farmlands shrunk by -2.50 Km2 and -3.51 Km2 respectively. Like the situation in Yaya reserve, only the area of forest cover expanded for the period under review (2000-2012). All the gain in forest cover between 2000 and 2012 came from settlements and farmlands (see figure 5.10 above). 229 Table 5.4.2: Nsemre Land Cover Change Statistics in pixels and Km2 from 2000 to 2012 2012 "to" Class Number of Pixels Percentage of total changed pixels 2000"from" class Number of Pixels Percentage of total changed pixels Net gain/loss pixels Net gain/loss pixels in percentage Net gain/loss Km2 Settlements 8412 38.70 Settlements 11189 49.6 -2777 -10.9 -2.4993 Forest 11816 54.37 Forest 5949 26.4 5867 28.0 5.2803 Forest Intercrop 1506 6.93 Forest Intercrop 5415 Total Pixels 21734 100 Total Pixels 22553 24.0 100 -3909 -17.1 -3.5181 To and from class change statistics with net gain/loss expressed in pixels and square kilometers (Km2). Conversion factor used in calculating land area is Km2 = number of pixels x 0.0009. 230 Figure 5.10: Nsemre Reserve Land Cover Change in Km2 from 2000 to 2012 Nsemre Land Cover Change in Square Kilometers Between 2000 and 2012 10.07 10.63 7.57 Settlements/Open Area/ Road Forest Farms/ Forest Intercrop 5.35 4.87 5.28 1.36 Year-2000 Year-2012 Change -2.50 -3.52 12.00 10.00 8.00 6.00 4.00 2.00 0.00 -2.00 -4.00 -6.00 231 m K q S ( ) a e r A 5.5 Landcover changes in Sawsaw Reserve 5.5.1 Land-cover changes in Sawsaw Reserve between 1990 and 2000 Figure 5.11 below presents the land-cover situation in Sawsaw in 1990, 2000 and 2012. In 1990 forests represented about 27% (15.17 Km2) of the total landmass in Sawsaw however by 2000 forest cover declined by about 48% and occupied only 13% (7.84 Km2). Table 5.5.1 and figure 5.11 capture net changes in all three land-cover classes between 1990 and 2000. Figure 5.11: Sawsaw Reserve Land Cover Change from 1990 to 2012 232 Table 5.5.1: Sawsaw Land Cover Change Statistics in pixels and Km2 from 1990 to 2000 2000 "to" Class Number of Pixels Percentage of total changed pixels 1990"from" class Number of Pixels Percentage of total changed pixels Net gain/loss pixels Net gain/loss pixels in percentage Net gain/loss Km2 Settlements 33614 52.92 Settlements 32500 51.17 1114 1.75 1.0026 Forest 8708 13.71 Forest 16852 26.53 -8144 -12.82 -7.3296 Forest Intercrop 21197 33.37 Total Pixels 63519 100 Forest Intercrop Total Pixels 14167 22.30 7030 11.07 6.327 63519 To and from class change statistics with net gain/loss expressed in pixels and square kilometers (Km2). Conversion factor used in calculating land area is Km2 = number of pixels x 0.0009. 233 Figure 5.12: Sawsaw Reserve Land-Cover Change in Km2 from 1990 to 2000 Sawsaw Land Cover Change in Square Kilometers Between 1990 and 2000 29.25 30.25 Settlements/Open Area/ Road Forest 19.08 Farms/ Forest Intercrop 15.17 12.75 7.84 6.33 1.00 Year-1990 Year-2000 Change -7.33 35.00 30.00 25.00 20.00 15.00 10.00 5.00 0.00 -5.00 -10.00 m K q S ( ) a e r A 5.5.2 Land-cover changes in Sawsaw Reserve between 2000 and 2012 In 2000 forests represented about 13% (7.84 Km2) of Sawsaw’s total landmass however by 2012 forest cover increased by 47% from 7.84 Km2 to 11.52 Km2. In 12 years between 1990 and 2012 settlements and open areas dominated Sawsaw’s landscape reaching a peak of about 71% (40.44 Km2) in 2012 (figure 5.13 and table 5.5.2 below). 234 Table 5.5.2: Sawsaw Land Cover Change Statistics in pixels and Km2 from 2000 to 2012 2012 "to" Class Number of Pixels Percentage of total changed pixels 2000"from" class Number of Pixels Percentage of total changed pixels Net gain/loss pixels Net gain/loss pixels in percentage Net gain/loss Km2 Settlements 44932 70.72 Settlements 33557 52.90 11375 17.82 10.2375 Forest Forest Intercrop 12803 20.15 Forest 8702 13.72 4101 6.43 3.6909 5696 8.97 Forest Intercrop 21172 33.38 -15476 -24.41 -13.9284 63431 100 Total Pixels To and from class change statistics with net gain/loss expressed in pixels and square kilometers (Km2). Conversion factor used in calculating land area is Km2 = number of pixels x 0.0009. Total Pixels 63431 100 235 Figure 5.13: Sawsaw Reserve Land Cover Change in Km2 from 2000 to 2012 Sawsaw Land Cover Change in Square Kilometers Between 2000 and 2012 50.00 40.00 30.00 20.00 10.00 0.00 -10.00 -20.00 30.20 19.05 7.83 40.44 Settlements/Open Area/ Road Forest Farms/ Forest Intercrop 11.52 5.13 10.24 3.69 Year-2000 Year-2012 Change -13.93 m K q S ( ) a e r A 5.6.1 Annual changes in forest cover in Yaya before and after MTS 5.6 Summary of findings The findings in section 5.3.1 and 5.3.2 above suggests that in 2012 approximately 76% (31 Km2) of Yaya’s 41 Km2 landmass was occupied by some type of forest. Settlements and crop farms occupied approximately 19% and 5% respectively representing a total combined landmass of about 9.6 Km2 in 2012. The results also suggest that forests in Yaya increased at an annual rate of 0.26 Km2 between 1990 and 2000 (i.e. the non-MTS period) and then more than doubled to a rate of about 0.52 Km2 between 2000 and 2012 (i.e. the MTS period). The sharp increase (more than two folds) in the rate of forest cover between 2000 and 2012 relative to the previous decade (1990 to 2000) may be attributed in part to some deliberate intervention within Yaya possibly the MTS program. The Difference in Difference (DID) method in chapter eight established that about 10% of perceived changes in natural capital assets among Yaya 236 communities between 2000 and 2012 were attributed directly to the MTS (refer to chapter eight for more details). The finding from the DID suggests that the MTS program played a large role in increasing trend of reforestation in Yaya. Assuming the annual growth rate of 0.52 Km2 in forest cover experienced between 2000 and 2012 continues “ceteris paribus” all 9.6 Km2 of Yaya’s landmass occupied by settlements and farmlands in 2012 would be converted back into forest by 2029 (17 years). While it is possible in theory to convert Yaya’s entire 41 Km2 landmass back into forests, this may not be possible under the current land-use pattern as portions of Buoku, Abrefakrom and Ayigbekrom communities physically occupying portions of the reserve. As shown in Figure 5.14 below, portions of Yaya are occupied by settlements hence these occupied areas remained largely deforested between 1990 and 2012. A new government policy may be required to either relocate communities currently resident within Yaya or redefine the boundaries of the reserve. 237 Figure 5.14: Yaya Forest Reserve and Communities in 2012 5.6.2 Annual changes in forest cover in Nsemre before and after MTS Unlike Yaya that experienced positive growth in forest cover between 1990 and 2000 forests Nsemre experienced a -0.37 Km2 annual decline during the same period (see section 5.4.1 above). Between 2000 and 2012 however forest cover in Nsemre increased sharply at an annual rate of about 0.45 Km2. The changes in forests cover observed in Nsemre between 2000 and 2012 may in part be attributed to the Japan International Cooperation Agency’s (JICA) Participatory Forest Resource Management (PAFORM) project interventions that encouraged fruit tree intercrops in parts of the reserve to provide safety nets during lean cropping seasons and discourage residents from deliberately setting fire to portions of the reserve in an attempt to trap wild game particularly grasscutters. In 2012, approximately 54% (11 Km2) of Nsemre reserve’s 20 Km2 total landmass was occupied by forests. Settlements and crop farms occupied 238 approximately 39% and 1% respectively representing a total landmass of about 8.9 Km2 in 2012. Assuming the annual growth rate of 0.45 Km2 in forest cover experienced between 2000 and 2012 continues “ceteris paribus” all 8.9 Km2 of Nsemre’s landmass occupied by settlements/bare ground and farmlands in 2012 would be converted back into forest in approximately 20 years. Figure 5.15 below shows the state of the three land-cover types in Nsemre and the location of the four surrounding communities included in the research. Together all four communities have fewer than 1500 individual housing units and are situated on the outskirts of the reserve. The relative location of Nsemre communities from the reserve suggests that reforestation programs such as PAFORM or MTS have a relatively higher chance of succeeding in converting degraded portions of the reserve back into forest. While the location of forest communities at safe distances from the forest means that most degraded areas could be easily converted back into forests, policing trees from illegal loggers presents a bigger challenge that requires a lot more effort from both the communities and the Ghana Forest Services Department (FSD). 239 Figure 5.15: Nsemre Forest Reserve and Communities in 2012 240 5.6.3 Annual changes in forest cover in Sawsaw before and after MTS Like Nsemre, Sawsaw also experienced negative growth in forest cover between 1990 and 2000 before reversing trends between 2000 and 2012. Between 1990 and 2000 tree cover in Sawsaw declined at an annual rate of -0.73 Km2 (see section 5.5.1 above) and then increased sharply at an annual rate of 0.31 Km2 between 2000 and 2012. While the MTS and PAFORM programs to a large extent help explain some of the changes in forests cover in Yaya and Nsemre between 2000 and 2012, these interventions were lacking in Sawsaw. It appears communities surrounding Sawsaw may still be engaged in the old/traditional Taungya. While the gains in tree cover between 2000 and 2012 may be attributed to the traditional Taungya, it is likely that these gains may be reversed as a result of a lack of incentive on the part of community members to see the trees grow (see chapter 1 section 1.2 for more details on Taungya Systems). Table 5.5.1 and figure 5.13 under section 5.5.1 above suggest that in 2012 approximately 20% (11.52 Km2) of Sawsaw reserve’s 57 Km2 landmass was occupied by forests. Settlements (including open areas) and croplands occupied approximately 71% and 9% respectively representing a total landmass of about 45.57 Km2 in 2012. Figure 5.16 below shows the condition of Sawsaw in 2012. 241 Figure 5.16: Sawsaw Forest Reserve and Communities in 2012 Assuming the annual growth rate of 0.31 Km2 in forest cover experienced between 2000 and 2012 continues “ceteris paribus” all 45.57 Km2 of Sawsaw’s landmass occupied by settlements and farmlands in 2012 would be converted back into forest in approximately 150 years. Assuming zero growth in all other land classes except forests, a 0.52 Km2 annual growth in forest cover (like that experienced in Yaya between 2000 and 2012) will convert all 45.57 Km2 occupied by settlements and farmlands in 2012 into forests in 87 years. As shown in figure 5.14 above, Papasu community as of the time of this research was situated completely inside Sawsaw reserve. As of the time of this research, Papasu had only 12 housing unit (see table 4.7 under chapter 4 section 4.8.1) however if left unchecked will most likely expand and present an 242 00.000010.000020.000030.000040.000005Kilometers®SAWSAW FOREST RESERVE AND COMMUNITIES IN 2012NtemaLEGENDMajor HighwayDomeabraPipotrimAyayoPapasuLegendForest CoverSettlements/Open Areas/Road/Grass and ShrubCrop Farms/ Forest IntercropSource of Base Map: Ghana Survey DepartmentSource of Image: Lansat January 1990 and January 2000, Source of Image: DMC January 2012Data Enriched and Map Produced by Doe Adovor (2013)2°10'0"W2°10'0"WMap of Brong Ahafo Region Showing Sawsaw ReserveSawsaw Reserve Land Cover Change Detection200019902012 even bigger challenges to future reforestation/forest conservation efforts. Like what was observed in Yaya (see figure 5.12), future forest policy may either acknowledge the permanent presence of Papasu within the Sawsaw or relocate this community to the outskirts of the reserve. Alternatively, Papasu could be empowered by the FSD to serve as a satellite community tasked with caretaker duties within the reserve. The sheer size of Sawsaw (almost the size of Yaya and Nsemre combined) coupled with the relatively slow growth of forest cover and the presence of a community within its boundaries presents exceptional challenges to future reforestation programs. 5.7 Conclusions and Implications for forest policy The statement below from a member of the Konsoa Community non-MTS Focus Group sheds light on the importance of community involvement in Ghana’s forest management, the competition problem and the potential for forest policies and management strategies to address or exacerbate these problems. I migrated to this village in 1967 and at that time the Yaya forest was still in its pristine state and had never been farmed. It was Dr. Busia that granted us permission to farm portions of the reserve for the first time and since then we have been farming there until now. Back in those days, the forest stretched beyond the limits of the eyes. A wide variety of wildlife littered the forest floor. Wildlife was so abundant that even the undergrowth could hardly conceal them…. wildlife was literally everywhere. I think it is the recent forest fires that have caused the forest and all the wildlife to disappear and leave us. Now we are trying to rehabilitate our forest. But even as we put forth our best efforts towards planting trees others are bent on destroying them because they prefer to plant corn. They keep on killing the trees so that the land can be used for planting food crops. That’s all I have to say. (Konsoa Non-MTS Focus Group Member, 2009). Together with the other findings in this chapter, the quote from Konsoa Non-MTS Focus Group demonstrate that a community centered reforestation program that confer some level of forest ownership to farmers has tremendous promise for success. As the quote suggests, Non- 243 MTS groups allocated portions of forest reserves under the traditional Taungya system often are torn between; a) protecting trees for which they have virtually no economic benefits after timber sales; and b) deliberately destroy the tree seedlings to free up the land for continues crop production and ensure some level of food security for their households. In the absence of the right incentives as embodies in an equitable benefit sharing arrangement, most rational actors would opt for the later strategy that frees up land for continuous crop production. Section 5.5 above thus demonstrated the impact of two reforestation strategies; the first in Yaya where through a benefit sharing agreement, farmers given some level of ownership in plantations established under the MTS program protected the forest. In the second example (in Nsemre Reserve), farmers supported under an alternative livelihood programs in exchange were tasked with establishing polyculture forests with fruit tree intercrops. Since fruit trees provide some level of food security during the lean production seasons, they served as an important safety net which guaranteed their protection and to a large extent the protection of other timber species as well. The case of Nsemre proves that moral suasion (i.e. encouraging rural households to do the right thing) bundled with the necessary incentives such as training and persistent behavioral change awareness can be just as effective in curbing the rate of deforestation in Ghana’s forest reserves (Kerr, Milne, Chhotray, et al, 2007 also provides insight into how moral suasion is used to effectively manage watershed in India). The two approaches described above provide the necessary incentives for the rapid forest growth observed in both Yaya and Sawsaw between 2000 and 2012 (see figures 5.7 and 5.10). 5.7.1 Monoculture Forest Plantations under MTS As shown in the analysis above, the monoculture teak plantation system implemented under Ghana’s MTS in Yaya produced the highest annual growth in forest cover, however this 244 management system raises serious questions that need to be addressed by forest policy. First, the current MTS raises questions regarding the long-term availability of land for producing seasonal staples and perennial crops that communities depend on for livelihoods. Currently, the MTS mandates that intercropping teak plantations should be done with only annual crops. Additionally, the MTS allows intercropping for a maximum of three years after which all such activities must completely stop. Halting all intercropping activities after three years means farmers must seek new production areas/acreage for their everyday staple crops (e.g., corn, beans, groundnuts, cassava). Without a plan or policy directive to relocate the 7.57 SqKm2 (see fig 5.7) of settlement areas outside of Yaya Reserve, very little or no forest area would be left to allocate under the current MTS until the end of a rotation. With no land left to allocate and all intercropping activities halted, farmers seeking new production areas either must lease non-reserve land outside of Yaya or travel to distant reserves such as Sawsaw to search for new staple crop production areas. The strict policy of halting intercropping after three years will put untold burden on household who have to migrate outside their communities (whether temporary or permanently) to new locations in search of new production areas. Secondly, forest policy must address the effect of the MTS’ current monoculture teak plantation system on biodiversity in national forest reserves. It is well documented that monoculture teak plantations support very few flora and fauna particularly on the forest bed (see Amoah, 2009). In most rural communities in Ghana, NTFPs such as small game, wild mushrooms, snails, herbs, and tree backs and raffia provides an essential source of proteins, natural herbs, medicines and building materials for most households. Losing these essential NTFPs affects entire forest communities that depend on these resources for their livelihoods. Under Ghana’s current MTS, benefits from the sale of timber accrues largely to the FSD and 245 individual farmers while the costs measured in terms of loss of biodiversity (NTFPs) accrue largely to the entire community including non-MTS households. Future MTS programs must safeguard NTFPs while simultaneously produce commercial timber species of value. Thirdly, Ghana’s forest policy must consider the potential environmental impacts of clear-cutting activities under the current MTS and its implications for biodiversity of the country’s national forests. Planting even-aged monoculture plantations require that relatively large tracts of forest area be cleared at the end of a rotation period (whether eight, fifteen or twenty-five years). While clearcutting is used in most temperate regions as an effective tool to regenerate shade-intolerant tree species, diversify stand structure and habitat, manage forest insects and diseases and improve the general productivity of even-aged forests (see SAF, 1997 and Bowyer, Fernholz, and Lindburg et. al., 2009) this practice may not be suitable for all forests and tree species. It is well documented that clear-cutting (or clear-felling or clear-logging) vast areas of forests in most instances results in loss of topsoil, loss of habitat for flora and fauna and a disruption of the hydrological cycle of entire watersheds. Fourthly, Ghana’s forest policy must consider the financial burden on the FSD and resource poor farmers of implementing a monoculture teak plantation system. Additionally, forest policy must interrogate the ways in which the current MTS can either address or exacerbate youth unemployment in forest communities. In terms of the economic impact of monoculture plantations, it is well documented that where there is abundance supply of heavy forest equipment, clear-cut harvesting of even-aged forests is known to be more economical. It is therefore not a coincident that monoculture timber plantations dominate forest landscapes of temperate industrialized nations where excess capital and heavy equipment makes forest management and clear-cut harvests the most efficient and economical method of harvesting. 246 Managing even-aged monoculture teak plantations under Ghana’s current MTS in the absence of cheap and abundant capital and machinery present huge financial challenges to the FSD and local farmers. The general lack of financial resources within the FSD to acquire and regularly maintain heavy forestry equipment and the huge financial burden imposed on MTS farmers who may have to rent these equipment presents long-term challenges to even-aged monoculture tree plantations in Ghana’s forest reserves. The current monoculture plantation system and associated sylvicultural practices implemented under Ghana’s current MTS raises important questions as to whether the MTS can serve as an effective development tool which simultaneously generates revenue from commercial tree species, provide food security and income opportunities (from annual intercrops) and increase the biodiversity of national forest reserves. As discussed above, the three-year maximum period imposed on intercropping teak plantations coupled with the disappearance of essential NTFPs raises serious policy questions regarding the effectiveness of Ghana’s current MTS to simultaneously address food insecurity and declining biodiversity within forest reserves. There is an urgent need for forest policy to consider alternative forest management systems (possibly polyculture or agroforestry) that allows staple and fruit crops to be cultivated permanently alongside commercial timber species. 5.7.2 Polyculture Forest Plantations under MTS Polyculture plantation systems involving a wide variety of commercial tree species permanently intercropped with annual and perennial fruit trees have proven to simultaneously increase forest cover, provide safety nets for vulnerable households particularly during lean cropping seasons (see section 5.3.1 and 5.3.2 above), and address rural unemployment. It is well documented that most rural communities in Ghana are plagued with disproportionately high 247 youth unemployment and food insecurity. Food insecurity concerns are particularly acute during the lean production season. The relatively high food insecurity coupled with high youth unemployment to a large extent explains the increasing trends in rural-urban migration patterns in recent times. Ghana’s MTS program can capitalize on the availability of cheap and abundant rural labor force to implement polyculture plantation or agroforestry systems in degraded portions of national forest reserves. It is well documented (see Gajaseni and Jordan, 1992) that where labor is abundant and cheap, the advantages of mechanization in even aged monoculture plantations often disappears, thus making polyculture plantations just as easily and efficiently managed as mechanized monocultures. By promoting polyculture systems as part of a multi- pronged forest policy, Ghana’s MTS can essentially address the need for increased biodiversity, food security and rural employment. Polyculture plantations established under the current MTS can serve as a conduit for addressing rural employment and associated rural-urban migration out of forest reserve communities. 5.8 Recommended areas for further research To manage Ghana’s forests sustainably, it is imperative that forest policy clearly defines what constitute a healthy forest as well as the livelihood and ecosystem benefits derived from healthy forests. Maintaining a healthy forest under the MTS would thus require forest policy to balance the needs for different forestry systems (even-aged monoculture, polyculture, and agroforestry) within national forest reserves. The research questions below are intended to help shape the debate surrounding how Ghana manages her forests for the benefits of the forest ecosystem and the life it supports. a) Which forestry system or combination is most likely to generate the highest ecosystem benefits for Ghana’s forest reserves and surrounding communities? 248 b) Which locally adapted fruit trees (e.g. citrus, mango, banana, cashew, cocoa and coconut) combine most effectively with commercial timber species to produce the largest safety-net for Ghana’s vulnerable forest populations during lean seasons and pre-rotation periods? c) What combination of forest area allocations to fruit tree plantations and commercial timber plantations is likely to produce the highest financial and natural capital benefits to forest reserve communities and households? d) Which forestry system or combination will produce the highest carbon sequestration potential for Ghana’s forest reserves? e) Which locally adapted tree species (commercial or non-commercial) present the highest potential for sequestered carbon and generation of carbon credits on the international carbon market. f) What is the willingness of MTS members and the Ghana’s Forest Service Department (FSD) to pay for crop/tree protection insurance and local financial institutions’ willingness to sell crop protection insurance that covers commercial timber species? Answers to this research questions provides a first step to establishing a tree crop receipt system that can be used by farmers and tree plantation owners to access financial resources from banks and other financial institutions between tree rotation periods. g) What are the forest ecology and silvicultural knowledge gaps and training needs of FSD personnel charged with training and monitoring commercial tree plantation under the MTS? h) What are the forest ecology and silvicultural knowledge gaps and training needs of farmers participating in the MTS? i) What is the GIS/Remote Sensing knowledge gap and training needs of Forest Research Institute of Ghana (FORIG) and FSD personnel charged with establishing and monitoring 249 commercial tree plantations under Ghana’s MTS and what short and long-term training and resources are needed to bridge potential knowledge gaps? 250 CHAPTER 6: COMMUNITY AND HOUSEHOLD PLACEMENT INTO GHANA’S MTS PROGRAM 251 Part I: Impact of Socio-Economic Factors on Community and Household Placement into Ghana’s Modified Taungya System Abstract This research used Binomial Probit Models (BPM) to isolate factors that influenced placement of Ghana’s Modified Taungya System (MTS) into communities and households as a first step to investigating livelihood changes since launching the national reforestation program in 2001/2002. In 2009, eight hundred and seventy-eight (878) household surveys were conducted in 19 forest communities in Yaya, Nsemre and Sawsaw forest reserves in Ghana’s Brong Ahafo Region. Included in the Yaya survey were 406 MTS participant and 240 non-participants. Included also are 232 households in Nsemre and Sawsaw where the MTS did not exist. The BPM results suggest that four factors (assistance from religious organization, access to transportation, rural-urban migration, and type of roof) significantly increased the predicted probability of MTS community placement while two factors (size of land and annual household income) significantly decreased the predicted probability of community placement. Extension support from public and civil society groups, and commercial crop production significantly increased the predicted probability of program placement within households following community selection. Also, an increase in the number of individuals in a household with at least a junior secondary education and those receiving help from relatives outside of the village significantly decreased the probability of program placement within a household. 252 6.1 Binary Probit Analysis This section assumes that like most development programs the MTS policy at the onset was confronted with the two major tasks of selecting the right: a) communities and b) households for effective program implementation. Hence under section 6.1 I use Binary Probit Models (BPM) to investigate factors that influence: a) MTS placement in communities and b) program placement within households following community selection. MTS community selection and household participation are modeled as dependent variables (DV) regressed onto other explanatory livelihood indicators such as education, disease among household members older than 12 years, help from relatives outside of the village and migration. The dichotomous or binary (0,1) nature of the dependent variables MTS community (1) vs. non-MTS community (0) and MTS participant household (1) vs. non-MTS participant household (0) suggests that the underlying properties of Ordinary Least Square (OLS) methods such as the Linear Probability Model (LPM) may be violated when used in modeling MTS community and program participant selection. Söderbom (2011), Pindyck and Rubinfeld, (1997) and Wooldridge, (2002), demonstrate that for certain combinations of explanatory variables, it is possible for binary DVs to take on values less than zero or greater than one in which case OLS estimates can no longer produce the best linear unbiased estimator (BLUE). OLS estimations with binary DV thus may become biased and inefficient (see Pindyck and Rubinfeld, 1997, Park, 2010, Nagler, 1994 and Söderbom, 2011). Probit models also known as “Binary Probit Models (BPM)” or “Binomial Probit” presents the best alternative to OLS regressions in analyzing the relationships between binary dependent variables (DV) and a set of explanatory or independent variables (IV) such as in this study. Section 6.1.1 below describes the general estimation procedure for running a BPM and interpreting the results. Section 6.1.2 presents the findings from the survey data analysis. 253 6.1.1 Functional form for Probit Model In the binary choice situations confronted in this research, the probit model as explained above is expected to outperform the LPM simply because in the BPM the original linear model typical in OLS regressions (see equation 1 below) is transformed such that predicted probabilities fall strictly between the range 0 and 1 for all values of explanatory variables (EV). Hence for the BPM increases in explanatory variables for example may be associated with increase (or decreases) in the dependent variable . This requirement thus suggests the use of a cumulative probability function F, with the resulting probability function written as: --------1 Where: : Is the constant representing values of the DV with no change in all other EVs. : Is the coefficient that explains how the DV is affected by a unit change in the EV holding all other factors constant. : Is a vector of ith explanatory variable. Transforming the model using a cumulative uniform probability function, yields the constrained version of the linear probability model: -------2 The probit model developed in this section assumes the existence of a theoretically continuous latent variable which is neither observed nor measured directly yet influences community and household selection into the MTS (see Pindyck and Rubinfeld, 1997). Hence while is not observed nor measured, a dichotomous indicator of may be determined by observing changes in the EVs . Thus if represent a dummy variable which takes on the 254 iY)()(iijiiZFXFP=+=iijXijiiXP+=iZiZiZijXiY value 1 if a community or household is selected to participate MTS and 0 otherwise, then for each community or household in our sample, represents a critical cut off value which translates the underlying latent index MTS community or household selection. From the preceding discussion equation 2 above can be re-written as: --------3 This livelihoods study assumed that MTS program implementers or the Forest Service Department (FSD) were faced with questions regarding: 1) Which communities should be selected into the MTS and 2) which households should be selected into the MTS following community selection. The index represents the underlying propensity of community or household to have been selected to participate in MTS given . If we assume that the index representing the underlying propensity to participate is a linear function of (livelihood indicators) then the probit model simply provides a suitable means of estimating the slope and intercept parameters of the relationship between the index and . In other words the probit analysis solves the problem of how to obtain estimates for the parameters and while at the same time obtaining information about the underlying index (Pindyck and Rubinfeld, 1997). From equation 3 we conclude that the Ghana Forest Services Department’s (FSD) decision to select communities and also households into the MTS is based on the following assumptions: Since our probit model assumes is a normally distributed random variable, then the probability that can be computed using the cumulative normal probability function: 255 *iZiZijiiXZ+=iZi*iiZZijXiZijXiiZ** if MTSinto householdor community select not Do if MTSin eparticipat tohouseholdor community Select iiiiZZZZ*iZ*iiZZ ---------4 ---------5 where (s) in equation 4 represents a normally distributed random variable with a zero mean and a unit variance. By construction the predicted probability will lie in the (0, 1) interval. We can therefore interpret the predicted probability resulting from the probit model in equations 4 as an estimate of the conditional probability that a community or household will be selected into MTS, given that the that community or household’s livelihoods indicators is . The research data collected on household livelihood indicators thus allows for the determination of whether individual observations falls in one of the two categories ( = select MTS community or household) or a second category ( = do not select community or household into MTS). 6.1.2 Probit results from household livelihood survey Tables 6.1.1 and 6.1.2 below describe variables used in the BPM analysis to determine factors that influence MTS community selection. A total of 878 households were surveyed in June and July of 2009 for the BPM (439 recall for 1999 and another 439 for 2009) however only the 1999 (before MTS) recall data of 439 that was used in the BPM. In order for the BPM to be successfully estimated, all 439 responses were recoded to take on values of zeros (0) and ones (1). Fifteen explanatory variables were included in the model however five were found to significantly influence the predicted probability of MTS community selection and their expected marginal effects. Of the five variables that significantly influence MTS selection, four increase the predicted probability of MTS community selection while one significantly decreases the 256 dseZFPiZsii−−==2221)(ijiiiXPFZ+==−1iPiP*ijijXX*iiZZ*iiZZ predicted probability pf community selection. The BPM in table 6.1.3 is discussed in detail under section 6.1.4 and where appropriate findings from focus group discussions are used to triangulate those from the probit analysis. Table 6.1.2 below provides clarity on the signs on BPM coefficients and the rational for their hypothesized effects on MTS selection. Table 6.1.1: Variable Definitions Variable Name Variable Definition MTS Community Selection MTS Community or Household (HH) Selection 0=No, 1=Yes 1. Primary Education 2. Secondary Education 3. Illness and disease 4. Help from relatives 5. Extension Support 6. Help from religious organizations Highest level of education in the HH is fist six years of basic education or less (i.e. Nursery, Kindergarten or Primary School)? 0=No, 1=Yes Highest level of education in the HH is Junior or Senior secondary? 0=No, 1=Yes How often HH members >12yrs old fall sick are hospitalized and cannot work? 0=Hardly (at most once a year), 1=Often (at least once a month) HH receives food, cash, labor assistance etc. from relatives outside of the village but in Brong Ahafo? 0=No, 1=Yes HH normally receives gov’t or NGO support such as training or technical advice on crop production or forest management? 0=No 1=Yes HH normally receives cash, farm inputs, labor assistance etc. from local church, mosque or religious organization? 0=Never, 1=At least once a year 7. Size of household How many people live in the HH? 0=Few (<5) 1=Many (>5). 8. Migrant Work At least one HH member >12yrs works outside of village daily or long term (>3months)? 0=No, 1=Yes 9. Size farmland owned by the household Total land owned by HH? 0=Small (<2acrs) 1=Medium to Large (>2acres). 10. Crops are mostly produced for markets Crops are mostly produced for the market? 0=Disagree 1=Agree 257 Table 6.1.1 (cont’d) Variable Name Variable Definition 11. Trend or History of Goat Production 12. Annual Household Income 13. Type of Roof 14. Transportation Services 15. Ownership of Sewing Machine 16. Ownership of Functioning Bicycle Trends in HH Goat and Sheep Production? 0=Increased/Deceased/Stable But Not Sufficient 1= Increased/Deceased/Stable but Sufficient Combined annual HH income in the last 10 year? 0=<500 GHC 1=>500 GHC Type of Roof? 0=Thatch/Palm Leaf/Raffia 1=Corrugated/Aluminum Sheet How long it takes for HH to access any public or commercial transport service? 0=Long(>1hr) 1=Short(<1hr) Does anyone in the HH have a functioning sewing machine? 0=No 1=Yes Does anyone in the HH have a functioning bicycle? 0=No 1=Yes 258 Table 6.1.2: Hypothesized Effects of Explanatory Variables on Selection Variable Name Hypothesized Effect on Selection Community Household Explanation of Hypothesized Effects on MTS Selection 1. Primary Education (PE) + + 2. Secondary Education (SE) 3. Illness and disease 4. Help from relatives outside village - - - - - - This study assumes that in Ghana individuals with some basic education (BE) to a large extent can read and write but have no significant difference in income from others with similar or no formal education that choose to migrate to larger cities in search for off- farm employment (see Haar, 2009). Unlike BE holders, individuals with at least some secondary education (SE) have higher off-farm employment opportunities with relatively higher compensation in larger cities hence more likely than BE holders to out- migrate in search for greener pastures in larger cities. The relatively higher tendency for individuals with SE to migrate out of rural communities and thus abandon farming and forestry work makes them poor candidates for the MTS program. The foregoing discussion suggests that an increase in the number of individuals in a community/household with only BE increases the predicted probability of MTS community and household selection while an increase in the number of individuals with SE or higher decreases the predicted probability of selection. MTS by its nature is labor intensive; hence frequent illness among individuals in a within households older than12 years was hypothesized to decrease the predicted probability of community and household selection. Help particularly in the form of cash from relatives outside of the household provides a cushioning effect against extreme hardship hence this variable was hypothesized to significantly decrease the predicted probability of MTS program placement. 259 Variable Name Hypothesized Effect on Selection Community Household Explanation of Hypothesized Effects on MTS Selection Table 6.1.2 (cont’d) 5. Extension Support + + In most rural communities in Ghana, the most influential and resource rich communities generally and households specifically tend to have better developed social networks with public, private and civil society actors. However, the major determinant of which communities receive support is the transaction costs of providing services. Generally, easy access to road networks and proximity to government and NGO offices reduces significantly these transaction costs (including cost of extension and advisory services). This study thus assumed that communities receiving frequent extension support from public and civil society actors must have reasonable transaction costs associated with the provision of such services. Hence a marginal increase in the number of households in a community already receiving extension support from public and civil society was hypothesized to increase significantly the predicted probability of MTS community and household placement. 6. Help from religious organizations + + 7. Size of household + + Religious organizations play a key role in providing not only spiritual and moral support to the rich and the poor in rural communities but also material and financial support especially to the resource poor. This study assumes that the poorer a community the more responsibility religious groups assume in supporting households within that communities (be it financial or material). Hence it is hypothesized that an increase in the number of households receiving support from religious organizations increase significantly the predicted probability of both MTS community and household selection. Resource poor households in rural communities in Ghana often tend to have larger families mostly because this provides a poor household with cheap labor force needed for farming activities. Since the MTS relies on cheap abundant labor for success, this variable was hypothesized to increase the predicted probability of both MTS community and household selection. 260 Variable Name Hypothesized Effect on Selection Community Household Explanation of Hypothesized Effects on MTS Selection Table 6.1.2 (cont’d) 8. Migrant Work + - Rural-urban migration in Ghana particularly among the youth is well documented (see Twumasi-Ankrah, 1995 and Boakye-Yiadom and McKay, 2007). Migration outside of the village for work in the long-term may provide alternative sources of income for households that don’t have to depend or deplete the forest for sustenance; however in the short term it reduces the labor force needed for MTS to succeed. Though migrating for work outside of rural communities provides an alternative source of income sometimes remitted back into rural communities, “most rural-urban migrants often do not possess employable skills other than farming hence end up in the informal sector where wages are low and irregular (Haar, 2009, p.46).” Having taken themselves out of the rural labor force and most probably not able to remit sufficient financial resources back into their rural communities, poorly educated rural-urban migrants leave their communities even worse-off. The MTS program through its benefit sharing scheme and guaranteed tenure rights is intended to attract land insecure households particularly the youth back into agriculture and forestry. Hence it is expected that the program will be biased in favor of forest communities with large unemployed youth populations with a relatively good chance of migrating to urban centers. Hence Migrant Work is hypothesized to significantly increase the predicted probability of MTS community selection. However, once a community is selected it is expected that Migrant Work will significantly decrease household placement as household populations significantly reduced by migration reduces the chances of success of the MTS program. 9. Size of farmland owned by household - - Since MTS is meant to temporally provide resource poor households with cheap access to farmlands, average household land ownership of more than two acres was hypothesized to decrease the predicted probability of community and household selection. 261 Variable Name Hypothesized Effect on Selection Community Household Explanation of Hypothesized Effects on MTS Selection Table 6.1.2 (cont’d) 10. Crops mostly produced for markets + + More often than not what discourages commercial production in most rural communities is poor access to road and transportation networks and services coupled with long distances from major commercial centers/markets not the lack of livelihoods development projects such as the MTS. Extremely high transaction costs of market participation among remotely situated communities often restrict production to mere subsistence. Hence while most livelihoods development programs such as the MTS may wish to target the poorest communities (with majority of households engaged in subsistent production), program placement decisions are often weighed against returns on program resources (e.g. financial and human). It is thus logical that the MTS like other livelihoods programs will target communities with relatively lower transaction costs of market participation so that land allocated under the MTS may be used for commercial crop production. It is assumed that commercialize crops produced from plots allocated under the MTS generates income for program beneficiaries and discourages illegal harvest and sale of teak trees before their designated rotation periods. From the foregoing discussion market-oriented production was thus hypothesized to significantly increase the predicted probability of MTS community and household placement. 262 Variable Name Hypothesized Effect on Selection Community Household Explanation of Hypothesized Effects on MTS Selection Table 6.1.2 (cont’d) Part of the objective of the Ghana’s MTS program is to assist participants establish alternative livelihood programs so as to ensure an alternative source of subsistence between canopy closure and scheduled rotation periods. Three types of livestock (goats, pigs and cane rats/grasscutters) were promoted among participating MTS communities to provide a source of alternative income during the period between canopy closure (three years after initial plantation establishment) and the first (eight years), second (15 years) and third (25 years) possible rotation periods. Since success of the MTS’ alternative livelihoods strategy hinges on rapid multiplication of the selected livestock during the first three years of the policy, an initial parent stock of goats, pigs and cane rats (Thryonomys swinderianus) (locally known as grasscutter or bush meat) were distributed among selected MTS households believed to have substantial experience breeding/raising these livestock. To produce sufficient livestock populations for future redistribution however required the active participation of local “master breeders” and a dedicated MTS group willing to provide labor to support the breeding efforts. Given the importance of the alternative livelihoods program (i.e. commercial livestock production) in the overall success of Ghana’s MTS it was assumed that communities with a large number of individuals with substantial experience with commercial goat, pig and grasscutter production will be highly favored for program placement. Hence trend or history of goat production was hypothesized to significantly increase the predicted probability of MTS community and household placement. 263 11. Trend/ history of goat production + + Variable Name Hypothesized Effect on Selection Community Household Explanation of Hypothesized Effects on MTS Selection Table 6.1.2 (cont’d) 12. Annual household income + - Annual household income and expenditure have been used for several decades by economist to help understand socio-economic and financial well-being of urban and rural households. Hence combined annual household income was included in the BPM to enrich our understanding of how a marginal increase in the number of households in a community with average annual income above a certain threshold affects MTS community selection. This study assumed that annual household income is relatively higher among communities with easy access to major road networks and transportation services because easy access to roads significantly lower transaction costs of market participating at the district and regional levels. The preceding argument suggests that access to major road networks as well as low transaction cost in a large part is what dictates the level of household income and not necessarily the lack of development projects such as the MTS. Hence it was expected that the MTS program in 1999 will disproportionately favor communities with relatively lower transaction costs of market participation and hence those with higher income. It was thus hypothesized that a marginal increase in the number of households in a community with average annual income greater than 500 Ghana Cedis (GHC) (about $500 US Dollars in 2009) significantly increased the predicted probability of MTS community placement. It was also hypothesized that once a community is selected households in that community with annual income less than 500 Ghana Cedis (GHC) will be favored for program placement. 13. Type of roof + - In the 1990s, most houses were roofed with thatch, leaves or bamboo. Having aluminum or corrugated roof is considered a status symbol and hence an indicator of wealth thus having aluminum roof over houses was expected to decrease the predicted probability of community and household selection. 264 Variable Name Hypothesized Effect on Selection Community Household Explanation of Hypothesized Effects on MTS Selection Table 6.1.2 (cont’d) 14. Access to road and transportation services + + This study assumed that the “poorest of the poor” remote communities and households are located more than one hour walk away from major roads and commercial transport services. The study also assumes that extremely high transaction costs involved in market participation (and not lack of development project) coupled with the high transaction costs of MTS project support to remote communities decrease the predicted probability of MTS selection. Communities situated less than an hour’s walk to major roads thus increase the predicted probability of MTS selection. 15. Ownership of sewing machine - - In most rural Ghana, a sewing machine is a luxury item often included in the bride prize by well-to-do grooms. Since they provide fulltime employment and hence alternative income for women in rural communities, increase in the ownership of sewing machine is hypothesized to decrease the predicted probability of community and household selection. 265 Variable Name Hypothesized Effect on Selection Community Household Explanation of Hypothesized Effects on MTS Selection Table 6.1.2 (cont’d) 16. Ownership of bicycle + + In most rural communities in Ghana bicycles play an important role in transporting people and commodities between communities and markets. In the early 1990s and years prior the presence of motorized vehicles in rural Ghana were very few and scattered while phones (including landlines and cellphones) were virtually non-existent. In the absence of motorized vehicles and phones, bicycles provided a convenient means of shuttle between places and also a means to relay vital information to distant communities. Hence bicycles on the one hand may be considered a luxury item in communities where few exist while on the other hand, they are necessary for facilitating the flow of vital information for private and communal benefit. Because bicycles facilitate commerce at the household, village and regional levels by facilitating transport of commodities as well the flow of timely market information it is reasonable to argue that the presence of bicycles will positively influence MTS community placement. By commercializing crop and livestock production MTS program participants may have a higher chance of sustaining their livelihoods until the teak rotation periods thus preventing premature and illegal harvesting of teak before scheduled rotations. Because bicycles also provide a means for participating communities to also engage frequently with program planners to resolve group and individual member issues a marginal increase in the number of bicycles in a community or household was hypothesized to significantly increase the predicted probability of MTS community and household placement. 266 6.1.3 Hypothesized effects of BPM variables on MTS community selection Of the 15 explanatory variables included in the BPM in table 6.3 below, 11 were hypothesized to significantly increase (positively influence) the predicted probability of MTS community placement while four were hypothesized to significantly decrease (negatively influence) the predicted probability of MTS community placement. Out of the 11 variables hypothesized to significantly increase MTS community placement, only four (Help from local church, mosque or religious organization, Time it takes for HH to access any transport service, Member of HH > 12 years works outside of the village and Housing unit has aluminum roof) significantly increased MTS placement as hypothesized. Of the 11 variables, three others (Primary Education, Trends in goat and sheep production, Household has a functioning bicycle) also increased MTS program placement however their impact was insignificant. Four other variables (Extension support from Government and NGOs, Size of household, Crops are mostly produced for markets and Combined annual household income) hypothesized to significantly increase MTS community placement rather decreased program placement albeit insignificantly. Out of the 15 variables included in the BPM four were hypothesized to significantly decrease the predicted probability of MTS community placement. Out of the four variables predicted to significantly decrease program placement, only one variable (Size of total land owned by the household) impacted program placement as hypothesized. While the three other variables (Illness and disease among HH members >12 years, Help from relatives outside the village but in BA and Household has a functioning sewing machine) also decreased the predicted probability of MTS community placement their impact was insignificant. Though not all the 15 variables included in the BPM influenced the predicted probability of MTS community selection as hypothesized, an attempt is made to discuss each variable so as 267 to gain a better understanding of their importance in affecting MTS community placement. The six variables that significantly influenced the predicted probability of MTS community selection are discussed next under section 6.1.4 while the rest of the nine are discussed under section 6.1.5. The sections that follow first provide a brief summary statistics of each variable followed by an interpretation of the marginal effects on community selection. 268 Table 6.1.3: BPM Analysis of Factors that Influence MTS Community Selection -1999 MTS Community Selection Hypothesized Effect on Selection Impact of EV on MTS Community Selection Marginal Mean Effects after Probit (dy/dx) X 0.440 0.219 0.688 0.606 0.085 (0.172) -0.026 (0.172) -0.034 (0.159) -0.202 (0.158) 0.023 (0.047) -0.007 (0.048) -0.009 (0.043) -0.054 (0.042) 1. Primary Education 2. Illness and disease among HH members >12 years 3. Help from relatives outside the village but in BA 4. Extension support from Government and NGOs 5. Help from local religious organization 6. Size of household 7. Size of total land owned by the household 8. Time it takes for HH to access any transport service 9. Crops are mostly produced for markets 10. Trends in household goat and sheep production 11. HH Member > 12 yrs and works outside village 1) + Positively associated, - Negatively associated with MTS Community Selection; 2) dy/dx is for discrete change of dummy variable from non-MTS selection (0) to selection (1); 3.) HH=household; 4) Significant levels **** p<0.001, *** p<0.01, **p<0.05, *p<0.10 with robust standard errors in parentheses; 5) X= Mean; 6) Wald test of rho=0: chi2(15) = 115.52, Prob > chi2 = 0.000 + - - + + + - + + + + (0.044) -0.050 (0.044) 0.004 (0.042) (0.159) -0.180 (0.154) 0.014 (0.153) (0.043) -0.006 (0.053) (0.159) -0.023 (0.189) 0.225**** 0.146**** 0.202**** 0.795**** 1.072**** -0.433*** -0.115*** 0.588 (0.157) (0.040) (0.310) (0.034) 0.551 0.399 0.499 0.257 0.533**** 0.392 0.141 269 Table 6.1.3 (cont’d) MTS Community Selection 12. Combined annual HH income 13. Housing unit has thatch or aluminum roof 14. Household has a functioning sewing machine 15. Household has a functioning bicycle 16. Constant Observations R-Squared (%) Hypothesized Effect on Selection + + - + 439 22.79 Impact of EV on MTS Community Selection -0.310* (0.173) 0.939**** Marginal Mean Effects after Probit (dy/dx) X -0.091* (0.054) 0.232 0.247**** 0.451 (0.166) -0.280 (0.194) 0.142 (0.172) 0.106 (0.224) (0.041) -0.083 (0.061) 0.038 (0.045) 0.182 0.278 Percent correctly predicted (%) 1) + Positively associated, - Negatively associated with MTS Community Selection; 2) dy/dx is for discrete change of dummy variable from non-MTS selection (0) to selection (1); 3.) HH=household; 4) Significant levels **** p<0.001, *** p<0.01, **p<0.05, *p<0.10 with robust standard errors in parentheses; 5) X= Mean; 6) Wald test of rho=0: chi2(15) = 115.52, Prob > chi2 = 0.000 78.36 270 6.1.4 Factors that significantly influence MTS community selection All else equal, factors that significantly increased the predicted probability of MTS community selection, and their expected marginal effects (evaluated at their Means X) on the probability of MTS community selection include; household that normally receives support in the form of cash, farm inputs and labor assistance from their local church, mosque or religious organizations (0.146, p<0.001), household has nearby access (i.e. spends less than one hour) to public or commercial transportation service (0.225, p<0.001), at least one household member older than 12 years works outside of the village daily or long term />3months (0.202, p<0.001) and the average household in the community has aluminum or corrugated roof (0.247, p<0.001). Holding all other factors equal, a marginal increase in the number of households in a community with land ownership of two acres or more decreases significantly the predicted probability (- 0.115, p<0.001) of MTS community selection and the expected marginal effects (evaluated at their Means X). Similarly, a marginal increase in the number of households in community having an annual income of 500 Ghana Cedis or more decreased significantly the predicted probability (-0.091, p<0.05) of MTS community selection and the expected marginal effects (evaluated at their Means X). Subsections 6.1.4.1 through 6.1.4.5 below discuss in detail all six factors that significantly influenced the predicted probability of MTS community selection and the expected marginal effects (evaluated at their Means X). 6.1.4.1 Help from religious organizations From the BMP results above, approximately 50% (n=220) of respondents said they received assistance in the form of cash, farm inputs, and farm labor from local church, mosque or religious organization in 1999. The BPM results also suggest that the expected marginal effect resulting from the average household in a community receiving any kind of support from 271 religious organizations in 1999, ceteris paribus, on the probability of community selection is 0.146. This results suggests that a marginal increase in the number of households in a community receiving support from religious organization significantly increases the predicted probability of MTS community selection (0.146, p<0.001) and the expected marginal effects (evaluated at their Means X). The quote below from a focus group interview in Asuofre community helps explain the importance of religious groups in helping the poor in rural communities in the research area. “I came to this town in 1953. I make it a point to attend church every Saturday because this was part of my upbringing. The church provides us with valuable advice on life and also helps to bring us together to support each other during times of need (Nana Kwame Tsreme, Asuofre Key informant interview, 2009).” The above quote confirms that an important source of livelihood support in most rural communities in Ghana is faith-based organizations. Typically, the type and level of support received by households provides a glimpse into the livelihood status of households and the community as a whole. Hence communities with a large number of households receiving regular assistance to meet basic needs such as food, clothing a shelter may be deemed extremely poor and thus ideal candidates for the MTS or similar programs. Also, since the success of the MTS hinges on available cheap labor to establish and monitor plantation forests, communities with labor support programs or networks provided through faith-based organizations may be viewed more favorably by the MTS as this helps fulfil the basic needs for program success (cheap labor). 6.1.4.2 Size of land owned by household The research findings reported in table 6.1.3 indicates that approximately 59% (n=258) of the study population owns at least two acres of land in 1999. Total land size owned by households was hypothesized to significantly decrease the predicted probability of MTS community selection. According to the BPM results, the expected marginal effect resulting from 272 additional household units in a community owning more than two acres of farmland, ceteris paribus, on the probability of community selection is -0.115. As hypothesized for the model, the predictive power on the explanatory variable for land size is negative and significant (-0.115, p<0.01) suggesting that land ownership is an important criteria in the MTS community selection. The quotes below from three of the research communities help shed light on availability and ownership/tenure of land within the research communities. “My parents founded this town sometime in 1942 or 1943 but the land we have been cultivating since then belongs to other chiefs. Recently, we the elders petitioned the government to allow our people to cultivate portions of the forest so that we the landless can also feed our families. Thankfully, the government granted us permission to farm portions of the Yaya forest (Amangoase, MTS focus group interview, 2009).” “Most of the farmlands in Konsua and Malamkrom communities were allocated to the people by the government within the Yaya Forest Reserve. A number of households in these two communities own land through inheritance or leases. Land for settlement is under the powers of the chief and as custom demands, anyone who needs land for settlement must consult the chief and his elders (UDS- ITTFPP Group 55 Development Report on Konsua and Malamkrom, 2007, p. 30)” The above statements from Amangoase focus group interview as well as the UDS- ITTFPP (2007) suggest that secure tenure rights are a serious issue in the research communities. Though the MTS through its benefit sharing scheme attempts to address the problem of tenure security, proper selection of communities to include the most land insecure households is the first step to addressing tenure security problem and stem illegal encroachments on forest reserves. The variable land ownership is thus included in the model to help understand the impact of land ownership and tenure security on MTS program placement. 6.1.4.3 Access to transportation services According to the survey about 55% (n=241) of respondents said they spent less than an hour (relatively short time) accessing major roads and transportation services in 1999 compared 273 to another 45% that spent more than an hour (relatively long time) in the same year. By 2009 however access to major roads improved among an additional 10% of respondents thus increasing the number of households that spend less than hour accessing major roads from a 1999 baseline of 55% to approximately 65% (286) in 2009. The descriptive statistics suggests that in 1999 a relatively larger number of households in the study communities may have been located far away from major paved roads and thus walked several miles to access these facilities. The quote below describes access to transportation in one of the research communities: “On average it takes us at least one and half hours to walk from our village to the nearest major/paved road and another half hour to get to the nearest market. For those carrying load, the average time to the nearest major market is roughly three hours. Sometimes if you are not lucky you can wait for several hours without getting transport. On those days we just return home with our loads (Ahwene, focus group interview, 2009).” Compared to communities with easy access to major road and transportation networks and services remote villages with little or no access have a relatively lower socio-economic status. The extremely high transaction costs of market participation among remote communities/ households located further away (at least one hour’s walk) from major paved roads/tarmacs prohibits them from participating in the district and regional economy. Hence it is argued that high transaction costs of market participation and not the presence of fewer development projects is what accounts for the relatively higher poverty levels among remotely situated communities in the research area. The quote below from Chambers (2006) explains how high transaction costs of participating in development assistance not only prevents remote villages from participation but also limits development projects to communities concentrated along major roads networks. “Most learning about rural conditions is mediated by vehicles. Starting and ending in urban centers, visits follow networks of roads. With rural development tourism, the hazards of dirt roads, the comfort of the visitor, the location of places to visit and places 274 for spending the night, and shortages of both time and fuel dictate a preference for tarmac roads and for travel close to urban centers. The result is overlapping urban, tarmac and roadside biases (Chambers, 2006, p.17).” Access to major roads and transportation services was thus included in this study to gauge communities and households’ relative distance from major road networks and transportations services and placement into the MTS. Hence in line with what was hypothesized for the model, the average time taken by households in a community to access major roads and commercial transport services significantly increased the predicted probability (0.225, p<0.001) of MTS selection and the expected marginal effect resulting from the average household in a community spending less than one hour to access major transports service in 1999. 6.1.4.4 Member of household migrates outside of community for work The household survey result indicates that in 1999 approximately 14% (n=62) of respondents had at least one member of the household older than 12 years working outside of the community either daily or long-term (> 3 months). From the BPM results in table 6.1.3, the expected marginal increase in the number of households in a community having a household member older than 12 migrating to work outside of the community in 1999, ceteris paribus, on the probability of community selection is 0.202. In line with what was hypothesized, the predictive power of the explanatory variable migrant work is positive and significant (1.072, p<0.001) suggesting that an increase in the number of households with at least one individual commuting for work outside of the community significantly increased the predicted probability of MTS community selection and the expected marginal effects (evaluated at their Means X). The quotes below provide a glimpse into the migration situation of the research communities. “Some of our young ones have travelled to seek greener pastures….they have travelled outside the community to find jobs (Malamkrom, focus group interview one, 2009)”. 275 “Some of the women from the village travel to the next town Tromeso every day and sometimes weekly to do hair dressing. Other residents also leave the town for relatively long periods from a few weeks to six months at a time…..they often go to Techiman and Wenchi. Trends in migrant work have not changed much over the last 10 years (Ntema, focus group interview, 2009).” There are a lot of people who migrate from this town on a daily basis to work in Chraa, Sunyani and Wenchi Township. Also, there are those who migrate elsewhere to work for at least six months at a time before returning to the community (Buoku, focus group interview, 2009).” The fact that “Migrant Work” significantly increased the predicted probability of MTS community selection suggests as hypothesized that the MTS program in-deed favored communities with relatively high rural-urban migration particularly among the youth thus suggesting also that this particular variable was an important predictor of community placement. 6.1.4.5 Combined annual household income According to table 6.1.3 approximately 23% (n=102) of households in all 19 research communities had annual income greater than 500 GHC in 1999. The BPM results also suggests contrary to what was hypothesized that the expected marginal increase in the number of households in a community earning more than 500 GHC in 1999, ceteris paribus, on the probability of community selection is -0.091. In contrast with what was hypothesized, the effect of annual household income is negative and significant (-0.091, p<0.10) suggesting that the MTS program contrary to what was hypothesized favored the poorest communities. 6.1.4.6 Housing unit roofed with corrugated iron or aluminum roofing sheets The expected marginal effect resulting from an increase in the average number of houses in a community roofed with corrugated iron sheets in 1999, ceteris paribus, on the predicted probability of community selection was 0.247. In line with what was hypothesized, the marginal effect on MTS community selection resulting from an improvement in the explanatory variable 276 “roof type” from thatch to corrugate sheets was positive and significant (0.247, p<0.001) suggesting that MTS program prioritized communities with relatively higher percentage of houses with corrugated roofing sheets. The discussion below provides clarity on the importance of roof type in predicting livelihood status and the rationale behind the hypothesized effect of roof type on MTS program placement. The type of roofing material used by rural households in roofing their houses (thatch or corrugated sheets) has been used in the past as a proxy to gauge household’s livelihood status (see Maleson et al, 2008). In the 1990s and years prior, majority of houses in rural parts of Ghana were roofed with thatch from raffia (leaves from palm trees), straw, hay and mud/clay. In recent years however, there has been a gradual shift away from thatch towards aluminum or corrugated roofing materials. Compared to corrugated sheets, thatch roofing materials have been touted to keep the interior of houses relatively cooler during the hotter hours of the day. In spite of the advantage regulating room temperatures, thatch roofs are vulnerable to forest fires particularly during the dry season and frequent leaks during torrential rainfall. Thatch roofs are also known to harbor biting insects and crawlers. The combined effects of insects chewing on thatch roofing materials coupled with rainwater and heat hastens the decay of the plant materials thus making thatch less desirable compared to aluminum sheets. Additionally, aluminum roofing sheets offer opportunities for rainwater harvesting and storage for communities with limited water availability (Adovor, 2012). The quotes below sheds light on the shifting patterns from traditional thatch and mud to corrugate or aluminum sheets among rural households in Ghana. “Years ago, we built our homes with mud. Today, we build with bricks and blocks. Most houses these days are roofed with aluminum sheets (Asuakwa non-MTS focus group interview, 2009).” “We received roofing sheets after the 1982 fires. It was the Roman Catholic Missionaries who came to our aid with these, but the majority of homes are roofed with thatch 277 (Abrefakrom, MTS focus group interview, 2009).” “We used thatch, bamboo and wood but recently we have begun using aluminum roofing sheets (Amangoase, MTS focus group interview, 2009).” From the foregoing discussion, and the selected quotes from focus groups, it is logical to assume that recent shifts in roofing materials from thatch to aluminum signify an improvement in household’s livelihoods/physical capital status. Hence the expectation that the MTS program will target predominantly poorer communities with relatively larger number of households with thatch roofs in an effort to improve their physical capital status and hence livelihoods. A close examination of the distributional patterns of roof types in the research communities suggests that proximity to urban centers and possible participation in market economies most likely explains the observed patterns of distribution and not necessarily the presence of fewer livelihood improvement programs such as the MTS. For example, in 2009 approximately 52% (n=228) of households surveyed in the research communities had corrugated roofs compared to only 45% (n=198) that had it in 1999. A closer observation of the distributional pattern of households with corrugated roofs in the research communities also reveals that communities further away from Sunyani (the Brong Ahafo regional capital) had fewer houses with corrugated roofs compared to those closer. For example, the 2009 ground trotting georeferenced exercise conducted on all housing units in the 19 research communities revealed that approximately 57% of the 1,041 houses around Yaya reserve had corrugated roofs compared the 31% of the 192 houses around Nsemre. Similarly, of the 213 housing units georeferenced in the five research communities around the Sawsaw reserve only 12% had corrugated sheets (Section 6.2 and Appendix M). The fact that Yaya communities are closer to Sunyani and its urban markets relative to Nsemre and Sawsaw (been the furthers) appears to suggest that proximity to urban centers where corrugated 278 sheets are sold coupled with relatively reduced transaction costs of market participation accounts for the differences in roof type and not necessarily the presence or absence of development projects such as the MTS. The preceding discussion suggests that forest communities closer to Sunyani have relatively low transaction costs of market participation thus develop faster socio- economically and accumulate sufficient financial capital to change their roofs. Also, since the MTS program is operated by the Ghana Forest Services Division (FSD) located in Sunyani, it is likely that the high transaction costs of implementing the project in remote communities created a natural preference for communities much closer to Sunyani and relatively easy to access. In the BPM above having aluminum roof was thus hypothesized to significantly increase the predicted probability of MTS community selection and this was confirmed by the model results. 6.1.5 Other factors that influence the probability of MTS community selection Section 6.1.5 discusses nine other independent variables in the BPM in table 6.1.5 above that influence MTS community selection albeit insignificantly. While these variables according to the BPM may not significantly impact the predicted probability of MTS community selection, they nonetheless enrich our understanding of how they may influence future program placement. 6.1.5.1 Primary Education Ghana has since independence (in 1957) implemented several educational reforms. However, none has had more impact on basic education than the 1996 “Free Compulsory Universal Basic Education (FCUBE),’ first initiated in 1987 and fully implemented beginning 1996. Basic education within the Ghanaian context is the first six years of primary education and three years of secondary education (see Little, 2010 for more insight into Ghana’s educational reforms). From table 6.3 above approximately 44% (n=193) of respondents had at least one individual in the households with some primary education (i.e. first 6 years of basic education) in 279 1999. The quotes below from the focus group interviews sheds light on the state of primary education in three of the research communities. “About ten years ago, our children had to walk nearly seven kilometers every day to get to school. The situation is different today as we have a school here. Child education has improved and about 80% of our children are currently in school. (Abrefakrom, non-MTS focus group interview, 2009).” “The reason why some children are not in school today is because some parents don’t realize the value of education hence their refusal to send their wards to school. There were not many children in school a few years back, but the situation today has changed. About 70% of children are in school (Kofitsumkrom, focus group interview one, 2009)”. “About 90% of our children are now in school (Malamkrom, focus group interview one, 2009)”. This study assumes that individuals with only some primary education are more likely than those with secondary education or higher to remain in their rural communities and contribute to the agricultural and forestry workforce. The premise for this assumption is that individuals with at least some secondary have a greater chance of securing off-farm employment in urban cities hence more likely to migrate out of the rural communities. Primary education was thus hypothesized on to significantly increase the predicted probability of MTS community placement whiles secondary education was hypothesized to decrease the probability of placement. According to the BPM results in table 6.3 the expected marginal effect from the average households in a community having some primary education, ceteris paribus, on the probability of community selection is 0.023. Though the explanatory variable primary education increased the predicted probability of program placement as predicted, the impact on the predicted probability of community placement was insignificant (0.023, p>0.10). In spite of the BPM results, there are several reasons to believe that the variable primary 280 education is an important selection criterion worthy of consideration in future reforestation programs. First, the MTS program stands to benefit from selecting program participants with at least some basic education because these individuals often tend to read and write and thus may be able to comprehend some of the language in the contractual “benefit sharing” documents (legal documentation governing tenure rights) signed prior to joining the MTS program. Secondly, the technical nature of forest plantation establishment may require periodic participation in technical forest extension training programs hence having some basic reading and writing skills may help program participant derive the outmost benefits from these extension programs. Thirdly if one of the goals of the MTS program is to improve rural livelihoods then one way to achieve this may be through creating employment opportunities for the youth particularly those with some primary and secondary education who are most likely to migrate to urban centers in search of greener pastures. Of the two groups of potential rural-urban migrants in the research communities, those with only some primary education most probably have fewer “employable skills other than farming Haar (2009)” hence have a lesser chance of securing meaning employment and remitting income back to their families left behind in their rural communities. Hence targeting youth with at least come primary education in the long run will minimize the rural-urban migration among forest reserve communities and to a large extent guarantee that the MTS program has the necessary manpower needed for program success. 6.1.5.2 Illness and Disease In most Ghanaian villages particularly those heavily dependent on forests, crop production and animal husbandry activities, and taking on adult responsibilities starts at a rather early age of around 12. Hence children 12 years and older are often expected to take on major responsibilities of tending crop and livestock, helping on Taungya plantations and harvesting 281 fuel wood and other forest products for home construction. The variable “illness and disease among household members older than 12” is included in the BPM to gauge how physical fitness of the potential labor force for the MTS in a community in 1999 affected program placement at the community level. As reported in table 6.3 above, approximately 22% (n=96) of respondents said every month in 1999, at least children 12 years and older within their households “fall sick are hospitalized and cannot work.” From the BPM result, the expected marginal effect resulting from frequent illness among children 12 years and older, ceteris paribus, on the probability of community selection is -0.007. Due to the reliance of MTS on cheap, abundant but healthy wage labor, frequent illness and disease affliction particularly among the average household members 12 years and older in a community was hypothesized to significantly decrease the predicted probability of MTS community placement. As hypothesized frequent “illness and disease among household members older than 12” decreased the predicted probability of MTS program placement however the impact was insignificant suggesting this variable may not have been an important criteria for selecting MTS communities in 1999. 6.1.5.3 Help from relatives outside the village but resident in Brong Ahafo The results from table 6.1.3 suggests that in 1999 an average of 69% (n=302) of households received assistance in the form of food, cash or labor occasionally from their relatives living outside of the village but resident in Brong Ahafo Region. It was assumed in the BPM that communities that have significant number of households receiving cash and material assistance from relatives in the Brong Ahafo region would be less likely to be selected for the MTS compared to those that do not. The foregoing discussion is based on the assumption that households’ receiving cash and material support will depend less on the forest for their livelihoods relative to those that don’t and hence will be less favored by the MTS. The variable 282 help from relatives outside of the village but in BA was thus hypothesized to significantly decrease the predicted probability of MTS community placement. Though not significant, the BMP results confirmed as hypothesized that the expected marginal effect resulting from a marginal increase in the number of households in a community receiving any form of assistance from relative outside of the village but in the Brong Ahafo Region, ceteris paribus, on the predicted probability of MTS community selection is -0.054. Knowing the level of support received from relatives outside of the village but in Brong Ahafo may be important in determining which communities desperately need the MTS program to support their livelihoods. However, the fact that this variable is insignificant may be an indication that this particular variable was important in determining community placement into the MTS program in 1999. 6.1.5.4 Extension support from Government and Non-government Organizations (NGOs) The results from the BMP indicate that approximately 61% of (n=268) of respondents received extension support in the form of training or technical advice periodically on agriculture related activities from the public and civil society groups in 1999. From the BPM results, the expected marginal effect resulting from a marginal increase in the number of households in a community receiving extension support from public and civil society groups in 1999, ceteris paribus, on the predicted probability of community selection is -0.032. Contrary to what was hypothesized, the coefficient on the explanatory variable is negative and insignificant suggesting that periodic extension support from public and civil society groups may not have been an important factor influencing MTS community selection in 1999. 6.1.5.5 Household size According to the results from the BMP approximately 26% (n=114) of respondents had relatively larger household size (more than five members) in 1999 compared to the majority 74% 283 (n=325) who have small family units (less than five members). From the BPM results, the expected marginal effect resulting from an increase in the number of households in a community having more than five members, ceteris paribus, on the predicted probability of community selection is -0.074. Contrary to what was predicted for the BPM the coefficient on the explanatory variable household size was negative and insignificant suggesting that household population may not be an important factor influencing MTS community placement in 1999. Historically traditional and modified Taungya systems around the globe depend on cheap and abundant peasant labor to establish forest plantations. In the case of Thailand’s Forest Village System (FVS), entirely new villages, each consisting of at least 100 family units were established for the purpose of establishing Taungya plantations (see Gajaseni, 1992, p.5). Each household or family unit was responsible for planting and tending 1.6 ha of Taungya plantation (see Gajaseni, 1992, p.5). In the case of Ghana, migrant communities (mostly from the three Northern regions of Ghana) provide a cheap source of labor for the government’s taungya efforts. Pressure on family units to complete field clearing and planting on a given plot within specified time periods put larger family units at greater advantage over smaller ones. Timely field clearing and planting are particularly important plantation management practice that guarantees uniformity in maturity indices (such as age and size) for the purpose of harvesting. Hence one of the goals of the MTS is to have vast areas established within specified time periods even if that meant reallocating un- cleared fields to different family units. The statement below sheds light on one families struggle to meet the terms specified under Ghana’s national Taungya program: “I was allocated a small parcel of land in the Nsemre Reserve under the national Taungya system, but since I did not have the manpower to clear and plant trees on all of it, the leftover parcel was re-demarcated and assigned to a different household under the ADB Taungya program (Ahwene focus group interview, 2009).” 284 From the statement above, it was expected that communities with the average household size larger than five would have a higher likelihood of selection into the MTS. Household size larger than five was thus hypothesized to significantly increase the predicted probability of community selection however this was not the case. It may be possible that the range of household size chosen for the BPM might have been lower than the threshold necessary to influence MTS community selection as predicted. 6.1.5.6 Crops mostly produced for markets (commercial production) The result from the BPM suggests that in 1999 majority of respondents (60% or n=263) produced crops on a subsistent level (i.e. primarily for consumption within the household) while the minority (40% or n=176) commercialized or sold parts of their harvests in nearby markets. Among the numerous factors that influence commercial crop production, access to markets (which also depends on access to road networks and transportation services) and the size of farmland available for production are probably the two most important factors. This study assumed that “all things equal” crop farms two acres or larger is sufficient to sustain commercial crop production activities for the average family of five or less individuals. According to the BPM results in table 6.3 only 26% (n=114) of respondents had more than five individuals within their household in 1999. Hence the fact that in 1999 more than 70% of households had five or less individuals and approximately 60% (n=263) had farm sizes greater than two acres yet only 40% (n=176) sold parts of their crop harvests in markets (see table 6.3) suggests that some other factor other than small land size explains the high level of subsistence crop production among majority of the respondents. Although majority of respondents (70%) cultivated more than two acres in 1999 yet only 55% (n=241) had access to roads and transport services suggests that the greatest bottleneck to commercial crop production in the research 285 communities is road and transportation access not farm size. The explanatory variable market-oriented production was thus hypothesized to increase the predicted probability of MTS community placement in the same way the variable access to major roads and transportation services impacted program placement (see section 6.1.4.3 above). Compared to communities situated closer to markets and therefore have easy access, remote villages with little or no access are relatively poorer socio-economically. Extremely high transaction costs of market participation among remote communities prohibit them from participating in vibrant city or urban market economies. Hence it is argued that high transaction costs of market participation and not the presence of fewer development projects is what accounts for the relatively higher poverty levels among remotely situated communities in the research area. From the statement above and the discussion in table 6.2, it was hypothesized that “all things equal” a marginal increase in the number of households in a community selling parts of their crops in a market will significantly increase the predicted probability of MTS community placement and the expected marginal effects (evaluated at their Means X). The BPM results however indicate that the expected marginal effect resulting from a marginal increase in the number of households in a community selling parts of their crops in a market in 1999, ceteris paribus, on the probability of MTS community selection is -0.050. The fact that the impact of market-oriented crop production was negative and insignificant probably suggests that this particular variable was important in determining MTS community placement in 1999. 6.1.5.7 Trends in goat production In order to select the appropriate indicator for livestock that influences MTS community placement, goats, pigs, and cane rats (Thryonomys swinderianus also locally known as 286 grasscutter or bush meat) production levels among research households were first compared to determine whether production levels sufficiently met household meat demands in 1999. The survey results suggests that only 7% (n=30) and 8% (n=36) of respondents respectively produced sufficient grasscutters and pigs to meet their households’ meat demands in 1999 compared to approximately 35% of households that produced sufficient goats meat for their households in the same year. The importance of commercial livestock production to the success of the MTS program coupled with the fact that a relatively larger number of households produced enough goats to meet their households’ meat requirements in 1999 explains why goat production trends was included in the BPM in table 6.3. Trend in goat production was hypothesized to significantly increase the predicted probability of MTS community placement and the expected marginal effects (evaluated at their Means X). According to table 6.1.3 the expected marginal effect resulting from a marginal increase in the number of households in a community producing sufficient levels of goats and sheep for the household in 1999, ceteris paribus, on the probability of community selection is 0.032. The coefficient on the explanatory variable “trends in household goat production” was positive as predicted however the effect on MTS community placement in 1999 was insignificant suggesting that this particular variable may not have been an important factor influencing selection in 1999. 6.1.5.8 Ownership of functioning sewing machine Summary statistics in table 6.1.3 above suggests that in 1999 approximately 18% (n=79) of households had at least one functioning sewing machine in their household. Because sewing machines are relatively expensive, they are generally considered a luxury item reserved for the privileged few in most rural communities in Ghana. A common route by which sewing machines 287 make it into rural communities is by being included in the prize paid by affluent grooms in exchange for their bride’s hand in marriage. Sewing machines generally provide an alternative source of off-farm income for a household hence ownership of this physical asset was hypothesized to significantly decrease the predicted probability of MTS community selection. The BPM result in table 6.1.3 appears to support the hypothesized impact of sewing machine ownership on MTS community selection though the effect was not significant. From the BMP results, the expected marginal effect resulting from an increase in the number of households in a community owning a functional sewing machine, ceteris paribus, on the probability of community selection is -0.083. The coefficient on the explanatory variable sewing machine ownership is negative and insignificant suggesting that ownership of a sewing machine may not have been an important factor influencing community placement in 1999. 6.1.5.9 Ownership of functioning bicycle Table 6.1.3 suggests that in 1999 approximately 28% (n=123) of households in the study had at least one functioning bicycle in their household. Because bicycles facilitate commerce at the household, village, and regional levels (in terms of timely flow of information that benefited the community as a whole) it was hypothesized that ownership of bicycles will significantly influence MTS community placement. While the hypothesized effect of bicycle ownership was positive as predicted the effect on program placement at the community level was not significant. According to the BPM results the expected marginal effect resulting from an increase in bicycles in a community, ceteris paribus, on the probability of community selection is 0.038. The coefficient on the explanatory variable bicycle ownership was positive as hypothesized however it was insignificant suggesting the variable bicycle ownership may not be an important factor in determining MTS community placement in 1999. 288 6.1.6 Power of BPM to predict MTS community placement The results presented in table 6.1.3 suggest that the BPM had a rather weak explanatory power as indicated by the relatively low adjusted R2 of 22.79%. The R2 value suggests that approximately 23% of the variations in the model are explained however the model is relatively weak in predicting factors that determine MTS community placement. Additionally, the model also suggests that approximately 78% of predicted outcomes were actually true whereas only 22% were false. 6.1.6.1 Binomial Probit Classifications Table 6.1.4 below suggests that 66.51% (n= 292) of the households surveyed for the BPM analysis lived in a community that participated in the MTS program and the probit model predicted accurately that this is indeed true. Similarly, for 11.85% of households said their communities did not participate in the MTS program the BPM again predicted accurately that this was also true. For 7.06% (n= 31) of households however that said their communities participated in the MTS program in 1999 the BPM predicted wrongly that their communities did not have the program. Similarly, for 14.58% (n= 64) households that said their communities did not participate in the MTS program the BPM predicted wrongly that they had the program. On the whole approximately 78% of the predicted outcomes were accurate while 22% were not. 289 Table 6.1.4: Classification of Predicted and Actual MTS Community Placement Binomial Probit Model Classifications N = 1446 True Participation True Non-Participation Total Frequency % Frequency % Frequency % Predicted Participation Predicted Non-Participation Total 292 31 323 Correctly Classified: 78.36% = (292+52)/(439/100) 66.51 7.06 64 52 73.58 116 14.58 11.85 26.42 356 83 439 81.09 18.91 100.00 290 6.1.7 Probit Results for MTS household placement The BPM results in table 6.5 below describe factors that influenced MTS household selection in 1999 following community placement in the same year. Out 878 households surveyed in 2009 for the BPM (439 recall for 1999 and another 439 for 2009) only the 1999 (before MTS) recall data for 323 households residing in MTS program communities was used in the BPM analysis in this section. In order for the BPM to be successfully estimated, all 323 responses were recoded to take on values of zeros (0) and ones (1). Sixteen explanatory variables were included in the model however only four were found to significantly influence the predicted probability of MTS household selection and their expected marginal effects. Of the four variables two significantly increased the predicted probability of MTS household selection while two others significantly decreased the predicted probability of selection. Sixteen explanatory variables were included in the BPM in table 6.5 below and out of the 16, eight were hypothesized to significantly increase (positively influence) the predicted probability of MTS household placement while eight others were hypothesized to significantly decrease (negatively influence) the predicted probability of MTS program placement within a household. Out of the eight variables hypothesized to significantly increase MTS household placement, only two (Extension support from Government and NGO, and Crops are mostly produced for markets) significantly increased program placement as hypothesized. Six others (Primary Education, Help from local church, mosque or religious organization, Size of household, Trends in household goat and sheep production , Access to road and transportation services, and Household has a functioning bicycle) also increased MTS household placement albeit insignificant. Of the eight other variables hypothesized to significantly decrease the predicted probability of MTS household placement only two (Junior or Senior Secondary Education and 291 Help from relatives outside the village but in BA) significantly decreased program placement as hypothesized. Four others (Frequent Illness and disease among children less than 12 years, Member of household older than 12 years works outside of the village, combined annual household income and Housing unit has corrugated iron/aluminum roof) decreased program placement however their impact was insignificant. Contrary to what was hypothesized, size of land and household ownership of a functioning sewing machine increased the predicted probability of MTS household placement although the impact was insignificant. Because all explanatory variables included in the BPM in table 6.3 above were previously discussed (regardless of whether they were significant or not) no attempt is made to repeat a discussion of all variables in this section. Hence the discussion of table 6.5 below focuses on only the four significant variables and their marginal effects evaluated at their Means X. The sections that follow discuss the findings from the BPM analysis. 6.1.7.1 Junior or Senior Secondary Education From the BMP results above, approximately 35% (n=113) of respondents said they had one member of the household with at least a Junior or Senior Secondary Education in 1999. The BPM results also suggest that the expected marginal effect resulting from a household in an MTS community having at least a Junior or Senior Secondary Education in 1999, ceteris paribus, on the probability of household selection is -0.163. This results suggests as hypothesized that a marginal increase in the number of individuals within a households with a Junior or Senior Secondary Education significantly decreased the predicted probability of MTS household placement (-0.163, p<0.10) and the expected marginal effects (evaluated at the Mean X). 292 Table 6.1.5: BPM Analysis of Factors that Influence Household Selection into MTS-1999 MTS Household Selection 1. Primary Education 2. Junior and Senior Secondary Education 3. Illness and disease among children < 12 yrs 4. Help from relatives outside the village but in BA 5. Extension support from Government and NGOs 6. Help from local religious organizations 7. Size of household 8. HH Member > 12 yrs and works outside village 9. Size of total land owned by the household 10. Crops are mostly produced for markets 11. Trends in household goat and sheep production Hypothesized Effect on Selection Impact of EV on MTS Household Selection Marginal Mean Effects after Probit (dy/dx) X + - - - + + + - - + + 0.114 (0.189) -0.429* (0.187) -0.222 (0.179) -0.320* (0.171) 0.582*** (0.158) 0.103 (0.158) 0.253 (0.180) -0.059 (0.205) 0.030 (0.159) 0.322* (0.169) 0.185 (0.161) 0.043 (0.071) -0.163* (0.071) -0.085 (0.069) -0.116* (0.060) 0.218*** (0.059) 0.039 (0.059) 0.093 (0.064) -0.022 (0.078) 0.011 (0.059) 0.118* (0.061) 0.068 (0.059) 0.477 0.353 0.217 0.700 0.579 0.529 0.272 0.180 0.551 0.381 0.381 1) + Positively associated, - Negatively associated with MTS Community Selection; 2) dy/dx is for discrete change of dummy variable from non-MTS selection (0) to selection (1); 3.) HH=household; 4) Significant levels **** p<0.001, *** p<0.01, **p<0.05, *p<0.10 with robust standard errors in parentheses; 5) X= Mean; 6) Wald test of rho=0: chi2(15) = 115.52, Prob > chi2 = 0.000 293 Table 6.1.5 (cont’d) Hypothesized Effect on Selection Impact of EV on MTS Household Selection Marginal Mean Effects after Probit (dy/dx) X -0.032 (0.073) -0.013 (0.061) 0.059 (0.061) 0.064 (0.073) 0.016 (0.064) 0.214 0.545 0.632 0.173 0.288 MTS Household Selection 12. Combined annual HH income in the last 10 year 13. Housing unit has corrugated iron/aluminum roof 14. Time it takes to access any transport service 15. Household has a functioning sewing machine 16. Household has a functioning bicycle 17. Constant - - + - + -0.084 (0.192) -0.035 (0.163) 0.157 (0.160) 0.175 (0.206) 0.043 (0.173) -0.038 (0.243) Observations R-Squared (%) Percent correctly predicted (%) 1) + Positively associated, - Negatively associated with Household Selection; 2) dy/dx is for discrete change of dummy variable from non- MTS selection (0) to selection (1); 3.) HH=household; 4) Significant levels **** p<0.001, *** p<0.01, **p<0.05, *p<0.10 with robust standard errors in parentheses; 5) X= Mean; 6) Wald test of rho=0: chi2(16) = 33.03, Prob > chi2 = 0.007 323 7.75 65.94 294 6.1.7.2 Help from relatives outside the village but resident in Brong Ahafo The results from table 6.1.5 above suggests that in 1999 an average of 70% (n=226) of households received assistance in the form of food, cash and labor occasionally from their relatives living outside of the village but resident in Brong Ahafo Region. Receiving assistance particularly in the form of cash and food from relative outside of the village but in the Brong Ahafo Region was hypothesized to significantly decrease the predicted probability of MTS household placement and the expected marginal effects (evaluated at their Means X). The expected marginal effect resulting from an increase in assistance received from relative outside of the village but resident in the Brong Ahafo Region in 1999, ceteris paribus, on the probability of household selection is -0.032. The fact that the variable in question decreased significantly (- 0.320, p<0.10) MTS household placement suggests that it must have been important in deciding MTS household placement in 1999. 6.1.7.3 Extension support from Government and Non-government Organizations According to table 6.1.5 above approximately 58% of (n=187) of respondents received extension support in the form of training or technical advice periodically on agriculture related activities public and civil society groups in 1999. From the BPM results, the expected marginal effect resulting from an increase in support from public and civil society groups by a household in 1999, ceteris paribus, on the predicted probability of community selection is 0.582. The fact that household support from public and civil society groups significantly increased (0.582, p<0.01) the predicted probability of MTS household placement as hypothesized suggests that this particular variable may have been an important factor influencing MTS program placement within a household in 1999. 295 6.1.7.4 Crops mostly produced for markets (commercial production) The result from the BPM suggests that in 1999 only 38% (n=123) of households in MTS communities produced crops for markets. Household commercial crop production was hypothesized to increase the predicted probability of MTS household placement (see section 6.1.4.3 above). In accordance with what was hypothesized “all things equal” a marginal increase in commercial crop production activity within the household (from non-commercial to commercial) significantly increased (0.322, p<0.10) the predicted probability of MTS household placement and the expected marginal effects (evaluated at their Means X). The fact that the impact of market-oriented crop production was positive and significant as predicted suggests that this particular variable was important in determining MTS community placement in 1999. 6.1.8 Power of BPM to Predict MTS Community Placement The results presented in table 6.1.5 suggest that the BPM had a very weak explanatory power as indicated by the relatively low adjusted R2 of 7.75%. The R2 value implies that only approximately 8% of the variations in the model are explained suggesting also that the model is relatively weak in predicting factors that determine MTS household placement. 6.1.8.1 Binomial Probit Classifications Table 6.1.6 below suggests that 54.14% (n=175) of the households surveyed for the BPM analysis said their household was selected to participate in the MTS program and the probit model predicted accurately that this is indeed true. Similarly, for 11.76% (n= 38) of households that said their communities did not participate in the MTS program the BPM again predicted accurately that this was indeed true. 296 For 8.67% (n= 28) of households however that said they also participated in the MTS program in 1999 the BPM predicted wrongly that they did participate in the program. Similarly, for 25.39% (n= 82) of households that said their household did not participate in the MTS program the BPM predicted wrongly that they had they participated in the program. On the whole approximately 66% of the predicted outcomes were accurate while 44% were inaccurate. 297 Table 6.1.6: Classification of Predicted and Actual MTS Household Placement Binomial Probit Model Classifications N = 1446 True Participation True Non-Participation Total Frequency % Frequency % Frequency % Predicted Participation Predicted Non-Participation Total 175 28 203 Correctly Classified: 65.94% = (175+38)/(323/100) 54.18 8.67 62.85 82 38 120 25.39 11.76 37.15 257 66 79.57 20.43 100.00 298 6.1.9 Summary, conclusion, and policy implications In this research, Binomial Probit Models were used to determine factors that influenced Ghana’s Modified Taungya System (MTS) placement into community and households. Included in the analysis was a total of 439 household surveys conducted in 19 forest fringe communities in Yaya (10), Nsemre (4) and Sawsaw (5) forest reserves in Ghana’s Brong Ahafo Region. Included in the Yaya survey were 203 MTS participant and 120 non-participant households. A total of 116 household from Nsemre and Sawsaw forest reserves where the MTS program did not exist were also included in the analysis. Fifteen explanatory variables were included in the first model to determine factors that influenced community placement while 16 were included in the second model to determine factors that influenced household placement after community selection. The BPM results suggest that four factors (help from religious organization, access to transportation services, migration among household members 12 years and older, and type of roof) increased significantly the predicted probability of MTS community placement while two others (size of land owned by the households and annual household income) significantly decreased the predicted probability of community placement and the marginal effects evaluated at the Mean X. According to BPM two variables (extension support from public and civil society groups, and commercial crop production) increased significantly the predicted probability of program placement within the household following community selection. However having individuals with a junior or secondary school education within the household and receiving help from relatives living outside of the village significantly decreased the predicted probability of program placement within a household. 299 Though it may appear from the BPM results that only a few variables in both models were significant, this exercise is nevertheless extremely important since knowing the impact of the MTS program cannot be done independently of examining the determinants of community and household selection where the program operates. The fact that this study found very few variables in both models (particularly the BMP to determine factors that influenced household placement) to significantly influence program placement suggest that the distribution of the program among communities and households may have been somewhat random thus guaranteeing a fair chance of program placement in communities and households. 300 Part II: Impact of Spatiotemporal Factors on Community and Household Placement into Ghana’s Modified Taungya System Abstract This study used four primary datasets collected in 2009 to construct Binomial Probit Models (BPM) to investigate factors that influenced placement of Ghana’s MTS reforestation program into communities and households. Included in the dataset are: 1) demographic information, 2) proximity of households to communal physical capital assets, 3) condition of household dwellings and 4) household livestock production activities. To determine proximity of households to other communal physical assets, a total of 7,011 georeferenced still photographs of houses and other physical assets in Yaya (4,915), Nsemre (1,182) and Sawsaw (914) were analyzed. Also included in the analysis were questions on the number of adult males and females as well as children in the household. Household livestock ownership was also included in the survey instrument. The survey results suggest that 1,225 families lived in 1,041 housing units in Yaya, 244 Nsemre (213 houses) and 219 in Sawsaw (192 houses). Seven explanatory variables including number of adult males, type of roof and bathroom, proximity to mosques, communal markets, provision stores and toilets were found to significantly increase the predicted probability of MTS community placement. Conversely, nine variables including number of adult females, type of kitchen, proximity to church, corn mill, borehole, kindergarten and primary school, communal dumpster and major roads significantly decreased the predicted probability of community placement. Number of adult females in the household, type of bathroom as well as proximity to mosques and major roads significantly increased the predicted probability of MTS household selection. Proximity to communal dumpster and market significantly decreased the predicted probability of household selection/participation. 301 6.2 Binary Probit Models Generated from Community Maps Under section 6.2, recall livelihoods data was used to determine factors that influenced MTS households and community placement/selection in 1999. The BMP in this section uses four different sets of primary data generated in 2009 to investigate factors that influenced MTS community and household placement. The data used to construct BPMs under section 6.2 was thus generated from community mapping exercises conducted in 19 forest communities (see Appendices K, L, and M). The BPMs in table 6.2.3 and 6.2.4 thus predict the effect of 21 explanatory variables on MTS community and household participation. Four categories of explanatory variables including: 1) demographics, 2) communal physical capital assets, 3) condition of household dwellings and 4) household livestock production activity was used to predict MTS participation. The demographic variable for example investigated the influence of adult male and female populations as well as children of both sexes on MTS community and household participation. Household physical capital assets such as type of roof, bathroom, and kitchen as well as the physical construction of houses were also used to predict participation at both the community and household levels. The third set of explanatory variables explains how access in terms of availability and proximity to communal physical assets such as corn mills, churches, mosques, boreholes and schools influences program participation. Since alternative livelihood activities particularly livestock is integral to Ghana’s MTS, a fourth set of variables was included to capture the influence of pig, goat and sheep ownership on MTS community and household selection. Data on all four groups of explanatory variables were collected from 1,446 georeferenced household units and used to construct 184 community maps for the purpose of the BPM analysis. Before discussing the BMP results in tables 6.2.3 and 6.2.4, the explanatory 302 variables included in both tables are first defined in table 6.2.1 and their hypothesized effects on MTS community and individual household participation discussed in table 6.2.2. 303 Table 6.2.1: Household Variable Definitions Explanatory variables used in BPM Variable definition 1. Number of adult males in the household Total number of males in the household older than 12yrs old 2. Number of adult females in the household Total number of females in the household older than 12yrs old 3. Number of male children in the household Total number of male children in the household younger than 12yrs old 4. Number of female children in the household Total number of female children in household younger than 12yrs old 5. Type of roof Type of roof over housing unit (1=Corrugated, 0=Raffia Leaves/Thatch) 6. Type of housing construction Physical Construction of housing unit (1=Cement/Concrete, 0=Earth/Mud) 7. Type of bathroom Type of Bathroom (1=Cemented floor, 0=Mud/gravel) 8. Type of kitchen Type of Kitchen (1=Secure roofed space in HH for Kitchen, 0=No kitchen/household cooks under tree or in an open space on compound) 304 Explanatory variables used in BPM Variable definition Table 6.2.1 (cont’d) 9. Proximity of HH to Corn Mill (Nika Nika) 10. Proximity of HH to Church 11. Proximity of HH to Mosque 12. Proximity of HH to Borehole Relative distance of household or housing unit from various communal physical assets. 0=Non-Available 1=>180meters 2=121-180meters 3=61-120meters 4=0-60meters 13. Proximity of HH to Kindergarten 14. Proximity of HH to Primary School 15. Proximity of HH to Communal Dumpster 16. Proximity of HH to Communal Market 17. Proximity of HH to Kiosks/ Provision stores 18. Proximity of HH to Communal Toilet 19. Proximity of HH to Major Roads 20. Household rears Goats and Sheep Does household have any Goats or Sheep (2=Yes, 0=No) 21. Household rears Pigs Does household have any Pigs (3=Yes, 0=No) 305 Table 6.2.2: Hypothesized Effects of Explanatory Variables on Participation Explanatory variables used in BPM Hypothesized Effect on Participation Community Household Explanation of Hypothesized Effects on MTS Participation 1. Number of adult males in household 2. Number of adult females in household 3. Number of male children in household 4. Number of female children in household 5. Type of roof + - - - - Success of the MTS depends on the ability of individual and plantation owners’ to effectively protect the trees from illegal loggers. In most Taungya communities, men are expected to provide the muscle power needed for policing hence it was hypothesized that an increase in the number of adult men in a community or household increases significantly the predicted probability of MTS participation at both the community and household levels. Forestry work in rural Ghana generally involves manual tree felling and timber hauling perceived as labor intensive and thus engaged in mostly by men. The labor-intensive nature forestry activities coupled with the need to physically confront illegal loggers, led to the hypothesis that an increase in the number of males significantly increases the predicted probability of participation while an increase in the number of females and children in a community or household significantly decreases the predicted probability of program participation. Prior to the 1990s, most households in rural communities were roofed with thatch, leaves or bamboo. Having aluminum roof was considered an improvement in livelihood status and an indication that a household was well- off. Aluminum roof was thus considered a symbol wealth hence hypothesized to significantly decreases the predicted probability of MTS participation. 306 Table 6.2.2 (cont’d) Explanatory variables used in BPM Hypothesized Effect on Participation Community Household Explanation of Hypothesized Effects on MTS Participation 6. Type of housing construction 7. Type of bathroom 8. Type of kitchen - - - Like roof-type, prior to the 1990s the main structure of most houses in rural communities in Ghana were built with clay/mud bricks. Constructing building from cement blocks or concrete is a relatively recent development considered a symbol of wealth and thus hypothesized to significantly decrease the predicted probability of MTS participation. Due to the high cost of cement, most resource poor households particularly in rural communities are not able to afford cement to construct their homes let alone their bathrooms. In the research communities, cemented bathroom floors were considered a luxury hence hypothesized to significantly decrease the predicted probability of MTS participation. Some households in the research communities have a dedicated space or rooms within their households that serve as a kitchen. The resource poor generally have no such space and often cook in open spaces or under trees. For this reason, it was hypothesized that an increase in the number of households with no dedicated space for cooking decreases significantly the predicted probability of MTS participation. 307 Table 6.2.2 (cont’d) Explanatory variables used in BPM: Proximity to: Hypothesized Effect on Participation Community Household - + - - - - - - - 9. Church 10. Mosque 11. Corn Mill/Nika Nika 12. Borehole 13. Kindergarten 14. Primary School 15. Communal Dumpster 16. Communal Market 17. Kiosks/Provision stores Explanation of Hypothesized Effects on MTS Participation Availability and geographic dispersion of critical physical assets in rural communities are often not random. Several factors including the socioeconomic status influence availability and access to assets. In this study it is assumed that households with relatively higher socioeconomic status influence relevant public, private, and civil society groups to provide communal assets and then weigh in heavily on the location of these assets within their communities. It is therefore expected that communal assets would be clustered around the most influential households and or those with relatively higher socioeconomic status. This non-random distribution of assets helps explain the unequal access to assets (access gap) between resource rich and poor households. Unlike other physical assets, mosques are unique in that they tend to be concentrated in communities with predominately migrant populations from Ghana’s Northern Region. Ghana’s 2010 census suggests that 60% of the region’s population is Moslem and the region has the highest poverty rate nationally. Hence the study assumes that migrant communities from Ghana’s Northern Region have a high per capita concentration of mosques and are also relatively poorly endowed with other critical livelihood assets Thus contrary to what was hypothesized for other physical assets, an increase in households’ access to mosques in a community is expected to increase the predicted probability of MTS community and household participation. 308 Hypothesized Effect on Participation Community Household - - + + Table 6.2.2 (cont’d) Explanation of Hypothesized Effects on MTS Participation Due to the persistent foul odor associated with most communal toilets in rural communities it is expected that relatively wealthy households would be situated further away from these communal assets while poorer households would be clustered around these assets. The discussion in the preceding table suggests that communities with relatively lower socio-economic status generally have lower per capita concentration of critical physical assets (except mosques) and thus were prioritized by the MTS policy. Once a community is selected however, it is again assumed that the decision of household participation in the MTS is influenced by socioeconomic status. Hence households further removed from communal assets by virtue of their relatively poor socio-economic status will be more inclined to participate in the MTS. Thus, relatively easy access to transportation services for example should decrease the predicted probability of MTS participation. To ensure that MTS participants have an alternative source of subsistence between canopy closure (1-3 years) and scheduled rotation periods (8, 15 and 25 years), goat and pig/hog production was introduced as alternative livelihood options as part of the MTS. For the purpose of this study it is assumed that initial success in the goat and hog program hinges on first distributing the initial breeder heard to households with prior experience in raising goats, hence a unit increase in a household’s goat and pig ownership was hypothesized to increase the predicted probability of MTS participation. 309 Explanatory variables used in BPM 18. Communal Toilet 19. Proximity of HH to Major Roads 20. Household rears Goats and Sheep 21. Household rears Pigs 6.2.1 Binomial Probit Analysis Results- Community Participation in MTS 6.2.1.1 Hypothesized Effects of on MTS Community Participation Of the four demographic variables included in the BPM in table 6.2.3 one variable (male children under the age 12) impacted the predicted probability of MTS community participation differently from what was hypothesized for the model. Also, the result in table 6.2.3 suggests that three of the four physical characteristics of a housing unit (type of roof, type of construction of house and bathroom floors) influenced the predicted probability of community participation differently from what was hypothesized for the model. Of the 11 communal assets included in table 6.2.3 only three (communal market, kiosk/provision store, and communal toilet) influenced the predicted probability of MTS community participation differently from what was hypothesized for the model. In the sections that follow, a summary statistic of each significant variable is provided followed by a discussion of the variable’s effect on community participation. 310 Table 6.2.3: Impact of 21 Selected Livelihood Assets on MTS Community Participation Hypothesized Effect on Participation Impact of EV on MTS Community Participation Marginal Effects Mean after Probit (dy/dx) 0.007* 0.004 -0.003 0.005 0.000 0.003 X 1.769 1.677 1.660 0.136* 0.078 -0.053 0.096 0.004 0.050 -0.182*** -0.009** 1.631 0.066 0.005 0.325**** 0.077*** 0.467 0.232 0.056 0.238 0.333*** 0.503 -0.393*** 0.153 0.025 0.003 0.012 0.401 0.035*** 0.169 0.014 -0.021** 0.488 0.011 -0.657**** -0.034**** 1.412 0.131 0.192** 0.088 0.010 0.010** 0.005 1.357 MTS Community Participation 1. Number of adult males in the household 2. Number of adult females in the household 3. Number of male children in household 4. Number of female children in household 5. Type of roof over housing unit 6. Type of housing construction 7. Type of bathroom 8. Type of kitchen 9. Proximity of HH to Church 10. Proximity of HH to Mosque + - - - - - - - - + 1) + Positively associated, - Negatively associated with MTS Community Selection; 2) dy/dx is for discrete change of dummy variable from non- MTS selection (0) to selection (1); 3.) HH=household; 4) Significant levels **** p<0.001, *** p<0.01, **p<0.05, *p<0.10 with robust standard errors in parentheses; 5) X= Mean; 6) Wald test of rho=0: chi2(15) = 115.52, Prob > chi2 = 0.000 311 MTS Community Participation 11. Proximity of HH to Corn Mill (Nika Nika) 12. Proximity of HH to Borehole 13. Proximity of HH to Kindergarten 14. Proximity of HH to Primary School 15. Proximity of HH to Communal Dumpster 16. Proximity of HH to Communal Market 17. Proximity of HH to Kiosk (Provision store) 18. Proximity of HH to Communal Toilet Table 6.2.3 (cont’d) Hypothesized Effect on Participation Impact of EV on MTS Community Participation Marginal Effects Mean after Probit (dy/dx) X - - - - - - - - - -0.413**** -0.021*** 2.165 0.069 0.007 -0.657**** -0.034**** 1.983 0.134 0.009 -0.592**** -0.030*** 0.974 0.117 -0.198** 0.100 0.010 -0.010* 0.006 1.205 -0.859**** -0.044**** 0.976 0.120 0.013 0.531**** 0.027*** 0.839 0.121 0.011 0.545**** 0.079**** 2.416 0.185 0.023 0.417**** 0.073**** 1.492 0.181 0.021 -0.367**** -0.019** 2.623 0.081 19. Proximity of HH to Major Roads 1) + Positively associated, - Negatively associated with MTS Community Selection; 2) dy/dx is for discrete change of dummy variable from non-MTS selection (0) to selection (1); 3.) HH=household; 4) Significant levels **** p<0.001, *** p<0.01, **p<0.05, *p<0.10 with robust standard errors in parentheses; 5) X= Mean; 6) Wald test of rho=0: chi2(15) = 115.52, Prob > chi2 = 0.000 0.008 312 MTS Community Participation 20. Household rears Goats and Sheep 21. Household rears Pigs 22. Cons Observations R-Squared (%) Percent correctly predicted (%) Table 6.2.3 (cont’d) Hypothesized Effect on Participation Impact of EV on MTS Community Participation + + 1,446 81.49 95.85 0.028 0.080 0.081 0.144 0.653** 0.296 Marginal Effects Mean after Probit (dy/dx) 0.001 0.004 0.004 0.008 X 0.759 0.069 1) + Positively associated, - Negatively associated with Community Participation; 2) dy/dx is for discrete change of dummy variable from non-MTS participation (0) to participation (1); 3.) HH=household; 4) Significant levels **** p<0.001, *** p<0.01, **p<0.05, *p<0.10 with robust standard errors in parentheses; 5) X= Mean; 6) Wald test of rho=0: chi2 (21) =234.23, Prob > chi2 = 0.000 313 6.2.1.2 Factors that significantly influence MTS community participation 6.2.1.2.1 Demographics Two demographic variables; average number of adult males and average number of female children residing in a household significantly influenced the predicted probability of MTS community participation, and their expected marginal effects evaluated at their Means X. According to table 6.2.3 an increase in the number of adult males in a household significantly increased the predicted probability of MTS community participation (0.136, p<0.10), while an increase in the average number of female children in a household significantly decreased the predicted probability of community participation (-0.182, p<0.01). The signs on both explanatory variables conformed to what was predicted for the model. 6.2.1.2.2 Physical condition of household dwelling Out of the four variables describing the effect of a household’s physical condition on community and household participation, three variables (type of roof, type of bathroom and type of kitchen) significantly influence the probability of MTS community participation, and their expected marginal effects evaluated at their Means X. According to table 6.2.3 above, all things equal, an increase in the number of houses in a community with aluminum roofing sheets significantly increased the predicted probability of MTS community participation (0.325, p<0.001). Similarly, communities with a relatively higher number of households with cemented bathroom floors had a significantly higher predicted probability (0.333, p<0.001) of participation in the MTS program. Unlike improved roof-type and bathroom floors that increased the predicted probability of MTS community participation, all things equal, an increase in the average number of households with dedicated space or room within the households for cooking significantly decreased the predicted probability (-0.393, p<0.01) of MTS community 314 participation. Of the three explanatory variables discussed under section 6.2.2.2 only kitchen type influenced the predicted probability of MTS community participation in a manner that what was predicted for the model. Since it is assumed and generally accepted that aluminum roofing materials, cemented bathroom floors and dedicated space for cooking (improved kitchen) are signs of a relatively higher livelihood status, it was assumed that communities with relatively fewer households with these physical attributes will have a greater urge to participate in the MTS as a way to improve their livelihood status. It was also expected that as a strategy to effectively target the poorest of the poor communities the MTS policy will have a natural bias toward selection of communities that have relatively lower levels of these physical capital assets. 6.2.1.2.3 Availability and access to essential communal physical capital assets Out of the 11 variables describing availability and access to communal physical assets, eight variables (proximity to church, mosque, corn-mill, borehole, kindergarten, primary school, dumpster, and major roads) significantly influence the predicted probability of MTS community participation and their expected marginal effects (evaluated at their Means X) in a way that was predicted for the model. Three other variables (communal markets, kiosks and toilets facilities) however significantly influenced the predicted probability of MTS community participation contrary to that predicted for the model. According to table 6.2.3 above, all things equal, an increase in the number of households with access to a church in a community significantly decreased the predicted probability of MTS community participation (-0.657, p<0.001) while an increase in the number of households with access to mosques significantly increased the predicted probability (0.192, p<0.05) of community participation. The influence of availability and access to places of worship (churches and mosques) on community participation was just as predicted for the model. It should be noted 315 that while a few of the 19 research communities predate Ghana’s independence (1957) others were constituted fairly recently by migrants from Ghana’s three northern regions in search of agricultural and forestry work. Older communities with predominantly native populations tend to have a higher Christian population and thus more churches. Newer communities with predominantly migrants from the north tend to have a relatively higher Moslem population. The assumption is that migrant populations with weaker social and economic ties to their communities tend to have relatively lower livelihood assets and thus likely to participant in the MTS relative to predominantly native communities with majority resource rich Christian populations. The forgoing argument explains why an increase in the availability and access to churches was expected to decrease the predicted probability of MTS participation while a similar trend for mosques was expected to increase the predicted probability of participation. All things equal, an increase in the number of households with access to a corn-mill in a community significantly decreased the predicted probability of MTS community participation (- 0.413, p<0.001). Similarly, an increase in the number of households with access to boreholes significantly decreased the predicted probability (-0.657, p<0.001) of community participation. It is not uncommon for relatively poorer communities with a predominantly large migrant population to have very few or no corn-mill or boreholes. Hence communities without these basic physical capital assets tend to walk several miles to access them or in the case of water depend on streams, creeks and gutters in or around the community. A lack of access to corn-mill and borehole was thus predicted to increase a household’s predicted probability of participating in the MTS program. It is generally expected that an increase in the availability of basic education facilities such as kindergarten and primary schools in a community improves the population’s basic 316 educational levels and literacy (two indicators of human capital). However, in young migrant communities these educational facilities were very few or nonexistent and may take several years to be established. Communities without basic formal education facilities were thus expected to have relatively lower scores/levels of human capital indicators such as basic educational and literacy. Since one of the surest means of rural livelihoods improvement is investment in basic education it was expected that relatively poorer communities in need of basic educational facilities such as kindergarten and primary schools would be more likely to participate in the MTS and thus prioritized for the program. The results from table 6.2.3 suggests that all things equal an increase in the number of households in a community with access to kindergarten (- 0.592, p<0.001) and primary school (-0.198, p<0.05) decreases the predicted probability of MTS community participation. According to table 6.2.3 above, all things equal, an increase in the number of households with access to a communal dumpster significantly decreased the predicted probability of MTS community participation (-0.859, p<0.001) while an increase in the number of households with access to improved communal toilet facilities significantly increased the predicted probability (0.417, p<0.001) of community participation. All things equal an increase in the availability and access (in terms of proximity) of communal markets significantly increase the predicted probability (0.531, p<0.001) of MTS community participation. Similarly, an increase in the availability and access to neighborhood kiosks (provision stores) in a community also significantly increase the predicted probability (0.545, p<0.001) of community participation in the MTS program. Both indicators influenced MTS community participation in a direction different from what was predicted for the model. 317 Holding all other factors constant, an increase in a community’s access (in terms of distance and time) to major roads significantly decreased the predicted probability of participation (-0.367, p<0.001) in the MTS program. It is generally accepted that access to major road networks facilitates economic activities thus this study assumed that a community closer to major roads would have more economic opportunities and thus relatively higher financial livelihoods assets than those remotely situated. With much higher economic opportunities (made possible by the road) and relatively remote location from the forest reserves it was correctly predicted that an increase in a community’s access to major roads decrease the predicted probability of MTS participation. 6.2.1.2.4 Livestock production activity While both livestock variables goat and sheep as well as pig ownership/production influenced MTS community participation in conformity with predictions for the model, none of the two explanatory variables significantly influenced MTS community participation. 6.2.1.3 Predictive Power of BPM on MTS community participation Overall, the BPM in table 6.2.3 has a very strong explanatory power as indicated by the high adjusted R2 of 81.49%. The R2 value thus suggests that 81.49% of the variations in the model are explained and also that this model is very strong in predicting factors that determine MTS community participation. Additionally, the model suggests that about 96% of predicted outcomes were true (see table 6.2.4). 318 6.2.1.4 Binomial Probit Classifications Table 6.2.4 below suggests that 70.12% (n= 1014) of the households georeferenced live in an MTS community and the probit model predicted this to be actually true. However, for the 1.87% (n= 47) of georeferenced households that said they lived in non-MTS community the BPM predicted these wrongly. Similarly, 25.73% (n= 372) of georeferenced households said they live in a non-MTS community and the models predicted correctly while for another 2.28% (n= 33) that also said they live in a non-MTS community the model wrongly predicted their residence as MTS community. In all 95.85% of the predicted outcomes were accurately predicted while only 4.1% were falsely predicted. 319 Table 6.2.4: Classification of Predicted and Actual MTS Community Participation Binomial Probit Model Classifications True Participation True Non-Participation Total % 2.28 25.73 28.01 Frequency % 1047 399 1446 54.47 45.53 100.00 N = 1446 Frequency % Frequency Predicted Participation Predicted Non-Participation Total 1014 27 1041 Correctly Classified: 95.85% = (1014+372)/(1446/100) 70.12 1.87 71.99 33 372 405 320 6.2.2 Binomial Probit Analysis Results- Household Participation in MTS This study assumes that policy decision for recruitment within MTS program occurs at two different levels: first at the community level and second at the household level. While entirely different factors may influence policy decision to select and beneficiaries’ willingness to participate at both levels, the BPMs in this study evaluated the influence of the same 21 variables on policy decision to select and both community and household’s willingness to participate. Table 6.2.5 unlike 6.2.3 presents the findings of how these 21 variables influence households’ decision to participate in the MTS following community selection. 6.2.2.1 Hypothesized Effects of IVs on MTS Household Participation Out of the 21 independent variables included in the BPM in table 6.2.5 five significantly increased the predicted probability of MTS household participation while two decreased significantly the predicted probability of participation with an improvement in these explanatory variables. Of the four demographic variables in the model, only one (number of male children) influenced the predicted probability of MTS household participation in accordance with what was predicted for the model albeit insignificantly. Though not significant the model suggests, in accordance with what was predicted, that ceteris paribus an increase in the number of male children in a household reduces the predicted probability of household MTS participation. While it was predicted that an increase in the number of both male and female children under 12 years decreases the predicted probability of household MTS participation the model suggests the contrary for female children though the influence of children generally was insignificant. Similarly, only one out of the four variables included to gauge the influence of the “physical condition of household dwellings” influenced MTS household participation as predicted. Though not significant the model suggests in accordance with what was predicted, that 321 ceteris paribus an improvement in the type of kitchen in the household decreased the predicted probability of MTS household participation. For the purpose of this study it was assumed that households with improved kitchen (described in table 6.2.2) have relatively higher livelihood status most likely supported by other alternative income sources other than forestry thus might have a lesser interest in the MTS. In order to gauge the influence of a household’s relative proximity to communal physical assets on MTS participation, four different buffer zones (0- 60meters, 61-120meters, 121-180 and more than 180meters) were created around each physical asset. The BPM was then used to determine how a households’ location within a buffer zone influenced that household’s predicted probability of participation. Out of the 11 communal assets included in the model, households’ relative location to six communal assets influenced the predicted probability of MTS participation as predicted. Of the six communal physical assets found to influence participation as predicted, all things equal, two (proximity to communal dumpster and open-air markets) significantly influence the predicted probability of households’ participation in the MTS. Of the five communal physical assets that influence households’ MTS participation contrary to what was predicted, two (proximity to communal kiosks/provision stores) significantly influence the predicted probability of participation. 322 MTS Household Participation 1. Number of adult males in the household 2. Number of adult females in the household 3. Number of male children in household 4. Number of female children in household 5. Type of roof over housing unit 6. Type of housing construction 7. Type of bathroom 8. Type of kitchen 9. Proximity of HH to Corn Mill (Nika Nika) 10. Proximity of HH to Church 11. Proximity of HH to Mosque 12. Proximity of HH to Borehole 13. Proximity of HH to Kindergarten 14. Proximity of HH to Primary School 15. Proximity of HH to Communal Dumpster 16. Proximity of HH to Communal Market 17. Proximity of HH to Kiosk (Provision store) + - - - - - - - - - + - - - - - - 0.018 0.036 0.005 0.009 1.817 0.109**** 0.028*** 1.741 0.043 -0.016 0.034 0.049 0.035 0.051 0.147 0.189 0.146 0.270** 0.121 -0.008 0.097 0.011 0.044 0.015 0.050 0.088* 0.047 -0.060 0.056 0.005 0.059 -0.061 0.054 0.011 -0.004 0.009 0.013 0.009 0.013 0.038 0.050 0.038 1.689 1.609 0.568 0.497 0.076** 0.213 0.036 -0.002 0.025 0.003 0.012 0.004 0.013 0.023* 0.012 -0.016 0.015 0.001 0.016 -0.016 0.014 0.476 2.176 1.298 1.457 2.002 0.968 1.339 -0.138** -0.036** 1.030 0.062 0.016 -0.156*** -0.041*** 1.047 0.057 0.105* 0.015 0.028* 2.946 Table 6.2.5: Impact of 21 Selected Livelihood Assets on MTS Household Participation Hypothesized Effect on Participation Impact of EV on MTS Community Participation Marginal Mean Effects after Probit (dy/dx) X 1) + Positively associated, - Negatively associated with Community Participation; 2) dy/dx is for discrete change of dummy variable from non-MTS participation (0) to participation (1); 3.) HH=household; 4) Significant levels **** p<0.001, *** p<0.01, **p<0.05, *p<0.10 with robust standard errors in parentheses; 5) X= Mean; 6) Wald test of rho=0: chi2 (21) =234.23, Prob > chi2 = 0.000 0.015 0.058 323 MTS Household Participation 18. Proximity of HH to Communal Toilet 19. Proximity of HH to Major Roads 20. Household rears Goats and Sheep 21. Household rears Pigs 22. Cons Observations R-Squared (%) Percent correctly predicted (%) Table 6.2.5 (cont’d) Hypothesized Effect on Participation Impact of EV on MTS Community Participation Marginal Mean Effects after Probit (dy/dx) X - - + + 1,041 8.69 80.50 -0.009 0.064 -0.002 0.017 1.893 0.118** 0.031** 2.507 0.054 0.055 0.050 -0.041 0.107 -1.668**** 0.205 0.014 0.014 0.013 -0.011 0.028 0.770 0.075 1) + Positively associated, - Negatively associated with Community Participation; 2) dy/dx is for discrete change of dummy variable from non-MTS participation (0) to participation (1); 3.) HH=household; 4) Significant levels **** p<0.001, *** p<0.01, **p<0.05, *p<0.10 with robust standard errors in parentheses; 5) X= Mean; 6) Wald test of rho=0: chi2 (21) =234.23, Prob > chi2 = 0.000 324 6.2.2.2 Factors that significantly influence MTS household participation 6.2.2.2.1 Demographics Only one demographic variable; average number of adult females residing in a household significantly influenced the predicted probability of MTS household participation, and the expected marginal effects evaluated at the Means X. According to table 6.2.5 an increase in the number of adult females in a household significantly increased the predicted probability of MTS community participation (0.109, p<0.001). The influence of adult females in a household on MTS participation is contrary to what was predicted for the model. 6.2.2.2.2 Physical condition of household dwelling Out of the four variables describing the effect of a household’s physical condition on household participation in the MTS, only one variable (type of bathroom) significantly influence the probability of MTS community participation, and their expected marginal effects evaluated at their Means X. According to table 6.2.5 above, all things equal, an increase in the number of households with cemented bathroom floors significantly increased the predicted probability (0.270, p<0.05) of participating in the MTS program. 6.2.2.2.3 Availability and access to essential communal physical capital assets Mosques All things equal an increase in households’ access to communal mosques significantly increases the predicted probability of MTS household participation (0.088, p<0.10) and the expected marginal effects evaluated at the Mean X. This finding supports claims in the study of a positive relationship between households’ access to mosques and willingness to participate in the MTS (see explanation in table 6.2.2). However, to fully appreciate the the important of mosques in shaping MTS policy regarding community and households’ selection and participation a brief 325 background of the population distribution, religion and poverty situation in Ghana’s three northern most regions is provided in Appendix N. Communal Dumpster As predicted, the BPM results in table 6.2.5 suggests that an increase in the number of households with access to a communal dumpster decreases the predicted probability (-0.138, p<0.05) of household participation and the marginal effects evaluated at the Mean X. Access to communal physical assets within the context of this study is measured not only in terms of availability but also proximity to these physical assets. Communal Market Ceteris paribus an increase in the number of households with relatively easy access to communal markets significantly decreases the predicted probability (-0.156, p<0.001) of MTS household participation and the marginal effects evaluated at the Mean X. As explained in table 6.2.2 above, this study assumes that resource poor households are generally live in relatively remote area away from critical functioning physical assets thus limiting their access to these assets. Following the assumption in table 6.2.2 and the BPM results in table 6.2.5 leads to the conclusion that the further away households are situated from communal markets the higher the predicted probability of participation in a livelihood improvement project such as the MTS. Kiosk (Provision store) All things equal, an increase in households’ access to community/neighborhood kiosks contrary to what was hypothesized for the model significantly increased the predicted probability (0.105, p<0.10) of MTS household participation and the expected marginal effects evaluated at the Mean X. A possible explanation to the effect of kiosks on the predicted probability of participation may be explained by how this variable was treated in the study. In this study, all 326 kiosks were assumed to be equal with no regard for the value of items sold in each kiosk. It is thus logical that kiosks will typically carry items that surrounding households or passersby can afford. It was thus expected that kiosks in resource poor communities will carry less pricy items relative to those in resource rich communities. Further analysis that disaggregate stores based on the value of each kiosk taking into consideration the value of the items in the store may likely produce results in line with what was hypothesized for the model. Major/Paved Roads (Asphalted and Red Roads) All things equal, an increase in households’ access to major roads contrary to what was hypothesized for the model significantly increased the predicted probability (0.118, p<0.05) of MTS household participation and the expected marginal effects evaluated at the Mean X. A possible explanation to the effect of roads on the predicted probability of participation stems from the fact that both asphalted and non-asphalted roads for the purpose of simplicity were treated as major roads in the study. In the study region, asphalt roads generally connect major towns and markets and have relatively higher traffic, while recently paved red dirt/non-asphalt forest roads connect smaller towns and markets and have less traffic. Also, the immediate and primary function of gravel roads are to facilitate movements of timber from the reserves though with time these become major trading and commercial transport routes as communities expand along these routes. Further analysis using disaggregated non-asphalted and asphalted routes may likely produce results in line with what was hypothesized for the model. 6.2.1.3 Predictive Power of BPM on MTS household participation Overall, the BPM in table 6.2.3 may be said to have relatively weak explanatory power as indicated by the low adjusted R2 of 8.69%. The R2 value thus suggests that only 8.69% of the 327 variations in the model are explained and that this model is very weak in predicting factors that determine household participation. 6.2.1.4 Binomial Probit Classifications Table 6.2.6 below suggests that 1.54% (n= 16) of the houses georeferenced in the study said they had at least one member participating in the MTS program and the BPM predicted this to be true. Similarly, 78.96% (n= 1,014) of georeferenced houses said they did not any member participating in the MTS program and BPM predicted correctly that no MTS members were resident in these houses. However, for 18.44% (n= 192) of houses that had a resident MTS participant the BPM predicted falsely that there were no MTS participants in the household. Similarly, for 1.06% (n= 11) of houses that had no resident MTS participants, the BPM predicted wrongly that they had participants in the household. In summary, the model suggests that 80.50% of predicted outcomes were indeed true whereas only 19.50% were false. 328 Table 6.2.6: Classification of Predicted and Actual MTS Household Participation Binomial Probit Model Classifications N = 1446, n = 1041 Predicted Participation Predicted Non-Participation Total True Participation True Non-Participation Total Frequency 16 192 208 % 1.54 18.44 19.98 Frequency 11 822 833 % 1.06 78.96 80.02 Frequency % 27 2.59 1014 1041 97.41 100 Correctly Classified: 80.50% = (16+822)/(1041/100) 329 6.2.3 Summary, conclusions, and policy implications Sixteen out of the 21 variables thought to influence MTS Community selection (thus included in the model) indeed influenced community selection significantly (table 6.2.5). Ten out of these 16 variables influenced participation in manner predicted for the BPM while six influenced participation contrary to what was predicted. Of the six variables, five increased the predicted probability of selection (contrary to predictions) while one decreased the predicted probability of selection (contrary to predictions). Table 6.2.7: Impact of 21 Selected Livelihood Assets on MTS Community Participation No Explanatory Variable Hypothesized Effect BPM Result 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Number of adult males in the household Number of female children in household Type of roof over housing unit Type of bathroom Type of kitchen Proximity of HH to Church Proximity of HH to Mosque Proximity of HH to Corn Mill (Nika Nika) Proximity of HH to Borehole Proximity of HH to Kindergarten Proximity of HH to Primary School Proximity of HH to Communal Dumpster Proximity of HH to Communal Market Proximity of HH to Kiosk (Provision store) Proximity of HH to Communal Toilet Proximity of HH to Major Roads + - - - - - + - - - - - - - - - + --- ++++ +++ --- ---- -- ---- ---- ---- -- ---- ++++ ++++ ++++ ---- 1) Significant improvements in livelihood index: +++p<0.001, ++ p<0.01, + p<0.05, 2) Significant decline in livelihood index: ---p<0.001, --p<0.01, - p<0.05, 3) Insignificant improvements in livelihood index: + and 4) Insignificant decline in livelihood index: - 330 In terms of household participation following community selection (table 6.2.6), only seven out of the 21 variables included in the BPM significantly influenced the predicted probability of MTS household selection (thus included in the model). Four out of these seven variables influenced participation in manner predicted for the BPM while three influenced participation contrary to what was predicted. Of the four variables that influenced participation as predicted, two significantly increased the predicted probability of MTS household participation while two significantly decreased the predicted probability of participation. Table 6.2.8: Impact of 21 Selected Livelihood Assets on MTS Household Participation No Explanatory Variable 1 2 3 4 5 6 7 Number of adult females in the household Type of bathroom Proximity of HH to Mosque Proximity of HH to Communal Dumpster Proximity of HH to Communal Market Proximity of HH to Kiosk (Provision store) Proximity of HH to Major Roads Hypothesized Effect BPM Result - + + - - - - ++++ ++ + -- --- + ++ 1) Significant improvements in livelihood index: +++p<0.001, ++ p<0.01, + p<0.05, 2) Significant decline in livelihood index: ---p<0.001, --p<0.01, - p<0.05, 3) Insignificant improvements in livelihood index: + and 4) Insignificant decline in livelihood index: - The BPM results for community selection (table 6.2.3) suggests that the poorest communities may have been selected for the MTS. However, the results from household participation (table 6.2.5) following community selection does not provide enough evidence to suggest that the poorest households were effectively targeted. While it may possible that the forest policy targeted households with the highest likelihood of succeeding in the MTS, this selection criteria is just the first step to targeting the poorest households. Thus, the critical question for forest policy is whether it is possible to select households with the highest likelihood 331 of succeeding in the MTS that also happen to the poorest. To effectively target the poorest households within a community, it is important for forest policy to insist on a socio-economic survey that scores each households based on a set of resource/socioeconomic indicators. Following the survey and scoring, households that fall below a pre-determined threshold may then be prioritized for selection into the MTS. To give all resource-poor households within the predetermine threshold a fair chance of being included in the MTS, a randomized selection process may be implemented to select MTS households within the threshold. Household level socio-economic surveys of MTS communities allows for the creation of community maps such as those presented in Appendices A, B and C showing the spatial distribution of households within each community. Community maps showing the spatial distribution of select indicators can be used to create poverty hot-spots pinpointing priority areas within a community from which potential MTS households may be targeted for recruitment into the program. 332 Part III: Patterns of Local Spatial Autocorrelation that Influenced Household Placement into Ghana’s Modified Taungya System – Insight from Ayigbekrom Community Abstract This study investigated patterns of local spatial autocorrelation among households in Ayigbekrom a Yaya Forest Reserve community included in the study. A combination of Binomial Probit Model (BPM) and Exploratory Cluster and Outlier Analysis (Anselin’s Local Moran’s I) as well as Hot and Cold Spot Analysis maps of Ayigbekrom were used to describe the spatial distribution of four indexes within the community and how each index influenced MTS placement into households. The four indexes include: 1) Housing Infrastructure Index (HII), 2) Household Livestock Production Index (HLI), 3) Household Proximity Index (HPI), and 4) Household Male Index (HMI). A significant p-value (**** p<0.001) associated with the positive Global Moran’s I for HII suggests a highly significant probability of spatial clustering of like HII values among households in Ayigbekrom. The non-significant p-value associated with Global Moran’s I for HLI suggests that there was not enough evidence to reject the null hypothesis of spatial randomness of HLI within households in Ayigbekrom. Global Moran’s I for HPI was positive and highly significant (**** p<0.001) suggesting clustering of like HPI values within the community. The results also reveal two clusters of HPI Hot Spots (one on the North and the other on the South side) and two clusters of Cold Spots (one on the East and the other on the North-West side of the community). The Global Moran’s Index however did not reveal any clustering of like HMI values in Ayigbekrom. The results of the BPM suggest that a marginal increase in a household’s HII index significantly increased the predicted probability of MTS household selection (0.2878, p<0.001) and the expected marginal effect evaluated at the Mean X. All things equal, a marginal increase in HLI significantly increased the predicted probabilities of 333 MTS household selection (0.153, p<0.10) and the expected marginal effect evaluated at the Mean X. The BPM result for HMI is similar to HLI (0.140, p<0.10). 6.3 Data Generation and Analysis The 2009 household level surveys in Ayigbekrom covered four broad categories: 1) demographic information (number of male and female children as well as adult male and females in a household), 2) proximity of the households to communal physical capital assets, 3) physical condition of household dwellings (whether roof was constructed with corrugated metal sheets or covered with thatch/leaves or mud, whether household was constructed with cement blocks or mud etc.) and 4) household livestock ownership. The thus used the 2009 field data on household physical, human, and natural capital assets to generate four indexes and then investigated trends in spatial autocorrelation of each indexes within the community. For example, to determine proximity of households to communal physical assets, all 171 individual households in Ayigbekrom together with all other physical capital assets within the community were georeferenced and photographed using a handheld Garmin GPS unit and digital camera. A total of 611 still photographs of houses in the community were analyzed and the results used to construct the Household Infrastructure Index (HII). Choropleth maps, Hot and Cold Spot as well as Cluster and Outlier Analysis maps (Anselin’s Local Moran’s I) supported by a Global Moran’s I provided the basis for investigating spatial distribution of livelihood assets within Ayigbekrom. In order to create the choropleth maps and other exploratory analysis maps of the indexes, the “Proximity” function in ArcGIS’ ArcToobox was used to first convert the household point-shapefile for the community into Thiessen Polygons with each polygon representing a household. The polygon file was then “Joined” to the survey data and used to create a choropleth map showing the distribution of each index within the community. Exploratory Cluster and 334 Outlier Analysis (Anselin’s Local Moran’s I) as well as Hot and Cold Spot Analysis maps were then created and used to describe the spatial distribution of each index within the community. The exploratory analysis helped explain patterns of spatial distribution of household indexes within the community and how these patterns may have influenced placement of the national reforestation program into a household. In order to determine the nature of spatial clustering among the indexes the Global Moran’s I (i.e., spatial autocorrelation index) was generated. To determine the extent to which each of the five livelihood assets influence program selection, the study used a Binomial Probit Model to predict the probability of program placement associated with a marginal increase in each livelihood asset. 335 Figure 6.3.1: Spatial Distribution of Households Relative to Physical Assets in Ayigbekrom 336 Legend1a) MAJOR ROAD NETWORK82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEî4a) CHURCHö4b) MOSQUE5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5h) SCHOOL PLAY GROUND6a) COMMUNITY CENTERSB7c) BLACK SMITH SHOPKL7f) LOTTERY KIOSK8b) CHOP BAR/ RESTAURANT8c) CORN MILL (NIKA NIKA)10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHEDÞ12c) SEEDLING NURSERY#*13a) PROVISION KIOSK#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET15a) DUMPSTER×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYALegend1a) MAJOR ROAD NETWORK82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEî4a) CHURCHö4b) MOSQUE5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5h) SCHOOL PLAY GROUND6a) COMMUNITY CENTERSB7c) BLACK SMITH SHOPKL7f) LOTTERY KIOSK8b) CHOP BAR/ RESTAURANT8c) CORN MILL (NIKA NIKA)10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHEDÞ12c) SEEDLING NURSERY#*13a) PROVISION KIOSK#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET15a) DUMPSTER×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYA 6.3.1 Housing Infrastructure Index (HII) Four categories of household Physical Capital Assets including roof type, material used in constructing household physical structure, bathroom floor and availability of dedicated kitchen space within the household was used to construct HII. The information below describes in further detail the four key variables used in constructing HII. The HII was included in a Binomial Probit Model to determin its effect on a household’s probability of selection into the MTS program in Ayigbekrom community. 6.3.1.1 Type of Roof Prior to the 1990s, most households in rural communities in Ghana were roofed with thatch, leaves or bamboo. The national, regional, and local appeal of corrugated/aluminum roofing materials in Ghana specifically and in Africa in general may be due to its durability, resistance to insects/bugs and dry season fires as well as providing opportunities for households to harvest rainwater during the raining season. Because corrugated roof is considered an improvement in livelihood status, households constructed with corrugated roofs were awarded one point while those with traditional roofs such as thatch scored zero points. 6.3.1.2 Type of housing construction Like roof-type, prior to the 1990s the main structure of most houses in rural communities in Ghana were built with clay/mud bricks. Constructing building from cement blocks or concrete may be considered a is a relatively recent development in most rural communities and often associated with an improvement in livelihood status. Houses constructed from concrete/cement blocks earned one point while those with mud earned zero points. 337 6.3.1.3 Type of bathroom Due to the relatively high cost of cement across the country and a general lack of availability in rural communities, most resource poor households particularly in rural communities are unable to afford cement to construct their bathrooms. Hence most household bathroom floors in the homes of the resource poor are bear as shown in the picture on the right while those of the resource rich are cemented as shown in figure 6.3.2. Households with cemented bathroom floors earned one point while those without cement earned zero points. 6.3.1.4 Type of kitchen Some households in the research communities have a dedicated space or rooms within the household that served as a kitchen. The resource poor generally have no such space and often cook in open spaces or under trees. Figure 6.3.2 provides as an example of a dedicated space/room in a household with an overhead roof used as a kitchen and one that has not such facility. Households with a dedicated space thus earned one point while those with no such space (i.e. cook in the open) earned zero points. Figure 6.3.2: Physical Features used in Constructing HHI 338 Corrugated Roof (1)Cement Structure (1)Closed-SpaceKitchen (1)Open-Space Kitchen (0)Cemented Bathroom Floor(1)Bare-Floor Bathroom(0)Thatch Roof (0)Mud Structure (1) All responses from the housing infrastructure section of the survey were summed up to obtain a single HII value for each household. Thus, households that answered YES to all four housing infrastructure questions scored four out of four points on the HII while another that answered NO has a zero HII (table 6.3.1). The HII index was included in a Binomial Probit Model (BMP) to determine its effect on the probability of household selection into the MTS program. Table 6.3.1: Housing Infrastructure Index Housing Infrastructure Survey Responses Variable/Weight definition a. Corrugated roof Yes= 1, No=0 b. Concrete housing construction Yes= 1, No=0 c. Cemented bathroom floor Yes= 1, No=0 All yes responses to the housing infrastructure question were weighted=1 while Nos were weighted=0 d. Dedicated kitchen Yes= 1, No=0 6.3.2 Household Livestock Production Index (HLI) To generate information for the HLI, individuals in all the households in Ayigbekrom were asked if they owned any of the following livestock: poultry (chicken/fowls), goat or sheep, pigs, and cattle. The questions only focused on ownership of livestock and not the quantity owned. Different weights were assigned to each livestock category based on current market prices as well as the relative cost of raising each of the livestock category. Hence chicken which cost the least to raise (in terms of market value and husbandry) was assigned the least weight of one, followed by goats which were assigned a weight of two. Responses from households were summed up to obtain a single HLI value for each household. A household raising only chicken/fowls thus scored one point on the HLI while another raising all four livestock (i.e. 339 chicken, goats, pigs and cows) scored 10 HLI points. The Livestock Index was included in a Binomial Probit Model (BMP) to determine its effect on the probability of household selection into the MTS program in Ayigbekrom. Figure 6.3.2 provides pictorial descriptions while table 6.3.2 describe the HLI variables and their associated weights. Figure 6.3.3: Livestock Categories used in Constructing HLI 340 Chicken (1)Goats/Sheep (2)Pigs(3)Cattle(4) Table 6.3.2: Household Variable Definitions Household livestock ownership Survey Responses a. Household with chicken (poultry) Yes= 1, No=0 b. Household rears Goats and Sheep Yes= 2, No=0 c. Household rears Pigs Yes= 3, No=0 d. Household raises cattle Yes= 4, No=0 6.3.3 Household Proximity Index (HPI) Variable/Weight definition Variables were assigned weights based on the relative expense in husbandry as well as market value of the unit of livestock in question. Hence the unit of chicken being least expensive to raise and was assigned a weight=1 while cattle being the most expensive was assigned a weight=4 The HPI was generated by assigning weights to households based on their relative distances from four essential communal physical assets. In most rural communities in Ghana, access to portable water is usually provided in the form of borehole fitted with manual hand or foot pumps to draw the water from natural underground wells. Lack of tap water to most rural communities and the relatively high cost of bottled water makes boreholes a vital livelihood asset. Major roads were also included in the list of essential communal assets since these connect households to other essential facilities such as markets, hospitals, schools and district administration offices outside of the community. The fourth asset included in HPI is the Corn mill (locally called Nika Nika). Corn is staple crop in Ghana and constitutes an essential component of most Ghanaian diet as it is used in preparing breakfast porridge and other meals consumed in nearly every household. Nearly all corn-based food products in Ghana are prepared from milled corn thus making the corn mill a vital communal asset. The fourth physical asset included in the study is communal toilet facilities. Since most household in rural communities in Ghana do not have private toilet facilities within the home, communal toilet facilities occupy a 341 vital space in rural communities. To assign weights to each of the four communal assets, 60, 120 and180 meter buffer zones are created around each of the four assets and weights assigned to households based on which buffer zone they are situated relative to a particular physical asset. Hence household within a 60-meter buffer zone of a communal asset scored 4 points while those between 61 and 120 meters scored 3 point. Similarly, households between 121 and 180 meters scored 2 points while those further than 180 meters scored only one point. A final HPI was generated for each household by multiplying the scores earned for each physical asset by the buffer zone then summing that up then divide by four to obtain an average (i.e. HPI) for the household. Figure 6.3.4 below provides a pictorial description of the four assets and their respective buffer zones. 342 Figure 6.3.4: Communal Assets used in Constructing HPI Major Road Borehole Public Toilet Corn Mill 343 6.3.4 Household Male Index (HMI) The fourth index is the Household Male Index (HMI) which simply tallied the total number of all adult males within each household. In most rural communities in Ghana most males assume major household responsibilities including working in crop fields and tending to livestock by the time they turn 12 year. This study thus categorized all males 12 years and above as adults. Hence the total number of adult males in a household represents the household’s HMI. 6.3.5 Results This section is divided into four parts with each part representing each of the Household Livelihood Indexes described above. The frequency distribution for each index across MTS and non-MTS participation households in Ayigbekrom is first presented following which the spatial distribution of each indexes is described using a choropleth map. The Cluster and Outlier Analysis Maps, Hot and Cold Spot Analysis Maps as well as the Global Moran's Index are also presented in this section. 6.3.5.1 Spatiotemporal Analysis of HII The frequency distribution for HII presented in table 6.3.2 below suggests that a disproportionately larger percentage of MTS households (19%) had houses roofed with corrugated roofing sheets, walls built with concrete/cement blocks, had cemented bathroom floors as well as dedicated space in the house for cooking (represented by the maximum HII score of 4). In comparison, only 3% of Non-MTS households in the community scored four- points on the HMI index. Similarly, 6% of MTS member households within the community scored zero on HMI compared to 36% of non-MTS households scored a zero on HII. Table 6.3.3 344 and figure 6.3.5 presents the frequency distribution of HII index across both MTS and non-MTS households in Ayigbekrom. Table 6.3.3: Frequency Distribution of HII Housing Infrastructure Index MTS Membership 0 1 2 3 4 Total 46 (36) 6 (14) 52 (30) 32 (25) 8 (19) 40 (23) 19 (15) 9 (21) 28 (16) 28 (22) 12 (28) 40 (23) 3 (3) 8 (19) 11 (6) 128 (100) 43 (100) 171 (100) No Yes Total Note: Percentages in parenthesis While some clustering of like HII values was observed in the choropleth map in figure 6.3.5 below, clustering was further confirmed by the results of the Global Moran’s I for Spatial Autocorrelation (see figure 6.3.5). The highly significant p-value associated with a positive Global Moran’s I (i.e. Spatial Autocorrelation Coefficient) suggests a highly significant probability (**** p<0.001) of spatial clustering of like HII values among households in Ayigbekrom. To further understand the distribution patterns of HII in the community, a Cluster and Outlier Analysis map (i.e. Anselin’s Local Moran I) together with a Hot and Cold Spot Analysis maps were generated. The Hot and Cold Spot Analysis map below confirms (at a 99% confidence level) a high degree of spatial clustering of High HII values (Hot Spots). As shown in figure 6.3.5, high HII values are clustered in Hot Spots on the Eastern portion of the community away from the Yaya forest reserve (also see figures 6.3.4). The Hot and Cold Spot Analysis map also suggest (at a 95% confidence level) clustering of low HII values (Cold Spots) on the 345 southern and western tips of the community the majority of which are situated well within the Yaya forest reserve. As shown below three out of the 43 MTS households are located within the Cold Spots providing some clues into how HII may have influenced MTS program placement into households in Ayigbekrom. 346 Figure 6.3.5: HII Distribution and Cluster Analysis Maps Global Moran's Index: 0.227990 P-value: 0.000000 Positive and significant Moran’s I suggest spatial clustering of like values. 347 6.3.5.2 Spatiotemporal Analysis of HLI The choice of teak trees for Ghana’s MTS dictates that farmers would have between 1-3 years to intercrop the tree plantations before canopy closure and shading renders any agricultural production uneconomical. With teak farmers are expected to wait anywhere between 8 to 25 years before deriving any financial benefits from harvested trees. The current rotation periods for teak are clearly beyond most rural farmers’ economic horizon hence the need for an alternative source of livelihood during the long waiting period. Goat, pig/hog and grasscutter husbandry was introduced as alternative livelihood options for MTS project participants. It was thus expected that the HLI among MTS would be relatively higher than non-member households given that most MTS households might have already received some livestock from the program. The frequency distribution of HLI in Ayigbekrom as shown in table 6.3.4 below confirms a relatively higher HLI index among MTS households. According to table 6.3.4, 37% of MTS households scored three-points on the HLI compared to 26% for the non-MTS group. On the contrary, 37% of the non-MTS had no livestock compared to 19% of for MTS households. Table 6.3.4: Frequency Distribution of HLI MTS Membership No Yes Total 0 47 (37) 8 (19) 55 (32) 1 41 (32) 15 (35) 56 (33) Note: Percentages in parenthesis Household Livestock Index 3 33 (26) 16 (37) 40 (29) 5 1 (1) 0 (0) 1 (1) 6 Total 1 (1) 0 (0) 1 (1) 128 (100) 43 (100) 171 (100) 2 5 (4) 4 (9) 49 (5) 348 Figure 6.3.6 below describes the spatial distribution patterns of HLI in Ayigbekrom. The stripped polygons represent MTS households while the non-stripped represents non-MTS member group. The non-significant p-value associated with the Global Moran’s I suggests that not enough evidence exists to reject the null hypothesis of spatial randomness. What the Global Moran’s I suggest essentially is that HLI among households in Ayigbekrom may be randomly distributed without any significant clustering of like values. The results of the Global Moran’s I appears to be supported by the Cluster and Outlier Analysis (Anselin’s Local Moran’s I) as well as the Hot and Cold Spot Analysis maps in figure 6.3.6 below. It is also possible from these results that HLI does not exert a strong influence on MTS household selection in Ayigbekrom. 349 Figure 6.3.6: HLI Distribution and Cluster Analysis Maps Global Moran's Index: -0.000671 P-value: 0.854275 Not sufficient evidence to reject the Null Hypothesis of spatial randomness. 350 6.3.5.3 Spatiotemporal Analysis of HPI The results presented in table 6.3.5 suggests that 23% of MTS participants are located relatively closer (75-95 meters) to essential physical capital assets compare to the non-MTS group (16%). On the contrary, a larger percentage of the non-MTS group (13%) are located between 196 and 240 meters compared to only 5% of the MTS households. Table 6.3.5: Frequency Distribution of HPI (Meters) Household Proximity Index (HPI) MTS Membership No Yes Total 75-90 91-120 121-150 151-195 196-240 Total 20 (16) 10 (23) 30 (18) 25 (20) 8 (19) 33 (19) 31 (24) 8 (19) 39 (23) 35 (27) 15 (35) 50 (29) 17 (13) 2 (5) 19 (11) 128 (100) 43 (100) 171 (100) Note: Percentages in parenthesis Table 6.3.5 clearly suggests that MTS households had a higher representation within the higher end of the HPI than their non-MTS counterparts and this finding appears to be supported by the choropleth map in figure 6.3.7 below. The Global Moran’s I for HPI is positive and highly significant suggesting clustering of like values. Further investigation into HPI clustering show two clusters of Hot Spots (one on the North and the other on the South side) and two clusters of Cold Spots (one on the East and the other on the North-West side of Ayigbekrom). Anselin’s Local Moran’s I confirm the clustering of high HPI values in Hot Spots and those of low HPI values in the Cold Spots. While the results in figure 3f suggests that HPI may likely be an important variable in determining MTS household selection, the fact that we also have a 351 relatively large clustering of non-MTS household within these Hot and Cold Spots suggests caution in making assumptions regarding the influence of HPI on MTS household selection. 352 Figure 6.3.7: HPI Distribution and Cluster Analysis Maps Global Moran's Index: 0.650863 P-value: 0.000000 Positive and significant Moran’s I suggest spatial clustering of like values. 353 6.3.5.4 Spatiotemporal Analysis of HMI The frequency distribution of households in HMI presented in table 6.3.6 below suggest that majority of households in Ayigbekrom have only one or two adult males in the household. As show below, about 80% of non-MTS households have between one and two adults compared to 65% for the MTS group. In terms of higher values of HMI, 33% of MTS households have more three or more adult males in the household compared to 17% for their non-MTS counterparts. The choropleth map in figure 6.3.8 below describes the spatial distribution of HMI within Ayigbekrom. Like the HLI discussed under section 4.2, the Global Moran’s Index for HMI is insignificant, suggesting there is not sufficient evidence to reject the null hypothesis of spatial randomness. Again, the Global Moran’s I suggest a lack of clustering of like HMI values among household in Ayigbekrom. Hence while we observe a Hot Spot on the eastern portion of the community with a few MTS households clustered within this spot, this level of clustering according to the Moran’s I does not rise to the level of being significant. Table 6.3.6: Frequency Distribution of HMI Household Male Index (HMI) 4 7 (5) 3 (7) 10 (6) 5 2 (2) 0 (0) 2 (1) 6 14 Total 1 (1) 2 (5) 3 (2) 0 (0) 1 (2) 1 (1) 128 (100) 43 (100) 171 (100) 1 2 36 (28) 10 (23) 46 (27) 66 (52) 18 (42) 84 (49) 3 11 (9) 8 (19) 19 (11) 354 MTS Membership No Yes Total 0 5 (4) 1 (2) 6 (4) Note: Percentages in parenthesis Figure 6.3.8: HMI Distribution and Cluster Analysis Maps Global Moran's Index: 0.030810 P-value: 0.165161 Not sufficient evidence to reject the Null Hypothesis of spatial randomness. 355 6.3.5.5 Binomial Probit Model (BPM) This section uses a BPM to determine the extent to which a marginal change in a household livelihood index influences the predicted probability of MTS household selection/placement and the expected marginal effect evaluated at the Mean X (i.e. the mean of the index). Table 6.3.7 describes the explanatory variables used in creating the four indexes in generating the BPM. 356 Table 6.3.7: Variables included in BPM Explanatory variables used in BPM Variable definition 1a. Type of roof 1b. Type of housing construction 1c. Type of bathroom 1d. Type of kitchen Type of roof over housing unit (1=Corrugated, 0=Raffia Leaves/Thatch) Physical Construction of housing unit (1=Cement/Concrete, 0=Earth/Mud) Type of Bathroom (1=Cemented floor, 0=Mud/gravel) Type of Kitchen (1=Secure roofed space in HH for Kitchen, 0=No kitchen/household cooks under tree or in an open space on compound) 2a. Household raises chicken (poultry) Yes= 1, No=0 2b. Household rears Goats and Sheep Yes= 2, No=0 2c. Household rears Pigs Yes= 3, No=0 2d. Household raises cattle Yes= 4, No=0 3a. Proximity of HH to Corn Mill (Nika Nika) 3b. Proximity of HH to Borehole (Portable water) 3c. Proximity of HH to Communal Toilet 3d. Proximity of HH to Major Roads 4. Number of adult males in the household Relative distance of household or housing unit from various communal physical assets. 0=Non-Available 1=>180meters 2=121-180meters 3=61-120meters 4=0-60meters Total number of males in the household older than 12yrs old The dependent variable (MTS Participation) is codded such that MTS households were assigned a weight of 1 for participation while non-MTS households were assigned a weight of 0. The dichotomous/binary nature of the dependent variable (i.e. MTS participant household =1 vs. non-MTS participant household = 0) suggests the likelihood for the underlying properties of 357 Ordinary Least Square (OLS) methods (e.g. Linear Probability Model) to be violated when used in modeling MTS household selection. Specifically, for certain combinations of explanatory variables, it is likely for the binary dependent variable to take on values outside of the 0 -1 range thus rendering the OLS incapable of producing the best linear unbiased estimator (BLUE) for program selection. Table 6.3.8 represents the four livelihood indexes used in Table 6.3.8: Livelihood Indexes used in BPM Explanatory variables used in BPM Variable definition 1. Household Infrastructure Index (HII) Measures roof type, Physical Construction of housing, Type of Bathroom and Type of Kitchen. 2. Household Livestock Index (HLI) Measure household ownership/production of chicken, goats/sheep, pigs or cattle. 3. Household Proximity Index (HPI) Measures the relative distance of household from a borehole, major road, corn mill and/or a communal toilet. 4. Household Male Index (HMI) Measures the total number of males older than 12yrs of age in the household. Before discussing the BPM results it is important to first discuss the hypothesized effects of each of the four livelihood indexes. Understanding the signs associated with each livelihood index in the BPM and what that means in terms of being validated or contradicted by the BPM would help in understanding the importance of an index in influencing program placement. At best the MTS program should disproportionately include the resource poor and at worse it should be seen as recruiting households randomly such that all households in a community have an 358 equal opportunity of being selected. Table 6.3.9 below describes the rationale behind the hypothesized effects of each index. Table 6.3.9: Hypothesized Effects of Livelihood Indexes Explanatory variables used Hypothesized Effect on Program in BPM Placement 1. HII - 2. HLI + 3. HPI - 4. HMI + Explanation of Hypothesized Effects on MTS Program Placement Because of the assumption that resource rich households in Ayigbekrom would have relatively higher HII (corrugated roof, concrete walls, cemented bathroom floors and dedicated kitchen space), it was hypothesized that a marginal increase in HII will significantly decreases the predicted probability of household selection thus directing the program disproportionately into relatively poorer households. Because the success of the MTS livestock program depends on the program’s ability to rapidly multiplier the start-up livestock so that it can be distributed widely to other program members, prior ownership of livestock and thus high HLI was considered a plus for the program hence hypothesized to significantly increase the predicted probability of selection. Because most resource rich households in rural communities generally tend to be clustered around major road networks, it was hypothesized that a marginal increase in HPI should significantly decrease the predicted probability of household selection thus directing the program towards relatively poorer households. In most Taungya communities, men are expected to provide the muscle power needed for logging as well as policing the plantations hence it was hypothesized that an increase in the number of adult males in a household significantly increases the predicted probability of MTS placement within that household. 359 Household Infrastructure Index (HII): According the BPM result in table 8, all things equal, a marginal increase in a household’s HII index significantly increased the predicted probability of MTS household selection (0.2878, p<0.001) and the expected marginal effect evaluated at the Mean X. This result suggests that contrary to the hypothesized effect of HII on household selection, the MTS program disproportionately favored households with relatively higher HII (i.e., resource rich households). Household Livestock Index (HLI): All things equal, a marginal increase in HLI significantly increases the predicted probability of MTS household selection (0.153, p<0.10) and the expected marginal effect evaluated at the Mean X. This BPM result validates the assumptions made in table 6.3.7c regarding the need to rapidly multiply livestock herds for future distribution to program members. Household Proximity Index (HPI): Though the negative sign associated with the HPI in the BPM is as predicted in table 6.3.7c, this index did not have any significant impact on MTS household selection. The BPM result suggests that HPI may not be an important predicted of household selection into the MTS program in Ayigbekrom. Household Male Index (HMI): All things equal, a marginal increase in HMI significantly increases the predicted probability of MTS household selection (0.140, p<0.10) and the expected marginal effect evaluated at the Mean X. Once again, the BPM result validates the assumption made in table 6.3.8 regarding the perceived importance of adult males in guaranteeing success of the MTS program. As shown in table 6.3.8, the sign on the HLI coefficient is positive and significant as hypothesized. 360 The R2 value of 10.99 in the BPM suggests that only about 11% of the variations in the model are explained suggesting that this model may be a relatively weak predictor of factors that determine MTS household selection. Additionally, the BPM recorded approximately a 77% accurate prediction rate for MTS household selection and non-selection in Ayigbekrom. Table 6.3.10 below presents the BPM results for all 171 households in Ayigbekrom. Table 6.3.10: Effects of Livelihood Indexes on Household Selection MTS Household Selection Hypothesized Effect on Participation 1. Household Infrastructure Index (HII) 2. Household Livestock Index (HLI) 3. Household Proximity Index (HPI) - + - Impact of EV on MTS Marginal Effects Community Participation 0.2878**** after Probit (dy/dx) 0.087**** 0.086 0.153* 0.086 -0.001 0.026 0.046* 0.026 -0.000 Mean X 1.52 1.357 145.0 08 + 0.002 0.140* 0.083 -0.001 0.043* 0.025 4. Household Males (HMI) 5. Cons Observations R-Squared (%) Percent correctly predicted (%) 1) + Positively associated, - Negatively associated with Household Selection; 2) dy/dx is for discrete change of dummy variable from non-MTS Selection (0) to Selection (1); 3.) HH=household; 4) Significant levels **** p<0.001, *** p<0.01, **p<0.05, *p<0.10 with robust standard errors in parentheses; 5) X= Mean; 6) Wald test of rho=0: chi2 (4) =21.20, Prob > chi2 = 0.000 171 10.99 76.61 -1.527*** 0.477 1.84 6.3.6 Summary, conclusions, and Policy Implications The BPM results suggest that selection of Ghana’s MTS project households in Ayigbekrom was driven largely by factors perceived to significantly increase the likelihood of program success and not necessarily by the livelihood status of the poorest households. This conclusion is based on marginal increases in HII, HLI and HMI which significantly increased the predicted probability of selection of MTS households in the community. From a policy 361 standpoint, household selection into the MTS should disproportionately target the poorest of the poor households. Not deliberately targeting resource poor households (which in most cases are already blamed for most of the forest degradation) means that the problem of forest loss in the Yaya may likely persist if not worsen. In the absence of an equitable and just selection process, it is likely that resource poor in Ayigbekrom and other similar communities would continue to seek sustenance from national forest reserves in ways that may be deemed social or economically unacceptable by the Forest Services Department. Long-term success of reforestation programs including the MTS thus require thoughtful program placement strategies that considers factors that currently affects the livelihood status of potential participants and the likelihood of participants deriving the outmost from program interventions. 362 APPENDICES 363 APPENDIX A: Yaya Forest Reserve Community Maps Yaya Reserve Community Maps Figure 6.3.9: Yaya Reserve Map Showing the Location of Abrefakrom 364 Yaya Reserve Community Legend Figure 6.3.9 (cont’d) 365 Legend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEKJ3b) RIVER/STREAM (COMMUNITY WATER SOURCE)UT3c) COMMUNITY WATER WELLî4a) CHURCHö4b) MOSQUE5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5c) JUNIOR HIGH SCHOOL (JHS)5d) SCHOOL/CLASSROOM UNDER TREE¹½5e) ABANDONED SCH STRUCTUREÆV5f) SCHOOL FEEDING PROGRAM KITCHEN²³5g) TEACHERS' BUNGALLOW5h) SCHOOL PLAY GROUND6a) COMMUNITY CENTERß6b) COMMUNITY/TOWN BELL7a) BUSINESS AND TELECOMMUNICATION CENTERPB7b) PRIVATE BUSINESS ENTSB7c) BLACK SMITH SHOPTS7d) TAILORING SHOP7e) SHOE REPAIR SHOPKL7f) LOTTERY KIOSK)7g) CAPENTRY SHOP½7h) BICYCLE REPAIR SHOPô7i) HAIR SALON AND BABER SHOP8a) FOOD PROCESSING CENTER (GARI)8b) CHOP BAR/ RESTAURANT8c) CORN MILL (NIKA NIKA)AD8d) ALOCOHOL DISTILERY"u9a) COMMUNBITY CLINIC9b) DRUG STORE10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHED"M11c) PAFORM DEMOSÒ11d) WOMENS MUSHROOM PROGRAMkj12a) BACKYARD FARMSõôó12b) TEAK PLANTATIONSÞ12c) SEEDLING NURSERY#*13a) PROVISION KIOSK#*13b) CLUSTER OF KIOSKS/PROVISION SHOPS#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET$+14b) MARKET ADMINISTRATION BUILDING15a) DUMPSTER×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESþ17a) PETROL STATIONForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYALegend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEKJ3b) RIVER/STREAM (COMMUNITY WATER SOURCE)UT3c) COMMUNITY WATER WELLî4a) CHURCHö4b) MOSQUE5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5c) JUNIOR HIGH SCHOOL (JHS)5d) SCHOOL/CLASSROOM UNDER TREE¹½5e) ABANDONED SCH STRUCTUREÆV5f) SCHOOL FEEDING PROGRAM KITCHEN²³5g) TEACHERS' BUNGALLOW5h) SCHOOL PLAY GROUND6a) COMMUNITY CENTERß6b) COMMUNITY/TOWN BELL7a) BUSINESS AND TELECOMMUNICATION CENTERPB7b) PRIVATE BUSINESS ENTSB7c) BLACK SMITH SHOPTS7d) TAILORING SHOP7e) SHOE REPAIR SHOPKL7f) LOTTERY KIOSK)7g) CAPENTRY SHOP½7h) BICYCLE REPAIR SHOPô7i) HAIR SALON AND BABER SHOP8a) FOOD PROCESSING CENTER (GARI)8b) CHOP BAR/ RESTAURANT8c) CORN MILL (NIKA NIKA)AD8d) ALOCOHOL DISTILERY"u9a) COMMUNBITY CLINIC9b) DRUG STORE10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHED"M11c) PAFORM DEMOSÒ11d) WOMENS MUSHROOM PROGRAMkj12a) BACKYARD FARMSõôó12b) TEAK PLANTATIONSÞ12c) SEEDLING NURSERY#*13a) PROVISION KIOSK#*13b) CLUSTER OF KIOSKS/PROVISION SHOPS#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET$+14b) MARKET ADMINISTRATION BUILDING15a) DUMPSTER×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESþ17a) PETROL STATIONForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYA Abrefakrom Legend Figure 6.3.9 (cont’d) 366 Legend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEö4b) MOSQUE5b) PRIMARY SCHOOL5c) JUNIOR HIGH SCHOOL (JHS)½7h) BICYCLE REPAIR SHOP8c) CORN MILL (NIKA NIKA)10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHEDkj12a) BACKYARD FARMS#*13a) PROVISION KIOSK#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET$+14b) MARKET ADMINISTRATION BUILDING15a) DUMPSTER16b) PRIVATE DUGOUT LATERINESþ17a) PETROL STATIONForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYA Figure 6.3.10: Abrefakrom Community Map 367 Figure 6.3.11: Abrefakrom Community – Buffer on Public Toilets 368 Figure 6.3.12: Abrefakrom Community – Buffer on Primary Schools 369 Figure 6.3.13: Abrefakrom Community – Buffer on Corn Mill 370 Figure 6.3.14: Abrefakrom Community – Buffer on Mosque 371 Figure 6.3.15: Abrefakrom Community – Buffer on Weekly Market 372 Figure 6.3.16: Abrefakrom Community – Buffer on Major Roads 373 Figure 6.3.17: Abrefakrom Community – Buffer on Provision Kiosks 374 Figure 6.3.18: Abrefakrom Community – Buffer on Junior Secondary School 375 Figure 6.3.19: Abrefakrom Community – Buffer on Dumpster 376 Figure 6.3.20: Abrefakrom Community – Buffer on Borehole 377 Ahyiem Community Figure 6.3.21: Yaya Reserve Map Showing the Location of Ahyiem 378 Ahyiem Legend Figure 6.3.21 (cont’d) 379 Legend1a) MAJOR ROAD NETWORK82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE3a) BOREHOLEö4b) MOSQUE5b) PRIMARY SCHOOL5d) SCHOOL/CLASSROOM UNDER TREE²³5g) TEACHERS' BUNGALLOW5h) SCHOOL PLAY GROUND½7h) BICYCLE REPAIR SHOP8c) CORN MILL (NIKA NIKA)10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHED"M11c) PAFORM/MTS DEMOSkj12a) BACKYARD FARMS#*13a) PROVISION KIOSK#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET16b) PRIVATE DUGOUT LATERINESþ17a) PETROL STATIONForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYA Figure 6.3.22: Ahyiem Community 380 Figure 6.3.23: Ahyiem Community – Buffer on Borehole 381 Figure 6.3.24: Ahyiem Community – Buffer on Market 382 Figure 6.3.25: Ahyiem Community – Buffer on Road 383 Figure 6.3.26: Ahyiem Community – Buffer on Mosque 384 Figure 6.3.27: Ahyiem Community- Buffer on Corn Mill 385 Figure 6.3.28: Ahyiem Community – Buffer on Nursery and Kindergarten 386 Figure 6.3.29: Ahyiem Community – Buffer on Primary School 387 Figure 6.3.30: Ahyiem Community – Buffer on Provision Kiosk 388 Figure 6.3.31: Ahyiem Community – Buffer on Public Toilet 389 Amangoase Community Figure 6.3.32: Yaya Reserve Map Showing the Location of Amangoase 390 Amangoase Legend Figure 6.3.32 (cont’d) 391 Legend1a) MAJOR ROAD NETWORK82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEî4a) CHURCHö4b) MOSQUE5b) PRIMARY SCHOOL5h) SCHOOL PLAY GROUND6a) COMMUNITY CENTER)7g) CAPENTRY SHOP½7h) BICYCLE REPAIR SHOPô7i) HAIR SALON AND BABER SHOP8a) FOOD PROCESSING CENTER (GARI)8c) CORN MILL (NIKA NIKA)"u9a) COMMUNBITY CLINIC9b) DRUG STORE10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHEDkj12a) BACKYARD FARMSõôó12b) TEAK PLANTATIONSÞ12c) SEEDLING NURSERY#*13a) PROVISION KIOSK15a) DUMPSTER×16a) PUBLIC TOILETþ17a) PETROL STATIONForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYALegend1a) MAJOR ROAD NETWORK82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEî4a) CHURCHö4b) MOSQUE5b) PRIMARY SCHOOL5h) SCHOOL PLAY GROUND6a) COMMUNITY CENTER)7g) CAPENTRY SHOP½7h) BICYCLE REPAIR SHOPô7i) HAIR SALON AND BABER SHOP8a) FOOD PROCESSING CENTER (GARI)8c) CORN MILL (NIKA NIKA)"u9a) COMMUNBITY CLINIC9b) DRUG STORE10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHEDkj12a) BACKYARD FARMSõôó12b) TEAK PLANTATIONSÞ12c) SEEDLING NURSERY#*13a) PROVISION KIOSK15a) DUMPSTER×16a) PUBLIC TOILETþ17a) PETROL STATIONForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYA Figure 6.3.33: Amangoase Community 392 Figure 6.3.34: Amangoase Community – Buffer on Borehole 393 Figure 6.3.35: Amangoase Community – Buffer on Church 394 Figure 6.3.36: Amangoase Community – Buffer on Clinic 395 Figure 6.3.37: Amangoase Community – Buffer on Public Dumpster 396 Figure 6.3.38: Amangoase Community – Buffer on Provision Kiosks 397 Figure 6.3.39: Amangoase Community – Households with Livestock 398 Figure 6.3.40: Amangoase Community – Buffer on Corn Mill 399 Figure 6.3.41: Amangoase Community – Buffer on Primary School 400 Figure 6.3.42: Amangoase Community – Buffer on Major Road 401 Figure 6.3.43: Amangoase Community – Buffer on Public Toilet 402 Amoahkrom Community Figure 6.3.44: Yaya Reserve Map Showing the Location of Amoahkrom 403 Amoahkrom Legend Figure 6.3.44 (cont’d) 404 Legend1a) MAJOR ROAD NETWORK82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLE5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5h) SCHOOL PLAY GROUND8c) CORN MILL (NIKA NIKA)10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHEDÞ12c) SEEDLING NURSERY#*13a) PROVISION KIOSK14a) COMMUNITY MARKET×16a) PUBLIC TOILETForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYA Figure 6.3.45: Amoahkrom Community 405 Figure 6.3.46: Amoahkrom Community – Buffer on Borehole 406 Figure 6.3.47: Amoahkrom Community – Buffer on Preschool and Kindergarten 407 Figure 6.3.48: Amoahkrom Community – Buffer on Provision Kiosks 408 Figure 6.3.49: Amoahkrom Community – Households with Livestock 409 Figure 6.3.50: Amoahkrom Community – Buffer on Market 410 Figure 6.3.51: Amoahkrom Community – Buffer on MTS Livestock Program 411 Figure 6.3.52: Amoahkrom Community – Buffer on Corn Mill 412 Figure 6.3.53: Amoahkrom Community – Buffer on Primary School 413 Figure 6.3.54: Amoahkrom Community – Buffer on Major Road 414 Figure 6.3.55: Amoahkrom Community – Buffer on Public Toilet 415 Asuakwa Community and Legends Figure 6.3.56: Yaya Reserve Map Showing the Location of Asuakwa 416 Asuakwa Legend Figure 6.3.56 (cont’d) 417 Legend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION3a) BOREHOLEî4a) CHURCH5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5c) JUNIOR HIGH SCHOOL (JHS)ÆV5f) SCHOOL FEEDING PROGRAM KITCHEN²³5g) TEACHERS' BUNGALLOW5h) SCHOOL PLAY GROUNDß6b) COMMUNITY/TOWN BELLTS7d) TAILORING SHOP7e) SHOE REPAIR SHOPô7i) HAIR SALON AND BABER SHOP8c) CORN MILL (NIKA NIKA)10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHEDõôó12b) TEAK PLANTATIONS#*13a) PROVISION KIOSK#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYALegend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION3a) BOREHOLEî4a) CHURCH5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5c) JUNIOR HIGH SCHOOL (JHS)ÆV5f) SCHOOL FEEDING PROGRAM KITCHEN²³5g) TEACHERS' BUNGALLOW5h) SCHOOL PLAY GROUNDß6b) COMMUNITY/TOWN BELLTS7d) TAILORING SHOP7e) SHOE REPAIR SHOPô7i) HAIR SALON AND BABER SHOP8c) CORN MILL (NIKA NIKA)10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHEDõôó12b) TEAK PLANTATIONS#*13a) PROVISION KIOSK#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYA Figure 6.3.57: Asuakwa Community 418 Figure 6.3.58: Asuakwa Community – Buffer on Borehole 419 Figure 6.3.59: Asuakwa Community – Buffer on Church 420 Figure 6.3.60: Asuakwa Community – Buffer on Market 421 Figure 6.3.61: Asuakwa Community – Buffer on Junior Secondary School 422 Figure 6.3.62: Asuakwa Community – Buffer on Major Road 423 Figure 6.3.63: Asuakwa Community – Buffer on Corn Mill 424 Figure 6.3.64: Asuakwa Community – Buffer on Preschool and Kindergarten 425 Figure 6.3.65: Asuakwa Community – Buffer on Primary School 426 Figure 6.3.66: Asuakwa Community – Provision Kiosks 427 Figure 6.3.67: Asuakwa Community – Buffer on Public Toilet 428 Ayigbekrom Community Figure 6.3.68: Yaya Reserve Map Showing the Location of Ayigbekrom 429 Ayigbekrom Legend Figure 6.3.68 (cont’d) 430 Legend1a) MAJOR ROAD NETWORK82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEî4a) CHURCHö4b) MOSQUE5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5h) SCHOOL PLAY GROUND6a) COMMUNITY CENTERSB7c) BLACK SMITH SHOPKL7f) LOTTERY KIOSK8b) CHOP BAR/ RESTAURANT8c) CORN MILL (NIKA NIKA)10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHEDÞ12c) SEEDLING NURSERY#*13a) PROVISION KIOSK#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET15a) DUMPSTER×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYALegend1a) MAJOR ROAD NETWORK82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEî4a) CHURCHö4b) MOSQUE5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5h) SCHOOL PLAY GROUND6a) COMMUNITY CENTERSB7c) BLACK SMITH SHOPKL7f) LOTTERY KIOSK8b) CHOP BAR/ RESTAURANT8c) CORN MILL (NIKA NIKA)10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHEDÞ12c) SEEDLING NURSERY#*13a) PROVISION KIOSK#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET15a) DUMPSTER×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYA Figure 6.3.69: Ayigbekrom Community 431 Figure 6.3.70: Ayigbekrom Community – Buffer on Boreholes 432 Figure 6.3.71: Ayigbekrom Community – Buffer on MTS Livestock Program 433 Figure 6.3.72: Ayigbekrom Community – Buffer on Church 434 Figure 6.3.73: Ayigbekrom Community – Buffer on Dumpster 435 Figure 6.3.74: Ayigbekrom Community – Buffer on Preschool and Kindergarten 436 Figure 6.3.75: Ayigbekrom Community – Buffer on Provision Kiosk 437 Figure 6.3.76: Ayigbekrom Community – Livestock Ownership 438 Figure 6.3.77: Ayigbekrom Community – Buffer on Mosque 439 Figure 6.3.78: Ayigbekrom Community – Buffer on Corn Mill 440 Figure 6.3.79: Ayigbekrom Community – Buffer on Major Road 441 Figure 6.3.80: Ayigbekrom Community – Buffer on Public Toilet 442 Buoku Community Figure 6.3.81: Yaya Reserve Map Showing the Location of Buoku Community 443 Buoku Legend Figure 6.3.81 (cont’d) 444 Legend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEî4a) CHURCHö4b) MOSQUE5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5c) JUNIOR HIGH SCHOOL (JHS)²³5g) TEACHERS' BUNGALLOW5h) SCHOOL PLAY GROUND7a) BUSINESS AND TELECOMMUNICATION CENTERPB7b) PRIVATE BUSINESS ENTTS7d) TAILORING SHOPKL7f) LOTTERY KIOSKô7i) HAIR SALON AND BABER SHOP8b) CHOP BAR/ RESTAURANT8c) CORN MILL (NIKA NIKA)10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDSõôó12b) TEAK PLANTATIONS#*13a) PROVISION KIOSK#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET15a) DUMPSTER×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYALegend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEî4a) CHURCHö4b) MOSQUE5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5c) JUNIOR HIGH SCHOOL (JHS)²³5g) TEACHERS' BUNGALLOW5h) SCHOOL PLAY GROUND7a) BUSINESS AND TELECOMMUNICATION CENTERPB7b) PRIVATE BUSINESS ENTTS7d) TAILORING SHOPKL7f) LOTTERY KIOSKô7i) HAIR SALON AND BABER SHOP8b) CHOP BAR/ RESTAURANT8c) CORN MILL (NIKA NIKA)10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDSõôó12b) TEAK PLANTATIONS#*13a) PROVISION KIOSK#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET15a) DUMPSTER×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYA Figure 6.3.82: Buoku Community Map 445 Figure 6.3.83: Buoku Community – Buffer on Boreholes 446 Figure 6.3.84: Buoku Community – Buffer on Church 447 Figure 6.3.85: Buoku Community – Buffer on Public Dumpster 448 Figure 6.3.86: Buoku Community – Buffer on Junior Secondary School 449 Figure 6.3.87: Buoku Community – Buffer on Provision Kiosks 450 Figure 6.3.88: Buoku Community – Buffer on Market 451 Figure 6.3.89: Buoku Community – Buffer on Mosque 452 Figure 6.3.90: Buoku Community – Buffer on Corn Mill 453 Figure 6.3.91: Buoku Community – Buffer on Preschool and Kindergarten 454 Figure 6.3.92: Buoku Community – Buffer on Primary School 455 Figure 6.3.93: Buoku Community – Buffer on Public Toilet 456 Figure 6.3.94: Buoku Community – Buffer on Major Roads 457 Konsua Community Figure 6.3.95: Yaya Reserve Map Showing the Location of Konsua 458 Konsua Legend Figure 6.3.95 (cont’d) 459 Legend1a) MAJOR ROAD NETWORK82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEî4a) CHURCH5b) PRIMARY SCHOOL²³5g) TEACHERS' BUNGALLOW5h) SCHOOL PLAY GROUNDKL7f) LOTTERY KIOSK10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDSÞ12c) SEEDLING NURSERY#*13a) PROVISION KIOSK×16a) PUBLIC TOILETForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYA Figure 6.3.96: Konsua Community Map 460 Figure 6.3.97: Konsua Community – Buffer on Borehole 461 Figure 6.3.98: Konsua Community – Buffer on Church 462 Figure 6.3.99: Konsua Community – Buffer on Provision Kiosks 463 Figure 6.3.100: Konsua Community – Buffer on Mosque 464 Figure 6.3.101: Konsua Community – Buffer on Corn Mill 465 Figure 6.3.102: Konsua Community – Buffer on Primary School 466 Figure 6.3.103: Konsua Community – Buffer on Major Roads 467 Figure 6.3.104: Konsua Community – Livestock Ownership 468 Figure 6.3.105: Konsua Community – Buffer on Public Toilet 469 Malamkrom Community Figure 6.3.106: Yaya Reserve Map Showing the Location of Malamkrom 470 Malamkrom Legend Figure 6.3.106 (cont’d) 471 Legend1a) MAJOR ROAD NETWORK82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEö4b) MOSQUE5b) PRIMARY SCHOOL10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDSõôó12b) TEAK PLANTATIONS#*13a) PROVISION KIOSK×16a) PUBLIC TOILETForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYA Figure 6.3.107: Malamkrom Community Map 472 Figure 6.3.108: Malamkrom Community – Buffer on Borehole 473 Figure 6.3.109: Malamkrom Community – Buffer on Church 474 Figure 6.3.110: Malamkrom Community – Buffer on Provision Kiosks 475 Figure 6.3.111: Malamkrom Community – Livestock Ownership 476 Figure 6.3.112: Malamkrom Community – Buffer on Mosque 477 Figure 6.3.113: Malamkrom Community – Buffer on Corn Mill 478 Figure 6.3.114: Malamkrom Community – Buffer on Primary School 479 Figure 6.3.115: Malamkrom Community – Buffer on Major Roads 480 Figure 6.3.116: Malamkrom Community – Buffer on Public Toilet 481 Sewiah Community Figure 6.3.117: Yaya Reserve Map Showing the Location of Sewiah 482 Sewiah Legend Figure 6.3.117 (cont’d) 483 Legend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION3a) BOREHOLEKJ3b) RIVER/STREAM (COMMUNITY WATER SOURCE)10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDSkj12a) BACKYARD FARMSõôó12b) TEAK PLANTATIONS×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYA Figure 6.3.118: Sewiah Community Map 484 Figure 6.3.119: Sewiah Community – Magnified 485 Figure 6.3.120: Sewiah Community – Buffer on Major Roads 486 Figure 6.3.121: Sewiah Community – Buffer on Borehole 487 Figure 6.3.122: Sewiah Community – Buffer on Church 488 Figure 6.3.123: Sewiah Community – Buffer on Corn Mill 489 Figure 6.3.124: Sewiah Community – Buffer on Primary School 490 Ahwene Community APPENDIX B: Nsemre Forest Reserve Community Maps Figure 6.3.125: Nsemre Reserve Map Showing the Location of Ahwene 491 Ahwene Legend Figure 6.3.125 (cont’d) 492 Legend1a) MAJOR ROAD NETWORK82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2f) ABANDONED HOUSE3a) BOREHOLEî4a) CHURCHö4b) MOSQUE5a) NURSERY AND KINGDERGARTEN8c) CORN MILL (NIKA NIKA)10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES"M11c) PAFORM DEMOS#*13c) ABANDONED KIOSK$+14b) MARKET ADMINISTRATION BUILDING15a) DUMPSTERForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYA Figure 6.3.126: Ahwene Community Map 493 Figure 6.3.127: Ahyiem Community – Buffer on Borehole 494 Figure 6.3.128: Ahyiem Community – Buffer on Church 495 Figure 6.3.129: Ahyiem Community – Buffer on Preschool and Kindergarten 496 Figure 6.3.130: Ahwene Community – Buffer on Mosque 497 Figure 6.3.131: Ahwene Community – Buffer on Corn Mill 498 Figure 6.3.132: Ahwene Community – Buffer on PAFORM Alternative Livelihood Program 499 Figure 6.3.133: Ahwene Community – Buffer on Major Road 500 Figure 6.3.134: Ahwene Community – Buffer on Public Toilet 501 Asuofre Community Figure 6.3.135: Nsemre Reserve Map Showing the Location of Asuofre 502 Asuofre Legend Figure 6.3.135 (cont’d) 503 Legend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEKJ3b) RIVER/STREAM (COMMUNITY WATER SOURCE)î4a) CHURCH5b) PRIMARY SCHOOL5h) SCHOOL PLAY GROUND)7g) CAPENTRY SHOP8c) CORN MILL (NIKA NIKA)10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES"M11c) PAFORM DEMOSõôó12b) TEAK PLANTATIONS×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYA Figure 6.3.136: Asuofre Community Map 504 Figure 6.3.137: Asuofre Community – Buffer on Borehole 505 Figure 6.3.138: Asuofre Community – Livestock Ownership 506 Figure 6.3.139: Ahyiem Community – Buffer on Corn Mill 507 Figure 6.3.140: Ahyiem Community – Buffer on PAFORM Alternative Livelihood Program 508 Figure 6.3.141: Ahyiem Community – Buffer on Primary School 509 Figure 6.3.142: Ahyiem Community – Buffer on Major Roads 510 Figure 6.3.143: Asuofre Community – Buffer on Public Toilet 511 Kofitsum Community Figure 6.3.144: Nsemre Reserve Map Showing the Location of Kofitsum 512 Kofitsum Legend Figure 6.3.144 (cont’d) 513 Legend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE3a) BOREHOLEî4a) CHURCH5b) PRIMARY SCHOOL5h) SCHOOL PLAY GROUND10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES"M11c) PAFORM DEMOSForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYA Figure 6.3.145: Kofitsum Community Map 514 Figure 6.3.146: Kofitsum Community – Buffer on Borehole 515 Figure 6.3.147: Kofitsum Community – Buffer on Church 516 Figure 6.3.148: Kofitsum Community – Livestock Ownership 517 Figure 6.3.149: Kofitsum Community – PAFORM Alternative Livelihood Program 518 Figure 6.3.150: Kofitsum Community – Buffer on Primary School 519 Figure 6.3.151: Kofitsum Community – Major Road 520 Pepewase Community Figure 6.3.152: Nsemre Reserve Map Showing the Location of Pepewase 521 Pepewase Legend Figure 6.3.152 (cont’d) 522 Legend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE3a) BOREHOLEî4a) CHURCHAD8d) ALOCOHOL DISTILERY10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES"M11c) PAFORM DEMOSkj12a) BACKYARD FARMSõôó12b) TEAK PLANTATIONS16b) PRIVATE DUGOUT LATERINESForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYA Figure 6.3.153: Pepewase Community – Map 523 Figure 6.3.154: Pepewase Community – Buffer on Borehole 524 Figure 6.3.155: Pepewase Community – Buffer on Church 525 Figure 6.3.156: Pepewase Community – Livestock Ownership 526 Figure 6.3.157: Pepewase Community – PAFORM Alternative Livelihood Program 527 Figure 6.3.158: Pepewase Community – Buffer on Major Road 528 Ayaayo Community APPENDIX C: Sawsaw Forest Reserve Community Maps Figure 6.3.159: Yaya Sawsaw Map Showing the Location of Ayaayo 529 Ayaayo Legends Figure 6.3.159 (cont’d) 530 Legend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEKJ3b) RIVER/STREAM (COMMUNITY WATER SOURCE)UT3c) COMMUNITY WATER WELLî4a) CHURCHö4b) MOSQUE5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5c) JUNIOR HIGH SCHOOL (JHS)5d) SCHOOL/CLASSROOM UNDER TREE¹½5e) ABANDONED SCH STRUCTUREÆV5f) SCHOOL FEEDING PROGRAM KITCHEN²³5g) TEACHERS' BUNGALLOW5h) SCHOOL PLAY GROUND6a) COMMUNITY CENTERß6b) COMMUNITY/TOWN BELL7a) BUSINESS AND TELECOMMUNICATION CENTERPB7b) PRIVATE BUSINESS ENTSB7c) BLACK SMITH SHOPTS7d) TAILORING SHOP7e) SHOE REPAIR SHOPKL7f) LOTTERY KIOSK)7g) CAPENTRY SHOP½7h) BICYCLE REPAIR SHOPô7i) HAIR SALON AND BABER SHOP8a) FOOD PROCESSING CENTER (GARI)8b) CHOP BAR/ RESTAURANT8c) CORN MILL (NIKA NIKA)AD8d) ALOCOHOL DISTILERY"u9a) COMMUNBITY CLINIC9b) DRUG STORE10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHED"M11c) PAFORM DEMOSÒ11d) WOMENS MUSHROOM PROGRAMkj12a) BACKYARD FARMSõôó12b) TEAK PLANTATIONSÞ12c) SEEDLING NURSERY#*13a) PROVISION KIOSK#*13b) CLUSTER OF KIOSKS/PROVISION SHOPS#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET$+14b) MARKET ADMINISTRATION BUILDING15a) DUMPSTER×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESþ17a) PETROL STATIONForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYALegend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEKJ3b) RIVER/STREAM (COMMUNITY WATER SOURCE)UT3c) COMMUNITY WATER WELLî4a) CHURCHö4b) MOSQUE5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5c) JUNIOR HIGH SCHOOL (JHS)5d) SCHOOL/CLASSROOM UNDER TREE¹½5e) ABANDONED SCH STRUCTUREÆV5f) SCHOOL FEEDING PROGRAM KITCHEN²³5g) TEACHERS' BUNGALLOW5h) SCHOOL PLAY GROUND6a) COMMUNITY CENTERß6b) COMMUNITY/TOWN BELL7a) BUSINESS AND TELECOMMUNICATION CENTERPB7b) PRIVATE BUSINESS ENTSB7c) BLACK SMITH SHOPTS7d) TAILORING SHOP7e) SHOE REPAIR SHOPKL7f) LOTTERY KIOSK)7g) CAPENTRY SHOP½7h) BICYCLE REPAIR SHOPô7i) HAIR SALON AND BABER SHOP8a) FOOD PROCESSING CENTER (GARI)8b) CHOP BAR/ RESTAURANT8c) CORN MILL (NIKA NIKA)AD8d) ALOCOHOL DISTILERY"u9a) COMMUNBITY CLINIC9b) DRUG STORE10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHED"M11c) PAFORM DEMOSÒ11d) WOMENS MUSHROOM PROGRAMkj12a) BACKYARD FARMSõôó12b) TEAK PLANTATIONSÞ12c) SEEDLING NURSERY#*13a) PROVISION KIOSK#*13b) CLUSTER OF KIOSKS/PROVISION SHOPS#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET$+14b) MARKET ADMINISTRATION BUILDING15a) DUMPSTER×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESþ17a) PETROL STATIONForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYA Figure 6.3.160: Ayaayo Community Map 531 Figure 6.3.161: Ayaayo Community – Buffer on Boreholes 532 Figure 6.3.162: Ayaayo Community – Buffer on Church 533 Figure 6.3.163: Ayaayo Community – Buffer on Clinic 534 Figure 6.3.164: Ayaayo Community – Buffer on Dumpster 535 Figure 6.3.165: Ayaayo Community – Buffer on Preschool and Kindergarten 536 Figure 6.3.166: Ayaayo Community – Buffer on Provision Kiosks 537 Figure 6.3.167: Ayaayo Community – Buffer on Market 538 Figure 6.3.168: Ayaayo Community – Buffer on Mosque 539 Figure 6.3.169: Ayaayo Community – Buffer on Corn Mills 540 Figure 6.3.170: Ayaayo Community – Buffer on Primary School 541 Figure 6.3.171: Ayaayo Community – Buffer on Major Road 542 Domeabra Community Figure 6.3.172: Sawsaw Reserve Map Showing the Location of Domeabra 543 Domeabra Legends Figure 6.3.172 (cont’d) 544 Legend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEKJ3b) RIVER/STREAM (COMMUNITY WATER SOURCE)UT3c) COMMUNITY WATER WELLî4a) CHURCHö4b) MOSQUE5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5c) JUNIOR HIGH SCHOOL (JHS)5d) SCHOOL/CLASSROOM UNDER TREE¹½5e) ABANDONED SCH STRUCTUREÆV5f) SCHOOL FEEDING PROGRAM KITCHEN²³5g) TEACHERS' BUNGALLOW5h) SCHOOL PLAY GROUND6a) COMMUNITY CENTERß6b) COMMUNITY/TOWN BELL7a) BUSINESS AND TELECOMMUNICATION CENTERPB7b) PRIVATE BUSINESS ENTSB7c) BLACK SMITH SHOPTS7d) TAILORING SHOP7e) SHOE REPAIR SHOPKL7f) LOTTERY KIOSK)7g) CAPENTRY SHOP½7h) BICYCLE REPAIR SHOPô7i) HAIR SALON AND BABER SHOP8a) FOOD PROCESSING CENTER (GARI)8b) CHOP BAR/ RESTAURANT8c) CORN MILL (NIKA NIKA)AD8d) ALOCOHOL DISTILERY"u9a) COMMUNBITY CLINIC9b) DRUG STORE10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHED"M11c) PAFORM DEMOSÒ11d) WOMENS MUSHROOM PROGRAMkj12a) BACKYARD FARMSõôó12b) TEAK PLANTATIONSÞ12c) SEEDLING NURSERY#*13a) PROVISION KIOSK#*13b) CLUSTER OF KIOSKS/PROVISION SHOPS#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET$+14b) MARKET ADMINISTRATION BUILDING15a) DUMPSTER×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESþ17a) PETROL STATIONForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYALegend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEKJ3b) RIVER/STREAM (COMMUNITY WATER SOURCE)UT3c) COMMUNITY WATER WELLî4a) CHURCHö4b) MOSQUE5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5c) JUNIOR HIGH SCHOOL (JHS)5d) SCHOOL/CLASSROOM UNDER TREE¹½5e) ABANDONED SCH STRUCTUREÆV5f) SCHOOL FEEDING PROGRAM KITCHEN²³5g) TEACHERS' BUNGALLOW5h) SCHOOL PLAY GROUND6a) COMMUNITY CENTERß6b) COMMUNITY/TOWN BELL7a) BUSINESS AND TELECOMMUNICATION CENTERPB7b) PRIVATE BUSINESS ENTSB7c) BLACK SMITH SHOPTS7d) TAILORING SHOP7e) SHOE REPAIR SHOPKL7f) LOTTERY KIOSK)7g) CAPENTRY SHOP½7h) BICYCLE REPAIR SHOPô7i) HAIR SALON AND BABER SHOP8a) FOOD PROCESSING CENTER (GARI)8b) CHOP BAR/ RESTAURANT8c) CORN MILL (NIKA NIKA)AD8d) ALOCOHOL DISTILERY"u9a) COMMUNBITY CLINIC9b) DRUG STORE10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHED"M11c) PAFORM DEMOSÒ11d) WOMENS MUSHROOM PROGRAMkj12a) BACKYARD FARMSõôó12b) TEAK PLANTATIONSÞ12c) SEEDLING NURSERY#*13a) PROVISION KIOSK#*13b) CLUSTER OF KIOSKS/PROVISION SHOPS#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET$+14b) MARKET ADMINISTRATION BUILDING15a) DUMPSTER×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESþ17a) PETROL STATIONForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYA Figure 6.3.173: Domeabra Community Map 545 Figure 6.3.174: Domeabra Community – Buffer on Boreholes 546 Figure 6.3.175: Domeabra Community – Buffer on Dumpster 547 Figure 6.3.176: Domeabra Community – Buffer on Provision Kiosks 548 Figure 6.3.177: Domeabra Community – Livestock Ownership 549 Figure 6.3.178: Domeabra Community – Buffer on Corn Mill 550 Figure 6.3.179: Domeabra Community – Buffer on Major Road 551 Ntema Community Figure 6.3.180: Sawsaw Reserve Map Showing the Location of Ntema 552 Ntema Legends Figure 6.3.180 (cont’d) 553 Legend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEKJ3b) RIVER/STREAM (COMMUNITY WATER SOURCE)UT3c) COMMUNITY WATER WELLî4a) CHURCHö4b) MOSQUE5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5c) JUNIOR HIGH SCHOOL (JHS)5d) SCHOOL/CLASSROOM UNDER TREE¹½5e) ABANDONED SCH STRUCTUREÆV5f) SCHOOL FEEDING PROGRAM KITCHEN²³5g) TEACHERS' BUNGALLOW5h) SCHOOL PLAY GROUND6a) COMMUNITY CENTERß6b) COMMUNITY/TOWN BELL7a) BUSINESS AND TELECOMMUNICATION CENTERPB7b) PRIVATE BUSINESS ENTSB7c) BLACK SMITH SHOPTS7d) TAILORING SHOP7e) SHOE REPAIR SHOPKL7f) LOTTERY KIOSK)7g) CAPENTRY SHOP½7h) BICYCLE REPAIR SHOPô7i) HAIR SALON AND BABER SHOP8a) FOOD PROCESSING CENTER (GARI)8b) CHOP BAR/ RESTAURANT8c) CORN MILL (NIKA NIKA)AD8d) ALOCOHOL DISTILERY"u9a) COMMUNBITY CLINIC9b) DRUG STORE10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHED"M11c) PAFORM DEMOSÒ11d) WOMENS MUSHROOM PROGRAMkj12a) BACKYARD FARMSõôó12b) TEAK PLANTATIONSÞ12c) SEEDLING NURSERY#*13a) PROVISION KIOSK#*13b) CLUSTER OF KIOSKS/PROVISION SHOPS#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET$+14b) MARKET ADMINISTRATION BUILDING15a) DUMPSTER×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESþ17a) PETROL STATIONForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYALegend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEKJ3b) RIVER/STREAM (COMMUNITY WATER SOURCE)UT3c) COMMUNITY WATER WELLî4a) CHURCHö4b) MOSQUE5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5c) JUNIOR HIGH SCHOOL (JHS)5d) SCHOOL/CLASSROOM UNDER TREE¹½5e) ABANDONED SCH STRUCTUREÆV5f) SCHOOL FEEDING PROGRAM KITCHEN²³5g) TEACHERS' BUNGALLOW5h) SCHOOL PLAY GROUND6a) COMMUNITY CENTERß6b) COMMUNITY/TOWN BELL7a) BUSINESS AND TELECOMMUNICATION CENTERPB7b) PRIVATE BUSINESS ENTSB7c) BLACK SMITH SHOPTS7d) TAILORING SHOP7e) SHOE REPAIR SHOPKL7f) LOTTERY KIOSK)7g) CAPENTRY SHOP½7h) BICYCLE REPAIR SHOPô7i) HAIR SALON AND BABER SHOP8a) FOOD PROCESSING CENTER (GARI)8b) CHOP BAR/ RESTAURANT8c) CORN MILL (NIKA NIKA)AD8d) ALOCOHOL DISTILERY"u9a) COMMUNBITY CLINIC9b) DRUG STORE10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHED"M11c) PAFORM DEMOSÒ11d) WOMENS MUSHROOM PROGRAMkj12a) BACKYARD FARMSõôó12b) TEAK PLANTATIONSÞ12c) SEEDLING NURSERY#*13a) PROVISION KIOSK#*13b) CLUSTER OF KIOSKS/PROVISION SHOPS#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET$+14b) MARKET ADMINISTRATION BUILDING15a) DUMPSTER×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESþ17a) PETROL STATIONForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYA Figure 6.3.181: Ntema Community – Buffer on Public Toilet 554 Figure 6.3.182: Ntema Community – Buffer on Public Toilet 555 Figure 6.3.183: Ntema Community – Buffer on Public Toilet 556 Figure 6.3.184: Ntema Community – Buffer on Public Toilet 557 Figure 6.3.185: Ntema Community – Buffer on Public Toilet 558 Papasu Community Figure 6.3.186: Sawsaw Reserve Map Showing the Location of Papasu 559 Papasu Legend Figure 6.3.186 (cont’d) 560 Legend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEKJ3b) RIVER/STREAM (COMMUNITY WATER SOURCE)UT3c) COMMUNITY WATER WELLî4a) CHURCHö4b) MOSQUE5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5c) JUNIOR HIGH SCHOOL (JHS)5d) SCHOOL/CLASSROOM UNDER TREE¹½5e) ABANDONED SCH STRUCTUREÆV5f) SCHOOL FEEDING PROGRAM KITCHEN²³5g) TEACHERS' BUNGALLOW5h) SCHOOL PLAY GROUND6a) COMMUNITY CENTERß6b) COMMUNITY/TOWN BELL7a) BUSINESS AND TELECOMMUNICATION CENTERPB7b) PRIVATE BUSINESS ENTSB7c) BLACK SMITH SHOPTS7d) TAILORING SHOP7e) SHOE REPAIR SHOPKL7f) LOTTERY KIOSK)7g) CAPENTRY SHOP½7h) BICYCLE REPAIR SHOPô7i) HAIR SALON AND BABER SHOP8a) FOOD PROCESSING CENTER (GARI)8b) CHOP BAR/ RESTAURANT8c) CORN MILL (NIKA NIKA)AD8d) ALOCOHOL DISTILERY"u9a) COMMUNBITY CLINIC9b) DRUG STORE10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHED"M11c) PAFORM DEMOSÒ11d) WOMENS MUSHROOM PROGRAMkj12a) BACKYARD FARMSõôó12b) TEAK PLANTATIONSÞ12c) SEEDLING NURSERY#*13a) PROVISION KIOSK#*13b) CLUSTER OF KIOSKS/PROVISION SHOPS#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET$+14b) MARKET ADMINISTRATION BUILDING15a) DUMPSTER×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESþ17a) PETROL STATIONForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYALegend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEKJ3b) RIVER/STREAM (COMMUNITY WATER SOURCE)UT3c) COMMUNITY WATER WELLî4a) CHURCHö4b) MOSQUE5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5c) JUNIOR HIGH SCHOOL (JHS)5d) SCHOOL/CLASSROOM UNDER TREE¹½5e) ABANDONED SCH STRUCTUREÆV5f) SCHOOL FEEDING PROGRAM KITCHEN²³5g) TEACHERS' BUNGALLOW5h) SCHOOL PLAY GROUND6a) COMMUNITY CENTERß6b) COMMUNITY/TOWN BELL7a) BUSINESS AND TELECOMMUNICATION CENTERPB7b) PRIVATE BUSINESS ENTSB7c) BLACK SMITH SHOPTS7d) TAILORING SHOP7e) SHOE REPAIR SHOPKL7f) LOTTERY KIOSK)7g) CAPENTRY SHOP½7h) BICYCLE REPAIR SHOPô7i) HAIR SALON AND BABER SHOP8a) FOOD PROCESSING CENTER (GARI)8b) CHOP BAR/ RESTAURANT8c) CORN MILL (NIKA NIKA)AD8d) ALOCOHOL DISTILERY"u9a) COMMUNBITY CLINIC9b) DRUG STORE10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHED"M11c) PAFORM DEMOSÒ11d) WOMENS MUSHROOM PROGRAMkj12a) BACKYARD FARMSõôó12b) TEAK PLANTATIONSÞ12c) SEEDLING NURSERY#*13a) PROVISION KIOSK#*13b) CLUSTER OF KIOSKS/PROVISION SHOPS#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET$+14b) MARKET ADMINISTRATION BUILDING15a) DUMPSTER×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESþ17a) PETROL STATIONForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYALegend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEKJ3b) RIVER/STREAM (COMMUNITY WATER SOURCE)UT3c) COMMUNITY WATER WELLî4a) CHURCHö4b) MOSQUE5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5c) JUNIOR HIGH SCHOOL (JHS)5d) SCHOOL/CLASSROOM UNDER TREE¹½5e) ABANDONED SCH STRUCTUREÆV5f) SCHOOL FEEDING PROGRAM KITCHEN²³5g) TEACHERS' BUNGALLOW5h) SCHOOL PLAY GROUND6a) COMMUNITY CENTERß6b) COMMUNITY/TOWN BELL7a) BUSINESS AND TELECOMMUNICATION CENTERPB7b) PRIVATE BUSINESS ENTSB7c) BLACK SMITH SHOPTS7d) TAILORING SHOP7e) SHOE REPAIR SHOPKL7f) LOTTERY KIOSK)7g) CAPENTRY SHOP½7h) BICYCLE REPAIR SHOPô7i) HAIR SALON AND BABER SHOP8a) FOOD PROCESSING CENTER (GARI)8b) CHOP BAR/ RESTAURANT8c) CORN MILL (NIKA NIKA)AD8d) ALOCOHOL DISTILERY"u9a) COMMUNBITY CLINIC9b) DRUG STORE10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHED"M11c) PAFORM DEMOSÒ11d) WOMENS MUSHROOM PROGRAMkj12a) BACKYARD FARMSõôó12b) TEAK PLANTATIONSÞ12c) SEEDLING NURSERY#*13a) PROVISION KIOSK#*13b) CLUSTER OF KIOSKS/PROVISION SHOPS#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET$+14b) MARKET ADMINISTRATION BUILDING15a) DUMPSTER×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESþ17a) PETROL STATIONForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYA Figure 6.3.187: Papasu Community Map 561 Figure 6.3.188: Papasu Community – Livestock Ownership 562 Figure 6.3.189: Papasu Community – Buffer on Major Road 563 Pipotrim Community Figure 6.3.190: Sawsaw Reserve Map Showing the Location of Pipotrim 564 Pipotrim Legend Figure 6.3.190 (cont’d) 565 Legend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEKJ3b) RIVER/STREAM (COMMUNITY WATER SOURCE)UT3c) COMMUNITY WATER WELLî4a) CHURCHö4b) MOSQUE5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5c) JUNIOR HIGH SCHOOL (JHS)5d) SCHOOL/CLASSROOM UNDER TREE¹½5e) ABANDONED SCH STRUCTUREÆV5f) SCHOOL FEEDING PROGRAM KITCHEN²³5g) TEACHERS' BUNGALLOW5h) SCHOOL PLAY GROUND6a) COMMUNITY CENTERß6b) COMMUNITY/TOWN BELL7a) BUSINESS AND TELECOMMUNICATION CENTERPB7b) PRIVATE BUSINESS ENTSB7c) BLACK SMITH SHOPTS7d) TAILORING SHOP7e) SHOE REPAIR SHOPKL7f) LOTTERY KIOSK)7g) CAPENTRY SHOP½7h) BICYCLE REPAIR SHOPô7i) HAIR SALON AND BABER SHOP8a) FOOD PROCESSING CENTER (GARI)8b) CHOP BAR/ RESTAURANT8c) CORN MILL (NIKA NIKA)AD8d) ALOCOHOL DISTILERY"u9a) COMMUNBITY CLINIC9b) DRUG STORE10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHED"M11c) PAFORM DEMOSÒ11d) WOMENS MUSHROOM PROGRAMkj12a) BACKYARD FARMSõôó12b) TEAK PLANTATIONSÞ12c) SEEDLING NURSERY#*13a) PROVISION KIOSK#*13b) CLUSTER OF KIOSKS/PROVISION SHOPS#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET$+14b) MARKET ADMINISTRATION BUILDING15a) DUMPSTER×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESþ17a) PETROL STATIONForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYALegend1a) MAJOR ROAD NETWORK1b) JUNCTION82a) HOUSE WITH STRAW ROOF AND MUD/EARTH CONSTRUCTION2b) HOUSE WITH IRON ROOF AND CEMENT CONSTRUCTION82c) HOUSE WITH STRAW ROOF AND CEMENT CONSTRUCTION2d) HOUSE WITH IRON ROOF AND MUD/EARTH CONSTRUCTION2e) UNCOMPLETED HOUSE2f) ABANDONED HOUSE3a) BOREHOLEKJ3b) RIVER/STREAM (COMMUNITY WATER SOURCE)UT3c) COMMUNITY WATER WELLî4a) CHURCHö4b) MOSQUE5a) NURSERY AND KINGDERGARTEN5b) PRIMARY SCHOOL5c) JUNIOR HIGH SCHOOL (JHS)5d) SCHOOL/CLASSROOM UNDER TREE¹½5e) ABANDONED SCH STRUCTUREÆV5f) SCHOOL FEEDING PROGRAM KITCHEN²³5g) TEACHERS' BUNGALLOW5h) SCHOOL PLAY GROUND6a) COMMUNITY CENTERß6b) COMMUNITY/TOWN BELL7a) BUSINESS AND TELECOMMUNICATION CENTERPB7b) PRIVATE BUSINESS ENTSB7c) BLACK SMITH SHOPTS7d) TAILORING SHOP7e) SHOE REPAIR SHOPKL7f) LOTTERY KIOSK)7g) CAPENTRY SHOP½7h) BICYCLE REPAIR SHOPô7i) HAIR SALON AND BABER SHOP8a) FOOD PROCESSING CENTER (GARI)8b) CHOP BAR/ RESTAURANT8c) CORN MILL (NIKA NIKA)AD8d) ALOCOHOL DISTILERY"u9a) COMMUNBITY CLINIC9b) DRUG STORE10a) ) HOUSEHOLDS THAT OWN LIVESTOCKQ10b) HUSEHOLD LIVESTOCK STRUCTURES^_10c) HOUSEHOLD GRAIN STORAGE STRUCTURES!(11a) MTS HOUSEHOLDS11b) MTS/CFMP COMMUNITY LIVESTOCK SHED"M11c) PAFORM DEMOSÒ11d) WOMENS MUSHROOM PROGRAMkj12a) BACKYARD FARMSõôó12b) TEAK PLANTATIONSÞ12c) SEEDLING NURSERY#*13a) PROVISION KIOSK#*13b) CLUSTER OF KIOSKS/PROVISION SHOPS#*13c) ABANDONED KIOSK14a) COMMUNITY MARKET$+14b) MARKET ADMINISTRATION BUILDING15a) DUMPSTER×16a) PUBLIC TOILET16b) PRIVATE DUGOUT LATERINESþ17a) PETROL STATIONForest Reserve BoundariesName of ReserveNSEMRESAWSAWYAYA Figure 6.3.191: Pipotrim Community Map 566 Figure 6.3.192: Pipotrim Community – Buffer on Church 567 Figure 6.3.193: Pipotrim Community – Buffer on Dumpster 568 Figure 6.3.194: Pipotrim Community – Buffer on Preschool and Kindergarten 569 Figure 6.3.195: Pipotrim Community – Buffer on Provision Kiosks 570 Figure 6.3.196: Pipotrim Community – Livestock Ownership 571 Figure 6.3.197: Pipotrim Community – Buffer on Mosque 572 Figure 6.3.198: Pipotrim Community – Buffer on Corn Mill 573 Figure 6.3.199: Pipotrim Community – Buffer on Primary School 574 Figure 6.3.200: Pipotrim Community – Buffer on Major Road 575 APPENDIX D: Population, Religion and Poverty in the Northern Ghana Population, Religion and Poverty in the Northern Parts of Ghana In 2010 Ghana’s Upper East (UE), Upper West (UW) and Northern Region (NR) (from here on northern most regions) represented about 17% (4,216,659) of the country’s 2,465,8823 million population (Ghana Statistical Service, 2010). The NR alone is about twice the landmass of UE and UW combined (figure 6.3.282 below) and in 2010 represented about 59% of the population of the entire area while UE and UW represented approximately 25% and 16% respectively (figure 6.3.282 below). The dominant religion in the NR is Islam and Moslems make up about 60% of the population while in UW and UE, Christians dominate making up about 44% and 42% respectively (figure 6.3.282). The Moslems in the NR (60%) significantly outnumber those in UW (38%) and UE (27%). Brong Ahafo (BA) has a relatively high Christian population (72%) creating a much more favorable conditions for Christian migrants into BA to assimilate and develop socio-economically faster than migrants from other faiths. In their new communities in BA, strong social networks mediated through religion often acts as a precursor for the development of other essential livelihood assets. The quote below provides a glimpse into the critical role of churches in developing migrant communities in BA. At the time of this study the community in question comprised predominately of tribes from Ghana’s three northernmost regions and at the time of this study had two mosques and no church (appendix 6.2.2). “The first family to settle and build a house here in Abrefakrom was one palm wine tapper Yaw Kyeremeh and his wife who also operated a chop bar. It was after their initial settlement here that other settlers joined him from the three Northern Regions. Nana Abrefa the Omanhene of Wenchi traditional area and the custodian of this land later came to name the place Abrefakrom meaning Abrefa’s town or village (krom). At present we have the Dargartis, Sissalas, Dagombas and Waala (northern tribes) who happen to be the dominant groups. As we speak there is no single school in this town. An NGO called the Church of Pentecost is building a three-room Day Nursery and Kindergarten, but this is not enough to cater for all our children especially the older ones. Hence some of our children walk several kilometers to Buoko or Chiraa to school. It is not surprising that our town has only a handful of educated adults. The very few that 576 have any education brought it along from their hometowns. So, you see there are only two of us in this town who can read and write (Abrefakrom Non-MTS Focus Group Interview, 2009).” Figure 6.3.201: Ghana’s Population Distribution by Region in 2010 577 Figure 6.3.202: Dominant Religions in Ghana’s three Northmost Region in 2010 578 Poverty in the Northern Parts of Ghana Figure 6.3.283 below describes differences in 2010 poverty rates in Ghana’s three northernmost regions and Brong Ahafo. Of the four regions under consideration, Brong Ahafo experienced the least poverty rate of 28% followed by Upper East (44%) and the North (50%). Upper West region recorded by far the highest poverty rate of 71% in 2010 yet was the second least contributor to both national (8%) and regional (17%) poverty levels. The relatively small contribution of Upper East and Upper West to national and regional poverty levels compared to North is due to their relatively small population size (see figure 6.3.281 above). Thus, Upper East with a 44% poverty rate contributed the least to national (7%) and regional poverty (16%) level while the North contributed the highest to national (21%) and regional (44%) poverty levels in 2010 (Ghana Census, 2010). Like poverty, extreme poverty rates in 2010 followed similar patterns for all four regions in terms of contributions to national and regional rates (figure 6.2.4). The disproportionately large differences in poverty and extreme poverty rates between Brong Ahafo and the three northernmost regions may be traced to large disparities in natural and man-made resources in the four regions. Several studies (see Poku-Boansi, and Amoako, Clifford, 2015, Qui, 2012 and Tsikata and Seini, 2004 and Bukari, Aabeyir and Basommi, 2014) document the large disparities in resources (natural, physical, financial, and human capital assets) between northern and southern parts of Ghana. The relatively large north-south resource gaps thus explain the long history of north-south migratory patterns observed in Ghana particularly between the three northernmost regions and Brong Ahafo. These resource gaps inadvertently create a natural tendency for north-south migration among Ghana’s resource poor in search of productive agricultural land as well as employment in cocoa production and forest establishment (see Adaawen and Owusu, 2013, and van der Geest, 2008 and 2011). 579 Figure 6.3.203: Poverty Rates in Ghana’s three Northern-most Region in 2010 580 Figure 6.3.204: Extreme Poverty Rates in Ghana’s three Northmost Region in 2010 581 Religion Migration and Poverty in the Northern Parts of Ghana Perched in the center of the country and closest to NR, UE and UW, the BA region has for decades played home to hundreds of migrant tribes from Ghana’s three northern regions. According to Ghana’s 2010 census, UW alone contributed the highest net migration of 95,713 into BA region followed by the NR (77,170) and then Upper East (56,482). This study assumes that migrants from the UW and UE due to their predominantly Christian backgrounds are more likely to assimilate into BA region and improve their livelihoods faster than those from the NR. The preceding argument assumes that a general improvement in key livelihood assets in migrant communities in BA is mediated through social capital assets such as religion and strong friendship and family ties in the host region. Hence Christian migrants from UW, UE and NR migrating into a predominantly Christian BA region stand a higher chance of assimilating and breaking out of poverty faster than Moslem migrants into BA. The discussion above explains the hypothesized effect of mosques on both MTS community and household selection or participation (see table 6.2.2 above). The presence of mosques in a community thus signals the presence of a group of households likely to be socio- economically disadvantaged due to their religion and thus need to be disproportionately targeted by livelihoods improvement programs such as the MTS. The BPM under sections 6.1 and 6.2 thus hypothesized that an increase in access to mosques increases the predicted probability of MTS community selection and household participation. Access to mosques was thus included in the BPM to help predict the effect poverty status (masked by religion) on a household’s willingness to participate in MTS. 582 CHAPTER 7: BETWEEN AND WITHIN GROUP CHANGES IN HOUSEHOLD LIVELIHOOD ASSETS IN YAYA, NSEMRE AND SAWSAW FOREST RESERVE COMMUNITIES 583 Abstract Chapter seven used household survey data to track changes in livelihood assets among MTS and non-MTS households since the launching of the MTS project in 2002 (research question three). Since respondents were asked to recall information going back 10 years from the time of data collection in 2009, the period 1999 was conveniently used in this study to represent the start date of Ghana’s MTS project. Hence chapter seven investigates differences in five key livelihood assets among both MTS and non-MTS households at the start of the MTS program (1999) and 10 years after MTS implementation (2009). Survey data from both Nsemre and Sawsaw reserve communities were combined and households within these two forest reserve communities treated as a single group of non-MTS households outside of Yaya (from here on referred to as NsemSaw). The analysis thus compares NsemSaw with MTS and non-MTS households in Yaya communities where the MTS project was under a pilot test. Data for the analysis was drawn from 878 household surveys focus group interviews conducted in the 19 research communities. The result from the household livelihoods index analysis suggests that, aggregate household physical capital index increased by 17% between 1999 and 2009 while social and financial capital both increased by 2% for the same period. Also, human capital index increased by 3% for the same time period while natural capital index decreased by 5%. In order to track changes in household livelihood assets before and after the MTS, a set of livelihood indexes were generated for the five major livelihood assets groups (financial, physical, human, natural and social) and then used to construct household livelihood pentagons for the purposes of comparing within and between group differences for two the time periods before and after MTS. 584 Introduction Under chapter seven, a two-sample t-test was used to determine between and within group differences in household livelihood indexes and the extent to which these differences are significant for the two time periods before MTS (1999) and after MTS (2009). The results of the two-sample t-test for each livelihood index are then discussed following the t-test results. The summary descriptive statistics for all five asset/indexes including financial capital indexes are presented in Appendices O through S. Appendix O through S thus includes the questions asked in the survey questionnaire (LMT), the response rate, and mean responses to each question and the computation of indexes from the means. Both the pooled and disaggregated mean responses for each question are presented. At the end of each subsection, a single index is generated for each livelihood asset category (Financial, Physical Natural, Human and Social) for MTS, non- MTS in Yaya and NsemSaw research groups. Since the tables in Appendices O, P, Q, R, and S primarily demonstrate how indexes were constructed; only the pooled results for the livelihood index in question in interpreted. Also, the disaggregated results are discussed only for the t-test results and livelihood pentagons in the main text (section 7.1 through 7.6). Interpretation of the Two-Sample T-test results The Two-Sample t-test used in this study is interpreted within the framework of two hypotheses: whether differences exist; 1) between groups and 2) within groups for the two time periods (1999 and 2009) and the level of significance of any observed differences in livelihood assets. Essentially the Two-Sample t-test investigates whether the mean difference between any two groups being compared is statistically significant (and possibly explained by some deliberate action such as the MTS project) or simply due to chance or some random event. 585 Hypothesis One: Between Group Differences H0: There is no statistically significant difference in mean livelihoods indexes between MTS, Non-MTS in Yaya and NsemSaw research groups in 1999 (the base year) and 2009 (10 years after MTS). The null hypothesis thus suggests that the difference in mean for a particular livelihoods index in 1999 and 2009 is equal or insignificant across all three research groups. H1: There is statistically significant differences in mean livelihoods indexes between MTS, Non- MTS in Yaya and NsemSaw research groups in 1999 (the base year) and 2009 (10 years after MTS). The alternate hypothesis thus suggests that the mean for a particular livelihoods index in 1999 and 2009 across all three research groups is significantly different at either 99.99% (*** p<0.001), 99% (*** p<0.01) or 95% (*** p<0.05) confident levels and may be explained by a non-random even in this case the implementation of the MTS or similar project. Hypothesis Two: Within Group Differences H0: There is no statistically significant difference in mean livelihoods indexes within MTS, Non- MTS in Yaya and NsemSaw research groups between 1999 (the base year) and 2009 (10 years after MTS). The null hypothesis thus suggests that the values for any of the livelihood indexes for each of three research groups remained unchanged (or whatever change occurred was insignificant) between 1999 and 2009 across all three research groups. H1: There is statistically significant differences in mean livelihoods indexes between MTS, Non- MTS in Yaya and NsemSaw research groups in 1999 (the base year) and 2009 (10 years after MTS). The alternate hypothesis thus suggests that for any of the research groups there was a significantly change (99.99% (*** p<0.001), 99% (*** p<0.01) or 95% (*** p<0.05) confident level) in any particular livelihoods index between 1999 and 2009. The non-random nature of the 586 within group difference may be attributed to a deliberate intervention such as the implementation of the MTS or similar projects. 587 7.1 Financial Capital To best understand households’ stock of financial capital within the research communities, section 7.1 first examines the different ways in which the basic unit of financial capital/asset (cash) is generated (e.g., from the sale of farm produce, livestock, wage labor or remittances), saved and expended by the household. Because cash may be conveniently saved in formal and informal financial institutions and readily withdrawn and exchanged for goods and services it remains the most common and versatile form of financial capital in both urban and rural households. The lack of financial institutions in most rural communities in Ghana in part explain the preference among most households in the research communities to hold their financial capital in relatively lumpy forms (e.g., livestock, grains, timber and jewelry) that require much more elaborate processes to convert into cash or used in transactions. Refer to section 3.3.1 of chapter three for a broader discussion of the literature on financial capital and how it helps shapes rural households and livelihoods. Section 7.1 discusses changes in 11 financial capital indexes in the three research communities before and after the MTS project. Thus table 7.1.1 below thus describes briefly how each of the 11 financial capital indexes were generated while table 7.1.2 presents the results of a Two Sample T-test used to track (within and between group) changes in each of the 11 indexes before (1999) and after (2009) the MTS project. The tables presented in Appendix O describes in greater detail how each financial capital index was constructed from the household survey results. 588 Table 7.1.1: Definition of Household Financial Capital Indexes No. Index Name of Index/ Variable 1 PHISI Primary HH Income Sources Index 2 SHISI Secondary HH Income Sources Index Description of Indexes Computation of Indexes PHISI captures the contribution of crop and livestock production (primary source of income) towards the household’s financial needs. A household scores 5 points for either crop or livestock if any of these sources is the highest contributor to total HH income and 0 if there is no income from that source. The highest attainable PHISI is thus 10 and the lowest is 0 and the PHISI index expressed in terms of level of income contribution from crop and livestock (primary sources) ranges from 0 to 1. SHISI captures house income contributions from six sources (sales of labor, remittance, pension, NTFP, Forestry including MTS and Other sources). Since the primary employment and source of income in all 19 research communities are crop and livestock production these six income sources are considered secondary sources. Like PHISI, a household scores 5 points for each of the secondary sources if that particular source is the highest contributor to total HH income and 0 if there is no income from the source. The highest attainable SHISI is thus 30 and the lowest is 0 and the SHISI index expressed in terms of level of income contribution from secondary sources ranges from 0 to 1. 589 10_2EAPHISI21itt===ni30_2EASHISI83itt===ni No. Index Name of Index/ Variable 3 HEI1 Primary HH Expenditure Index1 4 HHE2 Secondary HH Expenditure Index Table 7.1.1 (cont’d) Description of Indexes Computation of Indexes HEI1 captures a house’s expenditure on seven different items including: (1) staple foods, 2) education, 3) health, 4) consumer goods, cloths and cosmetics, 5) transportation, 6) funerals and weddings and 7) farm equipment). These seven items were considered primary expenditure items because they consume the bulk of the average household’s income. In order of magnitude, a household earns 10 points on a particular item if that item consumed the most of total HH income and 4 if that item consumed very little or no HH income. For each HH the highest attainable HEII is thus 49 and the lowest is 0 and the HEI1 index expressed in terms of level of spending on primary expenditure items ranges from 0.48 to 1. HEI2 captures a house’s expenditure on the following six items: (1) fruits & vegetables including spices 2) alcoholic beverages, 3) fuel (wood and kerosene), 4) loan repayments 5) home repair, and 6) other items. These seven items were considered secondary expenditure items because they generally consume relatively less of average household’s income. In order of magnitude, a household earned 7 points on a particular item if that item consumed the most of total HH income (for the period under consideration) and 1 if little or no income is expended on that source. The highest attainable HEI2 is thus 28 and the lowest is 7. The HEI2 index expressed in terms of level of spending on secondary expenditure items ranges from 0.25 to 1. 590 49b H_Exp_HEI14924itt===ni28bH_Exp_HEI2287itt===ni No. Index Name of Index/ Variable 5 HHBAI HH Bank Account Index 6 FHHBAI Frequency of HH Banking Activities Index Table 7.1.1 (cont’d) Description of Indexes Computation of Indexes HHBAI capture local and formal bank account ownership by the household. Hence if at least one member of the HH has a bank account with any of the formal financial institutions (Ghana Commercial Bank, Agricultural Development Bank, SGSSB, etc.) the household earns a score of 1 and 0 if no one in the HH has any such account. Likewise, if any member of the HH has s Susu account (local/traditional banking system) the HH receives a score of 1 and 0 if no such account exists within the HH. The total attainable score for HH bank account ownership is thus 2 and the minimum is 0. HHBAI index expressed in terms of HH bank account ownership ranges from 0 to 1. FHHBAI capture the frequency with which HHs use formal (banks) and local (Susu Collectors) financial institutions. Hence HHs that saves or withdraws money from a bank at least 5 times in a year scores 3 points for frequency of activity with a bank and 0 if no such activity ever occurs in year. The same scoring is repeated for frequency of activities with the Susu Collector. Also, HH’s that obtain loans from a bank at least five times in a year scores 3 points and 0 if no such loan services were used. The total attainable score for HH banking activities is 12 and the minimum is 0. FHHBAI index expressed in terms of frequency of HH banking activities thus ranges from 0 to 1. 591 2(L_Sav_1a)HHBAI21itt===ni12QFSLA_FHHBAI120itt===ni No. Index Name of Index/ Variable 7 TLSAI Total Saving and Loan Amount Index 8 AHHIS Annual HH Income and Sufficiency Index Table 7.1.1 (cont’d) Description of Indexes Computation of Indexes TLSAI captures a HH’s total savings with either a Formal Bank or a Susu Collector for the two time periods (10 years) before MTS (1990-1999) and after MTS (2000-2009). Households that did not save with either of these institutions scores 1 while those that saved between 1 and 500 Ghana Cedis (GHC) scored 2. Similarly, households that saved between 501 and 1000 GHC scored 3 and those that save more than 1000GHC scored 4 points. The same scoring matrix was applied to loans received from either the Formal Bank or a Susu Collector. The highest attainable score for HH total saving and loans is 8 and the minimum is 2. TLSAI index expressed in terms of HH’s total savings and loans for the periods in question thus ranges from 0.25 to 1. AHHIS captures HH annual income for the two time periods 1999 and 2009 and trends in income sufficiency before (1990-1999) and after (2000-2009) MTS. Households earning no income scored 1 while those with income between 1 and 500 GHC scored 2. Households with income between 501 and 1000 GHC scored 3 while those earning more than 1000 GHC scored 4 points. On the income sufficiency scale, HHs that claimed their income “decreased and was not sufficient” for the HH scored 1 while those who experienced “increased but not sufficient” income scored 2. Similarly, HHs that claimed their income “decreased but was sufficient” scored 3 while those who experienced “stable but sufficient” income scored 4. The highest attainable score for AHHIS is 9 and the minimum is 2 and AHHIS index expressed in terms of HH’s annual income and sufficiency for the periods in question thus ranges from 0.22 to 1. 592 8S_L_1Q TLSAI21itt===ni9GTA_HHQ CA_HHQ AHHIS51it41itt+=====nini No. Index Name of Index/ Variable Table 7.1.1 (cont’d) Description of Indexes Computation of Indexes 9 VSEI Village Shop/Store Expenditure Index VSEI captures HHs’ frequency of purchase of the three most expensive items in the village store (from here on shop) in a month and the average monthly expenditure at the store for the two time periods before (1999) and after (2009) MTS. Households that “Never” purchase any of the 3 most expensive items score 1 while those that purchase them a “Few” times (1-3 times) score 2. Similarly, HHs that purchase “Many times (3-5 times) score 3 while those that purchase “Very Many” times (more than 5 times) score 4. In terms of average monthly expenditure (regardless of items purchased), HHs that claim to have zero expenditure (0 GHC) at the village store score 1 while those that spend between 1-2 GHC each month score 2 points. Similarly, HHs that spend between 6-10 GHC each month at the village store score 3 while those that spend more than 10 GHC each month regardless of items purchased score 4 points for this indicator. The maximum attainable VSEI score is thus 8 and the lowest is 2. The VSEI index expressed in terms of level of HH monthly expenditure activities at the village store rages from 0.25 to 1. 593 8AEVSS FP_WQVSEI41it41itt+=====nini No. Index Name of Index/ Variable Table 7.1.1 (cont’d) Description of Indexes Computation of Indexes 10 HHILI HH Item Liquidati on Index HHILI captures frequency of sales of HH possessions to supplement HH income during hard times in 1999 (before MTS) and 2009 (after MTS) and general trends in accumulation of HH possessions during the periods before (1990-1999) and after MTS (2000-2009). Households that “Never” sold any HH items to run the HH scored 4 while those that claimed to have sold only “A Few” times (1-3 times) scored 3. Similarly, HHs that sold their possessions “Many” times (4-5 times) scored 2 while those that sold “Very Many” (more than 5 times) scored 1. In terms of trends in possessions, HH that experienced a “decline and not sufficient” trend scored 1 while those who experienced “increased but not sufficient” trend in HH possessions scored 2. Similarly, HHs that claimed their possessions “decreased but remained sufficient” scored 3, while those that experienced “stable but sufficient” trends scored 4. The highest attainable score of 5 on the sufficiency scale was among HH that experienced “increased and sufficient” levels of HH possessions. The maximum attainable HHILI score is thus 9 and the lowest is 2. The HHILI index expressed in terms of trends in HH item accumulation and liquidation rages from 0.22 to 1. 11 CHHF CI Combine HH Financial Capital Index CHHFCI captures a HH’s aggregate or combined Financial Capital endowment. This index is simply the average of all the 10 indexes hence the maximum attainable score on a household’s combined Financial Capital endowment is 10 and the minimum is 0. CHHFCI index thus ranges from 0 to 1. 594 9THH_PQ DHHI_CQHHILI51it41itt+=====nini10] HHILI+VSEI +AHHIS +TLSAI + FHHBAI+HHBAI +HHE2+ HEI1+SHISI +[PHISICHHFCIt= 7.1.1 Two Sample T-test of between and within group differences in 1999 and 2009 Section 7.1.1 investigates between and within group differences for all 11 financial capital indexes described in tables 7.1.1 above. The Two-Sample t-test determined what changes if any occurred in household financial capital indexes between 1999 and 2009 and also the degree to which these changes (between and within group) were significant using three different confident levels 99.99% (*** p<0.001), 99% (**p<0.01) and 95% (*p<0.05). Within group difference for each livelihood index is computed by subtracting the mean result obtained for that particular index in 2009 from the mean in 1999 for the same research group (MTS households in Yaya or non-MTS groups in Yaya). Since the Two-Sample t-test is computed by differencing the mean of the current year (2009) from the initial year (1999), a positive difference suggests a decline in livelihood index while a negative difference suggests an improvement. The computation for between group differences is similar to within group differences only in the former, the changes were computed for two different groups for a particular time period either 1999 or 2009. Finally, section 7.6 constructed household livelihood pentagons in order to visually display (using graphs) between and within group changes in all the five livelihood assets including Combine Combined Financial Capital Index (CHHFCI). The household livelihood pentagons in section 7.6 thus compares the between and within group changes across all the three research groups for two time periods before MTS in 1999 and after MTS in 2009. The section that follows discusses the results of the Two sample t-test presented in table 7.1.2a and 7.1.2b. 595 7.1.1.1 Primary Household Income Sources Index (PHISI) In 1999 and 2009 there was no significant difference in the mean Primary Household Income Sources Index (PHISI) among all three research groups suggesting that whatever difference is observed may be attributed to some random chance. Among all three groups the MTS recorded the highest increase of 2.9% in PHISI between 1999 and 2009 followed by NsemSaw (2.2%) and Non-MTS (1.3%). Though all three groups experienced an increased in PHISI these differences according to the t-test were also insignificant. These results are not entirely unexpected given that the predominant economic activity in most rural communities in Ghana small and medium-scale crop and livestock production. 7.1.1.2 Secondary Household Income Sources Index (SHISI) In 1999 and 2009 there was no statistically significant difference in mean Secondary Household Income Sources Index (SHISI) between MTS and Non-MTS. In 1999 the SHISI index among MTS was significantly (*** p<0.001) higher (5.5%) than that of NsemSaw. By 2009 however the gap in SHISI between MTS and NsemSaw increased to 8.4% and remained significant (*** p<0.001). Similarly, in 1999 the mean SHISI index among the Non-MTS group was significantly (*** p<0.001) higher (6.2%) than NsemSaw and by 2009 the observed difference between both groups was similar both in magnitude and significance to that observed between MTS and NsemSaw. Of the three research groups only MTS and Non-MTS experienced an increased in SHISI between 1999 and 2009 though these changes were insignificant. NsemSaw experienced an insignificant decline (1.5%) in SHISI between 1999 and 2009. 596 7.1.1.3 Primary Household Expenditure Index (HEI1) In 1999 and 2009 there was no significant difference in the mean Primary Household Expenditure Index (HEI1) between any of the three research groups. Among the three groups however, MTS in Yaya experienced the highest (3.8%) and significant (*p<0.05) within group increase between 1999 and 2009. Both Non-MTS in Yaya and NsemSaw experienced respectively 1.7% and 0.6% increases in HEI1 between 1999 and 2009 however these differences were insignificant according to the t-test results. 7.1.1.4 Secondary Household Expenditure Index (HHEI2) In 1999 the mean HHEI2 among MTS was 1.2% higher than the Non-MTS however this difference was not significant however by 2009 the gap had increased significantly (**p<0.01) to 3.4%. In 1999 the HHEI2 index among MTS was significantly (***p<0.001) higher (6.3%) than that of the NsemSaw groups and by 2009 the difference increased to 7.9%. In 1999 the mean HHEI2 among Non-MTS households was significantly (***p<0.001) higher (5.1%) than NsemSaw. By 2009 however the gap in HHEI2 between Non-MTS households and NsemSaw decreased to 4.5% however the differences was still significant (***p<0.001). Among all three research groups, only MTS in Yaya experienced an increase (1.4%) in within group HHEI2 index between 1999 and 2009. Both Non-MTS in Yaya and NsemSaw experienced respectively a 0.8% and 0.2% decline in within group HHEI2 index between 1999 and 2009 though these changes were insignificant. 597 7.1.1.5 Household Bank Account Index (HHBAI) In both 1999 and 2009 there was no significant difference in the mean Household Bank Account Index (HHBAI) between MTS and Non-MTS households. Though in 1999 the difference in HHBAI index between MTS and NsemSaw was insignificant by 2009 the MTS group experienced a significantly (***p<0.001) higher (14.4%) HHBAI index compared to NsemSaw. The trend in between group differences in HHBAI for Non-MTS and NsemSaw was similar to MTS and NsemSaw. Among all three research groups, the MTS and Non-MTS households experienced respectively 8.1% and 10.4% increases in within group HHBAI index between 1999 and 2009 and both these changes were significant (*p<0.05). Unlike the two research groups in Yaya, the NsemSaw group experienced no change in mean HHBAI between 1999 and 2009. 7.1.1.6 Frequency of Household Banking Activity Index (FHHBAI) In both 1999 and 2009 there was no significant difference in the mean Frequency of Household Banking Activity Index (FHHBAI) between MTS and Non-MTS households. In 1999 there was no significant difference in the mean FHHBAI between MTS and NsemSaw however by 2009 there the MTS group experienced a significantly (***p<0.001) higher (3.6%) increase in FHHBAI index relative to the NsemSaw group. The trend in between group differences in FHHBAI for Non-MTS and NsemSaw was similar to MTS and NsemSaw. In terms of differences within groups, there was no statistically significant difference in mean FHHBAI within all three research groups from 1999 to 2009. Though the t-test results suggest that none of the changes within each of the research groups were significant, it is worth noting that both MTS and Non-MTS groups in Yaya averaged 1.2% and 1.5% increases respectively in their mean FHHBAI between 1999 and 20009 while NsemSaw declined by 0.5%. 598 7.1.1.7 Total Saving and Loan Amount Index (TLSAI) Table 7.1.2b below suggests that for the two time periods (1999 and 2009) there was no significant difference in the mean Total Saving and Loan Amount Index (TLSAI) amongst all three research groups. Among all three research groups only the MTS in Yaya group experienced a significant (*p<0.05) increase (3.5%) in mean TSAI between 1999 and 2009. For the same time period between 1999 and 2009 Non-MTS experienced a 2.4% increase in TLSAI while NsemSaw declined by 2.0% however these changes experienced within both groups were insignificant according to the t-test and may be attributed to chance. 7.1.1.8 Annual HH Income and Sufficiency Index (AHHIS) In 1999 the mean Annual HH Income and Sufficiency Index (AHHIS) among Non-MTS households was significantly (**p<0.01) higher (4.4%) than those of the MTS group. Similarly, in 1999 the mean AHHIS for the NsemSaw groups was significantly (***p<0.001) higher (9.2%) than those of MTS households. By 2009 however the mean AHHIS for MTS surpassed those of Non-MTS by 1.9% and NsemSaw by 0.2% though these differences according to the t-test results were insignificant. Among all the three research groups only the MTS group in Yaya experienced a significant (*p<0.05) increase (3.8%) in mean AHHIS between 1999 and 2009. Between 1999 and 2009 Non-MTS households experienced a 2.5% decline in mean AHHIS however this was change was insignificant. Among the three groups, NsemSaw was the only one to experience a decline (7.4%) in mean AHHIS between the two time periods and this change according to the t-test was significant (***p<0.001). 599 7.1.1.9 Village Shop/Store Expenditure Index (VSEI) In 1999 the mean Village Shop/Store Expenditure Index (VSEI) among MTS households was significantly (*p<0.05) higher (3.7%) than those of NsemSaw and by 2009 the gap in VSEI between both groups increased significantly (**p<0.01) from 3.7% to 5.5%. Similarly, in 1999 the mean VSEI among the MTS groups was significantly (***p<0.001) higher (6.4%) than those of NsemSaw and by 2009 the gap between both groups both increased significantly (*p<0.001) from 6.7% to 8.8%. Between 1999 and 2009 the both MTS and Non-MTS groups experienced respectively 7.2% and 7.9% increases in mean VSEI and these changes according to the t-test results were significant (***p<0.001). Like the other two groups, NsemSaw also experienced a significant (**p<0.001) increase (5.5%) in mean VSEI between 1999 and 2009. 7.1.1.10 Household Item Liquidation Index (HHILI) In 1999 the mean Household Item Liquidation Index (HHILI) among NsemSaw households was significantly (***p<0.001) higher (8.9%) than those of MTS. By 2009 however the trends had reversed with MTS experiencing a 3% increase in HHILI than NsemSaw though this change was according to the t-test results was insignificant. Similarly, in 1999 the mean HHILI among NsemSaw was significantly (**p<0.01) higher (5.4%) than those of Non-MTS. By 2009 however the mean HHILI among Non-MTS was 2.1% higher than those of NsemSaw though this change was insignificant. Among all three groups only MTS experienced a significant (*p<0.05) within group increase (3.6%) in HHILI between 1999 and 2009. Both MTS and NsemSaw experienced respectively 0.8% and 8.3% decline in HHILI between 1999 and 2009 though only the change experienced by NsemSaw was significant (***p<0.001). 600 7.1.1.11 Combine HH Financial Capital Index (CHHFCI) In 1999 there was no statistically significant difference in the mean Combine HH Financial Capital Index (CHHFCI) among all three research groups. By 2009 however the MTS group experienced a significantly (*** p<0.001) higher (4.8%) mean CHHFCI relative to the NsemSaw group. Similarly, in 2009 the Non-MTS group also experienced a significantly (*** p<0.001) higher (4.1%) mean CHHFCI than NsemSaw. The estimated within group difference for MTS and Non-MTS suggests that both groups experienced significant (***p<0.001 and **p<0.01) increases (3.8% and 2.3%) in CHHFCI between 1999 and 2009 with the most profound change observed among the MTS group. Among all three groups however only NsemSaw experienced a decline (1.1%) in CHHFCI between 1999 and 2009 though this change according to the t-test was insignificant. Table 7.1.2 below summarizes the t-test statistics described above. Using a Difference- in-Difference technique chapter eight isolate changes in CHHFCI index that may be attributed directly to the MTS project or spillover from the project. Section 7.2 that follows discusses changes in human capital indexes among the three research groups. 601 7.1.2 Summary Results Among all three research groups only MTS in Yaya experienced improvements in all 11 financial capital indexes between 1999 and 2009 (see tables 7.1.1 and 7.1.2 above and 7.1.3 below). Between 1999 and 2009 the MTS group experienced significant improvements in seven out of the 11 indexes with the two most significant (*** p<0.001) improvements recorded for HHBAI (8.1%) and VSEI (7.1%). Of the 11 indexes the Non-MTS group in Yaya experienced improvements in eight with the most significant (*** p<0.001) improvement recorded for VSEI (7.9%) between 1999 and 2009. Unlike the MTS the Non-MTS group experienced a decline in three indexes (HHE2, AHHIS and HHILI) between 1999 and 2009 though these changes according to the t-test were insignificant. Between 1999 and 2009 the NsemSaw group experienced an increase in three out of the 11 financial capital indexes and out of these three the most significant (*** p<0.001) change was with VSEI (5.5%). For NsemSaw there was a general decline in eight financial capital indexes between 1999 and 2009 and of the eight the most significant (*** p<0.001) decline was experienced for the HHILI (8.3%) index. NsemSaw also experienced a significant (** p<0.01) decline in AHHIS (7.4%) index between 1999 and 2009. The discussions in this section suggest that between 1999 and 2009 the MTS group in Yaya experienced significantly higher improvements in financial capital indexes than Non-MTS and NsemSaw. Also, the Non-MTS group appears to perform better than NsemSaw for the same time period. Since the CHHFCI index captures changes in all the 10 different financial capital indexes (see table 7.1.1) changes in CHHFCI index among all three research groups is further analyzed in section 7.6 using livelihood pentagons. Following section 7.6, the question of what might have caused the MTS group to record significantly higher improvements in several financial capital indexes relative to both Non-MTS and NsemSaw is addressed in chapter eight. 602 Table 7.1.2: T-Test for Estimated Differences Between and Within Groups MTS HH in Yaya (n=203) Non-MTS HH in Yaya (n=120) NsemSaw (n=116) MTS and Non-MTS Estimated Estimated Differences Between Groups Differences Within Groups MTS in Yaya & NsemSaw Non-MTS in Yaya & NsemSaw MTS HH in Yaya Non- MTS in Yaya NsemSaw HH in Yaya M -0.026 (0.025) 0.011 (0.028) 0.001 (0.011) 0.007 (0.012) -0.019 (0.016) 0.003 (0.019) M 0.006 (0.025) 0.013 (0.028) M 0.032 (0.028) 0.024 (0.030) -0.084*** -0.084*** (0.011) (0.013) -0.055*** -0.062*** (0.013) (0.014) -0.015 (0.016) 0.018 (0.019) 0.037 (0.016) 0.018 (0.019) M M M -0.029 (0.023) -0.013 (0.029) -0.022 (0.029) -0.014 (0.010) -0.008 (0.012) 0.015 (0.015) -0.038* (0.016) -0.017 (0.017) -0.006 (0.017) -0.014 (0.010) 0.008 (0.012) 0.002 (0.010) -0.034** (0.011) -0.12 (0.012) -0.079*** -0.045*** (0.011) (0.011) -0.063*** -0.051*** (0.011) (0.012) FINANCIAL CAPITAL INDEX 1. PHISI Year M SD M SD M SD 2009 0.727 0.215 0.701 0.213 0.733 0.214 1999 0.699 0.250 0.688 0.241 0.711 0.226 2009 0.185 0.097 0.185 0.095 0.101 0.098 2. SHISI 1999 0.171 0.107 0.178 0.096 0.116 0.123 2009 0.508 0.141 0.490 0.122 0.493 0.129 3. HEI1 1999 0.470 0.185 0.473 0.137 0.488 0.031 2009 0.122 0.104 0.089 0.091 0.044 0.076 4. HHE2 1999 0.108 0.099 0.097 0.102 0.046 0.082 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) Standard Errors in parentheses; 6) *** p<0.001, ** p<0.01, * p<0.05 603 Table 7.1.2 (cont’d) Estimated Estimated Differences Between Groups Differences Within Groups MTS HH in Yaya (n=203) Non-MTS HH in Yaya (n=120) NsemSaw (n=116) MTS & Non-MTS MTS in Yaya & NsemSaw Non-MTS in Yaya & NsemSaw MTS HH in Yaya Non- MTS in Yaya NsemSaw M M M M M -0.144*** -0.145*** (0.038) -0.063 (0.037) (0.043) -0.040 (0.039) -0.036*** -0.036 ** (0.010) -0.019 (0.010) -0.029 (0.017) 0.025 (0.016) -0.019 (0.021) 0.092*** (0.023) (0.012) -0.016 (0.011) -0.023 (0.018) 0.022 (0.018) -0.002 (0.024) 0.048 (0.026) -0.081* (0.036) -0.104* (0.048) 0.000 (0.032) -0.012 (0.010) -0.015 (0.014) 0.005 (0.008) -0.035** (0.014) -0.024 (0.018) 0.020 (0.018) -0.038* (0.019) 0.025 (0.025) 0.074** (0.026) HH in Yaya M 0.001 (0.044) -0.022 (0.040) -0.000 (0.012) -0.003 (0.010) -0.006 (0.018) 0.004 (0.016) -0.017 (0.022) 0.044* (0.021) FINANCIAL CAPITAL INDEX 5. HHBAI Year M SD M SD M SD 2009 0.291 0.368 0.292 0.397 0.147 0.238 1999 0.209 0.351 0.188 0.337 0.147 0.256 2009 0.320 0.098 0.320 0.116 0.284 0.060 6. FHHBAI 1999 0.308 0.095 0.304 0.096 0.288 0.067 2009 0.406 0.158 0.400 0.153 0.377 0.119 7. TLSAI 1999 0.370 0.126 0.374 0.131 0.395 0.052 2009 0.541 0.187 0.524 0.197 0.522 0.177 8. AHHIS 1999 0.504 0.187 0.548 0.181 0.596 0.218 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) Standard Errors in parentheses; 6) *** p<0.001, ** p<0.01, * p<0.05 604 MTS HH in Yaya (n=203) Non-MTS HH in Yaya (n=120) NsemSaw (n=116) MTS & Non-MTS FINANCIAL CALIPTAL INDEX 9. VSEI Year M SD M SD M SD 2009 0.541 0.156 0.574 0.127 0.486 0.165 1999 0.467 0.154 0.494 0.141 0.430 0.147 2009 0.643 0.196 0.634 0.178 0.613 0.156 10. HHILI 1999 0.605 0.175 0.641 0.167 0.694 0.156 2009 0.428 0.084 0.421 0.075 0.380 0.071 11. CHHFCI 1999 0.391 0.087 0.398 0.072 0.391 0.070 HH in Yaya M 0.033* (0.017) 0.026 (0.017) -0.009 (0.022) 0.035 (0.019) -0.008 (0.009) 0.007 (0.009) -0.055** (0.019) -0.037* (0.018) -0.030 (0.022) 0.089*** (0.020) (0.019) -0.021 (0.022) 0.054** (0.021) -0.048*** -0.041*** (0.009) -0.000 (0.009) (0.009) -0.007 (0.009) Table 7.1.2 (cont’d) Estimated Estimated Differences Between Groups Differences Within Groups MTS in Yaya & NsemSaw Non-MTS in Yaya & NsemSaw MTS HH in Yaya Non-MTS in Yaya NsemSaw M M M M M -0.088*** (0.019) -0.072*** -0.079*** -0.064*** (0.015) (0.017) -0.055** (0.021) -0.036* (0.019) 0.008 (0.022) 0.083*** (0.021) -0.038*** (0.008) -0.023** (0.010) 0.011 (0.009) 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) Standard Errors in parentheses; 6) *** p<0.001, ** p<0.01, * p<0.05 605 Table 7.1.3: Summary of T-Test for Estimated Differences Between and Within Groups MTS & Non-MTS Mean (MTS – Non- No. Index MTS) MTS & NsemSaw Non & MTS-NsemSaw Mean (MTS – NsemSaw) Mean (Non-MTS – MTS Non-MTS NsemSaw NsemSaw) 1999 2009 1999 2009 1999 2009 2009-1999 2009- 1999 2009-1999 - +++ + +++ + + - - +++ -- + - +++ + +++ +++ ++ + + +++ + + + + + + + ++ + +++ + + + + - + + + + +++ - + - + - - - - -- ++ --- - - + 11. CHHFCI 1) Significant improvements in livelihood index: +++p<0.001, ++ p<0.01, + p<0.05, 2) Significant decline in livelihood index: ---p<0.001, --p<0.01, - p<0.05 3) Insignificant improvements in livelihood index: + and 4) Insignificant decline in livelihood index: - +++ +++ ++ 1. 2. 3. 4. 5. 6. 7. 8. 9. PHISI SHISI HEI1 HHE2 HHBAI FHHBAI TLSAI AHHIS VSEI 10. HHILI - - - + + - - - + - + - + ++ - - + + + + - +++ - +++ + - - --- + --- + - +++ + +++ +++ +++ + + ++ + +++ 606 7.2 Human Capital This section discusses changes in 18 human capital indexes among the three research groups (MTS, Non-MTS and NsemSaw) before (1999) and after (2009) the MTS project. Human capital within the context of this research may be described as the “sum of individual or household’s skills, experiences; knowledge, abilities and health status” (see chapter three sections 3.3.5 for a broader discussion of the literature on human capital and how it helps shapes rural households and livelihoods). The 18 human capital indexes described in this section encompass four broad categories including a) formal education and literacy, b) migration, c) dietary diversity and c) health status of the research groups. Household Education and Literacy Index (EDULIT) index for example captures the level of households’ education and literacy while Household Migration Index (MIGI) captures both short and long-term migration patterns of household members out of their immediate communities in search for work. Human capital indexes three through 14 describe dietary patterns and frequency of consumption and sufficiency of different food groups while indexes 15 (Household Health and Disease Index-HHDI) and 16 (Household Mortality Index-HHMI) describes the general health status of households within each research group. The last two human capital indexes (Combined Household Human Capital Index 1 and 2) represent the aggregate of all 16 human capital indexes. Thus table 7.2.1 below describes how each of the 18 human capital indexes were generated while table 7.2.2 presents the results of a Two Sample T-test used to track (within and between group) changes in each of the 18 indexes before (1999) and after (2009) the MTS project. The tables presented in Appendix P describes in greater detail how each human capital index was constructed from the household survey results. 607 No. Index Name of Index/ Variable Household Education 1 EDULIT and Literacy Index 2 MIGI Household Migration Index Table 7.2.1: Definition of Household Human Capital Indexes Description of Indexes Computation of Indexes EDULIT captures both formal and informal education levels within a household. A HH attains a score of 1 each if at least an individual attended “Day Nursery, Kindergarten or Primary (EDUCATION1),” “JSS or SSS (EDUCATION2) or can read and write English or any local Ghanaian language (LITERACY). The total attainable score for Education and Literacy levels of a HH is thus 3 and the EDULIT index expressed in terms of level of education and literacy ranges from 0 to 1. MIGI captures short and long-term HH migration activities. Migration for work is viewed as a risk mitigating strategy against crop failure in the research communities. Hence HHs with at least one individual migrating on a daily basis for work attains a score of 1. Also, if at least a HH member migrates for work for more than three months (i.e. long term) attain a score of 1. The total attainable score for migration is thus 2. MIGI index expressed in terms of level of HH migration ranges from 0 to 1. 608 3LITERACY)EDUCATION21(EDUCATIONEDULIT31tit==++=ni2ode)MIGWK_2reccode(MIGWK_1reMIGI21itt==+=ni No. Index Name of Index/ Variable 3 FSFCI Staple Food Consumption Frequency Index Fruits and Vegetables 4 FFVCI Consumption Frequency Index Table 7.2.1 (cont’d) Description of Indexes Computation of Indexes FSFCI capture the frequency with which three categories of major staples diets are consumed within the household. Foods made from cereals such as corn and rice constitutes one group while those from root crops such as cassava and yam also constitute another. Foods with an orange color such as sweet potato also form another group. HHs that always eat food items from any of the three groups attains a score of 3 for the group while those that “Never” eat foods from a particular group scores 0 for the group. The total attainable score for staple food consumption is 9 and FSFCI index expressed in terms of level of HH staple food consumption ranges from 0 to 1. FFVCI capture the frequency with which fruits and vegetables are consumed. Households that eat dark green leafy vegetables such as potato, cocoyam or cassava leaves at least once a week score 3npoints on this indicator. Consuming vitamin A rich fruits such as ripe mangoes or papaya at least once a week also earns the household 3 points. Consuming any other vegetables and fruits at least once a week also earns households 3 points each for the two different categories. HH that eat any food group at most once in a year (Never) scores 0 for the group. The total attainable score for fruit and vegetable consumption is 12 and FFVCI index expressed in terms of level of HH fruits and vegetable consumption ranges from 0 to 1. 609 9 Freq)HHDD3_ +HHDD2_Freq +q(HHDD1_FreFSFCI31itt===ni12 Freq)HHDD7_ + FreqHHDD6_ + FreqHHDD5_ + Freq(HHDD4_FFVCI41itt===ni No. Index Name of Index/ Variable 5 FPAPCI Plants and Animal Protein Consumption Frequency Index 6 FOFC Other Foods Consumption Frequency Index (FOFC) Table 7.2.1 (cont’d) Description of Indexes Computation of Indexes FPAPC capture the frequency with which plant and animal proteins are consumed in the household. Four groups of animal proteins (meat, egg, fish, and milk products) and plant protein (leguminous seeds) were included used in the computation of FPAPC index. Households that consume any one of the protein groups at least once a week (Always) scores 3 points for that particular group and 0 if they consume only once a year (Never). The total attainable score for the five plants and animal protein groups is thus 15 and the minimum is 0. FPAPC index expressed in terms of level of HH protein consumption ranges from 0 to 1. FOFC capture the frequency with which foods labeled broadly as “Other Foods” are consumed within the HH. Other Foods consist of three different food groups: (foods made from oil, fat, or butter) or (sugar or honey) or (cocoa, coffee or tea). Households that consume any of the three groups at least once a week (Always) scores 3 points for that particular indicator group and 0 if consumed only once a year (Never). A household’s maximum attainable score on FOFC is thus 9 and the minimum is 0. FPAPC index as a measure of the level of consumption of the “Other Foods” ranges from 0 to 1. 610 15 Freq)HHDD12_ + FreqHHDD11_ + FreqHHDD10_ + Freq HHDD9_+ Freq(HHDD8_FPAPC51itt===ni9 Freq)HHDD15_ + FreqHHDD14_ + Freq(HHDD13_FOFC31itt===ni No. Index 7 DDAFC 8 CHHDDI1 Name of Index/ Variable Dietary Diversity Aggregate Frequency of Consumption Index (DDAFC) Combined HH Dietary Diversity Index for Trend & Sufficiency of Consumption (CHHDDI1) DDAFC capture a HH’s aggregate frequency of consumption of the 15 individual food items that make up the four food groups described above (Staple Food, Fruits and Vegetables, Plant and Animal Proteins, and Other Foods). Depending on the frequency with each of the 15 food items is consumed, a HH’s DDAFC as a measure of a household’s aggregate frequency of consumption ranges between 0 and 45 points and the DDAFC index also ranging between 0 and 1. CHHDDI1 is very similar to DDAFC above however the only difference is that instead of including the individual 15 food items in the computation, CHHDDI1 is computed from the aggregate values obtained for the four major food groups (Staple Food, Fruits and Vegetables, Plant and Animal Proteins, and Other Foods). Since it is almost impossible as human beings to survive without eating from any one of the four major food groups, the lowest attainable score on CHHDDI1 is greater than 0. The highest attainable score for households with a highly diversified diet is closer to 4. 611 Table 7.2.1 (cont’d) Description of Indexes Computation of Indexes 45eq) HHDD15_Fr+... +q(HHDD1_FreDDAFC151itt===ni4 FOFC)+FPAPC +FFVCI +(FSFCICHHDDI141itt===ni No. Index Name of Index/ Variable Table 7.2.1 (cont’d) Description of Indexes Computation of Indexes SFSI captures changing trends in the level of sufficiency of the three categories of major staples diets available to the household. Foods made from cereals such as corn and rice constitutes one group while those from root crops such as cassava and yam also constitute another. Foods with an orange color such as sweet potato also form another group. For each food group, households that experience an increasing and sufficient levels of availability score 5 for the particular group while a “Declining and Insufficient” levels of supply or availability earns the HH a score of 1. Between the two extreme sufficiency scores are three other levels: Increasing but not Sufficient (2), Decreasing but Sufficient (3) and Stable but Sufficient (4). The total attainable SFS score is thus 15 and the lowest is 3. The SFSI index expressed in terms of level of HH staple food sufficiency ranges from 0.2 to 1. 9 SFSI Staple Foods Sufficiency Index (SFSI) 612 15d)HHDD3_Tren +nd HHDD2_Tre+nd(HHDD1_TreSFSI31itt===ni No. Index Name of Index/ Variable Table 7.2.1 (cont’d) Description of Indexes Computation of Indexes FVSI captures changing trends in the level of consumption and sufficiency of four categories of fruits and vegetables available to the household. Dark, green, leafy vegetables such as cassava leaves, bean leaves and spinach constitute one group while any other vegetable constitutes another group. Food rich in vitamin A for example ripe mangoes and papaya form another group while any other fruits consumed within the household constitute a fourth group. For each food group, households that experience an increasing and sufficient level of availability score 5 for the particular group while a “Declining and Insufficient” levels of supply or availability earns the HH a score of 1. Between the two extreme sufficiency scores are three other levels: Increasing but not Sufficient (2), Decreasing but Sufficient (3) and Stable but Sufficient (4). The total attainable FVSI score is thus 20 and the lowest is 4. The FVSI index expressed in terms of level of HH staple food sufficiency ranges from 0.2 to 1. 613 10 FVSI Fruits and Vegetables Sufficiency Index (FVSI) 20d)HHDD7_Tren +dHHDD6_Tren +nd HHDD5_Tre+nd(HHDD4_TreFVSI31itt===ni No. Index Name of Index/ Variable 11 PAPSI Plants and Animal Protein Sufficiency Index (PAPSI) 12 OFSI Other Foods Sufficiency Index (OFSI) Table 7.2.1 (cont’d) Description of Indexes Computation of Indexes PAPSI captures changing trends in the level of consumption and sufficiency of plant and animal proteins the household. Four groups of animal (meat, egg, fish, and milk products) and plant protein (leguminous seeds) were included in the computation of PAPSI index. For each of the five protein groups, HHs that experience an increasing and sufficient level of availability score 5 for the particular group while a “Declining and Insufficient” levels of supply or availability translates into a score of 1 for the HH. The total attainable PAPSI score is thus 25 and the lowest is 5. The PAPSI index expressed in terms of level of HH plant and animal protein sufficiency ranges from 0.2 to 1. OFSI capture changing trends in the level of consumption and sufficiency of foods broadly labeled as “Other Foods.” Other Foods consist of three different food groups: (foods made from oil, fat, or butter) or (sugar or honey) or (cocoa, coffee or tea). HHs that experience an increasing and sufficient level of availability score 5 for the particular group while a “Declining and Insufficient” levels of supply or availability translates into a score of 1. The maximum attainable OFSI score is thus 15 and the lowest is 3. The OFSI index expressed in terms of level of HH “Other Foods” sufficiency ranges from 0.2 to 1. 614 25nd)HHDD12_Tre +ndHHDD11_Tre+ndHHDD10_Tre +nd HHDD9_Tre+nd(HHDD8_TrePAPSI51itt===ni15nd)HHDD15_Tre +ndHHDD14_Tre +end(HHDD13_TrOFSI31itt===ni No. Index Name of Index/ Variable Dietary Diversity Aggregate 13 DDAFS Food Sufficiency Index (DDAFS) 14 CHHDDI2 Combined HH Dietary Diversity Index -Trend in Sufficiency of Consumption (CHHDDI2) Table 7.2.1 (cont’d) Description of Indexes Computation of Indexes DDAFS is similar to Dietary Diversity Aggregate Frequency of Consumption Index (DDAFC). However instead of a HH’s aggregate frequency of consumption, DDAFS capture changing trends in the level of consumption and sufficiency of the 15 food items that make up the four food groups (Staple Food, Fruits and Vegetables, Plant and Animal Proteins, and Other Foods). Hence depending on the trend in consumption and level of sufficiency of a particular food item within the HHs, a HH may obtain a DDAFS score between 15 and 75. The DDAFC index as a measure of a household’s aggregate captures changing trends in the level of consumption and sufficiency of the 15 food items ranges between 0.2 and 1. CHHDDI2 is similar to DDAFS above however unlike DDAFS that includes individual 15 food items in the computation, CHHDDI2 is computed from the aggregate values obtained for the four major food groups (Staple Food, Fruits and Vegetables, Plant and Animal Proteins, and Other Foods). Since it is almost impossible as human beings to survive without eating from any one of the four major food groups, the lowest attainable score on CHHDDI2 is 0.8 (i.e., greater than 0). The highest attainable score for households with a highly diversified diet is 4. 615 75end) HHDD15_Tr+... +nd(HHDD1_TreDDAFS151itt===ni4) OFSI + PAPSI+ FVSI+(SFSICHHDDI241itt===ni No. Index Name of Index/ Variable 15 HHDI Household Health and Disease Index (HHDI) 16 HHMI Household Mortality Index (HHMI) Table 7.2.1 (cont’d) Description of Indexes Computation of Indexes HHDI captures the frequency of illness and disease within the household. In order to adequately capture trends in health and disease within a household, HHDI focuses on three different age categories of household members: individuals 6 years or younger, individuals between 6 and 12 and individuals 12 years or older. Household’s that reported individual “Always” sick and unable to perform daily functions such as going to school, or work in the case of adults or bed-ridden or not able to eat in the case of children younger than 6 years score 1. Household that reported that their members fall sick at most once in a year (Never) score 4. The highest attainable HHDI score is thus 12 and the lowest is 3. HHDI index as a measure of a HH’s health and disease status thus ranges from 0.25 to 1. HHMI captures the incidence of mortality and physical disability within the household. HHMI reveals mortality among two age categories, children younger than 6 years and individual between 7 and 57 (i.e. life expectancy of the average Ghanaian at the time of the study). HHMI also reveal captures physical disability among all individuals within the HH (regardless of age). Households with more than 4 individuals diseased score 1 and those with no deaths score 4. Similarly, HHs with more 4 handicapped or disabled individuals of any age score 1 and those with none score 4. The highest attainable HHMI score is thus 12 and the lowest is 3. HHMI index as a measure of a HH’s mortality and disability index ranges from 0.25 to 1. 616 12ID_3) +ID_2 +(ID_1HHDI31itt===ni12DM_3) +DM_2 +(DM_1HHMI31itt===ni Table 7.2.1 (cont’d) Description of Indexes Computation of Indexes No. Index 17 CHCI1 Name of Index/ Variable Combined Household Human Capital Index1 CHCI1 captures a HH’s aggregate or combined human capital endowment. This index is simply the average of six human capital indexes (Education and Literature, Migration, Dietary Diversity Aggregate Frequency of Consumption, Dietary Diversity Aggregate Food Sufficiency, Household Health and Disease and Household Mortality Indices). The maximum attainable score on a household’s combined Human Capital endowment is thus 6 and the minimum is 0. CHCI1 index thus ranges from 0 to 1. Combined Household Human Capital Index2 CHCI2 is very similar to CHCI1 however the only difference is that in CHCI2, DDAFC and DDAFS are replaced with Combined HH Dietary Diversity Index for Trend & Frequency of Consumption (CHHDDI1) and Combined Household Dietary Diversity Index for Trend in Sufficiency of Consumption (CHHDDI2). Just like CHCI1, a HH’s maximum attainable CHCI2 score is 6 and the minimum is 0. CHCI2 index also ranges from 0 to 1. 18 CHCI2 617 ]6[] HHMI+HHDI + DDAFS+DDAFC +MIGI+[EDULITCHCI1t=]6[] HHMI+ HHDI+CHHDDI2+CHHDDI1 + MIGI+[EDULITCHCI2t= 7.2.1 Two Sample T-test of between and within group differences in 1999 and 2009 Section 7.2.1 investigates between and within group differences for all 18 human capital indexes described in tables 7.2.1 above. The Two-Sample t-test determined changes that occurred in household human capital indexes between 1999 and 2009 and also the degree to which these changes (between and within group) were significant using three different confident levels 99.99% (*** p<0.001), 99% (**p<0.01) and 95% (*p<0.05). Within group difference for each livelihood index is computed by subtracting the mean result obtained for that particular index in 2009 from the mean in 1999 for the same research group (MTS households in Yaya or non-MTS groups in Yaya). Since the Two-Sample t-test is computed by differencing the mean value of the human capital index of the current year (2009) from the initial year (1999), a positive difference suggests a decline in livelihood index while a negative difference suggests an improvement. The computation for between group differences is similar to within group differences only in the former, the changes were computed for two different groups for a particular time period either 1999 or 2009. 7.2.1.1 Household Education and Literacy Index (EDULIT) In 1999 and 2009 there was no statistically significant difference in the mean Household Education and Literacy Index (EDULIT) between MTS and Non-MTS research groups. However, in 1999 the EDULIT index among MTS was significantly (*** p<0.001) higher (15.6%) than NsemSaw with the gap increasing to 16% in 2009. The gap in EDULIT index between Non-MTS in Yaya and NsemSaw was wider in 1999 with the Non-MTS group experiencing a significantly (*** p<0.001) higher (20.3%) mean EDULIT index than NsemSaw. In 2009 EDULIT index among Non-MTS households was 19.7% higher than NsemSaw and the difference according to the t-test was still significant (*** p<0.001). Though all three research 618 groups experienced a general increase in their EDULIT index between 1999 and 2009, only the change (9.2%) experienced among the MTS group was significant (* p<0.05). 7.2.1.2 Household Migration Index (MIGI) In 1999 and 2009 there was no statistically significant difference in mean Household Migration Index (MIGI) between MTS and Non-MTS. In 1999 the MIGI index among MTS was significantly (*** p<0.001) higher (9.2%) than that of NsemSaw. By 2009 however the gap in MIGI between MTS and NsemSaw increased significantly (*** p<0.001) from 9.2% to 16.5%. Similarly in 1999 the mean MIGI index among the Non-MTS group was significantly (*** p<0.001) higher (10.7%) than NsemSaw and by 2009 the observed difference between both groups increased significantly (*** p<0.001) from 10.7% to 19.5%. All three research groups experienced an increased in their MIGI indexes between 1999 and 2009 though these changes were significant only for the MTS and the Non-MTS groups. Between 1999 and 2009 the MTS group experienced a significant (** p<0.01) increase (8.1%) in MIGI while the Non-MTS group also experienced a significant (*p<0.05) increase (9.6%) between the same time period. The relatively larger standard error associated with the mean MIGI value obtained for the Non-MTS (4.2%) relative to MTS (3.1%) explains the discrepancy in p-values obtained for both groups. 7.2.1.3 Staple Food Consumption Frequency Index (FSFCI) In 1999 there was no significant difference in Staple Food Consumption Frequency Index (FSFCI) among all three research groups. While the difference in FSFCI between MTS and Non- MTS still remained insignificant in 2009, both MTS and Non-MTS experienced significantly (** p<0.01) higher (5.1%) increases in FSFCI relative to NsemSaw. While all the three research groups experienced a general decline in their FSFCI indexes between 1999 and 2009, only the 619 NsemSaw group experienced a significantly (*p<0.05) higher increase (9.6%) in FSFCI for the same time period. 7.2.1.4 Fruits and Vegetables Consumption Frequency Index (FFVCI) In 1999 the mean Fruits and Vegetables Consumption Frequency Index (FFVCI) among the MTS group was significantly (*** p<0.001) higher (5.5%) than that of NsemSaw. By 2009 however the gap in FFVCI between both groups declined to insignificant levels. Similarly, in 1999 Non-MTS households experienced a significantly (*** p<0.001) higher (6.2%) mean FFVCI relative to NsemSaw. In spite of the decline in the FFVCI gap between both groups the mean FFVCI among the Non-MTS group was still significantly (* p<0.05) higher (4.0%) than those of NsemSaw in 2009. Among all three groups the NsemSaw experienced the highest increase of 5.7% in FFVCI between 1999 and 2009. Non-MTS and MTS groups respectively experienced increases of 5.5% and 5.4% in FFVCI over the ten-year period. The changes observed among all three research groups between 1999 and 2009 were significant (*** p<0.001). 7.2.1.5 Plant and Animal Protein Consumption Frequency Index (FPAPC) In 1999 there was no significant difference in the mean Plant and Animal Protein Consumption Frequency Index (FPAPC) among all three research groups. While the difference in FPAPC between MTS and Non-MTS still remained insignificant in 2009, both MTS and Non- MTS experienced significantly (*** p<0.001) higher (5.8%) increases in FPAPC relative to NsemSaw. Between 1999 and 2009 all the three research groups experienced an increase in their FPAPC indexes however the highest (7.0%) and significant (*** p<0.001) increase in mean 620 FPAPC occurred among the MTS group. Like the MTS group, the Non-MTS also experienced a significant (**p<0.01) increase (6.4%) in mean FPAPC between 1999 and 2009. 7.2.1.6 Other Foods Consumption Frequency Index (FOFC) In 1999 the mean Other Foods Consumption Frequency Index (FOFC) among the MTS group was significantly (*** p<0.001) higher (12.6%) than that of NsemSaw. Similarly in 1999 the mean FOFC index among Non-MTS households was significantly (*** p<0.001) higher (12.3%) that that of NsemSaw. By 2009 both MTS and Non-MTS groups still maintained significantly (*** p<0.001) higher (11.9%) mean FOFC relative to the NsemSaw group. Among all three groups the NsemSaw experienced the highest and significant (*** p<0.001) increase (11.4%) 5.7% in FOFC index. Between 1999 and 2009 the FOFC index among MTS and Non- MTS groups increased significantly (*** p<0.001) by 10.7% and 10.9% respectively. 7.2.1.7 Dietary Diversity Aggregate Frequency of Consumption Index (DDAFC) In 1999 there was no significant difference in the mean Dietary Diversity Aggregate Frequency of Consumption Index (DDAFC) among all three research groups. While the difference in mean DDAFC index between MTS and Non-MTS remained insignificant in 2009, the DDAFC index among the MTS group was significantly (***p<0.001) higher (5.9%) than that of NsemSaw. Similarly, in 2009 the mean DDAFC index among the Non-MTS group was significantly (***p<0.001) higher (6.4%) than that of NsemSaw. Among all the research groups only NsemSaw experienced a decline in DDAFC index between 1999 and 2009 though this change like those experienced among MTS and Non-MTS was insignificant. 621 7.2.1.8 Combined Household Dietary Diversity Index for Trend & Sufficiency of Consumption (CHHDDI1) In 1999 the mean CHHDDI1 index among the MTS group was significantly (*p<0.05) higher (4.1%) than that of NsemSaw and by 2009 the gap in CHHDDI1 between both groups increased significantly (***p<0.001) from 4.1% to 6.3%. Like the MTS, the Non-MTS group experienced a significantly (**p<0.01) higher (5.0%) CHHDDI1 index than NsemSaw in 1999 and by 2009 the difference between both groups increased significantly (***p<0.001) from 5.0% to 6.7%. All three research groups experienced an increased in their CHHDDI1 indexes between 1999 and 2009 though these changes were significant only for the MTS and the Non-MTS groups. Between 1999 and 2009 the MTS group experienced a significant (***p<0.001) increase (17.5%) in CHHDDI1 while the Non-MTS group also experienced a significant (***p<0.001) increase (17.1%) between the same time period. 7.2.1.9 Staple Foods Sufficiency Index (SFSI) In 1999 there was no significant difference in the mean Staple Foods Sufficiency Index (SFSI) among all three research groups. While the difference in mean SFSI index between MTS and Non-MTS remained insignificant in 2009, the SFSI index among the MTS group was significantly (***p<0.001) higher (8.8%) than that of NsemSaw. Similarly in 2009 the mean SFSI among the Non-MTS group was significantly (*p<0.05) higher (5.8%) than that of NsemSaw. Between 1999 and 2009, all three groups experienced a decline in their mean SFSI though the largest (9.1%) and significant decline occurred among the NsemSaw group. While the Non-MTS group also experienced a significant (*p<0.05) decline (5.4%) in SFSI index between 1999 and 2009 the change among MTS was insignificant. 622 7.2.1.10 Fruits and Vegetables Sufficiency Index (FVSI) The Fruits and Vegetables Sufficiency Index (FVSI) appears to be the only human capital index for which there was no statistically significant difference between all three research groups in 1999 and 2009. Though all three groups experienced a decline in FVSI index between 1999 and 2009 these changes were not statistically significant. 7.2.1.11 Plant and Animal Protein Sufficiency Index (PAPSI) In 1999 the mean Plant and Animal Protein Sufficiency Index (PAPSI) among the MTS group was significantly (**p<0.001) higher (5.7%) than that of NsemSaw and by 2009 the gap in PAPSI between both groups increased significantly (***p<0.001) to 9.4%. Like the MTS, the Non-MTS group also experienced a significantly (*p<0.05) higher (5.2%) PAPSI index than NsemSaw in 1999 and by 2009 the difference between both groups increased significantly (***p<0.001) to 8.9%. All the three research groups experienced a general decline in PAPSI between 1999 and 2009 though these changes according to the t-test were insignificant. 7.2.1.12 Other Foods Sufficiency Index (OFSI) In 1999 the mean Other Foods Sufficiency Index (OFSI) among the MTS group was significantly (**p<0.001) higher (5.6%) than that of NsemSaw and by 2009 the gap in OFSI between both groups increased significantly (***p<0.001) from to 10.6%. Like the MTS, the Non-MTS group also experienced a significantly (*p<0.05) higher (6.6%) OFSI index than NsemSaw in 1999 and by 2009 the difference between both groups increased to12.5% and was significant (***p<0.001). Among all the three research groups only NsemSaw experienced a decline in OFSI between 1999 and 2009 though these change according to the t-test was 623 insignificant. Between 199 and 2009 both the MTS and Non-MTS experienced an increase in their OFSI index however these changes were also insignificant. 7.2.1.13 Dietary Diversity Aggregate Food Sufficiency Index (DDAFS) In In 1999 there was no statistically significant difference in the mean Dietary Diversity Aggregate Food Sufficiency Index (DDAFS) among all three research groups. While the difference in mean DDAFS index between MTS and Non-MTS remained insignificant in 2009, the MTS group experienced a significantly (***p<0.001) higher (7.8%) DDAFS index than that of NsemSaw. Similarly, in 2009 the mean DDAFS index among the Non-MTS group was significantly (***p<0.001) higher (6.9%) than that of NsemSaw. While all three research groups experienced a general decline in their DDAFS index between 1999 and 2009 only the change (4.8%) among the NsemSaw group was significant (*p<0.05). 7.2.1.14 Combined Household Dietary Diversity Index for Trend in Sufficiency of Consumption (CHHDDI2) In 1999 there was no statistically significant difference in the mean Combined Household Dietary Diversity Index for Trend in Sufficiency of Consumption (CHHDDI2) among all three research groups. While the difference in mean CHHDDI2 index between MTS and Non-MTS remained insignificant in 2009 that between the MTS and NsemSaw was not. By 2009 the MTS group experienced a significantly (***p<0.001) higher (7.9%) CHHDDI2 index than NsemSaw. Similarly in 2009 the mean CHHDDI2 index among the Non-MTS group was significantly (***p<0.001) higher (7.1%) than that of NsemSaw. While all three research groups experienced a general decline in their CHHDDI2 index between 1999 and 2009 only the change (5.1%) among the NsemSaw group was significant (*p<0.05). 624 7.2.1.15 Household Health and Disease Index (HHDI) In 1999 there was no statistically significant difference in the mean Household Health and Disease Index (HHDI) among all three research groups. By 2009 however the MTS group experienced a significantly (*p<0.05) higher (4.2%) HHDI index than the Non-MTS group. Similarly, in 2009 the mean HHDI index among the MTS group was significantly (***p<0.001) higher (13.0%) than that of NsemSaw. By 2009 the Non-MTS group also experienced a significantly (**p<0.01) higher (8.7%) HHDI index than NsemSaw. Among all the three research groups only NsemSaw experienced a decline in HHDI between 1999 and 2009 though these change according to the t-test was insignificant. Between 199 and 2009 both the MTS and Non-MTS experienced an increase in their HHDI index however these changes were also insignificant. 7.2.1.16 Household Mortality Index (HHMI) In 1999 the mean Household Mortality Index (HHMI) among NsemSaw was significantly (*p<0.05) higher (2.2%) than that of the MTS group and by 2009 the gap in between both groups increased significantly (**p<0.01) from to 2.9%. In 1999 NsemSaw experienced a higher (1.4%) mean HHMI index than the Non-MTS group however this difference was not statistically significant. By 2009 however the difference in mean HHMI between both groups increased from 1.4% to 2.2% and this difference was significant (*p<0.05). Among all the three research groups only NsemSaw experienced an increase (albeit insignificantly) in HHMI index between 1999 and 2009. 625 7.2.1.17 Combined Household Human Capital Index1 (CHCI1) In 1999 the Combined Household Human Capital Index1 (CHCI1) among MTS was significantly (*** p<0.001) higher (5.5%) than that of NsemSaw. By 2009 however the gap in CHCI1 between both groups increased significantly (*** p<0.001) from 5.5% to 9.4%. Similarly, in 1999 the mean CHCI1 index among the Non-MTS group was significantly (*** p<0.001) higher (5.8%) than NsemSaw and by 2009 the observed difference between both groups increased significantly (*** p<0.001) from 5.8% to 9.8%. Among all three research groups, only NsemSaw experienced a decline (albeit insignificantly) in CHCI1 index between 1999 and 2009. Between 1999 and 2009 the MTS groups experienced a significant (**p<0.01) increase (3.6%) in CHCI1 index. The change in CHCI1 among the Non-MTS group between 1999 and 2009 was similar to the MTS group. 7.2.1.18 Combined Household Human Capital Index1 (CHCI2) In 1999 the Combined Household Human Capital Index2 (CHCI2) among MTS was significantly (*** p<0.001) higher (5.6%) than that of NsemSaw and by 2009 the gap increased significantly (*** p<0.001) to 9.5%. Similarly, in 1999 the mean CHCI2 index among the Non- MTS group was significantly (*** p<0.001) higher (6.3%) than NsemSaw and by 2009 the observed difference increased significantly (*** p<0.001) to 9.9%. While all the three research groups experienced an increase in CHCI2 index between 1999 and 2009 only the changes experienced among MTS (6.3%) and Non-MTS (6.0%) groups were significant (***p<0.001). Table 7.2.2 below presents a summary of the t-test results discussed under sections 7.2.1.1 through 7.2.1.18 above. 626 Table 7.2.2: T-Test for Estimated Differences in Human Capital Indexes Between and Within Groups MTS HH in Yaya (n=203) Non-MTS HH in Yaya (n=120) NsemSaw (n=116) MTS & Non-MTS HUMAN CAPITAL INDEX 1. EDULIT Year M SD M SD M SD 2009 0.489 0.403 0.525 0.405 0.328 0.362 1999 0.397 0.394 0.444 0.417 0.241 0.319 2009 0.200 0.335 0.229 0.348 0.034 0.158 2. MIGI 1999 0.118 0.278 0.133 0.295 0.026 0.145 2009 0.670 0.148 0.669 0.140 0.619 0.169 3. FSFCI 1999 0.680 0.242 0.705 0.209 0.714 0.217 2009 0.706 0.146 0.723 0.153 0.683 0.142 4. FFVCI 1999 0.171 0.107 0.178 0.096 0.116 0.123 Estimated Estimated Differences Between Groups Differences Within Groups MTS in Yaya & NsemSaw Non-MTS in Yaya & NsemSaw MTS HH in Yaya Non-MTS in Yaya NsemSaw M M M M M -0.162*** -0.197*** (0.045) (0.050) -0.156*** -0.203*** (0.043) (0.048) -0.092* (0.040) -0.081 (0.053) -0.086 (0.045) -0.165*** -0.195*** (0.033) (0.035) -0.092*** -0.107*** (0.028) (0.030) -0.051** (0.018) 0.033 (0.027) -0.023 (0.017) -0.051** (0.020) 0.009 (0.028) -0.040* (0.019) -0.081** (0.031) -0.096* (0.042) -0.009 (0.020) 0.011 (0.020) 0.036 (0.023) 0.096*** (0.026) -0.536*** -0.545*** -0.567*** -0.055*** -0.062*** (0.013) (0.016) (0.017) (0.013) (0.014) HH in Yaya M 0.036 (0.046) 0.047 (0.046) 0.030 (0.039) 0.015 (0.033) -0.001 (0.017) 0.024 (0.027) 0.017 (0.017) 0.007 (0.012) 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) Standard Errors in parentheses; 6) *** p<0.001, ** p<0.01, * p<0.05 627 Table 7.2.2 (cont’d) Estimated Estimated Differences Between Groups Differences Within Groups MTS HH in Yaya (n=203) Non-MTS HH in Yaya (n=120) NsemSaw (n=116) MTS & Non-MTS MTS in Yaya & NsemSaw Non-MTS in Yaya & NsemSaw MTS HH in Yaya Non-MTS in Yaya NsemSaw M M M M M -0.058*** (0.017) -0.016 (0.024) -0.058** (0.019) -0.023 (0.025) -0.119*** -0.119*** -0.070*** (0.019) -0.064** (0.023) -0.028 (0.022) (0.023) (0.024) -0.107*** -0.109*** -0.114*** -0.126*** -0.123*** (0.023) (0.028) (0.031) (0.032) (0.034) -0.059*** -0.064*** (0.014) -0.029 (0.022) (0.015) -0.019 (0.023) -0.063*** -0.067*** (0.014) -0.041* (0.018) (0.015) -0.050** (0.018) -0.021 (0.016) -0.036 (0.019) 0.008 (0.020) -0.175*** -0.171*** (0.014) (0.017) -0.154 (0.017) HH in Yaya M -0.000 (0.017) 0.007 (0.025) 0.000 (0.021) -0.003 (0.030) 0.004 (0.013) -0.011 (0.021) 0.004 (0.014) 0.009 (0.018) HUMAN CAPITAL INDEX 5. FPAPC Year M SD M SD M SD 2009 0.705 0.147 0.705 0.140 0.647 0.156 1999 0.635 0.224 0.642 0.208 0.619 0.176 6. FOFC 2009 0.687 0.194 0.688 0.172 0.569 0.202 1999 0.581 0.271 0.578 0.248 0.455 0.271 7. DDAFC 2009 0.695 0.120 0.699 0.112 0.635 0.118 1999 0.674 0.190 0.663 0.174 0.645 0.175 2009 0.692 0.121 0.696 0.113 0.630 0.120 8. CHHDDI1 1999 0.517 0.167 0.526 0.145 0.476 0.137 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) Standard Errors in parentheses; 6) *** p<0.001, ** p<0.01, * p<0.05 628 Table 7.2.2 (cont’d) Estimated Estimated Differences Between Groups Differences Within Groups MTS HH in Yaya (n=203) Non-MTS HH in Yaya (n=120) NsemSaw (n=116) MTS & Non-MTS MTS in Yaya & NsemSaw Non-MTS in Yaya & NsemSaw MTS HH in Yaya Non- MTS in Yaya NsemSaw M -0.088*** (0.020) -0.013 (0.025) -0.028 (0.021) -0.008 (0.023) M -0.058* (0.024) -0.021 (0.029) -0.011 (0.026) -0.002 (0.026) -0.094*** -0.089*** (0.019) -0.057** (0.022) (0.024) -0.052* (0.025) -0.106*** -0.125*** (0.023) -0.056* (0.025) (0.027) -0.066* (0.029) M M M 0.015 (0.019) 0.054* (0.026) 0.091*** (0.027) 0.001 (0.018) 0.013 (0.025) 0.022 (0.027) 0.005 (0.017) 0.005 (0.025) 0.042 (0.024) -0.002 (0.019) -0.010 (0.026) 0.048 (0.031) HH in Yaya M -0.030 (0.021) 0.008 (0.023) -0.017 (0.021) -0.006 (0.022) -0.005 (0.021) -0.005 (0.021) 0.019 (0.022) 0.010 (0.023) HUMAN CAPITAL INDEX 9. SFSI Year M SD M SD M SD 2009 0.621 0.171 0.591 0.197 0.533 0.167 1999 0.636 0.204 0.644 0.200 0.624 0.237 10. FVSI 2009 0.700 0.173 0.684 0.192 0.672 0.203 1999 0.702 0.194 0.696 0.193 0.694 0.213 11. PAPSI 2009 0.593 0.168 0.588 0.197 0.499 0.165 1999 0.598 0.178 0.593 0.184 0.541 0.201 12. OFSI 2009 0.551 0.192 0.570 0.199 0.445 0.216 1999 0.549 0.198 0.559 0.202 0.494 0.249 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) Standard Errors in parentheses; 6) *** p<0.001, ** p<0.01, * p<0.05 629 Table 7.2.2 (cont’d) Estimated Estimated Differences Between Groups Differences Within Groups MTS HH in Yaya (n=203) Non-MTS HH in Yaya (n=120) NsemSaw (n=116) MTS & Non-MTS MTS in Yaya & NsemSaw Non-MTS in Yaya & NsemSaw MTS HH in Yaya Non- MTS in Yaya NsemSaw M M M M M -0.078*** -0.069*** (0.016) -0.035 (0.020) (0.019) -0.035 (0.023) -0.079*** -0.071*** (0.016) -0.033 (0.020) -0.130*** (0.025) -0.037 (0.027) 0.029** (0.011) 0.022* (0.010) (0.019) -0.035 (0.023) -0.087** (0.030) 0.001 (0.031) 0.022* (0.011) 0.014 (0.010) 0.005 (0.015) 0.014 (0.020) 0.048* (0.022) 0.005 (0.015) 0.016 (0.019) 0.051* (0.023) -0.030 (0.020) -0.026 (0.028) 0.062 (0.033) 0.000 (0.011) 0.001 (0.013) -0.008 (0.006) HH in Yaya M -0.008 (0.017) 0.001 (0.018) -0.008 (0.017) 0.002 (0.018) -0.042* (0.021) -0.038 (0.027) 0.007 (0.013) 0.008 (0.012) HUMAN CAPITAL INDEX 13. DDAFS Year M SD M SD M SD 2009 0.619 0.140 0.610 0.153 0.541 0.141 1999 0.624 0.160 0.624 0.151 0.589 0.194 2009 0.616 0.140 0.608 0.151 0.537 0.143 14. CHHDDI2 1999 0.621 0.160 0.623 0.149 0.588 0.198 15. HHDI 2009 0.797 0.175 0.755 0.192 0.667 0.268 1999 0.767 0.232 0.729 0.240 0.730 0.235 2009 0.959 0.110 0.966 0.106 0.988 0.043 16. HHMI 1999 0.958 0.105 0.966 0.099 0.980 0.043 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) Standard Errors in parentheses; 6) *** p<0.001, ** p<0.01, * p<0.05 630 MTS HH in Yaya (n=203) Non-MTS HH in Yaya (n=120) NsemSaw (n=116) MTS & Non-MTS Table 7.2.2 (cont’d) Estimated Estimated Differences Between Groups Differences Within Groups MTS in Yaya & NsemSaw Non-MTS in Yaya & NsemSaw MTS HH in Yaya Non-MTS in Yaya NsemSaw M M M M M -0.094*** -0.098*** (0.013) (0.015) -0.055*** -0.058*** (0.014) (0.015) -0.036** (0.012) -0.038** (0.016) 0.003 (0.013) -0.095*** -0.099*** (0.013) (0.015) -0.063*** -0.060*** -0.056*** -0.063*** (0.012) (0.016) -0.024 (0.013) (0.013) (0.015) HH in Yaya M 0.004 (0.014) 0.004 (0.014) 0.004 (0.014) 0.007 (0.014) HUMAN CAPITAL INDEX 17. CHCI1 Year M SD M SD M SD 2009 0.626 0.116 0.631 0.126 0.532 0.100 1999 0.590 0.123 0.593 0.125 0.535 0.103 2009 0.626 0.117 0.630 0.127 0.531 0.101 18. CHCI2 1999 0.563 0.122 0.570 0.123 0.507 0.102 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) Standard Errors in parentheses; 6) *** p<0.001, ** p<0.01, * p<0.05 631 7.2.2 Summary Results Among all three research groups MTS and Non-MTS in Yaya experienced significant improvements in eight out of the 18 human capital indexes between 1999 and 2009 while NsemSaw experienced significant improvements in four (see tables 7.2.2 above and 7.2.3 below). The t-test results suggested that improvements in FFVCI and FOFC among all three research groups between 1999 and 2009 were indeed significant (*** p<0.001). Similarly, the results suggest that improvements in CHHDDI1 and CHCI2 among MTS and Non-MTS between 1999 and 2009 were also (*** p<0.001). The t-test result for FPAPC index suggests that only the improvements recorded among the MTS group between 1999 and 2009 can be supported at the highest degree of confidence (99.99%). While neither the MTS nor Non-MTS experienced significant declines in any human capital indexes, the NsemSaw experienced significant decline in FSFCI, SFSI, DDAFS and CHHDDI2 indexes. The discussions in this section suggest that between 1999 and 2009 the MTS group in Yaya experienced significantly higher improvements in financial capital indexes than Non-MTS and NsemSaw. Also, the Non-MTS group appears to perform better than NsemSaw for the same time period. Since the CHHFCI index captures changes in all the 10 different financial capital indexes (see table 7.1.1) changes in CHHFCI index among all three research groups is further analyzed in section 7.6 using livelihood pentagons. Following section 7.6, the question of what might have caused the MTS group to record significantly higher improvements in several financial capital indexes relative to Non-MTS the NsemSaw is addressed in chapter eight. Using a Difference-in-Difference technique chapter eight isolate changes in CHHFCI index that may be attributed directly to the MTS project or spillover from the project. Section 7.3 that follows discusses changes in physical capital indexes among the three research groups. 632 Table 7.2.3: Summary of T-Test for Estimated Differences Between and Within Groups MTS & Non-MTS Mean (MTS – Non- No. Index MTS) MTS & NsemSaw Mean (MTS – NsemSaw) Non & MTS- NsemSaw Mean (Non-MTS – NsemSaw) 1999 2009 1999 2009 1999 2009 MTS Non-MTS NsemSaw CHCI2 18. 1) Significant improvements in livelihood index: +++p<0.001, ++ p<0.01, + p<0.05, 2) Significant decline in livelihood index: ---p<0.001, --p<0.01, - p<0.05 3) Insignificant improvements in livelihood index: + and 4) Insignificant decline in livelihood index: - - - +++ +++ +++ +++ ++ +++ ++ +++ +++ +++ +++ +++ 633 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. EDULIT MIGI FSFCI FFVCI FPAPC FOFC DDAFC CHHDDI1 SFSI FVSI PAPSI OFSI DDAFS CHHDDI2 HHDI HHMI CHCI1 - - - - - + + - - + + - - - + - - - - + - + - - - + + + - + + + - - +++ +++ - +++ + +++ + + + + ++ + + + + - +++ +++ ++ + +++ +++ +++ +++ +++ + +++ +++ +++ +++ +++ -- +++ +++ - +++ + +++ + ++ + + + + + + - - +++ +++ ++ + ++ +++ +++ +++ + + +++ +++ +++ +++ +++ - 2009- 1999 + ++ - +++ +++ +++ + +++ - - - + - - + - 2009-1999 + + - +++ ++ +++ + +++ + - - + - - + - 2009- 1999 + + --- +++ + +++ + + --- +++ + +++ - - - + - + 7.3 Physical Capital Section 7.3 discusses changes in 12 physical capital indexes among the three research groups (MTS, Non-MTS and NsemSaw) before (1999) and after (2009) the MTS project. Physical capital within the context of this research describes the physical infrastructure that provides the basic foundations for individual/households or communities to engage in economic activities. Chapter three section 3.3.3 provides a broader discussion of the literature on physical capital and how it shapes household livelihoods. The 12 physical capital indexes in this section may be grouped into the following five broad categories: a) shelter, b) water and energy infrastructure, c) availability/proximity to communal infrastructure such as roads, markets and health centers, d) communication facilities/networks and e) basic household possessions (e.g., cellphones, bicycle, motorbike and radio). Table 7.3.1 below describes how each of the 12 physical capital indexes were generated while table 7.3.2 presents the results of a Two Sample T- test used to track (within and between group) changes in each of the 12 indexes before (1999) and after (2009) the MTS project. The tables presented under Appendix Q describes in greater detail how each physical capital index was constructed from the household survey results. 634 Table 7.3.1: Definition of Household Physical Capital Indexes No. Index Name of Index/ Variable Description of Indexes Computation of Indexes 1 HORO Home Ownership and Room Occupancy Index HORO captures home ownership status and the average number of individuals in a room. Households that own their own home earn a score of 1 and those that rent score 0. Depending on the number persons in the house, HHs with more than 3 persons per room score 0 and those with less than 3 earn a score of 1. The maximum attainable HORO score is 2 and the minimum is 0. HORO index thus ranges from 0 to 1. Home Constructi 2 HCRT on and Roof-Type Index HCRT captures the physical construction of the main structure of the house and the type of roof over the house. If the main structure of the house is made from cement and concrete the household scores 1 and if it is made from mud/earth the HH scores 0. Houses with aluminum roof score 1 while those with thatch, raffia or mud roofs score 0. Households living in wooden kiosks or metal shipping containers were ranked similar to those in earth/mud houses with thatch roofs. The maximum attainable score for HCRT is 2 and the HCRT index ranges from 0 to 1. 635 21)RoomOccup_wn_1Hse_Rent_OHORO21ititt==+=ni2)RoofType_1u_1(HseConstrHCRT21ititt==+=ni No. Index Name of Index/ Variable 3 LSWS Light Energy and Water Source Index Table 7.3.1 (cont’d) Description of Indexes Computation of Indexes LSWS captures drinking water and energy sources used by the HH. Households that obtain their drinking water from portable sources such as tap or boreholes earn a score of 1 while those that fetch from streams, rivers or open gutters score 0. Households that use electric, propane gas or lithium batteries to light-up their homes earn a score of 1 while those that use firewood and kerosene score 0. The maximum attainable score for LSWS is 2 and the minimum is 0. LSWS index expressed in terms of energy and water sources ranges from 0 to 1. 4 KBTF Kitchen, Bathroom and Toilet Facility Index KBTF captures the type of kitchen, bathroom and toilet facilities within a HH. Households with a private space or room exclusively used as a kitchen score 1 while those who cook in an open space with no roof score 0. Similarly, HHs with a private space used exclusively as a bathrooms score 1 while those with a shared space outside of the HH score 0. Households with access to a private or communal KVIP facility score 1 while those without any toilet facility score 0. The maximum attainable score for KBTF is 3 and the minimum is 0. KBTF index thus ranges from 0 to 1. 636 2e_1)WaterSourcce_1(LightSourLSWS21ititt==+=ni3y_1a)ToiletFact +_1BathnFacty +pe_1(KitchenTyKBTF31itt===ni No. Index Name of Index/ Variable Table 7.3.1 (cont’d) Description of Indexes Computation of Indexes 5 LSWD Liquid and Solid Waste Disposal Index LSWD captures liquid and solid waste disposal facilities available to a HH. Households with gutter facilities to damp their liquid waste score 1 while those without gutters sore 0. Similarly, HHs with access to a public dumpster score 1 while those without access score 0. The maximum attainable score for LSWD is 2 and the minimum is 0 and the LSWD index ranges from 0 to 1. 6 MHPT Access to Markets, Health, Post Office, and Transport Facilities Index MHPT captures a HH’s access to market outlets, health care facilities, post office and transport facilities (e.g. major roads). Households that live within an hour’s travel by foot or motor vehicle from any of the four facilities earn a score of 1 for each of those facilities while those that live more than one hour’s walk or drive score 0. The maximum attainable score for MHPT is 4 and the minimum is 0. The MHPT index thus ranges from 0 to 1. 637 2isp_1)SolidWastD +sp_1(LQDWastDiLSWD21itt===ni1*4erv)itTransportS +Offce_1AccessPost +AccessHealthServ +ess(MarketAccMHPT41t===ni No. Index Name of Index/ Variable 7 HCPO Household Cell-Phone Ownership Index 8 CNWR Cell-phone Reliability Index Table 7.3.1 (cont’d) Description of Indexes Computation of Indexes HCPO captures the availability of five major cellphones and networks within each HH surveyed. Households that have functional cell-phones from any of the five cellular network providers and are able to catch signals within their households from any of these providers score 1 for the particular provider and those that do score 0. The maximum attainable score on HCPO is 5 and the minimum is 0. HCPO index expressed in terms of the range of functional cell-phone and cellular network availability within a household ranges from 0 to 1. CNWR captures cellular network reliability of each of the five service providers/networks within a household. Hence HHs that cannot access a particular network signal within their HH or report that the network is unreliable (i.e. takes at least 6 trials to make a call) receive a score of 0 for the particular network. Households that reported having reliable network signal score 1 for the service provided in question. The maximum attainable CNWR score is 5 and the minimum is 0. CNWR index expressed in terms of the level of cellular network reliability within a household ranges from 0 to 1. 638 1*5 Kasapa)it+Tigo +Zaine + MTN+(OneTouchHCPO41t===ni1*1ability)it(NetwkReliCNWR11t===ni No. Index Name of Index/ Variable 9 BHP Basic Household Possessions Index 10 LHP Luxury Household Possessions Index Table 7.3.1 (cont’d) Description of Indexes Computation of Indexes BHP captures a household’s ownership of 7 basic household items including lantern, touch/flashlight, radio, cell-phone, sewing machine, carte/four-legged trolley, and bicycle. Since some of these items cost more than others, each of the 7 items was weighted in accordance with their economic value. In order of magnitude, households that have a lantern score 1for this item while those that have a bicycle score 7. Touch/flashlight, radio, cellphone, sewing machine and carte are weighted 2, 3, 4, 5 and 6 respectively. The maximum attainable BHP score on is thus 28 and the minimum is 0. BHP index expressed in terms availability of basic household items ranges from 0 to 1. LHP captures a household’s ownership of 5 items (TV, modern furniture, motor, car and tractor) considered as luxury household items. Since some of these items cost more than others, each of the 5 items is weighted in accordance with their economic value. In order of magnitude, households that have a TV score 8 for this item while those that have a tractor score 12. Items in between the minimum and maximum thresholds including modern furniture, motor, car and tractor are weighted 9, 10, and 11 respectively. The maximum attainable LHP score on is thus 50 and the minimum is 0. BHP index expressed in terms availability of luxury household items ranges from 0 to 1. 639 287]it*e_1)(HHPossBik +6*te_1)(HHPossCar +5*Mchne_1)(HHPossSew +4*lPhne_1)(HHPossCel +3*io_1)(HHPossRad +2*)ashlight_1(HHPoss_Fl +antern_1)[(HHPoss_LBHP71t===ni5012it*ctor_1)](HHPossTra +11*_1)(HHPossCar +10*or_1)(HHPossMot +9*nture_1)(HHPossFur +8*_1)[(HHPossTVLHP71t===ni No. Index Name of Index/ Variable Table 7.3.1 (cont’d) Description of Indexes Computation of Indexes 11 CHPCI1 Physical Combine HH Capital Index CHPCI1 captures a HH’s aggregate or combined physical capital endowment. This index is simply the average of all the 10 indexes hence the maximum attainable score on a household’s combined Physical Capital endowment is 10 and the minimum is 0. CHPCI1 index thus ranges from 0 to 1. 12 CHPCI2 Physical Combine HH Capital Index CHPCI2 is similar to CHPCI1 only in CHPCI2 Luxury Household Possessions Index was not included in the computation thus making the maximum attainable sore for CHPCI2 9. 640 ]10[] LHP+ BHP+CNWR +CNWA + MHPT+ LSWD+KBTF + LSWS+ HCRT+[HOROCHPCI1tt=]9[] BHP+CNWR +HCPO + MHPT+LSWD + KBTF+LSWS + HCRT+[HOROCHPCI2tt= 7.3.1 Two Sample T-test of between and within group differences in 1999 and 2009 Section 7.3.1 investigates between and within group differences for all 12 physical capital indexes described in tables 7.3.1 above. The Two-Sample t-test determined changes that occurred in household physical capital indexes between 1999 and 2009 and also the degree to which these changes (between and within group) were significant using three different confident levels 99.99% (*** p<0.001), 99% (**p<0.01) and 95% (*p<0.05). Within group difference for each physical capital index is computed by subtracting the mean result obtained for that particular index in 2009 from the mean in 1999 for the same group (MTS households in Yaya or non-MTS groups in Yaya). Because the Two-Sample t-test is computed by differencing the mean value for physical capital index in the current year (2009) from the base year (1999), a positive difference suggests a decline in livelihood index while a negative difference suggests an improvement. The computation for between group differences and within group differences are similar however in the former, the changes were computed for the different research groups for a particular time period (either 1999 or 2009). 7.3.1.1 Home Ownership and Room Occupancy Index (HORO) In 1999 and 2009 there was no statistically significant difference in the mean Home Ownership and Room Occupancy Index (HORO) between MTS and Non-MTS research groups. However, in 1999 the HORO index among NsemSaw was significantly (* p<0.05) higher (7.1% and 4.6% respectively) than the MTS and non-MTS groups. By 2009 both the MTS and Non- MTS groups recorded a higher HORO index than NsemSaw however these differences were not significant. Among all three groups only Non-MTS experienced an increase in HORO index between 1999 and 2009 though this increase was insignificant. Both MTS and NsemSaw 641 experienced a decline (0.8% and 7.8% respectively) in HORO index between 1999 and 2009 however only the decline experienced by NsemSaw was significant (*p<0.05). 7.3.1.2 Home Construction and Roof-Type Index (HCRT) In 1999 and 2009 there was no statistically significant difference in mean Home Construction and Roof-Type Index (HCRT) between MTS and Non-MTS. In 1999 the HCRT index among MTS was significantly (*** p<0.001) higher (21.4%) than that of NsemSaw. By 2009 however the gap in HCRT between MTS and NsemSaw increased and remained significantly (*** p<0.001) higher (26.4%) than NsemSaw. Similarly, in 1999 the mean HCRT index among the Non-MTS group was significantly (*** p<0.001) higher (23.4%) than NsemSaw and by 2009 the observed difference between both groups increased and remained significantly (*** p<0.001) higher (27.1%) than that of NsemSaw households. All three research groups experienced an increased in their HCRT indexes between 1999 and 2009 though these changes were significant only for the MTS group. Between 1999 and 2009 the MTS group experienced a significant (*p<0.05) increase (7.1%) in mean HCRT index. 7.3.1.3 Light Energy and Water Source Index (LSWS) In 1999 and 2009 there was no statistically significant difference in the mean Light Energy and Water Source Index (LSWS) between the MTS and Non-MTS research groups. According to the T-test results presented below, the MTS group experienced a significantly (*p<0.05) higher (6.5%) LSWS index than NsemSaw in 1999 and by 2009 the difference between both groups increased significantly (***p<0.001) from 6.6% to 12.9%. Among all three groups the MTS group experienced the highest and significant (*** p<0.001) increase (21.9%) in 642 LSWS index. Between 1999 and 2009 the LSWS index among non-MTS and NsemSaw groups increased significantly (*** p<0.001) by 18.3% and 15.5% respectively. 7.3.1.4 Kitchen, Bathroom and Toilet Facility Index (KBTF) In 1999 the mean Kitchen, Bathroom and Toilet Facility Index (KBTF) among the NsemSaw group was significantly (*p<0.05) higher (6.1% and 5.7% respectively) than those of MTS and Non-MTS groups. By 2009 the KBTF gap between both NsemSaw and both MTS and Non-MTS closed and was insignificant. All three research groups experienced an improvement in KBTF index between 1999 and 2000 however these changes according to the t-test was insignificant. 7.3.1.5 Liquid and Solid Waste Disposal Index (LSWD) In 1999 the there was no significant difference in mean LSWD index between MTS and Non-MTS however by 2009 the Non-MTS group experienced a higher (5.5%) and significant (*p<0.05) difference in LSWD relative to the MTS group. Between 1999 and 2009 none of the research groups experienced a significant change in LSWD index. 7.3.1.6 Access to Markets, Health, Post Office and Transport Facilities Index (MHPT) In 1999 the mean MHPT index among the MTS group was significantly (*p<0.05) higher (7.8%) than that of the Non-MTS group. The MHPT index among MTS was significantly (*** p<0.001) higher (27.2%) than that of NsemSaw and by 2009 the gap in MHPT between both groups increased to 31.2% and remained significantly (*** p<0.001). Similarly, in 1999 the mean MHPT index among the Non-MTS group was significantly (***p<0.001) higher (22%) than NsemSaw and by 2009 the observed difference increased to 23.3% and remained significant (*** p<0.001). All the three research groups experienced an increase in MHPT index between 643 1999 and 2009 however only the changes experienced among MTS (9%) was significant (**p<0.01). 7.3.1.7 Household Cell-Phone Ownership (HCPO) Index In 1999 there was no significant difference in the mean HCPO index among MTS and Non-MTS households however by 2009 the Non-MTS group experienced a significantly (*p<0.05) higher (6.8%) HCPO index than the MTS group. The HCPO index among MTS households was significantly (*p<0.05) higher (4.6%) in 1999 than that of the NsemSaw group. Though both groups experienced an increase in HCPO index between 1999 and 2009, the MTS group still maintained a significantly (*p<0.05) higher (5.3%) index than the NsemSaw group. Like the MTS group, the Non-MTS also experienced significantly (*p<0.05) higher (3.3%) HCPO than NsemSaw in 1999 and the gap in HCPO between both groups increased (12.2%) significantly (***p<0.001) by 2009. Between 1999 and 2009, all three groups experienced a significant (***p<0.001) increase (average of 62%) in HCPO index. 7.3.1.8 Cell-phone Network Reliability Index (CNWR) In 1999 the mean CNWR index among Non-MTS households was significantly (*p<0.05) higher (4.4%) than those of the MTS group. By 2009 however, the gap in CNWR index between both MTS and Non-MTS fell below 2% and was insignificant according to the t- test results. In 1999 there was no significant difference between MTS and NsemSaw however by 2009 the MTS group experienced a significantly (***p<0.001) higher (36.3%) CNWR index than NsemSaw. When the CNWR index was compared between Non-MTS and NsemSaw in 1999 and 2009 the trends were similar to those between MTS and NsemSaw groups. Among all three groups, the MTS experienced the highest (70%) increase in CNWR index between 1999 644 and 2009 followed by Non-MTS (64.2%) and NsemSaw (31.9%). The change in CNWR index among all three research groups between 1999 and 2009 was significant (***p<0.001). 7.3.1.9 Basic Household Possessions Index (BHP) Among all three research groups, MTS experienced the highest (13.7%) increase in Basic Household Possessions (BHP) Index between 1999 and 2009 followed by Non-MTS (12.7%) and NsemSaw (11.5%). The change in BHP index among all three research groups between 1999 and 2009 according to the t-test results was significant (***p<0.001). 7.3.1.10 Luxury Household Possessions Index (LHP) In 1999 there the mean LHP index among MTS households was higher (1.7%) than NsemSaw however this difference according to the t-test was not significant. By 2009 however, the gap in mean LHP index between MTS and NsemSaw increased to 6.2% and was significant (***p<0.001). In 1999 the mean LHP index among Non-MTS was significantly (**p<0.01) higher (2.9%) than NsemSaw and by 2009 gap in mean LHP between both groups increased to 6.2% and remained significant (**p<0.01). Between 1999 and 2009, all three groups experienced an increase in BHP index however MTS households experienced the highest (4.6%) and most significant (***p<0.001) increase. For the same time period, the Non-MTS group experienced a significant (**p<0.01) increase (3.4%) in LHP index. Though the NsemSaw group also experienced an increase (0.2%) in LHP this change according to the t-test was insignificant. 7.3.1.11 Combine Household Physical Capital Index (CHPCI1) In 1999 the Combine Household Physical Capital Index (CHPCI1) among MTS was significantly (*** p<0.001) higher (4.4%) than that of NsemSaw. By 2009 however the gap in CHPCI1 between both groups increased from 4.4% to 12.1% and this difference according to the 645 t-test was significant (*** p<0.001). Similarly, in 1999 the mean CHPCI1 index among the Non- MTS group was significantly (*** p<0.001) higher (5.0%) than NsemSaw and by 2009 the observed difference between both groups increased from 5.0% to 11.9% and the difference was significant (*** p<0.001). Between 1999 and 2009, all three research groups experienced a significant (*** p<0.001) improvement in CHPCI1. The MTS group recorded the highest improvement (19.5%) in CHPCI1 followed by the Non-MTS (18.6%) and NsemSaw (11.8%). 7.3.1.12 Combine HH Physical Capital Index (CHPCI2) In 1999 there was no significant difference in Combine Household Physical Capital Index (CHPCI2) between MTS and NsemSaw households however by 2009 the gap in CHPCI2 increased significantly (*** p<0.001) from 4.7% to 12.8% with the MTS experiencing a significantly higher CHPCI2 index. A similar trend was observed when the differences in CHPCI2 were compared between Non-MTS and NsemSaw for 1999 and 2009. Between 1999 and 2009, all three research groups experienced a significant (*** p<0.001) improvement in CHPCI2. The MTS group recorded the highest improvement (21%) in CHPCI1 followed by the Non-MTS (20.3%) and NsemSaw (12.9%). Table 7.3.2b below presents a summary of the results discussed under sections 7.3.1.7 through 7.3.1.12 above. Table 7.3.2a below presents a summary of the results discussed under sections 7.3.1.1 through 7.3.1.6 above. The t-test results for the next set of six the physical capital indexes are discussed in sections 7.3.1.7 through 7.3.1.12 following which the summary t-test results are presented in table 7.3.2b. 646 Table 7.3.2: T-Test for Estimated Differences in Physical Capital Indexes Between and Within Groups MTS HH in Yaya (n=203) Non-MTS HH in Yaya (n=120) NsemSaw (n=116) MTS & Non-MTS PHYSICAL CAPITAL INDEX 1. HORO Year M SD M SD M SD 2009 0.581 0.267 0.567 0.297 0.543 0.260 1999 0.549 0.307 0.575 0.308 0.621 0.314 2009 0.384 0.345 0.392 0.338 0.121 0.215 2. HCRT 1999 0.313 0.325 0.333 0.333 0.099 0.200 2009 0.603 0.305 0.563 0.370 0.474 0.279 3. LSWS 1999 0.384 0.338 0.379 0.343 0.319 0.312 2009 0.343 0.302 0.356 0.286 0.379 0.240 4. KBTF 1999 0.315 0.299 0.319 0.275 0.376 0.243 Estimated Estimated Differences Between Groups Differences Within Groups MTS in Yaya & NsemSaw Non-MTS in Yaya & NsemSaw M -0.038 (0.031) 0.071* (0.036) M -0.024 (0.036) 0.046 (0.040) -0.264*** -0.271*** (0.035) (0.037) -0.214*** -0.234*** (0.033) (0.036) -0.129*** (0.034) -0.065* (0.038) 0.036 (0.033) 0.061* (0.033) -0.088* (0.043) -0.060 (0.043) 0.024 (0.034) 0.057* (0.034) MTS HH in Yaya Non-MTS in Yaya NsemSaw M M M -0.032 (0.029) 0.008 (0.039) 0.078* (0.038) -0.071* (0.033) -0.058 (0.043) -0.022 (0.027) -0.219*** -0.183*** -0.155*** (0.032) (0.046) (0.039) -0.028 (0.030) -0.036 (0.036) -0.003 (0.032) HH in Yaya M -0.015 (0.032) 0.026 (0.035) 0.007 (0.039) 0.021 (0.038) -0.041 (0.038) -0.005 (0.039) 0.012 (0.034) 0.004 (0.033) 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) Standard Errors in parentheses; 6) *** p<0.001, ** p<0.01, * p<0.05 647 Table 7.3.2 (cont’d) Estimated Estimated Differences Between Groups Differences Within Groups MTS HH in Yaya (n=203) Non-MTS HH in Yaya (n=120) NsemSaw (n=116) MTS & Non-MTS MTS in Yaya & NsemSaw Non-MTS in Yaya & NsemSaw M 0.017 (0.023) 0.044 (0.025) M -0.038 (0.027) 0.013 (0.028) -0.312*** -0.233*** (0.040) (0.050) -0.272*** -0.220*** (0.041) (0.049) MTS HH in Yaya Non-MTS in Yaya NsemSaw M M M -0.027 (0.022) -0.050 (0.032) 0.000 (0.021) -0.090** (0.033) -0.063 (0.050) -0.050 (0.049) -0.053* (0.033) -0.046*** (0.015) -0.122*** (0.035) -0.033* (0.018) -0.377*** -0.363*** (0.053) 0.004 (0.023) (0.061) -0.040 (0.032) -0.592*** -0.673*** -0.584*** (0.024) (0.032) (0.023) -0.700*** -0.642*** -0.319*** (0.034) (0.048) (0.049) HH in Yaya M 0.055* (0.026) 0.032 (0.028) -0.079* (0.040) -0.052 (0.042) 0.068* (0.034) -0.013 (0.020) -0.014 (0.051) 0.044* (0.026) PHYSICAL CAPITAL INDEX 5. LSWD Year M SD M SD M SD 2009 0.483 0.213 0.538 0.244 0.500 0.162 1999 0.456 0.234 0.488 0.255 0.500 0.162 2009 0.756 0.324 0.677 0.380 0.444 0.386 6. MHPT 1999 0.666 0.350 0.615 0.393 0.394 0.358 2009 0.655 0.301 0.723 0.297 0.602 0.242 7. HCPO 1999 0.063 0.157 0.050 0.189 0.017 0.006 2009 0.739 0.440 0.725 0.448 0.362 0.483 8. CNWR 1999 0.039 0.195 0.083 0.278 0.043 0.204 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) Standard Errors in parentheses; 6) *** p<0.001, ** p<0.01, * p<0.05 648 Table 7.3.2 (cont’d) Estimated Estimated Differences Between Groups Differences Within Groups MTS HH in Yaya (n=203) Non-MTS HH in Yaya (n=120) NsemSaw (n=116) MTS & Non-MTS MTS in Yaya & NsemSaw Non-MTS in Yaya & NsemSaw M -0.028 (0.027) -0.006 (0.025) M -0.011 (0.030) 0.001 (0.028) -0.062*** -0.062*** (0.015) -0.017 (0.011) (0.015) -0.029** (0.013) -0.121*** -0.119*** MTS HH in Non-MTS in Yaya Yaya NsemSaw M M M -0.137*** -0.127*** -0.115*** (0.023) (0.030) (0.028) -0.046*** (0.012) -0.034* (0.017) -0.002 (0.011) (0.014) (0.015) -0.195*** -0.186*** -0.118*** -0.044*** -0.050*** (0.012) (0.016) (0.011) (0.012) (0.013) -0.128*** -0.125*** (0.015) -0.047 (0.013) (0.016) -0.052 (0.014) -0.210*** -0.203*** -0.129*** (0.013) (0.018) (0.012) HH in Yaya M -0.017 (0.027) -0.007 (0.025) 0.000 (0.017) 0.012 (0.012) -0.002 (0.016) 0.006 (0.013) -0.003 (0.017) 0.005 (0.015) PHYSICAL CAPITAL INDEX 9. BHP Year M SD M SD M SD 2009 0.443 0.236 0.426 0.237 0.415 0.221 1999 0.307 0.217 0.300 0.222 0.301 0.203 2009 0.081 0.147 0.081 0.146 0.019 0.078 10. LHP 1999 0.035 0.096 0.047 0.110 0.017 0.092 2009 0.507 0.134 0.505 0.136 0.386 0.092 11. CHPCI1 12. CHPCI2 1999 0.313 0.116 0.319 0.116 0.269 0.076 2009 0.554 0.142 0.552 0.146 0.427 0.101 1999 0.344 0.126 0.349 0.128 0.297 0.083 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) Standard Errors in parentheses; 6) *** p<0.001, ** p<0.01, * p<0.05 649 7.3.2 Summary Results Among all three research groups MTS experienced significant improvements in nine out of 12 physical capital indicators between 1999 and 2009 while Non-MTS in Yaya and NsemSaw experienced significant improvements in seven and six respectively (see tables 7.3.2a and 7.3.2b above and 7.3.3 below). According to the t-test results, all three research groups experienced significant (*** p<0.001) improvements in six of the 12 physical capital indexes (LSWS, HCPO, CNWR, BHP, CHPCI1 and CHPCI2) between 1999 and 2009. Of the three groups, NsemSaw experienced a significant (*p<0.05) decline in one physical capital index (HORO). While the NsemSaw group also experienced a decline in LSWD index between 1999 and 2009 this change was not significant. Like NsemSaw, the Non-MTS group also experienced a decrease in HORO index between 1999 and 2009 however this change according to the t-test was insignificant among the Non-MTS group. The MTS group was the only group that did not experience a decline in any of the 12 physical capital indexes. The discussions in this section suggest that between 1999 and 2009 the MTS group in Yaya appear to fair better on most of the physical capital indexes than the Non-MTS in Yaya and NsemSaw groups. Also, the Non-MTS group appears to perform better than NsemSaw for the same time period. Since the CHPCI1 index captures changes in the nine key physical capital indexes across all three research groups, CHPCI1 is further analyzed in section 7.6 using livelihood pentagons. Following section 7.6, the question of what might have caused the MTS group to record significantly higher improvements in several physical capital indexes relative to both Non-MTS and NsemSaw is addressed in chapter eight. Using a Difference-in-Difference technique chapter eight isolate changes in CHPCI1 index that may be attributed directly to the MTS project or spillover from the project. 650 Section 7.4 that follows discusses changes in natural capital indexes among the three research groups. 651 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. HORO HCRT LSWS KBTF LSWD MHPT HCPO CNWR BHP LHP CHPCI1 - - + - - + + - + - - + - + - - + - + + - + - +++ + - - +++ +++ - + + +++ + +++ +++ - - +++ + +++ + +++ +++ +++ - + +++ +++ + - - +++ + - - ++ +++ + - + +++ +++ +++ + +++ +++ + 2009- 1999 + + +++ + + ++ +++ +++ +++ +++ +++ +++ 2009- 1999 - + +++ + + + +++ +++ +++ + 2009- 1999 - + +++ + - + +++ +++ +++ + +++ +++ +++ +++ Table 7.3.3: Summary of T-Test for Estimated Differences Between and Within Groups MTS & Non-MTS Mean (MTS – Non- No. Index MTS) MTS & NsemSaw Non & MTS-NsemSaw Mean (MTS – NsemSaw) Mean (Non-MTS – MTS Non-MTS NsemSaw NsemSaw) 1999 2009 1999 2009 1999 2009 - + + CHPCI2 12. 1) Significant improvements in livelihood index: +++p<0.001, ++ p<0.01, + p<0.05, 2) Significant decline in livelihood index: ---p<0.001, --p<0.01, - p<0.05 3) Insignificant improvements in livelihood index: + and 4) Insignificant decline in livelihood index: - + 652 7.4 Natural Capital This section discusses changes in 17 Natural Capital Indexes (NCI) among the three research groups (MTS, Non-MTS and NsemSaw) before (1999) and after (2009) the MTS project. Within the context of this research NCIs are the different natural resource stock from which households within the research communities build and sustained their livelihoods (refer to chapter three sections 3.3.4 for a broader discussion on natural capita). The 17 NCIs discussed thus encompass four broad categories including: a) crop land ownership, crop production and marketing, b) livestock production and marketing, c) soil fertility and d) Non-Timber Forest Products (NTFPs). For example, Crop Land Ownership Index (CLOI) describes land ownership by the household while Household Crop Diversity Index (HHCDI) describes the diversity of crops produced by the household. Subsistence Oriented Crop Production Index (SOCPI) and Market Oriented Crop Production Index (MOCPI) both describe the nature of crop production activities within the household. Like crop production, the study uses six indexes to measure changes in livestock ownership, diversity, populations as well as market and subsistent oriented production for the two time periods before and after MTS implementation. Under chapter seven section 7.4 the study discusses changes in soil fertility and NTFPs before and after the MTS project. Table 7.4.1 below describes how each of the 17 NCIs were generated while table 7.4.2 presents the results of a Two Sample T-test used to track (within and between group) changes in NCIs before (1999) and after (2009) the MTS project. The tables presented under Appendix R describes in greater detail how each NCI was constructed from the household survey results. 653 Table 7.4.1: Definition of Household Natural Capital Indexes No. Index Name of Index/ Variable Description of Indexes Computation of Indexes 1 CLOI Crop Land Ownership Index (CLOI) CLOI captures four categories (total land owned, size of crop farm, size of uncultivated land, and land farmed that does not belong to the household) of land ownership status. A household owning more than 5 acres on a particular category earns a score of 4 on that category while those with 0 acres earn a score of 0. The maximum attainable score on Crop Land Ownership is thus 16 and the minimum is 0. The CLOI index thus ranges from 0 to 1. 2 HHCDI Household Crop Diversity Index (HHCDI) HHCDI captures the diversity of 6 major crops (maize/corn, cassava, plantain, yam, cocoyam and tomatoes) produced by a household. A household producing any of the 6 crops earns a score of 1 for that crop and 0 if not producing. The maximum attainable score on HHCD is thus 6 and the minimum is 0. HHCDI expressed in terms of the level crop diversity ranges from 0 to 1. 654 16CP4Q) + CP3Q + CP2Q + (CP1QCLOI41tit===ni6CPY_6) + CPY_5 + CPY_4+ CPY_3 + CPY_2 + (CPY_1HHCDI61tit===ni Table 7.4.1 (cont’d) Description of Indexes Computation of Indexes No. Index Name of Index/ Variable Subsistence Oriented Crop 3 SOCPI Production Index (SOCPI) SOCPI captures the level of subsistence crop production within a HH. Households that strongly agree that crops are usually produced for HH consumption or that they usually consume all their harvest with little left for sale earn a score of 4 on each of those questions while those that strongly disagree earn a score of 1. The maximum attainable score on SOCP is thus 8 and the minimum is 0. SOCPI expressed in terms of the level crop diversity ranges from 0.25 to 1. 4 MOCPI Market Oriented Crop Production Index (MOCPI) MOCPI captures the level of commercial crop production within a HH. Households that strongly agree that crops are usually produced for markets or that they usually sell all their harvest with little left for home consumption earn a score of 4 on each of those questions while those that strongly disagree earn a score of 1. The maximum attainable score on MOCP is thus 8 and the minimum is 0. MOCPI expressed in terms of the level crop diversity ranges from 0.25 to 1. 655 8CLU_3Q) +(CLU_1QSOCPI21tit===ni8CLU_4Q) +(CLU_2QMOCPI21tit===ni Table 7.4.1 (cont’d) Description of Indexes Computation of Indexes No. Index 5 CPTI Name of Index/ Variable Crop Production Trend Index (CPTI) CPTI captures trend in production (i.e. decreasing, stable or increasing) and level of sufficiency of 6 major crops (maize/corn, cassava, plantain, yam, cocoyam and tomatoes) produced by a household. Households that record an increasing level of production and self-sufficiency on any of the 6 crops score 5 for that particular crop. While households recording decreasing levels of production and sufficiency score 1 for the crop in question. The maximum attainable score on CPT is thus 30 and the minimum is 6. CPTI expressed in terms of the trend in production and sufficiency levels ranges from 0.20 to 1. 6 HHLDI1 HH Livestock Diversity Index 1 (HHLDI1) HHLDI1 captures the diversity of 5 major livestock groups (poultry, rabbits, guinea pigs and grasscutter, goats and sheep, pigs and cattle) produced by a household. A household producing any one of the 5 livestock groups earns a score of 1 for that group and 0 if not producing. The maximum attainable score on HHLD1 is thus 5 and the minimum is 0. HHLDI1 expressed in terms of the level livestock diversity ranges from 0 to 1. 656 30CPY_A6Q) + CPY_A5Q + CPY_A4Q+ CPY_A3Q + CPY_A2Q + (CPY_A1QCPTI61tit===ni5LPP_5Q) LPP_4Q LPP_3Q LPP_2Q (LPP_1QHHLDI151tit==++++=ni No. Index Name of Index/ Variable Table 7.4.1 (cont’d) Description of Indexes Computation of Indexes 7 HHLDI2 HH Livestock Diversity Index 2 (HHLDI2) HHLDI2 captures the diversity of 3 major livestock groups (poultry, goats and sheep, pigs) produced by a household. A household producing any one of the 3 livestock groups earns a score of 1 for that group and 0 if not producing. The maximum attainable score on HHLD2 is thus 3 and the minimum is 0. HHLDI2 expressed in terms of the level livestock diversity ranges from 0 to 1. 8 HHLPI1 HH Livestock Population Index 1 (HHLPI1) HHLPI1 captures the population of 5 major livestock groups (poultry, rabbits, guinea pigs and grasscutter, goats and sheep, pigs and cattle) produced by a household. A household producing more than 20 of each livestock group earns a score of 4 for the group and 1 if not producing. The maximum attainable score on HHLP1 is thus 20 and the minimum is 5. HHLPI1 expressed in terms of the level livestock population ranges from 0.25 to 1. 657 3) LPP_4Q LPP_3Q (LPP_1QHHLDI231tit==++=ni20LP_5Q) LP_4Q LP_3Q LP_2Q (LP_1QHHLPI151tit==++++=ni No. Index Name of Index/ Variable Table 7.4.1 (cont’d) Description of Indexes Computation of Indexes 9 HHLPI2 HH Livestock Population Index 2 (HHLPI2) HHLPI2 captures the population of 3 major livestock groups (poultry, goats and sheep and pigs) produced by a household. Household’s producing more than 20 of each livestock group scores 4 for that group and 1 if not producing. The maximum attainable score on HHLP1 is thus 12 and the minimum is 3. HHLPI1 expressed in terms of the level livestock population ranges from 0.25 to 1. 10 LPTI Livestock Production Trend Index 1 (LPTI) LPTI captures the trend in production (i.e. decreasing, stable or increasing) and level of sufficiency of the 5 major livestock groups produced by a household. Households that produce any of the 5 major livestock at an increasing level of production and sufficiency earns a score of 5 for that particular group. While households producing at a decreasing levels of production and sufficiency earn a score of 1 for the livestock in question. Sufficiency in this context measures the level with which production meets the HH’s needs. The maximum attainable score for LPT is thus 25 and the minimum is 5. CPTI expressed in terms of the trend in production and sufficiency levels ranges from 0.20 to 1. 658 12) LP_4Q LP_3Q (LP_1QHHLPI231tit==++=ni25LP_B5Q) +LP_B4Q +LP_B3Q +LP_B2Q +(LP_B1QLPTI51tit===ni No. Index Name of Index/ Variable Table 7.4.1 (cont’d) Description of Indexes Computation of Indexes 11 SOLPI Subsistence Oriented Livestock Production Index (SOCPI) SOLPI captures the level of subsistence livestock production within a HH. Households that strongly agree that livestock are usually produced for HH consumption or that they usually consume all their livestock with little left for sale earn a score of 4 on each of those questions while those that strongly disagree earn a score of 1. The maximum attainable score on SOLP is thus 8 and the minimum is 2. SOLPI expressed in terms of the level subsistence ranges from 0.25 to 1. 12 MOLPI Market Oriented Livestock Production Index (MOLPI) MOLPI captures the level of commercial livestock production within a HH. Households that strongly agree that livestock are usually produced for markets or that they usually sell all their livestock with little left for home consumption earn a score of 4 while those that strongly disagree earn a score of 1. The maximum attainable score on MOLP is thus 8 and the minimum is 2. MOCPI expressed in terms of the level crop diversity ranges from 0.25 to 1. 659 8CLU_7Q) +(CLU_5QSOLPI21tit===ni8CLU_8Q) +(CLU_6QMOLPI21tit===ni No. Index Name of Index/ Variable Soil Fertilizer 13 SFAI Application Index (SFAI) NTFP Harvests 14 NHMTI and Market Trend (NHMTI) Table 7.4.1 (cont’d) Description of Indexes Computation of Indexes SFAI captures fertilizer use at the time of the research, perceived changes/trend and sufficiency in soil fertility and fertilizer application. Households that applied any type of fertilizer (organic or inorganic) within a year of the study, scored 1 on fertilizer use and those that did not scored 0. HH’s that believed soil fertility on their farms was increasing and sufficient for crop production scored 5 while those who believed soil fertility was declining and not sufficient for crop production scored 1. Households experiencing increasing trend in quantity and frequency of fertilizer use scored 5 while those experiencing declining trends scored 1. The maximum attainable score SFA score is 11 and the minimum is 2. SFAI index ranges from 0.18 to 1. NHMTI captures NTFP harvest patterns as well as trends in quantities harvested and sufficiency for the HH. If a household collected any NTFPs within the past year that HH scored 1 for NTFP harvests and 0 if no. HHs experiencing increasing trends in area of key NTFPs as well as harvests scores 5 points each for this indicators and 1 if the trend is declining and not sufficient. Since firewood appears to be the most important NTFP it was given special attention and scored for trends in harvest and sufficiency. The maximum attainable NHMT score is 16 and the minimum is 3. NHMTI index thus ranges from 0.19 to 1. 660 112Q)FertzerUse SoilFertQ 1(FertzeUseSFAI31tit==++=ni16NTFP_1D) NTFP_1C NTFP_1B (NTFP_1ANMHT41tit==+++=ni No. Index Name of Index/ Variable 15 NHTI NTFP Harvesting Time Index (NHTI) 16 CHHNCI 1 Combined HH Natural Capital Index 1 (CHHNCI1) Table 7.4.1 (cont’d) Description of Indexes Computation of Indexes NHTI captures the average time in hours spent each week harvesting four major NTFPs (mushrooms, firewood, snails and bush-meat including grasscutter). It is assumed that as NTFPs become scare, more time is spent trying to find them hence households that spend more than 20hours a week hunting for any of the four NTFP receives 1 point for the NTFP in question. Households that spend less than 10hrs a week searching for their usual NTFP requirements earn a score of 4 for the NTFP is question. The maximum attainable NHT score is 16 and the minimum is 4. NHTI index thus ranges from 0.25 to 1. CHHNCI1 captures a HH’s aggregate Natural capital endowment. This index takes the average of 13 of the 15 indexes listed above. HHLDI and HHLPI both of which capture respectively the diversity and population of five major livestock categories was included in the computation of CHHNCI1 while HHLD2 and HHLP2 which capture diversity and population of only three of the five most common livestock categories was not included in the computation. The maximum attainable score on CHHNCI1is thus 13 and the minimum is 2.27. CHHNCI1 index ranges from 0.17 to 1. 661 16NTFP_2Dd) NTFP_2Cc NTFP_2Bb (NTFP_2AaNHTI41tit==+++=ni13NHTI) +NHMTI +SFAI +MOLPI+SOLPI + LPTI+HHLPI1 + HHLDI1+CPTI +MOCPI+SOCPI + HHCDI+(CLOICHHNCI1131tit====nni Table 7.4.1 (cont’d) Description of Indexes Computation of Indexes CHHNCI2 captures a HH’s aggregate Natural capital endowment. This index takes the average of 13 of the 15 indexes listed above. HHLD2 and HHLP2 both of which capture respectively the diversity and population of only three major livestock categories (poultry, goats and sheep, pigs) was included in the computation of CHHNCI2 while HHLD1 and HHLP1 which capture diversity and population of all the five most common livestock categories was not included in the computation. The maximum attainable score on CHHNCI2is thus 13 and the minimum is 2.27. CHHNCI21 index ranges from 0.17 to 1. No. Index Name of Index/ Variable 17 CHHNCI2 Combined HH Natural Capital Index 2 (CHHNCI2) 662 13NHTI) +NHMTI +SFAI +MOLPI+SOLPI +LPTI+HHLPI2 + HHLDI2+CPTI +MOCPI+SOCPI + HHCDI+(CLOICHHNCI2131tit====nni 7.4.1 Two Sample T-test of between and within group differences in 1999 and 2009 Section 7.4.1 investigates between and within group differences for all 17 natural capital indexes (NCI) described in tables 7.4.1 above. The Two-Sample t-test determined changes in household NCIs between 1999 and 2009 and the degree to which between and within group difference for each NCI was significant. Three confident levels; 99.99% (*** p<0.001), 99% (**p<0.01) and 95% (*p<0.05) were used to determine the degree to which changes in NCIs were significant. Like section 7.1 through 7.3, within group difference for each NCI is computed by subtracting the mean result for each NCI in 2009 from the mean in 1999 for the same group (MTS households in Yaya or non-MTS groups in Yaya). Because the Two-Sample t-test is computed by differencing the mean value for NCI in the current year (2009) from the base year (1999), a positive difference suggests a decline in NCI while a negative difference suggests an improvement. The computation for between group differences and within group differences are similar however in the former, the changes in NCI were computed for the different research groups for a particular time period (either 1999 or 2009). 7.4.1.1 Crop Land Ownership Index (CLOI) In 1999 there was no significant difference in CLOI between MTS and Non-MTS groups however by 2009 the MTS group experienced a significantly (**p<0.01) higher (4.5%) CLOI index than Non-MTS. In 1999 NsemSaw experienced significantly (***p<0.001) higher CLOI (6.1% and 8.1% respectively) than MTS and Non-MTS. By 2009 however the gap in CLOI had closed to the point where no significant difference existed between MTS and NsemSaw as well as Non-MTS and NsemSaw. Among the three groups, only MTS experienced a significant (***p<0.001) increase (5.1%) in CLOI between 1999 and 2009. NsemSaw was the only group 663 among the three groups to experience a significant (*p<0.05) decline in CLOI between 1999 and 2009. 7.4.1.2 Household Crop Diversity Index (HHCDI) Household Crop Diversity Index is the only index among all 17 NCIs that did not experience any significant change between and within groups before (1999) and after (2009) the MTS project. 7.4.1.3 Subsistence Oriented Crop Production Index (SOCPI) In 1999 the mean SOCPI among NsemSaw households was significantly (***p<0.001) higher (7% and 6.7% respectively) than both MTS and Non-MTS groups. By 2009, the NsemSaw group experience a significantly (***p<0.001) higher SOCPI than both MTS and Non-MTS groups (8.1% and 7.9% respectively). Among all the three groups NsemSaw was the only group that experienced an increase in SOCPI between 1999 and 2009 though this change was insignificant. While the MTS and Non-MTS groups experienced a general decline in within group SOCPI between 1999 and 2009 these changes were also insignificant. 7.4.1.4 Market Oriented Crop Production Index (MOCPI) In 1999 the MTS group experienced a significantly (*p<0.05) higher (3.4%) MOCPI than the Non-MTS. By 2009 however the gap in MOCPI between both groups declined to 3% and deemed insignificant according to the t-test result. In 1999 the NsemSaw group experienced a significantly (*p<0.05) higher (4%) MOCPI than the MTS however by 2009 the gap in MOCPI between both groups declined to 1.6% and deemed insignificant by the t-test result. While there was no significant within group difference among all three groups between 1999 and 2009 it is worth noting that only NsemSaw experienced a decline in MOCPI between the two time periods. 664 7.4.1.5 Crop Production Trend Index (CPTI) In 1999 the mean CPTI among MTS was 1.2% higher than those of the Non-MTS however this difference was not significant. By 2009 however the gap in mean CPTI between both groups increase from 1.2% to 4.7% and became significant (*p<0.05). Similarly, in 1999 the mean CPTI among MTS was 0.7% higher than those of NsemSaw however this difference was insignificant. By 2009, the gap in CPTI between MTS and NsemSaw increased significantly (***p<0.001) to 12.3%. In 1999 the mean CPTI among NsemSaw was 0.6% higher but insignificant than those of the Non-MTS however by 2009 the trend reversed, and the MTS group experienced a significantly (***p<0.001) higher (7.7%) mean CPTI relative to NsemSaw. Between 1999 and 2009 all three research groups experienced a decline in mean CPTI though the change among MTS households was insignificant. The highest (12.7%) and significant (***p<0.001) within group decline in CPTI occurred among the NsemSaw group. The Non-MTS also experienced a significant (*p<0.05) decline (4.5%) in mean CPTI between 1999 and 2009. 7.4.1.6 Household Livestock Diversity Index 1 (HHLDI1) In 1999 the mean HHLDI1 among MTS was 1.8% higher than those of the Non-MTS. By 2009 however, the gap in mean HHLDI1 between both groups had increased significantly (***p<0.001) from 1.8% to 7.2%. In 1999 NsemSaw experienced a significantly (***p<0.001) higher (7.1%) mean HHLDI1 compared to MTS households. By 2009 the difference in mean HHLDI1 decreased from 7.2% to 1.4% and was insignificant. In 1999 and 2009 the mean HHLDI1 among NsemSaw were respectively 8.9% and 8.5% higher than those of Non-MTS and these differences were both significant (***p<0.001). Among all three research groups, only MTS households experienced a significant (*p<0.05) increase (3.9%) in HHLDI1 between 1999 665 and 2009. Between 1999 and 2009, both Non-MTS and NsemSaw experienced a decline (1.3% and 1.7% respectively) in HHLDI1 albeit insignificant. 7.4.1.7 Household Livestock Diversity Index 2 (HHLDI2) In 1999 the mean HHLDI2 among MTS was 2.3% higher than those of the Non-MTS group. By 2009 however, the gap in mean HHLDI2 between both groups had increased significantly (***p<0.001) from 2.3% to 9.7%. In 1999 NsemSaw experienced a significantly (**p<0.01) higher (10.3%) mean HHLDI2 compared to MTS households. By 2009 the difference in mean HHLDI2 decreased from 10.3% to 1.4% and was insignificant. In 1999 and 2009 the mean HHLDI1 among NsemSaw were respectively 12.6% and 11.1% higher than those of Non-MTS and these differences were both significant (***p<0.001 and **p<0.01 respectively). Among all three research groups, only MTS households experienced a significant (*p<0.05) increase (5.8%) in HHLDI2 between 1999 and 2009. Both Non-MTS and NsemSaw experienced a decline (1.7% and 3.2% respectively). 7.4.1.8 Household Livestock Population Index 1 (HHLPI1) In 1999 the mean HHLPI1 among MTS was 0.7% higher than those of the Non-MTS however by 2009 the gap increased significantly (***p<0.001) from 0.7% to 3.4%. Similarly, in 1999 NsemSaw experienced a significantly (***p<0.001) higher (4.7%) mean HHLPI1 compared to MTS households however by 2009 the MTS group experienced a higher (0.5%) mean HHLPI1. In 1999 the mean HHLPI1 among NsemSaw was significantly (***p<0.001) higher (5.4%) than those of the Non-MTs group. While the difference in mean HHLPI1 declined from 5.4% and 2.8% between 1999 and 2009 this difference remained significant (**p<0.01). 666 Between 1999 and 2009, all three groups experienced a decline in mean HHLPI1 however only NsemSaw experienced a significant (***p<0.001) decline (5.6%) in HHLPI1. 7.4.1.9 Household Livestock Population Index 2 (HHLPI2) In terms of significance, the intergroup differences observed for mean HHLPI2 between the two time periods (1999 and 2009) were very similar to those of mean HHLPI1 discussed in section 7.4.1.8 above (also see table (7.4b below). Hence this section skips the discussion of between group differences for HHLPI2. For the intra group difference, all three groups experienced a general decline in mean HHLPI2 between 1999 and 2009. NsemSaw experienced the largest (9%) and significant (***p<0.001) decline in HHLPI2 between the two time periods. The Non-MTS group also experienced a significantly (**p<0.01) high decline in HHLP12 between 1999 and 2009. 7.4.1.10 Livestock Production Trend Index 1 (LPTI) In 1999 the mean LPTI among MTS was significantly (*p<0.05) higher (3.2%) than those of the Non-MTS group and by 2009 the gap increased significantly (***p<0.001) from 3.2% to 4.9%. In 1999 NsemSaw had a 3.0% higher mean LPTI compared to MTS households but this difference was insignificant. By 2009 however, there was a reverse trend with the MTS group experiencing a significantly (**p<0.01) higher (3.6%) mean LPTI than NsemSaw. In 1999 the mean LPTI among NsemSaw was significantly (***p<0.001) higher (6.2%) than those of Non-MTS households. By 2009 however, the mean LPTI among Non-MTS was 1.4% higher than NsemSaw though this difference was insignificant. All three research groups experienced significant decline in their mean LPTI between 1999 and 2009 however the highest (8.7%) and most significant (***p<0.001) decline occurred among NsemSaw households. Between the two 667 time periods in question, Non-MTS also experienced a significant (**p<0.01) decline of 3.9% in mean LPTI. The least (2.1%) yet significant (*p<0.05) decline in mean LPTI occurred among MTS household. 7.4.1.11 Subsistence Oriented Livestock Production Index (SOCPI) In 1999 the mean SOCPI among Non-MTS was 2.4% higher than those of the MTS group however this difference was not significant. By 2009 however, the gap in mean SOCPI between both groups increased significantly (*p<0.05) from 2.4% to 3.8%. In 1999 NsemSaw experienced a significantly (**p<0.01) higher (5.5%) SOCPI than the MTS group and by 2009 the SOCPI gap between both groups increased significantly from 5.5% to 7.1%. According to the t-test none of the three research groups experienced any significant change in SOCPI between the two time periods. 7.4.1.12 Market Oriented Livestock Production Index (MOLPI) In 1999 the mean MOLPI among NsemSaw was significantly (*p<0.05) higher (4.1%) than those of MTS. By 2009 however the difference in MOLPI declined to an insignificant level (2.0%). In 1999 mean MOLPI among NsemSaw was (**p<0.01) higher (7.3%) than those of Non-MTS. By 2009 however MOLPI gab between both groups declined to 5.7% but still remained significant (*p<0.05). Between 1999 and 2009 all three research groups experienced a decline in mean MOLPI however these changes were all not significant. 7.4.1.13 Soil Fertilizer Application Index (SFAI) In 1999 the mean SFAI among the MTS group was higher (2.9%) than that for the Non- MTS group. By 2009 the SFAI gap between both groups increased significantly (*p<0.05) from 2.9% to 3.7%. In 1999 the MTS group experienced a significantly (**p<0.01) higher (5.5%) 668 mean SFAI than the NsemSaw and by 2009 the SFAI gap increased significantly (***p<0.001) from 5.5% to 10.5%. Like the MTS group Non-MTS households experienced a higher (2.6%) mean SFAI than NsemSaw and by 2009 the SFAI gap between both groups increased significantly (***p<0.001) from 2.6% to 6.8%. Between 1999 and 2009 all three research groups experienced significant declines in mean SFAI however the NsemSaw group recorded a significantly (***p<0.001) steeper (9.1%) decline than the MTS and Non-MTS groups. Both MTS and Non-MTS experienced respectively 4% and 4.8% decline in SFAI between 1999 and 2009 and these changes were significant (*p<0.05). 7.4.1.14 NTFP Harvests and Market Trend (NHMTI) In 1999 NsemSaw experienced a significantly (***p<0.001) higher (8.8%) mean NHMTI compared to MTS households. By 2009 however, the trends had reversed with the MTS group experiencing a significantly (**p<0.01) higher (5.8%) mean NHMTI than NsemSaw. Similar to the differences observed between MTS and NsemSaw, in 1999 NsemSaw experienced a significantly (**p<0.01) higher (7.2%) mean NHMTI compared to Non-MTS households. By 2009 however, the trend had reversed with the MTS group rather experiencing a significantly (*p<0.05) higher (5.4%) mean NHMTI than NsemSaw. Between 1999 and 2009 all three research groups experienced significant (***p<0.001) declines in mean NHMTI. The largest decline (38.1%) in NHMTI between the two time periods occurred among the NsemSaw group followed by Non-MTS (25.4%) and MTS (23.5%). 7.4.1.15 NTFP Harvesting Time Index (NHTI) In 1999 the mean NHTI among the NsemSaw group was significantly (*p<0.05) higher (5.0%) than those of the MTS group however by 2009 the trends reversed with MTS 669 experiencing a significantly (***p<0.001) higher (8.9%) mean NHTI than NsemSaw. Similarly, in 1999 NsemSaw experienced a 2.0% higher mean NHTI than Non-MTS. By 2009 however, the trends reversed with the Non-MTS experiencing a significantly (**p<0.01) higher (7.1) mean NHTI than NsemSaw. Between 1999 and 2009 all the three research groups experienced significantly (***p<0.001) high decline in mean NHTI. Among the three groups NsemSaw recorded the largest decline of 34.3% followed by Non-MTS (25.3%) and MTS (20.5%). 7.4.1.16 Combined HH Natural Capital Index 1 (CHHNCI1) In 1999 the MTS households experienced a 0.9% higher mean CHHNCI1 than the Non- MTS group. By 2009 the CHHNCI1 between both groups increased significantly (**p<0.01) from 0.9% to 2.6%. In 1999 NsemSaw experienced a significantly (***p<0.001) higher (3.6%) mean CHHNCI1 than MTS households. By 2009 the CHHNCI1 gap between both groups decreased from 3.6% to 2.1% however this difference was still significant (*p<0.05). Similarly, in 1999 NsemSaw experienced a significantly (***p<0.001) higher (4.5%) mean CHHNCI1 than Non-MTS households. By 2009 however, the CHHNCI1 gap between NsemSaw declined from 4.5% to 0.5% and this difference was not significant. Between 1999 and 2009 all three research groups experienced a significant (***p<0.001) declines in mean CHHNCI1. The NsemSaw group recorded the highest decline of 8.9% followed by Non-MTS (4.9%) and MTS (3.2%). 7.4.1.17 Combined HH Natural Capital Index 2 (CHHNCI2) In 1999 the MTS households experienced a 1% higher mean CHHNCI2 than the Non- MTS group. By 2009 the CHHNCI2 between both groups increased significantly (**p<0.01) from 1% to 2.1%. In 1999 NsemSaw experienced a significantly (***p<0.001) higher (4%) mean CHHNCI2 than MTS households. By 2009 the CHHNCI2 gap between both groups decreased 670 from 4% to 2.1% however this difference remained significant (*p<0.05). Similarly, in 1999 NsemSaw experienced a significantly (***p<0.001) higher (5%) mean CHHNCI2 than Non- MTS households. By 2009 however, the CHHNCI2 gap between NsemSaw declined from 5% to 0.8%. Between 1999 and 2009 all three research groups experienced a significant (***p<0.001) declines in mean CHHNCI2. The NsemSaw group recorded the highest decline of 9.3% followed by Non-MTS (5%) and MTS (3.2%). Between and within group differences for CHHNCI2 for the two time periods (1999 and 2009) are very similar to those of CHHNCI1 discussed under section 7.4.1.16 above. Table 7.4.2 below presents a summary of the results discussed under sections 7.4.1.1 through 7.4.1.17 above. 671 Table 7.4.2: T-Test for Estimated Differences in Natural Capital Indexes Between and Within Groups MTS HH in Yaya (n=203) Non-MTS HH in Yaya (n=120) Estimated Estimated Differences Between Groups Differences Within Groups NsemSaw (n=116) MTS & Non-MTS HH in Yaya MTS in Yaya & NsemSaw Non-MTS in Yaya & NsemSaw MTS HH in Yaya Non-MTS in Yaya NsemSaw Year M SD M SD M SD M 2009 0.595 0.157 0.550 0.174 0.570 0.157 NATURAL CAPITAL INDEX 1. CLOI 1999 0.542 0.170 0.523 0.183 0.604 0.157 2009 0.744 0.281 0.739 0.243 0.746 0.234 2. HHCDI 1999 0.743 0.284 0.736 0.249 0.743 0.237 2009 0.571 0.176 0.573 0.180 0.652 0.134 3. SOCPI 1999 0.577 0.174 0.581 0.175 0.648 0.137 2009 0.633 0.193 0.603 0.180 0.619 0.147 4. MOCPI 1999 0.627 0.181 0.594 0.172 0.640 0.150 -0.045** (0.019) -0.020 (0.020) -0.005 (0.031) -0.007 (0.031) 0.002 (0.020) 0.579 (0.010) -0.030 (0.022) -0.034* (0.020) M -0.025 (0.018) 0.061*** (0.019) 0.002 (0.031) 0.000 (0.031) M 0.020 (0.022) 0.081*** (0.022) 0.007 (0.031) 0.007 (0.032) 0.081*** (0.019) 0.070*** (0.019) 0.079*** (0.021) 0.067*** (0.021) -0.014 (0.021) 0.012 (0.020) 0.016 (0.021) 0.046* (0.021) M M M -0.051*** (0.016) -0.025 (0.023) 0.036* (0.021) -0.001 (0.028) -0.003 (0.032) -0.003 (0.031) 0.006 (0.017) 0.008 (0.023) -0.004 (0.018) -0.006 (0.019) -0.010 (0.023) 0.020 (0.019) 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) Standard Errors in parentheses; 6) *** p<0.001, ** p<0.01, * p<0.05 672 Table 7.4.2 (cont’d) Estimated Estimated Differences Between Groups Differences Within Groups MTS HH in Yaya (n=203) Non-MTS HH in Yaya (n=120) NsemSaw (n=116) MTS & Non-MTS HH in Yaya M -0.047* (0.023) -0.012 (0.023) -0.072*** (0.011) -0.018 (0.022) -0.097** (0.034) -0.023 (0.034) MTS in Yaya & NsemSaw Non-MTS in Yaya & NsemSaw MTS HH in Yaya Non-MTS in Yaya NsemSaw M -0.123*** (0.024) -0.007 (0.024) 0.014 (0.024) 0.071*** (0.024) 0.014 (0.035) 0.103** (0.036) -0.005 (0.010) 0.047*** (0.012) M -0.077** (0.027) 0.006 (0.027) 0.085*** (0.026) 0.089*** (0.026) 0.111** (0.039) 0.126*** (0.039) 0.028** (0.010) 0.054*** (0.014) M M M 0.010 (0.020) 0.045* (0.026) 0.127*** (0.028) -0.039* (0.019) 0.013 (0.023) 0.017 (0.029) -0.058* (0.030) 0.017 (0.038) 0.032 (0.040) 0.004 (0.009) 0.030 (0.011) 0.056*** (0.013) NATURAL CAPITAL INDEX 5. CPTI Year M SD M SD M SD 2009 0.588 0.198 0.542 0.207 0.465 0.211 1999 0.598 0.205 0.586 0.200 0.592 0.217 2009 0.300 0.192 0.228 0.178 0.314 0.221 6. HHLDI1 1999 0.260 0.194 0.242 0.179 0.331 0.216 2009 0.475 0.298 0.378 0.293 0.489 0.303 7. HHLDI2 1999 0.417 0.304 0.394 0.290 0.520 0.312 8. HHLPI1 2009 0.361 0.088 0.327 0.067 0.355 0.081 1999 0.365 0.099 0.358 0.096 0.411 0.117 -0.034*** (0.009) -0.007 (0.011) 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) Standard Errors in parentheses; 6) *** p<0.001, ** p<0.01, * p<0.05 673 Table 7.4.2 (cont’d) Estimated Differences Between Estimated Differences Within MTS HH in Yaya (n=203) Non-MTS HH in Yaya (n=120) NsemSaw (n=116) MTS and Non-MTS NATURAL CAPITAL INDEX 9. HHLPI2 Year M SD M SD M SD 2009 0.426 0.141 0.378 0.111 0.417 0.125 1999 0.436 0.158 0.426 0.157 0.507 0.183 2009 0.324 0.132 0.275 0.091 0.288 0.108 10. LPTI 1999 0.345 0.134 0.313 0.119 0.375 0.146 2009 0.474 0.183 0.512 0.187 0.546 0.175 11. SOLPI 1999 0.476 0.183 0.500 0.190 0.531 0.176 2009 0.546 0.207 0.508 0.203 0.565 0.201 12. MOLPI 1999 0.547 0.192 0.515 0.208 0.588 0.202 Groups MTS in Yaya & NsemSaw SD -0.009 (0.016) 0.072*** (0.020) -0.036** (0.014) 0.030 (0.016) 0.071*** (0.021) 0.055** (0.021) 0.020 (0.024) 0.041* (0.023) HH in Yaya M -0.048*** (0.015) -0.009 (0.018) -0.049*** (0.014) -0.032* (0.015) 0.038* (0.021) 0.024 (0.021) -0.037 (0.024) -0.032 (0.023) Groups Non-MTS in Yaya & NsemSaw MTS HH in Yaya Non- MTS in Yaya NsemSaw M SD M M 0.039** (0.015) 0.081*** (0.022) 0.009 (0.015) 0.049** (0.018) 0.090*** (0.021) 0.021* (0.013) 0.039** (0.014) 0.002 (0.018) -0.012 (0.024) 0.087*** (0.017) -0.015 (0.023) 0.001 (0.020) 0.006 (0.027) 0.023 (0.026) 0.014 (0.013) 0.062*** (0.017) 0.034 (0.024) 0.030 (0.024) 0.057* (0.026) 0.073** (0.027) 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) Standard Errors in parentheses; 6) *** p<0.001, ** p<0.01, * p<0.05 674 MTS HH in Yaya (n=203) Non-MTS HH in Yaya (n=120) NsemSaw (n=116) MTS and Non-MTS Table 7.4.2 (cont’d) Estimated Differences Between Estimated Differences Within Groups MTS in Yaya & NsemSaw Non-MTS in Yaya & NsemSaw Groups MTS HH in Yaya Non-MTS in Yaya NsemSaw SD M SD M M -0.105*** -0.068*** (0.022) -0.055** (0.021) -0.058** (0.024) 0.088*** (0.026) -0.089*** (0.026) 0.050* (0.025) -0.021* (0.011) 0.036*** (0.010) (0.021) -0.026 (0.024) -0.054* (0.026) 0.072** (0.029) -0.071** (0.029) 0.020 (0.025) 0.005 (0.012) 0.045*** (0.011) 0.040* (0.020) 0.048* (0.023) 0.091*** (0.022) 0.235*** (0.022) 0.254*** (0.029) 0.381*** (0.027) 0.205*** (0.023) 0.253*** (0.028) 0.343*** (0.025) 0.032*** (0.009) 0.049*** (0.011) 0.089*** (0.012) HH in Yaya M -0.037* (0.023) -0.029 (0.021) -0.003 (0.024) 0.016 (0.027) -0.018 (0.027) 0.030 (0.026) -0.026** (0.011) -0.009 (0.010) NATURAL CAPITAL INDEX 13. SFAI Year M SD M SD M SD 2009 0.383 0.210 0.346 0.171 0.278 0.154 1999 0.424 0.180 0.395 0.185 0.369 0.178 2009 0.472 0.211 0.468 0.208 0.414 0.199 14. NHMTI 1999 0.706 0.236 0.721 0.237 0.794 0.207 2009 0.626 0.234 0.608 0.230 0.537 0.213 15. NHTI 16. CHHNCI1 1999 0.830 0.237 0.860 0.208 0.880 0.165 2009 0.509 0.091 0.483 0.092 0.488 0.096 1999 0.541 0.089 0.532 0.086 0.577 0.087 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) Standard Errors in parentheses; 6) *** p<0.001, ** p<0.01, * p<0.05 675 Table 7.4.2 (cont’d) Estimated Differences Between Estimated Differences Within NATURAL CAPITAL INDEX 17. CHHNCI2 MTS HH in Yaya (n=203) Non-MTS HH in Yaya (n=120) NsemSaw (n=116) MTS and Non-MTS Year M SD M SD M SD 2009 0.527 0.098 0.498 0.100 0.506 0.101 1999 0.558 0.097 0.548 0.093 0.599 0.097 Groups MTS in Yaya & NsemSaw SD -0.021* (0.012) 0.040*** (0.011) Groups Non-MTS in Yaya & NsemSaw MTS HH in Yaya Non-MTS in Yaya NsemSaw M SD M M 0.008 (0.013) 0.050*** (0.012) 0.032*** (0.010) 0.050*** (0.012) 0.093*** (0.013) HH in Yaya M -0.029** (0.011) -0.010 (0.011) 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) Standard Errors in parentheses; 6) *** p<0.001, ** p<0.01, * p<0.05 676 7.4.2 Summary Results Among all three research groups only the MTS group experienced significant improvements in at least one natural capital index between 1999 and 2009. Out of the 17 natural capital indexes (NCIs) included in the study, the MTS group experienced significant improvements in CLOI, HHLDI1 and HHLDI2 indexes (see tables 7.3.2a and 7.3.2b above and 7.3.3 below). Of the three research groups, NsemSaw experienced significant decline in 10 out of the 17 NCIs while the Non-MTS group in Yaya experienced significant decline in nine NCIs. The MTS group experienced significant decline in six out of the 17 NCIs. According to the t-test, all three research groups experienced significant (*** p<0.001) decline in NHMTI, NHTI, CHHNCI1 and CHHNCI2 indexes between 1999 and 2009. Similarly, all three groups experienced significant (*p<0.05) decline in SFAI between the two time periods. Of the three groups, NsemSaw experienced a significant (*** p<0.001) decline in four other NCIs (CPTI, HHLPI1, HHLPI2 and LPTI) between the two time periods. Between 1999 and 2009, NsemSaw was the only group that experienced a significant (*p<0.05) decline in CLOI. Like NsemSaw, Non-MTS households also experienced a significant decline in CPTI (*p<0.05), HHLPI1 (*p<0.05), HHLPI2 (**p<0.01) and LPTI (**p<0.01). The discussions in this section suggest that between 1999 and 2009 all three research groups performed rather poorly on majority of the NCIs however the NsemSaw households performed the worst followed by Non-MTS and then the MTS group. Since CHHNCI1 and CHHNCI2 both captures changes in 13 of the 15 key NCIs (see table 7.4.1 above) and appear to be very similar in magnitude and significance across all three research groups further analysis in section 7.6 uses livelihood pentagons to visually illustrate changes in CHHNCI1 between and within all three groups for the two time periods. Following section 7.6, the question of what 677 might have caused the MTS group to record significantly higher improvements in three NCIs relative to both Non-MTS and NsemSaw is addressed in chapter eight. Using a Difference-in- Difference technique chapter eight isolate changes in CHHNCI1 index that may be attributed directly to the MTS project or spillover from the project. 678 Table 7.4.3: Summary of T-Test for Estimated Differences Between and Within Groups MTS & Non-MTS Mean (MTS – Non- No. Index MTS) MTS & NsemSaw Non & MTS-NsemSaw Mean (MTS – NsemSaw) Mean (Non-MTS – MTS Non-MTS NsemSaw NsemSaw) 1999 2009 1999 2009 1999 2009 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. CLOI HHCDI SOCPI MOCPI CPTI HHLDI1 HHLDI2 HHLPI1 HHLPI2 LPTI SOLPI MOLPI SFAI NHMTI NHTI CHHNCI1 CHHNCI2 + + - + + + + + + + - + + - - + + ++ + - + + +++ ++ +++ +++ +++ - + + + + ++ ++ --- - --- - + --- -- --- --- - -- - ++ --- - --- --- + - --- + +++ - - + + ++ --- - +++ ++ +++ + + --- - --- - - --- --- --- --- ++ - -- + -- - --- --- - - --- - ++ --- -- -- -- +++ - - +++ + ++ + - 2009- 1999 +++ + - + - + + - - - - - - --- --- --- --- 2009-1999 2009- 1999 + + - + - - - - -- -- + - - --- --- --- --- - + + - --- - - --- --- --- + - - --- --- --- --- 1) Significant improvements in livelihood index: +++p<0.001, ++ p<0.01, + p<0.05, 2) Significant decline in livelihood index: ---p<0.001, --p<0.01, - p<0.05 3) Insignificant improvements in livelihood index: + and 4) Insignificant decline in livelihood index: - 679 7.5 Social Capital This section discusses changes in 14 Social Capital Indexes (SCI) among the three research groups (MTS, Non-MTS and NsemSaw) before (1999) and after (2009) the MTS project. Within the context of the study, SCIs capture the different social networks, and relationships households develop with other family members and friend, local organizations, public and civil society groups and how these networks are harnessed to solve different socio- economic and financial problems that arise within the household (refer to chapter three sections 3.3.2 for a broader discussion on social capita). The 14 SCIs discussed thus encompass five broad categories including; a) family networks developed within and outside of the village and support drawn from such networks; b) networks developed with public and civil society groups; c) networks developed within local community groups that are not religions in nature (e.g. local savings and loans groups, farming groups and water user groups); d) networks developed within religious groups, and e) others (population/size of the household, joint household activities and the impact of migration on the household). Table 7.5.1 below describes how each of the 14 SCIs were generated while tables 7.5.2a through 7.5.2c presents the results of a Two Sample T-test used to track (within and between group) changes in SCIs before (1999) and after (2009) the MTS project. The tables presented in Appendix S describes in greater detail how each SCI was constructed from the household survey results. Section 7.5.2 and table 7.5.2d summarize all the changes that occurred between and within each of the research groups before (1999) and after (2009) the MTS project. 680 Table 7.5.1: Definition of Household Social Capital Indexes No . Index Name of Index/ Variable 1 FNWI Family Network Index 2 HRITVI Help from Relatives Inside the Village Index 3 HROTV 4 HROTB Help from Relatives Outside the Village Index Help from Relatives Outside the BA Region Index Description of Indexes Computation of Indexes FNWI captures the number of close relatives that do not reside within the HH but live in the same village. HH with no close relatives or immediate family member in their village earn a score of 1 while those with more than 10 relatives earn 4. FNWI thus take on scores ranging from 0.25 to 1. HRITVI captures three types of assistance or help from relatives residing in the same village as the HH. Depending on how often a HH receives assistance in the form of “farm labor, food or cash” from relatives in the village, the household may attain a total score of 12. HRITVI expressed in terms of the level of assistance from close relatives in the village takes on a values ranging from 0.25 to 1. HROTV is very similar to HRITVI only difference is HROTV assistance from relatives residing outside of the village but in the in the same BA Region. HROTB is also similar to HRITVI only this index emphasizes help from relatives outside of the BA region but resident in some part of the country. 681 4(FNWKS)FNWI11itt===ni12RITV3Q) +RITV2Q +(RITV1QHRITV31tit===ni12ROTV_B3Q) ROTV_B2Q (ROTV_B1QHROTV31tit==++=ni12ROTB_G3Q) ROTB_G2Q (ROTB_G1QHROTB31tit==++=ni Table 7.5.1 (cont’d) Description of Indexes Computation of Indexes No. Index 5 HROGI Name of Index/ Variable Help from Relatives Outside of Ghana Index HROGI captures two types of assistance or help from relatives residing abroad/outside of Ghana. Depending on how often a HH receives assistance in the form of “food or cash” from relatives living abroad, the household may attain a total score of 8. HROGI expressed in terms of the level of assistance takes on a values ranging from 0.25 to 1. 6 GNSI Government and NGO Support Index GNSI captures Government support to HH in the form of extension advice, production material, food and cash. Depending on the level of government support, HH may attain a maximum score of 16 on this particular indicator. GNSI expressed in terms of the level of government ranges from 0.25 to 1. 7 OFANF Other forms of Support from Neighbors and Friends Index OFANF captures support in the form of extension advice, production material, food and cash from neighbors and friends. Depending on how often a HH receives this type of assistance the household may attain a total score of 8. OFANF expressed in terms of the level of assistance ranges from 0.25 to 1. 8 CAMI Community and Association Membership Index CAMI captures association memberships and group meetings. Depending on whether HH belong to a group or association and/or how often they attend association meetings at least once a year they attain a total score of two. CAMI ranges from 0 to 1. 682 8ROG2Q)+ (ROG1QHROGI21tit===ni16GNGO4Q) GNGO3Q GNGO2Q (GNGO1QGNSI41tit==+++=ni8SFNF2Q) +(SFNF1QOFANF21tit===ni2ngs)AssocMeeti +r(AsscMembeCAMI21tit===ni No. Index 9 RGMMA 10 RGSI Name of Index/ Variable Religious Group Membershi p and Meeting Attendance Index Religious Group Support Index HH 11 HHPOPI Population Index Table 7.5.1 (cont’d) Description of Indexes Computation of Indexes RGMMA captures the extent of religious group membership and meeting attendance. Households attain a total score of 2 if they belong to a religious organization and attend group meetings at least once a year. The RGMMA index thus ranges from 0 to 1. RGSI captures the extent to which the HH receives three categories of support from religious organizations. Support in the form of counseling or advice in resolving family conflicts, support in the form of cash, food and cloths and support in the form farm labor contributions may earn a household a total score of 12. A household’s RGSI ranges from 0.25 to 1. HHPOPI captures the number of individuals living in a HH. Four response categories with boundaries: 1=less than five (Very Few) and 4=more than 15 (Very Many) capture the number of individuals living in a HH. Because the survey communities are predominantly engaged in farming and forestry which requires manual labor, households with very many individuals earns a higher total score of 4 while those with very few individuals earn a score of 1. A household’s HHPOPI index thus ranges from 0.25 to 1. 683 2)FCMA1_Freq +(GA3aRGMMA21tit===ni12FCMA4Q) + FCMA3Q+(FCMA2QRGSI31tit===ni4*1(HHPOPI)HHPOPI11itt===ni No. Index 12 JHHAI Name of Index/ Variable Joint HH Activity Index HH 13 HMIGI Migration Index Table 7.5.1 (cont’d) Description of Indexes Computation of Indexes JHHAI captures the degree to which individuals within the HHs depend on each other for farming, cooking, and sharing of postharvest storage facilities. HHs members with a high degree of interdependence amongst themselves attain a total score of 6 and while those with the least attain 3. JHHAI index ranges from 0.5 to 1. HMIGI captures a HH’s work related migration activities. While work related migration for more than three months (long-term) negatively impacts farming or forestry activities, this serves as a risk mitigating strategy against crop failures and forest canopy closure (canopy closure renders crop production activities uneconomical). A HH earns a score of 1 if at least one member migrates outside for work or 0 if no one migrates for work. Depending on the number of HH members migrate for work, HHs may earn 1 (None) or 4 (very Many). The maximum attainable score for work related migration is t5 and the minimum is 1. HMIGI index ranges from 0.20 to 1. 14 AHHSCI Aggregate HH Social Capital Index AHHSCI captures a HH’s aggregate social capital endowment. This index is simply the average of all the 13 indexes hence the maximum attainable score on Social Capita is 13 and the minimum is. AHHSCI index this ranges from 0.08 and 1. 684 6T4Q) HH_LIV_AC+T2Q HH_LIV_AC+T1Q(HH_LIV_ACJHHAI31tit===ni5ABSENTQ) +(ABSENT_1QHMIGI21tit===ni13 HMIGI)+JHHAI + HHPOPI+ RGSI+ RGMMA+CAMI +OFANF +GNSI + HROGI+ HROTB+ HROTV+ HRITVI+(FNWIAHHSCI131tit====nni 7.5.1 Two Sample T-test of between and within group differences in 1999 and 2009 Section 7.5.1 investigates between and within group differences for all 14 social capital indexes (SCI) described in tables 7.5.1 above. The Two-Sample t-test determined changes in household SCIs between 1999 and 2009 and the degree to which between and within group difference for each SCI was significant. Three confident levels; 99.99% (*** p<0.001), 99% (**p<0.01) and 95% (*p<0.05) were used to determine the degree to which changes in SCIs were significant. Like section 7.1 through 7.4, within group difference for each SCI is computed by subtracting the mean result for each SCI in 2009 from the mean in 1999 for the same group (MTS households in Yaya or non-MTS groups in Yaya). Because the Two-Sample t-test is computed by differencing the mean value for SCI in the current year (2009) from the base year (1999), a positive difference suggests a deterioration in SCI while a negative difference suggests an improvement. The computation for between group differences and within group differences are similar however in the former, the changes in SCI were computed for the different research groups for a particular period (either 1999 or 2009). 7.5.1.1 Family Network Index (FNWI) The results of the t-test suggest no statistically significant difference between the three research groups in 1999 and 2009. Within group comparison however suggests that only NsemSaw households experienced a significant (*p<0.05) increase (5.6%) in FNWI between 1999 and 2009. While both MTS and Non-MTS also experienced an increase in FNWI between the two time periods these changes were not significant. 685 7.5.1.2 Help from Relatives Inside the Village Index (HRITVI) In 1999 HRITVI was significantly (**p<0.01) higher (5.3%) among Non-MTS than then the MTS group. In 2009 however, the gap in HRITVI between both groups had declined from 5.3% to 3.6% and this difference according to the t-test was not significant. Similarly, in 1999 HRITVI among NsemSaw was significantly (**p<0.01) higher (6%) than those of MTS. While HRITVI gap between both groups declined from 6% to 5.1% by 2009 the difference remained significant (*p<0.05). 7.5.1.3 Help from Relatives Outside the Village Index (HROTV) Among all three groups, only the MTS group recorded an increase (0.9%) in HROTV between 1999 and 2009. Both the Non-MTS in Yaya and NsemSaw groups experienced a decline of 1.3% and 0.1% respectively in HROTV index. Though the MTS group appears to have performed better on the HROTV index, the t-test result suggests that none of the within group differences were significant. Similarly, the between group differences in 1999 and 2009 were insignificant among all three groups according to the t-test results. 7.5.1.4 Help from Relatives Outside of the Brong Ahafo Region (HROTB) The Non-MTS and NsemSaw groups appeared to have slightly improved HROTB index in 1999 and 2009 relative to the MTS group however none of these between group differences were significant according to the t-test results. With the exception of NsemSaw that experienced an increase in HROTB between 1999 and 2009, both MTS and non-MTS groups in Yaya experienced a slight decline in HROTB for the same period. According to the t-test results, the within group difference recorded for all three groups was insignificant. 686 7.5.1.5 Help from Relatives Outside of Ghana Index (HROGI) In 1999 the mean HROGI among Non-MTS in Yaya was significantly (**p<0.01) higher (3.9%) higher than that of NsemSaw. Though the HROGI index between both groups declined from 3.9% to 3.7% by 2009 this difference according to the t-test was still significant. Between 1999 and 2009 there was no statistically significant change in mean HROG within all three groups. 7.5.1.6 Government and NGO Support Index (GNSI) In 1999 the mean GNSI among the MTS households was significantly (***p<0.001) higher (4.1%) than the Non-MTS group. By 2009 the GNSI gap between both groups increased from 4.1% to 5.5% and remained highly significant (***p<0.001). In 1999 the mean GNSI among NsemSaw was significantly (**p<0.01) higher (4.1%) than Non-MTS and by 2009 the GNSI gap between both groups remained the same in both magnitude and significance. While all three research groups recorded an increase in mean GNSI between 1999 and 2009 only the MTS group experienced a significant (*p<0.05) increase (2.1%) between the two time periods. 7.5.1.7 Other forms of Support from Neighbors and Friends Index (OFANF) In 1999 the mean OFANF index among the MTS group was significantly (**p<0.001) higher (3.6%) than that of Non-MTS and by 2009 the gap in OFANF index between both groups increased significantly (***p<0.001) from 3.6% to 5.7%. Similarly, in 1999 the MTS group also experienced a higher (2.1%) albeit insignificant mean OFANF than NsemSaw and by 2009 the OFANF gap increased significantly (**p<0.01) from 2.1% to 4.8%. 687 7.5.1.8 Community and Association Membership Index (CAMI) In 1999 the mean CAMI index among MTS was significantly (**p<0.001) higher (24.1%) than that of Non-MTS and by 2009 the gap in the mean CAMI index between both groups increased significantly (***p<0.001) from 24.1% to 42%. Similarly, in 1999 the MTS group also experienced a significantly (*p<0.05) higher (8.3%) mean CAMI index than NsemSaw and by 2009 the gap CAMI index between both groups increased significantly (***p<0.001) from 8.3% to 16.7%. Like the MTS group, NsemSaw also experienced a significantly (**p<0.01) higher (15.8%) mean CAMI index than the Non-MTS group in 1999. By 2009 however, the gap in mean CAMI index between the NsemSaw and Non-MTS households increased significantly (***p<0.001) from 15.8% to 25.3%. Between 1999 and 2009 all three research groups experienced an increase in their mean CAMI index however the change among the MTS group was the highest (18.7%) and most significant (***p<0.001). Next to the MTS group, the NsemSaw group also experienced a significantly (*p<0.05) high (10.3%) increase in mean CAMI between 1999 and 2009. While the Non-MTS group also experienced an increase in mean CAMI between 1999 and 2009, this change was insignificant. 7.5.1.9 Religious Group Membership and Meeting Attendance Index (RGMMA) In 1999 the mean RGMMA index among MTS households was higher (3.5%) than that of the Non-MTS group however the difference was not significant. By 2009 however, the mean RGMMA gap between both groups increased significantly (*p<0.05) from 3.5% to 5.2%. Similarly, in 1999 the mean RGMMA among MTS households was significantly (*p<0.05) higher (6.7%) than the NsemSaw group and by 2009 the gap had increased significantly (***p<0.001) from 6.7% to 10.1%. 688 7.5.1.10 Religious Group Support Index (RGSI) Between 1999 and 2009 only the MTS group recorded an increase (1%) in the mean RGSI index. Both Non-MTS and NsemSaw registered a decline of 0.6% and 1.3%. The t-test result suggests that none of the within group changes recorded between 1999 and 2009 were significant. The within group changes in RGSI however were not significant according to the t- test result. In 1999, the RGSI index for the MTS group was lower than both non-MTS in Yaya and the NsemSaw group albeit not significant. However, in 2009, the RGSI index for the MTS group was 1% higher than the Non-MTS and 0.9% higher than the NsemSaw groups albeit not significant according to the t-test results. 7.5.1.11 Household Population Index (HHPOPI) In 1999 the mean HHPOPI was significantly (*p<0.05) higher (4.9%) among Non-MTS households than the NsemSaw. By 2009, the HHPOPI gap shrunk from 4.9% to 1.8% and this difference according to the t-test was not significant. Between 1999 and 2009, the MTS and NsemSaw groups experienced significant (*p<0.05) increases (4.9% and 6% respectively) in their mean HHPOPI. While the Non-MTS group also recorded a 2.9% increase in HHPOPI between 1999 and 2009 this change was insignificant. 7.5.1.12 Joint Household Activity Index (JHHAI) In 1999 the mean JHHAI was significantly (***p<0.001) higher (7.3%) among NsemSaw households than the Non-MTS group. By 2009 however the JHHAI gap between the two groups decreased from 7.3% to 5.7% and still remained significant (**p<0.01). In 1999 the mean JHHAI was significantly (**p<0.01) higher (6.5%) among NsemSaw than the Non-MTS group and by 2009 the JHHAI gap increased from 6.5% to 6.6% and still remained significant. 689 7.5.1.13 Household Migration Index (HMIGI) In 1999 the mean HMIGI was higher (2.0%) among the MTS households than NsemSaw however this difference was insignificant. By 2009 however, the mean HMIGI gap between both groups increased significantly (**p<0.01) from 2% to 7%. In 1999 the mean HMIGI among Non-MTS households was significantly (*p<0.05) higher (4.7%) than NsemSaw. By 2009 the HMIGI gap between both groups still remained significant (*p<0.05) having increased from 4.7% to 5.1%. Between 1999 and 2009 all three research groups experienced an increase in HMIGI however only the change (7.6%) among MTS group was significant (***p<0.001). 7.5.1.14 Aggregate Household Social Capital Index (AHHSCI) In 1999 the mean AHHSCI among MTS households was higher (1.3%) than the Non- MTS group however the difference was not significant. By 2009 however, the gap in AHHSCI between both groups increased significantly(***p<0.001) from 1.3% to 4.2%. Similarly, in 1999 the MTS group experienced a slightly higher (0.8%) mean AHHSCI index than NsemSaw however this difference was not significant. By 2009 however, the AHHSCI gap between both groups increased significantly (**p<0.01) from 0.8% to 2.6%. Between 1999 and 2009 all three research groups experienced an increase in their AHHSCI index however only the change (3.5%) among the MTS group was significant (***p<0.001). Table 7.4.2 below presents a summary of the results discussed under sections 7.5.1.13 through 7.5.1.17. 690 Table 7.5.2: T-Test for Estimated Differences in Natural Capital Indexes Between and Within Groups MTS HH in Yaya (n=203) Non-MTS HH in Yaya (n=120) NsemSaw (n=116) MTS & Non-MTS SOCIAL CAPITAL INDEX 1. FNWI Year M SD M SD M SD 2009 0.685 0.267 0.696 0.264 0.700 0.241 1999 0.648 0.270 0.669 0.275 0.644 0.247 2009 0.468 0.199 0.504 0.206 0.519 0.217 2. HRITVI 1999 0.460 0.193 0.513 0.210 0.520 0.217 2009 0.424 0.163 0.423 0.145 0.419 0.155 3. HROTV 1999 0.415 0.151 0.435 0.149 0.420 0.160 4. HROTB 2009 0.386 0.145 0.399 0.148 0.381 0.143 1999 0.388 0.146 0.406 0.150 0.379 0.144 Estimated Estimated Differences Between Groups Differences Within Groups MTS in Yaya & NsemSaw Non-MTS in Yaya & NsemSaw MTS HH in Yaya Non-MTS in Yaya NsemSaw M 0.016 (0.030) -0.003 (0.030) 0.051* (0.024) 0.060** (0.024) -0.005 (0.019) 0.005 (0.018) -0.005 (0.017) -0.008 (0.017) M 0.005 (0.033) -0.024 (0.034) 0.015 (0.028) 0.008 (0.028) -0.004 (0.020) -0.015 (0.020) -0.018 (0.019) -0.026 (0.019) M M M -0.037 (0.027) -0.027 (0.035) -0.056* (0.032) -0.008 (0.019) 0.009 (0.027) 0.001 (0.029) -0.009 (0.016) 0.013 (0.019) 0.001 (0.021) 0.001 (0.014) 0.007 (0.019) -0.002 (0.019) HH in Yaya M 0.011 (0.031) 0.021 (0.031) 0.036 (0.023) 0.053** (0.023) -0.001 (0.018) 0.020 (0.017) 0.013 (0.017) 0.018 (0.017) 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) Standard Errors in parentheses; 6) *** p<0.001, ** p<0.01, * p<0.05 691 Table 7.5.2 (cont’d) Estimated Estimated Differences Between Groups Differences Within Groups MTS HH in Yaya (n=203) Non-MTS HH in Yaya (n=120) NsemSaw (n=116) MTS & Non-MTS HH in Yaya M 0.021 (0.013) 0.021 (0.013) -0.055*** (0.014) -0.041*** (0.013) MTS in Yaya & NsemSaw Non-MTS in Yaya & NsemSaw MTS HH in Yaya Non-MTS in Yaya NsemSaw M -0.017 (0.012) -0.017 (0.012) -0.013 (0.016) 0.000 (0.013) -0.048** (0.017) -0.021 (0.017) (0.049) -0.083* (0.052) M -0.037** (0.015) -0.039** (0.016) 0.041** (0.016) 0.041** (0.014) 0.009 (0.018) 0.015 (0.019) 0.253*** (0.058) 0.158** (0.057) M M M 0.000 (0.010) 0.001 (0.018) -0.001 (0.013) -0.021* (0.012) -0.008 (0.014) -0.008 (0.016) -0.021 (0.014) 0.000 (0.017) 0.007 (0.019) -0.187*** (0.041) -0.008 (0.055) -0.103* (0.061) SOCIAL CAPITAL INDEX 5. HROGI Year M SD M SD M SD 2009 0.299 0.101 0.320 0.136 0.282 0.098 1999 0.299 0.102 0.320 0.139 0.281 0.103 2009 0.397 0.133 0.342 0.106 0.384 0.134 6. GNSI 1999 0.374 0.117 0.334 0.108 0.374 0.010 7. OFANF 2009 0.464 0.140 0.407 0.131 0.416 0.147 1999 0.442 0.139 0.406 0.136 0.421 0.014 -0.057*** (0.016) -0.036** (0.016) 2009 0.766 0.396 0.346 0.424 0.599 0.463 8. CAMI 1999 0.579 0.435 0.338 0.421 -0.420*** -0.167*** (0.047) -0.241*** 0.496 0.459 (0.050) 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) Standard Errors in parentheses; 6) *** p<0.001, ** p<0.01, * p<0.05 692 Table 7.5.2 (cont’d) Estimated Differences Between Estimated Differences Within MTS HH in Yaya (n=203) Non-MTS HH in Yaya (n=120) NsemSaw (n=116) MTS and Non-MTS SOCIAL CAPITAL INDEX 9. RGMMA Year M SD M SD M SD 2009 0.877 0.253 0.825 0.308 0.776 0.338 1999 0.847 0.279 0.813 0.297 0.780 0.332 2009 0.457 0.160 0.447 0.145 0.448 0.166 10. RGSI 1999 0.448 0.159 0.453 0.145 0.461 0.178 11. HHPOPI 2009 0.544 0.212 0.540 0.213 0.522 0.185 1999 0.495 0.242 0.510 0.235 0.461 0.215 2009 0.854 0.200 0.844 0.198 0.911 0.175 12. JHHAI 1999 0.844 0.197 0.851 0.194 0.917 0.171 Groups MTS in Yaya & NsemSaw SD -0.101*** (0.033) -0.067* (0.035) -0.009 (0.019) 0.013 (0.019) -0.023 (0.024) -0.034 (0.027) 0.057** (0.022) 0.073*** (0.022) Groups Non-MTS in Yaya & NsemSaw MTS HH in Yaya Non- MTS in Yaya NsemSaw M -0.049 (0.042) -0.032 (0.041) 0.001 (0.020) 0.008 (0.021) -0.018 (0.026) -0.049* (0.029) 0.066** (0.024) 0.065** (0.024) SD M M -0.030 (0.026) -0.013 (0.039) 0.004 (0.044) -0.010 (0.016) 0.006 (0.019) -0.049* (0.023) -0.029 (0.029) 0.013 (0.023) -0.060* (0.026) -0.010 (0.020) 0.007 (0.025) 0.006 (0.023) HH in Yaya M -0.052* (0.032) -0.035 (0.033) -0.010 (0.018) 0.006 (0.018) -0.005 (0.024) 0.015 (0.028) -0.009 (0.023) 0.007 (0.023) 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) Standard Errors in parentheses; 6) *** p<0.001, ** p<0.01, * p<0.05 693 MTS HH in Yaya (n=203) Non-MTS HH in Yaya (n=120) NsemSaw (n=116) MTS and Non-MTS Table 7.5.2 (cont’d) Estimated Differences Between Estimated Differences Within Groups MTS in Yaya & NsemSaw SD -0.070** (0.026) -0.020 (0.024) -0.026** (0.009) -0.008 (0.010) Non-MTS in Yaya & NsemSaw MTS HH in Yaya Groups Non- MTS in Yaya NsemSaw M -0.051* (0.027) -0.042* (0.025) 0.016 (0.011) 0.005 (0.011) SD M M -0.076*** (0.022) -0.035 (0.025) -0.026 (0.027) -0.035*** (0.008) -0.006 (0.010) -0.018 (0.012) HH in Yaya M -0.019 (0.025) 0.022 (0.023) -0.042*** (0.009) -0.013 (0.009) SOCIAL CAPITAL INDEX 13. HMIGI Year M SD M SD M SD 2009 0.442 0.225 0.423 0.196 0.372 0.212 1999 0.367 0.209 0.388 0.191 0.347 0.193 2009 0.543 0.075 0.501 0.075 0.518 0.091 14. AHHSCI 1999 0.508 0.079 0.495 0.082 0.500 0.086 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) Standard Errors in parentheses; 6) *** p<0.001, ** p<0.01, * p<0.05 694 7.5.2 Summary Results Out of the 14 Social Capital Indexes (SCIs) included in this study, the MTS group experienced significant improvements in six indexes (GNSI, CAMI, HHPOPI, JHHAI, HMIGI and AHHSCI) while NsemSaw saw improvements in three (FNWI, CAMI and HHPOPI) between 1999 (before MTS) and 2009 (after MTS). Among the three research groups only the Non-MTS group in Yaya did not experience significant improvement in any SCI. According to the t-test, the MTS group experienced significant (*** p<0.001) improvement in CAMI, HMIGI and AHHSCI between 1999 and 2009. Between the two time periods, only the MTS group experienced significant (*p<0.05) improvements in GNSI and JHHAI while NsemSaw was the only group that experienced significant (*p<0.05) improvements in FNWI. Between 1999 and 2009 both MTS and NsemSaw experienced significant (*p<0.05) improvements in HHPOPI while NsemSaw also experienced significant improvements (*p<0.05) in RGMMA for the same period. The discussions in this section suggest that between 1999 and 2009 the MTS group outperformed both the Non-MTS and NsemSaw group on most of the SCIs. The Non-MTS group by virtue of not experiencing any significant change in any of the 14 SCIs may be said to have performed the worst on the 14 SCIs. Since AHHSCI captures changes in all 13 key SCIs (see table 7.5.1 above) across all three research groups further analysis in section 7.6 uses livelihood pentagons to visually illustrate changes in AHHSCI between and within all three groups for the two time periods. Following section 7.6, questions regarding how much of the observed or perceived changes in SCIs among the three research groups may be attributed to the MTS project is addressed in chapter eight using the Difference-in-Difference technique. 695 Table 7.5.3: T-Test for Estimated Differences Between and Within Groups MTS & Non-MTS Mean (MTS – Non- No. Index MTS) MTS & NsemSaw Mean (MTS – NsemSaw) Non & MTS- NsemSaw Mean (Non-MTS – NsemSaw) MTS Non-MTS NsemSaw 1999 2009 1999 2009 1999 2009 2009-1999 2009- 1999 1999 1. 2. 3. 4. 5. 6. 7. 8. 9. FNWI HRITVI HROTV HROTB HROGI GNSI OFANF CAMI RGMMA 10. RGSI 11. HHPOPI 12. JHHAI 13. HMIGI - -- - - - - - + - - +++ ++ +++ +++ +++ +++ + - - - - + + + + + + -- - + + - + + + - + --- + - - + + + + ++ +++ +++ + + -- ++ ++ + - + + - - + + +++ ++ +++ ++ - -- + - + -- + - --- + - + -- + 14. AHHSCI 1) Significant improvements in livelihood index: +++p<0.001, ++ p<0.01, + p<0.05, 2) Significant decline in livelihood index: ---p<0.001, --p<0.01, - p<0.05 3) Insignificant improvements in livelihood index: + and 4) Insignificant decline in livelihood index: - + + - - +++ + + + - - + + +++ + + + + +++ +++ + - - - - + - + + - + - + + + - - + + + - + - - + - + + 696 7.6 Livelihood Asset Pentagons In the first five sections of chapter seven, I demonstrate how the asset-based approach can be used to generate livelihood indexes across five major asset categories. In the current section, I demonstrate how the different assets presented in sections 7.1 through 7.5 may be compared across all three research groups before and after MTS periods using livelihood pentagons. Using livelihood pentagons allow changes in any of the livelihood asset categories to be visually displayed and comparisons made between MTS and Non-MTS participants before and after the project implementation. 7.6.1 Comparative Analysis of Household Livelihood Indexes/Asset (HLIs) In the comparative analysis of household assets below, I first discuss within group changes in livelihood assets for MTS households in Yaya, non-MTS households in Yaya and NsemSaw households for the two time periods before MTS (1999) and after MTS (2009). For the first within group analysis, the mean aggregate livelihood assets for all three research groups in 1999 are compared with 2009 means to determine what changes if any have occurred between the two time periods for the same group. Following the aggregate analysis, a similar within group comparison is done for the disaggregated research groups starting with MTS households in Yaya for 1999 and 2009 asset levels followed by non-MTS in Yaya and finally for NsemSaw households. Unlike the within group differences that examine changes in a particular research group, between group differences investigates how two different groups for example MTS and non- MTS in Yaya differ in terms of their livelihood assets in the initial year (1999) and the final year (2009). The same between group analyses is done for MTS in Yaya versus NsemSaw and also non-MTS in Yaya versus NsemSaw for both time periods 1999 and 2009. 697 7.6.2 Interpretation of Livelihood Pentagons Each livelihood pentagon like the name suggests has five corners with each corner representing one of five household livelihood capital or asset (financial, physical, natural, human, or social capital assets). A total of five pentagons are used in each analysis with the corners of the outermost thicker (black) pentagon representing the maximum attainable level for a particular livelihood asset. Since the analysis used the combined household financial, physical, natural, human and social capital assets expressed as a total for the particular asset category (see how each combined asset group were generated 7.1 through 7.5), the maximum attainable score for each asset group is one or 100%. The outer corners of each pentagons thus represent the levels of each livelihood asset accumulated by the average household in the three research groups before (1999) and after (2009) MTS. According to Messer and Townsley (2003), households with relatively large, well-balanced, and regular livelihood pentagons have a stronger asset base while those with small, distorted pentagons have fewer assets from which to build their livelihoods. Thus, by varying access or claims to external resources such as the alternative livelihoods support under the current MTS for example, research communities may be able to alter the shape of their livelihood pentagons and thus their livelihood outcomes. While the outermost pentagon (thick black line) represents the maximum attainable asset levels (1), the innermost pentagon (thick yellow line) represents the minimum attainable livelihood asset levels (0) for each asset category. Throughout the analysis, both the outermost (thick black line) and innermost (thick yellow line) pentagons do not change but rather remain the same (1 and 0) and used as benchmarks against which improvements or decline in household livelihood assets are measured for the two time periods before and after MTS for all three research categories. The red and blue pentagons represent the three research groups and how 698 they change relative to the maximum and minimum attainable indexes for the two time periods. The green pentagon captures differences between the red and blue pentagons. Thus, when a corner of the green pentagon rises above the yellow pentagon, an improvement in livelihood asset is observed between the two time periods (1999 and 2009) or between the blue and red pentagons. When the green pentagon dips or falls below the yellow pentagon, a decline in livelihood is registered between the two time periods for a particular livelihood asset. 7.6.3 Changes in Household Livelihood Indexes for all Groups Between 1999 and 2009 Figure 7.6.3 below represents aggregate changes in all five-household livelihood asset within the three research communities (i.e., 878 households) before (1999) and after MTS (2009). The result from the green pentagon which captures changing asset levels suggests that the combined physical asset index increased by 0.17 or 17% the largest increase of all the other four livelihood asset indexes. Compared to the physical capital index, social and financial capital both increased by 2% while human capital index increased by 3% from 1999 to 2009. Of all household livelihood indexes represented in the household livelihood pentagon (HLP) above, only natural capital index experienced a decline for the period under review (-5%). The rest of section 7.6 compares within and between changes in household livelihood indexes (HLI) for the three research groups before and after MTS and tries to draw inferences whether the changes may be attributable to the MTS or not. 699 Figure 7.6.1: Changes in Aggregate Asset Indexes between 1999 and 2009 700 0.300.580.390.550.500.470.600.410.500.530.170.030.02-0.050.02-0.200.000.200.400.600.801.00Physical (chpci1)Human (chci1)Financial (chhfci)Natural (chhnci1)Social (ahhsci)Changes in Aggregate Household Livelihood Asset Indexes between 1999 and 2009 99_Pooled (N=439)09_Pooled (N=439)Change in Index (2009)-(1999)Max Attainable IndexMin Attainable Index Figure 7.6.2: Within Group Changes in MTS and non-MTS Asset Indexes between 1999 and 2009 701 0.320.590.400.530.500.500.630.420.480.500.190.040.02-0.050.01-0.200.000.200.400.600.801.00Physical (chpci1)Human (chci1)Financial (chhfci)Natural (chhnci1)Social (ahhsci)Changes in Non-MTS Household Livelihood Asset Indexes between 1999 and 200999_Non-MTS in Yaya (n=120)09_Non-MTS in Yaya (n=120)Change in Index (2009)-(1999)Max Attainable IndexMin Attainable Index0.310.590.390.540.510.510.630.430.510.540.190.040.04-0.030.04-0.200.000.200.400.600.801.00Physical (chpci1)Human (chci1)Financial (chhfci)Natural (chhnci1)Social (ahhsci)Changes in MTS Household Livelihood Asset Indexes between 1999 and 2009 99_MTS in Yaya (n=203)09_MTS in Yaya (n=203)Change in Index (2009)-(1999)Max Attainable IndexMin Attainable Index 7.6.4 Within Group Changes in MTS and non-MTS Asset Indexes between 1999 and 2009 Within group changes observed for both MTS and non-MTS households in Yaya as shown in figure 7.6.2 above suggest very similar changes in trend and magnitude for all five HLIs. For example, combined household physical capital index (CHPCI) and combined household human capital index (CHHCI) were equal for both groups in 1999 and 2009 and increased by 19% and 4% respectively for both groups from their 1999 baseline levels to 2009. It may be argued that but for spillover effects resulting from the MTS policy, non-MTS households would have experienced relatively lower CHPCI and CHHCI in 2009. Further analysis in chapter eight attempts to answer the questions of MTS project and spill-over effects on HLIs among both MTS and non-MTS research groups in Yaya. In terms of aggregate household social capital index (AHHSCI), and combined household financial capital index (CHFCI), MTS households recorded slight improvements in both indexes compared to their non-MTS counterparts. For the AHHSCI index, MTS households experienced an increase of 4% relative to 1% recorded for non-MTS households. Figure 7.6.2 shows that both groups experienced an increase in CHFCI however the trend for MTS household was 2% higher than non-MTS households. Since one of the areas of household livelihoods targeted by the MTS policy is income improvement through implementation of alternative livelihood programs and also profit sharing from harvested timber, increases in CHFCI among MTS households is a plus for the MTS. For both MTS and non-MTS groups, combined household natural capital index declined between the non-MTS period (1999) and MTS period (2009). The decline in CHHNCI was however slightly higher among non-MTS (-5%) compared to MTS groups (-3%). Using a Difference-in-Difference (DID) approach chapter eight will attempt to investigate how much decline in CHHNCI would have occurred among both groups without the MTS policy. 702 Figure 7.6.3: Between Group Changes in MTS and non-MTS Asset Indexes for 1999 and 2009 703 0.310.590.390.540.510.320.590.400.530.50-0.010.00-0.010.010.01-0.200.000.200.400.600.801.00Physical (chpci1)Human (chci1)Financial (chhfci)Natural (chhnci1)Social (ahhsci)Comparison Between MTS & Non-MTS Household Livelihood Asset Indexes for Year 199999_MTS in Yaya (n=203)99_Non-MTS in Yaya (n=120)Change in Index (MTS)-(Non-MTS)Max Attainable IndexMin Attainable Index0.510.430.510.540.500.630.420.480.500.000.000.010.030.04-0.200.000.200.400.600.801.00Physical (chpci1)Human (chci1)Financial (chhfci)Natural (chhnci1)Social (ahhsci)Comparison Between MTS & Non-MTS Household Livelihood Asset Indexes for Year 200909_MTS in Yaya (n=203)09_Non-MTS in Yaya (n=120)Change in Index (MTS)-(Non-MTS)Max Attainable IndexMin Attainable Index 7.6.5 Between Group Changes in MTS, non-MTS and NsemSaw HLIs for 1999 and 2009 The pentagon on the left in figure 7.6.3 above compares pre-MTS (1999) values of each of the five HLIs between MTS and non-MTS households while the one on the right compare HLIs values of both groups in 2009. According to the figure above, in 1999 all the corners of the green livelihood pentagon superimpose almost perfectly on the corners of the yellow pentagon suggesting very little if any difference in HLIs between both groups in 1999. By 2009 MTS households experienced slightly higher AHHSCI (4%), CHHNCI (3%) and CHFCI (1%) than non-MTS households in Yaya (see HLI pentagon on the right in figure 7.6.3). Though both MTS and non-MTS groups experienced significant increases of about 20% and 14% in CHPCI and CHCI respectively between 1999 and 2009 the 2009 CHPCI and CHCI were almost the same for both groups (see HLI pentagon on the right in figure 7.6.3). Using a DID approach, chapter eight attempts to isolate the magnitude of the changes in CHPCI and CHCI resulting from a direct MTS policy effect on participating households or a spillover effect on non-MTS households in Yaya. 704 Figure 7.6.4: Within Group Changes in MTS and NsemSaw Asset Indexes between 1999 and 2009 705 0.270.540.390.580.500.390.530.380.490.520.120.00-0.01-0.090.02-0.200.000.200.400.600.801.00Physical (chpci1)Human (chci1)Financial (chhfci)Natural (chhnci1)Social (ahhsci)Changes in NsemSaw Household Livelihood Asset Indexes between 1999 and 200999_NsemSaw (n=116)09_NsemSaw (n=116)Change in Index (2009)-(1999)Max Attainable IndexMin Attainable Index0.310.590.390.540.510.510.630.430.510.540.190.040.04-0.030.04-0.200.000.200.400.600.801.00Physical (chpci1)Human (chci1)Financial (chhfci)Natural (chhnci1)Social (ahhsci)Changes in MTS Household Livelihood Asset Indexes between 1999 and 2009 99_MTS in Yaya (n=203)09_MTS in Yaya (n=203)Change in Index (2009)-(1999)Max Attainable IndexMin Attainable Index 7.6.6 Within Group Changes in MTS and NsemSaw Asset Indexes between 1999 and 2009 Figure 7.6.4 above examines within group changes in HLIs among MTS households in Yaya and NsemSaw between 1999 and 2009. While the pentagon on the left-hand side in figure 7.6.4 examines changes in HLIs among MTS households that on the right examines those for NsemSaw. According to figure 7.6.4 between 1999 and 2009, MTS households in Yaya experienced a 19% increase in CHPCI while NsemSaw experienced a 12% increase (a 7% difference in CHPCI between both groups). In terms of changes in AHHSCI between 1999 and 2009, MTS households experienced a 4% increase in AHHSCI compared to 2% among NsemSaw. While the MTS group in Yaya experienced a 4% increase in CHCI and CHHFCI between 1999 and 2009, households in Yaya experienced no change in CHCI for the same time period. Unlike MTS households that experienced a 4% improvement in CHHFCI, NsemSaw households experienced a 1% decline in CHHFCI between 1999 and 2009. Both MTS and NsemSaw groups experienced a decline in their CHHNCI between 1999 and 2009. The decline in CHHNCI was however slightly higher among NsemSaw (-5%) compared to MTS groups (- 3%). Using NsemSaw as a control chapter 8 employs the DID approach to investigate how much of the observed changes the five HLIs among MTS households may be attributed to the MTS policy. Since HLIs were very similar for both MTS and non-MTS households, the foregoing discussion on MTS versus NsemSaw would be similar to one between non-MTS in Yaya and NsemSaw. Hence, I skip the discussion of figures 7.6.6 below which presents within group changes for non-MTS in Yaya and NsemSaw. 706 Figure 7.6.5: Between Group Changes in MTS and NsemSaw Asset Indexes for 1999 and 2009 707 0.310.590.390.540.510.270.540.390.580.500.040.050.00-0.040.01-0.200.000.200.400.600.801.00Physical (chpci1)Human (chci1)Financial (chhfci)Natural (chhnci1)Social (ahhsci)Comparison Between MTS & NsemSaw Household Livelihood Asset Indexes for Year 199999_MTS in Yaya (n=203)99_NsemSaw (n=116)Change in Index (MTS)-(NsemSaw)Max Attainable IndexMin Attainable Index0.510.630.430.510.540.390.530.380.490.520.120.090.050.020.03-0.200.000.200.400.600.801.00Physical (chpci1)Human (chci1)Financial (chhfci)Natural (chhnci1)Social (ahhsci)Comparison Between MTS & NsemSaw Household Livelihood Asset Indexes for Year 200909_MTS in Yaya (n=203)09_NsemSaw (n=116)Change in Index (MTS)-(NsemSaw)Max Attainable IndexMin Attainable Index 7.6.7 Between Group Changes in MTS and NsemSaw HLIs for 1999 and 2009 The pentagon on the left-hand side in figure 7.6.5 above compares pre-MTS (1999) values of each HLI between MTS households in Yaya and NsemSaw while the one on the right compare HLIs values of both groups in 2009. Different the shapes of the center green pentagon between 1999 and 2009 captures differences in HLIs between MTS and NsemSaw households. In 1999 the values of CHFCI among both MTS and NsemSaw households were equal since both corners of the green and yellow pentagons representing CHFCI superimposed perfectly on each other. In the post-MTS period (2009) however, the MTS group experienced a 5% increase in CHFCI over NsemSaw (depicted by the 5% outward shift of the inner green livelihood pentagon). Compared to the NsemSaw group in 1999 MTS households experienced relatively higher values for AHHSCI (1%), CHPCI (4%) and CHCI (5%). By 2009 the differences in all three HLIs between both groups increased with the highest gap recorded for CHPCI (12%) followed by CHCI (9%) and AHHSCI (3%). The relatively large gaps observed for CHPCI, CHCI and AHHSCI between MTS and NsemSaw households in the post-MTS values is readily observed by the larger outward shift in the corners of these three HLIs in figure 7.6.5. In 1999 NsemSaw households experienced a 4% higher CHHNCI compared to MTS households in Yaya however by 2009 a reversal in trend was observed with the MTS group recording a 2% higher CHHNCI than NsemSaw. The question of whether the MTS policy is responsible for any of the changes in HLI observed among project participants is addressed in the next chapter. Since 1999 and 2009 HLI values were very similar for both MTS and non- MTS households, it is anticipated that the foregoing discussion on MTS versus NsemSaw would be similar to the analysis between non-MTS in Yaya and NsemSaw. Hence, I skip the discussion of figures 7.6.6 and 7.6.7 below. 708 Figure 7.6.6: Within Group Changes in non-MTS and NsemSaw Asset Indexes between 1999 and 2009 709 0.270.540.390.580.500.390.530.380.490.520.120.00-0.01-0.090.02-0.200.000.200.400.600.801.00Physical (chpci1)Human (chci1)Financial (chhfci)Natural (chhnci1)Social (ahhsci)Changes in NsemSaw Household Livelihood Asset Indexes between 1999 and 200999_NsemSaw (n=116)09_NsemSaw (n=116)Change in Index (2009)-(1999)Max Attainable IndexMin Attainable Index0.320.590.400.530.500.500.630.420.480.500.190.040.02-0.050.01-0.200.000.200.400.600.801.00Physical (chpci1)Human (chci1)Financial (chhfci)Natural (chhnci1)Social (ahhsci)Changes in Non-MTS Household Livelihood Asset Indexes between 1999 and 200999_Non-MTS in Yaya (n=120)09_Non-MTS in Yaya (n=120)Change in Index (2009)-(1999)Max Attainable IndexMin Attainable Index Figure 7.6.7: Between Group Changes in non-MTS and NsemSaw Asset Indexes for 1999 and 2009 710 0.320.590.400.530.500.270.540.390.580.500.050.060.01-0.04-0.01-0.200.000.200.400.600.801.00Physical (chpci1)Human (chci1)Financial (chhfci)Natural (chhnci1)Social (ahhsci)Comparison Between Non-MTS & NsemSaw Household Livelihood Asset Indexes for Year 199999_Non-MTS in Yaya (n=120)99_NsemSaw (n=116)Change in Index (Non-MTS)-(NsemSaw)Max Attainable IndexMin Attainable Index0.500.630.420.480.500.390.530.380.490.520.120.100.04-0.01-0.02-0.200.000.200.400.600.801.00Physical (chpci1)Human (chci1)Financial (chhfci)Natural (chhnci1)Social (ahhsci)Comparison Between Non-MTS & NsemSaw Household Livelihood Asset Indexes for Year 200909_Non-MTS in Yaya (n=120)09_NsemSaw (n=116)Change in Index (Non-MTS)-(NsemSaw)Max Attainable IndexMin Attainable Index APPENDICES 711 APPENDIX A: Descriptive Analysis of Financial Capital Indexes A7.1.1 Primary Sources of Household Income Tables 7.5.4 and 7.5.5 below was generated from a list of questions that ask respondents to rank different income sources according to their importance to the household’s total annual income. The results were later grouped into two different categories comprising of primary and secondary income sources. Table 7.5.4 thus presents results on sources that respondent said contributed the most to their household’s income. Respondents who answered the question in table 7.5.4 ranked on a scale of one to five (where 1 is the highest contributor and 5 the least) the contributions of both crop and livestock production to household income. No income contributions from neither crop nor livestock production were ranked zero. The results in table 7.5.4 thus indicate an approximate ranking of each of the income sources (crop and livestock) in terms of their contributions towards total annual income. From the pooled responses below 77% (n=338) of survey respondents said crop production was the highest contributor to their total household income in 1999 compared to 85.9% (n=377) who also reported that crops were the highest contributor in 2009. For majority of respondents (44%, n=193), livestock was the second highest contributor to total household income in 1999 compared to approximately 45% (n=198) in 2009. Table 7.5.4 provides the basis to computing the “Primary Household Income Source Index” in table 7.5.5. To generate the index, all the responses to the questions in table 7.5.4 were weighted such that households that listed a particular source as the highest contributor to household income received a score of five while those who ranked a source second received a score of four for that source. This rating scale, among other things, captures the importance of different income sources in contributing to households’ income/finances. Hence for the two income sources, the maximum attainable score 712 for a household is 10. A score of 10 suggests that both livestock and crop production are very important contributors to household income and also that both contribute equally to the household’s income. Table 7.5.4: Primary Sources of Household Income - Frequency Frequency (Percentage) Highest contributor (s)=5 2nd Highest=4 3rd Highest=3 4th Highest=2 5th Highest=1 Zero from Source= 0 Sub-total Missing System Total Which of the following contributes the most to the HH financial needs? 1. Crop Production (EA1_2) 2. Livestock Production (EA2_2) 1999 338 (77.0) 54 (12.3) 4 (0.9) 12 (2.7) 12 (2.7) 0 (0) 420 (95.7) 19 (4.3) 439 (100) 2009 377 (85.9) 43 (9.8) 5 (1.1) 2 (0.5) 0 (0) 0 (0) 427 (97.3) 12 (2.7) 439 (100) 1999 15 (3.4) 193 (44.0) 71 2009 6 (1.4) 198 (45.1) 75 (16.2) (17.1) 21 (4.8) 12 (2.8) 0 (0) 312 (71.1) 127 (28.9) 439 (100) 16 (3.6) 12 (2.7) 0 (0) 307 (69.9) 132 (30.1) 439 (100) 713 Table 7.5.5: Primary Sources of Household Income - Descriptive NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 M SD M SD M SD M SD 1999 4.3707 1.2754 4.5917 1.1339 4.4138 1.3520 4.4510 1.2756 2009 4.6983 0.9800 4.7750 0.7720 4.7192 0.9092 4.7289 0.8923 Which of the following contributes the most to the HH financial needs? SCALE Highest contributor (S)=5 2nd Highest=4 3rd Highest=3 4th Highest=2 5th Highest=1 Zero from Source= 0 YR 1. Crop Production (EA1_2) 2. Livestock Production (EA2_2) 1999 2.7414 1.7048 2.2833 1.8752 2.5714 1.7316 2.5376 1.7693 2009 2.6293 1.7621 2.2333 1.8230 2.5517 1.7151 2.4852 1.7606 Primary HH Income Sources Index (PHISI) A7.1.2 Secondary Sources of Household Income Table 7.5.6 and 7.5.7 like the previous tables lists all secondary sources of household income within the research communities and the extent to which each source contributes to total household income or finances. As shown in tables 7.5.6 and 7.5.7 below, of all secondary income sources, sale of daily or hourly farm labor and income from forestry/MTS appears to be the two most important secondary income sources within the research communities. On average 60% (n=261) of respondents did not receive any income from farm labor in 1999 and of the 40% (n=173) who did, approximately 46% (n=79) also said sale of farm labor was the third largest contributor to their total household income. Compared to 1999, in 2009, 714 10_2EA5*2_2EA*_2EAPHISI21it21it21itt=========nininiSn approximately 59% (n=257) of respondents said they did not earn any income from farm labor and of the 41% (n=177) who did, majority (45%, n=79) also said this was the third largest income generator for their households. Table 7.5.6: Secondary Sources of Household Income - Frequency Which of the following contributes the most to the HH financial needs? Freque ncy (Perce ntage) Highest contribu tor (s)=5 2nd Highest =4 3rd Highest =3 4th Highest =2 5th Highest =1 Sub- total 3. Sale of Farm Labor (EA3_2) 4. Remittance Income (EA4_2) 5. Pension (EA5_2) 6. Sale of NTFP (EA6_2) 7. Income from Forestry/MT S (EA7_2) 8. Other (EA8_2) 1999 2009 1999 2009 1999 2009 1999 2009 1999 2009 1999 2009 5 5 3 1 (1.1) (1.1) (0.7) (0.2) 0 (0) 1 8 2 26 17 31 23 (0.2) (1.8) (0.5) (5.9) (3.9) (7.1) (5.2) 33 39 14 12 2 5 24 16 27 31 37 48 (7.5) (8.9) (3.2) (2.7) (0.5) (1.1) (5.5) (3.6) (6.2) (7.1) (8.4) (10.9) 79 79 41 39 7 5 44 34 33 59 23 26 (18.0) (18.0) (9.3) (8.9) (1.6) (1.1) 10.0) (7.7) (7.5) (13.4) (5.2) (5.9) 34 38 39 59 5 5 46 57 28 29 27 28 (7.7) (8.7) (8.9) (13.4) (1.1) (1.1) (10.5) (13.0) (6.4) (6.6) (6.2) (6.4) 27 21 32 28 6 5 26 30 17 28 24 24 (6.1) (4.8) (7.3) (6.4) (1.3) (1.1) (5.7) (6.8) (3.9) (6.4) (5.5) (5.5) 178 182 129 139 20 21 (40.5) (41.5) (29.4) (31.7) (4.6) (4.8) 148 33.7) 139 131 (31.7) (29.8) 164 (37.4) 142 149 (32.3) (33.9) Missing System 261 59.5) 257 (58.5) Total 439 (100) 439 (100) 310 (70.6) 439 (100) 300 419 418 291 300 308 (68.3) (95.4) (95.2) (66.3) (68.3) (70.2) 275 (62.6) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 297 (67.7) 439 (100) 290 (66.1) 439 (100) The results for farm labor for both 1999 and 2009 are remarkably similar, indicating very little change in income generation activities across all research communities for the two time periods. When respondents were asked to rank income contributions from forestry or Taungya activities, 70% (n=308) and 63% (n=275) said they did not receive any income from these sources in 1999 and 2009 respectively. Of those 30% (n=131) households that earned income 715 from forestry in 1999 25% (n=33) said it was their third largest income earner compared to 35% (n=59) who ranked income from forestry third place in 2009.The mean results presented in table 7.5.7 provided the basis for computing the “secondary Household Income Source Index” (SHISI). To calculate the SHISI index, the means for all secondary income sources were aggregated for each response group and divided by the total attainable score of 30 for the SHISI (six secondary income sources multiplied by five the highest score on the five-point rating scale). 716 Table 7.5.7: Secondary Sources of Household Income - Descriptive NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 M SD M SD M SD M SD Which of the following contributes the most to the HH financial needs? SCALE Highest contributor=5 2nd Highest=4 3rd Highest=3 4th Highest=2 5th Highest=1 Zero from Source= 0 YR 3. Sale of Farm Labor (EA3_2) 4. Remittance Income (EA4_2) 1999 0.7414 1.4333 1.4333 1.5652 1.1379 1.4524 1.1139 1.4976 2009 0.8103 1.4857 1.5500 1.6080 1.1576 1.4674 1.1731 1.5326 1999 0.2500 0.8635 1.1583 1.4225 0.6700 1.1708 0.6925 1.2140 2009 0.2500 0.8223 1.1750 1.3821 0.7192 1.1103 0.7198 1.1748 5. Pension (EA5_2) 6. Sale of NTFP (EA6_2) 7. Income from Forestry/MTS (EA7_2) 8. Other (EA8_2) 1999 0.0259 0.2785 0.1500 0.6346 0.1182 0.5412 0.1025 0.5155 2009 0.0776 0.4961 0.1917 0.7592 0.1133 0.5903 0.1253 0.6195 1999 0.7069 1.4627 0.9917 1.3811 0.9113 1.3578 0.8793 1.3936 2009 0.4397 1.0574 0.8917 1.3080 0.7980 1.1997 0.7289 1.2057 1999 0.8534 1.7558 0.3000 0.8562 1.3547 1.7213 0.9339 1.6027 2009 0.7069 1.4447 0.3000 0.8462 1.7438 1.7013 1.0752 1.5765 1999 0.9052 1.6887 1.3830 1.8571 0.9261 1.5668 1.0251 1.6858 2009 0.7414 1.4985 1.4500 1.9222 1.0099 1.5606 1.0592 1.6693 Secondary HH Income Sources Index (SHISI) 717 30_2EA5*6_2EA*_2EASHISI83it83it61itt=========nininiSn A7.1.3 Primary Household Expenditure Items In table A7.1.3a, respondents were provided with a list of 13 expenditure items and asked to rank a maximum of 10 items on which the spend their household income. The highest expenditure item was raked 10 and the lowest ranked 1. All other items for which very little or no income was expended received a zero ranking. From table A7.1.3a below, the mean responses for the pooled data suggests that education, health and consumer goods expenditure were ranked the highest among the list of expenditure items that consume the lion’s share of household income. The results also suggest an increasing trend in expenditure on all primary expenditure items with the highest increases recorded for education between 1999 and 2009. Table A7.1.3a provides a basis for constructing the “Household Primary Expenditure Index 1” (HEI1). The HEI1 formula under table A7.1.3b below suggests that the only way a household is able to attain a maximum score of 70 on HEI is if all seven items are ranked equally and assigned 10-point ratings. There were two computations under household expenditure. The first presented in table A7.1.3a ranked the top five expenditure items with any items ranked higher than five assigned a value of zero on the expenditure scale. Thus, table A7.1.3a was included in the analysis to help readers to understand the response rates and their weights for each category expenditure item. Section A7.1.4 that follows describes how the same approach was used to generate a secondary expenditure index based only on items categorized as secondary household expenditure items. 718 Table 7.5.8: Primary Sources of Household Expenditure - Frequency Which of the following consumes the most to the HH annual income? Frequency (Percentage) 1. Staple Foods (H_Exp_1b) 4. Education (H_Exp_4b) 5. Health (H_Exp_5b) 6. Consumer goods, cloths and cosmetic products (H_Exp_6b) 8. Transportatio n (H_Exp_8b) 9. Funerals and Weddings (H_Exp_9b) 11. Farm Equipment (H_Exp_11b) 1999 2009 1999 2009 1999 2009 1999 2009 1999 2009 1999 2009 1999 2009 Highest Expenditure (s)=5 53 66 172 190 48 44 21 15 20 13 31 34 24 24 (12.1) (15.0) (39.2) (43.3) (10.9) (10.0) (4.8) (3.4) (4.6) (3.0) (7.1) (7.7) (5.5) (5.5) 57 61 65 80 80 78 54 53 46 38 31 32 28 28 (13) (13.9) (14.8) (18.2) (18.2) (17.8) (12.3) (12.1) (10.5) (8.7) (7.1) (7.3) (6.4) (6.4) 40 (9.1) 36 50 47 77 84 60 70 61 73 46 47 26 (8.2) (11.4) (10.7) (17.5) (19.1) (13.7) (15.9) (13.9) (16.6) (10.5) (10.7) (5.9) 30 (6.8) 27 32 30 28 46 47 75 89 63 65 57 57 29 29 (6.2) (7.3) (6.8) (6.4) (10.5) (10.7) (17.1) (20.3) (14.4) (14.8) (13) (13.0) (6.6) (6.6) 30 25 14 15 32 38 54 50 67 77 57 64 44 40 (6.8) (5.7) (3.2) (3.4) (7.3) (8.7) (12.3) (11.4) (15.3) (17.5) (13.0) (14.6) (10) (9.1) 2nd Highest=4 3rd Highest=3 4th Highest=2 5th Highest=1 6th thru 10th= 0 Sub-total 13 (3) 221 13 (3) 233 (50.3) (53.1) 7 4 19 22 21 25 34 35 40 43 44 52 (1.6) (0.9) (4.4) (4.5) (4.8) (5.8) (7.8) (8.1) (9.1) (9.8) (10) (11.8) 338 (77) 101 (23) 364 302 313 285 302 291 301 262 277 195 203 (82.9) (68.8) (71.3) (64.9) (68.8) (66.3) (68.6) (59.7) (63.1) (44.4) (46.2) 75 137 126 154 137 148 138 177 162 244 236 (17.1) (31.2) (28.7) (35.1) (31.2) (33.7) (31.4) (40.3) (36.9) (55.6) (53.8) Zero Expenditure /Missing System 219 206 (49.9) (46.9) Total 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 719 Table 7.5.9: Primary Sources of Household Expenditure - Descriptive Which of the following consumes the most to the HH annual income? SCALE Zero Expenditure= 0 Least Expenditure=1 Highest Expenditure=10 (S) 1. Staple Foods (H_Exp_1b) 4. Education (H_Exp_4b) 5. Health (H_Exp_5b) 6. Consumer goods, cloths and cosmetic products (H_Exp_6b) 8. Transportation (H_Exp_8b) 9. Funerals and Weddings (H_Exp_9b) 11. Farm Equipment/ Tools (H_Exp_11b) NSEMSAW N=116 NON-MTS in YAYA N=120 MTS in YAYA N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 4.4655 4.2456 3.9250 4.2150 3.8966 4.2501 4.0547 4.2369 2009 4.7586 4.4206 4.1250 4.3222 4.2857 4.2388 4.3667 4.3073 1999 6.3966 4.1867 7.3750 3.6970 6.8571 4.0154 6.8770 3.9850 2009 6.6379 4.2147 7.8583 3.3514 7.7783 3.3397 7.4989 3.6232 1999 4.7586 4.2254 5.8000 3.9530 5.7241 3.6847 5.4897 3.9227 2009 4.6207 4.1821 5.8083 3.9028 6.0099 3.4655 5.5877 3.8227 1999 4.5776 3.9302 5.2830 3.6050 4.6601 3.7201 4.8087 3.7490 2009 4.6638 3.8311 5.5333 3.4664 4.9852 3.5551 5.0501 3.6128 1999 5.4052 3.7393 4.8330 3.6350 4.3695 3.5609 4.7699 3.6458 2009 5.1810 3.5913 4.7833 3.5247 4.6798 3.4685 4.8405 3.5148 1999 5.3448 3.7396 4.0500 3.7840 3.7685 3.6656 4.2620 3.7675 2009 5.5345 3.5957 4.3333 3.8180 4.0148 3.6308 4.5034 3.7196 1999 3.1810 3.9926 1.8080 3.0330 3.6207 3.7243 3.0091 3.6956 2009 3.1466 4.0114 1.8417 3.0651 3.8276 3.6634 3.1048 3.6941 HH Expenditure Index1 (HEI1) (Where n is the total number of expenditure items and N is the total number of respondents) 720 49bH_Exp_b H_Exp_HEI14924it10..47*4itt==++====ninin A7.1.4 Secondary Household Expenditure Items Table 7.5.10: Secondary Household Expenditure - Frequency Which of the following consumes the most to the HH annual income? Frequency (Percentage) 2. Fruits, Vegetables and Spices (H_Exp_2b) 3. Alcoholic Drinks (H_Exp_3b) 7. Fuel (Firewood, Kerosene) (H_Exp_7b) 10. Loan Repayments (H_Exp_10b) 12. Home Repair (H_Exp_12b) 13. Other (H_Exp_13b) 1999 2009 1999 2009 1999 2009 1999 2009 1999 2009 1999 2009 Highest Expenditure (s)=5 4 5 (0.9) (1.1) 13 (3) 10 10 (2.3) 13 8 7 (1.8) (1.6) (2.3) (3.0) 6 (1.4) 6 (1.4) 5 (1.1) 2 (0.5) 6 7 2 2 16 18 2 4 (1.4) (1.6) (0.5) (0.5) (3.6) (4.1) (0.5) (0.9) 13 (3) 18 (4.1) 15 5 5 16 14 2 3 (3.4) (1.1) (1.1) (3.6) (3.2) (0.5) (0.7) 13 (3.0) 2 (0.5) 3 (0.7) 6 (1.4) 8 (1.8) 2 (0.5) 2 (0.5) 2nd Highest=4 3rd Highest=3 4th Highest=2 5th Highest=1 6th thru 10th= 0 Sub-total 5 8 8 4 18 19 10 8 11 13 1 1 (1.1) (1.8) (1.8) (0.9) (4.1) (4.3) (2.3) (1.8) (2.5) (3.0) (0.2) (0.2) 5 3 4 7 28 30 9 8 (1.1) (0.7) (0.9) (1.6) (6.4) (6.8) (2.1) (1.8) 36 (8.3) 40 (9.1) 22 22 23 24 10 15 (5.2) (5.0) (5.3) (5.5) (2.3) (3.5) (8.4) (9.4) 64 69 62 58 106 108 (14.6) (15.7) (14.1) (13.2) (24.1) (24.6) 38 (8.7) 41 (9.3) 99 106 (22.6) (24.1) 16 (3.6) 13 (3) 37 12 (2.7) 41 0 (0) 9 (2) 1 (0.2) 11 (2.5) 22 (5.0) Zero Expenditure /Missing System 375 370 377 381 333 331 401 398 340 333 423 417 (85.4) (84.3) (85.9) (86.8) (75.9) (75.4) (91.3) (90.7) (77.4) (75.9) (96.4) (95.0) Total 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 721 Table 7.5.11: Secondary Household Expenditure – Descriptive NSEMSAW N=116 NON-MTS in YAYA N=120 MTS in YAYA N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.3103 1.6546 0.8170 2.4040 1.0049 2.1992 0.7699 2.1453 2009 0.3879 1.8499 0.7333 2.2445 1.1576 2.3557 0.8383 2.2205 1999 0.6810 2.3760 1.1080 2.8190 0.8916 2.4268 0.8952 2.5262 2009 0.4741 2.0450 0.8750 2.5588 0.9951 2.4826 0.8246 2.4012 1999 1.0345 2.6373 2.1250 3.3670 1.5517 2.9012 1.5718 2.9914 2009 0.8534 2.3415 1.7667 3.1429 1.8867 3.1813 1.5809 2.9973 1999 0.0603 0.6499 0.5920 2.0720 0.7192 2.1049 0.5103 1.8430 2009 0.1379 1.0786 0.5333 1.9226 0.7882 2.1363 0.5467 1.8670 1999 0.5690 2.0567 1.0580 2.7300 2.0690 3.3353 1.3964 2.9474 2009 0.6897 2.2240 1.1917 2.9199 2.1084 3.3148 1.4829 3.0090 1999 0.0862 0.9285 0.1080 0.8960 0.2709 1.3681 0.1777 1.1469 2009 0.0862 0.9285 0.2333 1.3456 0.4089 1.7019 0.2756 1.4395 Which of the following consumes the most to the HH annual income? SCALE Zero Expenditure= 0 Least Expenditure=1 Highest Expenditure=10 (S) 2. Fruits, Vegetables and Spices (H_Exp_2b) 3. Alcoholic Drinks (H_Exp_3b) 7. Fuel (Firewood, Kerosene) (H_Exp_7b) 10. Loan Repayments (H_Exp_10b) 12. Home Repair (H_Exp_12b) 13. Other (H_Exp_13b) HH Expenditure Index2 (HEI2) (Where n is the total number of expenditure items and N is the total number of respondents) Table 7.1.4 above describes respondents’ household expenditure on other items categorized as secondary household expenditure items because of the relative low percentage of household income expended on these expenditure items. Of the six items included in this particular expenditure category, home repair and fuel wood appear to be the two most important 722 28bH_Exp_b H_Exp_HEI2287it7..17*1itt==++====ninin secondary expenditure items that consume part of the household income. Assuming each secondary expenditure item received the maximum ranking of six, a maximum score of 60 would be achieved for the list of items presented in the “Household Expenditure Index 2” (HHI2) category. A7.1.5 Household Savings and Loan Activities Table 7.5.12: Household Savings and Loan Account Ownership - Frequency Does any member of the HH have any of the following? 1. Bank Account (LL_Sav_1) 2. Local Susu Account (LL_Sav_2) 1999 325 (74.0) 101 (23.0) 426 (97.0) 13 (3.0) 439 (100) 2009 289 (65.8) 141 (32.1) 430 (97.9) 9 (2.1) 439 (100) 1999 360 (82.0) 63 (14.4) 423 (96.4) 16 (3.6) 439 (100) 2009 346 (78.8) 81 (18.5) 427 (97.3) 12 (2.7) 439 (100) Frequency (Percentage) No=0 Yes=1 Sub-total Missing System Total 723 Table 7.5.13: Household Savings and Loan Account Ownership - Descriptive NSEMSAW N=116 NON-MTS in YAYA N=120 MTS in YAYA N=203 POOLED N=439 YR 1999 0.2348 0.4257 0.2203 0.4162 0.2487 0.4334 0.2371 0.4258 SD SD M SD M M SD M 2009 0.2609 0.4410 0.3475 0.4782 0.3553 0.4798 0.3279 0.4700 1999 0.0631 0.2442 0.1610 0.3691 0.1907 0.3939 0.1489 0.3565 2009 0.0351 0.1848 0.2458 0.4324 0.2462 0.4319 0.1897 0.3925 Does any member of the HH have any of the following? SCALE No=0 Yes=1 1. Bank Account (LL_Sav_1) 2. Local Susu Account (LL_Sav_2) HH Bank Account Index (HHBAI) In table 7.1.5 respondents were asked if they have a bank account with any financial institution or whether they had any other forms of accounts with the local money lenders also called “Susu” collectors. A yes response to any of the two questions is ranked one while a no is ranked zero. From the above table, approximately 33% of respondents in the pooled data set have bank accounts in 2009 relative to approximately 24% who had it in 1999. In terms of local Susu account ownership, approximately 4% more had Susu accounts in 2009 (19%) compared to 1999 (15%). Table 7.1.5 was used to construct the “Household Bank Account Index” (HHBAI) by dividing the total score obtained by each respondent by the total attainable score of two for both bank and Susu account ownership. 724 2(L_Sav_1a)1*2(L_Sav_1a)*(L_Sav_1a)HHBAI21it21it21itt=========nininiSn A7.1.6 Frequency of Household Savings and Loan Activities Table 7.1.6 above tries to understand respondents banking and saving behavior for the two time periods before MTS (1999) and after MTS (2009). Of the two savings and loan options presented in the table, respondents across all three survey categories appear to patronize formal banks more than they do local “Susu” baking programs. Table 7.1.6 was used to construct the “Frequency of Household Banking Activity Index (FHHBAI).” From the index formula above, the maximum attainable FHHBAI score of 12 suggests that households in the particular survey group whether MTS or no-MTS, use the services of both formal and local banking institutions at least five times in a year. Maximum patronage of financial institutions attracts a score of three on the rating scale while no usage attracts a zero score. The number of options (four) multiplied by the maximum rating scale (three) results in the maximum attainable score for FHHBAI. 725 Table 7.5.14: Frequency of Household Savings and Loan Activities - Frequency How often in a year have member of the HH done any of the following? 1. Save or withdraw money from the bank? (FSLA_1) 2. Save or withdraw money from the Susu Collector (FSLA_2) 3. Take a bank loan (FSLA_3) 4. Borrow money from the Susu Collector (FSLA_4) 1999 2009 1999 2009 1999 2009 1999 2009 315 293 358 345 366 363 401 393 (71.8) (66.7) (81.5) (78.6) (83.4) (82.7) (91.3) (89.5) 68 94 44 47 53 57 23 29 (15.5) (21.4) (10.0) (10.7) (12.1) (13.0) (5.2) (6.6) 29 29 12 17 6 6 2 3 (6.6) (6.6) (2.7) (3.9) (1.4) (1.4) (0.5) (0.7) 13 12 (3.0) (2.7) 425 428 (96.8) (97.5) 11 (2.5) 425 14 (3.2) 423 (96.8) (96.4) 1 2 (0.2) (0.5) 0 (0) 1 (0.2) 426 428 426 426 (97.0) (97.5) (97.0) (97.0) 14 11 14 16 13 11 13 13 (3.2) (2.5) (3.2) (3.6) (3.0) (2.5) (3.0) (3.0) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) Frequency (Percentage) Never/None= 0 A Few Times (1-3 times/Yr)=1 Many Times (3-5 times/Yr)=2 Very Many (>5 times/Yr)=3 Sub-total Missing System Total 726 Table 7.5.15: Frequency of Household Savings and Loan Activities - Descriptive How often in a year have member of the HH done any of the following? SCALE Never= 0 A Few Times (1-3 times)=1 Many Times (3-5 times)=2 Very Many (>5 times)=3 1. Save money in the bank? (FSLA_1) 2. Save money with the Susu Collector (FSLA_2) 3. Take a bank loan (FSLA_3) 4. Borrow money from the Susu Collector (FSLA_4) NON-MTS in NSEMSAW N=116 YAYA N=120 MTS in YAYA N=203 POOLED N=439 YR 1999 1.3947 0.7481 1.3697 0.6872 1.3958 0.7857 1.3882 0.7476 SD SD M SD M M SD M 2009 1.3982 0.7263 1.4417 0.7537 1.4615 0.7477 1.4393 0.7425 1999 1.0877 0.4324 1.3025 0.7196 1.2865 0.6522 1.2376 0.6276 2009 1.0442 0.3102 1.3729 0.8453 1.3854 0.7289 1.2908 0.6974 1999 1.1228 0.3555 1.1667 0.4734 1.1771 0.4218 1.1596 0.4205 2009 1.1062 0.3371 1.2083 0.5484 1.1949 0.4343 1.1752 0.4484 1999 1.0175 0.1319 1.0336 0.1810 1.1088 0.3440 1.0634 0.2625 2009 1.0088 0.0941 1.1008 0.3990 1.1289 0.3655 1.0892 0.3312 Frequency of HH Banking Activities Index (FHHBAI) 727 12QFSLA_3*4QFSLA_3*QFSLA_FHHBAI21it21it41itt=========nininin TOTAL HOUSEHOLD SAVINGS AND LOAN AMOUNT Table 7.5.16: Total Household Savings and Loan Amount Taken - Frequency Describe the following? 1. Total HH Savings? (S_L_1) 2. Total Loans? (S_L_2) 1999 2009 1999 2009 158 129 288 281 (36.0) (29.4) (65.6) (64.0) 195 206 88 85 (44.4) (46.9) (20.0) (19.4) 49 68 22 (11.2) (15.5) (5.0) 7 7 1 31 (7.1) 2 (1.6) (1.6) (0.2) (0.5) 409 410 399 399 (93.2) (93.4) (90.9) (90.9) 30 (6.8) 29 (6.6) 40 (9.1) 40 (9.1) 439 (100) 439 (100) 439 (100) 439 (100) Frequency (Percentage) 0GHC=1 1-500GHC=2 501-1000GHC=3 >1000GHC=4 Sub-total Missing System Total 728 A7.1.7 Total Household Savings and Loan Amount Table 7.5.17: Total Household Savings and Loan Amount Taken - Descriptive Describe the following? SCALE 0GHC=1 1-500GHC=2 501- 1000GHC=3 >1000GHC=4 1. Total HH Savings? (S_L_1) 2. Total Loans (S_L_2) NSEMSAW N=116 NON-MTS in YAYA N=120 MTS in YAYA N=203 POOLED N=439 YR 1999 1.8624 0.7873 1.7009 0.6600 1.7541 0.7184 1.7677 0.7224 SD SD M SD M SD M M 2009 1.7798 0.5987 1.8609 0.7479 1.9624 0.7875 1.8854 0.7328 1999 1.3832 0.6819 1.3190 0.5532 1.3239 0.5583 1.3383 0.5917 2009 1.3084 0.5030 1.3913 0.6582 1.4237 0.7199 1.3835 0.6503 Total Saving and Loan Amount Index (TSLAI) In table 7.1.6, respondents were asked to select based on a four-point scale their household’s total savings and loans for 1999 and 2009. The mean results suggest very little differences in total household savings and loans/ credit amounts for the two time periods. The mean results obtained from table 7.1.7 were used to compute the “Total Savings and Loan Amount Index (TSLAI)” using the above formula for TSLAI. 729 8S_L_1Q 4*2S_L_1Q *S_L_1Q TLSAI21it21it21itt=========nininiSn ANNUAL HOUSEHOLD INCOME Table 7.5.18: Annual Household Income - Frequency Frequency (Percentage) 0GHC=1 1-500GHC=2 501-1000GHC=3 >1000GHC=4 Sub-total Missing System Total Describe the following? 1. Annual HH income from all sources including MTS? (CA_HH) 1999 33 (7.5) 252 (57.4) 78 (17.8) 24 (5.4) 387 (88.2) 52 (11.8) 439 (100) 2009 12 (2.7) 221 (50.3) 135 (30.8) 21 (4.8) 389 (88.6) 50 (11.4) 439 (100) A7.1.8 Annual Household Income Table 7.5.19: Annual Household Income - Descriptive NSEMSAW N=116 NON-MTS in YAYA N=120 MTS in YAYA N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 2.2672 0.7385 2.1835 0.6690 2.2593 0.67387 2.2403 0.6916 2009 2.3534 0.5787 2.3670 0.7027 2.5122 0.6410 2.4242 0.6442 730 Describe the following? SCALE 0GHC=1 1-500GHC=2 501-1000GHC=3 >1000GHC=4 1. Annual HH income? (CA_HH) TREND IN ANNUAL HOUSEHOLD INCOME Table 7.5.20: Annual Household s Income Sufficiency - Frequency Describe the following? 2. Trend in Annual HH income? (GTA_HH) 1999 104 (23.7) 82 (18.7) 76 (17.3) 89 (20.3) 69 (15.7) 420 (95.7) 19 (4.3) 439 (100) 2009 120 (27.3) 92 (21.0) 106 (24.1) 52 (11.8) 52 (11.8) 422 (96.1) 17 (3.9) 439 (100) Frequency (Percentage) Decreased and not sufficient=1 Increased but not sufficient=2 Decreased but sufficient=3 Stable but sufficient=4 Increased and sufficient=5 Sub-total Missing System Total 731 A7.1.9 Annual Household Income Table 7.5.21: Annual Household Income and Sufficiency - Descriptive Describe the following? SCALE Decreased and not sufficient=1 Increased but not sufficient=2 Decreased but sufficient=3 Stable but sufficient=4 Increased and sufficient=5 2. Trend in Annual HH income? (GTA_HH) NSEMSAW N=116 NON-MTS in YAYA N=120 MTS in YAYA N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 3.0948 1.6311 2.906 1.30637 2.6631 1.34359 2.8500 1.4273 2009 2.3448 1.3391 2.5128 1.3105 2.7725 1.3433 2.5829 1.3424 Annual HH Income and Sufficiency Index (AHHIS) Table 7.1.7 describes households’ annual income and trends in household earnings. To gauge the trends in household earnings, table 7.1.7 uses a five-point scale to try to understand trends in household income as well as the degree to which earnings are sufficient for the household. Though annual household income increased by nearly 250 Ghana Cedis between 1999 and 2009, average annual household income for the two time periods still remained within 732 9]GTA_HHQ[CA_HHQ]54[]GTA_HHQ[CA_HHQ][]GTA_HHQ[CA_HHQAHHIStttt21ttt+=++=++=SS9GTA_HHQ CA_HHQ AHHIS51it41itt+=====nini the lower range of 500 and 1000 Ghana Cedis. In terms of income sufficiency, the average household was slightly worse of in 2009 compared to 1999 due to an almost 0.30 points decline on the income sufficiency scale between the two time periods. On the average trends in annual income for the average household in the study range from increase but not sufficient and decreased but sufficient. The two questions in the table 7.1.7 were used to generate the “Annual Household Income and Sufficiency Index” (AHHIS). A high score on the AHHIS means annual income of the average household within the study group in question improved in cash and sufficiency terms. The maximum attainable score on the AHHIS index is 9 hence a household’s score on AHHIS is obtained by expressing the total score obtained for the household on the two question in table 7.1.7 over the total attainable score of 9 (4+5). 733 VILLAGE STORE Table 7.5.22: Number of Village Stores or Kiosks that Sell Consumer Goods - Frequency Frequency (Percentage) 1. How many shops sell consumer goods in your village? (CG_1) None (0)=1 Few (1-3 )=2 Many (3-5)=3 Very Many (>5)=4 Sub-total Missing System Total 1999 148 (33.7) 211 (48.1) 40 (9.1) 20 (4.6) 419 (95.4) 20 (4.6) 439 (100) 2009 62 (14.1) 212 (48.3) 108 (24.6) 44 (10) 426 (97.0) 13 (3.0) 439 (100) A7.1.10 Village Store and Average Price of Consumer Goods Table 7.5.23: Number of Village Stores or Kiosks that Sell Consumer Goods - Descriptive NSEMSAW N=116 NON-MTS in YAYA N=120 MTS in YAYA N=203 POOLED N=439 M SD M SD M SD M SD 1999 1.5391 0.5664 2.0504 0.8621 1.8865 0.7961 1.8377 0.7841 Describe the following? SCALE None (0)=1 Few (1-3 )=2 Many (3-5)=3 Very Many (>5)=4 YR 1. How many shops sell consumer goods in your village? (CG_1) 2009 1.7391 0.5313 2.5500 0.7762 2.5131 0.8817 2.3146 0.8455 734 AVERAGE PRICE OF ITEMS IN THE VILLAGE STORE Table 7.5.24: Average Price of Items Sold in Village Store- Frequency Frequency (Percentage) 2. Average price of the three most expensive items in the village shop? (PECGV) 0GHC=1 1-5GHC=2 6-10GHC=3 >10GHC=4 Sub-total Missing System Total 1999 99 (22.6) 4 (0.9) 46 (10.5) 255 (58.1) 404 (92.0) 35 (8.0) 439 (100) 2009 48 (10.9) 38 (8.7) 76 (17.3) 253 (57.6) 415 (94.5) 24 (5.5) 439 (100) A7.1.11 Village Store and Average Price of Consumer Goods Table 7.5.25: Average Price of Items Sold in Village Store- Frequency - Descriptive Describe the following? SCALE 0GHC=1 0-5GHC=2 6-10GHC=3 >10GHC=4 2. Average price of the three most expensive items in the village shop? (PECGV) NSEMSAW N=116 NON-MTS in YAYA N=120 MTS in YAYA N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 1.6283 0.5378 1.9658 0.5862 2.0057 0.6496 1.8886 0.6222 2009 1.8333 0.6228 2.3866 0.6652 2.4044 0.7917 2.2506 0.7770 735 Table 7.1.8 above describes the number of shops, store or kiosks in the research communities that sell everyday consumer goods such as soap, cooking oil, sugar, light or touch batteries and canned food products. While this table was not included in the livelihood indexes computations, it provides a background to understanding table 7.1.9 which describes household expenditure and frequency with which certain items were purchased from the village store. It can be deduced from table 7.1.8, that fewer shops in most of the communities surveyed sold consumer good in 1999 compared to 2009. From the survey results, approximately 35% (n=148) of respondents do not recall having a consumer goods shop in their communities in 1999 compared to only 14% (62) who said they do not have a single shop in 2009. Table 7.1.8 shows that increase in the number of stores or kiosks may be a direct response to the upward prices of basic consumer goods between 1990 and 2009. A7.1.12 Household Expenditure in the Village Store In table 7.1.9, survey respondents were asked how often in a month they purchased any of the three most expensive items in their village store and also the average amount in Ghana Cedis (GHC) spent per visit to the store in 1999 and 2009. From the results, approximately 60% (n=243) of respondents said on average they purchase any of the three most expensive items in the village store anywhere from one to three times (M=1.9951) in a month while another 18% (n=74) purchase them more than 3 times in a month. Compared to 1999, in 2009 approximately 55% (n=230) purchase the three most expensive items between one and three times (M=2.2976) each month while about 34% (n=144) purchase those same three items more than three times in a month. The nearly 100% increase in household’s monthly expenditure on any of the three most expensive items in village stores between 1999 and 2009 may be an indication of a general improvement in household’s income status and also the relative abundance of consumer good 736 and shops in 2009 relative to 1999. How much of these improvements may be attributed to the MTS afforestation program will be discussed in chapter eight. FREQUENCY OF PURCHASE OF THREE MOST EXPENSIVE ITEMS IN VILLAGE STORE Table 7.5.26: Household Expenditure in Village Store - Frequency Frequency (Percentage) 3. How often in a month does HH purchased the three highest priced items in the village store? (FP_W) 1999 91 (20.7) 243 (55.4) 59 (13.4) 15 (3.4) 408 (92.9) 31 (7.1) 439 (100) 2009 46 (10.5) 230 (52.4) 117 (26.7) 27 (6.2) 420 (95.7) 19 (4.3) 439 (100) Never/None (0 )=1 Few (1-3) =2 Many (3-5) =3 Very Many (>5) = 4 Sub-total Missing System Total 737 Table 7.5.27: Household Expenditure in Village Store - Descriptive NSEMSAW N=116 NON-MTS in YAYA N=120 MTS in YAYA N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 1.7982 0.6807 2.0513 0.7052 2.0847 0.7298 1.9951 0.7183 2009 2.0614 0.7903 2.4153 0.7434 2.3850 0.7415 2.2976 0.7470 Describe the following? SCALE None (0)=1 Few (1-3 )=2 Many (3-5)=3 Very Many (>5)=4 3. How often in a month does HH purchased the three highest priced items in the village store? (FP_W) 738 AVERAGE MONTHLY EXPENDITURE AT THE VILLAGE STORE Table 7.5.28: Household Expenditure in Village Store – Frequency Frequency (Percentage) 4. Average HH monthly expenditure at village shop? (AEVSS) 0GHC=1 1-5GHC=2 6-10GHC=3 >10GHC=4 Sub-total Missing System Total 2009 44 (10.0) 5 (1.1) 78 (17.8) 286 (65.1) 413 (94.1) 26 (5.9) 439 (100) 1999 89 (20.3) 1 (0.2) 35 (8.0) 275 (62.6) 400 (91.1) 39 (8.9) 439 (100) Table 7.5.29: Household Expenditure in Village Store - Descriptive NSEMSAW N=116 NON-MTS in YAYA N=120 MTS in YAYA N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 1.6670 0.5431 1.9570 0.5500 1.9470 0.5247 1.8700 0.5512 2009 1.8600 0.5780 2.2370 0.5650 2.1770 0.5393 2.1070 0.5773 739 Describe the following? SCALE 1= 0GHC, 2= 1-5GHC 3= 6-10GHC 4= >10GHC 4. Average HH monthly expenditure at village shop? (AEVSS) Village Shop/Store Expenditure Index (VSEI) The second question in table 7.1.9 investigates average household’s expenditure per visit to the village store for both periods 1999 and 2009. According to the survey results, majority (M=1.8700) of households in 1999 between 1-5GHC per visit to the village store compare to a majority (M=2.1070) who spend within the same range in 2009. A closer review of the summary statistics indicates that in 1999, approximately 69% (n=275) on average spend in the range of 1- 5GHC per visit to the village store. Compared to 1999, in 2009 about the same number of respondents (69%, n=286) spend within the same range of 1-5GHC per visit to the village store. The difference in village expenditure between 1999 and 2009 is that in 1999 22% (n=89) of respondents/households had no village store expenditure compare to only 10% (n=44) who did not spend in the village store in 2009 (a 12% point increase over 1999). The mean responses in table 7.1.9 were used to generate the “Village Store Expenditure Index” (VSEI) for the purpose of comparing disaggregated between and within group difference among the three research groups. In order to obtain VSEI the total score attained by each household is divided by the maximum attainable score of 8 (4+4) in table 7.1.9 above. 740 8]AEVSS[FP_WQ]44[]AEVSS[FP_WQ][]AEVSS[FP_WQVSEItttt21ttt+=++=++=SS8AEVSS FP_WQVSEI41it41itt+=====nini SALE OF HOUSEHOLD POSSESSIONS Table 7.5.30: Sale of Household Possessions - Frequency Describe the following? 1. Frequency of sales of HH items due to dire needed cash? (DHHI_C) 1999 26 (5.9) 47 (10.7) 112 (25.5) 245 (55.8) 430 (97.9) 9 (2.1) 439 (100) 2009 20 (4.6) 45 (10.3) 117 (26.7) 250 (56.9) 432 (98.4) 7 (1.6) 439 (100) Frequency (Percentage) Very Many (>5 times)=1 Many Times (3-5 times)=2 A Few Times (1-3 times)=3 Never (0 times)= 4 Sub-total Missing System Total 741 A7.1.13 Sale of Household Possession Table 7.5.31: Sale of Household Possessions - Descriptive NSEMSAW N=116 NON-MTS in YAYA N=120 MTS in YAYA N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 3.4569 0.8064 3.3025 0.9165 3.2923 0.9370 3.3395 0.8986 2009 3.5043 0.7300 3.3109 0.9183 3.3535 0.8704 3.3819 0.8508 Describe the trends in sales/ accumulations of HH possessions? SCALE Very Many (>5 times)=1 Many Times (3-5 times)=2 A Few Times (1-3 times)=3 Never (0 times)= 4 1. Frequency of sales due to badly needed cash in the HH? (DHHI_C) 742 SALE OF HOUSEHOLD POSSESSIONS Table 7.5.32: General Trend in Accumulation of HH Possessions - Frequency Describe the following? 2. General trend in accumulation of HH possessions? (THH_P) 1999 108 (24.6) 96 (21.9) 131 (29.8) 87 (19.8) 5 (1.1) 427 (97.3) 12 (2.7) 439 (100) 2009 139 (31.7) 95 (21.6) 94 (21.4) 82 (18.7) 16 (3.6) 426 (97.0) 13 (3.0) 439 (100) Frequency (Percentage) Decreased and not sufficient=1 Increased but not sufficient=2 Decreased but sufficient=3 Stable but sufficient=4 Increased and sufficient=5 Sub-total Missing System Total 743 A7.1.14 Sale of Household Possession Table 7.5.33: General Trend in Accumulation of HH Possessions - Descriptive Describe the trends in sales/ accumulations of HH possessions? SCALE Decreased and not sufficient=1 Increased but not sufficient=2 Decreased but sufficient=3 Stable but sufficient=4 Increased and sufficient=5 2. General trend in HH possession? (THH_P) NSEMSAW N=116 NON-MTS in YAYA N=120 MTS in YAYA N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 2.7931 1.1907 2.4958 1.0486 2.3177 1.0674 2.4965 1.1120 2009 2.0345 1.1416 2.4286 1.1902 2.5864 1.2572 2.3920 1.2267 HH Item Liquidation Index (HHILI) “Afflicted by sickness, pregnancies and birth, short of food, with food prices high, and with the high need for energy for agricultural work, the poor people are often driven to distress, sales or borrowing. They sell or mortgage land, livestock, jewelry, their future crop, or their future labor; they beg from patrons; they become indebted to money lenders. The seasonal crisis drives them into dependence. Moreover, the knowledge that there will be the future crisis constraints them to be on good terms with their patrons. They are thus seasonally screwed down into, and kept down in, subordinate and dependent relationships in which they are vulnerable to exploitations. The poor are subordinated to the less poor or the rich; and within the family, the women may be subordinated to the men. Seasonal 744 9]THH_PQ [DHHI_CQ]54[]THH_PQ [DHHI_CQ][]THH_PQ [DHHI_CQHHILItttt21ttt+=++=++=SS9THH_PQ DHHI_CQHHILI51it41itt+=====nini stress may, in fact, be passed on down the line from the stronger to the weaker, culminating in the women and children, the indigent and the aged. This is, then, a time of the year when many dependent and exploitative relationships begin and are reinforced and deepened (Chambers, 1979, p.5).” In table 7.1.10, survey respondents were asked how often they sold their personal belongings out of need for cash and the general trend in household possessions between 1999 and 2009. The two questions in table 7.1.10 were included in an attempt to capture desperation within household trapped in extreme financial poverty or challenges faced between two time periods. The mean results obtained for 1999 (M=3.3395) and 2009 (M=3.3819) suggests that households’ almost never sell their personal belongings when hit with hard times. From the summary statistics, approximately 56% (n=245) of households said they never sold their personal belonging in 1999 compared to 57% (a slight increase) who never sold in 2009. Often times for cultural reason and personal uneasiness discussing issues of extreme household poverty with non-family members, households may hesitate to divulge pertinent information about desperate financial conditions particular if it leads to the sale of household personal effects or possessions. Hence to triangulate the first question in table 7.1.10, a second question asked differently was used to further probe households’ financial situations in case the first question failed to elicit an honest response. In the second question, respondents were asked to describe using a five point scale the general trends in household possessions and also the degree to which household possessions were sufficient for livelihood support. When the response to the second question is collapsed onto a two-point scale (0=Increased, decreased, stable but not sufficient, 1= Increased, decreased, stable but sufficient), approximately 49% (n=215) said regardless of the trends in their basic household possessions, on the whole in 1999 their household possessions was not sufficient. Compared to 1999, approximately 56% (n=246) of respondents said regardless of the 745 trends, their household possessions, were not sufficient for the household in 2009 (7% increase over 1999). The responses in table 7.1.10 was used to compute the “Household Item Liquidation Index” (HHILI) by summing up the mean responses to questions one and two and dividing by the maximum attainable score of 9 (4+5). A7.1.15 Combined Household Financial Capital Index After computing all the 10 indexes for the various financial capital indicators presented in tables 7.1.1 through 7.1.10 above, the next step aggregated all the indexes and expressed them as a fraction of their total. Hence assuming a household achieved the maximum score of 1 on each of the 10 indexes above that household achieves a perfect score of 10 thus when expressed as a fraction of 10 the household’s “Combined Household Financial Capital Index” (CHHFCI) would be 1. The formula below described how CHHFCI is computed using the rest of the other financial capital indexes above. Combine HH Financial Capital Index (CHHFCI) 746 10)(n Indeces Capital FinancialofNumber TotalIndeces) Capital Financial(All CHHFCI101tt====ni]10[] HHILI+VSEI +AHHIS +TLSAI + FHHBAI+ HHBAI+ HHE2+ HEI1+SHISI +[PHISICHHFCIt= APPENDIX B: Descriptive Analysis of Human Capital Indexes A7.2.1 Education and Literacy Table 7.5.34: English and Local Language Literacy Is there anyone in the Household 1. Older than15yrs that can read and write English? (E1) 2. Older than15years that can read and write at least one Ghanaian language? (E2) 3. Between 6-15 that has never been to school? (E3) 4. Older than15yrs that has completed at least primary education? (E4) 1999 2009 1999 2009 1999 2009 1999 2009 260 (59.2) 234 (53.3) 256 (58.3) 221 (50.3) 292 (66.5) 299 (68.1) 252 (57.4) 240 (54.7) 138 (31.4) 179 (40.8) 140 (31.9) 168 (38.3) 86 90 (19.6) (20.5) 136 (31.0) 163 (37.1) 398 (90.7) 413 (94.1) 396 389 378 389 388 403 (90.2) (88.6) (86.1) (88.6) (88.4) (91.8) 41 (9.3) 26 (5.9) 43 (9.8) 50 61 50 51 (11.4) (13.9) (11.4) (11.6) 36 (8.2) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) Frequency (Percentage) No (1) Yes (2) Sub-total Missing System (N/A) Total 747 Table 7.5.35: Primary, Secondary and College Level Education - Frequency Is there anyone in the Household Frequency (Percentage) 5. Older than 18yrs that has completed at least JSS? (E5) 6. Older than 18yrs has completed at least SSS? (E6) 7. Younger than 6yrs in pre- school (Nursery and Kindergarten)? (E7) 8. Older than 18yrs and has attained post- secondary school level education? (Including training college, nursing, Polytechnics, University, etc. (E8) No (1) Yes (2) Sub-total 1999 2009 1999 2009 1999 2009 1999 2009 263 (59.9) 147 (33.5) 304 (69.2) 290 (66.1) 268 (61.0) 251 (57.2) 336 (76.5) 340 (77.4) 120 (27.3) 252 (57.4) 67 97 (15.3) (22.1) 100 (22.8) 137 (31.2) 29 (6.6) 37 (8.4) 383 399 371 387 368 388 (87.2) (90.9) (84.5) (88.2) (83.8) (88.4) 365 (83.1) 377 (85.9) Missing System (N/A) 56 (12.8) 40 (9.1) 68 52 71 51 74 62 (15.5) (11.8) (16.2) (11.6) (16.9) (14.1) Total 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 748 Is there anyone in the household that has accomplished the following? SCALE No=0 Yes=1 1. Is there anyone in the HH that has attended Day Nursery, Kindergarten or Primary School? (EDUCATION1) 2. Is there anyone in the HH that has attended JSS or SSS? (EDUCATION2) 3. Is there anyone in the HH that can read or write English or any Ghanaian language? (LITERACY) Table 7.5.36: Education and Literacy - Descriptive NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.3362 0.4745 0.4833 0.5018 0.4729 0.5005 0.4396 0.4969 2009 0.4224 0.4961 0.5833 0.4951 0.5961 0.4919 0.5467 0.4984 1999 0.1810 0.3867 0.4167 0.4951 0.3153 0.4658 0.3075 0.4620 2009 0.2414 0.4298 0.4583 0.5004 0.4089 0.4928 0.3781 0.4855 1999 0.2069 0.4068 0.4333 0.4976 0.4039 0.4919 0.3599 0.4805 2009 0.3190 0.4681 0.5333 0.5010 0.4631 0.4999 0.4442 0.4974 749 Household Education and Literacy Index (EDULIT) 750 1*3LITERACY)EDUCATION21(EDUCATION*LITERACY)EDUCATION21(EDUCATIONEDULIT3131titit====++=++=niniSn3LITERACY)EDUCATION21(EDUCATIONEDULIT31tit==++=ni A7.2.2 Household Migration Table 7.5.37: Household Migration out of the Village - Frequency Does any member of the Household Frequency (Percentage) 1. Older than 12yrs work outside of the village on a daily basis? (MIGWK_1) 2. Older than 12yrs migrate from the village to work outside for a long period of time (i.e. >3mths at a time)? (MIGWK_2) No (1) Yes (2) Sub-total Missing System Total 1999 321 (73.1) 42 (9.5) 363 (82.7) 76 (17.3) 439 (100) 2009 328 (74.7) 60 (13.6) 388 (88.4) 51 (11.6) 439 (100) 1999 317 (72.2) 44 (10.1) 361 (82.2) 78 (17.8) 439 (100) 2009 300 (68.4) 84 (19.1) 384 (87.5) 55 (12.5) 439 (100) 751 Table 7.5.38: Household Migration out of the Village - Descriptive Does any member of the household older than 12 years of age engage in any of the following? SCALE No=0 Yes=1 1. Work outside of the village on a daily basis? (MIGWK_1recode) 2. Work outside for a long period of time (i.e. >3mths at a time)? (MIGWK_2recode) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.0172 0.1307 0.1500 0.3586 0.1084 0.3116 0.0957 0.2945 2009 0.0259 0.1594 0.2167 0.4137 0.1527 0.3606 0.1367 0.3439 1999 0.0345 0.1833 0.1167 0.3224 0.1281 0.3350 0.1002 0.3007 2009 0.0431 0.2040 0.2417 0.4299 0.2463 0.4319 0.1913 0.3938 Household Migration Index (MIGI) 752 1*2ode)MIGWK_2reccode(MIGWK_1re*ode)MIGWK_2reccode(MIGWK_1reMIGI21it21itt====+=+=niniSn2ode)MIGWK_2reccode(MIGWK_1reMIGI21itt==+=ni A7.2.3 Household Dietary Diversity (HDD) Table 7.5.39: HDD-Frequency of Consumption of Major Staples - Frequency Frequency (Percentage) Never=0 Occasionally=1 Often=2 Always=3 Sub-total Missing System Total 1. Foods made from cereals/grains e.g. Banku, bread, rice noodles? (HHDD1_Freq) 2. . Foods with yellow or orange inside e.g. carrots, squash, or sweet potato? (HHDD2_Freq) 3. Foods made from roots yams and cassava e.g. Fufu? (HHDD3_Freq) 2009 1999 2009 61 (13.9) 69 (15.7) 88 (20.0) 220 (50.1) 438 (99.8) 1 (0.2) 439 (100) 103 (23.4) 74 (16.9) 128 (29.2) 98 (22.3) 403 (91.8) 36 (8.2) 439 (100) 140 (31.9) 70 (15.9) 112 (25.5) 80 (18.2) 402 (91.6) 37 (8.4) 439 (100) 1999 10 (2.3) 15 (3.4) 70 (15.9) 339 (77.2) 434 (98.9) 5 (1.1) 439 (100) 2009 12 (2.7) 10 (2.3) 39 (8.9) 374 (85.2) 435 (99.1) 4 (0.9) 439 (100) 1999 27 (6.2) 81 (18.5) 115 (26.2) 214 (48.8) 437 (99.5) 2 (0.5) 439 (100) 753 Table 7.5.40: HDD-Frequency of Consumption of Major Staples - Descriptive How often does the household eat the following foods? SCALE Never (At most ones a year)=0 Occa (At most ones in 6months)=1 Often (At least ones a month)=2 Always (At least ones a week)=3 1. Foods made from cereals/grains e.g. banku, bread, rice noodles? (HHDD1_Freq) 2. Foods with yellow or orange inside e.g. carrots, squash, or sweet potato? (HHDD2_Freq) 3. Foods made from roots yams and cassava e.g. fufu? (HHDD3_ Freq) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 2.4140 0.7810 2.2170 0.9454 2.0050 1.0172 2.1710 0.9535 2009 2.0690 0.6944 2.0000 0.5186 2.1080 0.5786 2.0680 0.5967 1999 1.3450 1.1580 1.4670 1.0764 1.4380 1.1940 1.4210 1.1518 2009 1.4740 0.9732 1.9670 0.8593 1.8180 0.9127 1.7680 0.9318 1999 2.6640 0.7336 2.6580 0.6673 2.6800 0.7113 2.6700 0.7040 2009 2.0260 0.3836 2.0580 0.3734 2.1030 0.4605 2.0710 0.4188 Staple Food Consumption Frequency Index (FSFCI) 754 3*3 Freq) HHDD3_+nd HHDD2_Tre+q(HHDD1_Fre* Freq) HHDD3_+nd HHDD2_Tre+q(HHDD1_FreFSFCI31it31itt======niniSn9 Freq) HHDD3_+q HHDD2_Fre+q(HHDD1_FreFSFCI31itt===ni Table 7.5.41: HDD- Frequency of Consumption of Vegetables and Fruits - Frequency 4. Dark, green, leafy vegetables such as cassava leaves, bean leaves, spinach etc. (HHDD4_Freq) 1999 26 (5.9) 23 (5.2) 104 (23.7) 281 (64.1) 434 (98.9) 5 (1.1) 439 (100) 2009 32 (7.3) 20 (4.6) 78 (17.8) 303 (67.0) 433 (98.6) 6 (2.1) 439 (100) Frequency (Percentage) Never=0 Occasionally =1 Often=2 Always=3 Sub-total Missing System Total 5. Any other vegetables? (HHDD5_Freq) 2009 10 (2.3) 33 (7.5) 106 (24.1) 245 (55.8) 394 (89.7) 45 (10.3) 439 (100) 1999 8 (1.8) 35 (8.0) 135 (30.8) 220 (50.1) 398 (90.7) 41 (9.3) 439 (100) 6. Foods rich in vitamin A e.g. 7. Any other ripe mangoes and fruits? (HHDD7_Freq) papaya? (HHDD6_Freq) 1999 6 (1.4) 80 2009 12 (2.7) 77 (18.2) (17.5) 83 1999 8 (1.8) 37 (8.4) 160 2009 10 (2.3) 47 (10.7) 132 96 (21.9) 249 (56.7) 431 (98.2) 8 (1.8) 439 (100) (18.9) (36.4) (30.1) 263 (59.9) 435 (99.1) 4 (0.9) 439 (100) 201 (45.8) 406 (92.5) 33 (7.5) 439 (100) 222 (50.6) 411 (93.6) 28 (6.4) 439 (100) 755 Table 7.5.42: HDD- Frequency of Consumption of Vegetables and Fruits - Descriptive How often does the household eat the following foods? SCALE Never (At most ones a year)=0 Occa (At most ones in 6months)=1 Often (At least ones a month)=2 Always (At least ones a week)=3 4. Dark, green, leafy vegetables such as cassava leaves, bean leaves, spinach etc? (HHDD4_ Freq) 5. Any other vegetables? (HHDD5_ Freq) 6. Foods rich in vitamin A e.g. ripe mangoes and papaya? (HHDD6_ Freq) 7. Any other fruits? (HHDD7_ Freq) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 2.5260 0.8389 2.3920 0.8431 2.4330 0.9277 2.4460 0.8817 2009 2.0600 0.5483 2.1670 0.5397 2.0790 0.5833 2.0980 0.5629 1999 2.1120 1.0448 2.0920 0.9958 2.3100 0.9582 2.1980 0.9951 2009 1.9480 0.8831 2.1170 0.8519 2.0100 0.7771 2.0230 0.8273 1999 2.2840 0.9022 2.1250 0.9309 2.4580 0.8157 2.3210 0.8807 2009 2.1120 0.4328 2.1500 0.5290 2.2120 0.5165 2.1690 0.5001 1999 2.2160 0.9492 2.0670 0.9591 2.2410 0.9205 2.1870 0.9396 2009 2.0780 0.7360 2.2420 0.7559 2.1720 0.7609 2.1660 0.7538 756 Fruits and Vegetables Consumption Frequency Index (FFVCI) 757 Snni* Freq) HHDD7_+ Freq HHDD6_+ Freq HHDD5_+ Freq(HHDD4_FFVCI41itt===3*4 Freq) HHDD7_+ Freq HHDD6_+ Freq HHDD5_+ Freq(HHDD4_FFVCI41itt===ni12 Freq) HHDD7_+ Freq HHDD6_+ Freq HHDD5_+ Freq(HHDD4_FFVCI41itt===ni Table 7.5.43: HDD- Frequency of Consumption of Animal and Plant Proteins - Frequency 8. Meat e.g. Beef, pork, lamb, goat, wild game, Frequency (Percentage) 12. Cheese, yoghurt, milk, or other 10. Fresh or dried fish or 11. Foods made from 9. Eggs (HHDD9_ Freq) rabbit, chicken, duck, any other bird? (HHDD8_ Freq) 1999 2009 1999 2009 1999 2009 1999 2009 1999 2009 beans, lentils, peas, or nuts? (HHDD11_ Freq) products? (HHDD12_ Freq) shellfish? (HHDD10_ Freq) milk Never=0 18 (4.1) 38 (8.7) 33 (7.5) 51 7 (11.6) (1.6) 21 (4.8) 39 (8.9) 47 (10.7) 159 (36.3) 158 (36.0) Occasionally=1 75 61 96 87 (17.1) (13.9) (21.9) (19.8) 28 (6.4) 17 75 66 (3.9) (17.1) (15.0) 122 (27.8) 108 (24.6) Often=2 Always=3 Sub-total 149 (33.9) 147 (33.5) 143 (32.6) 142 (32.3) 129 (29.4) 99 (22.6) 139 (31.7) 120 (27.3) 83 (18.9) 100 (22.8) 193 (44.0) 185 (42.1) 155 (35.3) 146 (33.2) 271 (61.7) 295 (67.2) 169 (38.8) 189 (43.0) 40 (9.1) 37 (8.5) 435 (99.1) 431 (98.2) 427 (97.3) 426 (97.0) 435 (99.1) 432 (98.4) 422 (96.1) 422 (96.1) 404 (92.0) 403 (91.8) Missing System Total 4 8 (0.9) (1.8) 439 (100) 439 (100) 12 (2.7) 439 (100) 13 (3.0) 439 (100) 4 7 (0.9) (1.6) 439 (100) 439 (100) 17 (3.9) 439 (100) 17 (3.9) 439 (100) 35 (8.0) 439 (100) 36 (8.2) 439 (100) 758 Table 7.5.44: HDD- Frequency of Consumption of Animal and Plant Proteins - Descriptive How often does the household eat the following foods? SCALE Never (At most ones a year)=0 Occa (At most ones in 6months)=1 Often (At least ones a month)=2 Always (At least ones a week)=3 8. Meat e.g., beef, pork, lamb, goat, wild game, rabbit, chicken, duck, any other bird? (HHDD8_ Freq) 9. Eggs (HHDD9_ Freq) 10. Fresh or dried fish or shellfish? (HHDD10_ Freq) 11. Foods made from beans, lentils, peas, or nuts? (HHDD11_ Freq) 12. Cheese, yoghurt, milk, or other milk products? (HHDD12_ Freq) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 2.2930 0.7462 2.1330 0.9069 2.1180 0.9418 2.1690 0.8856 2009 2.1290 0.7971 2.2330 0.6181 2.2710 0.6297 2.2230 0.6759 1999 1.9570 0.9636 1.9420 0.9553 1.9060 1.0274 1.9290 0.9895 2009 2.2240 0.7470 2.1750 0.6438 2.1130 0.7847 2.1590 0.7384 1999 2.5520 0.6507 2.4420 0.7970 2.5120 0.7269 2.5030 0.7274 2009 2.0950 0.6183 2.1580 0.5345 2.1970 0.5634 2.1590 0.5711 1999 1.9220 0.9970 1.9750 1.0805 1.9700 1.0288 1.9590 1.0329 2009 1.9570 0.8484 2.1330 0.6208 2.1630 0.7163 2.1000 0.7334 1999 0.5600 0.7944 1.1330 1.0202 1.0200 1.0339 0.9290 0.9964 2009 1.3020 0.8259 1.8750 0.8751 1.8330 0.9342 1.7040 0.9210 759 Plants and Animal Protein Consumption Frequency Index (FPAPC) 760 Snni* Freq) HHDD12_+ Freq HHDD11_+ Freq HHDD10_+ Freq HHDD9_+ Freq(HHDD8_FPAPC51itt===3*5 Freq) HHDD12_+ Freq HHDD11_+ Freq HHDD10_+ Freq HHDD9_+ Freq(HHDD8_FPAPC51itt===ni15 Freq) HHDD12_+ Freq HHDD11_+ Freq HHDD10_+ Freq HHDD9_+ Freq(HHDD8_FPAPC51itt===ni Table 7.5.45: HDD- Frequency of Consumption of Vegetables and Fruits - Frequency Frequency (Percentage) 13. Foods made from oil, fat or butter? (HHDD13_ Freq) 14. Sugar or honey? (HHDD14_ Freq) 15. Foods such as condiments, cocoa, coffee, tea? (HHDD15_ Freq) Never=0 Occasionally= 1 Often=2 Always=3 Sub-total Missing System Total 1999 93 (21.2) 77 (17.5) 115 (26.2) 134 (30.5) 419 (95.4) 20 (4.6) 439 (100) 2009 95 (21.6) 70 (15.9) 108 (24.6) 147 (33.5) 420 (95.7) 19 (4.3) 439 (100) 1999 26 (5.9) 71 (16.2) 146 (33.3) 183 (41.7) 426 (97.0) 13 (3.0) 439 (100) 2009 48 (10.9) 51 (11.6) 131 (29.8) 197 (44.9) 427 (97.3) 12 (2.7) 439 (100) 1999 114 (26.0) 114 (26.0) 100 (22.8) 75 (17.1) 403 (91.8) 36 (8.2) 439 (100) 2009 122 (27.8) 115 (26.2) 100 (22.8) 71 (16.2) 408 (92.9) 31 (7.1) 439 (100) 761 Table 7.5.46: HDD- Frequency of Consumption of Other Food Items - Descriptive How often does the household eat the following foods? SCALE Never (At most ones a year)=0 Occa (At most ones in 6months)=1 Often (At least ones a month)=2 Always (At least ones a week)=3 13. Foods made from oil, fat or butter? (HHDD13_ Freq) 14. Sugar or honey? (HHDD14_ Freq) 15. Foods such as condiments, cocoa, coffee, tea? (HHDD15_ Freq) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 1.3450 1.2237 1.7170 1.1090 1.7090 1.1515 1.6150 1.1684 2009 1.7330 0.8481 2.0330 0.7440 2.0250 0.7863 1.9500 0.8008 1999 1.9140 1.0920 2.1580 0.9074 2.1230 0.9118 2.0770 0.9644 2009 1.9400 0.8875 2.2080 0.5926 2.2170 0.6389 2.1410 0.7107 1999 0.8360 0.9952 1.3250 1.0781 1.3940 1.1135 1.2280 1.0971 2009 1.4480 0.8978 1.9500 0.7430 1.9460 0.8743 1.8150 0.8733 Other Foods Consumption Frequency Index (FOFC) 762 3*3 Freq) HHDD15_+ Freq HHDD14_+ Freq(HHDD13_* Freq) HHDD15_+ Freq HHDD14_+ Freq(HHDD13_FOFC31it31itt======niniSn9 Freq) HHDD15_+ Freq HHDD14_+ Freq(HHDD13_FOFC31itt===ni 3e) Dietary Diversity Aggregate Frequency of Consumption Index (DDAFC) 3f) Combined Household Dietary Diversity Index for Trend in Sufficiency of Consumption (CHHDDI1) 763 Snni*eq) HHDD15_Fr+eq HHDD14_Fr+qHHDD13_Fre+qHHDD12_Fre +eq HHDD11_Fr+qHHDD10_Fre+q HHDD9_Fre+q HHDD8_Fre+HHDD7_Freq +q HHDD6_Fre+q HHDD5_Fre+q HHDD4_Fre+q HHDD3_Fre+q HHDD2_Fre+q(HHDD1_FreDDAFC151itt===45eq) HHDD15_Fr+eq HHDD14_Fr+qHHDD13_Fre+qHHDD12_Fre +eq HHDD11_Fr+qHHDD10_Fre+q HHDD9_Fre+q HHDD8_Fre+HHDD7_Freq +q HHDD6_Fre+q HHDD5_Fre+q HHDD4_Fre+q HHDD3_Fre+q HHDD2_Fre+q(HHDD1_FreDDAFC151itt===ninni===41itt FOFC)+ FPAPC+ FFVCI+(FSFCICHHDDI14 FOFC)+ FPAPC+ FFVCI+(FSFCICHHDDI141itt===ni 4a. HOUSEHOLD DIETARY DIVERSITY- CONSUMPTION TREND OF MAJOR STAPLE A7.2.4 Household Dietary Diversity (HDD) FOODS Table 7.5.47: HDD-Trend of Consumption of Major Staple Foods - Frequency Frequency (Percentage) Decreased and not sufficient =1 Increased but not sufficient =2 Decreased but sufficient =3 Stable but sufficient =4 Increased and sufficient=5 Sub-total Missing System Total 1. Foods made from cereals/grains e.g. Banku, bread, rice noodles? (HHDD1_Trend) 2. . Foods with yellow or orange inside e.g. carrots, squash, or sweet potato? (HHDD2_ Trend) 1999 2009 1999 2009 91 (20.7) 57 (13.0) 76 (17.3) 136 (31.0) 76 (17.3) 436 (99.3) 3 (0.7) 439 (100) 174 (39.6) 54 (12.3) 108 (24.6) 59 (13.4) 17 (3.9) 412 (93.8) 27 (6.2) 439 (100) 134 (30.5) 55 (12.5) 90 (20.5) 86 (19.6) 50 (11.4) 415 (94.5) 24 (5.5) 439 (100) 108 (24.6) 77 (17.5) 106 (24.1) 85 (19.4) 60 (13.7) 436 (99.3) 3 (0.7) 439 (100) 3. Foods made from roots yams and cassava e.g. fufu? (HHDD3_ Trend) 1999 37 (8.4) 19 (4.3) 70 (15.9) 155 (35.3) 157 (35.8) 438 (99.8) 1 (0.2) 439 (100) 2009 19 (4.3) 37 (8.4) 80 (18.2) 136 (31.0) 165 (37.6) 438 (99.8) 1 (0.2) 439 (100) 764 Table 7.5.48: HDD-Trend of Consumption of Major Staple Foods - Descriptive In terms of sufficiency, how would you describe trends in quantity of the following foods available to the household (1990- 1999 and 2000- 2009)? SCALE Decreased and not sufficient=1 Increased but not sufficient=2 Decreased but sufficient=3 Stable but sufficient=4 Increased and sufficient=5 YR 1. Foods made from cereals/grains e.g. bread, rice noodles? (HHDD1_Trend) 2. Foods with yellow or orange inside e.g. carrots, squash, or sweet potato? (HHDD2_Trend) 3. Foods made from roots yams and cassava? (HHDD3_Trend) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 M SD M SD M SD M SD 1999 3.1120 1.5647 3.1330 1.3593 3.0690 1.3480 3.0980 1.4084 2009 2.4740 1.4474 2.8500 1.3821 2.9260 1.3010 2.7860 1.3732 1999 2.4400 1.4879 2.5580 1.4189 2.6700 1.4017 2.5790 1.4295 2009 1.6900 0.9993 2.2000 1.2203 2.4330 1.3198 2.1730 1.2504 1999 3.8020 1.2936 3.9750 1.1187 3.8080 1.2013 3.8520 1.2044 2009 3.8280 1.1962 3.8080 1.2918 3.9560 1.0211 3.8820 1.1466 Staple Foods Sufficiency Index (SFSI) 765 5*3nd) HHDD3_Tre+nd HHDD2_Tre+nd(HHDD1_Tre*nd) HHDD3_Tre+nd HHDD2_Tre+nd(HHDD1_TreSFSI31it31itt======niniSn15nd) HHDD3_Tre+nd HHDD2_Tre+nd(HHDD1_TreSFSI31itt===ni 4b. HOUSEHOLD DIETARY DIVERSITY- CONSUMPTION TREND OF VEGETABLES AND FRUITS Table 7.5.49: HDD- Trend of Consumption of Vegetables and Fruits- Frequency 4. Dark, green, leafy vegetables such as cassava leaves, bean leaves, spinach etc? (HHDD5_Trend) 5. Any other vegetables? (HHDD5_ Trend) 6. Foods rich in vitamin A e.g. ripe mangoes and papaya? (HHDD6_ Trend) 7. Any other fruits? (HHDD7_ Trend) 1999 2009 1999 2009 1999 2009 1999 2009 50 (11.4) 43 (9.8) 24 (5.5) 46 (10.5) 34 (7.7) 26 (5.9) 26 (5.9) 34 (7.7) 21 (4.8) 15 (3.4) 58 64 (13.2) (14.6) 25 (5.7) 39 (8.9) 31 (7.1) 47 (10.7) 89 88 117 120 100 113 117 123 (20.3) (20.0) (26.7) (27.3) (22.8) (25.7) (26.7) (28.0) 125 119 143 138 136 133 146 132 (28.5) (27.1) (32.6) (31.4) (31.0) (30.3) (33.3) (30.1) 143 137 88 87 116 108 86 79 (32.6) (31.2) (20.0) (19.8) (26.4) (24.6) (19.6) (18.0) 431 433 408 405 431 433 413 412 (98.2) (98.6) (92.9) (92.3) (98.2) (98.6) (94.1) (93.8) 8 (1.8) 439 (100) 6 (1.4) 439 (100) 31 (7.1) 439 (100) 34 (7.7) 439 (100) 8 (1.8) 439 (100) 6 (1.4) 439 (100) 26 (5.9) 439 (100) 27 (6.2) 439 (100) Frequency (Percentage) Decreased and not sufficient =1 Increased but not sufficient =2 Decreased but sufficient =3 Stable but sufficient =4 Increased and sufficient=5 Sub-total Missing System Total 766 Table 7.5.50: HDD- Trend of Consumption of Vegetables and Fruits- Descriptive In terms of sufficiency, how would u describe trends in quantity of the following foods available to the household (1990-1999 and 2000- 2009)? SCALE Decreased and not sufficient=1 Increased but not sufficient=2 Decreased but sufficient=3 Stable but sufficient=4 Increased and sufficient=5 4. Dark, green, leafy vegetables such as cassava leaves, bean leaves, spinach etc? (HHDD4_Trend) 5. Any other vegetables? (HHDD5_Trend) 6. Foods rich in vitamin A e.g. ripe mangoes and papaya? (HHDD6_Trend) 7. Any other fruits? (HHDD7_Trend) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 3.5340 1.4352 3.6670 1.2790 3.6350 1.3220 3.6170 1.3393 2009 3.4570 1.4105 3.5250 1.3282 3.6550 1.2701 3.5670 1.3240 1999 3.3190 1.3358 3.3750 1.2640 3.3990 1.2678 3.3710 1.2825 2009 3.2590 1.3713 3.3670 1.2565 3.4140 1.2131 3.3600 1.2670 1999 3.6470 1.1960 3.4420 1.2216 3.6110 1.1863 3.5740 1.1987 2009 3.5090 1.2050 3.4750 1.1665 3.6260 1.1070 3.5540 1.1492 1999 3.3790 1.2205 3.4420 1.2353 3.3940 1.2358 3.4030 1.2291 2009 3.2240 1.2792 3.3080 1.3016 3.3150 1.2183 3.2890 1.2554 767 Fruits and Vegetables Sufficiency Index (FVSI) 768 Snni*nd) HHDD7_Tre+nd HHDD6_Tre+nd HHDD5_Tre+nd(HHDD4_TreFVSI41itt===5*4nd) HHDD7_Tre+nd HHDD6_Tre+nd HHDD5_Tre+nd(HHDD4_TreFVSI51itt===ni20nd) HHDD7_Tre+nd HHDD6_Tre+nd HHDD5_Tre+nd(HHDD4_TreFVSI31itt===ni 4c. HOUSEHOLD DIETARY DIVERSITY- CONSUMPTION TREND OF PLANT AND ANIMAL PROTEINS Table 7.5.51: HDD- Trend of Consumption of Plant and Animal Proteins - Frequency Frequency (Percentage) 8. Meat e.g. Beef, pork, lamb, goat, wild game, rabbit, chicken, duck, any other bird? (HHDD8_ Trend) 9. Eggs? (HHDD9_ Trend) 10. Fresh or dried fish or shell fish? (HHDD10_ Trend) 11. Foods made from beans, lentils, peas, or nuts? (HHDD11_ Trend) 12. Cheese, yoghurt, milk, or other milk products? (HHDD12_ Trend) 1999 2009 1999 2009 1999 2009 1999 2009 1999 2009 Decreased and not sufficient =1 Increased but not sufficient =2 Decreased but sufficient =3 Stable but sufficient =4 Increased and sufficient=5 Sub-total Missing System Total 67 (15.3 ) 63 (14.4 ) 103 (23.5 ) 134 (30.5 ) 64 (14.6 ) 431 (98.2 ) 8 97 74 88 40 36 (22.1) (16.9) (20.0) (9.1) (8.2) 69 61 72 37 43 (15.7) (13.9) (16.4) (8.4) (9.8) 106 115 108 98 98 (24.1) (26.2) (24.6) (22.3) (22.3) 110 (25.1) 127 112 160 145 (28.9) (25.5) (36.4) (33.0) 51 53 49 94 110 (11.6) (12.1) (11.2) (21.4) (25.1) 433 430 429 429 432 (98.6) (97.9) (97.7) (97.7) (98.4) 6 9 10 10 7 72 (16.4 ) 55 (12.5 ) 106 (24.1 ) 111 (25.3 ) 65 (14.8 ) 409 (93.2 ) 30 73 (16.6) 64 (14.6) 106 (24.1) 103 (23.5) 69 205 (46.7 ) 73 (16.6 ) 63 (14.4 ) 54 (12.3 ) 16 212 (48.3) 69 (15.7) 65 (14.8) 54 (12.3) 13 (15.7) (3.6) (3.0) 415 (94.5) 24 411 (93.6 ) 28 413 (94.1) 26 (1.8) (1.4) (2.1) (2.3) (2.3) (1.6) (6.8) (5.5) (6.4) (5.9) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 769 Table 7.5.52: HDD- Trend of Consumption of Plant and Animal Proteins - Descriptive In terms of sufficiency, how would u describe trends in quantity of the following foods available to the household (1990-1999 and 2000- 2009)? SCALE Decreased and not sufficient=1 Increased but not sufficient=2 Decreased but sufficient=3 Stable but sufficient=4 Increased and sufficient=5 8. Meat e.g. beef, pork, lamb, goat, wild game, rabbit, chicken, duck, any other bird? (HHDD8_Trend) 9. Eggs (HHDD9_Trend) 10. Fresh or dried fish or shell fish? (HHDD10_Trend) 11. Foods made from beans, lentils, peas, or nuts? (HHDD11_Trend ) 12. Cheese, yoghurt, milk, or other milk products? (HHDD12_Trend) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 2.9310 1.3751 3.1670 1.3178 3.1820 1.2552 3.1120 1.3064 2009 2.5090 1.2752 2.9500 1.4074 3.0000 1.3050 2.8560 1.3396 1999 2.7760 1.4451 3.0750 1.2580 3.1130 1.2072 3.0140 1.2927 2009 2.5520 1.3471 2.9750 1.3870 2.9850 1.2407 2.8680 1.3210 1999 3.4830 1.2954 3.5000 1.2636 3.4680 1.1952 3.4810 1.2383 2009 3.3880 1.3369 3.6170 1.2846 3.5760 1.1509 3.5380 1.2397 1999 2.7670 1.5057 2.9420 1.3799 3.0790 1.2950 2.9590 1.3793 2009 2.5860 1.4982 2.9830 1.3780 3.1630 1.2618 2.9610 1.3769 1999 1.5600 1.0737 2.1420 1.2853 2.0990 1.2187 1.9680 1.2234 2009 1.4400 0.8875 2.1670 1.3556 2.0940 1.1883 1.9410 1.2036 770 Plants and Animal Protein Sufficiency Index (PAPSI) 771 Snni*end) HHDD12_Tr+ndHHDD11_Tre+end HHDD10_Tr+nd HHDD9_Tre+nd(HHDD8_TrePAPSI51itt===5*5end) HHDD12_Tr+ndHHDD11_Tre+end HHDD10_Tr+nd HHDD9_Tre+nd(HHDD8_TrePAPSI51itt===ni25end) HHDD12_Tr+ndHHDD11_Tre+end HHDD10_Tr+nd HHDD9_Tre+nd(HHDD8_TrePAPSI51itt===ni 4d. HOUSEHOLD DIETARY DIVERSITY- CONSUMPTION TREND OF OTHER FOODS Table 7.5.53: HDD- Trend of Consumption of Other Foods - Frequency Frequency (Percentage) Decreased and not sufficient =1 Increased but not sufficient =2 Decreased but sufficient =3 Stable but sufficient =4 Increased and sufficient=5 Sub-total Missing System Total 13. Foods made from oil, fat or butter? (HHDD13_ Trend) 14. Sugar or honey? (HHDD14_ Trend) 15. Foods such as condiments, cocoa, coffee, tea? (HHDD15_Trend) 1999 2009 1999 2009 1999 2009 111 (25.3) 50 (11.4) 81 (18.5) 101 (23.0) 53 (12.1) 396 (90.2) 43 (9.8) 439 (100) 114 (26.0) 64 (14.6) 83 (18.9) 97 (22.1) 58 (13.2) 416 (94.8) 23 (5.2) 439 (100) 64 (14.6) 45 (10.3) 125 (28.5) 124 (28.2) 61 (13.9) 419 (95.4) 20 (4.6) 439 (100) 81 (18.5) 54 (12.3) 111 (25.3) 126 (28.7) 51 (11.6) 432 (96.4) 16 (3.6) 439 (100) 147 (33.5) 62 (14.1) 79 (18.0) 84 (19.1) 28 (6.4) 400 (91.1) 39 (8.9) 439 (100) 160 (36.4) 73 (16.6) 68 (15.5) 78 (17.8) 27 (6.2) 406 (92.5) 33 (7.5) 439 (100) 772 Table 7.5.54: HDD- Trend of Consumption of Other Foods – Descriptive In terms of sufficiency, how would u describe trends in quantity of the following foods available to the household (1990-1999 and 2000- 2009)? SCALE Decreased and not sufficient=1 Increased but not sufficient=2 Decreased but sufficient=3 Stable but sufficient=4 Increased and sufficient=5 13. Foods made from oil, fat or butter? (HHDD13_Trend) 14. Sugar or honey? (HHDD14_Trend) 15. Foods such as condiments, cocoa, coffee, tea? (HHDD15_Trend) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 2.6120 1.6565 2.7080 1.4225 2.6500 1.3537 2.6560 1.4548 2009 2.5090 1.6445 2.9000 1.3744 2.7240 1.3360 2.7150 1.4378 1999 2.7840 1.4967 3.2330 1.2349 3.1480 1.2055 3.0750 1.3055 2009 2.3880 1.4251 3.2250 1.2466 3.1180 1.2210 2.9540 1.3268 1999 2.0090 1.3220 2.4500 1.3711 2.4430 1.3426 2.3300 1.3558 2009 1.7840 1.1855 2.4250 1.3389 2.4240 1.3672 2.2550 1.3408 773 Other Foods Sufficiency Index (OFSI) 4e) Dietary Diversity Aggregate Food Sufficiency Index (DDAFS) 4f) Combined Household Dietary Diversity Index for Trend in Sufficiency of Consumption (CHHDDI2) 774 Snni*end) HHDD15_Tr+end HHDD14_Tr+end(HHDD13_TrOFSI31itt===5*3end) HHDD15_Tr+end HHDD14_Tr+end(HHDD13_TrOFSI31itt===ni15end) HHDD15_Tr+end HHDD14_Tr+end(HHDD13_TrOFSI31itt===niSnni*end) HHDD15_Tr+end HHDD14_Tr+end HHDD13_Tr+end HHDD12_Tr+ndHHDD11_Tre +rend HHDD10_T+nd HHDD9_Tre+nd HHDD8_Tre+nd HHDD7_Tre+dHHDD6_Tren +nd HHDD5_Tre+nd HHDD4_Tre+nd HHDD3_Tre+nd HHDD2_Tre+nd(HHDD1_TreDDAFS151itt===75end) HHDD15_Tr+end HHDD14_Tr+end HHDD13_Tr+end HHDD12_Tr+ndHHDD11_Tre +rend HHDD10_T+nd HHDD9_Tre+nd HHDD8_Tre+nd HHDD7_Tre+dHHDD6_Tren +nd HHDD5_Tre+nd HHDD4_Tre+nd HHDD3_Tre+nd HHDD2_Tre+nd(HHDD1_TreDDAFS151itt===ninni===41itt) OFSI+ PAPSI+ FVSI+(SFSICHHDDI2nni===41itt) OFSI+ PAPSI+ FVSI+(SFSICHHDDI2 5. HEALTH AND DISEASE Table 7.5.55: Frequency of Disease and Illnesses within the Household - Frequency Frequency (Percentage) 1. Children <6yrs old fall ill or are hospitalized? (ID_1) 2. Children between 6 and 12yrs old fall ill and are hospitalized? (ID_2) 3. HH members >12yrs old fall ill, cannot work and/or are hospital? (ID_3) 1999 2009 1999 2009 1999 Always =1 Often =2 74 (16.9) 61 (13.9) Occasionally =3 140 (31.9) Never =4 Sub-total Missing System Total 135 (30.8) 411 (93.6) 29 (6.6) 439 (100) 62 (14.1) 72 (16.4) 175 (39.9) 130 (29.6) 439 (100) 0 (0) 439 (100) 58 (13.2) 63 (14.4) 141 (32.1) 147 (33.5) 409 (93.2) 30 (6.8) 439 (100) 46 (10.5) 55 (12.5) 180 (41.0) 158 (36.0) 439 (100) 0 (0) 439 (100) 52 (11.8) 44 (10.0) 164 (37.4) 164 (37.4) 424 (96.6) 15 (3.4) 439 (100) 2009 40 (9.1) 35 (8.0) 187 (42.6) 177 (40.3) 439 (100) 0 (0) 439 (100) 775 Table 7.5.56: Frequency of Disease and Illnesses within the Household - Descriptive NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 2.8020 1.0972 2.8420 1.1152 2.9850 1.0600 2.8970 1.0858 2009 2.5260 1.1829 2.8830 0.8905 3.0150 0.9091 2.8500 1.0024 1999 2.9220 1.0397 2.9750 1.0569 3.0490 1.0281 2.9950 1.0381 2009 2.6900 1.1452 3.0670 0.8957 3.1920 0.8069 3.0250 0.9517 1999 3.0340 0.9863 2.9330 1.0186 3.1720 0.9672 3.0710 0.9895 2009 2.7930 1.0995 3.1080 0.8481 3.3600 0.7537 3.1410 0.9107 How often does each of the following occur? SCALE Always (At least once a week)=1 Often (At most ones a month)=2 Occa (At most once in 6months)=3 Never (At most once a year)=4 1. Children <6yrs old fall ill or are hospitalized? (ID_1) 2. Children between 6 and 12yrs old fall ill and are hospitalized? (ID_2) 3. HH members >12yrs old fall ill, cannot work and/or are hospital? (ID_3) Household Health and Disease Index (HHDI) 776 Snni*ID_3) +ID_2 +(ID_1HHDI31itt===12ID_3) +ID_2 +(ID_14*3ID_3) +ID_2 +(ID_1HHDI31it31itt======nini 6. MORTALITY Table 7.5.57: Frequency of Mortality cases within the Household Frequency (Percentage) Very Many (>4) =1 Many (3-4) =2 Very Few (1-2) =3 None (0) =4 Sub-total Missing System Total 4. No. of children 5. Individuals <6yrs diseased in the between ages of 7- HH? (DM_1) 57yrs? (DM_2) 6. No. of disabled or physically handicapped in the HH? (DM_3) 2009 5 (1.1) 4 (0.9) 26 (5.9) 384 (87.5) 419 (95.4) 20 (4.6) 439 (100) 1999 9 (2.1) 5 (1.1) 41 (9.3) 364 (82.9) 419 (95.4) 20 (4.6) 439 (100) 2009 7 (1.6) 7 (1.6) 42 (9.6) 368 (83.8) 424 (96.6) 15 (3.4) 439 (100) 1999 4 (0.9) 3 (0.7) 10 (2.3) 365 (83.1) 382 (87.0) 57 (13.0) 439 (100) 2009 6 (1.4) 3 (0.7) 17 (3.9) 359 (81.8) 385 (87.7) 54 (12.3) 439 (100) 1999 5 (1.1) 8 (1.8) 44 (10.0) 359 (81.8) 416 (94.8) 23 (5.2) 439 (100) 777 Table 7.5.58: Frequency of Mortality cases within the Household - Descriptive Describe the mortality and disease prevalence in the household for periods1990- 1999 and 2000-2009? SCALE Very Many (>4)=1 Many (3-4)=2 Very Few (1-2)=3 None (0)=4 4. No. of children <6yrs diseased in the HH? (DM_1) 5. Individuals between ages of 7-57yrs? (DM_2) 6. No. of disabled or physically handicapped in the HH? (DM_3) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 3.8710 0.3852 3.8170 0.5018 3.8130 0.5490 3.8290 0.4970 2009 3.9390 0.2401 3.8960 0.4055 3.8410 0.5421 3.8830 0.4409 1999 3.9140 0.3112 3.8500 0.4962 3.7540 0.6513 3.8220 0.5409 2009 3.9480 0.2597 3.8350 0.5287 3.7320 0.6278 3.8180 0.5304 1999 3.9740 0.1594 3.9170 0.4412 3.9260 0.3847 3.9360 0.3582 2009 3.9650 0.1840 3.8360 0.5671 3.8810 0.4936 3.8940 0.4525 Household Mortality Index (HHMI) 778 Snni* DM_3)+ DM_2+(DM_1HHMI31itt===12 DM_3)+ DM_2+(DM_14*3 DM_3)+ DM_2+(DM_1HHMI31it31itt======nini 7a) Combined Household Human Capital Index1 7b) Combined Household Human Capital Index2 779 6)(n Indeces Capital HumanlofNumber TotalIndeces) Capital FHuman(All CHCI161tt====ni]6[] HHMI+ HHDI+ DDAFS+ DDAFC+ MIGI+[EDULITCHCI1t=6)(n Indeces Capital HumanlofNumber TotalIndeces) Capital FHuman(All CHCI261tt====ni]6[] HHMI+ HHDI+CHHDDI2 +CHHDDI1 + MIGI+[EDULITCHCI2t= APPENDIX C: Descriptive Analysis of Physical Capital Indexes A7.3.1 Home Ownership and Room Occupancy Table 7.5.59: Home Ownership Status – Frequency Frequency (Percentage) Rent =0 Own =1 Sub-total Missing System Total 1. Which of the following best describes your home ownership status? (Hse_Rent_Own) 1999 80 (18.2) 351 (80.0) 431 (98.2) 1.8 (8) 439 (100) 2009 60 (13.7) 375 (85.4) 435 (99.1) 4 (0.9) 439 (100) Table 7.5.60: Home Ownership Status – Descriptive NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.8534 0.3552 0.8000 0.40168 0.8079 0.3949 0.8178 0.3865 2009 0.8793 0.3272 0.8250 0.3816 0.8768 0.3294 0.8633 0.3439 Describe the following? SCALE Rent=0 Own=1 1. Home ownership status? (Hse_Rent_Own_1) 780 Table 7.5.61: Average Number of Person per Room - Frequency Frequency (Percentage) More than 10=1 Between 6 to 9=2 Between 3 to 5=3 Between 1 to 2=4 Sub-total Missing System Total What is the average number of persons per room in your HH? (RoomOccup) 1999 8 (1.8) 38 (8.7) 241 (54.9) 146 (33.3) 433 (98.6) 6 (1.4) 439 (100) 2009 7 (1.6) 49 (11.2) 262 (59.7) 119 (27.1) 437 (99.5) 2 (0.5) 439 (100) Table 7.5.62: Average Number of Person per Room - Descriptive NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 M SD M SD M SD M SD 1999 0.3879 0.4894 0.3500 0.4790 0.2906 0.4552 0.3326 0.4717 2009 0.2069 0.4068 0.3083 0.4637 0.2857 0.4529 0.2711 0.4450 Describe the following? Room occupancy? SCALE Large=(>3 Occupants/room)=0 Small(<3 Occupants/room)=1 YR 4. Average number of persons per room? (RoomOccup_1) 781 Home Ownership and Room Occupancy Index (HORO) PHYSICAL CONSTRUCTION OF HOUSE Table 7.5.63: Physical Construction of the House - Frequency 1. Physical construction of home? (HseConstru_1) 1999 2009 0 (0) 1 (0.2) 395 (90.0) 39 (8.9) 436 (99.3) 4 (0.9) 439 (100) 0 (0) 0 (0) 385 (87.7) 49 (11.2) 434 (98.9) 5 (1.1) 439 (100) Frequency (Percentage) Wooden Kiosk =1 Metal Container =2 Earth/Mud Brick=3 Concrete/Cement=4 Sub-total Missing System Total 782 1*21)RoomOccup_wn_1Hse_Rent_O*1)RoomOccup_Own_1(Hse_Rent_HORO21itit21ititt====+=+=niniSn21)RoomOccup_wn_1Hse_Rent_OHORO21ititt==+=ni Describe the following? Structure of house/home? SCALE Earth/Mud House=0 Concrete/Cement House=1 2. Physical construction of home? (HseConstru_1) Table 7.5.64: Physical Construction of the House - Descriptive NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.0086 0.0929 0.1261 0.3333 0.1133 0.3177 0.0890 0.2851 2009 0.0086 0.0929 0.1500 0.3586 0.1478 0.3558 0.1116 0.3153 Table 7.5.65: Type of Roofing Material/Construction - Frequency Frequency (Percentage) Slate =1 Thatch /Raffia=2 Corrugate Metal Sheet=3 Cement/Roofing Tiles=4 Sub-total Missing System Total 1. Which of the following best describes the roof of your house? (RoofType) 1999 0 (0) 223 (50.8) 198 (45.1) 14 (3.2) 435 (99.1) 4 (0.9) 439 (100) 2009 0 (0.0) 209 (47.6) 229 (52.2) 0 (0.0) 438 (99.8) 1 (0.2) 439 (100) 783 Type of roofing material? SCALE Thatch/Palm Leaf/Raffia=0 Corrugated Metal Sheet=1 3. Type of roof? (RoofType_1) Table 7.5.66: Type of Roofing Material/Construction - Descriptive NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.1897 0.3937 0.6018 0.4917 0.5510 0.4987 0.4659 0.4994 2009 0.2328 0.4244 0.6333 0.4839 0.6207 0.4864 0.5216 0.5001 Home Construction and Roof-Type Index (HCRT) 784 1*2)RoofType_1u_1(HseConstr*)RoofType_1u_1(HseConstrHCRT21itit21ititt====+=+=niniSn2)RoofType_1u_1(HseConstrHCRT21ititt==+=ni HOUSEHOLD ENERGY AND DRINKING WATER SOURCES Table 7.5.67: Source of Energy for Lighting the House - Frequency Frequency (Percentage) Kerosene Lamp =1 Flashlight/ Touch =2 Gas Lamp =3 Electric =4 Solar Energy=5 Sub-total Missing System Total Energy/power source for lighting your HH? (LightSource) 1999 371 (84.5) 23 (5.2) 37 (8.4) 2 (0.5) 3 (0.7) 436 (99.3) 3 (0.7) 439 (100) 2009 307 (69.9) 42 (9.6) 85 (19.4) 0 (0.0) 0 (0.0) 434 (98.9) 5 (1.1) 439 (100) Table 7.5.68: Source of Energy for Lighting the House - Descriptive Describe the following? SCALE Kerosene/Wood=0 Electric/LPG/Lithium Battery=1 1. Energy/power source for lighting your HH? (LightSource_1) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.1034 0.3059 0.1500 0.3586 0.1478 0.3558 0.1367 0.3439 2009 0.1724 0.3794 0.3417 0.4763 0.3251 0.4696 0.2893 0.4540 785 Table 7.5.69: Source of Drinking Water in the House - Frequency Frequency (Percentage) Source of drinking water for your HH? (WaterSource) Gutter =1 River/Stream =2 Duggout/Open Well =3 Tap (Potable Water)=4 Potable Water (Borehole)=5 Sub-total Missing System Total 1999 3 (0.7) 165 (37.6) 10 (2.3) 27 (6.2) 231 (52.6) 436 (99.3) 3 (0.7) 439 (100) 2009 0 (0.0) 68 (15.5) 8 (1.8) 0 (0.0) 363 (82.7) 439 (100) 0 (0.0) 439 (100) Table 7.5.70: Source of Drinking Water in the House - Descriptive Describe the following? SCALE Non-portable (River, streams etc)=0 Portable(Borehole or Tap)=1 2. Source of drinking water? (WaterSource_1) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.5345 0.5010 0.6083 0.4902 0.6207 0.4864 0.5945 0.4915 2009 0.7759 0.4188 0.7833 0.4137 0.8818 0.3237 0.8269 0.3788 786 Light Energy and Water Source Index (LSWS) KITCHEN, BATHROOM AND TOILET FACILITIES Table 7.5.71: Type of Kitchen in the House - Frequency Frequency (Percentage) Type of kitchen in your HH? (KitchenType) Enclosure without roof =1 Veranda in the Household =2 Hall/Living Room/Bedroom =3 Open Space in the Compound=4 Separate but shared room in the compound =4 Separate but private room in the compound for exclusive use by the HH =5 Sub-total Missing System Total 1999 25 (5.7) 14 (3.2) 2 (0.5) 221 (50.3) 136 (31.0) 40 (9.1) 438 (99.8) 1 (0.2) 439 (100) 2009 0 (0.0) 0 (0.0) 0 (0.0) 235 (53.5) 203 (46.2) 0 (0.0) 438 (99.8) 1 (0.2) 439 (100) 787 1*2e_1)WaterSourcce_1(LightSour*e_1)WaterSourcce_1(LightSourLSWS21itit21ititt====+=+=niniSn2e_1)WaterSourcce_1(LightSourLSWS21ititt==+=ni Table 7.5.72: Type of Kitchen in the House - Descriptive NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.5259 0.5015 0.4167 0.4951 0.3990 0.4909 0.4374 0.4966 2009 0.5345 0.5010 0.4667 0.5010 0.4187 0.4946 0.4624 0.4992 Which of the following best describes the kitchen facility in the HH? SCALE Open space in the compound=0 Private enclosed space/room =1 1. Type of kitchen? (KitchenType_1) Table 7.5.73: Type of Bathroom in the House - Frequency Frequency (Percentage) Type of bathroom? (BathnFacty_1) River/Lake/Pond =1 Bathroom in another House =2 Shared/Open Bathing cubicles in the HH =3 Private Open cubicles in the HH =4 Shared Separate Facility in the HH =4 Own bathroom for exclusive use Sub-total Missing System Total 1999 4 (0.9) 34 (7.7) 110 (25.1) 62 (14.1) 103 (23.5) 125 (28.5) 438 (99.8) 1 (0.2) 439 (100) 2009 0 (0) 32 (7.3) 0 (0) 0 (0) 209 (47.6) 194 (44.2) 435 (99.1) 4 (0.9) 439 (100) 788 Table 7.5.74: Type of Bathroom in the House - Descriptive NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.5948 0.4931 0.3917 0.4902 0.3498 0.4781 0.4260 0.4951 2009 0.5948 0.4931 0.4083 0.4936 0.3744 0.4852 0.4419 0.4972 Describe the following? Structure of bathroom? SCALE Shared space outside of the HH=0 Private/exclusive use by HH =1 2. Type of bathroom? (BathnFacty_1) Table 7.5.75: Type of Toilet Facility in the House - Frequency Frequency (Percentage) No toilet facility (Done in bush) =1 Dugout pits or buckets =2 Kumasi Ventilated Improved Pit (KVIP)=3 Sub-total Missing System Total Type of toilet facility available to your HH? (ToiletFacty) 1999 78 (17.8) 301 (68.6) 59 (13.4) 438 (99.8) 1 (0.2) 439 (100) 2009 72 (16.4) 294 (67.0) 72 (16.4) 438 (99.8) 1 (0.2) 439 (100) 789 Table 7.5.76: Type of Toilet Facility in the House - Descriptive NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 M SD M SD M SD M SD Describe the following? Type of Toilet? SCALE No Toilet/Dug-out pit or Latrine/Bucket=0 Communal KVIP=1 YR 4. Type of toilet? (ToiletFacty_1a) 1999 0.0086 0.0929 0.1500 0.3586 0.1970 0.3988 0.1344 0.3415 2009 0.0086 0.0929 0.1917 0.3953 0.2365 0.4260 0.1640 0.3707 Kitchen Bathroom and Toilet Facility Index (KBTF) 790 1*3y_1a)ToiletFact +y_1 BathnFact+pe_1(KitchenTy*y_1a)ToiletFact +y_1 BathnFact+pe_1(KitchenTyKBTF31it31itt======niniSn3y_1a)ToiletFact +y_1 BathnFact+pe_1(KitchenTyKBTF31itt===ni HOUSEHOLD TRASH/GARBAGE DISPOSAL Table 7.5.77: Disposal Site for Household Liquid Waste - Frequency Frequency (Percentage) Liquid waste is disposed of onto the street or outside of HH =1 Disposed into Gutters =2 Sub-total Missing System Total Where does your HH normally dispose of liquid Waste? (LQDWastDisp) 1999 384 (87.5) 37 (8.4) 425 (96.8) 18 (4.1) 439 (100) 2009 369 (84.1) 52 (11.8) 421 (95.9) 18 (4.1) 439 (100) Table 7.5.78: Disposal Site for Household Liquid Waste - Descriptive Which one of the following best describes where the HH normally dispose of liquid waste? SCALE Thrown inside or around the HH=0 Throw inside gutters=1 1. Liquid garbage (LQDWastDisp_1) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.0517 0.2224 0.1167 0.3224 0.0837 0.2777 0.0843 0.2781 2009 0.0517 0.2224 0.1833 0.3886 0.1182 0.3237 0.1185 0.3235 791 Table 7.5.79: Disposal Site for Household Solid Waste - Frequency Frequency (Percentage) Burnt in or immediately around the HH =1 Buried in or around the HH =2 Disposed elsewhere in the community =3 Disposed in a public Dumpster= 4 Sub-total Missing System Total Where does your HH normally dispose of solid Waste? (SolidWastDisp_1) 1999 40 (9.1) 11 (2.5) 188 (42.8) 193 (44.0) 432 (98.4) 7 (1.6) 439 (100) 2009 36 (8.2) 7 (1.6) 183 (41.7) 206 (46.9) 432 (98.4) 7 (1.6) 439 (100) Table 7.5.80: Disposal Site for Household Solid Waste - Descriptive Describe your HH solid and liquid garbage disposal habits? SCALE Burnt, buried/disposed anywhere inside or around the HH=0 Public Dumpster=1 2. Solid garbage? (SolidWastDisp_1) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.9483 0.2224 0.8583 0.3502 0.8276 0.3787 0.8679 0.3390 2009 0.9483 0.22243 0.8917 0.3121 0.8473 0.3606 0.8861 0.3181 792 Liquid and Solid Waste Disposal Index (LSWD) PROXIMITY OF HOUSHOLD TO DIFFERENT SERVICES Table 7.5.81: Proximity of Household to Different Markets- Frequency Frequency (Percentag e) >3hrs 2-3hrs 1-2hrs <1hr Sub-total Missing System Total 1. Roadside Market? (RdSideMktAccess) 2. Weekly Market? (WeeklyMktAccess) 3. Regional Market? (RegnalMktAccess) 2009 39 (8.9) 35 (8.0) 99 (22.6) 264 (60.1) 437 (99.5) 2 (0.5) 439 (100) 1999 77 (17.5) 70 (15.9) 158 (36.0) 130 (29.6) 435 (99.1) 4 (0.9) 439 (100) 2009 56 (12.8) 60 (13.7) 155 (35.3) 158 (36.0) 429 (97.7) 10 (2.3) 439 (100) 1999 42 (9.6) 43 (9.8) 74 (16.9) 277 (63.1) 436 (99.3) 3 (0.7) 439 (100) 2009 33 (7.5) 22 (5.0) 65 (14.8) 317 (72.2) 437 (99.5) 2 (0.5) 439 (100) 1999 50 (11.4) 48 (10.9) 111 (25.3) 225 (51.3) 434 (98.9) 5 (1.1) 439 (100) 793 1*2isp_1)SolidWastD +sp_1(LQDWastDi*isp_1)SolidWastD +sp_1(LQDWastDiLSWD21it21itt======niniSn2isp_1)SolidWastD +sp_1(LQDWastDiLSWD21itt===ni Table 7.5.82: Proximity of Household to Different Markets - Descriptive How long does it take to travel to/access the following services from your HH? SCALE Long(>1hr to any market outlet)=0 Short(<1hr to any market outlet)=1 1. Any Market Outlet? (MarketAccess) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.6810 0.4681 0.8083 0.3953 0.8768 0.3294 0.8064 0.3956 2009 0.6983 0.4610 0.8500 0.3586 0.9113 0.2850 0.8383 0.3686 794 Table 7.5.83: Proximity of Household to Different Health Facilities - Frequency How long does it take to travel from your house to the nearest… Frequency (Percentage) 1. Regional Hospital? (AccessRegHosp) 2. District Hospital? (AccessDistHosp) 3. Rural/ Community Clinic? (AccessRurlClnc) 4. Maternity 4. Drug Store/ Home? (AccessMatHme) Pharmacy? (AccessPhmcy) >3hrs 2-3hrs 1-2hrs <1hr 1999 2009 1999 2009 1999 2009 1999 2009 1999 2009 102 (23.2) 58 59 (13.2) (13.4) 43 (9.8) 47 (10.7) 37 (8.4) 50 (11.4) 80 67 74 49 61 46 48 (18.2) (15.3) (16.9) (11.2) (13.9) (10.5) (10.9) 38 (8.7) 43 (9.8) 54 (12.3) 47 (10.7) 40 (9.1) 34 (7.7) 159 (36.2) 189 (43.1) 124 (28.2) 116 (26.4) 108 (24.6) 100 (22.8) 102 (23.2) 84 (19.1) 102 (23.2) 93 (21.2) 97 (22.1) 123 (28.0) 177 (40.3) 222 (50.6) 195 (44.4) 226 (51.5) 179 (40.8) 214 (48.7) 200 (45.6) 217 (49.4) Sub-total 438 (99.8) 437 (99.5) 434 (98.9) 430 (97.9) 411 (93.6) 409 (93.2) 379 (86.3) 379 (86.3) 403 (91.8) 384 (87.5) Missing System Total 1 (0.2) 439 (100) 2 (0.5) 439 (100) 5 (1.1) 439 (100) 9 (2.1) 439 (100) 28 (6.4) 439 (100) 30 (6.8) 439 (100) 60 60 36 55 (13.7) (13.7) (8.2) (12.5) 439 (100) 439 (100) 439 (100) 439 (100) Table 7.5.84: Proximity of Household to Different Health Facilities - Descriptive How long does it take to travel to or access the following services from your HH? SCALE Long(>1hr to any market outlet)=0 Short(<1hr to any market outlet)=1 2. Any Health Facilities? (HealthServAccess) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.2931 0.4572 0.5667 0.4976 0.6453 0.4796 0.5308 0.4996 2009 0.3793 0.4873 0.6167 0.4882 0.7340 0.4430 0.6082 0.4887 795 Table 7.5.85: Proximity of Household to any Postal Service - Frequency How long does it take to travel from your house to the nearest… Frequency (Percentage) 1. Any Postal Service? (AccessPostOffce) >3hrs 2-3hrs 1-2hrs <1hr Sub-total Missing System Total 1999 58 (13.2) 58 (13.2) 100 (22.8) 190 (43.3) 406 (92.5) 33 (7.5) 439 (100) 2009 40 (9.1) 39 (8.9) 102 (23.2) 224 (51.0) 405 (92.3) 34 (7.7) 439 (100) Table 7.5.86: Proximity of Household to any Postal Service - Descriptive How long does it take to travel to/access the following services from your HH? SCALE Long(>1hr to any market outlet)=0 Short(<1hr to any market outlet)=1 3. Any Postal Service? (AccessPostOffce_1) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.2759 0.4489 0.4667 0.5010 0.5025 0.5012 0.4328 0.4960 2009 0.3276 0.4714 0.5333 0.5010 0.6010 0.4909 0.5103 0.5005 796 Table 7.5.87: Proximity of Household to Transportation Services - Frequency Frequency (Percentage) >3hrs 2-3hrs 1-2hrs <1hr Sub-total Missing System Total How long does it take to get the following transportation services from your household? 1. Taxi Service? (TimeTaxiServ) 2. Minivan/Bus (Trotro)? (TimeTrotroServ) 1999 2009 1999 2009 70 40 61 (15.9) (9.1) (13.9) 42 (9.6) 50 50 56 49 (11.4) (11.4) (12.8) (11.2) 84 69 84 72 (19.1) (15.7) (19.1) (16.4) 231 277 233 272 (52.6) (63.1) (53.1) (62.0) 435 436 434 435 (99.1) (99.3) (98.9) (99.1) 4 (0.9) 439 (100) 3 (0.7) 439 (100) 5 (1.1) 439 (100) 4 (0.9) 439 (100) Access to Transportation Services Table 7.5.88: Proximity of Household to Transportation Services - Descriptive How long does it take to travel to/access the following services from your HH? SCALE Long(>1hr to any market outlet)=0 Short(<1hr to any market outlet)=1 4. Transportation Service? (TransportServ) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.3276 0.4714 0.6167 0.4882 0.6404 0.4811 0.5513 0.4979 2009 0.3707 0.4851 0.7083 0.4564 0.7783 0.4164 0.6515 0.4771 797 Access to Markets, Health, Post Office and Transport Facilities Index (MHPT) Cell Phone Ownership Table 7.5.89: Cell Phone Ownership Which of the following cellular phones or service do you have in your household? Frequency (Percenta ge) 1. One- Touch/Vodaph one (OneTouch) 2. MTN (MTN) 3. Zaine/Airtel? (Zaine) 4. Tigo (Tigo) 5. Kasapa/Expres so (Kasapa) 1999 2009 1999 2009 1999 2009 1999 380 101 365 45 386 279 381 (86.6) (23.0) (83.1) (10.3) (87.9) (63.6) (86.8) (16.2) 2009 1999 379 (86.3 71 22 (5.0) 331 (75.4) 40 (9.1) 387 (88.2) 11 (2.5) 138 (31.4) 21 (4.8) 349 (79.5) Sub-total 402 432 405 432 397 417 402 420 (91.6) (98.4) (92.3) (98.4) (90.4) (95.0) (91.6) (95.7) No=0 Yes=1 Missing System Total 2009 166 (37.8) 243 (55.4) 409 (93.2) 30 (6.8) ) 10 (2.3) 389 (88.6 ) 50 (11.4 ) 37 7 34 7 42 22 37 19 (8.4) (1.6) (7.7) (1.6) (9.6) (5.0) (8.4) (4.3) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 798 Snni*erv)itTransportS +Offce_1AccessPost +vAccess HealthSer+ess(MarketAccMHPT41t===1*4erv)itTransportS +Offce_1AccessPost +vAccess HealthSer+ess(MarketAccMHPT41t===ni Table 7.5.90: Cell Phone Network/Services NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.0086 0.0929 0.0417 0.2007 0.0788 0.2701 0.0501 0.2184 2009 0.7845 0.4130 0.8083 0.3953 0.7044 0.4574 0.7540 0.4312 1999 0.0517 0.2224 0.0667 0.2505 0.1281 0.3350 0.0911 0.2881 2009 0.9310 0.2545 0.8417 0.3666 0.8768 0.3294 0.8815 0.3235 1999 0.0086 0.0929 0.0500 0.2189 0.0197 0.1393 0.0251 0.1565 2009 0.1638 0.3717 0.4500 0.4996 0.3202 0.4677 0.3144 0.4648 1999 0.0086 0.0929 0.0583 0.2354 0.0640 0.2454 0.0478 0.2137 2009 0.7500 0.4349 0.8583 0.3502 0.7833 0.4131 0.7950 0.4042 1999 0.0086 0.0929 0.0333 0.1803 0.0246 0.1554 0.0228 0.1494 2009 0.3793 0.4873 0.6583 0.4763 0.5911 0.4928 0.5535 0.4977 Do you have the following telephone services in your HH? SCALE No =0 Yes=1 1. One-Touch Cellular Network/Service? (OneTouch) 2. MTN Cellular Network/Service? (MTN) 3. Zaine Cellular Network/Service? (Zaine) 4. Tigo Cellular Network/Service? (Tigo) 5. Kasapa Cellular Network/Service? (Kasapa) Household Cell-Phone Cell Phone Ownership Index (HCPO) 799 Snni* Kasapa)it+Tigo +Zaine + MTN+(OneTouchHCPO51t===1*5 Kasapa)it+Tigo +Zaine + MTN+(OneTouchHCPO41t===ni How reliable is your telephone network? SCALE Not Reliable/Unavailable =0 Reliable =1 Network Reliability (NetwkReliability) CELL PHONE RELIABILITY Table 7.5.91: Cell Phone Network Reliability NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.0431 0.2040 0.0833 0.2775 0.0394 0.1950 0.0524 0.2230 2009 0.3621 0.4827 0.7250 0.4484 0.7389 0.4403 0.6355 0.4818 Cell-phone Network Reliability Index (CNWR) = 800 Snni*ability)it(NetwkReliCNWR11t===1*1ability)it(NetwkReliCNWR11t===niability)it(NetwkReliCNWRt= BASIC HOUSEHOLD POSSESSIONS Table 7.5.92: Basic Household Possessions One - Frequency Does anyone in the HH have any of the following….? Frequency (Percentage) 1. Lantern (HHPoss_Lantern) 2. Touch/Flash Light (HHPoss_Flashlight) 3. Radio? (HHPossRadio) 4. Cell phone (HHPossCellPhne) No=0 Yes=1-7 1-12 Sub-total Missing System Total (100%) 1999 2009 1999 2009 1999 2009 1999 2009 49 (11.2) 384 40 (9.1) 394 88 48 133 65 343 172 (20.0) (10.9) (30.3) (14.8) (78.1) (39.2) 347 385 297 362 87 242 (87.5) (89.7) (79.0) (87.7) (67.7) (82.5) (19.8) (55.1) 6 (1.4) 439 (100) 0 (0) 439 (100) 0 (0) 434 3 (0.7) 438 0 (0) 433 4 (0.9) 434 0 (0) 427 5 (1.1) 435 0 (0) 414 (98.9) (99.8) (98.6) (98.9) (97.3) (99.1) (94.3) 5 (1.1) 439 (100) 1 (0.2) 439 (100) 6 (1.4) 439 (100) 5 (1.1) 439 (100) 12 (2.7) 439 (100) 4 (0.9) 439 (100) 25 (5.7) 439 (100) Table 7.5.93: Basic Household Possessions Two - Frequency Does anyone in the HH have any of the following….? Frequency (Percentage) 5. Sewing Machine (HHPossSewMchne) 6. Four legged carte (HHPossCarte) 7. Functioning Bicycle (HHPossBike) No=0 Yes=1-7 1-12 Sub-total Missing System Total 1999 352 (80.2) 76 (17.3) 4 (0.9) 432 (98.4) 7 (1.6) 439 (100) 2009 276 (62.9) 139 (31.7) () 415 (94.5) 24 (5.5) 439 (100) 1999 412 (93.8) 13 (3.0) 6 (1.3) 431 (98.2) 8 (1.8) 439 (100) 801 2009 384 (87.5) 25 (5.7) () 409 (93.2) 30 (6.8) 439 (100) 1999 311 (70.8) 117 (26.7) 5 (1.2) 433 (98.6) 6 (1.4) 439 (100) 2009 237 (54.0) 176 (40.1) () 413 (94.1) 26 (5.9) 439 (100) Table 7.5.94: Basic Household Possessions - Descriptive Does anyone in the HH have any of the following? SCALE No=0 Yes=1-12 1. Lantern? (HHPoss_Lantern_1) 2. Touch/Flash light? (HHPoss_Flashlight_1) 3. Radio? (HHPossRadio_1) 4. Cell Phone? (HHPossCellPhne_1) 5. Sewing machine? (HHPossSewMchne_1) 6. Carte/Four legged trolley? (HHPossCarte_1) 7. Bicycle? (HHPossBike_1) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.9052 0.2943 0.8500 0.3586 0.9015 0.2988 0.8884 0.3153 2009 0.9386 0.2411 0.8750 0.3321 0.8966 0.3053 0.8975 0.3037 1999 1.7759 0.6336 1.5667 0.8274 1.5074 0.8639 1.5945 0.805 2009 1.9304 0.3681 1.7000 0.7171 1.6946 0.7212 1.7540 0.6576 1999 1.9397 1.4403 2.0750 1.3912 2.1133 1.3723 2.0569 1.3944 2009 2.4159 1.1932 2.4500 1.1657 2.5567 1.0673 2.4738 1.1422 1999 0.8621 1.6519 0.9000 1.6773 0.7882 1.5950 0.8383 1.6299 2009 2.2745 1.9909 2.2667 1.9905 2.2857 1.9844 2.2050 1.9917 1999 1.0435 2.0408 0.7917 1.8329 0.9113 1.9351 0.9112 1.9324 2009 1.4904 2.2981 1.5833 2.3356 1.7241 2.3824 1.5831 2.3285 1999 0.1565 0.9606 0.4000 1.5029 0.2365 1.1703 0.2597 1.2223 2009 0.3564 1.4254 0.6000 1.8076 0.2069 1.0975 0.3417 1.3920 1999 1.7500 3.0442 1.8083 3.0769 2.1379 3.2321 1.9453 3.1393 2009 3.1569 3.5003 2.4500 3.3528 3.0345 3.4775 2.8064 3.4345 Basic Household Possessions Index (BHP) 802 7.......217]it*e_1)(HHPossBik +6*te_1)(HHPossCar +5*Mchne_1)(HHPossSew +4*lPhne_1)(HHPossCel +3*io_1)(HHPossRad +2*)ashlight_1(HHPoss_Fl +antern_1)[(HHPoss_LBHP71t+++===ni287]it*e_1)(HHPossBik +6*te_1)(HHPossCar +5*Mchne_1)(HHPossSew +4*lPhne_1)(HHPossCel +3*io_1)(HHPossRad +2*)ashlight_1(HHPoss_Fl +antern_1)[(HHPoss_LBHP71t===ni LUXURY HOUSEHOLD POSSESSIONS Table 7.5.95: Luxury Household Possessions - Frequency Frequency (Percentage) 1. TV (HHPossTV) Does anyone in the HH have any of the following….? 2. Modern Furniture (HHPossFurnture) 3. Functioning Motocycle (HHPossMotor) 4. Functioning Car (HHPossCar) 5. Functioning Tractor and Plough (HHPossTractor) No=0 Yes=8-12 Sub-total Missing System Total 1999 2009 1999 2009 1999 2009 1999 2009 1999 2009 395 (90.0) 332 (75.6) 394 (89.7) 350 (79.7) 419 (95.4) 390 (88.8) 424 (96.6) 392 (89.3) 427 (97.3) 399 (90.9) 34 (7.8) 429 (97.7) 10 (2.3) 439 (100) 76 (17.3) 408 (92.9) 31 (7.1) 439 (100) 35 (8.0) 429 (97.7) 10 (2.3) 439 (100) 56 (12.8) 406 (92.5) 33 (7.5) 439 (100) 9 (2.0) 428 (97.5) 11 (2.5) 439 (100) 15 (3.4) 405 (92.3) 34 (7.7) 439 (100) 4 (0.9) 428 (97.5) 11 (2.5) 439 (100) 13 (3.0) 405 (92.3) 34 (7.7) 439 (100) 1 (0.2) 428 (97.5) 11 (2.5) 439 (100) 1 (0.2) 400 (91.1) 39 (8.9) 439 (100) 803 Does anyone in the HH have any of the following? SCALE No=0 Yes=1-12 8. Television? (HHPossTV_1) 9. Modern Furniture? (HHPossFurnture_1) 10. Motorcycle? (HHPossMotor_1) 11. Car? (HHPossCar_1) 12. Tractor & Plough? (HHPossTractor_1) Table 7.5.96: Luxury Household Possessions - Descriptive NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.2783 1.4723 0.8667 2.4968 0.6700 2.2215 0.6196 2.1409 2009 0.4000 1.7523 1.9333 3.4391 1.6552 3.2487 1.3850 3.0303 1999 0.2348 1.4408 1.1250 2.9890 0.7537 2.4992 0.7175 2.4406 2009 0.1800 1.2664 1.4250 3.2993 1.5517 3.4081 1.1481 3.0059 1999 0.1739 1.3130 0.2500 1.5678 0.1970 1.3933 0.205 1.4187 2009 0.3000 1.7145 0.4167 2.0066 0.3448 1.8292 0.3417 1.8187 1999 0.1913 1.4443 0.0917 1.0042 0.0542 0.7721 0.1002 1.0464 2009 0.2200 1.5478 0.2750 1.7246 0.4335 2.1455 0.3257 1.8668 1999 0.0000 0.0000 0.0000 0.0000 0.0591 0.8422 0.0273 0.5727 2009 0.0000 0.0000 0.0000 0.0000 0.0591 0.8422 0.0273 0.5727 Luxury Household Possessions Index (LHP) 804 12.......812it*ctor_1)](HHPossTra +11*_1)(HHPossCar +10*or_1)(HHPossMot +9*nture_1)(HHPossFur +8*_1)[(HHPossTVLHP71t++===ni5012it*ctor_1)](HHPossTra +11*_1)(HHPossCar +10*or_1)(HHPossMot +9*nture_1)(HHPossFur +8*_1)[(HHPossTVLHP71t===ni Combine HH Physical Capital Index (CHPCI1I) Combine HH Physical Capital Index (CHPCI2) 805 10)(n Indices Capital PhysialofNumber TotalIndices) Capital Physical(All CHHFCI101tt====ni]10[] LHP+ BHP+CNWR +CNWA + MHPT+ LSWD+ KBTF+ LSWS+ HCRT+[HOROCHPCI1tt=10)(n Indices Capital PhysialofNumber TotalIndices) Capital Physical(All CHHFCI91tt====ni]9[] BHP+CNWR + HCPO+ MHPT+ LSWD+ KBTF+ LSWS+ HCRT+[HOROCHPCI2tt= APPENDIX D: Descriptive Analysis of Natural Capital Indexes 7.4.1 Crop Land Ownership Table 7.5.97: Cropland Ownership Frequency (Percentag e) None (0) Small (0.1- 2acres) Medium (2.1- 5acres) Large (>5acres) Sub-total Missing System Total 1. Total land owned by HH? (CP1) 2. Size of crop farm? (CP2) 1999 2009 1999 75 52 46 (17.1) (11.8) (10.5) 91 101 103 2009 15 (3.4) 108 3. Size of uncultivated land? (CP3) 1999 254 2009 243 4. Land farmed that does not belong to HH? (CP4) 1999 213 2009 204 (57.9) (55.4) (48.5) (46.5) 52 62 68 80 (20.7) (23.0) (23.5) (24.6) (11.8) (14.1) (15.5) (18.2) 118 135 146 167 57 63 73 71 (26.9) (30.8) (33.3) (38.0) (13.0) (14.4) (16.6) (16.2) 140 141 132 144 (31.9) (32.1) (30.1) (32.8) 424 429 427 434 31 (7.1) 394 27 (6.2) 395 40 (9.1) 394 44 (10.0) 399 (96.6) (97.7) (97.3) (98.9) (89.7) (90.0) (89.7) (90.9) 15 (3.4) 439 (100) 10 (2.3) 439 (100) 12 (2.7) 439 (100) 5 (1.1) 439 (100) 45 (10.3) 439 (100) 44 (10.0) 439 (100) 45 (10.3) 439 (100) 40 (9.1) 439 (100) 806 Table 7.5.98: Cropland Ownership NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 2.9310 1.1169 2.5750 1.1643 2.6453 1.0865 2.7016 1.1225 2009 2.7838 1.0216 2.6949 1.0822 2.9800 0.9561 2.8508 1.0144 1999 3.2069 0.8497 2.5667 1.0670 2.7094 1.0046 2.8018 1.0133 2009 3.0614 0.8232 2.8475 0.9751 3.0842 0.7713 3.0138 0.8487 1999 1.5603 0.9717 1.5833 0.9401 1.6108 0.9756 1.5900 0.9630 2009 1.5981 0.9602 1.6486 0.9691 1.7514 0.9859 1.6810 0.9740 1999 1.9655 1.1567 1.6417 0.9597 1.7143 0.9836 1.7608 1.0313 2009 1.8818 1.0982 1.7818 1.0081 1.9553 1.0643 1.8872 1.0585 Crop Land Ownership Index (CLOI) 807 Describe the following? SCALE None (0)=1 Small (0.1- 2acres)=2 Medium (2.1- 5acres)=3 Large (>5acres)=4 1. Total land owned by HH?(CP1) 2. Size of crop farm? (CP2) 3. Size of uncultivated land? (CP3) 4. Land farmed that does not belong to HH? (CP4) 4*4CP4Q) + CP3Q + CP2Q + (CP1Q*CP4Q) + CP3Q + CP2Q + (CP1QCLOI4141titit======niniSn16CP4Q) + CP3Q + CP2Q + (CP1QCLOI41tit===ni Table 7.5.99: Crop Production 1. Frequency (Percentage) Maize/Corn (CPY_1) 2. Cassava (CPY_2) 3. Plantain (CPY_3) 4. Yam (CPY_4) 5. Cocoyam (CPY_5) 6. Tomato (CPY_6) 1999 2009 1999 2009 1999 2009 1999 2009 1999 2009 1999 2009 18 25 22 27 65 71 60 67 115 133 228 230 (4.1) (5.7) (5.0) (6.2) (14.8) (16.2) (13.7) (15.3) (26.2) (30.3) (51.9) (52.4) 404 401 394 397 348 351 352 353 278 272 176 183 (92.0) (91.3) (89.7) (90.4) (79.3) (80.0) (80.2) (80.4) (63.3) (62.0) (40.1) (41.7) No (0) Yes (1) Sub-total 422 426 416 424 413 422 412 420 393 405 404 413 (96.1) (97.0) (94.8) (96.6) (94.1) (96.1) (93.8) (95.7) (89.5) (92.3) (92.0) (94.1) Missing System 17 13 23 15 26 17 27 19 46 34 35 26 (3.9) (3.0) (5.2) (3.4) (5.9) (3.9) (6.2) (4.3) (10.5) (7.7) (8.0) (5.9) Total 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 808 Did you produce any of the following crops? SCALE No=0 Yes=1 1. Maize (CPY_1) 2. Cassava (CPY_2) 3. Plantain (CPY_3) 4. Yam (CPY_4) 5. Cocoyam (CPY_5) 6. Tomato (CPY_6) Table 7.5.100: Crop Production NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.9655 0.1833 0.9083 0.2898 0.9015 0.2988 0.9203 0.2712 2009 0.9655 0.1833 0.8833 0.3224 0.9015 0.2988 0.9134 0.2815 1999 0.9483 0.2224 0.8750 0.3321 0.8818 0.3237 0.8975 0.3037 2009 0.9569 0.2040 0.8833 0.3224 0.8867 0.3177 0.9043 0.2945 1999 0.6983 0.4610 0.8500 0.3586 0.8128 0.3910 0.7927 0.4058 2009 0.6983 0.4610 0.8667 0.3414 0.8177 0.3870 0.7995 0.4008 1999 0.8793 0.3272 0.7583 0.4299 0.7833 0.4131 0.8018 0.3991 2009 0.8793 0.3272 0.7667 0.4247 0.7833 0.4131 0.8041 0.3973 1999 0.5690 0.4974 0.6750 0.4703 0.6453 0.4796 0.6333 0.4825 2009 0.5776 0.4961 0.6333 0.4839 0.6355 0.4825 0.6196 0.4860 1999 0.3966 0.4913 0.3500 0.4790 0.4335 0.4968 0.4009 0.4906 2009 0.3966 0.4913 0.4000 0.4920 0.4384 0.4974 0.4169 0.4936 Household Crop Diversity Index (HHCDI) 809 1*6CPY_6) + CPY_5 + CPY_4 + CPY_3 + CPY_2 + (CPY_1*CPY_6) + CPY_5 + CPY_4 + CPY_3 + CPY_2 + (CPY_1HHCDI6161titit======niniSn6CPY_6) + CPY_5 + CPY_4 + CPY_3 + CPY_2 + (CPY_1HHCDI61tit===ni Table 7.5.101: Trends in Quantity of Specific Crops Produced 1. Maize/Corn (CPY_A1) 2. Cassava (CPY_A1) 3. Plantain (CPY_A3) 4. Yam (CPY_A4) 5. Cocoyam (CPY_A5) 6. Tomato (CPY_A6) 1999 2009 1999 2009 1999 2009 1999 2009 1999 2009 1999 2009 51 (11.6) 111 (25.3) 52 77 56 75 71 78 60 63 (11.8) (17.5) (12.8) (17.1) (16.2) (17.8) (13.7) (14.4) 94 (21.4) 91 (20.7) 48 48 (10.9) (10.9) 43 (9.8) 46 44 47 (10.5) (10.0) (10.7) 43 (9.8) 50 50 47 (11.4) (11.4) (10.7) 33 (7.5) 36 (8.2) 43 (9.8) 117 (26.7) 57 90 61 94 53 84 72 78 51 54 (13.0) (20.5) (13.9) (21.4) (12.1) (19.1) (16.4) (17.8) (11.6) (12.3) 126 (28.7) 78 (17.8) 110 (25.1) 92 (21.0) 130 (29.6) 98 (22.3) 110 (25.1) 96 88 73 52 63 (21.9) (20.0) (16.6) (11.8) (14.4) 158 (36.0) 76 (17.3) 158 (36.0) 117 (26.7) 94 76 (21.4) (17.3) 106 (24.1) 81 54 53 (18.5) (12.3) (12.1) 29 (6.6) 24 (5.5) Frequency (Percentage ) Decreased and not sufficient (1) Increased but not sufficient (2) Decreased but sufficient( 3) Stable but sufficient (4) Increased and sufficient (5) Sub-total 426 (97.0) 430 (97.9) 420 (95.7) 422 (96.1) 385 (87.7) 390 (88.8) 383 (87.2) 389 (88.6) 324 (73.8) 314 (71.5) 259 (59.0) 268 (61.0) Missing System 13 (3.0) 9 19 (2.1) (4.3) 17 (3.9) 54 49 56 50 (12.3) (11.2) (12.8) (11.4) 115 (26.2) 125 (28.5) 180 (41.0) 171 (39.0) Total 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 810 Table 7.5.102: Trends in Quantity of Specific Crops Produced NSEMSAW N=116 NON-MTS in YAYA N=120 MTS in YAYA N=203 POOLED N=439 M M SD M M SD SD SD YR 1999 3.7759 1.4391 3.4333 1.4245 3.6108 1.4321 3.6059 1.4341 2009 2.5351 1.4028 2.9483 1.4377 3.0950 1.3949 2.9070 1.4243 1999 3.6552 1.4392 3.3917 1.4968 3.5813 1.4546 3.5490 1.4624 2009 2.9115 1.4794 3.1681 1.4875 3.5969 1.3343 3.2986 1.4428 1999 2.7845 1.5649 3.1417 1.4968 3.3054 1.4335 3.1230 1.4984 2009 2.4725 1.3278 3.2273 1.3725 3.4021 1.3117 3.1359 1.3809 1999 3.2414 1.5358 2.9833 1.5930 2.9951 1.5811 3.0569 1.5729 2009 2.6916 1.3696 3.0700 1.3945 3.4286 1.3838 3.1337 1.4134 1999 2.2845 1.4314 2.8167 1.5228 2.5123 1.4602 2.5353 1.4799 2009 2.5270 1.2630 3.0761 1.4842 3.2297 1.2836 3.0191 1.3658 1999 2.0086 1.3862 1.8167 1.3028 1.9458 1.3468 1.9271 1.3444 2009 2.6349 1.3113 2.4500 1.4399 2.6800 1.4005 2.6007 1.3906 Crop Production Trend Index (CPTI) 811 Describe the trends in crop production between 1990-1999 and 2000- 2009? SCALE 1=Decreased and not sufficient, 2=Increased but not sufficient, 3=Decreased but sufficient, 4=Stable but sufficient, 5=Increased and sufficient 1. Maize/Corn (CPY_A1) 2. Cassava (CPY_A2) 3. Plantain (CPY_A3) 4. Yam (CPY_A4) 5. Cocoyam (CPY_A5) 6. Tomato (CPY_A6) Snni*CPY_A6Q) + CPY_A5Q + CPY_A4Q + CPY_A3Q + CPY_A2Q + (CPY_A1QCPTI61tit===5*6CPY_A6Q) + CPY_A5Q + CPY_A4Q + CPY_A3Q + CPY_A2Q + (CPY_A1QCPTI61tit===ni30CPY_A6Q) + CPY_A5Q + CPY_A4Q + CPY_A3Q + CPY_A2Q + (CPY_A1QCPTI61tit===ni Table 7.5.103: Subsistence Oriented Crop Production Frequency (Percentage) 1. Usually produced for HH consumption? (CLU_1) 3. Usually consume in HH with very little left for sale? (CLU_3) Strongly Disagree=1 Disagree=2 Agree=3 Strongly Agree=4 Sub-total Missing System Total 1999 51 (11.6) 170 (38.7) 158 (36.0) 53 (12.1) 432 (98.4) 7 (1.6) 439 (100) 2009 58 (13.2) 150 (34.2) 185 (42.1) 43 (9.8) 436 (99.3) 3 (0.7) 439 (100) 1999 75 (17.1) 171 (39.0) 161 (36.7) 24 (5.5) 431 (98.2) 8 (1.8) 439 (100) 2009 79 (18.0) 161 (36.7) 148 (33.7) 30 (6.8) 418 (95.2) 21 (4.8) 439 (100) 812 Table 7.5.104: Subsistence Oriented Crop Production NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 2.6957 0.7513 2.3729 0.8351 2.4472 0.9079 2.4931 0.8562 2009 2.7043 0.6066 2.3475 0.8611 2.4483 0.9287 2.4885 0.8455 1999 2.4957 0.5360 2.2966 0.8604 2.2121 0.9154 2.3109 0.8217 2009 2.6204 0.5590 2.2957 0.9078 2.1436 0.9194 2.3086 0.8583 Describe how crop harvests are used to support the HH? SCALE Strongly Disagree=1 Disagree=2 Agree=3 Strongly Agree=4 1. Usually produced for HH consumption? (CLU_1) 3. Usually consume in HH with very little left for sale?(CLU_3) Subsistence Oriented Crop Production Index (SOCPI) 813 4*2CLU_3Q) +(CLU_1Q*CLU_3Q) +(CLU_1QSOCPI2121titit======niniSn8CLU_3Q) +(CLU_1QSOCPI21tit===ni Table 7.5.105: Market Oriented Crop Production Frequency (Percentage) 2. Usually produced for 4. Usually have enough left markets? (CLU_2) for sale or barter? (CLU_4) Strongly Disagree=1 Disagree=2 Agree=3 Strongly Agree=4 Sub-total Missing System Total 1999 47 (10.7) 207 (47.2) 128 (29.2) 47 (10.7) 429 (97.7) 10 (2.3) 439 (100) 2009 49 (11.2) 206 (46.9) 124 (28.2) 53 (12.1) 432 (98.4) 7 (1.6) 439 (100) 1999 28 (6.4) 151 (34.4) 193 (44.0) 52 (11.8) (96.6) 15 (3.4) 439 (100) 2009 35 (8.0) 152 (34.6) 179 (40.8) 58 (13.2) 424 (96.6) 15 (3.4) 439 (100) 814 Describe how crop harvests are used to support the HH? SCALE Strongly Disagree=1 Disagree=2 Agree=3 Strongly Agree=4 2. Usually produced for markets?(CLU_2 ) 4. Usually have enough left for sale or barter?(CLU_4) Table 7.5.106: Market Oriented Crop Production NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR 199 9 200 9 199 9 200 9 M 2.566 4 2.451 3 2.619 5 2.580 4 SD 0.705 5 0.640 8 0.617 0 0.652 6 M 2.254 2 SD 0.797 3 M 2.409 1 SD 0.889 5 M 2.407 9 SD 0.825 4 2.305 0.832 2.467 0.948 2.419 0.847 1 1 7 8 0 0 2.551 0.848 2.692 0.823 2.634 0.781 7 0 3 4 4 5 2.577 0.886 2.653 0.872 2.613 0.823 6 1 1 4 2 1 Market Oriented Crop Production Index (MOCPI) 815 4*2CLU_4Q) +(CLU_2Q*CLU_4Q) +(CLU_2QMOCPI2121titit======niniSn8CLU_4Q) +(CLU_2QMOCPI21tit===ni 5a. Do you have any of the following livestock? Frequency (Percentage ) No (0) Yes (1) Sub-total Missing System Total Table 7.5.107: Livestock Production (Ownership) 1. Poultry (Chicken, Guinea fowls, Ducks, Turkey) (Poultry) 2. Rabbits and Guinea pigs, Grass-cutter? (Rab_Gpig_Gcutter) 3. Goats or Sheep? (Gt_Shp) 4. Pigs? (Pig) 5. Cattle? ( Cow) 1999 2009 1999 2009 1999 2009 1999 2009 1999 2009 91 105 363 372 139 157 316 330 369 388 (20.7) (23.9) (82.7) (84.7) (31.7) (35.8) (72.0) (75.2) (84.1) (88.3) 302 302 39 48 261 257 84 83 24 22 (68.8) (68.8) (8.8) (11.0) (59.4) (58.5) (19.2) (18.9) (5.5) (5.0) 393 407 402 420 400 414 400 413 393 410 (89.5) (92.7) (91.6) (95.7) (91.1) (94.3) (91.1) (94.1) (89.5) (93.4) 46 32 37 19 39 25 39 26 46 29 (10.5) (7.3) (8.4) (4.3) (8.9) (5.7) (8.9) (5.9) (10.5) (6.6) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 816 Table 7.5.108: Livestock Production (Ownership) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.8276 0.37938 0.6500 0.4790 0.6650 0.4732 0.7039 0.4571 2009 0.7739 0.4201 0.5798 0.4957 0.7083 0.4557 0.6901 0.4630 1999 0.0431 0.20397 0.0250 0.1568 0.0345 0.1829 0.0342 0.1819 2009 0.0522 0.2234 0.0000 0.0000 0.0695 0.2550 0.0460 0.2098 1999 0.5431 0.5003 0.4750 0.5015 0.4631 0.4999 0.4875 0.5004 2009 0.5130 0.5020 0.5169 0.5018 0.6378 0.4819 0.5711 0.4955 1999 0.1897 0.39373 0.0583 0.2354 0.1232 0.3294 0.123 0.3288 2009 0.1930 0.3964 0.0526 0.2243 0.1505 0.3586 0.1353 0.3424 1999 0.0517 0.22243 0.0000 0.0000 0.0148 0.1210 0.0205 0.1419 2009 0.0536 0.2262 0.0088 0.0937 0.0108 0.1034 0.0218 0.1464 HH Livestock Diversity Index 1 (HHLDI1) and 2 (HHLDI2) 817 5a. Do you have any of the following livestock? SCALE No=0 Yes=1 1. Poultry (Chicken, Guinea fowls, Ducks, Turkey) (LPP_1) 2. Rabbits and Guinea pigs, Grass- cutter? (LPP_2) 3. Goats or Sheep? (LPP_3) 4. Pigs? (LPP_4) 5.Cattle? (LPP_5) 1*5 LPP_5Q) LPP_4Q LPP_3Q LPP_2Q (LPP_1Q* LPP_5Q) LPP_4Q LPP_3Q LPP_2Q (LPP_1QHHLDI15151titit====++++=++++=niniSn5 LPP_5Q) LPP_4Q LPP_3Q LPP_2Q (LPP_1QHHLDI151tit==++++=ni3) LPP_4Q LPP_3Q (LPP_1QHHLDI231tit==++=ni Table 7.5.109: Livestock Production (Quantity) Frequency (Percentage) 1. Poultry (Chicken, Guinea fowls, Ducks, Turkey) (LPP_1) 2. Rabbits and Guinea pigs, Grass- cutter? (LPP_2) 3. Goats or Sheep? (LPP_3) 4. Pigs? (LPP_4) 5. Cattle? (LPP_5) 1999 2009 1999 2009 1999 2009 1999 2009 1999 2009 None (0)=1 114 (26.0) 132 (30.1) 396 (90.2) 394 (89.7) 207 (47.2) 184 (41.9) 357 (81.3) 358 (81.5) 404 (92.0) 403 (91.8) Small (1- 10)=2 99 (22.6) 175 (39.9) 9 (2.1) 15 (3.4) 111 (25.3) 186 (42.4) 37 46 7 9 (8.4) (10.5) (1.6) (2.1) Few (11-20)=3 94 70 5 3 64 48 15 (21.4) (15.9) (1.1) (0.7) (14.6) (10.9) (3.4) 8 (1.8) 2 (0.5) Large (>20)=4 116 (26.4) 49 1 1 (11.2) (0.2) (0.2) 39 (8.9) 11 (2.5) 2 (0.5) 2 (0.5) 0 (0) 0 (0) 0 (0) Sub-total Missing System Total 423 (96.4) 426 (97.0) 411 (93.6) 413 (94.1) 421 (95.9) 429 (97.7) 411 (93.6) 414 (94.3) 413 (94.1) 412 (93.8) 16 (3.6) 13 (3.0) 28 (6.4) 26 (5.9) 18 (4.1) 10 (2.3) 28 (6.4) 25 (5.7) 26 (5.9) 27 (6.2) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 818 Table 7.5.110: Livestock Production (Quantity) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 M SD M SD M SD M SD 1999 2.8017 1.1129 2.3167 1.1881 2.3202 1.1565 2.4465 1.1709 5e. How many of the following livestock do you have? SCALE None (0)=1 Small (1-10)=2 Few (11-20)=3 Large (>20)=4 YR 1. Poultry (Chicken, Guinea fowls, Ducks, Turkey etc (LP_1) 2. Rabbits, Guinea pigs, Grass- cutter? (LP_2) 3. Goats or Sheep? (LP_3) 2009 2.1478 0.9007 1.9076 0.9742 2.1563 0.9850 2.0845 0.9640 1999 1.0690 0.3420 1.0333 0.2220 1.0493 0.2944 1.0501 0.2902 2009 1.0522 0.2234 1.0000 0.0000 1.0963 0.3898 1.0581 0.2898 1999 2.0000 1.0955 1.7250 0.9255 1.7537 0.9536 1.8109 0.99 2009 1.6696 0.7577 1.5932 0.6303 1.8571 0.8102 1.7343 0.7577 4. Pigs (LP_4) 5.Cattle (LP_5) 1999 1.2845 0.6568 1.0750 0.3218 1.1527 0.4352 1.1663 0.4846 2009 1.2018 0.4245 1.0526 0.2243 1.2097 0.5542 1.1643 0.4532 1999 1.0690 0.3155 1.0000 0.0000 1.0148 0.1210 1.0251 0.1834 2009 1.0536 0.2262 1.0088 0.0937 1.0108 0.1034 1.0218 0.1464 HH Livestock Population Index 1 and 2 (HHLPI1) (HHLPI2) 819 4*5 LP_5Q) LP_4Q LP_3Q LP_2Q (LP_1Q* LP_5Q) LP_4Q LP_3Q LP_2Q (LP_1QHHLPI15151titit====++++=++++=niniSn20 LP_5Q) LP_4Q LP_3Q LP_2Q (LP_1QHHLPI151tit==++++=ni Table 7.5.111: Trend in Livestock Production (Quantity) 1. Poultry (Chicken, Guinea fowls, Ducks, Turkey etc? (LP_B1) 1999 2009 1999 2009 1999 2009 1999 2009 1999 2009 2. Rabbits and Guinea pigs, Grass- cutter? (LP_B2) 4. Pigs? (LP_B4) 5. Cattle? (LP_B5) 3. Goats or Sheep? (LP_B3) 94 (21.4) 168 (38.3) 123 (28.0) 121 (27.6) 88 (20.0) 111 (25.3) 129 (29.4) 133 (30.3) 121 (27.6) 120 (27.3) 43 40 7 6 45 46 9 15 1 4 (9.8) (9.1) (1.6) (1.4) (10.3) (10.5) (2.1) (3.4) (0.2) (0.9) 53 79 11 16 48 80 15 18 7 6 (12.1) (18.0) (2.5) (3.6) (10.9) (18.2) (3.4) (4.1) (1.6) (1.4) 80 37 10 11 72 45 17 10 4 1 (18.2) (8.4) (2.3) (2.5) (16.4) (10.3) (3.9) (2.3) (0.9) (0.2) Frequency (Percentage) Decreased and not sufficient (1) Increased but not sufficient (2) Decreased but sufficient(3) Stable but sufficient (4) Increased and sufficient (5) 100 (22.8) 33 6 3 52 28 8 8 (7.5) (1.4) (0.7) (11.8) (6.4) (1.8) (1.8) 0 (0) 0 (0) Sub-total Missing System Total 370 (84.3) 357 (81.3) 157 (35.8) 157 (35.8) 305 (69.5) 310 (70.6) 178 (40.5) 184 (41.9) 133 (30.3) 131 (29.8) 69 82 (15.7) (18.7) 282 (64.2) 282 (64.2) 134 (30.5) 129 (29.4) 261 (59.5) 255 (58.1) 306 (69.7) 308 (70.2) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 820 12) LP_4Q LP_3Q (LP_1Q4*3) LP_4Q LP_3Q (LP_1QHHLPI23131titit====++=++=nini Table 7.5.112: Trend in Livestock Production (Quantity) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR 1999 3.5243 1.5393 2.7157 1.5050 3.1455 1.5391 3.1324 1.5552 SD SD M SD M M SD M 2009 2.0000 1.1773 2.1170 1.4283 2.4534 1.4316 2.2353 1.3742 1999 1.8276 1.2837 1.2553 0.8715 1.5595 1.1338 1.5188 1.1044 2009 1.4444 0.8473 1.4000 1.0313 1.6235 1.1231 1.5287 1.0535 1999 3.2055 1.5089 2.6024 1.4562 2.8067 1.4687 2.8464 1.4863 2009 2.2933 1.1943 2.0779 1.1895 2.7278 1.4261 2.4613 1.3427 1999 2.3243 1.5644 1.3333 0.9092 1.6022 1.1242 1.674 1.2197 2009 1.8636 1.2500 1.3673 0.9507 1.6264 1.1513 1.6141 1.1347 1999 1.5833 1.1001 1.0698 0.4575 1.1429 0.5188 1.1971 0.6624 2009 1.2800 0.7371 1.0000 0.0000 1.1690 0.5342 1.1450 0.5132 Livestock Production Trend Index 1 and 2 (LPTI), (LPT2) 821 3f. Describe the trends in livestock production between 1990-1999 and 2000- 2009? SCALE Decreased and not sufficient=1 Increased but not sufficient=2 Decreased but sufficient=3 Stable but sufficient=4 Increased and sufficient=5 1. Poultry (Chicken, Guinea fowls, Ducks, Turkey etc?(LP_B1) 2. Rabbits and Guinea pigs, Grass- cutter? (LP_B2) 3. Goats or Sheep? (LP_B3) 4. Pigs (LP_B4) 5.Cattle (LP_B5) 5*5 LP_B5Q)+ LP_B4Q+ LP_B3Q+ LP_B2Q+(LP_B1Q* LP_B5Q)+ LP_B4Q+ LP_B3Q+ LP_B2Q+(LP_B1QLPTI5151titit======niniSn25 LP_B5Q)+ LP_B4Q+ LP_B3Q+ LP_B2Q+(LP_B1QLPTI51tit===ni Table 7.5.113: Subsistent Oriented Livestock Production Frequency (Percentage) 1. Usually produced for HH consumption? (CLU_5) 3. Usually consume in HH with very little left for sale? (CLU_7) 1999 110 (25.1) 155 (35.3) 126 (28.7) 19 (4.3) 410 (93.4) 29 (6.6) 439 (100) 2009 104 (23.7) 149 (33.9) 145 (33.0) 15 (3.4) 413 (94.1) 26 (5.9) 439 (100) 1999 121 (27.6) 201 (45.8) 75 (17.1) 16 (3.6) 413 (94.1) 26 (5.9) 439 (100) 2009 122 (27.8) 198 (45.1) 76 (17.3) 17 (3.9) 413 (94.1) 26 (5.9) 439 (100) Strongly Disagree=1 Disagree=2 Agree=3 Strongly Agree=4 Sub-total Missing System Total 822 Table 7.5.114: Subsistent Oriented Livestock Production NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 2.2703 0.7621 2.1273 0.8894 2.0529 0.8976 2.1317 0.8634 2009 2.3393 0.7775 2.1802 0.8442 2.0684 0.8792 2.1719 0.8488 1999 2.0631 0.7296 2.0268 0.8105 1.8737 0.8132 1.9661 0.7937 2009 2.1182 0.7630 2.0625 0.8196 1.8325 0.7968 1.9709 0.8030 Describe how livestock are used to support the HH? SCALE Strongly Disagree=1 Disagree=2 Agree=3 Strongly Agree=4 1. Usually produced for HH consumption? (CLU_5) 3. Usually consume in HH with very little left for sale?(CLU_7) Subsistence Oriented Livestock Production Index (SOLPI) 823 4*2CLU_7Q) +(CLU_5Q*CLU_7Q) +(CLU_5QSOLPI2121titit======niniSn8CLU_7Q) +(CLU_5QSOLPI21tit===ni Table 7.5.115: Subsistent Oriented Livestock Production Frequency (Percentage) 1. Usually produced for markets? (CLU_6) 3. Usually have enough left for sale or butter?(CLU_8) 1999 109 (24.8) 136 (31.0) 142 (32.3) 25 (5.7) 412 (93.8) 27 (6.2) 439 (100) 2009 110 (25.1) 133 (30.3) 142 (32.3) 26 (5.9) 411 (93.6) 28 (6.4) 439 (100) 1999 88 (20.0) 131 (29.8) 161 (36.7) 32 (7.3) 412 (93.8) 27 (6.2) 439 (100) 2009 96 (21.9) 142 (32.3) 144 (32.8) 30 (6.8) 412 (93.8) 27 (6.2) 439 (100) Strongly Disagree=1 Disagree=2 Agree=3 Strongly Agree=4 Sub-total Missing System Total 824 Table 7.5.116: Subsistent Oriented Livestock Production NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 2.4775 0.8404 2.0636 0.8703 2.1204 0.9240 2.2015 0.9019 2009 2.4091 0.8705 2.0545 0.8332 2.1728 0.9550 2.2044 0.9091 2.3578 1999 0.898 2009 2.2477 0.8295 2.1696 0.8993 2.3246 0.9344 2.2621 0.8984 0.8663 2.2143 0.9146 2.3874 0.9042 2.3325 Market Oriented Livestock Production Index (MOLPI) 825 Describe how livestock are used to support the HH? SCALE Strongly Disagree=1 Disagree=2 Agree=3 Strongly Agree=4 2. Usually produced for markets? (CLU_6) 4. Usually have enough left for sale or butter?(CLU_8) 4*2CLU_8Q) +(CLU_6Q*CLU_8Q) +(CLU_6QMOLPI2121titit======niniSn8CLU_8Q) +(CLU_6QMOLPI21tit===ni Table 7.5.117: Soil Fertility and Fertilizer Application Frequency (Percentage) No (0) Yes (1) Sub-total Missing System Total 1. Apply fertilizer? (FertzeUse1) 1999 2009 311 297 (70.8) (67.7) 84 106 (19.1) (22.8) 395 403 (90.0) (91.8) 44 (10.0) 36 (8.2) 439 (100) 439 (100) Table 7.5.118: Soil Fertility and Fertilizer Application NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 M SD YR 1999 0.1379 0.3463 0.1667 0.3742 0.2365 0.4260 0.1913 0.3938 SD SD M M SD M 2009 0.1638 0.3717 0.2333 0.4247 0.2906 0.4552 0.2415 0.4285 826 Did you do any of the following? SCALE No=0 Yes=1 1. Apply fertilizer? (FertzeUse1) Table 7.5.119: Trend in Fertilizer Application 1. Perceived soil fertility on your farm? (SoilFert) 2. Trend in quantity and frequency of fertilizer use? (FertzerUse2) 1999 2009 1999 2009 57 140 100 96 (13.0) (31.9) (22.8) (21.9) 29 27 14 17 (6.6) (6.2) (3.2) (3.9) Frequency (Percentage) Decreased and not sufficient (1) Increased but not sufficient (2) Decreased but sufficient(3) 62 86 12 11 (14.1) (19.6) (2.7) (2.5) Stable but sufficient (4) 95 63 16 24 (21.6) (14.4) (3.6) (5.5) Increased and sufficient (5) 105 32 8 14 (23.9) (7.3) (1.8) (3.2) Sub-total Missing System Total 348 348 150 162 (79.3) (79.3) (34.2) (36.9) 91 91 289 277 (20.7) (20.7) (65.8) (63.1) 439 (100) 439 (100) 439 (100) 439 (100) 827 Table 7.5.120: Trend in Fertilizer Application NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 3.2198 1.7049 3.4624 1.4785 3.5602 1.2331 3.4457 1.4368 2009 1.9032 1.2517 2.6092 1.3755 2.7381 1.4152 2.4828 1.4049 1999 1.8750 1.2619 1.5208 1.3045 1.8072 1.2922 1.7290 1.2909 2009 1.7407 1.1298 1.9800 1.4213 2.1529 1.5158 2.0309 1.4292 Soil and Fertilizer Application Index (SFAI) 828 How would you describe the following? SCALE Decreased and not sufficient=1, Increased but not sufficient=2 Decreased but sufficient=3, Stable but sufficient=4, Increased and sufficient=5 1. Perceived soil fertility on your farm? (SoilFert) 2. Trend in quantity and frequency of fertilizer use? (FertzerUse2) )5*2()1*1(e2Q) FertzerUsSoilFertQ 1(FertzeUse)*()*(e2Q) FertzerUsSoilFertQ 1(FertzeUseSFAI31221131titit+++=+++=====niniSnSn321 =n where21=+=+nn11e2Q) FertzerUsSoilFertQ 1(FertzeUseSFAI31tit==++=ni 11. TRENDS IN NTFP HARVEST AND MARKET VALUE Table 7.5.121: Harvest of NTFPs Frequency (Percentage) 1. Did anyone in your HH collect NTFPs from the forest? reserve? (NTFP_1A) No (0) Yes (1) Sub-total Missing System Total 2009 44 (10) 321 (73.1) 365 (83.1) 74 (16.9) 439 (100) 1999 24 (5.5) 297 (67.7) 321 (73.1) 118 (26.9) 439 (100) Table 7.5.122: Harvest of NTFPs NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 M YR 1999 0.9828 0.13073 0.9083 0.2898 0.9458 0.2270 0.9453 0.2276 SD SD M SD M M SD 2009 0.8916 0.3116 0.9310 0.2545 0.9833 0.5795 0.9271 0.3931 829 Did you do any of the following? SCALE No=0 Yes=1 1. Collect NTFPs? (NTFP_1A) Table 7.5.123: Trends in NTFP Harvest Frequency (Percentage) Decreased and not sufficient (1) Increased but not sufficient (2) Decreased but sufficient(3) 1. Trends in the area of key NTFPs? (NTFP_1B) 2. Trend in the value of key NTFPs collected? (NTFP_1C) 3. Trend in the value of firewood collected? (NTFP_1D) 1999 2009 1999 2009 1999 2009 43 243 36 192 19 111 (9.8) (55.4) (8.2) (43.7) (4.3) (25.3) 54 46 64 70 41 46 (12.3) (10.5) (14.6) (15.9) (9.3) (10.5) 53 96 72 79 58 101 (12.1) (21.9) (16.4) (18.0) (13.2) (23.0) Stable but sufficient (4) 121 22 125 45 149 90 (27.6) (5.0) (28.5) (10.3) (33.9) (20.5) Increased and sufficient (5) 149 12 124 35 154 47 (33.9) (2.7) (28.2) (8.0) (35.1) (10.7) Sub-total Missing System Total 420 419 421 421 421 395 (95.7) (95.4) (95.9) (95.9) (95.9) (90.0) 19 20 18 18 18 44 (4.3) (4.6) (4.1) (4.1) (4.1) (10.0) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 830 Table 7.5.124: Trends in NTFP Harvest NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 M M SD YR 1999 4.0000 1.2047 3.6053 1.4181 3.5000 1.3498 3.6643 1.3447 SD SD M SD M 2009 1.6460 0.9903 1.8421 1.1177 1.9531 1.1814 1.8401 1.1199 1999 3.8158 1.1941 3.4741 1.3218 3.4660 1.3049 3.5629 1.2869 2009 1.9391 1.1868 2.1842 1.3141 2.3542 1.4141 2.1948 1.3363 1999 4.0261 1.0717 3.8870 1.1605 3.8272 1.1592 3.8979 1.1366 2009 2.3070 1.2700 2.9266 1.4318 3.0174 1.3441 2.7873 1.3798 NTFP Harvests and Market Trend (NHMTI) 831 How would you describe the following? SCALE Decreased and not sufficient=1, Increased but not sufficient=2 Decreased but sufficient=3, Stable but sufficient=4, Increased and sufficient=5 1. Trends in the area of key NTFPs? (NTFP_1B) 2. Trend in the value of key NTFPs collected? (NTFP_1C) 3. Trend in the value of firewood collected? (NTFP_1D) )5*3()1*1(NTFP_1D) NTFP_1C NTFP_1B (NTFP_1A4)*()*(NTFP_1D) NTFP_1C NTFP_1B (NTFP_1ANMHT31221141titit++++=++++=====niniSnSn431 =n where21=+=+nn16NTFP_1D) NTFP_1C NTFP_1B (NTFP_1ANMHT41tit==+++=ni 12. TIME SPENT HARVESTING NTFP Table 7.5.125: Time Spent Harvesting NTFPs Frequency (Percentage) >20hrs=1 16-20hrs= 1. Mushroom s? (NTFP_2Aa) 2. Firewood? (NTFP_2Bb) 3. Snails? (NTFP_2Cc) 4. Glasscutters and other bush- meat?(NTFP_2Dd) 1999 2009 1999 2009 1999 2009 1999 2009 28 143 8 64 30 133 30 157 (6.4) (32.6) (1.8) (14.6) (6.8) (30.3) (6.8) (35.8) 22 58 18 47 30 78 30 69 (5.0) (13.2) (4.1) (10.7) (6.8) (17.8) (6.8) (15.7) 10-15hrs=3 43 134 45 120 65 126 91 133 (9.8) (30.5) (10.3) (27.3) (14.8) (28.7) (20.7) (30.3) <10hrs=4 Sub-total Missing System Total 322 78 350 189 278 67 257 54 (73.3) (17.8) (79.7) (43.1) (63.3) (15.3) (58.5) (12.3) 415 413 421 420 403 404 408 413 (94.5) (94.1) (95.9) (95.7) (91.8) (92.0) (92.9) (94.1) 24 26 18 19 36 35 31 (5.5) (5.9) (4.1) (4.3) (8.2) (8.0) (7.1) 26 (5.9) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 439 (100) 832 Table 7.5.126: Time Spent Harvesting NTFPs NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 M SD M M M SD SD YR 1999 3.6579 0.7018 3.5965 0.8695 3.5479 0.9606 3.5913 0.8706 2009 2.0783 1.0187 2.4248 1.1787 2.4865 1.1661 2.3559 1.1413 1999 3.7130 0.6320 3.8158 0.5077 3.7409 0.6810 3.7536 0.6248 2009 2.6053 1.1492 3.0965 1.0388 3.2500 0.9974 3.0333 1.0831 1999 3.4690 0.8026 3.5856 0.8471 3.4088 1.0321 3.4741 0.9241 2009 2.0265 0.9586 2.3784 1.1205 2.4556 1.1400 2.3144 1.0994 1999 3.4775 0.7367 3.3839 0.9702 3.3925 0.9763 3.4132 0.9144 SD 2009 1.9565 0.9947 2.1964 1.0725 2.3602 1.1315 2.2034 1.0894 NTFP Harvesting Time Index (NHTI) Combined HH Natural Capital Index 1 (CHHNCI1) Combined HH Natural Capital Index 2 (CHHNCI2) 833 What is the average time spent/week harvesting the following? SCALE >20hrs=1 16-20hrs=2 10-15hrs=3 <10hrs=4 1. Mushrooms? (NTFP_2Aa) 2. Firewood? (NTFP_2Bb) 3. Snails? (NTFP_2Cc) 4. Glasscutters and other bush- meat? (NTFP_2Dd) 4*4NTFP_2Dd) NTFP_2Cc NTFP_2Bb (NTFP_2Aa4)*(NTFP_2Dd) NTFP_2Cc NTFP_2Bb (NTFP_2AaNHTI4141titit====+++=+++=niniSn16NTFP_2Dd) NTFP_2Cc NTFP_2Bb (NTFP_2AaNHTI41tit==+++=ni13NHTI) +NHMTI +SFAI + MOLPI+SOLPI + LPTI+ HHLPI1+ HHLDI1+CPTI + MOCPI+SOCPI + HHCDI+(CLOICHHNCI1131tit====nni13NHTI) +NHMTI +SFAI + MOLPI+SOLPI + LPTI+ HHLPI2+ HHLDI2+CPTI + MOCPI+SOCPI + HHCDI+(CLOICHHNCI2131tit====nni APPENDIX E: Descriptive Analysis of Social Capital Indexes Table 7.5.127: Number of Close Relatives in the Village - Frequency Frequency (Percentage) None (0)=1 Very Few (1-5) = 2 Many (6-10) = 3 Very Many (>10) = 4 Sub-total Missing System Total 1. Number of close relatives in the village that do not reside in the HH? (FNWKS) 1999 73 (16.6) 125 (28.5) 120 (27.3) 113 (25.7) 431 (98.2) 8 (1.8) 439 (100) 2009 59 (13.4) 112 (25.5) 131 (29.8) 134 (30.5) 436 (99.3) 3 (0.7) 439 (100) 834 Table 7.5.128: Number of Close Relatives in the Village - Descriptive NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 M SD M SD M SD M SD Describe the following? SCALE None (0)=1 Very Few (1-5)=2 Many (6-10)=3 Very Many (>10)=4 YR 1. Number of close relatives in the village that do not reside in the HH? (FNWKS) 1999 2.5913 0.9814 2.689 1.095 2.6313 1.0618 2.6366 1.04863 2009 2.8174 0.9513 2.7983 1.0462 2.7475 1.0653 2.7798 1.0293 Family Network Index (FNWI) = 835 Snni*(FNWKS)FNWI11itt===4*1(FNWKS)11it==niitt(FNWKS)FNWI= Table 7.5.129: Help from Relatives in the Village - Frequency Frequency (Percentage) Never =1 Occasionally = 2 Often = 3 Always = 4 Sub-total Missing System Total 1. Farm or perform other work? (RITV1) 2. Give or receive 3. Give or receive food from your relatives? (RITV2) cash from your relatives? (RITV3) 2009 145 (33.0) 183 (41.7) 60 (13.7) 45 (10.3) 433 (98.6) 6 (1.4) 439 (100) 1999 127 (28.9) 192 (43.7) 71 (16.2) 41 (9.3) 431 (98.2) 8 (1.8) 439 (100) 2009 127 (28.9) 192 (43.7) 70 (15.9) 43 (9.8) 432 (98.4) 7 (1.6) 439 (100) 1999 171 (39.0) 170 (38.7) 59 (13.4) 26 (5.9) 426 (97.0) 13 (3.0) 439 (100) 2009 167 (38.0) 180 (41.0) 53 (12.1) 29 (6.6) 429 (97.7) 10 (2.3) 439 (100) 1999 143 (32.6) 183 (41.7) 59 (13.4) 47 (10.7) 432 (98.4) 7 (1.6) 439 (100) 836 Table 7.5.130: Help from Relatives in the Village - Descriptive NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR 1999 2.0696 0.9433 2.1429 0.9680 1.9242 0.9394 2.0231 0.9510 SD SD M SD M M SD M 2009 2.0435 0.9214 2.075 0.9452 1.9545 0.9575 2.0115 0.9440 1999 2.1754 0.9241 2.1933 0.9767 1.9141 0.8536 2.0603 0.9155 2009 2.1842 0.9175 2.1583 0.9701 1.9444 0.8853 2.0671 0.9228 1999 2.0531 0.9146 1.8475 0.9117 1.7538 0.8129 1.8592 0.8752 2009 2.0526 0.9298 1.8167 0.8791 1.7949 0.8305 1.8695 0.8763 Help from Relatives Inside the Village Index (HRITV) 837 Describe how often you and your relatives in the village help each other perform the following functions? SCALE Never=1 Occasionally=2 Often=3 Always=4 1. Farm or perform other work? (RITV1) 2. Give or receive food from your relatives? (RITV2) 3. Give or receive cash from your relatives? (RITV3) 4*3 RITV3Q)+ RITV2Q+(RITV1Q* RITV3Q)+ RITV2Q+(RITV1QHRITV3131titit======niniSn12 RITV3Q)+ RITV2Q+(RITV1QHRITV31tit===ni Table 7.5.131: Help from Relatives Outside the Village but in Brong Ahafo - Frequency Frequency (Percentage) Never=1 Occasionally=2 Often=3 Always=4 Sub-total Missing System Total 1. Farm or perform other work? (ROTV_B1) 2. Give or receive 3. Give or receive food from your cash from your relatives? (ROTV_B2) relatives? (ROTV_B3) 2009 197 (44.9) 178 (40.5) 44 (10.0) 9 (2.1) (97.5) 11 (2.5) 439 (100) 1999 173 (39.4) 201 (45.8) 51 (11.6) 6 (1.4) 431 (98.2) 8 (1.8) 439 (100) 2009 175 (39.9) 201 (45.8) 49 (11.2) 7 (1.6) 432 (98.4) 7 (1.6) 439 (100) 1999 188 (42.8) 199 (45.3) 40 (9.1) 4 (0.9) 431 (98.2) 8 (1.8) 439 (100) 2009 188 (42.8) 201 (45.8) 38 (8.7) 6 (1.4) 433 (98.6) 6 (1.4) 439 (100) 1999 193 (44.0) 179 (40.8) 47 (10.7) 7 (1.6) 426 (97.0) 13 (3.0) 439 (100) 838 Table 7.5.132: Help from Relatives Outside the Village but in Brong Ahafo - Descriptive Describe how often you and your relatives outside the village but in BA help each other perform the following functions? SCALE Never=1 Occasionally=2 Often=3 Always=4 1. Farm or perform other work? (ROTV_B1) 2. Give or receive food from your relatives? (ROTV_B2) 3. Give or receive cash from your relatives? (ROTV_B3) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 1.7273 0.7408 1.7203 0.7832 1.6515 0.6943 1.6901 0.731 2009 1.6937 0.6980 1.6639 0.7510 1.6919 0.7614 1.6846 0.7409 1999 1.7130 0.6592 1.8319 0.7515 1.7107 0.7232 1.7448 0.7152 2009 1.7130 0.6457 1.7731 0.7300 1.7374 0.7552 1.7407 0.7193 1999 1.6522 0.7014 1.6975 0.6319 1.6751 0.6970 1.6752 0.6795 2009 1.6609 0.6995 1.6500 0.6170 1.7121 0.7286 1.6813 0.6906 839 Help from Relatives Outside the Village Index (HROTV) Table 7.5.133: Help from Relatives Outside Brong Ahafo but in Ghana - Frequency Frequency (Percentage) Never=1 Occasionally=2 Often=3 Always=4 Sub-total Missing System Total 1. Farm or perform other work? (ROTB_G1) 2. Give or receive 3. Give or receive food from your cash from your relatives? (ROTB_G2) relatives? (ROTB_G3) 2009 267 (60.8) 134 (30.5) 23 (5.2) 7 (1.6) 431 (98.2) 8 (1.8) 439 (100) 1999 214 (48.7) 169 (38.5) 44 (10.0) 3 (0.7) 430 (97.9) 9 (2.1) 439 (100) 2009 214 (48.7) 178 (40.5) 35 (8.0) 4 (0.9) 431 (98.2) 8 (1.8) 439 (100) 1999 203 (46.2) 181 (41.2) 43 (9.8) 2 (0.5) 429 (97.7) 10 (2.3) 439 (100) 2009 205 (46.7) 183 (41.7) 40 (9.1) 2 (0.5) 430 (97.9) 9 (2.1) 439 (100) 1999 267 (60.8) 132 (30.1) 25 (5.7) 6 (1.4) 430 (97.9) 9 (2.1) 439 (100) 840 4*3 ROTV_B3Q) ROTV_B2Q(ROTV_B1Q* ROTV_B3Q) ROTV_B2Q(ROTV_B1QHROTV3131titit====++=++=niniSn12 ROTV_B3Q) ROTV_B2Q(ROTV_B1QHROTV31tit==++=ni Table 7.5.134: Help from Relatives Outside Brong Ahafo but in Ghana - Descriptive Describe how often you and your relatives outside of Brong Ahafo help each other perform the following functions? SCALE Never=1 Occasionally=2 Often=3 Always=4 1. Farm or perform other work? (ROTB_G1) 2. Give or receive food from your relatives? (ROTB_G2) 3. Give or receive cash from your relatives? (ROTB_G3) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 1.3628 0.5356 1.5085 0.7367 1.4975 0.6954 1.4651 0.6707 2009 1.3717 0.5378 1.5210 0.7462 1.4874 0.6953 1.4664 0.6739 1999 1.5965 0.6615 1.7119 0.7292 1.5758 0.6921 1.6186 0.6954 2009 1.6140 0.6850 1.6471 0.6838 1.5707 0.6703 1.6032 0.6771 1999 1.6228 0.6963 1.6838 0.6905 1.6162 0.6639 1.6364 0.6789 2009 1.6228 0.6834 1.6441 0.6859 1.6162 0.6562 1.6256 0.6702 Help from Relatives Outside the Region i.e. BA Index (HROTB) 841 4*3 ROTB_G3Q) ROTB_G2Q(ROTB_G1Q* ROTB_G3Q) ROTB_G2Q(ROTB_G1QHROTB3131titit====++=++=niniSn12 ROTB_G3Q) ROTB_G2Q(ROTB_G1QHROTB31tit==++=ni Table 7.5.135: Help from Relatives Outside Ghana - Frequency Frequency (Percentage) Never = 1 Occasionally = 2 Often = 3 Always = 4 Sub-total Missing System Total 1. Give or receive 2. Give or receive food from your relatives? (ROG1) cash from your relatives? (ROG2) 1999 368 (83.8) 49 (11.2) 8 (1.8) 1 (0.2) 426 (97.0) 13 (3.0) 439 (100) 2009 367 (83.6) 53 (12.1) 7 (1.6) 1 (0.2) 428 (97.5) 11 (2.5) 439 (100) 1999 344 (78.4) 63 (14.4) 19 (4.3) 2 (0.5) 428 (97.5) 11 (2.5) 439 (100) 2009 344 (78.4) 66 (15.0) 19 (4.3) 1 (0.2) 430 (97.9) 9 (2.1) 439 (100) 842 Table 7.5.136: Help from Relatives Outside Ghana - Descriptive NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 M SD YR 1999 1.1053 0.4072 1.2222 0.5271 1.1538 1.2487 1.1596 0.4315 SD SD M M SD M 2009 1.1053 0.3849 1.2288 0.5133 1.1582 0.3929 1.1636 0.4289 1999 1.1504 0.4672 1.3475 0.6715 1.2487 0.5092 1.2500 0.5523 2009 1.1579 0.4527 1.3361 0.6415 1.2487 0.5092 1.2488 0.5384 Help from Relatives Outside of Ghana Index (HROGI) 843 Describe how often you and your relatives outside of Ghana help each other perform the following functions? SCALE Never=1 Occasionally=2 Often=3 Always=4 1. Give or receive food from your relatives? (ROG1) 2. Give or receive cash from your relatives? (ROG2) 4*2ROG2Q)+ (ROG1Q*ROG2Q)+ (ROG1QHROGI2121titit======niniSn8ROG2Q)+ (ROG1QHROGI21tit===ni Table 7.5.137: Support from Government and Various NGOs - Frequency Frequency (Percentage) 1. GOV’NT 2. NGO extension advice on agricultural extension advice on agricultural production, production, forest forest management, management, health awareness? (GNGO1) health awareness? (GNGO2) 3. GOV’NT support in the form of cash, farm inputs, food, cloths, etc.? (GNGO3) 4. NGO support in the form of cash, farm inputs, food, cloths, etc.? (GNGO4) Never=1 Occasionally=2 Often=3 Always=4 Sub-total Missing System Total 1999 175 (39.9) 210 (47.8) 32 (7.3) 11 (2.5) 428 (97.5) 11 (2.5) 439 (100) 2009 167 (38.0) 204 (46.5) 42 (9.6) 19 (4.3) 432 (98.4) 7 (1.6) 439 (100) 1999 233 (53.1) 167 (38.0) 29 (6.6) 2 (0.5) 431 (98.2) 8 (1.8) 439 (100) 2009 219 (49.9) 176 (40.1) 31 (7.1) 8 (1.8) 434 (98.9) 5 (1.1) 439 (100) 1999 309 (70.4) 115 (26.2) 2009 292 (66.5) 132 (30.1) 1999 317 (72.2) 106 (24.1) 2009 305 (69.5) 120 (27.3) 9 8 8 6 (2.1) (1.8) (1.8) (1.4) 0 (0.0) 433 (98.6) 6 (1.4) 439 (100) 2 (0.5) 434 (98.9) 5 (1.1) 439 (100) 1 (0.2) 432 (98.4) 7 (1.6) 439 (100) 3 (0.7) 434 (98.9) 5 (1.1) 439 (100) 844 Table 7.5.138: Support from Government and Various NGOs - Descriptive How often you receive the following external support? SCALE Never=1 Occasionally=2 Often=3 Always=4 1. GOV’NT extension advice on agricultural production, forest management, health awareness? (GNGO1) 2. NGO extension advice on agricultural production, forest management, health awareness? (GNGO2) 3. GOV’NT support in the form of cash, farm inputs, food, cloths, etc.? (GNGO3) 4. NGO support in the form of cash, farm inputs, food, cloths, etc.? (GNGO4) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 1.8230 0.7100 1.5254 0.6757 1.7716 0.7169 1.7173 0.7125 2009 1.8407 0.7857 1.5798 0.7071 1.9050 0.8120 1.7986 0.7878 1999 1.7105 0.7130 1.3559 0.5473 1.5427 0.6252 1.536 0.6419 2009 1.7368 0.7763 1.4118 0.5736 1.6418 0.7079 1.6037 0.7027 1999 1.2456 0.4324 1.2437 0.4504 1.3800 0.5632 1.3072 0.505 2009 1.2895 0.4746 1.2521 0.4361 1.4527 0.6156 1.3548 0.5423 1999 1.2544 0.4944 1.2373 0.4468 1.3400 0.5440 1.2894 0.5071 2009 1.3158 0.5536 1.2437 0.4311 1.3781 0.5799 1.3249 0.5377 845 Government and NGO Support Index (GNSI) Table 7.5.139: Support from Friends Inside or Outside the Village - Frequency 1. Support in the form of extension, forest management, and health awareness advice, from neighbors and friends in or outside the village? (SFNF1) 2. Support in the form of cash, farm inputs, food, cloths, etc from neighbors and friends in or outside the village? (SFNF2) 1999 134 (30.5) 239 (54.4) 53 (12.1) 8 (1.8) 434 (98.9) 5 (1.1) 439 (100) 2009 123 (28.0) 238 (54.2) 66 (15.0) 9 (2.1) 436 (99.3) 3 (0.7) 439 (100) 1999 208 (47.4) 198 (45.1) 24 (5.5) 2 (0.5) 432 (98.4) 7 (1.6) 439 (100) 2009 210 (47.8) 198 (45.1) 26 (5.9) 2 (0.5) 436 (99.3) 3 (0.7) 439 (100) Frequency (Percentage) Never=1 Occasionally=2 Often=3 Always=4 Sub-total Missing System Total 846 4*4GNGO4Q) GNGO3Q GNGO2Q (GNGO1Q*GNGO4Q) GNGO3Q GNGO2Q (GNGO1QGNSI4141titit====+++=+++=niniSn16GNGO4Q) GNGO3Q GNGO2Q (GNGO1QGNSI41tit==+++=ni Table 7.5.140: Support from Friends Inside or Outside the Village - Descriptive Describe how often your HH receive help from friends whether inside or outside of the village? SCALE Never=1 Occasionally=2 Often=3 Always=4 1. Support in the form of extension, forest management, and health awareness advice, from neighbors and friends in or outside the village? (SFNF1) 2. Support in the form of cash, farm inputs, food, cloths, etc from neighbors and friends in or outside the village? (SFNF2) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 1.8870 0.7582 1.6864 0.6090 1.9254 0.6924 1.8502 0.6952 2009 1.8870 0.7582 1.7227 0.5955 2.0347 0.7288 1.9106 0.7136 1999 1.5000 0.6133 1.5847 0.6038 1.6300 0.6286 1.5833 0.6187 2009 1.4522 0.5959 1.5462 0.5785 1.6881 0.6516 1.5872 0.6246 847 Other forms of Support from Neighbors and Friends Index (OFANF) Table 7.5.141: Membership in Community Organizations/Associations - Frequency 1. Is any member of 2. Is any member of the HH part of a farming group or association? (GA1) the HH part of a collective marketing group/ association? (GA2) 1999 241 (54.9) 185 (42.1) 426 (97.0) 13 (3.0) 439 (100) 2009 177 (40.3) 258 (58.8) 435 (99.1) 4 (0.9) 439 (100) 1999 369 (84.1) 47 (10.7) 416 (94.8) 23 (5.2) 439 (100) 2009 372 (84.7) 54 (12.3) 426 (97.0) 13 (3.0) 439 (100) Frequency (Percentage) No=0 Yes=1 Sub-total Missing System Total 848 4*2SFNF2Q) +(SFNF1Q*SFNF2Q) +(SFNF1QOFANF2121titit======niniSn8SFNF2Q) +(SFNF1QOFANF21tit===ni Table 7.5.142: Membership in Community Organizations/Associations - Descriptive Is any member of the HH part of the following organizations? SCALE No=0 Yes=1 1. Farming or collective marketing group? (AsscMember) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.4138 0.4947 0.3167 0.4671 0.5123 0.5011 0.4396 0.4969 2009 0.4576 0.4993 0.3167 0.4671 0.7488 0.4348 0.5923 0.4920 849 Table 7.5.143: Attendance of Community Organizations/Associations Meetings - Frequency Frequency (Percentage) Never=1 Occasionally=2 Often=3 Always=4 Sub-total Missing System Total 1. How often do you attend farmer group or association meetings? (FGM1) 2. How often do you attend collective marketing group/association meetings? (FGM2) 1999 2009 1999 2009 187 (42.6) 123 (28.0) 79 (18.0) 31 (7.1) 420 (95.7) 19 (4.3) 439 (100) 157 (35.8) 131 (29.8) 96 (21.9) 44 (10.0) 428 (97.5) 11 (2.5) 439 (100) 326 (74.3) 44 (10.0) 12 (2.7) 2 (0.5) 384 (87.5) 55 (12.5) 439 (100) 311 (70.8) 47 (10.7) 17 (3.9) 2 (0.5) 377 (85.9) 62 (14.1) 439 (100) 850 Table 7.5.144: Attendance of Community Organizations/Associations Meetings - Descriptive NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 M SD M SD M SD M SD 1999 0.4397 0.4985 0.3583 0.4815 0.6453 0.4796 0.5421 0.4988 Is any member of the HH part of the following organizations ? CALE Never=0 At least once a year=1 YR 1. Frequency of farming or marketing meeting attendance? (AssocMeetings) 2009 0.5948 0.4931 0.375 0.4862 0.7833 0.4131 0.6219 0.4855 Community and Association Membership Index (CAMI) 851 )1*1()1*1(ngs)AssocMeeti +r(AsscMembe)*()*(ngs)AssocMeeti +r(AsscMembeCAMI21221121titit+=+=====niniSnSn211 =n where21=+=+nn2ngs)AssocMeeti +r(AsscMembeCAMI21tit===ni Table 7.5.145: Membership in Religious Groups - Frequency Frequency (Percentage) No=0 Yes=1 Sub-total Missing System Total Is any member of the HH part of a church, mosque or any religious organization inside or outside of the village? (GA3) 1999 100 (22.8) 322 (73.3) 422 (96.1) 17 (3.9) 439 (100) 2009 90 (20.5) 335 (76.3) 425 (96.8) 14 (3.2) 439 (100) Table 7.5.146: Membership in Religious Groups - Descriptive NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 M SD YR 1999 0.6810 0.4681 0.7083 0.4564 0.7783 0.4164 0.7335 0.4426 SD SD M M M SD 2009 0.6897 0.4646 0.7500 0.4348 0.8128 0.3910 0.7631 0.4257 852 Is any member of the HH part of the following organizations? SCALE No =0 Yes=1 1. Religious group such as church or mosque? (GA3a) Table 7.5.147: Attendance of Religious Group Meetings - Frequency Frequency (Percentage) Never=1 Occasionally=2 Often=3 Always=4 Sub-total Missing System Total 2. Church or mosque attendance? (FCMA1) 1999 36 (8.2) 63 (14.4) 64 (14.6) 271 (61.7) 434 (98.9) 5 (1.1) 439 (100) 2009 37 (8.4) 61 (13.9) 66 (15.0) 272 (62.0) 436 (99.3) 3 (0.7) 439 (100) Table 7.5.148: Attendance of Religious Group Meetings - Descriptive How often HH attends Religious Group meetings? SCALE Never=0 At least once a year=1 2. Church or mosque attendance? (FCMA1_Freq) NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 M SD M YR 1999 0.8793 0.3272 0.9167 0.2775 0.9163 0.2777 0.7722 0.4199 SD SD SD M M 2009 0.8621 0.3463 0.9000 0.3013 0.9409 0.2364 0.9089 0.2881 853 Religious Group Membership and Meeting Attendance Index (RGMMA) , Table 7.5.149: Support from Religious Groups - Frequency Frequency (Percentage) Never=1 Occasionally=2 Often=3 Always=4 Sub-total Missing System Total 2. Support in the form of advice on family matters or help resolving disputes in or outside of the HH? (FCMA2) 1999 95 (21.6) 156 (35.5) 86 (19.6) 97 (22.1) 434 (98.9) 5 (1.1) 439 (100) 2009 93 (21.2) 166 (37.8) 76 (17.3) 101 (23.0) 436 (99.3) 3 (0.7) 439 (100) 3. Support in the form of cash, farm inputs, food, clothes? (FCMA3) 4. Help in performing farming or other economic activities? (FCMA4) 1999 235 (53.5) 152 (34.6) 35 (8.0) 11 (2.5) 433 (98.6) 6 (1.4) 439 (100) 2009 236 (53.8) 152 (34.6) 36 (8.2) 11 (2.5) 435 (99.1) 4 (0.9) 439 (100) 1999 290 (66.1) 103 (23.5) 25 (5.7) 14 (3.2) 432 (98.4) 7 (1.6) 439 (100) 2009 297 (67.7) 101 (23.0) 25 (5.7) 12 (2.7) 435 (99.1) 4 (0.9) 439 (100) 854 )1*1()1*1(q) FCMA1_Fre+(GA3a)*()*(q) FCMA1_Fre+(GA3aRGMMA21221121titit+=+=====niniSnSn211 =n where21=+=+nn2q) FCMA1_Fre+(GA3aRGMMA21tit===ni Table 7.5.150: Support from Religious Groups - Descriptive NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 2.4435 1.1176 2.5126 1.0404 2.3650 1.0475 2.4263 1.0639 2009 2.3913 1.1371 2.4790 1.0565 2.4109 1.0340 2.4243 1.0662 1999 1.5913 0.8045 1.5294 0.6616 1.6231 0.7612 1.5889 0.7466 2009 1.5565 0.7515 1.5210 0.6993 1.6517 0.7734 1.5908 0.7485 1999 1.5217 0.8721 1.4202 0.6701 1.4293 0.7143 1.4514 0.7477 2009 1.4522 0.7637 1.3866 0.6524 1.4428 0.7470 1.4299 0.7258 Religious Group Support Index (RGSI) 855 How often HH receives the following assistance from religious groups? SCALE Never=1 Occasionally=2 Often=3 Always=4 2. Support in the form of advice on family matters or help resolving disputes in or outside of the HH? (FCMA2) 3. Support in the form of cash, farm inputs, food, cloths? (FCMA3) 4. Help in performing farming or other economic activities? (FCMA4) 4*3 FCMA4Q)+ FCMA3Q+(FCMA2Q* FCMA4Q)+ FCMA3Q+(FCMA2QRGSI3131titit======niniSn12 FCMA4Q)+ FCMA3Q+(FCMA2QRGSI31tit===ni Table 7.5.151: Size of Household - Frequency Frequency (Percentage) 1. Number of individuals in the HH? (HHPOP) Very Few (<5)=1 Few (5-10)=2 Many (11-15)=3 Very Many (>15)=4 Sub-total Missing System Total 1999 132 (30.1) 161 (36.7) 78 (17.8) 35 (8.0) 406 (92.5) 33 (7.5) 439 (100) 2009 65 (14.8) 224 (51.0) 95 (21.6) 30 (6.8) 414 (94.3) 25 (5.7) 439 (100) How would you describe the following? SCALE Very Few (<5)=1 Few (5-10)=2 Many (11-15)=3 Very Many (>15)=4 1. Number of individuals in the HH? (HHPOP) Table 7.5.152: Size of Household - Descriptive NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 1.8909 0.8604 2.1364 0.92347 2.0699 0.9645 2.0394 0.9288 2009 2.1150 0.7289 2.2523 0.8143 2.2579 0.8177 2.2174 0.7941 856 HH Population Index (HHPOPI) = Table 7.5.153: Joint Household Activities - Frequency Frequency (Percentage) No=0 Yes=1 Sub-total Missing System Total 1. Farm together? granary or storage 2. Use the same (HH_LIV_ACT1) facilities? (HH_LIV_ACT2) 4. Cook together as a HH? (HH_LIV_ACT4) 1999 135 (31.4) 295 (68.6) 430 (100) 9 (2.1) 439 (100) 2009 137 (31.4) 299 (68.6) 436 (100) 3 (0.7) 439 (100) 1999 113 (26.3) 317 (73.7) 430 (100) 9 (2.1) 439 (100) 2009 113 (26.2) 319 (73.8) 432 (100) 7 (1.6) 439 (100) 1999 78 (18.2) 350 (81.8) 428 (100) 11 (2.5) 439 (100) 2009 83 (19.3) 347 (80.7) 430 (100) 9 (2.1) 439 (100) 857 Snni*(HHPOPI)HHPOP11itt===4*1(HHPOPI)11it==niitt(HHPOPI)HHPOP= Answer yes if the following apply to your HH? SCALE No =1 Yes=2 1. Farm together? (HH_LIV_ACT1) 2. Use the same granary or storage facilities? (HH_LIV_ACT2) 4. Cook together as a HH? (HH_LIV_ACT4) Table 7.5.154: Joint Household Activities - Descriptive NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 1.8246 0.3820 1.6525 0.47819 1.6263 0.4850 1.686 0.4646 2009 1.8070 0.3964 1.6050 0.4909 1.6650 0.4732 1.6858 0.4647 1999 1.8246 0.3820 1.7119 0.45483 1.7020 0.4585 1.7372 0.4407 2009 1.8142 0.3907 1.7143 0.4537 1.7100 0.4549 1.7384 0.4400 1999 1.9027 0.2978 1.7731 0.42059 1.7959 0.4041 1.8178 0.3865 2009 1.9027 0.2978 1.7712 0.4219 1.7739 0.4194 1.8070 0.3951 Joint HH Activity Index (JHHAI) 858 2*3T4Q) HH_LIV_AC+T2Q HH_LIV_AC+T1Q(HH_LIV_AC*T4Q) HH_LIV_AC+T2Q HH_LIV_AC+T1Q(HH_LIV_ACJHHAI3131titit======niniSn6T4Q) HH_LIV_AC+T2Q HH_LIV_AC+T1Q(HH_LIV_ACJHHAI31tit===ni Table 7.5.155: Short-term Seasonal Migration - Frequency Frequency (Percentage) No=0 Yes=1 Sub-total Missing System Total 2. Are there HH members absent from the HH for a relatively long period of time (>3months)? (HH_LIV_ACT6) 1999 212 (48.3) 176 (40.1) 388 (88.4) 51 (11.6) 439 (100) 2009 194 (44.2) 220 (50.1) 414 (94.3) 25 (5.7) 439 (100) Table 7.5.156: Short-term Migration - Descriptive NSEMSAW NON-MTS in MTS in YAYA N=116 YAYA N=120 N=203 POOLED N=439 YR M SD M SD M SD M SD 1999 0.4569 0.5003 0.4583 0.50035 0.3350 0.4732 0.3212 0.4675 2009 0.4914 0.5021 0.5167 0.5018 0.5714 0.4961 0.4419 0.4972 859 Answer yes if the following apply to your HH? SCALE No =0 Yes=1 2. Are there HH members absent from the HH for a relatively long period of time (>3months)? (ABSENT_1) Table 7.5.157: Long-term Seasonal Migration - Frequency Frequency (Percentage) None=1 Few (1-3)=2 Many (4-5)=3 Very Many (>5)=4 Sub-total Missing System Total 3. Number of HH members usually absent for a relatively long period (>3months)? (ABSENT) 2009 212 (52.2) 139 (34.2) 44 (10.8) 11 (2.7) 406 (100) 33 (7.5) 439 (100) 1999 298 (67.9) 101 (23.0) 30 (6.8) 10 (2.3) 439 (100) 0 (0.0) 439 (100) 860 Table 7.5.158: Long-term Seasonal Migration - Frequency NSEMSAW N=116 NON-MTS in YAYA N=120 MTS in YAYA N=203 POOLED N=439 M SD M SD M SD M SD 1999 1.3190 0.6406 1.4833 0.68579 1.4729 0.7791 1.4351 0.7213 Answer yes if the following apply to your HH? SCALE None=1 Few (1-3)=2 Many (4-5)=3 Very Many (>5)=4 YR 3. Number of HH members usually absent for a relatively long period (>3months)? (ABSENT) 2009 1.4248 0.7417 1.7523 0.7717 1.7065 0.7898 1.6404 0.7817 HH Migration Index (HMIGI) Aggregate HH Social Capital Index (AHHSCI) 861 )4*1()1*1(ABSENTQ) +(ABSENT_1Q)*()*(ABSENTQ) +(ABSENT_1QHMIGI21221121titit+=+=====niniSnSn211 =n where21=+=+nn5ABSENTQ) +(ABSENT_1QHMIGI21tit===ni13 HMIGI)+JHHAI + HHPOPI+ RGSI+ RGMMA+CAMI +OFANF +GNSI + HROGI+ HROTB+ HROTV+ HRITVI+(FNWIAHHSCI131tit====nni CHAPTER 8: USING DIFFERENCE-IN-DIFFERENCE (DID) METHODOLOGY TO ISOLATE DIRECT AND SPILLOVER EFFECTS OF GHANA’S MTS PROJECT ON PARTICIPANT AND NON- PARTICIPANT PROJECT COMMUNITIES 862 Abstract In 2009, a survey of 878 households was carried out to track changes in five livelihood assets (financial, human, natural, social, and physical capital) among participants and non- participants of Ghana’s Modified Taungya System (MTS) reforestation project. This study compared changes in livelihood assets around the time of launching the MTS program (1999) and 10 years (2009) into the program. Data for the analysis was drawn from surveys conducted in 19 forest fringe communities around Yaya, Nsemre and Sawsaw forest reserves in Ghana’s Brong Ahafo Region. With an interest in attributing change in livelihood outcomes, the omitted variable problem was anticipated. Hence, a Difference-in-Difference (DID) method was used to compute how much of the observed changes in each livelihood asset may be attributed to the MTS program. The DID estimates suggests that on average 5.25% of the observed changes in livelihood assets among MTS participants may be attributed directly to the MTS program while 4.28% of changes among non-MTS participants in project communities may be attributed to spillover effects from the project. The DID results also suggest that the MTS program delayed the rate of decline in Natural Capital Index by 5.7% and by 4% respectively among MTS project participants and Non-participants in project communities. 863 8.1 Difference in Difference (DID) This dissertation is interested in the magnitude of change in five livelihood indicators/outcomes/assets between 1999 and 2009, as well as how much of these changes may be attributed to the MTS. With an interest in attributing change in livelihood outcomes between two time periods, the omitted variable problem is anticipated. Omitted variables in this case are factors other than the MTS that may have contributed to observed changes in livelihood outcomes during the MTS period but cannot be accounted for by simple differencing of livelihood outcomes of the two time periods. To effectively address the omitted variable problem in this study, the DID approach was used to isolate the impact of MTS on each livelihood outcome during the program implementation period. The DID approach generally requires that treatment (MTS) and control (Non-MTS) groups be similar in every respect in the base (1999) and treatment years (2009) except for the MTS treatment on participating households in 2009. Since this study uses household survey data and not experimental data, staking a claim that both MTS and Non-MTS households are similar in every respect but for the MTS treatment on program households in 2009 would be extremely difficult. With household survey research such as this study, it is highly unlikely to find the equivalent of treatment and control groups in which every variable except for MTS treatment are the same for the two time periods under consideration. A different but weaker assumption is thus warranted in the particular case presented in this study (see Imbens and Wooldridge, 2007). The alternative assumption (applied in this study) states that in the absence of treatment, the unobserved differences between treatment and control groups are the same over time. With this new assumption, data on MTS and Non-MTS households in the pre-treatment period (1999) are first used to estimate a Normal Difference between the two groups following which the after- 864 treatment difference in 2009 is computed. Figure 8.1 below uses hypothetical livelihood outcomes to illustrate how the Normal Difference and MTS treatment effect are computed in this study. Following the illustration in figure 8.1, the survey results from the five livelihood outcomes are used to compute MTS treatment effects. Figure 8.1.1: Illustration of Normal and Treatment Effect in a DID Approach 8.1.1 Background information on data used for DID Though not deliberate, most development projects for one reason or the other often neglect or delay collection of baseline information on key indicators needed for effective project monitoring and evaluation. In the absence of such baseline, for example on Ghana’s MTS project, this dissertation research used recall methodology to obtain pre-MTS livelihood outcomes around the time of the project’s initiation (approximately 10 years prior to implementation of this survey). Without baseline data on Pre-MTS (1999) livelihood outcomes, 865 Pre-MTS (1999)Post-MTS (2009)5104317BCALivelihood Outcome(Y)Treatment Effect(26)NormalDifference(5)12 the estimate of the treatment effect as demonstrated in figure 8.1 may be erroneously computed as simply the distance between A and B (i.e., 43-12=31). A treatment effect of AB (31) presupposes that the post-MTS outcome AB is a result of only MTS treatment with no contributions from other omitted variables or factors. This assumption about the MTS-treatment effect of AB thus fail to account for other omitted variables as well as any natural trends in Y that may have influenced changes in livelihood outcomes during the project implementation periods between 1999 and 2009. Difference-in-Difference accounts for omitted variables by assuming that for any given livelihood outcome (Y), there exist a Normal Difference that remains constant with time between the treatment (MTS) and control (Non-MTS) households. In other words, DID assumes a constant trend in livelihood outcomes Y for both treatment and control groups. From the illustration in figure 8.1 above, the Normal Difference may be computed by obtaining the difference in livelihood outcomes Y (CB=10-5=5 or 17-12=5) between the MTS and Non-MTS groups during the pre-MTS period (1999). Once a Normal Difference of five (CB=5) is computed between MTS and non-MTS households between 1999 and 2009 the MTS treatment effect of AC (43-17=26) is easily determined/isolated. Because a constant trend in Y is assumed for both MTS and Non-MTS groups, there is a risk of overestimating the MTS treatment effect AC if the trend in Y were greater among the MTS group compared to the non-MTS. Equally so, AC is likely to be underestimated if the trend in Y were smaller among the MTS group relative to the non-MTS. This is a major weakness of the DID approach since the identifying assumption of equal or similar trends in Y in the absence of MTS treatment can never be tested with data on only two time periods. With more than two observations on MTS and non-MTS however the assumption of a common trend may be tested. The sections that follow first defines the five 866 livelihood outcomes of interest and then for each outcome compute and graph their DID estimates. 867 Index Name of Index/ Variable 1. CHHFCI Combine HH Financial Capital Index 2a. CHCI1 Combined Household Human Capital Index1 2b. CHCI2 Combined Household Human Capital Index2 Table 8.1.1: Definition of Household Aggregate Livelihood Indexes Description of Indexes Computation of Indexes CHHFCI captures a HH’s aggregate or combined Financial Capital endowment. This index was generated by averaging of all the 10 financial capital indexes. Hence the maximum attainable score on a household’s combined Financial Capital endowment is 10 and the minimum is 0. CHHFCI index thus ranges from 0 to 1. CHCI1 captures a HH’s aggregate or combined human capital endowment/assets. This index is simply the average of six human capital indexes (Education and Literature, Migration, Dietary Diversity Aggregate Frequency of Consumption, Dietary Diversity Aggregate Food Sufficiency, Household Health and Disease and Household Mortality Indices). The maximum attainable score on a household’s combined Human Capital endowment is thus 6 and the minimum is 0. CHCI1 index thus ranges from 0 to 1. CHCI2 is very similar to CHCI1 however the only difference is that in CHCI2, DDAFC and DDAFS are replaced with Combined HH Dietary Diversity Index for Trend & Frequency of Consumption (CHHDDI1) and Combined Household Dietary Diversity Index for Trend in Sufficiency of Consumption (CHHDDI2). Just like CHCI1, a HH’s maximum attainable CHCI2 score is 6 and the minimum is 0. CHCI2 index also ranges from 0 to 1. 3a. CHPCI1 Combine HH Physical Capital Index CHPCI1 captures a HH’s aggregate or combined physical capital endowment. This index is simply the average of all the 10 indexes hence the maximum attainable score on a household’s combined Physical Capital endowment is 10 and the minimum is 0. CHPCI1 index thus ranges from 0 to 1. 868 10] HHILI+VSEI +AHHIS +TLSAI + FHHBAI+HHBAI +HHE2+ HEI1+SHISI +[PHISICHHFCIt=]6[] HHMI+HHDI + DDAFS+DDAFC +MIGI+[EDULITCHCI1t=]6[] HHMI+ HHDI+CHHDDI2+CHHDDI1 + MIGI+[EDULITCHCI2t=]10[] LHP+BHP +CNWR +CNWA +MHPT + LSWD+KBTF +LSWS + HCRT+[HOROCHPCI1tt==n Index Name of Index/ Variable Table 8.1.1 (cont’d) Description of Indexes Computation of Indexes Combine HH 3b. CHPCI2 Physical Capital Index CHPCI2 is similar to CHPCI1 only in CHPCI2 Luxury Household Possessions Index was not included in the computation thus making the maximum attainable sore for CHPCI2 9. 4. AHHSCI Aggregate HH Social Capital Index AHHSCI captures a HH’s aggregate social capital endowment. This index is simply the average of all the 13 indexes hence the maximum attainable score on Social Capita is 13 and the minimum is. AHHSCI index this ranges from 0.08 and 1. 5a. Combined HH Natural Capital CHHNCI1 Index 1 (CHHNCI1) 5b. Combined HH Natural Capital CHHNCI2 Index 2 (CHHNCI2) CHHNCI1 captures a HH’s aggregate Natural capital endowment. This index takes the average of 13 of the 15 indexes listed above. HHLDI and HHLPI both of which capture respectively the diversity and population of five major livestock categories was included in the computation of CHHNCI1 while HHLD2 and HHLP2 which capture diversity and population of only three of the five most common livestock categories was not included in the computation. The maximum attainable score on CHHNCI1is thus 13 and the minimum is 2.27. CHHNCI1 index ranges from 0.17 to 1. CHHNCI2 captures a HH’s aggregate Natural capital endowment. This index takes the average of 13 of the 15 indexes listed above. HHLD2 and HHLP2 both of which capture respectively the diversity and population of only three major livestock categories (poultry, goats and sheep, pigs) was included in the computation of CHHNCI2 while HHLD1 and HHLP1 which capture diversity and population of all the five most common livestock categories was not included in the computation. The maximum attainable score on CHHNCI2is thus 13 and the minimum is 2.27. CHHNCI21 index ranges from 0.17 to 1. 869 ]9[] BHP+CNWR +HCPO + MHPT+LSWD + KBTF+LSWS +HCRT+[HOROCHPCI2tt==n13 HMIGI)+JHHAI + HHPOPI+ RGSI+RGMMA +CAMI +OFANF +GNSI+ HROGI+ HROTB+HROTV + HRITVI+(FNWIAHHSCI131tit====nni13NHTI) +NHMTI +SFAI + MOLPI+SOLPI + LPTI+HHLPI1 + HHLDI1+CPTI + MOCPI+SOCPI +HHCDI+(CLOICHHNCI1131tit====nni13NHTI) +NHMTI +SFAI + MOLPI+SOLPI + LPTI+HHLPI2 + HHLDI2+CPTI +MOCPI+SOCPI + HHCDI+(CLOICHHNCI2131tit====nni 8.1.2 Definition of DID Estimators for Household Livelihood Indexes Let represent the mean livelihood outcome in group at time t Where: = 0 for the control group (NsemSaw) = 1 for the treatment group in treatment community (MTS in Yaya) and = 2 for none-treatment group in treatment community (Non-MTS in Yaya). t = 0 represent the pre-treatment period (1999) t = 1 represent the post treatment period (2009) The difference estimator uses the difference in means between treatment and control group post- treatment as the estimate of the treatment effect. Hence if we assume that any observed differences in the treatment (MTS) and control groups in Yaya (Non-MTS) between 1999 and 2009 are a result of the MTS treatment then DID estimators in these cases becomes and . Using DID estimators and assumes that the treatment (MTS) and control groups in Yaya (Non-MTS) and outside of Yaya (NsemSaw) have no other differences apart from the treatment. Due to the influx of development projects (local and international) in rural communities in Ghana, it is extremely difficult to assume that the only reason for observed differences in livelihood outcomes is a result of one particular intervention such as the MTS. This study thus uses a weaker assumption which states that any difference in the change in mean between treatment and control groups is the result of the treatment. This assumption allows for the use of: and as an estimate of the treatment effect, which is also the DID estimators. Since Non-MTS in Yaya households were not directly targeted by the MTS reforestation project it is expected that the MTS effects or the spillover effects of MTS will be of a lower magnitude relative to households 870 itiiii0111−0121−0111−0121−)()(00010111−−−)()(00012021−−− directly targeted by the MTS program. This study thus assumes that implementation of the MTS in a project community impacts project beneficiaries/participants directly and non-participants indirectly. The benefits that accrue to non-participants is thus described as the spillover effects of project activity on non-participant households. This study thus captures MTS spillover/indirect effects by including data from non-beneficiary households residing in MTS project communities. To measure the direct project effects, communities far removed from the project communities were used as control for capturing these effects. Table 8.2 below explains how the direct and spillover effects of MTS project on the five selected livelihood outcomes are computed. 871 8.1.3 Difference in Difference Estimates of Combined Livelihoods Indexes Table 8.1.2: DID Estimates of Combined Livelihoods Indexes LIVELIHOOD CALIPTAL INDEX MTS HH in Yaya (n=203) Non- MTS HH in Yaya (n=120) NsemSa w (n=116) Estimated Differences Within MTS HH in Yaya 1 Groups Non- MTS in NsemSaw Yaya 2 3 Difference-in-Difference Estimates MTS and Non-MTS HH in Yaya MTS in Yaya & NsemSaw Non-MTS in Yaya & NsemSaw Year M ( ) M ( ) M ( ) (1)-(2) (1)-(3) (2)-(3) 1. CHHFCI 2a. CHCI1 2b. CHCI2 3a. CHPCI1 3b. CHPCI2 2009 (t=1) 0.429 0.421 1999 (t=0) 0.391 0.392 2009 (t=1) 0.626 0.631 1999 (t=0) 0.590 0.593 2009 (t=1) 0.626 1999 (t=0) 0.563 0.63 0.57 2009 (t=1) 0.507 0.505 1999 (t=0) 0.313 0.319 2009 (t=1) 0.554 0.552 0.380 0.391 0.532 0.535 0.531 0.507 0.386 0.269 0.427 1999 (t=0) 0.344 0.349 0.297 0.038*** 0.023** -0.011 0.009 0.049 0.040 0.036** 0.038** -0.003 -0.002 0.039 0.041 0.063*** 0.06*** 0.024* 0.003 0.039 0.036 0.194*** 0.186*** 0.117*** 0.008 0.077 0.069 0.210*** 0.203*** 0.130*** 0.007 0.08 0.073 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) *** p<0.001, ** p<0.01, * p<0.05 872 t1t2t01011−2021−0001− Table 8.1.2 (cont’d) LIVELIHOOD CALIPTAL INDEX MTS HH in Yaya (n=203) Non-MTS HH in Yaya (n=120) Estimated Differences Within Difference-in-Difference Groups Estimates NsemSaw (n=116) MTS HH in Yaya Non-MTS in Yaya NsemSaw 1 2 3 MTS and Non-MTS HH in Yaya MTS in Yaya & NsemSaw Non-MTS in Yaya & NsemSaw Year M ( ) M ( ) M ( ) (1)-(2) (1)-(3) (2)-(3) 4. AHHSCI 5a. CHHNCI1 5b. CHHNCI2 2009 (t=1) 0.543 0.501 0.518 1999 (t=0) 0.508 0.495 0.500 2009 (t=1) 0.509 0.483 0.488 1999 (t=0) 0.541 0.532 0.577 2009 (t=1) 0.527 0.498 0.506 1999 (t=0) 0.558 0.548 0.599 0.035*** 0.006 0.018 0.029 0.017 -0.012 -0.032*** -0.049*** -0.089*** 0.017 0.057 0.04 -0.031*** -0.050*** -0.093*** 0.019 0.062 0.043 1) P>|t| represents the two-tailed significance probability under the null hypothesis that the means values of the various Index are equal across groups; 2) Between group Diff= Mean (Non-MTS) – Mean (MTS), 3) Diff= Mean (NsemSaw) – Mean (MTS), 4) Mean Within group Diff = Mean (1999) – Mean (2009), 5) *** p<0.001, ** p<0.01, * p<0.05 873 t1t2t01011−2021−0001− 8.2 Direct and Spillover MTS Project Effect on CHHFCI Figure 8.2 below explains how the DID approach was used to isolate the direct and spillover MTS project effects on CHHFCI among MTS and Non-MTS households between 1999 and 2009. The NsemSaw group was used as a control in computing the “Normal Differences (ND)” between MTS and NsemSaw as well as between Non-MTS and NsemSaw. By computing the ND for both MTS and Non-MTS groups in Yaya, the direct MTS project effect and spillover were determined while also accounting for possible effects of omitted variables. According to figure 8.2 the MTS and NsemSaw groups had the same mean CHHFCI of 0.391 during the pre- MTS period hence the ND between these two groups is zero. A zero ND suggests that in the absence of the MTS project the mean CHHFCI for both MTS and NsemSaw households in the post-MTS treatment period (2009) should be equal (i.e. M=0.38). With the foregoing assumption, the isolated impact of the MTS project on mean CHHFCI of MTS households is 0.049 (0.429-0.380=0.049) (i.e. the difference between post-MTS means for NsemSaw and that for the MTS group). A DID estimate of 0.049 suggests that 4.9% of the observed changes in mean CHHFCI among MTS households in Yaya may be attributed directly to the MTS project. Since the Non-MTS and NsemSaw groups registered mean CHHFCI of 0.392 and 0.391 respectively during the pre-MTS period, the ND between these two groups is the difference in their pre-MTS CHHFCI (0.392-0.391=0.001). A Normal Difference of 0.001 suggests that in the absence of the MTS project, the mean CHHFCI for Non-MTS households in Yaya would have been 0.001(equal to the mean for NsemSaw plus the ND). With the foregoing assumption, the spillover effect of the MTS project on mean CHHFCI of Non-MTS households is 0.040 (computed by first adding the ND to the post-MTS mean for NsemSaw and then subtracting the result from the post MTS means for the Non-MTS group [0.421-0.380+0.001] = 0.040). The 874 DID estimate of 0.040 suggests that 4.0% of the observed changes in mean CHHFCI among Non-MTS households in Yaya may be attributed to the MTS (figure 8.1.2). The 4.0% increase in mean CHHFCI among Non-MTS households may be considered a spillover effect since Non- MTS households were not specifically or directly targeted by the MTS project yet benefited somewhat indirectly from the MTS project by virtue of residing in close quarters with MTS households. Figure 8.1.2 below illustrates graphically the direct and spillover effects of the MTS project on CHHFCI. 875 Figure 8.1.2: DID Estimates of Direct and Spillover MTS Project Effects on CHHFCI 876 0.3910.4290.3800.3920.4210.3810.3700.3800.3900.4000.4100.4200.430Pre-MTS (1999)Post-MTS (2009)Combined Household Financial Capital Index(CHHFCI)Direct and Spillover MTS Project Effects on Combined Household Financial Capital IndexMTS (With Treatment)NsemsawMTS (Without Treatment)Non-MTS (Treatment)Non-MTS (With Treatment)Project Effect on MTS (0.049)Spillover Effect on Non-MTS(0.040)NormalDiff. between Non-MTS and NsemSaw(0.001) 8.2.1 Direct and Spillover MTS Project Effect on CHCI1 and CHCI2 This section discusses the direct and spillover MTS project effects of Household Human Capital Indexes one and two among MTS and Non-MTS households in Yaya communities. Figure 8.3 below helps determines the magnitude of the direct and spillover MTS project effects on CHCI1 among MTS and Non-MTS households between 1999 and 2009. The NsemSaw group was used as a control to compute the ND between MTS and NsemSaw as well as between Non- MTS and NsemSaw. By computing a ND for both MTS and Non-MTS groups in Yaya, the MTS project and spillover effects were determined while also accounting for possible effects of any omitted variables. According to figure 8.3 the MTS and NsemSaw groups experienced mean CHCI1 of 0.590 and 0.535 respectively during the pre-MTS period hence the ND between these two groups is 0.055 (i.e., 0.590 - 0.535). A 0.055 ND between MTS households in Yaya and NsemSaw suggests that in the absence of the MTS project the mean CHCI1 for MTS households in the post MTS treatment period (2009) is 0.587 (i.e., the mean CHCI1 of 0.532 for NsemSaw plus the 0.055 ND between both groups). With the foregoing assumption, the isolated impact of the MTS project on mean CHCI1 of MTS households is 0.039 (i.e., the difference between post-MTS means for MTS participant households (0.626) and 0.587). The DID estimate of 0.039 suggests that 3.9% of the observed changes in mean CHCI1 among MTS households in Yaya may be attributed directly to the MTS project (see figure 8.1.3 below). Applying a similar principle, DID estimate of 0.041 was computed for Non-MTS households in Yaya suggesting that 4.1% of the observed changes in CHCI1 may be attributed to spillover from the MTS project (see figure 8.1.4 below). Section 8.6 below discusses the direct and spillover MTS project effect on CHPCI1 and CHPCI2 indexes. 877 Figure 8.1.3: DID Estimates of Direct and Spillover MTS Project Effects on CHCI1 878 0.6260.5350.5320.5900.5870.6310.5930.5900.4800.5000.5200.5400.5600.5800.6000.6200.640Pre-MTS (1999)Post-MTS (2009)Combined Household Human Capital Index1(CHCI1)Direct and Spillover MTS Effects on Combined Human Capital Index1MTS (With Treatment)NsemSawMTS (Without Treatment)Non-MTS (Treatment)Non-MTS (With Treatment)Spillover Effect on Non-MTS(0.041)Project Effect on MTS(0.039)NormalDiff. between Non-MTS and NsemSaw(0.058)NormalDiff. between MTS and NsemSaw(0.055) Figure 8.1.4: DID Estimates of Direct and Spillover MTS Project Effects on CHCI2 879 0.5630.6260.5070.5310.5870.5700.6300.5940.5000.5200.5400.5600.5800.6000.6200.640Pre-MTS (1999)Post-MTS (2009)Combined Household Human Capital Index2(CHCI2)Direct and Spillover MTS Effects on Combined Human Capital Index2MTS (With Treatment)NsemSawMTS (Without Treatment)Non-MTS (Treatment)Non-MTS (With Treatment)Spillover Effect on Non-MTS(0.036)MTS Project Effect (0.039)NormalDiff. between Non-MTS and NsemSaw(0.063)NormalDiff. between MTS and NsemSaw(0.056) 8.2.2 Direct and Spillover MTS Project Effect on CHPCI1 and CHPCI2 During the pre-MTS period (1999) the MTS and NsemSaw groups experienced mean Combined Household Physical Capital Index (CHPCI1) of 0.313 and 0.0.269 respectively and a ND of 0.044 between both groups (i.e. 0.313 - 0.269). A 0.044 ND between MTS households in Yaya and the NsemSaw group suggests that in the absence of the MTS project the mean CHPCI1 for MTS households in the post-MTS treatment period (2009) would have been 0.430 (i.e., the mean CHPCI1 of 0.386 for NsemSaw plus the 0.044 ND between both groups). The foregoing assumption suggests that the isolated impact of the MTS project on mean CHPCI1 of MTS households is 0.077 (i.e., the difference between post-MTS means for MTS households (0.507) and 0.430). A DID estimate of 0.077 suggests that 7.7% of observed changes in mean CHPCI1 among MTS households in Yaya may be attributed directly to the MTS project (see figure 8.3 below). Similarly, a DID estimate of 0.069 was observed for Non-MTS households in Yaya suggesting that 6.9% of the observed changes in CHCI1 may be attributed to spillover from the MTS project (see figure 8.5 below). Figure 8.6 below shows similar trends in both direct MTS project and spillover effects of the MTS policy on CHPCI2 however the magnitude of the MTS policy effect on participant (8.0%) and non-participant households (7.3%) in Yaya were slightly higher for CHPCI2 relative to CHPCI1. Section 8.7 below discusses the direct and spillover MTS project effect on Aggregate Household Social Capital Index (AHHSCI) among MTS and Non-MTS households in Yaya. 880 Figure 8.1.5: DID Estimates of Direct and Spillover MTS Project Effects on CHPCI1 881 0.3130.5070.2690.3860.4300.5050.3190.4360.2500.3000.3500.4000.4500.5000.550Pre-MTS (1999)Post-MTS (2009)Combined Household Physical Capital Index 1(CHPC1)Direct and Spillover MTS Project Effects on Household Physical Capital 1MTS (With Treatment)NsemSawMTS (Without Treatment)Non-MTS (Treatment)Non-MTS (With Treatment)Project Effect on MTS (0.077)Spillover Effect on Non-MTS(0.069)NormalDiff. between Non-MTS and NsemSaw(0.050)NormalDiff. between MTS and NsemSaw(0.044) Figure 8.1.6: DID Estimates of Direct and Spillover MTS Project Effects on CHPCI2 882 0.3440.5540.2970.4270.4740.5520.3490.4790.2500.3000.3500.4000.4500.5000.5500.600Pre-MTS (1999)Post-MTS (2009)Combined Household Physical Capital Index 2(CHPCI2)Direct and Spillover MTS Project Effects on Households Physical Capital 2MTS (With Treatment)NsemSawMTS (Without Treatment)Non-MTS (Treatment)Non-MTS (With Treatment)Project Effect on MTS (0.080)Spillover Effect on Non-MTS(0.073)NormalDiff. between Non-MTS and NsemSaw(0.052)NormalDiff. between MTS and NsemSaw(0.047) 8.2.3 Direct and Spillover MTS Project Effect on AHHSCI During the pre-MTS period (1999), MTS and NsemSaw groups experienced mean AHHSCI of 0.508 and 0.500 respectively and a ND of 0.008 between both groups (i.e. 0.508 - 0.500). A 0.008 ND between MTS households in Yaya and the NsemSaw group suggests that in the absence of the MTS project the mean AHHSCI for MTS households in the post MTS treatment period (2009) would have been 0.526 (i.e., the mean post-MTS AHHSCI of 0.518 for NsemSaw plus the 0.008 ND between both groups). The foregoing assumption suggests that the isolated impact of the MTS project on mean AHHSCI of MTS households is 0.017 (i.e. the difference between post-MTS means for MTS participant households (0.543) and 0.526). The DID estimate of 0.017 suggests that 1.7% of observed changes in mean AHHSCI among MTS households in Yaya may be attributed directly to the MTS project (see figure 8.1.7 below). Unlike MTS households that experienced a positive ND and DID estimate, Non-MTS households in Yaya on the contrary experienced negative ND and DID estimates of -0.005 and - 0.012 respectively. The fact that the MTS policy provided a forum for program participant to meet regularly or periodically to discuss program issues and possibly any other matters outside of the MTS may help explain the results obtained for Non-MTS households in Yaya. While membership in religious organizations such as churches and mosques provide opportunities for community members to strengthen their AHHSCI, the positive influence of the MTS on AHHSCI as exemplified by periodic group meeting to find solutions to joint problems of forest management might in part explain why the MTS members experienced an increase in AHHSCI while Non-MTS households in the same communities experienced an average of 1.2% decline in AHHSCI (see figure 8.1.8 below). Section 8.8 below discusses the direct and spillover MTS project effect on CHHNCI1 among MTS and Non-MTS households in Yaya. 883 Figure 8.1.7: DID Estimates of Direct and Spillover MTS Project Effects on AHHSCI 884 0.5080.5430.5000.5180.5260.4950.5010.5130.4900.5000.5100.5200.5300.5400.550Pre-MTS (1999)Post-MTS (2009)Aggregate Household Social Capital Index(AHHSCI)Direct and Spillover MTS Project Effects on Aggregate Household Social Capital IndexMTS (With Treatment)NsemSawMTS (Without Treatment)Non-MTS (Treatment)Non-MTS (With Treatment)Spillover Effect on Non-MTS(-0.012)Project Effect on MTS(0.017)NormalDiff. between MTS and NsemSaw(0.008)NormalDiff. between Non-MTS and NsemSaw(-0.005) 8.2.4 Direct and Spillover MTS Project Effect on CHHNCI1 and CHHNCI2 Like the previous sections, figure 8.8 below uses the DID approach to determine the magnitude of direct and spillover MTS project effects on CHHNCI1and CHHNCI2 among MTS and Non-MTS households between 1999 and 2009. According to figure 8.1.8 the MTS and NsemSaw groups experienced mean CHHNCI1 of 0.541 and 0.577 respectively during the pre- MTS period with a ND of -0.036 (i.e., 0.541 - 0.577) between both groups. A -0.036 ND suggests that in the absence of the MTS project, the mean CHHNCI1 for MTS households in 2009 (i.e., post MTS treatment period) would have been 0.452 (i.e., the mean CHHNCI1 of 0.488 recorded for NsemSaw in 2009 minus the 0.036 ND between both groups). Figure 8.1.8 shows that in 2009 MTS households experienced a mean CHHNCI1 of 0.509 suggesting an MTS project effect of 5.7% (0.509 minus 0.452). Though a 0.509 mean CHHNCI1 recorded among MTS households in 2009 represents a 3.2% decline over the pre-MTS index of 0.541, figure 8.8 suggests that without the MTS project a much larger decline of 8.9% (i.e., pre-MTS value of 0.541 minus 0.452) in CHHNCI1 would have occurred among MTS households. The difference between what would have been the decline in CHHNCI1 without the MTS project (8.9%) and the actual decline of 3.2% represents the MTS policy effect of 5.7%. The foregoing analysis suggests that the MTS policy slowed down the decline in CHHNCI1 among MTS project households by 5.7%. A similar declining trend in CHHNCI1was observed among Non-MTS households in Yaya between 1999 and 2009 though the magnitude of the decline was much higher among this group. As shown in figure 8.8 a ND of -0.045 between Non-MTS and NsemSaw households suggests that without the MTS project CHHNCI1 among the Non-MTS group would have been 0.443 (i.e., CHHNCI1 of 0.483 among NsemSaw minus the ND of 0.045). A -0.045 ND suggests 885 Figure 8.1.8: DID Estimates of Direct and Spillover MTS Project Effects on CHHNCI1 886 0.5410.5090.5770.4880.4520.5320.4830.4430.4000.4500.5000.5500.600Pre-MTS (1999)Post-MTS (2009)Combined Household Natural Capital Index 1 (CHHNCI1)Direct and Spillover MTS Project Effects on Combined Household Natural Capital Index 1MTS (With Treatment)NsemSawMTS (Without Treatment)Non-MTS (Treatment)Non-MTS (With Treatment)PE on MTS (0.057)SoEon Non-MTS (0.040)PE= Project EffectSoE=Spillover EffectND=Normal DifferenceND between MTS and NsemSaw (-0.036)ND between Non-MTS and NsemSaw (-0.045) a 4% spillover effect on CHHNCI1among the Non-MTS group. Though a 0.483 mean CHHNCI1 recorded among Non-MTS households in 2009 represents a 4.9% decline over the pre-MTS index of 0.532, figure 8.1.8 suggests that without the MTS project, a much larger decline of 8.9% (i.e., pre-MTS value of 0.532 minus 0.443) in CHHNCI1 would have been observed among Non-MTS households. The difference between what would have been the decline in CHHNCI1 without the MTS policy (8.9%) and the actual decline of 4.9% represents the 4.0% spillover effect attributable to the MTS project. The foregoing analysis suggests that the MTS policy in effect slowed down the decline in CHHNCI1 among Non-MTS project households by 4.0%. Figure 8.9 below also shows similar trends in both direct MTS project and spillover effects of the MTS project on CHHNCI2 though the magnitude of the MTS policy effect on participant (6.2%) and non-participant households (4.3%) in Yaya were slightly higher for CHHNCI2 relative to CHHNCI1. Section 8.9 below presents a summary of the results presented in chapter eight and the implications for forest communities in which similar MTS projects are implemented. 887 Figure 8.9: DID Estimates of Direct and Spillover MTS Project Effects on CHHNCI2 888 0.5270.5990.5060.5580.4650.4980.5480.4550.4500.4700.4900.5100.5300.5500.5700.5900.610Pre-MTS (1999)Post-MTS (2009)Combined Household Natural Capital Index 2(CHHNCI2)Direct and Spillover MTS Project Effects on Household Natural Capital Index 2MTS (With Treatment)NsemSawMTS (Without Treatment)Non-MTS (Treatment)Non-MTS (With Treatment)PE on MTS (0.062)SoE on Non-MTS (0.043)ND between Non-MTS and NsemSaw (-0.051)ND between MTS and NsemSaw (-0.041)PE= Project EffectSoE=Spillover EffectND=Normal Difference 8.3 Summary, conclusions, and policy implications This chapter examined change in five livelihood assets between 1999 and 2009 and determined the magnitude of the change in each asset level among the MTS and Non-MTS groups in Yaya that may be attributed directly or indirectly to the MTS project. As the results of the DID analysis show, households experienced an increase in four out of the five household livelihood assets: Combined Household Financial, Human, Physical and Social Capital (CHHFCI, CHCI, CHPCI) and Aggregate Household Social Capital Index (AHHSC) between the pre (1999) and post (2009) MTS period and the DID analysis shows that the MTS project contributed positively to the observed changes between the two time periods. Contrary to the increasing trends observed for CHHFCI, CHCI, CHPCI and AHHSC, Combined Household Natural Capital Index (CHHNCI), declined between 1999 and 2009 among all the three research groups. The DID results however indicate that the decline in CHHNCI would have relatively larger had the MTS project not been implemented. The DID analysis under this chapter demonstrates how baseline data can be used to isolate both direct and indirect project impacts of a development project at different points in time. Hence, while there was no available baseline information available at the time of this study, this was generated from the research participants using recall information. Part of forest policy regarding the implementation of future MTS projects in other forest reserves in Ghana should include the implementation/establishment of baseline data on different livelihood assets as part of the MTS implementation plan. Based on the baseline livelihood status of household in MTS communities, project milestones would then be developed and used to periodically track both direct and spillover effects of the MTS reforestation program within a national forest reserve (as demonstrated in this chapter). 889 BIBLIOGRAPHY 890 BIBLIOGRAPHY Adesina, F. A. (1990). Planted Fallows for Sustained Fuelwood Supply in the Humid Tropics. Transactions of the Institute of British Geographers, 15(3), 323. doi:10.2307/622674 Adger, W. N., Kelly, P. M., Winkels, A., Huy, L. Q., & Locke, C. (2002). Migration, Remittances, Livelihood Trajectories, and Social Resilience. AMBIO: A Journal of the Human Environment, 31(4), 358-366. doi:10.1579/0044-7447-31.4.358 Agbeja, B. O. (2004). Valuation of Investment on Teak Plantation Establishment on a 15- Year- Rotation for Intending Small Holders in Nigeria (A Paper Presented at International Workshop on Small-Holder Timber Production, Rep.). Nairobi, Kenya: World Agroforestry Centre (ICRAF). Afful-Koomson, T. (2000). Implications of international timber certification for Ghanaian timber exports and sustainable forest management. Boston, MA: Fletcher School of Law and Diplomacy, Tufts University. Ampadu-Adjei, O. (1987). Dimensions of bush fire and management in Ghana. Survey Data. Accra: EPC. In Adam, I. (1996, October 1). Energy and deforestation in Ghana: a study of woodfuel-deforestation links in rural Ghana. Energy and deforestation in Ghana: a study of wood fuel-deforestation links in rural Ghana. Ann Arbor, MI: ProQuest Company. Amanor, K. S. (1999). Global Restructuring and Land Rights in Ghana: Forest Food Chains, Timber and Rural Livelihoods (Research Report No. 108). Sweden: Nordic Africa Institute's program on "The political and social context of Structural Adjustment in Sub-Saharan Africa. Afful-Koomson, T. (2000). Implications of international timber certification for Ghanaian timber exports and sustainable forest management. Boston, MA: Fletcher School of Law and Diplomacy, Tufts University. Agyarko, T (n.d). Forestry Outlook Study for Africa (FOSA), Forestry Outlook Study for Africa (FOSA). Accra, Ghana: Ministry of Lands and Forestry. Agyeman, V., Kasanga, K. R., Danso, E., Marfo, K. A., Asare, A., Yeboah, O. M., & Whiteman, A. (n.d.). Equitable Revenue Sharing From [Costs and Benefits Sharing On] Plantation Development: A Case Study Of Public, Private And Local Community Partnerships In Ghana (Rep.). Accra, Ghana: Forestry Commission. 891 Akpalu, W., & Parks, P. J. (2007). Natural resource use conflict: gold mining in tropical rainforest in Ghana. Environment and Development Economics, 12(1), 55–72. doi: 10.1017/s1355770x0600338x Ali, J., Benjaminsen, T. A., Hammad, A. A., & Dick, Ø B. (2005). The road to deforestation: An assessment of forest loss and its causes in Basho Valley, Northern Pakistan. Global Environmental Change, 15(4), 370-380. doi:10.1016/j.gloenvcha.2005.06.004 Allison, E. H., & Horemans, B. (2006). Putting the principles of the Sustainable Livelihoods Approach into fisheries development policy and practice. Marine Policy, 30(6), 757-766. doi:10.1016/j.marpol.2006.02.001 Amanor, K. S. (1999). Global Restructuring and Land Rights in Ghana: Forest Food Chains, Timber and Rural Livelihoods (Research Report No. 108). Sweden: Nordic Africa Institute's program on "The political and social context of Structural Adjustment in Sub-Saharan Africa. Amelung, T., & Diehl, M. (1992). Deforestation of tropical rain forests: economic causes and impact on development. Tübingen: Mohr. In Woollery, A. (1997). Tropical forest depletion: Assessing the forest products trade and economic influences on the forest area industrially logged. Ann Arbor, MI: ProQuest Company. Ampadu-Adjei, O. (1987). Dimensions of bush fire and management in Ghana. Survey Data. Accra: EPC. In Adam, I. (1996, October 1). Energy and deforestation in Ghana: a study of woodfuel-deforestation links in rural Ghana. Energy and deforestation in Ghana: a study of woodfuel-deforestation links in rural Ghana. Ann Arbor, MI: ProQuest Company. Asante, M. (2005). Deforestation in Ghana: Explaining the chronic failure of forest preservation policies in a developing country. Lanham, Boulder, New York, Toronto, Oxford: University Press of America. Ashurst, Morris Crisp and Company, (1911). Deposition on behalf of Amanhin and Ahinfu (Kings and Chiefs) of the Central and Western Provinces of the Gold Coast and the Aborigines Society. In Belfield, C. H. Secretary of State for the Colonies, (1912, p.103). Legislation governing the alienation of native lands in the Gold Coast colony and Ashanti: With some observations on the "Forest Ordinance" 1911 (Cd.6278). London, UK: Wynam and Sons Ltd. Baraclough, S., & Ghimire, (1995). Deforestation and historical perspective. In Forests and Livelihoods London, New York: MacMillan Press Ltd and St. Martin's Press Inc. Basu, J. P. (2012). Traditional Forest Related Knowledge Mitigation and Adaptation to Climate Change: A Study on Forest Dependent Communities in Drought Prone Areas of West 892 Bengal, India (S. Heok-Choh, L. Jinlong, Y. Shunbo, & S. A. Hamid, Eds.). In Traditional Forest-related Knowledge and Culture for Green Economy (pp. 64-70). Yangling, China: Yangling International Agricultural Forum. Baumann, P. (2002). Improving access to natural resources for the rural poor: A critical analysis of central concepts and emerging trends from a sustainable livelihoods perspective (LSP Working paper No. 1). Rome, Italy: FAO Livelihood Support Program (LSP). Beckett, I. F. W. (2009). Wolseley and Ashanti: the Asante war journal and correspondence of Major General Sir Garnet Wolseley 1873-1874. Stroud: History Press for the Army Records Society. Bedele, D. K. (1988). The Changing Use of Mountains by a Farming People: The Krobos of Ghana. Mountain Research and Development, 8(4), 297. doi:10.2307/3673551 Belfield, C. H. Secretary of State for the Colonies, (1912). Legislation governing the alienation of native lands in the Gold Coast colony and Ashanti: With some observations on the "Forest Ordinance" 1911 (Cd.6278). London, UK: Wynam and Sons Ltd. Benhin, J. K., & Barbier, E. B. (2004). Structural Adjustment Programme, Deforestation and Biodiversity Loss in Ghana. Environmental and Resource Economics, 27(3), 337– 366. doi: 10.1023/b:eare.0000017653.15107.0f Benneh, G. (1984, January). In Flt. Lt. J.J Rawlings (Chair). Suggestions for the management of Ghana's national resources. The thirty-fifth annual New Year school: The environment and national survival, Commonwealth Hall, University of Ghana. Bhatia, K. K. (1958). A Mixed Teak Forest of Central India. The Journal of Ecology, 46(1), 43. doi:10.2307/2256902 Blocher, J. (2006). Building on Custom: Land Tenure Policy and Economic Development in Ghana. Yale Human Rights and Development Law, 9, 177. Retrieved from https://papers.ssrn.com/sol3/papers.cfm?abstract_id=985353## Boahen, A. A. (2003). Yaa Asantewaa and the Asante-British War of 1900-1. Accra: Sub- Saharan. Boakye, K. A., & Baffoe, K. A. (2006). Trends in forest ownership, forest resource tenure and institutional arrangements. Case study of Ghana (FAO Regional Workshop on Trends in Forest Ownership, Forest Resource Tenure and Institutional Arrangements in Africa). Nukuru, Kenya: FAO. Boomgaard, P. (1992). Forest Management and Exploitation in Colonial Java, 1677-1897. Forest & Conservation History, 36(1), 4-14. doi:10.2307/3983978 893 Borggaard, O. K., Gafur, A., & Petersen, L. (2003). Sustainability Appraisal of Shifting Cultivation in the Chittagong Hill Tracts of Bangladesh. AMBIO: A Journal of the Human Environment, 32(2), 118-123. doi:10.1579/0044-7447-32.2.118 Bosetti, V., & Lubowski, R. N. (2010). Deforestation and climate change: reducing carbon emissions from deforestation and forest degradation. Cheltenham, UK: Edward Elgar. Bowyer, J. L., Fernholz, K., Lindburg, A., Howe, J., & Bratkovich, S. (2009). The power of silviculture: Employing thinning, partial cutting systems and other intermediate treatments to increase productivity, forest health and public support for forestry. Minneapolis, MN, USA: Dovetail Partners. Boyd, W. (2010). Ways of Seeing in Environmental Law: How Deforestation Became an Object of Climate Governance. Ecology Law, 37(843). Retrieved from https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1694612 Brown (1911). Deposition by Mr. Brown, President of the Aborigines Protection Society on behalf of Amanhin and Ahinfu (Kings and Chiefs) of the Central and Western Provinces of the Gold Coast. In Belfield, C. H. Secretary of State for the Colonies, (1912, p.111). Legislation governing the alienation of native lands in the Gold Coast colony and Ashanti: With some observations on the "Forest Ordinance" 1911 (Cd.6278). London, UK: Wynam and Sons Ltd. Brown, C. K. (1984). Social structure and poverty in selected rural communities in Ghana. ISSER Discussion Paper No.54: University of Ghana, Legon. Bryant, D., Nielsen, D., & Tangley, L. (1997). The last frontier forests: ecosystems and economies on the edge. Washington, D.C.: World Resources Institute. Bryant, R. L. (1993). Forest Problems in Colonial Burma: Historical Variations on Contemporary Themes. Global Ecology and Biogeography Letters, 3(4/6), 122. doi:10.2307/2997765 Bryant, R. L. (1994). From Laissez-faire to Scientific Forestry: Forest Management in Early Colonial Burma, 1826-85. Forest & Conservation History, 38(4), 160-170. doi:10.2307/3983602 Bryant, R. L. (1994). Shifting the Cultivator: The Politics of Teak Regeneration in Colonial Burma. Modern Asian Studies, 28(2), 225-250. doi:10.1017/s0026749x00012397 Bryant, R. L. (1996). Romancing Colonial Forestry: The Discourse of Forestry as Progress in British Burma. The Geographical Journal, 162(2), 169. doi:10.2307/3059874 Chakraborty, S., & Das, M. (2005). Mortality, Human Capital and Persistent Inequality. Journal of Economic Growth, 10(2), 159-192. doi:10.1007/s10887-005-1670-5 894 Chalk, L., Davy, B., Desch, H. E., & Hoyle, A. C. (1933). Twenty West African timber trees. Oxford: Oxford University Press. Carney, D. (1998). Sustainable rural livelihoods: what contribution can we make? London: Department for International Development. Carney, D. (2003). Sustainable livelihoods approaches: Progress and possibilities for change. London: DFID, Department for International Development. Carter, M. R., & Barrett, C. B. (2006). The economics of poverty traps and persistent poverty: An asset-based approach. Journal of Development Studies, 42(2), 178-199. doi:10.1080/00220380500405261 Currey, J. (1992). Africa from the seventh to the eleventh century (I. Hrbek, Ed.). Paris, France: UNESCO. Cattaneo, A. (2010). Incentives to reduce emissions from deforestation: a stock-flow approach with target reductions. In Bosetti and Lubowski (2010 pp. 93–120) Deforestation and Climate Change: Reducing Carbon Emissions from Deforestation and Forest Degradation. Cheltenham, UK: Edward Elgar Publications. Cattermoul, B., Townsley, P., & Campbell, J. (2008). Sustainable Livelihoods Enhancement and Diversification (SLED) : A Manual for Practitioners. Cambridge, UK: IUCN Gland, Switzerland and Colombo, Sri Lanka; CORDIO, Kalmar, Sweden; and ICRAN. Chambers, R. (1983). Rural development: Putting the last first. London, UK: Longman. Chambers, R. (2007). Poverty research: Methodologies, mindsets and multidimensionality. Brighton, U.K.: Institute of Development Studies (IDS). Chambers, R., & Conway, G. R. (1991). Sustainable rural livelihoods: Practical concepts for the 21st century (Discussion paper No. 296). Brighton, UK: Institute of Development Studies (IDS). Chambers, R., Longhurst, R., Bradley, D., & Feachem, R. (1979, February). Seasonal dimensions to rural poverty: Analysis and practical implications. Seasonal dimensions to rural poverty, Brighton, England. Chambers, R., Pacey, A., & Thrupp, L. A. (1989). Farmer first. UK: Intermediate Technology Publications. Chipp, T. F. (1922). The forest officer’s handbook of the Gold Coast, Ashanti and the Northern Territories. London, UK: Waterloo and Sons Limited. Chipp, T. F. (1927). The Gold Coast forest: A study in Synecology. In Oxford University Forestry School Memoirs Nos. 7-12. Oxford: Clarendon Press. 895 CSIR-FORIG. (2003). CSIR- Forestry Research Institute of Ghana: Annual Report (Rep.). FORIG. Culas, R. (2008). Non-forest sector policies, global environmental markets and the deforestation process. In I. Sanchez & C. Alonso (Eds.), Deforestation Research Progress. New York: Nova Science Publishers, Inc. Current, D., Lutz, E., & Scherr, S. J. (1995). The Costs And Benefits Of Agroforestry To Farmers. The World Bank Research Observer, 10(2), 151-180. doi:10.1093/wbro/10.2.151 Dadebo, M. A., & Shinohara, T. (1999). Forest Resources and Timber Production of Ghana: Current Instruments for Sustainable Development. Journal of Forest Research, 4(1), 1– 12. doi: 10.1007/bf02760317 Dakubu, M. E. K. (1988). The Languages of Ghana. London: Kegan Paul International for the International African Institute. In Fridy, K. S. (2007). "We only vote but do not know": the social foundations of partisanship in Ghana. S.l.: s.n. Damodaran, A., & Engel, S. (2003). Joint forest management in India: Assessment of performance and evaluation of impacts (Working paper No. 77). Bonn, Germany: Zentrum für Entwicklungsforschung Center for Development Research. Dasgupta, S., & Debnath, D. (2008). Community forest management in tribal states of India. Dehradun: International Book Distributors. Datta, S.K. and Sarkar, K (2010). Status of joint forest management in India: Socio-economic determinants of forest participation in a dynamic optimization setting. International Journal of Social Forestry, 3(2), 81-100 De Grassi, A. (2003). Environmental Governance in Africa: Constructing Subsidiarity, Consolidating Hegemony: Political Economy and Agro-Ecological Processes in Ghanaian Forestry (Working paper No. 13). Washington DC: World Resources Institute. Deacon, Robert T. “Deforestation and the Rule of Law in a Cross-Section of Countries.” Land Economics, vol. 70, no. 4, 1994, p. 414., doi:10.2307/3146638. Dei, G. J. (1992). A forest beyond the trees: Tree cutting in rural Ghana. Human Ecology, 20(1), 57-88. doi:10.1007/bf00889696 Dickson, K. B. (1969). A historical geography of Ghana. Cambridge: Cambridge University Press. In Owusu, J. H. (1993). Ghana's economic recovery program and changes in the structural and organizational patterns of the formal wood industry (1983-1991). Ann Arbor, MI: ProQuest. DOI: www.il.proquest.com 896 Dorward, A., Anderson, S., Clark, S., Keane, B., & Moguel, J. (2001). Asset Functions and Livelihood Strategies: A Framework for Pro-Poor Analysis, Policy and Practice (Working paper No. ADU). Kent, UK: Imperial College Wye. Dzokoto, G., & Opoku, K. (2010). Land tenure in Ghana: Making the case for incorporation of customary law in land administration and areas of intervention by the Growing Forest Partnership (International union for the conservation of nature and growing forest partnership). Edgerton, R. (1995). The fall of the Asante Empire: The hundred-year war for Africa's gold coast. New York, London, Toronto, Sydney, Tokyo, and Singapore: The Free Press. Egunjobi, J. K. (1974). Litter Fall and Mineralization in a Teak Tectona grandis Stand. Oikos, 25(2), 222. doi:10.2307/3543646 Elasha, B. O., Elhassan, N. G., Ahmed, H., & Zakieldin, S. (2005). Sustainable Livelihood Approach for Assessing Community Resilience to Climate Change: Case Studies from Sudan (Working paper No. 17). Assessments of Impacts and Adaptations of Climate Change (AIACC). Ellis-Jones, J., & Mason, T. (1999). Livelihood Strategies and Assets of Small Farmers in the Evaluation of Soil and Water Management Practices in the Temperate Inter-Andean Valleys of Bolivia. Mountain Research and Development, 19(3), 221-234. Retrieved July 1, 2010, from www.jstor.org/stable/25164028 England, P. (1993). Forest Protection and the Rights of Cocoa Farmers in Western Ghana. Journal of African Law, 37(2), 164-176. doi:10.1017/s0021855300011220 Ennion, R. C. (1998). Evaluation of Four Taungya Permanent Big-Leaf Mahogany Plots, Aged 36 to 38 Years, in Belize. Big-Leaf Mahogany Ecological Studies, 362-374. doi:10.1007/0-387-21778-9_19 Evans, J. (1982). Plantation forestry in the Tropics. Oxford: Clarendon Press. Ewusi, K. (1983). The regional dimensions of industrialization in Ghana. Institute of Statistical, Social and Economic Research (ISSER) Technical Publication No. 42, University of Ghana, Legon. Ewusi, K. (1983, November). The economy of Ghana in 1982-1983 and prospects for the future. Institute of Statistical, Social and Economic Research (ISSER), University of Ghana, Legon. Ewusi, K. Institute of Statistical, Social and Economic Research, University of Ghana, (1983). Accra: University of Ghana. 897 Ewusi, K. (1984a, April). The dimensions and characteristics of rural poverty in Ghana. Institute of Statistical, Social and Economic Research (ISSER) Technical Publication No. 43, University of Ghana, Legon. Ewusi, K. (1984b, November). The political economy of Ghana in the post-independence period: Description and analysis of the decadence of the political economy of Ghana and the survival techniques of her citizens. ISSER Discussion Paper No.14 24th anniversary proceeding of the Ghana academy of arts and sciences, University of Ghana, Legon. Ewusi, K. (1986, December). Economy trends in Ghana in 1984-1985 and prospects for 1986. Institute of Statistical, Social and Economic Research (ISSER), University of Ghana, Legon. Ewusi, K. (1988, December). Economy trends in Ghana in 1987-1988 and prospects for 1989. Institute of Statistical, Social and Economic Research (ISSER), University of Ghana, Legon. Ewusi, K. (1988). Trends in the economy of Ghana, 1986-1988. Institute of Statistical, Social and Economic Research (ISSER), University of Ghana, Legon. Ewusi, K. (1989, October). The impact of structural adjustment on the agricultural sector in Ghana. Institute of Statistical, Social and Economic Research (ISSER), University of Ghana, Legon. FAO. (1993). Forest Resources Assessment 1990: Tropical countries (FAO Forestry Paper 112). In Leach, M., & Fairhead, J. (2000). Challenging Neo-Malthusian Deforestation Analyses in West Africa's Dynamic Forest Landscapes. Population and Development Review, 26(1), 17-43. doi:10.1111/j.1728-4457.2000.00017.x FAO. (2006). Global Forest Resources Assessment 2005: Progress towards sustainable forest management (FAO Forestry Paper. No. 147). Rome, Italy: FAO. Fearnside, P. M. (2004). Are climate change impacts already affecting tropical forest biomass? Global Environmental Change, 14(4), 299–302. doi:10.1016/j.gloenvcha.2004.02.001 Fearnside, P. M., & Laurance, W. F. (2004). Tropical Deforestation and Greenhouse-Gas Emissions. Ecological Applications, 14(4), 982–986. doi: 10.1890/03-5225 Forestry Commission. (2002). Benefit Sharing Agreement for Modified Taungya Forest Plantations (Rep.). Accra, Ghana: Forestry Commission. Forestry Department. (n.d.). Annual Reports of Forestry Department, Ghana (1949-1994). (Rep.). Accra, Ghana: Forestry Commission. In Dadebo, M. A., & Shinohara, T. (1999). Forest Resources and Timber Production of Ghana: Current Instruments for Sustainable Development. Journal of Forest Research, 4(1), 1–12. doi: 10.1007/bf02760317 Fouracre, P. (2001). Transportation and Sustainable Rural Development (Rep.). TRL Limited. 898 Frankenberger, T. R., Drinkwater, M., & Maxwell, D. (2002). Operationalizing Household Livelihood Security: A Holistic Approach for Addressing Poverty and Vulnerability (Rep.). Georgia, USA: CARE International. Gajaseni, J. (1992). Socio-economic aspects of taungya. In C. F. Jordan, J. Gajaseni, & H. Watanabe (Eds.), Taungya: Forest Plantations with Agriculture in Southeast Asia (pp. 36-62). Wallingford, U.K.: C. A. B. International. Gajaseni, J., & Jordan, C. F. (1990). Decline of Teak Yield in Northern Thailand: Effects of Selective Logging on Forest Structure. Biotropica, 22(2), 114. doi:10.2307/2388402 George, S. (1990). A fate worse than debt. New York: Grove Weindenfeld in Shah, A. (2010, July 20). Structural Adjustment-a Major Cause of Poverty. Retrieved February 12, 2013, from https://www.globalissues.org/article/3/structural-adjustment-a-major-cause-of- poverty. Ghana Living Standards Survey (GLSS) 5+ (Rep.). (2000). Accra, Ghana: Institute of Statistical, Social and Economic Research. Ghana Living Standards Survey (GLSS) 5+ (Rep.). (2008). Accra, Ghana: Institute of Statistical, Social and Economic Research. Ghana Statistical Service (2013). 2010 Population & Housing Census National Analytical Report. Retrieved August 12, 2013, from: https://statsghana.gov.gh/gssmain/fileUpload/pressrelease/2010_PHC_National_Analytic al_Report.pdf Gill & Duffus Group (n.d). Cocoa Statistics. In Kolavalli, S., & Vigneri, M. (2011). Cocoa in Ghana: Shaping the success of an economy. In P. Chuhan-Pole & M. Angwafo (Authors), Yes Africa can: Success stories from a dynamic continent (pp. 201-217). Washington, D.C.: The World Bank. Gillis, M., & Rapetto, R. (1988). Deforestation and government policy. International Center for Economic Growth. Occasional papers, no.8, ICS Press, San Francisco, CA. Gordon, A., & Craig, C. (2001). Rural non-farm activities and poverty alleviation sub-Saharan Africa (Policy Series Rep. No. 14). Hathan, UK: Natural Resources Institute. Grainger, A. (1993). Controlling tropical deforestation. London, UK: Eathscan Publications Ltd. Guppy, N. (1984). Tropical Deforestation: A Global View. Foreign Affairs, 62(4), 928–965. doi: 10.2307/20041914 Guha, R., & Gadgil, M. (1989). State Forestry and Social Conflict in British India. Past and Present, 123(1), 141-177. doi:10.1093/past/123.1.141 899 Haskett, N. (1981). Kete-Krakye and the Middle Volta Basin, 1700-1914: Cockpit of African and European rivalry pt.2. Ann Arbor, MI, London, England: University Microfilms International. Hawthorne, W., & Abu-Juam, M. (1993). Forest protection in Ghana with particular reference to vegetation and plant species. In Hawthorne, W., & Abu-Juam, M. (1995) Forest protection in Ghana: With particular reference to vegetation and plant species. Gland: IUCN in collaboration with ODA and the Forest Department Republic of Ghana. Hayford, J. E. C. (1970). Gold Coast Native Institutions: With thoughts upon a healthy imperial policy for the Gold Coast and Ashanti. London, UK: Frank Cass and Company Ltd. Hilhorst, T., & Porchet, N. (2012). Food Security and Land Governance Fact Sheet: Ghana. The Netherlands: Royal Tropical Institute. Hilson, G. M., 2004 Structural Adjustment in Ghana: Assessing the Impacts of Mining-Sector Reform Author(s): Gavin M. Hilson Source: Africa Today, Vol. 51, No. 2 (Winter, 2004), pp. 53-77. Published by: Indiana University Press Stable URL: http://www.jstor.org/stable/4187650. Accessed: 09/10/2011 14:02 Hoddinott, J., & Yohannes, Y. (2002). Dietary Diversity as a Household Food Security Indicator. (Rep. No. Food and Nutrition Technical Notes No. 4). Washington, DC., USA: USAID. Horton, A. B. (1970). E. Ayandele (Ed.), Letters on the condition of the Gold Coast: Since the exchange of territories between the English and the Dutch Governments on January 1, 1868 together with a short account of the Asante War, 1862-4, and the Awoonah War, 1866 London, UK: Frank Cass and Company Ltd. Horton, J. A B (1868). West African Countries and Peoples. London, Edinburgh: Reprinted with an introduction by George Shepperson (1969). Edinburgh: Univ. Press. Hussein, K., & Nelson, J. (n.d.). Sustainability Livelihoods and Livelihood Diversification (Working paper No. 69). ODA - Institute of Development Studies and the Poverty Research Unit - University of Sussex. Ibhawoh, B. (1999). Structural Adjustment, Authoritarianism and Human Rights in Africa. Comparative Studies of South Asia, Africa and the Middle East, 19(1), 158– 167. doi: 10.1215/1089201x-19-1-158OAU, 2002, p. 31-32 Imbens, G., & Wooldridge, J. M. (2007). What's new in econometrics? S.l.: NBER. Imo, M. (2009). Interactions amongst trees and crops in Taungya systems of western Kenya. Agroforestry Systems, 76(2), 265-273. doi:10.1007/s10457-008-9164-z James Currey, California, & UNESCO, (1990). European partition and conquest of Africa: an overview. In A. Boahen (Ed.), General History of Africa. VII Africa under colonial 900 domination 1880-1935: James Currey Ltd., London, California, Berkeley & UNESCO, Paris. Jansen, H. G., Siegel, P. B., & Pichon, F. (2005). Identifying the Drivers of Sustainable Rural Growth and Poverty Reduction in Honduras (DSGD Discussion Paper No. 19). Washington, DC, USA: International Food Policy Institute (IFPRI). Janvry, A. D., & Sadoulet, E. (2000). Rural poverty in Latin America. Food Policy, 25(4), 389- 409. doi:10.1016/s0306-9192(00)00023-3 Jordan, C. F., Gajaseni, J., & Watanabe, H. (1992). Taungya: Forest plantations with agriculture in Southeast Asia. Oxon: C.A.B. International. Kanji, N., & Barrientos, S. (2002). Trade liberalization, poverty and livelihoods: Understanding the linkages. (Working paper No. 159). Sussex, U.K: Institute of Development Studies (IDS). Kerr, J., Milne, G., Chhotray, V. et al. Managing Watershed Externalities in India: Theory and Practice. Environ Dev Sustain 9, 263–281 (2007). https://doi.org/10.1007/s10668005- 9022-3 Kimble, D. (1963). A political history of Ghana the rise of the Gold Coast nationalism 1850- 1928. Oxford: Clarendon Press. Kinloch, D. (1945). Silvicultural notes on some of the more important Gold Coast trees. Accra: Government printing Dept. KJÆR, E. D., & Foster, G. S. (1996). The Economics of Tree Improvement of Teak (Tectona Grandis L.) (Rep.). Arboretum, Royal Veterinary and Agricultural University, Denmark and USDA Forest Service, USA. Kollmair, M., & St. Gamper, J. (2002). The sustainable livelihood approach. Input Paper for Integrated Training Course of NCCR North-South Aeshiried, Development Study Group University of Zurich, Switzerland. Kotey, E.N.A., J. Francois, J.G.K. Owusu, R. Yeboah, K.S. Amanor and L. Antwi. 1998. Falling into place: policy that works for forests and people. Series No. 4. London: International Institute for Environment and Development, pp. 80–94. Krantz, L. (2001). The Sustainable Livelihood Approach to Poverty Reduction: An Introduction (Rep.). Swedish International Development Cooperation Agency: Division for Policy and Socio-Economic Analysis. 901 Krishna, A. (2004). Understanding, measuring and utilizing social capital: Clarifying concepts and presenting a field application from India. Agricultural Systems, 82(3), 291-305. doi:10.1016/j.agsy.2004.07.003 Kufuor, K. O. (2000). Forest management in Ghana: Towards a sustainable approach. Journal of African Law, 44(1), 52-64. doi:10.1017/s0021855300012031 Kusimi, J. M. (2008). Assessing land use and land cover change in the Wassa West District of Ghana using remote sensing. GeoJournal, 71(4), 249–259. doi: 10.1007/s10708- 008- 9172-6 Kwakye-Nuako, K. (1999). Sika ‘Dwa Kofi invades Ghana politics: A study of the religious dimensions of the National Liberation Movement (NLM), 1954-1957. Ann Arbor, MI.: A Bell and Howell Company. DOI: www.bellhowell.infolearning.com Kyem, P. A. K. (1997). A GISs-based strategy for improving local community participation in resource management, allocation and planning: The case of institution building for collaborative forest management in Southern Ghana. Ann Arbor, MI.: ProQuest Company. Jindal, R., Kerr, J. M., & Carter, S. (2012). Reducing Poverty Through Carbon Forestry? Impacts of the N’hambita Community Carbon Project in Mozambique. World Development, 40(10), 2123–2135. doi: 10.1016/j.worlddev.2012.05.003 Lawson, S., & MacFaul, L. (2010). Illegal logging and related trade: indicators of the global response. London: Chatham House. Leach, M., & Fairhead, J. (2000). Challenging Neo-Malthusian Deforestation Analyses in West Africa's Dynamic Forest Landscapes. Population and Development Review, 26(1), 17-43. doi:10.1111/j.1728-4457.2000.00017.x Leonard, L., R. Kopp, and N. Purvis (2010). International Forest Carbon in the U.S. Congress: A Survey of Key Congressional Staff. In Deforestation and Climate Change: Reducing Carbon Emissions from Deforestation and Forest Degradation, edited by V. Bosetti and R. Lubowski. Cheltenham, UK: Edward Elgar. Lin, N. (1999). Building a Network Theory of Social Capital. Connections, 22(1), 28–51. doi: 10.4324/9781315129457-1 Lindquist, E. J., D’annunzio, R., Gerrand, A., Macdicken, K., Achard, F., Beuchle,, R., . . . Stibig, H. (2012). Global forest land-use change, 1990-2005. Rome: Food and Agriculture Organization of the United Nations. Lopez, R., & Galinato, G. I. (2005). Trade Policies, Economic Growth, and the Direct Causes of Deforestation. Land Economics, 81(2), 145–169. doi: 10.3368/le.81.2.145 902 Lowe, R. G. (1987). Development of Taungya in Nigeria. In H. L. Gholz (Ed.), Agroforestry: Realities, possibilities and potentials. (pp. 137-154). Dordrecht, The Netherlands: Martinus Nijhoff in cooperation with ICRAF. Malleson, R., Asaha, S., Sunderland, T., Burnham, P., Egot, M., Obeng-Okrah, K., . . . Miles, W. (2008). A Methodology for Assessing Rural Livelihood Strategies in West/Central Africa: Lessons from the Field. Ecological and Environmental Anthropology, 4(1). Maloka, E. (2002). Africa's Development Thinking since Independence: A Reader. Pretoria, South Africa: Africa Institute of South Africa. Marshal, R. C. Forestry Department, (1945). Forestry in the Northern Territories of the Gold Coast (111). Accra: Government Printing Department. Maung, T. M., & Yamamoto, M. (2008). Exploring the Socio-Economic Situation of Plantation Villagers: A Case Study in Myanmar Bago Yoma. Small-scale Forestry, 7(1), 29-48. doi:10.1007/s11842-008-9039-1 May, C. A., & Schoone, P. (1989, March). Structural adjustment in theory and practice: With reference to Ghana. McCarthy, M. (1983). Social change and the growth of British power in the Gold Coast: The Fante States 1807-1874. Lanham, MD: University Press of America. Mends, T. M., & De Meijere, J. (2006). A Study of the Institution of the Customary Land Tenure System in the Supply of Property Rights for Urban Development – An Example of Accra, Ghana. In Promoting Land Administration and Good Governance. (5th). Accra, Ghana. Menzies, N. (1988). Three hundred years of Taungya: A sustainable system of forestry in South China. Human Ecology, 16(4), 361-376. doi:10.1007/bf00891648 Messer, N., & Townsley, P. (2003). Local institutions and livelihoods: Guidelines for analysis. Rome: FAO. Mgeni, A. S. (1992). Farm and Community Forestry (Village Afforestation) Program in Tanzania: Can It Go beyond Lipservice? Ambio, 21(6), 426-430. Retrieved September 12, 2008, from http://www.jstor.org/stable/4313975 Milton, R. (1994). Some issues and observations arising from District Taungya studies … for use in discussion for development of FP4 – Forest Reserve Rehabilitation (Rep. No. FP4). Accra, GA: Forest Commission. In Agyeman, V. K., Kasanga, M. K., Danso, E., Yeboah, O. M., & Agyeman, F. (2003). Revising the Taungya plantation system: New revenue- sharing proposals from Ghana. Unasylva (An International Journal of Forestry and Forest Industries), 54(212), 40-43. 903 Mincer, J. (1996). Economic development, growth of human capital, and the dynamics of the wage structure. Journal of Economic Growth, 1(1), 29-48. doi:10.1007/bf00163341 Moloney, C. A. (1887). Sketch of the forestry of West Africa to its present commercial products. London: S. Low, Marston, Searle, & Rivington. Moran, E. F., & Ostrom, E. (2005). Seeing the forest and the trees: human-environment interactions in forest ecosystems. Cambridge, MA: Mit Press. Moser, M. O. (2006). Asset-based Approaches to Poverty Reduction in a Globalized Context: An introduction to asset accumulation policy and summary of workshop findings (Global Economy and Development Working Paper). Washington DC, USA: The Brookings Institute. Mulligan, C. B., & Sala-I-Martin, X. (2000). Measuring Aggregate Human Capital. Journal of Economic Growth, 5(3), 215-252. doi:10.1023/a:1009893417085 Naiman, R., & Watkins, N. (1999). A survey of the impacts of IMF structural adjustment in Africa: Growth, social spending, and debt relief. Washington, D.C.: Preamble Center.NFPDP. (2004). The Modified Taungya System and Private Developers (Rep.). Accra, Ghana: National Forest Plantation Development Program. Nicol, A. (2000). Adopting a Sustainable Livelihoods Approach to Water Projects: Implications for Policy and Practice (Working paper No. 133). London, UK: Overseas Development Institute (ODI). Nyanteng, V. K. (1980). In Ampem II Nana Wereko (Chair). Why is the cocoa industry declining and what can be done to resuscitate it? In Fordjor, P.K (Eds.), Restructuring the Economy of Ghana (pp.72-82). Accra: University of Ghana. Odoom, F. (2002). Hardwood plantations in Ghana (Forest Plantations Working Paper No. 24). Rome, Italy: FAO Forest Resources Division. OERL. (2008). OERL Definitions of terms. Retrieved 12/11/2008, from http://oerl.sri.com/definitions.html. Ollennu, N. A. (1962). Principles of customary land law in Ghana. London: Sweet & Maxwell. In Mends, T. M., & De Meijere, J. (2006). A Study of the Institution of the Customary Land Tenure System in the Supply of Property Rights for Urban Development – An Example of Accra, Ghana. In Promoting Land Administration and Good Governance. (5th). Accra, Ghana. Olukoshi, O. A. (1999). State, conflict, and democracy in Africa: The complex process of renewal. In R. Joseph (Ed.), State, Conflict and Democracy in Africa (pp. 451-465). Boulder: Lynne Rienner. 904 Olukoshi, O. A. (2000). Structural Adjustment and Social Policies in Africa: Some Notes. In Draft Paper Prepared for the GASPP 4 Seminar on Global Social Policies and Social rights. 8 - 10 November. New Delhi, India: GASPP 4. Omosini, O. (1972). The Gold Coast Land Question, 1894-1900: Some Issues Raised on West Africa’s Economic Development. The International Journal of African Historical Studies, 5(3), 453. doi:10.2307/217095 Opoku, K. (2006). Forest governance in Ghana: An NGO perspective. Brussels: FERN. Owusu, J. H. (1993). Ghana's economic recovery program and changes in the structural and organizational patterns of the formal wood industry (1983-1991). Ann Arbor, MI: ProQuest. DOI: www.il.proquest.com Owusu, J. H. (1998). The myth of development under ghana. In B. Adjibolosoo & B. Ofori- Amoah (Eds.), Addressing Misconceptions about Africa's Development: Seeing beyond the veil. Studies in African Economic and Social Development (1 ed., Vol. 11, pp. 193 216). Lewiston, Queenston, Lampeter: The Edwin Mellen Press. Owusu, J. H. (2001). Economic recovery, debt and dependence in African economies: the case of Ghana. In F. Columbus (Ed.), Politics and Economics of Africa (1 ed., Vol. 1, pp. 1-39). Huntington, New York: Nova Science Publishers, Inc. Owusu, J. H. (2006). Economic/political convenience and unprecedented deforestation in sub- Saharan Africa: Development lessons from Ghana. In K. Konadu-Agyemang & K. Pandford (Eds.), Africa's Development in the Twenty-first Century: Penitent Socio- Economic and Development Issues. Hampshire, England: Ashgate Publishing Limited. Owusu, J. H. (2010). Trends in Ghana's International Timber Trade: Implications for Local Livelihoods and Sustainable Forest Management (L. A. German, A. Karsenty, & A. Tiani, Eds.). In Governing Africa's forests in a globalized world (pp. 261-283). London: Earthscan. Palupi, E. R., & Owens, J. N. (1997). Pollination, Fertilization, and Embryogenesis of Teak (Tectona grandis L. f.). International Journal of Plant Sciences, 158(3), 259-273. doi:10.1086/297437 Pandey, N. (2005). Monitoring the Impact of Joint Forest Management on Rural Livelihoods (Rep. No. PGP-I). Rajastan and New Delhi, India: Aravali Institute of Management and Winrock International. Parren, M. P. (1994). French and British colonial forest policies: Past and present implications for Côte d'Ivoire and Ghana. Boston, MA: African Studies Center, Boston University. Parren, M., & de Graaf, N. R. (1995). The quest for natural forest management in Ghana, Cote D'Ivoire and Liberia. Wegeningen: Tropenbos Foundation. 905 Pearce, D. W., & Brown, K. (1994). The causes of tropical deforestation: the economic and statistical analysis of factors giving rise to the loss of the tropical forests. London: UCL Press. Peluso, N. L. (1991). The History of State Forest Management in Colonial Java. Forest & Conservation History, 35(2), 65-75. doi:10.2307/3983940 Peluso, N. (1993). Traditions of Forest Control in Java: Implications for Social Forestry and Sustainability. Global Ecology and Biogeography Letters. Vol. 3, No. 4/6, The Political Ecology of Southeast Asian Forests: Transdisciplinary Discourses (Jul. - Sep. - Nov., 1993), pp. 138-157 Peprah, K. K. (2008, September). Impact of Mining Activities on Deforestation In The Mining Communities Of Ghana: The Case Of Obuasi. In Awudi, G. B.K. (2002). CCNM Global Forum on International Investment. Conference on Foreign Direct Investment and the Environment Lessons to be Learned from the Mining Sector. In The role of foreign direct investment (FDI) in the mining sector of Ghana and the environment. Paris, France:OECD. Perman, R., Ma, Y., Common, M., Maddison, D., & Mcgilvray, J. (2003). Natural Resource and Environmental Economics. Harlow, United Kingdom: Addison Wesley. Postel, S., & Heise, L. (1988). Reforesting the earth. (83 ed.). Washington D.C: Worldwatch Institute. Portes, A. (1998). Social Capital: Its Origins and Applications in Modern Sociology. Annual Review of Sociology, 24(1), 1–24. doi: 10.1146/annurev.soc.24.1.1 Robbins, R. H. (1999). Global problems and the culture of capitalism. Boston: Allyn and Bacon in Shah, A. (2010, July 20). Structural Adjustment-a Major Cause of Poverty. RetrievedFebruary 12, 2013, from https://www.globalissues.org/article/3/structural- adjustment-a-major-cause-of-poverty. Ross, P. (1958). Herbicide Control of Coppicing in Teak. Botanical Gazette, 120(1), 59-61. doi:10.1086/336002 Ross, P. (1959). Teak in Trinidad. Economic Botany, 13(1), 30-40. doi:10.1007/bf02860123 Ross, P. (1961). The Plant Ecology of the Teak Plantations in Trinidad. Ecology, 42(2), 387-398. doi:10.2307/1932090 Rugmini, P., & Jayaraman, K. (2008). Intrinsic units of growth for teak trees. Trees, 23(1), 51- 58. doi:10.1007/s00468-008-0253-4 906 Sanderson, D. (2000). Cities, Disasters and Livelihoods. Risk Management, 2(4), 49-58. doi:10.1057/palgrave.rm.8240068 Saatchi, S. S., Harris, N. L., Brown, S., Lefsky, M., Mitchard, E. T., Salas, W., . . . Morel, A. (2011). Benchmark map of forest carbon stocks in tropical regions across three continents. Proceedings of the National Academy of Sciences, 108(24), 9899-9904. doi:10.1073/pnas.1019576108 Seini, W., Howell, J., & Commander, S. (1987, September). Agricultural policy adjustment in Ghana. Overseas Development Institute Conference: The design and impact of adjustment programs in agriculture and agricultural institutions. Sen, A. (1984, June). Beyond liberalization: Social Opportunity and Human capacity. D.T. Lakdwala Memorial Lecture Series, New Delhi, India. Sen, A. (1995). Choice, welfare and development. In K. Basu, P. Pattanaik & K. Suzumura (Eds.). Oxford: Clarendon Press. Sen, A., & Foster, J. E. (1997). On economic inequality. Oxford: Oxford Univ. Press. Sen, A. (1999). Commodities and capabilities. Oxford: Oxford University Press. Sen, A. (1981). Poverty and famines: An essay on entitlement and deprivation. Oxford: clarendon. Sen, A. (1985). A Sociological Approach To The Measurement Of Poverty: A Reply To Professor Peter Townsend. Oxford Economic Papers, 37(4), 669-676. doi:10.1093/oxfordjournals.oep.a041716 Sen, A. (1987). The Standard of Living. Cambridge, U.K: Cambridge University Press. Serrat, O. (2008). The Sustainable Livelihoods Approach (Rep.). Washington, DC: Asian Development Bank. Sarris, A., & Shams, H. (1991). Ghana under structural adjustment: The impact of agriculture and the rural poor. New York City: New York University Press. Siegel, P. B. (2005). Using an Asset-Based Approach to Identify Drivers of Sustainable Rural Growth and Poverty Reduction in Central America: A Conceptual Framework (World Bank Policy Research WP No. 3475). Washington, D.C., USA: The World Bank. Scoones, I. (1998). Sustainable Rural Livelihoods: A Framework for Analysis (Working paper No. 72). Brighton, U.K.: Institute of Development Studies (IDS). 907 Shah, A. (2010, July 20). Structural Adjustment-a Major Cause of Poverty. Retrieved February 12, 2013, from https://www.globalissues.org/article/3/structural-adjustment-a-major- cause-of-poverty. Sarpong, G. A. (2006). Improving tenure security for the rural poor Ghana – country case study. In Support to the Legal Empowerment of the Poor (LEP) Improving tenure security for the rural poor – with the financial assistance of Norway under a Program Cooperation Agreement (LEP Working Paper # 2). Rome: FAO. Sanwal, M. (1988). Community Forestry: Policy Issues, Institutional Arrangements, and Bureaucratic Reorientation. Ambio, 17(5), 342-346. Retrieved September 12, 2008, from: www.jstor.org/stable/4313491 Schlönvoigt, A., & Beer, J. (2001). Initial growth of pioneer timber tree species in a Taungya system in the humid lowlands of Costa Rica. Agroforestry Systems, 51(2), 97-108. doi:10.1023/A:1010674402907 Spate, O. H. (1945). The Burmese Village. Geographical Review, 35(4), 523. doi:10.2307/210794 Siegel, P. B., & Alwang, J. (1999). An asset-based approach to social risk management: A conceptual framework (Discussion paper No. 9926). Washington, DC: Human Development Network, Social Protection Unit of The World Bank. Sahn, D. E., & Stifel, D. (2000). Assets as a Measure of Household Welfare in Developing Countries (Working paper No. 00-11). St. Louis, USA: Center for Social Development - Washington University. Sivaramakrishnan, K. (1995). Colonialism and Forestry in India: Imagining the Past in Present Politics. Comparative Studies in Society and History, 37(1), 3-40. doi:10.1017/s0010417500019514 Sayer, J., Harcourt, C. S., & Collins, N. M. (1992). The Conservation atlas of tropical forests: Africa. Macmillan. In Leach, M., & Fairhead, J. (2000). Challenging Neo-Malthusian Deforestation Analyses in West Africa's Dynamic Forest Landscapes. Population and Development Review, 26(1), 17-43. doi:10.1111/j.1728-4457.2000.00017.x StatPac. (2008). Questionnaires and survey design. Retrieved 12/11/2008 from http://www.statpac.com/surveys. Swindale, A., & Bilinsky, P. (2006). Household Dietary Diversity Score (HDDS) for Measurement of Household Food Access: Indicator Guide (2nd ed.). Washington, DC, USA: FHI 360/FANTA. 908 Thompson, H. N. (1910). Report on forests, Colonial Reports - Miscellaneous (Rep. No. 66). London: His Majesty's Stationary Office. Takeda, S. (1992). Origins of Taungya. In J. Gajaseni, C. F. Jordan, & H. Watanabe (Eds.), Taungya: Forest Plantations with Agriculture in Southeast Asia (pp. 9-17). Wallingford, U.K.: C. A. B. International. Tockman, J (2002). Cote d’ Ivoire: IMP, Cocoa, coffee, logging and mining. World Rainforest Mov Bull 54 (January). Available on: http.wrm.org.uy/54/CoteIvoire.html. Tockman, J. (2002, April 30). How the IMF is Destroying the World's Forests. Earth First.Usher, A., Paranjpye, V., & Ramnath, M. (1999). Forests for the future: Local strategies for forest protection, economic welfare and social justice. London, UK: Zed Books. DOI: www.zedbooks.demon.co.uk United Nations Industrial Development Organization (UNIDO). Regional and Country Studies Branch, (1986). Industrial development review series: Ghana (PPD.18). UNDP-SEF, (2008). Joint Evaluation of the Global Environmental Facility Small Grant Program (Evaluation Report No. 39). Washington, DC, USA: UNDP GEF Evaluation Office. Unknown Author, (u.d). Difference-in-Difference: Method of Economic Investigation. Lecture 10. Lecture. Retrieved on Mar 31, 2015, from: https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved=2ahUK EwifnJGekpnvAhUaZc0KHVViAzYQFjAAegQIARAD&url=https%3A%2F%2Fwww.rese archgate.net%2Fprofile%2FZoubair_Boulahia%2Fpost%2FIs-it-necessary-to-test-for- parallel-trends-while-using-difference-in-difference- method%2Fattachment%2F59d638ec79197b80779961ce%2FAS%253A3993794259722 24%25401472230883669%2Fdownload%2FMEI_iyengar_DinDPanel.ppt&usg=AOvVa w34FfKvy92G_ChWFNHk5Wfx. Vemuri, A (2008). Joint Forest Management in India: An unavoidable and conflicting common property regime in natural resource management. Journal of Development and Social Transformation, 5:80-90. Visseren-Hamakers, I. J., & Glasbergen, P. (2007). Partnerships in forest governance. Global Environmental Change, 17(3-4), 408-419. doi:10.1016/j.gloenvcha.2006.11.003 Ward, W. E. F. (1948). A history of the Gold Coast. London, UK: George Allen and Unwin Limited. Wardell, A. D. (2005). Collision, collusion and muted resistance - contrasting early and later encounters with empire forestry in the Gold Coast, 1874-1957 (Historical and Contemporary Studies (October 2005), Working paper No. 8). University of Copenhagen. 909 Wardell, D. A., Reenberg, A., & Tøttrup, C. (2003). Historical footprints in contemporary land use systems: Forest cover changes in savannah woodlands in the Sudano-Sahelian zone. Global Environmental Change, 13(4), 235-254. doi:10.1016/s0959- 3780(03)00056-6 Wealth in Wood (1950). Wealth in wood: A brief description of forestry and the timber industry in the gold coast. Gold Coast Public Relations. Watford: Rembrandt Photogravure Ltd. Wolseley, G. (2009). Wolseley and Ashanti: The Asante war journals and correspondence of Major General Sir Garnet Wolseley 1873-1874. In I. Beckett (Ed.), England: The History Press and Army Records Society. Watanabe, H., Sahunalu, P. & Khemnark, C. (1988). Combinations of trees and crops in the Taungya method as applied in Thailand. Agroforest System 6, 169–177. https://doi.org/10.1007/BF02344754 Watanabe, H. (1992). ‘Tree-crop interactions in Taungya plantations. In C. F. Jordan, J. Gajaseni, & H. Watanabe (Eds.), Taungya: Forest Plantations with Agriculture in Southeast Asia (pp. 32-43). Wallingford, U.K.: C. A. B. International. Whitehead, A. (2002). Tracking Livelihood Change: Theoretical, Methodological and Empirical Perspectives from North-East Ghana. Journal of Southern African Studies, 28(3), 575- 598. doi:10.1080/0305707022000006521 Weatherly, L. A. (2003). Human Capital - The Elusive Asset: Measuring and Managing Human Capital: A Strategic Imperative for HR (SHRM Research Quarterly Rep.). Alexandria, VA, USA: The Society for Human Resource Management (SHRM). Xu, G. (1639). Nong Zheng Quan Shu (Complete Book of Agricultural Administration). Shanghai: Publisher not identified. In Menzies, N. (1988). Three hundred years of Taungya: A sustainable system of forestry in South China. Human Ecology, 16(4), 361- 376. doi:10.1007/bf00891648 Yao, S., Guo, Y., & Huo, X. (2010). An Empirical Analysis of the Effects of China’s Land Conversion Program on Farmers’ Income Growth and Labor Transfer. Environmental Management, 45(3), 502-512. doi:10.1007/s00267-009-9376-7 910