7.31 . x
4. 2:0. is

 

 

 

 

 

 

 

 

 

 

 

.
1.. . .5
33m: “3%. u
... 3: .3:
24...“...32'...
1 15.3

lxl‘:
.Hziiblgn’asll
. . f. :0!
0 ti 5! it‘

y: r; . .3: 5..
. Saxhyﬂuaixﬂur

'2 I‘I ‘-
3’}. " ‘01-
$cx5.u1_ ,.s.~c n.- 3... t a n ..
3‘!!! C . I'.’I :13,
113...: xtli:s:l.?i ti. .- .0:
sat-3...}; ‘0‘. ‘1 l .
. .20....0a33 t... :1...- Eli-iii...
b... .’W 5’31!“
.0. I

g

. III:
.2: n 131 {\Illilll
13:19.93 intxotli?

n... ..t..u~ 15113... ‘1‘...
L :55! a?" 3‘;

5.5.3.
v r .152-
. 3.}? 5'31}.

.‘u.’ 13.53;! .tlp.01llitb.‘hl.
.zxtll i‘xllzl1

.t..!.!.!!.
[2.3 ’13.???

 

 

" MICI'IIQa- “ State
University

 

 

 

This is to certify that the
thesis entitled

ECONOMIC IMPACT OF NEWCASTLE DISEASE CONTROL
IN VILLAGE CHICKENS: A CASE STUDY IN MOZAMBIQUE

presented by

ALDA ARMINDO TOMO

has been accepted towards fulfillment
of the requirements for the

MS. degree in Aﬂultural Economics

 

 

are“ oak

Major Professor’s Signature

ﬂag,“ 3% /7.ﬁ 9019?

Date

MSU is an Afﬁrmative Action/Equal Opportunity Employer

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE DATE DUE

 

 

99111

5

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K'IProjIAcc8PrelelRC/DaIeDueiindd

ECONOMIC IMPACT OF NEWCASTLE DISEASE CONTROL IN VILLAGE
CHICKENS: A CASE STUDY IN MOZAMBIQUE
By

Alda Armindo Tomo

A THESIS

Submitted to
Michigan State University
In partial fulﬁllment of the requirements
for the degree of

MASTER OF SCIENCE

Agricultural Economics

2009

ABSTRACT

ECONOMIC IMPACT OF NEWCASTLE DISEASE CONTROL IN VILLAGE
CHICKEN: A CASE STUDY IN MOZAMBIQUE

By
Alda Armindo Tomo
Newcastle disease (ND) is the main constraint for village chicken production in
Mozambique, and chicken vaccination is the only effective way to control ND. The I-2
vaccine has been locally produced since 1999, and it is suitable for small farmers, but it is
only used in some parts of the country, with very low levels of adoption. Vaccine
availability is the key bottleneck for increased adoption. The current level of vaccine
production is under the installed capacity, indicating potential for expansion of a
vaccination program; but, it is necessary to evaluate the economic viability of the
program and the feasibility of its expansion. This research evaluates the ﬁnancial
viability of the vaccination program at farm level, as well as the economic viability of the
program in Chibuto District and all the districts the program is implemented. Also, the
sensitivity of the proﬁtability measures to ND incidence, chicken price reduction and
adoption ceiling levels is evaluated. A simulation program VIPOSIM with parameters
adapted to the Mozambican case is used to simulate the beneﬁts of vaccination at farm
level, while beneﬁt-cost analysis is used to evaluate the proﬁtability of the program. For
households with ﬂocks above 10 chickens, the vaccination is shown to be proﬁtable, even
for low levels of ND infection. Production and extension of I-2 vaccine has positive
returns; and, expansion of the program to new areas should be explored and strategies put
in place to speed up the adoption in areas already covered. When planning expansion of

the program, areas with high incidence of ND should be prioritized.

To my parents, for their unconditional support and love.
To my siblings, for their encouragement and unending support.

To my nephew Elton, may this be an inspiration for the long way ahead in his life.

ACKNOWLEDGMENTS

I would ﬁrst like to acknowledge USAID and MSU Food Security Project in
Mozambique for funding my assistantship and ﬁeldwork. I am very grateful to my major
professor Dr. Cynthia Donovan, for her guidance and support throughout my master
degree studies. I would like to thank my thesis supervisor Dr. Eric Crawford, for his
valuable input throughout my Master’s thesis research. I am also grateful to Dr. Thomas
Walker, a non-voting member of my committee, for helping me to conceptualize this
study and for his insightful comments during the writing process. I would like to thank
Dr. James Lloyd and Dr. Duncan Boughton for serving my committee and for providing

valuable input on the thesis. Also, I thank Nicky Mason for her useful comments.

Additionally, I would like acknowledge the Animal Production Systems Group from the
Department of Animal Sciences in Wageningen University, and my special thanks to Dr.
Henk Udo and Dr. Theo Viets, for providing the Village Poultry Simulation Model, a tool
which was extremely valuable in this research. My gratitude is extended to the Southern
Africa Newcastle Disease Control Project, which was funded by the Australian Agency
for International Development and implemented by GRM International, with special
thanks to Dr. Robyn Alders, for providing bibliographic material. Also, I thank Dr.

Joanne Meers for providing bibliographic material.

In Mozambique several institutions and individuals direct or indirectly supported my
ﬁeld work. My special thanks to: Manuel Josefa and other staff from CSL for providing
the vaccination data set and other relevant information; Ellen Payongayong for her

valuable assistance during sampling process; Dr. Harun from Veterinary Medicine

College in Eduardo Mondlane University, for Sharing his extensive knowledge of ND and
I-2 vaccine; a team from Animal Science Directorate in IIAM, especially Quintino Lobo
and Amilcar, for sharing information on the production of I-2 vaccine in Mozambique; a
team from DINAP and DNER, especially He'lder Gemo, Titos Paris and Anabela dos
Muchangos, for providing relevant information on chicken vaccination in Mozambique;
VETAID, World Vision, SPP and Chibuto District Agricultural Services, for the sharing
information on chicken vaccination; Mr. Jorge Francisco from IIAM for editing the map;
teams from CESE, DFDTT and Mozambique MSU F 8111, for their logistic support; the
drivers, Mr. Reginaldo, Joao Dunduro and Mr. Manhica for their safe driving. Also, I’m

very grateﬁil to chicken farmers in Chibuto, for their cooperation and generosity.

The words might not be enough to thank my friends and fellow graduate students in
Michigan, especially Mukwiti, Wezi, Nicky, Kirimi, Tina, Martin, Kay, Alexandra,
Ramzi, Anna, Sean and Alex. With all we shared together, my time in USA was not only
about academic learning but also about school of life. I would also like to thank Tammy
Brown and her family, for the time we shared, their friendship and hospitality. Your
support in every way made it clear that I have family in Michigan. Lastly, but deﬁnitely
not the least, I am very thankful to God for the strength, wisdom, inspiration, courage,

and many other uncountable blessings, which are proof his unfailing love.

TABLE OF CONTENTS

 

 

 

LIST OF TABLES _ _ - -- _ - _ - -_ _- - _ - _ __viii
LIST OF FIGURES . -- - -- _- - - - - - - -_ _- -_ ix
LIST OF ACRONYMS _ ........................ _ _ - _ ................................ x
CHAPTER I: INTRODUCTION .................................................................................... 1
1.1. GENERAL BACKGROUND ........................................................................................... 1
1.2. PROBLEM STATEMENT .............................................................................................. 5
1.3. RESEARCH QUESTIONS .............................................................................................. 8
1.4. HYPOTHESES ............................................................................................................. 9
1.5. ORGANIZATION OF THESIS ...................................................................................... 10
CHAPTER 11: AN OVERVIEW OF VILLAGE CHICKEN PRODUCTION ......... 11
2.1. THE RELEVANCE OF POULTRY PRODUCTION ........................................................... l 1
2.2. POULTRY PRODUCTION SYSTEM IN MOZAMBIQUE .................................................. 13
2.3. EPIDEMIOLOGY OF NEWCASTLE DISEASE AND ITS CONTROL .................................. 14
2.4. SYSTEM OF PRODUCTION AND EXTENSION OF I-2 VACCINE IN MOZAMBIQUE ........ 17
2.4.1. Analysis of Critical Issues Relevant for Improvement of Production and
Extension of [-2 Vaccine in Mozambique ................................................................. I 9
CHAPTER III: LITERATURE REVIEW AND METHODS .................................... 23
3.1. OVERVIEW OF IMPACT ASSESSMENT STUDIES ........................................................ 23
3.2. ECONOMICS OF ANIMAL DISEASE ........................................................................... 27
3.3. THEORETICAL FRAMEWORK FOR IMPACT ASSESSMENT OF ND CONTROL .............. 29
3.4 DESCRIPTION OF STUDY AREA ................................................................................. 31
3.5. DESCRIPTION OF DATA AND METHODS OF COLLECTION ......................................... 34
3.6. MEASUREMENT OF IMPACTS OF ND CONTROL IN VILLAGE POULTRY: VIPOSIM .. 35
3.7. ADAPTATION OF VIPOSIM TO THE CONTEXT OF THE STUDY ................................. 39
3.8. ANALYSIS OF THE IMPACT OF CHICKENS VACCINATION ......................................... 42
3. 8. I . Measurement of Annual Net Beneﬁts of ND control at Farm Level ............... 44
3.8.1.1. Considerations of the Effect of the Flock size on the Beneﬁts ................ 46
3.8.1.2. Determination of Incremental Beneﬁts of Vaccination ........................... 48
3.8.1.3. Risk Analysis ........................................................................................... 50
3.8.2. Measurement of Returns to Investments in ND Control Program .................. 51
3.8.2.1. Deterministic Measurement of the Returns to ND Control ..................... 51
3.8.2.1.1. Estimation of Beneﬁts of ND Control at the Aggregate Level ......... 53
3.8.2.1 .2. Estimation of the Costs of the Vaccination Program ........................ 57
3.8.2.1.3. Measurement of Economic Proﬁtability ........................................... 58
3.8214. Sensitivity Analysis .......................................................................... 59
3.8.2.2. Stochastic Analysis of the Returns to ND Control .................................. 59

vi

CHAPTER IV: RESULTS OF ANALYSES OF THE IMPACT ND CONTROL... 60

 

4.1. BENEFITS OF ND CONTROL AT FARM LEVEL .......................................................... 60
4.1.1. Deterministic Incremental Net Beneﬁts at Farm Level .................................. 60
4.1.1.1. Sensitivity of Net Beneﬁts at Farm Level to ND Mortality Levels ......... 62
4.1.1.2. Sensitivity of Incremental Net Beneﬁts at Farm Level to Prices ............. 65

4.1.2. Stochastic Analysis of Incremental Net Benefits at Farm Level ..................... 67

4.2. PROFITABILITY OF VACCINATION PROGRAM AT DISTRICT LEVEL ........................... 68

4. 2. 1. Sensitivity Analysis .......................................................................................... 68
4.2.2. Stochastic Analysis .......................................................................................... 71

4.3. PROFITABILITY OF PRODUCTION AND EXTENSION OF [-2 IN MOZAMBIQUE ............. 72

4. 3. 1. Sensitivity Analysis .......................................................................................... 72
4.3.2. Stochastic Analysis .......................................................................................... 74
CHAPTER V: SUMMARY OF FINDINGS AND POLICY IMPLICATIONS ....... 76
5.1. FINDINGS AND POLICY IMPLICATIONs ..................................................................... 76
5.2. FURTHER RESEARCH ............................................................................................... 79
APPENDICES .................... - - .......... 81
REFERENCES .............................................................................................................. 118

Vii

LIST OF TABLES

Table 1: Percentage of Rural Households Owning Livestock in Mozambique ................ 11
Table 2: Reasons for Raising Chickens among Households in Chibuto District ............. 12
Table 3: Rates of Mortality Due to Predation and Other Causes ...................................... 41
Table 4: Deﬁnition of Households Categories Based on the Flock Size Distribution in
Chibuto .............................................................................................................................. 48
Table 5: Annual Incremental Beneﬁts and Costs of ND Control at Farm Level .............. 60
Table A 1: Average Chickens per Household and Total Chickens Vaccinated per
Vaccinator in Each Campaign .......................................................................................... 82
Table A 2: Estimated Adoption Rates in Chibuto District ............................................... 96
Table A 3: Estimated Adoption Rates in the 44 Districts where [-2 Vaccine is Currently
Used .................................................................................................................................. 98
Table A 4: Projected of Number of Total Households and Households Raising CthkCI;%O
Table A 5: Estimated Costs of Vaccination Program ..................................................... 102
Table A 6: Number of I—2 Doses Produced .................................................................... 102

Table A 7: Farm Annual Net Beneﬁts per Flock Category and ND Mortality Levels 103
Table A 8: Proﬁtability of the Vaccination Program at District Level ........................... 104

Table A 9: Sensitivity Analysis of the Beneﬁts of Vaccination to Overall Without-
Control ND Mortality at the District Level ..................................................................... 105

Table A 10: Sensitivity Analysis of the Beneﬁts of Vaccination to Assumption of
Adoption Ceiling Levels at District Level ...................................................................... 109

Table A 11: Proﬁtability of the Vaccination Program in the 44 Districts of
Implementation ............................................................................................................... 1 1 1

Table A 12: Sensitivity Analysis of the Beneﬁts of Vaccination to Overall Without-
Control ND Mortality in the 44 Implementation Districts ............................................. 1 12

Table A 13: Sensitivity Analysis of the Beneﬁts of Vaccination to Assumption of
Adoption Ceiling Level in the 44 Implementation Districts ........................................... 1 l6

viii

LIST OF FIGURES

Figure 1: The Effect of Disease on Livestock Production ................................................ 28
Figure 2: Impact of Vaccine Use in Village Chicken Production .................................... 30
Figure 3: Location of the Study Area ............................................................................... 33
Figure 4: Schematic Representation of Sequences of Events in the VIPOSIM for a
Reproduction Season. ....................................................................................................... 37
Figure 5: Distribution of Flock Size in Chibuto ............................................................... 47
Figure 6: Sensitivity of the Annual Net Beneﬁts at Farm Level to Without-Control ND
Mortality Levels ................................................................................................................ 63
Figure 7: Distribution of the Overall Annual Net Beneﬁts at Farm Level ....................... 67

Figure 8: Beneﬁts of Investments in the Extension of I-2 Vaccine in Chibuto District .. 68

Figure 9: Sensitivity of NPV to Without-Control ND Mortality Rates in Chibuto .......... 69
Figure 10: Sensitivity of NPV to Adoption Ceiling Levels in Chibuto ............................ 70
Figure 11: Distribution of NPV of Extension of I-2 Vaccine in Chibuto District ............ 71
Figure 12: Beneﬁts of Investments in the Production and Extension of I-2 Vaccine in the
44 Districts ........................................................................................................................ 72
Figure 13: Sensitivity of NPV to Without-control ND Mortality Rates in the 44 the
Districts ............................................................................................................................. 73
Figure 14: Sensitivity of NPV to Adoption Ceiling Levels in the 44 Districts ................ 74

Figure 15: Distribution of NPV of Production and Extension of I-2 Vaccine in the 44
Districts ............................................................................................................................. 75

Figure A 1: 1-2 Adoption Proﬁles at District Level Given the Adoption Ceiling Levels. 97

Figure A 2: 1-2 Adoption Proﬁles in the 44 Districts Given the Adoption Ceiling Levels
........................................................................................................................................... 99

Figure A 3: Total Number of Households and Number of Households Raising Chickens
at District Level ............................................................................................................... 101

Figure A 4: Total Number of Households and Number of Households Raising Chickens
in the 44 Districts ............................................................................................................ 101

ix

LIST OF ACRONYMS

ACIAR - Australian Center for International Agricultural Research
AIDS —- Acquired Immunodeﬁciency Syndrome

DINAP — National Livestock Directorate

DNER — National Extension Services Directorate

FAO — Food and Agriculture Organization

HIV — Human Immunodeﬁciency Virus

IIAM — Institute of Agricultural Research of Mozambique

IRR — Internal Rate of Return

MAE — Ministry of State Administration

MADER — Ministry of Agriculture and Rural Development (currently called MINAG)
MINAG — Ministry of Agriculture

MTN - Meticais, Mozambican currency.

ND — Newcastle Disease

NGO — Non-Governmental Agency

NPV — Net Present Value

PARPA — Government’s Strategic Plan to Reduce Poverty

SANDCP -- Southern Africa Newcastle Disease Control Program
SSA — Sub-Saharan Africa

TIA —Trabalho de Inquérito Agricola (National Agricultural Survey)

VIPOSIM — Village Poultry Simulation Model

CHAPTER I: INTRODUCTION

1.1. General Background

The Mozambican agricultural sector plays a very important role in the sociO-economic
development of the country. Most of the population in Mozambique lives in rural areas
and they depend directly or indirectly on agricultural activities for their livelihoods. The

agricultural sector is dominated by smallholders who farm in a risky environment that is

vulnerable to droughts and ﬂoods (World Bank, 2006).

Although the incidence of rural poverty decreased from 69 percent in 1996-97 to 54
percent in 2002-03 (Government of Mozambique 2006), rural poverty remains
widespread. Rural inequality appears to have increased moderately from 1996 to 2003, as
rural growth beneﬁted upper income quintiles relatively more than low-income quintiles.
The poorest households, particularly female-headed households, were signiﬁcantly
disadvantaged during this period of growth (World Bank 2006). According to Boughton
et al. (2006), rural household incomes remain very low, and they are critically low for the
poorest 60 percent of the population. Food security is a concern for rural households
given Mozambique’s vulnerability to weather and market induced risks (World Bank
2006). Absolute poverty is still a major challenge to the Mozambican government and
according to the government’s Strategic Plan to Reduce Poverty (known as PARPA), the
government seeks to reduce absolute poverty to 45 percent by 2009 (Government of
Mozambique 2006). The agricultural sector is seen as a critical component in the
PARPA. By promoting sustained agricultural production and productivity growth, the
government can help to ensure growth of the rural sector, improving livelihoods and

reducing risk and vulnerability of households who depend on agriculture for their

livelihoods. Development programs beneﬁting the poorest among the rural farmers
constitute the key for effective poverty eradication, and programs targeting these farmers

should be given special attention.

Livestock production constitutes an important component of the agricultural economy of
developing countries. The livestock sector makes signiﬁcant contributions to the
livelihood of smallholders and the rural poor. Livestock serves to increase farm income,
while it also diversiﬁes farm income, thereby reducing risk and vulnerability, especially
in regions where potential for crop production is limited (World Bank 2006; Branckaert
and Gueye 1999). The contribution of livestock goes beyond direct food production and
includes multipurpose products and uses, such as skins, feathers, ﬁber, manure for
fertilizer and fuel, power and transportation. Animals also provide a means of capital
accumulation and serve as barter products in societies where there is limited circulation
of currency (Branckaert and Gueye 1999). Furthermore, livestock holding is associated
with status in a community and other cultural and religious traditions (Branckaert and

Gueye 1999).

Chickens are the most signiﬁcant livestock species in terms of level of ownership, access
to animal protein, and the potential for earning cash income in Southern Africa
(SANDCP 2005). Rather than caged, large-scale poultry production, Mozambican
smallholders have “village poultry” systems that are generally owned and managed by
rural poor people, usually women. Village chickens are most often kept under a free-
range, low input management system (SANDCP 2005). In Sub-Saharan Africa (SSA)
about 85 percent of all households keep poultry, with women owning 70 percent of them

(Guéye, 1998). In Mozambique, the national agricultural survey data (TIA 2002; TIA
2

2003; TIA 2005) indicate that more than 65 percent of rural households keep chickens. In
their analysis of the determinants of rural income in Mozambique, Walker et a1. (2004)
found that possession of some chickens results in marked superiority in rural income:
households with 30 or more chickens have 33 percent higher agricultural income.
Furthermore, Walker et a1. (2006) Show that chickens are among the ﬁrst 10 agricultural
commodities with the highest contributions to total value of production, and that a 20
percent increase in chicken production could result in a four percent reduction in the
severity of poverty. Thus, productivity enhancements for village poultry have a
signiﬁcant bearing on the economic well being of a large portion of Mozambican rural
households. Increased poultry production can also result in improved nutritional
wellbeing, whether the chickens and eggs are consumed in the household or sold to
generate income. Village chicken production is a reliable and viable tool for eradication

of rural poverty and food insecurity in Mozambique.

Newcastle disease (ND) is the most serious constraint for village chicken production
throughout the world, particularly in developing countries (Branckaert and Gueye 1999).
ND is the most pathogenic of the annual poultry epidemics in Mozambique, and it is the
main cause of mortality of chickens accounting for between 50 and 100 percent of deaths
annually in rural households (Bangnol 2001; MADER 2004; MADER 2005). According
to TIA data, about 66 percent of households who raised chickens during the 2005-2006
cropping season reported losing birds because of disease. The high prevalence of ND in
Mozambique and its resulting losses suggest that improvement of household poultry

production requires capacity to ensure sustainable ND control.

Given the nutritional, cultural, and economic importance of village chickens and the
severity of the ND problem, several vaccines for ND control have been tested and used in
Mozambique. However, the level of chicken vaccination against ND is still very low.
According to National Agricultural Survey data (TIA 2006) in 2006, only six percent of

households who raise chickens, vaccinated their chickens against ND.

Sustainable ND control requires maintenance of high levels of quality control in
production, distribution and administration of the vaccine, all in a timely and low cost
manner. Village chicken production systems are based on minimum input use with low
productivity. Any cost-effective strategy without high capital costs that increases their
productivity will assist in poverty alleviation and food security improvement (Alders and
Spradbrow 2001). Thus, investments in production and extension of technologies for ND
control may have signiﬁcant returns and, more importantly, may have a relevant role in

the reduction of rural poverty and food insecurity.

There are both direct and indirect beneﬁciaries of ND control. The indirect beneﬁts may
result from reduced village chicken mortality which increases poultry and possibly egg
supply for the market and/or household consumption. Given that village chickens are one
of the main assets owned by the poor households, adoption of ND control by these
families may have a signiﬁcant effect on poverty reduction. Thus, ND control has a vital
role in increasing poultry production, contributing to the improvement of household food

security and poor rural income.

1.2. Problem Statement

Currently two vaccines are used by small-scale farmers for ND control. The ﬁrst is the I-
2 vaccine, a thermo stable live vaccine which was developed and tested locally with
technical and ﬁnancial support from the Australian Centre for International Agricultural
Research (ACIAR) ND Control Project. The second vaccine is called Itanew, an
inactivated ND vaccine imported through a livestock project funded by the African
Development Bank (Alders et al. 2000). Between the two vaccines, 1-2 is more suitable
for small-scale Mozambican farmers due to its lower costs (World Bank 2006; Alders
and Spradbrow 2001). In addition, I-2 does not require technical skills for its application,
and it stimulates all forms of immunity (Alders and Spradbrow 2001). Therefore, the I-2
vaccine has better prospects for enhancement of village poultry productivity in

Mozambique.

Even though the I-2 vaccine has been produced by the Institute of Agricultural Research
of Mozambique (IIAM) for use in village chickens since 1999, until now it has been used
only in selected districts in the provinces of Gaza, Inhambane, Tete, Nampula and
Zambezia. Program administrators working with the I-2 vaccine decided to launch the
program in these provinces with prospects of later expansion to other districts and
provinces. Selection of these provinces was based on stafﬁng levels for both the
Directorate of Rural Extension Services (DNER) and the National Directorate of
Livestock (DINAP)I, provincial levels of poultry production, and existence of Non-

governmental Organizations (N G03) and private companies interested in participating in

 

I DINAP, is the entity which coordinates and implements livestock disease control programs in the
Ministry of Agriculture (MINAG), in collaboration with DNER. For more details about the decision
making and management of 1-2 vaccine in Mozambique, see section 3.4.

the program 2. Availability of the vaccine constitutes the main constraint for adoption,
and IIAM production levels of the vaccine are based on requests from DINAP. Currently
about 2,500,000 doses are produced per vaccination campaign3. Given that three
vaccination campaigns are carried out per year, the actual level of production is far below
the installed capacity which is 23,000,000 doses per year, about three times the current
levels of production. This suggests that there is a potential to expand the vaccination

program to the other provinces.

However, decisions on coverage require additional information on the economic viability
of the vaccination program and on the likelihood that the program will be expanded
successfully and sustainably in the long term. Expansion of the ND control program to
other sites implies the need for considerably higher budgets (vaccines production inputs,
transportation, and refrigeration equipment), and will compete with other development
interventions for scarce resources. As demand for scarce ﬁrnds grows, better evidence is
needed to Show that investments in the program can generate attractive returns (Batz et
a1. 2003), or at least meet some of the main economic, political and social objectives.
Scientists, research administrators and policymakers face increasing pressure to justify
continued public investment in agricultural research and development programs (Alston
et -al. 1998). Socio-economic impact studies of investments can help policymakers to
assess the value Of past investments and to prioritize alternative future investments.
Evaluation studies of past, current and proposed development programs are very
important for effective and efﬁcient use of available resources. Empirical evidence of the

economic impact of investment in the ND control program using the 1-2 vaccine in

 

2 Hélder Gemo, conversation with author, Maputo, 18 August 2007.
3 Quitino Lobo, intemet conversation with author, 20 February 2009

6

already implemented zones is fundamental for proper formulation of policy and strategies

of expansion of the program of ND control using this vaccine.

Despite the relevance of this kind of study for the researchers and policy makers, few
studies of evaluation of the ND control program have been conducted in Mozambique.
Woolcock et al. (2004) studied the impact of ND vaccination on household welfare in
Mozambique using a static poultry model. Although the parameters they used in their
model to represent “without-control” situation were based on the literature and portrayed
a typical African village chicken flock, the parameters for “with-control” situation were
purely based on assumptions; no Speciﬁc data were used to get the parameters of the
beneﬁts of vaccination. Furthermore, researchers have given little attention to the
evaluation of efﬁciency of investments in production and extension of the I-2 vaccine at

aggregate levels.

This study ﬁlls the gap of the study of Woolcock et al. (2004) by using farmers’ surveys
and elicitation methodology in combination with parameters derived from empirical
literature to estimate the beneﬁts of ND control at the farm level. In addition, this study
will provide information on return of investments on ND control at the regional level.
The ﬁndings of this study will provide information about the proﬁtability of ND control
to farmers and to society in general. This information will be helpful in providing insights
relevant for expansion of ND disease control using the I-2 vaccine to the other provinces

and for prioritization of development activities which compete for scarce funds.

1.3. Research Questions

The study addresses the following questions:

ii.

iii.

iV.

vi.

Is it ﬁnancially viable for farmers in Chibuto District to invest in the vaccine I-2

to protect their chickens against ND?

How sensitive are the farm-level annual net beneﬁts of the vaccination to various

hypothetical disease mortality scenarios and possible chicken price reductions?

Are the investments of the government and NGO agencies to promote the I-2

vaccine in the District of Chibuto economically viable?

Are the investments of the government and NGO agencies to promote the 1-2

vaccine in the ﬁve provinces where it is currently used economically Viable?

Is it economically Viable to expand the production and use of the I-2 vaccine in

Mozambique?

How sensitive are the aggregate proﬁtability measures to changes in mortality

rates, chicken prices, and adoption ceiling levels?

1.4. Hypotheses

This study tests the hypothesis that investments in the program of vaccination of Village
chickens against ND have positive returns at the farm level as well as at the district and
national levels. This hypothesis will be veriﬁed if: i) the net beneﬁts associated with the
use of the vaccine at the farm level are positive and generate a sufﬁciently high marginal
rate of return to warrant adoption; and ii) the net present value (NPV) is positive or the
economic rate of return of the investments in the vaccine production and distribution at

the aggregate level is greater than the opportunity cost of capital at the aggregate level.

A Village Poultry Simulation Model (VIPOSIM) with parameters adapted to the
Mozambican context is used for determination of annual incremental beneﬁts resulting
from chicken vaccination at the farm level. Data on village poultry production systems
collected in Chibuto district are used for simulation. In both deterministic and stochastic
analyses, the distribution of ﬂock size in Chibuto is used in deﬁning different categories
of initial ﬂock size that are considered for determination of overall annual incremental
beneﬁts at the farm level. In stochastic analyses, @Risk software is used in combination
with VIPOSIM to measure the proﬁtability of the vaccination, and the without-control
ND mortality rate is treated as a random variable. For aggregate analyses, the results of
incremental annual beneﬁts at farm level are used, and the costs involved are

incorporated in the analysis for determination of aggregate measures of proﬁtability.

In deterministic analysis for both farm and aggregate levels, sensitivity of proﬁtability
measures to. without-control ND mortality rates and chicken price reductions is evaluated.
Also, for aggregate analysis, the sensitivity of the NPV to the assumptions of adoption

ceiling level is evaluated.

1.5. Organization of Thesis

This thesis is divided into ﬁve chapters. Chapter 2 gives an overview of the Mozambican
poultry production sector and information about the epidemiology of ND. Then, the
system of production and distribution of the I-2 vaccine in Mozambique is described and
their constraints discussed. In Chapter 3, the literature on impact assessment is brieﬂy
reviewed; the theoretical framework for analysis of animal diseases and their control is
presented, and methodologies of analysis as well as the data sources are discussed. Then,
procedures for the analyses are described in detail. In Chapter 4, the results of the
analyses are presented and discussed. Chapter 6 summarizes the main ﬁndings and their

policy implications, and then further research is recommended.

10

CHAPTER 11: AN OVERVIEW OF VILLAGE CHICKEN PRODUCTION

2.1. The Relevance of Poultry Production

Village chickens are the most signiﬁcant livestock Species in terms of level of ownership,
access to animal protein, and the potential for earning cash income (SANDCP 2005). In
SSA, 85 percent of all households keep poultry, with women owning 70 percent of them
(Gueye, 1998). Poultry provides approximately 20 percent of protein consumed in
developing countries (Jensen and Dolberg 2002). According to TIA data, poultry is the
most popular livestock kept by small and medium scale farmers in Mozambique (see

Table 1).

Table 1: Percentage of Rural Households Owning Livestock in Mozambique

 

 

Year

Livestock Species 2002 2005 2006
Chickens 74 65 70
Goats 37 38 34
Cattle 1 4 l9 1 7
Pigs 1 9 1 9 1 5
Ducks 1 9 1 3 l 3
sheep 3 4 3
Turkey 1 1 1

 

Data Source: National Agricultural Survey (TIA)

Village poultry play a Vital role in the improvement of nutritional status and income of
many poor rural households and are a global asset for many millions who live below the
poverty line (Copland and Alders 2005). Village poultry provide scarce animal protein in
the form of meat and eggs and provide the owners with a form of savings which can help
in times of need to meet essential family expenses such as medicines, clothing and school
fees. Families can also increase their income by taking advantage of seasonal peaks in

poultry demand, such as at religious festivals or celebrations (Johnston and Cumming

ll

1991). The beneﬁts of family poultry production include other functions for which it is
difﬁcult to assign any monetary value. These include pest control, provision of manure
and household contribution to traditional ceremonies and festivals. In addition, village
chickens provide some beneﬁts in terms of cleanliness and hygiene (Johnston and
Cumming 1991). According to the results of household survey in Chibuto, small-scale
farmers raise chickens for three primary reasons: regular consumption within the
household, serving guests (social obligations), and sales in emergency Situations

(Table 2).

Table 2: Reasons for Raising Chickens among Households in Chibuto District

 

 

Reason % of households
Regular Consumption 83
To Serve Guests 69
Sale in Emergency Situations 48
Regular Sale 32
Cultural Ceremonies 21
Consumption in Holidays 13
Exchange for Other Kinds Of Food 1 1
Other <1

 

Data Source: Chibuto Survey in 2007
Note: Each household could have up to 8 reasons during the interview

Furthermore, village chickens play an important role in households where there is a lack
of able-bodied workers due to war or HIV/AIDS, and in households that have a disabled
or elderly member (Copland and Alders 2005). According to the same authors, in
households headed by widows, children or grandparents, chickens represent the easiest
species to raise for sale and home consumption, providing a source of high quality

protein and vitamins that play an important role in the nutrition of HIV/AIDS patients.

12

2.2. Poultry Production System in Mozambique

Poultry production in Mozambique is carried out in two systems of production: the free-
range poultry system in rural and peri-urban areas and the caged poultry system which is
based on commercially improved poultry farms concentrated in urban or peri-urban areas.
The free range system dominates with scavenging chickens as an integral part of the
farming system. According to the Chibuto District survey, essentially all (99.6 percent) of
the households interviewed let their chickens scavenge. In general, the management
system of village chickens involves minimum input use, requiring the lowest capital
investment of any livestock Species with a Short production cycle (Copland and Alders
2005) and waste from household consumption is occasionally used to supplement the
diet. Only seven percent of interviewed households stated that they sometimes buy

supplements for their chickens.

In addition, most farmers in the Village poultry sector in Mozambique do not provide
housing for their chickens. About 26 percent of interviewed households only provide
housing to their chickens during the night and less than one percent of households
provide housing continuously. Although housing is usually suggested as a means to
improve chicken productivity, it appears to be controversial for several reasons. First, in
discussion groups in Chibuto, farmers revealed little interest in housing their chickens
because of the perceived high risk associated with housing. According to the farmers
when housing chickens, rather than keeping them free and sleeping in trees, there is high
risk of losing all the chickens at once either due to attack of predators where they are
conﬁned or due to theft. Also, when chickens are conﬁned, they are no longer able to

scavenge around for their own food, and the farmer will need to invest in supplements.

13

This can be a big problem for poor farmers-who most of the time do not have enough

resources even to meet their own food needs.

The village chicken production system is also characterized by low output levels, which
are mainly because of high mortality due to diseases and predation. ND is found to be the
most serious constraint for village chicken production in Mozambique; it is the main
cause of mortality of chickens (Bangnol 2001; MADER 2004; MADER 2005). ND is
endemic in Mozambique, occurring every year in the rural poultry sector (MADER
1992). According to the survey in Chibuto district, 87 percent of interviewed households
reported that they had lost chickens due to disease in the two previous years. ND is
particularly devastating for small village farmers who usually have limited means of
protecting their ﬂocks. It is commonly recognized that little progress could be made in

the village poultry industry unless ND is controlled.

2.3. Epidemiology of Newcastle Disease and Its Control

ND is a highly infectious Viral disease caused by a paramyxovirus which mainly affects
poultry (Alders and Spradbrow 2001). The ND virus can infect through the respiratory
tract, the ocular mucous membranes and the digestive tract. The incubation period usually
ranges from 2 to 15 days depending on the strain of virus. Under Village conditions, the

Virus is unlikely to survive outside the host for more than a month.

The virulence of the disease depends on the particular strain of the virus. Of the highly
virulent strains, which are particularly common in South-East Asia and Africa, some
grow in the gut (viscerotropic strains), while others grow mainly in the central nervous

system (neurotropic strains). The most common indication of a serious outbreak of a

14

neurotropic strain of the disease is seen in a nervous symptom exhibited in infected birds
where the neck twists right back and the chickens simply fall over and die (Alders and
Spradbrow 2001). Less virulent strains, such as those that are endemic in Australia, affect
only the respiratory system, with varying degrees of severity. Symptoms may include loss
of appetite, a dramatic drop in egg production, increased respiration, and coughing,
gasping and even rapid death without any exhibition of other symptoms. Most avian
species, domestic or feral, can be infected by strains of ND virus; however, the
consequences of these infections will vary with the strain of virus and the species of the
host infected (Spradbrow, 1999). Chickens are the most susceptible host (Alders and
Spradbrow 2001). Ducks and geese are also susceptible to infection, but these species

rarely succumb to the disease.

ND is highly transmissible: birds are normally infected through direct contact with
diseased or carrier birds, but the virus can also be carried on contaminated objects such as
chicken or egg crates, feed, vehicles, dust and clothing. The virus is usually inactivated
by direct sunlight, but in cool weather the virus can survive in feces and contaminated
housing for up to 21 days. The virus can also persist in poultry products (meat and eggs)
and can be carried by migrating wild birds. Human activity was found to inﬂuence the
occurrence of ND. Seasonal high sales of chickens and population movements are some

of the factors that lead to outbreaks.

Eradication of ND is unlikely and there are few poultry species which are resistant to the
disease. Continual vaccination programs currently offer the only effective way of
controlling ND (Udo et al. 2006; Alders and Spradbrow 2001). Thus, chicken vaccination

is one of the most important technical possibilities to improve village chicken production.

15

ND control was found to have positive effects on bird off-take, egg production, egg off-
take and ﬂock size (Udo et al. 2006; Woolcock et al. 2004). Of the households surveyed
in Chibuto who participated in a vaccination program, 80 percent perceived

improvements in their chicken production resulting from ND control.

Different types of vaccine were developed for ND control. Conventional vaccines used in
formal vaccination adopted in commercial chicken enterprises are inappropriate for
village chickens for several reasons. First, village chickens are raised in small, multi-age
and free-range ﬂocks; therefore it requires more effort and time to vaccinate them using
conventional vaccines than when the I-2 vaccine is used. Second, commercial vaccines
are heat-labile, requiring complex cold-chains to link the vaccine producers and users,
which is difﬁcult to maintain under village conditions. Finally, the cost of purchase is
higher than the I-2 vaccine because they can only be purchased in relatively higher
number of doses (Alders and Spradbrow 2001). However, new vaccines, including the I-
2, are thermo stable, such that they can withstand limited periods without refrigeration
prior to use. The I-2 vaccine was developed by ACIAR and it is free of commercial
ownership such that seed cultures of I-2 are made available without cost to countries that
wish to test or produce their own vaccine (Alders and Spradbrow 2001). I-2 is highly
suitable for village chickens, is very cheap, and does not require a continuous cold chain.
It is safe for both chickens and handlers (i.e., overdosing causes no ill effect) and
protection can spread horizontally from vaccinated to non-vaccinated birds. Furthermore,
it does not require technical Skills for its application. I-2 has proven to be effective in

trials under laboratory conditions and in villages. According to the results of ﬁeld trials in

16

Mozambique, vaccination of chickens with I-2 vaccine via eye drops in the areas affected

by virulent ND virus increases their survival rate to about 80 percent (Dias et al. 2001).

2.4. System of Production and Extension of [-2 Vaccine in Mozambique

Currently I-2 vaccine is used in 44 districts within 5 provinces: Gaza, Inhambane, Tete,
Nampula, and Zambezia. The process of production and extension of I-2 vaccine in
Mozambique involves a series of stakeholders. DNER, IIAM, DINAP and SANDCP are
involved in vaccine production, distribution, vaccinator training and monitoring of the
program at the national level. IIAM is responsible for production of the I-2 vaccine as
well as for monitoring and evaluation of performance of the vaccine itself. DINAP is
responsible for delivery of requested amounts of the vaccine to the Provincial Livestock
Services (SPP). DINAP also coordinates monitoring and evaluating the vaccination
program at national level. DNER is responsible for vaccinator training and vaccine
extension evaluation. SANDCP provided the vaccine seed, and also provided some
technical and ﬁnancial support to the production and distribution of the vaccine during

the initial stage of the program.

SPP is responsible for providing cold chains for the vaccines, and for delivery of the
vaccine in the districts. SPP collects vaccine sales revenues from the districts and delivers
the funds to DINAP. In addition, SPP is responsible for providing information on vaccine
requirements and vaccination data to DINAP. SSP participates in monitoring and

evaluation of the vaccination program at provincial level.

The Services of Agriculture in the District coordinate the vaccination programs at district

level. They are responsible for delivery of the vaccine to the vaccinators and participating

17

NGOs. In addition, they are responsible for collecting vaccine sale revenues from the
vaccinators or NGOs to deliver to SPP, and participate in monitoring and evaluation of

vaccination program at district level.

NGOs provide technical and ﬁnancial support for vaccinator training and vaccine
distribution in some villages. In addition, they monitor and evaluate the vaccination
program in the villages they cover. The main NGOs participating in the ND control
programs in Mozambique are World Vision, Heifer International, ADRA, and VETAID.
A private company for mineral extraction, CSL, has a very active role in providing
technical and ﬁnancial support for chicken vaccination and vaccine conservation in
Chibuto. This company participates in vaccination to fulﬁll some of the requirements of
the government, where private companies have to contribute directly in some

development activities in the communities where they operate.

The vaccinators are community members trained in ND and vaccination issues, so that
they can vaccinate chickens in their villages. The vaccinators are not paid any salary;
they earn the margin they get after paying the cost of vaccine doses. The bottles contain
about 250 doses: given that each dose is 0.50 MTN, if the vaccinator is able to use
efﬁciently all doses in the bottle, he/she gets a margin of 100 MTN after paying the cost
of the bottle which is 25 MTN. Thus, the proﬁts the vaccinator gets depend on his/her
vaccination performance. Also, as a result of participating in the program, vaccinators

usually tend to be the biggest chickens raisers in the communities.

18

 

2.4.1. Analysis of Critical Issues Relevant for Improvement of Production and
Extension of [-2 Vaccine in Mozambique

The limited coverage of vaccine distribution systems seems to be the bottleneck for
diffusion of the vaccine in Mozambique. According to Gemo et al. (2005), there are
limited resources to expand public extension in Mozambique, and some hard choices
must be made in terms of the size of the public extension system and the role of NGOs
and the private sector in delivering extension. Partnership with NGOs and other private
entities in delivering the vaccine to the ﬁnal users is a good strategy to overcome public
funding limitations. However, relying only on that strategy may pose some threats in
terms of sustainability of vaccine access by the farmers in the long term. For the
particular case of Chibuto, the vaccination program is consistently well functioning only
in areas covered by CSL; in the other areas the access to this program is negligible. If for
some reason CSL stops its actions in the area of vaccination, there will deﬁnitely be
problems in controlling ND in that district. This suggests that there is a need to look for
additional strategies in order to ensure sustainability of the extension of the vaccine in the

long term.

I-2 vaccine can be produced in wet or freeze dried forms. Although both forms of I-2 are
thermostable, they can only withstand limited periods without refrigeration prior to use.
They still require cold storage and cold transport to maintain their maximum viability.
However, the freeze dried form lasts longer out of cold storage while the wet form
requires a more reliable distribution system. This suggests that among both forms of 1-2
vaccine, the freeze dried form is more suitable for the Mozambican context.
Nevertheless, the 1-2 is currently produced only as a wet vaccine in Mozambique because

the freeze dryer is not operational. The amount of vaccine produced by IIAM is based on

19

the request from DINAP; they do not produce more than the requested amount to avoid

spoilage4. As a result, their level of Operation is below production capacity.

This suggests that purchase of a new freeze dryer or repairing the old one could be a
means of increasing the efﬁciency of production and of distribution systems. By
restarting production of dried vaccine, production levels will no longer be limited to short
term requests from DINAP because the dried vaccine can last longer; it will be possible
to produce and to transport larger quantities of the vaccine to the provinces at once.
Production of larger quantities will allow for exploration of economies of scale, reducing
the production costs per unit of vaccine. Also, for districts with a reliable cold storage
system, an amount of vaccine to cover more than one vaccination campaign would be
sent at once, reducing the unit cost of transportation. But, there will still be a need to
assess whether the resulting cost reduction will compensate the additional costs of storage

that will be incurred when keeping the vaccine longer in the provinces.

On the other hand, production of higher amounts of vaccine with a longer lifetime may
allow IIAM to explore prospective markets such as the commercial poultry sector and
other countries in the region, with the possibility of increasing its levels of production in
order to meet the efﬁcient capacity. But, for that to happen, ﬁrst the vaccine needs to be

commercially registered.

Making the vaccine available does not seem to be a sufﬁcient condition for guaranteeing
positive and sustainable results. For example, the data set on chicken vaccinations and

number of chickens provided by CSL shows that in communities vaccinated by one

 

4 Quitino Lobo, intemet conversation with author, 20 February 2009

20

vaccinator, Eduardo Mondlane, the ﬂock size increases are outstanding and consistent
over the time (see Table Al on the Appendices). Personal communication with this
vaccinator and the leader of those communities suggested that involvement of the leader
in the campaign and the fact that he always talked about the importance of vaccination in
all the community meetings was the key to the success of vaccination in this area, such
that vaccination in those communities is currently driven by the demand from the
farmers. On the other hand, visits and groups discussion in another community in which
the leader did not promote the vaccine revealed the opposite in terms of sucCess of
vaccination, even though the vaccine was made available there. The lack of interest in
vaccination shown by the leader of that community supports the idea that involvement of
community leaders is a crucial aspect in inﬂuencing ND control and ensuring
sustainability of the program. Thus, community leaders and district administrators can be
useﬁrl resources to speed up the adoption of the vaccine in the district when they address

the problem of ND and the importance of vaccination in public community meetings.

Vaccination of chickens in the community is a seasonal job for the vaccinators, since it is
just carried out for a total of three months during a year (a month in each of three
campaigns). In the remaining months, vaccinators do not have anything to do related to
the ND program. Limited seasonal employment may constitute a problem for
sustainability of the vaccination program because the vaccinators can easily sacriﬁce the
vaccination program in favor of better prospects for any given season. This programmatic
vulnerability mainly applies to male vaccinators who have high propensity of migrating
to other places seeking jobs. Making vaccination a full-time job in the form of other

related local employment opportunities in periods outside the vaccination campaigns

2]

could enhance the performance and sustainability of the program by lessening the chance
that the vaccinators commit themselves to other engagements when needed for the
vaccination campaign. While the actual application of the vaccine does not require
technical skills, the campaigns require organizational skills and vaccinators become more

efﬁcient in time, such that constant retraining would lower the efﬁciency of the program.

22

CHAPTER III: LITERATURE REVIEW AND METHODS

3.1. Overview of Impact Assessment Studies

A large number of impact assessment studies have been undertaken in the agricultural
sector in the developing world. Most of them are focused on the evaluation of programs
related to development and extension of improved crop varieties (Bellon et al. 2005;
Marasas et al. 2004; McSween et al. 2006). Impact assessment of animal health
interventions has captured the attention of agricultural socio-economists (Pritchett et al.
2005; Johnston and Cumming 1991; Otte and Chilonda 2000; Bennett 2003; McDermott

et al. 2001).

The majority of impact assessment studies evaluate, in ﬁnancial or economic terms, the
efﬁciency of development and/or extension of technologies using proﬁtability measures.
In the ﬁnancial valuation, the beneﬁts and costs are valued in terms of market prices
unadjusted for distortions; in the economic valuation, prices are adjusted to reﬂect the

economic values of inputs and outputs.

In addition to understanding the economic efﬁciency of the allocation of the resources in
development programs, researchers also want to assess the impacts of these programs on
poverty, food security, equity and the environment, among other aspects (McDermott et
al. 2001). Given a wide range of development objectives, some development programs
may be accepted when they are not economically efﬁcient because they beneﬁt
vulnerable or target groups, or meet speciﬁc policy objectives. Even though the other

components of evaluations are important, this study only addresses the economic

23

efﬁciency component of impact evaluation. Incorporating other components of evaluation

would require additional work, beyond the scope of the current research.

According to Joshi (2003), efﬁciency analyses of development programs can be
undertaken at three levels according to the phase of implementation of the program,
namely, ex ante, ex post and concurrent assessments. The ex ante assessment is typically
used to quantify the returns of proposed projects and prioritize the investment agenda; it
is usually done to justify funding for different investment options (Alston et al. 1998;
Joshi 2003). The ex post assessment approach is used to determine the impact of past
investments, and it is usually undertaken when the research outputs and technologies are
largely adopted (J oshi 2003). The concurrent evaluation is the intermediary level and it is
done to identify the impediments for larger adoption of the technology. Its purpose is to
correct the gaps and provide feedback and ﬁne-tuning of the technology according to
stakeholders’ requirements (Joshi 2003). Given that the producers have started adopting
the I-2 vaccine, but the levels of adoption are still very low, this study falls under

concurrent analysis.

Many economic approaches are used to quantify the economic effects of investments in
development programs. Quantitative techniques are divided into three broad groups: 1)
econometric; 2) programming; and 3) economic surplus (Masters et al. 1996; Alston et al.
1998; Maredia et al. 2000). The choice of the methods depends on data availability, the
objectives of the research and/or the nature of the problem (including the economic level
involved and the complexity of the problem), the timing of the study and the availability
of resources such as time, money and analytical tools (Alston et al. 1998; Otte and

Chilonda 2000). Furthermore, Pritchett et al. (2005) argue that in the ﬁeld of animal
24

disease the choice of the analytical approach for impact assessment is inﬂuenced by the
focus of the impact analysis (production level, market prices, welfare), level of analysis

(geographically, marketing phase), and proposed policy alternatives.

The econometric approaches aim to estimate marginal productivity of investments during
a long period of time for a variety of research activities (Masters et al. 1996). They are
based on ftmctional forms such as production, proﬁt, and supply functions to estimate a
change in productivity due to investment (Maredia et al. 2000). In these approaches,
investment is treated as a variable, allowing for the calculation of the marginal rate of
return on investment (Alston et al. 1998). In general, econometric approaches assess the
change in marginal productivity at the macro level. The programming approach tries to
identify one or more optimal technologies or research activities from a set of options

given the constraints. This approach is based in the constrained optimization problem.

Economic surplus approaches include consumer-producer surplus, beneﬁt-cost, and total
factor productivity. Total factor productivity is the ratio of the total value of the product
and the value of the whole set of inputs, and it shows the residual left after incorporating
the contribution of inputs (Joshi 2003). Total factor productivity can be decomposed into

the contribution of research resource allocation and other determinants.

The consumer-producer surplus framework is the most widely used means of ex ante
evaluation of the impacts of agricultural research in a partial equilibrium framework
(Alston et al. 1998). This approach estimates aggregate returns on investment in a
particular project by measuring the change in consumer and producer surplus from a shift

of supply curve due to the adoption of the technology. In this approach, it is assumed

25

implicitly that adopters of the new technology have higher productivity relative to non-
adopters. This approach can also be used in an ex post evaluation framework; in this case,
the changes can be measured using elicitation methods or econometric analysis (Alwang
and Siegel 2003). In economic surplus methods, the stream of beneﬁts is compared to the
stream of costs, and indicators of the return of the investments such as beneﬁt-cost ratio,
net present value (NPV), internal rate of returns (IR) and payback period are used for

assessment.

This approach has an advantage of requiring less information than the other methods. The
information on how much the technical change shifted the supply curve and the
elasticities of supply and demand for the commodity, are the only empirical information
required. In addition, it is a relatively simple and ﬂexible approach that can be applied to
the broadest range of situations (Masters et al. 1996). It can be modiﬁed to account for
the effect of trade and price policy on the distribution of beneﬁts between consumers and
producers. However, the consumer-producer surplus methods do not provide clear cut
evidence about the impact of a program on measures of aggregate poverty (Alwang and

Siegel 2003).

The beneﬁt-cost method of analysis is a variant of the consumer-producer surplus
method. In this method, the economic surplus changes may not be explicitly measured,
but economic surplus calculations are implicitly incorporated when IRR, NPV or beneﬁt-
cost ratios are calculated to place a value on the extra output or the inputs saved (cost
reduction) because of the technology use (Alston et al. 1998). According to the same
authors, the advantage of employing beneﬁt-cost methods is that the values of demand

and supply elasticities are simply imposed on the analysis by assumption, eliminating the

26

need to obtain elasticity estimations. For the valuation of extra output due to a technology
use, a single market price is used and it is assumed that the supply curve is vertical and
shifts against a horizontal demand curve. For valuation of inputs saved (cost reduction) at
current level of production, a horizontal supply curve which shift down against a vertical
demand curve is considered (Alston et al. 1998). However, the implicit economic surplus
calculations ignore all regional and international prices effects due to technology use, as
well as any distributional effects (Alston et al. 1998). In this study the beneﬁt-cost
method is used for the analysis. In addition, sensitivity analysis of prices is performed in
order to evaluate how much the prices of chickens can decrease without affecting the

overall proﬁtability of the program.

3.2. Economics of Animal Disease

Disease represents a negative input in the process of converting resources or production
factors into products, goods and services available to people; it causes direct economic
losses for the producer and a potential loss of value in the view of the consumer (Otte and
Chilonda 2000). According to Bennett (2003), disease in livestock has seven main
economic impacts, namely: i) reduction in the level of marketable outputs; ii) reduction in
(perceived or actual) output quality; iii) waste (or higher level of use) of inputs; iv)
resource costs associated with disease prevention and control; V) human health costs
associated with disease or disease control; vi) negative animal welfare associated with
disease; and vii) international trade restrictions due to disease and its control. In addition,
FAO (2001) considers animal diseases an example of invasive species, and categorizes

six areas of their impact, namely, production effects, markets and price effects, trade

27

effects, impact on food security and nutrition, human health and the environment, and

ﬁnancial costs.

, A
Livestock

 
 
 
  

 

 

product
output
Px/Py
Attainable Health
Y With Disease
Y.

 

 

 

A

X X' Input the

 

Figure l: The Effect of Disease on Livestock Production (Source: Bennett 2003)

In the production area, the presence of the disease results in an inefﬁcient production
process, that is, livestock producers operate on a lower production function than in the
situation of absence of disease. Figure 1 shows the physical effect of disease on livestock
production (inefﬁciencies in production process) in terms of both output losses and input
use. In the presence of disease, instead of operating at point A on the “attainable health”
production function with output level Y and input use of X, producers operate at new
equilibrium point B, on the disease production function with lower output level Y'
and/or higher input use X' , given output and input prices, PY and Px , respectively. This

suggests that producers readjust to the changed relationship between inputs and outputs

28

caused by the disease, achieving a new equilibrium. The effect of animal diseases in a
given production system is a reduction of the efﬁciency with which inputs/resources are

converted into outputs/products, decreasing productivity (Otte and Chilonda 2000).

Furthermore, Otte and Chilonda (2000) classify the effects of the disease as direct and
indirect effects. The direct losses may occur when: (i) disease destroys the basic resource
of the livestock production process (mortality of breeding or productive animals); (ii)
disease lowers the efﬁciency of the production process and the productivity of resources
employed (e. g. reduced feed conversion), and (iii) disease may reduce the quantity and/or
quality of output. The indirect losses include: (i) losses through additional costs incurred
to avoid or reduce the incidence of the disease; (ii) detriment of human health well-being;
(iii) sub-optimal exploitation of otherwise available resources through forced adoption of
production methods which do not allow the full exploitation of the available resources
and/or through revenue foregone as a result of denied access to better markets. According
to Bennett (2003), the presence or absence of disease may have an effect, not only on
production, but also on both output and input prices. For example, if the majority of
producers adhere to the programs of disease control, the output supplied in the market
increases, and as a result, the price of the product in the market may decrease. According
to Pritchett et al. (2005) disease impacts are generally easy to identify but may be

difﬁcult to quantify.

3.3. Theoretical Framework for Impact Assessment of ND Control
The measurement of economic beneﬁts of agricultural technology consists of comparing
the beneﬁts in the situation with a particular technology to a counterfactual that

represents what would have occurred without the technology, known as the “with” and

29

“without” situations. The difference is the incremental net beneﬁts due to investment in
the technology (Gittinger 1982; Alston et al. 1998). For this particular study, the “with
situation” is represented by a situation where the I-2 vaccine is used for ND control and,
in the “without situation”, no vaccine is used. The effective control of ND increases the
efﬁciency of resource use in the affected population, through avoidance of chicken
mortality due to ND, and consequently shifts the supply curve for chicken outputs to the
right. The ND vaccine can be considered a productivity-enhancing technology, and as a

result of its use, the consumer and producer surpluses change (Figure 2).

Supply without Vaccine

   

, A
Price

Supply with Vaccine

 

 

 

 

 

 

 

 

 

 

Demand

 

 

Quantity

Figure 2: Impact of Vaccine Use in Village Chicken Production (Adapted from
Masters et al. 1996)

As illustrated in Figure 2, for producers, the impact of vaccine use is to avoid the losses
due to the disease, reducing production cost. The beneﬁts to the producers are

represented by area A. However, vaccine use may also reduce the price received by

30

producers due to the increase in quantity supplied. This reduction in price reduces
producer surplus by area B. The net gain of producers is area A minus area B. Producers
net gain is positive only when the demand is relatively more elastic than the supply. In a
situation where demand is relatively inelastic compared to supply, only a limited quantity
of the good is wanted and the producers lose from technical change. The consumers
always beneﬁt from technical change: they gain what is lost by the producers due to price
reduction (area B in ﬁgure 2), plus the economic surplus of the increased quantity (area
C) (Masters et al. 1996). For the society as a whole, the impact of vaccine use is basically
the gains Of areas A and C. The area B is just a transfer from the producers to the
consumers. Given that small scale producers of village chickens are also the main

consumers, this study only focuses on total social gains.

The net social gains of investments in a technology depend not only on the elasticities of
demand and supply, but also on the productivity increase generated by the technology,
the equilibrium price of the product, the adoption rate, the costs of research and
implementation of the program, and the timeframe between research and adoption.
According to Bennett (2003), three basic types of information are required to quantify the
beneﬁts of disease control: (i) the disease incidence; (ii) the magnitude, incidence and
distribution of disease effects, and (iii) the treatment and/or prevention measures

undertaken.

3.4 Description of Study Area
The primary data collection was carried out in Chibuto, a district located in the Southern
region of Mozambique, in Gaza Province (see Figure 3). It has an area of about 5,653

sz and an estimated population of 203,910 in 2005. Climatically, the district is

31

characterized by a low level of rainfall ranging from 400 to 600 mm per year in the
coastal area; For the interior area, the level of rainfall is even lower (MAE 2005). In
general, the seasonal rainfall trend follows unimodal patterns with high rainfall levels
observed between November and March; however, the rainy season is quite irregular in
terms of starting time, length, levels and distribution. The Limpopo River has an
extremely important role in the performance of the agricultural sector in this district; it is
used for irrigation, enabling agriculture in large areas which would be unsuitable

otherwise.

The estimated poverty incidence in the district was about 60% of households in 2003
(MAE 2005). Agriculture is the basic economic activity in the district and it is dominated
by small-scale farmers who occupy about 52,000 ha (9 percent of total cultivated area).
The average cropping area per household is about 1 ha. The main subsistence crops are
maize, rice, cowpea, cassava and vegetables, which are generally produced in an
intercropped system using local varieties and traditional tools. Large-scale private sector
holdings occupy a signiﬁcant portion of the most productive land and employ about 15
percent of labor in the district. In some areas, animal traction is used in agricultural
production activities. The main cash crops are sugar-cane, cotton, tobacco, and cashew.
Cashew is a very important cash crop in the district; Chibuto is the second biggest
supplier of raw cashew to the cashew processing industry in the province. About 25,000
people keep small ruminants and swine. More than 40,000 people keep poultry, and most
of them are kept under a free-range scavenging system. Non-farm activities such as wood
and coal extraction and ﬁshing also contribute greatly to the income of thousands of

households in the district (MAE 2005).

32

 

 

  
   

      

 

 

  
   

  
    

   
 

 
 

 

 

 

I N j—f~ g-"I I": 7 4”.» . " l” u.
. , ,1 aux .1 u ‘ I‘» '.
ILHA‘IIICOM ’ ‘ " .‘ «.1. .. x
LAGO,NII:ISSA “ / cxeoje‘fétécgw; .
I. I‘\'I»...— . . , n. I . Nun“ F“ I .. ‘ “
. .' FINE; ,_.— 1' A“. ﬂiwcuméﬁ‘wr a
.. I~;5-‘ ’
'r ,, I . _
1513:05qu
t ‘ x v!
, ,,
. / I
— ‘7! ‘C-WGHILJJII,
u-u.‘ I A .
~ — W193 SOFALA ' ‘,
r .~ “N ' mu I...
P. .
Republic of Mozambique
/ Administrative Map
,/
Ndl’u- 111.4511.
-‘ (’Q;"(7CL7NM—Nﬁ
I: u.- -r
marmmmma‘nmams Legend. 345 09 ID 27 36
- Emu
m:mmmmoim “WWO"
V3804
mum Chibuto District !

 

 

 

 

 

 

 

Figure 3: Location of the Study Area (Source: Administrative Map of Mozambique)

33

3.5. Description of Data and Methods of Collection

A formal household level survey was conducted in July of 2007. The stratiﬁed sampling
technique was employed to select 226 households who raised village chickens in four
villages. In the group of interviewed households, 127 are participants and 95 are non-
participants in the vaccination program. One of the villages covered by the survey did not
implement a chicken vaccination program at all. The non-participant households were
selected randomly within the Village, based on the list provided by the local leaders.
Participant households were selected based on a vaccination data set provided by
Corridor Sands Limited (CSL) Company, which is one of the partners in the
implementation of the ND prevention program in Chibuto district. This survey collected
household information on demographics, assets, chicken production systems, marketing,
constraints faced in chicken production in the district, cost of the vaccine, and farmers’
perceptions of the usefulness of the vaccine. This information is valuable for evaluation

of the ND control program.

Additional data was collected in August of 2008. Semi-structured group interviews of the
village poultry producers in the four previously visited villages were conducted. The
objective of these interviews was to gather relevant information on village chicken life
cycle, growth trajectory, off-take, likelihood of ND occurrence in the region, and
dynamics of ﬂock size versus management and disease incidence. This information was
useful in modeling the village poultry production system and in determining the beneﬁts

Of ND control at the farm level.

TIA data sets provide additional information for farm level analyses as well as aggregate

analyses, including estimation of adoption rates and estimation of the levels of poultry

34

off-takes. Census data are also used to estimate the beneﬁts at aggregated levels.
Furthermore, estimates of costs involved in production and transfer of the vaccine are
based on information provided by IIAM, MINAG, SANDCP, CSL and NGOS involved
in the ND control program. These costs involve the production of the vaccine (salaries,
infrastructure, equipment, and administrative costs) and the costs of extension programs

required to distribute the vaccine and to speed up adoption.

3.6. Measurement of Impacts of ND Control in Village Poultry: VIPOSIM

Village poultry production systems are complex, thus their studies require insight in the
dynamics of the production system (Asgedom 2007; Udo et al. 2006). According to the
same authors, temporal variation in village poultry is a result of interaction of several
factors, including ﬂock mortality, egg production, reproduction, bird off—take and egg
off-take. In addition, the dynamics of village poultry are inﬂuenced by random
phenomena. Thus, measurement of the impacts of interventions in this complex and
dynamic system requires research tools that integrate the different processes and
management options involved. Simulation models are one type of research tool that can
be used to model the impacts of interventions in poultry production systems, especially if

probabilistic effects can be incorporated in the analysis (Asgedom 2007).

VIPOSIM (VIllage POultry SImulation Model) is a dynamic simulation model used by
Asgedom (2007) to assess the impacts of different management strategies in a poultry
production system. VIPOSIM was developed by a team from Wageningen University in

the Netherlands, and it was validated with data from Tigray, Ethiopia.

¥

5 .
Consumption and sales

35

Conceptualization of VIPOSIM includes Six processes related to production and
utilization of chickens: ﬂock mortality, ﬂock off-take, egg production, egg loss, egg off-
take and reproduction (incubation and hatching). VIPOSIM performs calculations in time
steps which represent reproduction cycles, i.e., a time needed by a broody hen to produce
and hatch eggs, and then to rear chicks. Each step has a length of a season of 3 months
and the maximum number of steps in the model is 12, which corresponds to a period of 3
years (Asgedom 2007). VIPOSIM was programmed in Microsoft Excel® and integrates
quantitative relationships of various elements of the system in a series of mathematical
equations. (For more details about mathematical procedures of conceptualization of

VIPOSIM see Appendix A 1.).

In the VIPOSIM model, a ﬂock is categorized in ﬁve categories of chickens according to
age and gender: i) the chicks group includes all chickens with age less than or equal to
three months; ii) cockerels are male chickens with age superior to three months but not
yet adult; iii) pullets are female chickens older than three months but not yet adult; iv)
hens are female adult chickens; and V) cocks are male adult chickens. According to the
farmers in Chibuto district, chickens are adult at the age of six months. There is a
variation of some of the input parameters across chicken categories. These variables
include the ﬂock size, mortality rate due to disease, mortality due to predation, and bird

off-takes.

36

MODEL

OUTPUTS

 

Bird
Manageme'"
nt

Egg
Production":
Values

Reproducti

on Values

Egg

Manageme:

nt Values

 

......................... >l A... j.

 

 

 

 

1

Net Flock Size and Structure

 

 

 

 

 

 

 

 

 

 

 

I

 

3'") Eggs Produced

-> Egg Off-take
~> Chicks Hatched

m) Flock

 

 

 

 

Figure 4: Schematic Representation of Sequences of Events in the VIPOSIM for a
Reproduction Season (adapted from Asgedom 2007).

37

Figure 4 illustrates the sequence of the events in VIPOSIM; the broken arrows represent
inputs and outputs variables of the model. The chicken production and utilization
parameters are introduced in VIPOSIM as input variables. Those variables include initial
size and composition of ﬂock, mortality rates, bird sales and consumption rates, egg
production, reproduction parameters (incubation and hatching), egg sales, egg loss, egg
consumption rates, and bird off-take limits. The inputs variables are believed to be related

to agro ecology and husbandry conditions, and they differ across the seasons.

Economic input parameters such as prices of birds and eggs and costs of production are
also introduced into model. The prices of birds and eggs used in this study are based on
TIA data. VIPOSIM categorizes cost input parameters into costs of labor and costs of
intervention. As the output, the VIPOSIM model gives the numbers and values of bird
off-take and egg off-take, and the ﬁnal composition of the ﬂock for each season during

the three-year period of simulation.

This simulation model has an advantage of allowing for incorporation of random
phenomena by using random numbers as coefﬁcients of standard deviations in the
simulations. It can be transformed easily into a deterministic model by setting all the
standard deviations to zero. However, the model is very restrictive because it assumes

that all the random parameters follow a normal distribution, which may not be true.

38

3.7. Adaptation of VIPOSIM to the Context of the Study

Given the limitations of VIPOSIM, the objective of the research, and data availability,
various changes were made in order to accommodate the model to the current context of
the study. Many changes were based on knowledge of Mozambican village chicken

production systems for Village chicken.

In this study, chicken production and utilization parameters are developed from the
household survey and farmer group discussions in Chibuto as well as from the National
TIA data. These parameters are then used as the inputs of the adapted VIPOSIM model.
The VIPOSIM model requires information on production and utilization, some of which
is not available for Mozambique. In the case of insufﬁcient information, original
parameters developed for the VIPOSIM model are used, based on the assumption that
production and utilization parameters are similar for village chickens in Mozambique or
elsewhere. For example, during the process of data collection in Mozambique, categories
of chickens were not taken into account for the parameters of ﬂock size, mortality and
bird off-takes. Thus, the parameters needed for each category were generated based on
general information collected and the relationship between the parameters across the
categories in the default input data of VIPOSIM. These data were used by Asgedom

(2007) to validate the model.

In the design of VIPOSIM, the input parameters of mortality are categorized in three
groups: mortality due to the disease, mortality due to predation, and mortality due to
other reasons. However, due to the nature of the data collected and objectives of the

analysis in this study, only two categories of mortality parameters are deﬁned: mortality

39

due to ND or mortality due to other reasons (combining predation and other causes). For

the third category designed in the model values of zero are entered.

In addition, VIPOSIM categorizes the input parameters for rates of bird and eggs off-take
into two groups: sales and home consumption. In the adapted VIPOSIM, bird off-take
input parameters are treated as one broad category of off-take, and TIA 02, TIA 03, TIA
05 and TIA 06 are used to compute the off-take rates. It is assumed that there is no Off-
take for chicks. Also, only consumed eggs are considered; the sale of eggs in Chibuto is

negligible.

Since in the village poultry production system there is very low use of production inputs,
only the cost of vaccination is included in the simulations as an additional cost of chicken
vaccination; it is assumed that the additional costs of labor or other inputs resulting from

chicken vaccination is negligible.

Similar to the original model, it is assumed that during a year, some input parameters
vary across the seasons. Their variation is based on Mozambican farmers’ perceptions of
the best and worst periods and the range of values they provided. In Chibuto, farmers
reported occurrence of ND over the last two quarters of the year, thus inputs of mortality
due to ND are only incorporated for the third and forth quarters in the adapted VIPOSIM.

For the other quarters zero values are entered.

Also, farmers reported seasonality in the losses due to predations and other causes.
According to them, in the hungry season (wet season, from October to February) there is
a lack of feed, and chickens tend to go far from the houses scavenging for food, making

them more vulnerable to predators than in the dry season. In addition, because of food

40

scarcity there is higher occurrence of theft. Moreover, the relatively denser vegetation in
the villages during the wet season harbors predators. Thus, the last quarter of the year has
the highest rates of bird losses due to predations and other causes, while the second
quarter. of the year has the lowest rates. The other two quarters have intermediary values

(Table 3).

Table 3: Rates of Mortality Due to Predation and Other Causes

Predation & Other Losses in Chibuto
Season Chicksj Pulletsl Cockerelsl Hensl Cocks AveraLe

 

 

1 42.3% 5.1% 3.1% 11.2% 1.5% 20.5%
2 32.5% 7.2% 9.0% 6.4% 0.4% 16.0%
3 39.2% 11.1% 6.3% 4.5% 2.4% 18.4%
4 49.2% 7.8% 1.7% 6.0% 6.6% 22.1%

 

 

Data Source: Group Discussion in Chibuto in 2008

Because it is believed that the dynamics of poultry systems will differ for “with-control”
and “without-control” situations, these two situations are simulated separately. However,
to capture only the impact of ND control for a given scenarioé of analysis, the input
parameters entered are the same for both situations, except for the level of ND mortality

and the cost of the ND control.

VIPOSIM was designed to generate both direct beneﬁts (bird and egg off-take) and
indirect beneﬁts (manure and the value of having birds in case of urgent cash or social
needs). However, in this study, only direct beneﬁts resulting from avoidance of bird loss

are considered for the analyses.

 

6 Different scenarios are deﬁned in the study based on the ND mortality rates

41

The initial ﬂock size is expected to have a positive effect on the size Of the beneﬁts of
vaccination obtained at farm level; the bigger the initial ﬂock size, the larger is the
number of chickens expected to be saved by vaccination, and consequently, the higher
are the expected beneﬁts of vaccination. Thus, for determination of overall beneﬁts of
ND control at farm level in the situation where initial ﬂock size is not treated as random
variable, there is a need to ensure that the initial ﬂock size incorporated as an input in
VIPOSIM is a representative ﬂock size. In the event of asymmetric distribution of initial
ﬂock size, different categories of initial ﬂock sizes are created. For a given scenario and
situation (with or without-control), a separate simulation was performed for each of the
initial ﬂock size categories to get the beneﬁts for each category. Then, the overall beneﬁt
of ND control of that scenario is given by the sum of the beneﬁts of each category
weighted by the respective frequency. This is one of the key modiﬁcations in the adapted
VIPOSIM. The identiﬁcation of ﬂock size categories and their use is discussed in depth

in section 3.8.

In the stochastic simulations, to overcome the restriction imposed by the assumption of
normal distribution of the parameters in the stochastic simulations of VIPOSIM, @Risk
component is incorporated. VIPOSIM is set to deterministic mode, and @Risk software
is used to generate the distribution of the results based on the best ﬁt probability
distribution to data of the variables treated as random variables (in this study the random

variable is ND mortality rates).

3.8. Analysis of the Impact of Chickens Vaccination
To assess the impact of vaccination, indicators of economic efﬁciency, such as Net

Present Value (NPV) and Internal Rate of Return (IRR), are estimated. These indicators

42

can be estimated in different levels of aggregation (farm level, district level and national
level), and data on costs and beneﬁts in each level are required. In addition, an
appropriate discount rate is required for the analyses. However, in this study, the
proﬁtability measures (IR and NPV) are only generated for aggregated levels of
analysis where the time value of money becomes important because the aggregate
analysis covers a long period of time and investments may change over the time. For the
household level analysis, since the value of investments in the vaccine is very low and the
costs do not change substantially from year to year, the results are presented in terms of
annual net beneﬁts associated with vaccination7. The net beneﬁts are given by the
difference in the off-take values (bird and eggs) from VIPOSIM results with and without

ND control and vaccination costs.

Investments in production and diffusion of ND control technology generate beneﬁts
when the technology is adopted by farmers. The beneﬁts of I-2 vaccine use could be in
terms of an increase in the quality of chickens produced, increased ﬂock size and/or
increased off-take. Regardless of the type of beneﬁts, their amount at the aggregate level
depends on the level of adoption of the technology. Thus, estimation of the levels of

adoption of the technology is one of the prerequisites for beneﬁt-cost analyses.

Due to the low level of quality differentiation in the market for Village chickens in
Mozambique, and to difﬁculties in getting quality improvement data on village chickens
resulting from vaccination, in this particular study, only beneﬁts related to increased off-

take of chickens and eggs are considered. There are additional beneﬁts not addressed here

 

7 The results at farm level are presented in terms of annualized total present value of the net beneﬁts over
the 12 seasons of simulation

43

(manure, social roles of chickens, etc.) that may increase due to vaccination, but they are
not considered in the study in part because it is difﬁcult to assign a monetary value to

these beneﬁts.

3.8.1. Measurement of Annual Net Benefits of ND control at Farm Level

The annual net beneﬁts of using ND control technology in a household depend on the
size of the initial ﬂock, the likelihood of being affected by ND, the mortality rates due to
ND, and the rate of vaccination against ND in that particular year. In the model, at the
household level, the rate of vaccination is given by the ratio between the numbers of

chickens vaccinated and the Size of the ﬂock in each vaccination campaign.

Estimation of the rate of vaccination at household level is complicated because it requires
very detailed information on the ﬂock size and numbers of chickens vaccinated in each
campaign during the year, and such data are hard to obtain. Rural households in
Mozambique are likely to vaccinate less than 100 percent of their ﬂocks because, among
other reasons, it is difﬁcult to catch all the chickens due to the feral nature of village
chickens. According to the results of the survey in Chibuto, about 20 percent of the
interviewed households who participated in the ND control program did not vaccinate
100 percent of their ﬂocks in their last vaccination. Not being able to catch all the
chickens on vaccination day was the main reason, stated by 90 percent of the households,
for incomplete vaccination. Nevertheless, estimation of the rate of vaccination at
household level may not be a big concern when a large proportion of the ﬂock is
vaccinated because of the nature of the 1-2 vaccine. 1-2 is a live vaccine and according to
Alders and Spradbrow (2001) and M. Harum (personal communication, August 20,

2007), the virus can be transmitted from vaccinated to non-vaccinated chickens in close

44

contact, protecting all the ﬂock, including non-vaccinated chickens. Furthermore,
vaccinating village chicken ﬂocks at 3-4 month intervals also provides protection for
newly hatched chicks (Alders and Spradbrow 2001). Thus, to simplify the analysis, in the
“with technology” scenario it is assumed that all birds were vaccinated, i.e. the rate of

vaccination at household level is 100 percent.

Adapted VIPOSIM Simulations8 are performed for the village poultry production system
in order to estimate the beneﬁts of vaccination at farm level. The beneﬁts of chicken
vaccination at farm level are determined by applying partial budgeting procedures to the
results of simulations. For each scenario considered in the analyses, two simulations are
performed: one of them represents the “without technology situation” in which the
without-control ND mortality levels are incorporated, and the other simulation represents
the “with technology situation”, where vaccination costs are incorporated, and ND
mortality rates are reduced to with-control levels. In the “with technology situation”, the
mortality due to ND is reduced by 80 percent (20% of “without-control” ND mortality
levels). The reduced mortality levels are based on the ﬁndings in ﬁeld trials in
Mozambique (Dias et al. 2001), which demonstrated that in the ﬁeld, vaccination reduces

but does not eliminate mortality due to ND in ﬂocks vaccinated.

Because in the presence of ND there might be a dramatic drop in number of egg laid per
clutch, an increase in egg production is another component that should be considered in
modeling the beneﬁts of ND control, by introducing higher reproduction parameters in
the “with-control” situation than in the “without-control” situation. However, there is no

information on how much egg produced per clutch increases as a result of ND control.

 

8 This refers to simulations performed with modiﬁcations described in section 3.7.

45

Furthermore, protection of non-vaccinated chickens by horizontal transmission from
vaccinated chickens is a potential positive extemality related to chickens vaccination, but

this is also not included in VIPOSIM due to lack of information.

Both stochastic and deterministic simulations are considered in the analysis. In the
deterministic analysis, all the standard deviations in VIPOSIM are set to zero; while in
stochastic analysis for this research, ND mortality rates are treated as a random variable,
VIPOSIM is set to deterministic mode, and @Risk software is used to generate a
distribution of the net beneﬁts based on the best ﬁt probability distribution of data

collected on ND mortality rates.

3.8.1.1. Considerations of the Effect of the Flock size on the Benefits

For a given overall level of mortality due to ND in a region, the mortality rates are
expected to vary across the ﬂock sizes. For households with bigger ﬂock sizes, the
mortality level is expected to be relatively higher than in households with smaller ﬂock
sizes. The reason for this expectation is that ND is transmitted through physical contact
between chickens, and the bigger is the ﬂock size, the higher is the contact between

chickens, and the higher are the chances of infecting each other.

Therefore, based on the distribution of the size of ﬂock in the data collected in Chibuto
distriCt (see Figure 5), in simulations of a given overall level of mortality due to ND in
the community, different levels of mortality are used across the different ﬂock sizes (see

Table 4).

46

Flock Size Distribution in Chibuto

Frequency
15
I

10
I

 

 

 

 

I I l I
50 100 150
Households Size of Flocks

Figure 5: Distribution of Flock Size in Chibuto (Data Source: Chibuto household
survey in 2007)

In addition, data collected in Chibuto district show that the ﬂock size has a skewed
distribution, with many households having small ﬂock sizes (Figure 5). This suggests that
the conventional mean is not a good measure of a typical ﬂock size. To overcome this
problem of asymmetric distribution of the ﬂock size across the households in the
deterministic as well as stochastic analyses, six categories of households were created
based on the size of their ﬂock, and the mean and standard deviation of ﬂock size. The
proportion of households in each category was calculated based on survey data collected
in Chibuto district (Table 4). The overall incremental beneﬁt at household level is given
by the sum of individual category beneﬁts weighted by the respective proportion of

households.

47

Table 4: Definition of Households Categories Based on the Flock Size Distribution
in Chibuto

 

 

Size of Flock Category Size of Flock Proportion of Mortality Rates
Averag SD Households Level SD
Category 1 (0-5 Chickens) 2.4 1.6 44% 33% 28%
Category 2 (6-10 Chickens) 7.8 1.4 27% 80% 4%
Category 3 (1 1-15 Chickens) 13.0 1.7 12% 88% 2%
Category 4 (16-20 Chickens) 18.1 1.5 6% 92% 91%
Category 5 (21-25 Chickens) 23.5 1.6 4% 94% < 1%
Categ016 (26 or more Chickens) 41.0 15.5 6% 96% 1%

 

Data Source: Survey in Chibuto District in 2007 and Group Discussion in 2008

3.8.1.2. Determination of Incremental Beneﬁts of Vaccination

From the outputs of adapted VIPOSIM, the net beneﬁts in each season for a given initial

ﬂock size category are computed as:

NB. = 3:” —Bl’0 —C. (1)
I I l l

where NB]. is the annual net beneﬁts of ND control at farm level in season i, in (MTN);9

Biwi are the beneﬁts (total value of off-take in MTN) for the “with-control” situation in

season i; BI?” are the beneﬁts (total value of off-take in MTN) for the “without-control”
situation in season i; Ci are the additional costs related to technology use that were

incurred in season i (costs of vaccination in MTN.)

From the net beneﬁts determined for each season, and assuming that the interest is

compounded quarterly given that a production season takes about 3 months, the total

 

9 There are approximately 25MTN per US Dollars
48

present value for the whole period of three years in a given category of ﬂock size is

calculated as (Hoy et al., 2001):

2 NB.
_ 1
PV .. Z (2)

1'
1:] (1+5)
4

where PV is the total present value of the net beneﬁts obtained during the three years of

.—i

simulation in MTN; NBi is the net beneﬁt in season i in MTN; and r is the discount rate.

A discount rate of ﬁve percent is used; according to T. Walker (personal communication,
May 11, 2009), this rate is increasingly common in the literature“). The annual net
beneﬁts at farm level for a given initial ﬂock size category are determined as (Ross et al.

zoosy

 

ANB=PV* M (3)

 

 

where ANB is the annual net beneﬁt in MTN; T is the number of years, and T=3 in this
based on the length of a simulation in VIPOSIM; and r is the discount rate. The overall

annual net beneﬁt of chicken vaccination per household is computed as:

6
_ a:
VB — lg] ANBI. Prl. (4)

 

'0 Thomas Walker, email to author, 11 May 2009
49

where VB is the overall annual net beneﬁt per household in MTN; ANB]. is the annual
net beneﬁt for a category of ﬂock Size i in MTN; and Prl. is the proportion of households

in the category of ﬂock size i.

3.8.1.3. Risk Analysis

There is uncertainty about some Of the relevant variables used for computation of the
beneﬁts at farm level. One such variable is the without-control ND mortality rate in
Mozambique, for which information is limited. For the deterministic analyses, to deal
with the uncertainty involved, the sensitivity of the annual net beneﬁts at farm level to the

assumptions about levels of ND mortality is evaluated.

Twenty scenarios of without-control ND mortality levels were deﬁned (19 hypothetical
plus the base scenario, varying from 5 percent to 95 percent, with the base at 63 percent)
and their respective beneﬁts estimated. Since for a given overall mortality level, the
mortality level is expected to vary across the categories of households and chickens
(larger ﬂock sizes categories tend to have relatively higher rates of infection), for each
assumption of overall level of mortality, mortality level in each category of household is
calculated based on household data set from Chibuto, then the calculations across birds
categories are based on the relationship of disease mortality rates between chickens
categories from Asgedom (2007). For stochastic analysis, to deal with risk, the ND

mortality rate is treated as random variable.

Another key variable in the beneﬁts estimates is the price of chickens. This becomes even
more important to evaluate because there might be price effects due to technology use.

Thus, sensitivity to prices is also evaluated for deterministic analysis. The analysis

50

determines how much the price of chickens can decrease without affecting the overall

farm level proﬁtability of vaccination.

3.8.2. Measurement of Returns to Investments in ND Control Program

Aggregate analysis of the returns to investments in ND control is performed in two
levels: the district level and a more aggregated level of analysis which includes all the 44
districts where the program is currently implemented. All the analyses are heavily based
on data collected in Chibuto. Data such as the costs of extension, the parameters of
chicken production and the levels of ND mortality among other information used in the
aggregate analysis may be different from one district to another, but data for each of the
districts covered were hard to obtain. Thus, parameters from Chibuto are mainly used as
general parameters to get overall values. Both stochastic with ND mortality levels treated
as random variable and deterministic analyses are also considered in those levels of

analysis.

3.8.2.1. Deterministic Measurement of the Returns to ND Control

At the aggregate levels of analysis, the beneﬁts of vaccination for each year depend on
aggregate adoption rates, the magnitude of the shift of the supply curve due to the use of
the technology, and elasticities of supply and demand. Logically, the increase of supply
would inﬂuence the price in the market; however, the data on elasticities are not available
for those calculations. Therefore, to simplify the analysis, a constant chicken price is
used, and the sensitivity of the results to chickens prices is evaluated. The beneﬁt-cost
method is used for the measurement of the beneﬁts at these levels with the underlying
assumptions that demand is perfectly elastic, supply is perfectly inelastic, the demand and

supply curves are linear, and the shift in the supply curve due to the use of the technology

51

is parallel. Then, a sensitivity analysis is conducted on how much the price can fall

without affecting the overall proﬁtability of the program.

The likely extent of adoption of technology has a strong inﬂuence on the efﬁciency of
investments; the beneﬁts will be zero if the technology is not adopted. On the other hand,
due to the time value of money, the longer the technology takes to be adopted, the higher
are the initial research and development costs relative to the beneﬁts, and consequently

the lower are the net beneﬁts (Batz et al. 2003).

There are two possible ways that the adoption rates can be considered for the analysis: in
terms of the proportion of farmers using the technology or in terms of the pTOportion of
chickens vaccinated. However, as previously mentioned, it is very difﬁcult to obtain
detailed and complete data on the ﬂock size and number of chickens vaccinated in all
campaigns. AS a result, it is difﬁcult to implement the second approach. Another reason
to use the proportion of farmers using the vaccine as the adoption rate is because the
farmers are not making the decision to vaccinate less than 100% of their chickens; they
vaccinate as many as they can catch when they have access to the vaccine, and some of
non-vaccinated chickens in a ﬂock may also get protection against ND through
transmission of vaccine virus from the vaccinated chickens. Thus, in this study the
diffusion of the technology is considered in terms of the proportion of farmers using the

technology on their ﬂocks.

The other component to consider in the determination of aggregate net beneﬁts of I-2
vaccine production and extension is the cost. The types of costs included are different in

the two aggregate levels of analysis. The district analysis will only include vaccine

52

extension costs while the more aggregated level of analysis will also include vaccine

production costs.

For the estimation of proﬁtability measures, the length of the period of analysis and the
discount rates are the other components that need to be considered. Taking into account
that 20 years is a reasonable lifetime for the equipment (vaccine production and
refrigeration equipment), this study covers a period of 20 years starting from 1999, the
year production of I-2 vaccine started in Mozambique. According to Gittinger (1982) for
economic analysis the “opportunity cost of capital”, i.e. the return on the marginal
investment made that uses up the last available capital, is the best discount rate. However,
no one really knows what the opportunity cost Of capital is (Gittinger 1982). In this study
a social discount rate of ﬁve percent per annum is used. According to T. Walker
(personal communication, May 11, 2009) this rate is increasingly common in the

literature. The discount rate is incorporated when estimating these beneﬁts over time.

3.8.2.1.1. Estimation of Beneﬁts of ND Control at the Aggregate Level
The annual beneﬁts of vaccination at aggregate levels are given by aggregating of the net
beneﬁts from all adopters of the technology in the level of analysis. The annual beneﬁts

are computed as:

8’] =ANB*HA{ =ANB*ytj *HCtj (5)

where Btj is the annual beneﬁts of vaccination for the aggregate level j during the year t

(in MTN); ANB is the overall annual incremental beneﬁts per household in MTN; HA}

is the number of households using the vaccine in the level of analysis j during the year t;

53

ytL is the proportion of households who use the vaccine in level j during the year t;

HCIL is the number of households who raise chickens in the level j during year t; and j

is either the district level or the sum of all 44 districts where the vaccine is currently used.

Estimation of A doption Patterns

The adoption rate at a certain point of time is given by the ratio between the number of
households participating in the ND control program and the total number of households

who raise chickens.

n
y, = N (6)

where n is the number of farmers who vaccinate their chickens and N is the number of

farmers who raise chickens.

Since the adoption rates over time are only known in Mozambique for a limited number
of years, the diffusion of the technology through time will be estimated using the
methodology described by Morris and Heisey (2003). The estimation of diffusion
patterns is based on the assumption that the cumulative proportion of households
adopting I-2 vaccine follows an S-shaped or logistic pattern. This pattern is commonly

used in adoption studies and mathematically is described as:

Y = L (7)
t 1+e-a_bt

 

where Yt is the cumulative percentage of adoption at time t; L is the upper bound of

adoption (ceiling); ais a constant related to the time when adoption begins; and b is a

54

constant related to the rate of adoption. According to Morris and Heisey (2003), given

sufﬁcient Observations on Y t , it is possible to estimate the unknown parameters L, a

and b using non-linear regression methods. In cases where at least three observations of

Y t ,are available, and L can be obtained independently, an ordinary least squares (OLS)

regression can be used to estimate a transformed version of the logistic curve:

Y
In[ t ]=a+bt (8)
L—Y

Logistic diffusion curves are appropriate for Situations where there is a large and non-

 

homogenous population of potential adopters who have unequal access to information
about innovations and who differ on their willingness to innovate (Morris and Heisey
2003). This condition may hold for the case of diffusion of the I—2 vaccine for village
chickens, which involves a large number of small-scale farmers with different levels of
access to the vaccine. In this study, the diffusion curve is estimated using a few adoption
points based on TIA data (TIA 2003, TIA 2005 and TIA 2006). Different ceiling levels L
are assumed. Sensitivity to those assumptions is evaluated. However, provincial adoption
rates are used instead of district rates, because TIA data estimates are not representative
at a more disaggregated level. For Chibuto district level analysis the adoption rate is
computed based on adoption in Gaza province, while for the more aggregated level of
analysis, adoption rate across the ﬁve provinces (Gaza, Inhambane, Tete, Zambézia and
Nampula) is used. The results of estimation of the rates of adoption are presented in

Tables A 2 and A 3, and Figures A 1 and A 2 in the Appendices.

55

Projection of Number of Households Raising Chickens

Census and TIA data are used to compute the number of households raising chickens in a
given level of aggregate analysis. First, based on Census data on household numbers in
1997 and 2007, the number of households is projected for each year over the period of

study using the formula:

H t = (1+k)Ht_1 (9)

where H t is the number of households in year t; H t—l is the number of households in

the year previous to t; k is the rate of increase in the number of households from year to
year. This rate is assumed to be constant over the period of analysis. Estimations resulted
in a rate of 3.9 percent at the district level and three percent for the overall area covered

by the vaccination program.

Then, the number of the households raising chickens for each year in a given level of

analysis is determined as:

HCI =aH (10)

t

where HC t is the number of households raising chickens in year t and H t is the

number of households in year t; a is the average rate of ownership of chickens in the
speciﬁc level of aggregation (computed from available TIA data sets: TIA 2002, TIA
2005 and TIA 2006). a is assumed to be constant over the years of analysis. See the
results of estimation of the total number of households and households raising chickens

in Table A4 and ﬁgures A 3 and A 4 in the appendices.

56

3.8.2.1.2. Estimation of the Costs of the Vaccination Program

To estimate costs of extension of the vaccine at the district level, information provided by
CSL and other institutions involved in the program is used. The costs of equipment for
vaccine conservation, extension material and vaccinators training were estimated based
on aggregated information provided by CSL. Given the problem of limited and seasonal
employment of vaccinators, there is a high risk of loss of trained vaccinators; thus, it is
assumed that there will be a major vaccinator training every three years. Also, the regular
extension costs associated with transportation of the vaccine to the communities during
the vaccination campaigns or for visits for assistance in other periods, as well as costs of
transportation of the vaccine from SPP to the district are included. In addition, salaries of
staff involved in the vaccination program are included, and it assumed that these workers
spend a quarter of the year devoted to the vaccination program during each year. See

Table A 5 for estimated costs of extension.

The extension costs at the national level are estimated through aggregation of the district
level extension costs over the 44 districts involved in the vaccination program. It is
recognized that the costs of extension may vary considerably from one district to another
according to the size of the population and distances covered. However, it is hard to
obtain cost data; so values estimated here are used just to give a general picture. At this
level of aggregate analysis, costs of production of vaccine (including quality control and
packaging) and transportation of the vaccine to the provinces are also included.
Computation of these costs is based on the estimates of SANDCP (2003). This author
estimated the costs per dose of the vaccine in Mozambique. See the estimated costs in

Table A 5, in the Appendices. The total costs of production and transportation of the

57

vaccine to the provinces per year is based on the amount of the vaccine produced in the
respective year. The amount of the vaccine produced per year is based on information
provided by IIAM. For the years in which this information is not available, it is estimated.
based on average yearly increase of production. For number of doses of vaccine produced

per year, see Table A 6 in the Appendices.

3.8.2.1.3. Measurement of Economic Proﬁtability
The proﬁtability analyses are carried out to assess the economic viability of the program.
In this study, NPV and IR are estimated. As previously mentioned, to account for the
time value of money during the 20 years of analysis, a discount rate of ﬁve percent is
used. NPV is equal to a ﬂow of net beneﬁts (beneﬁts minus costs) discounted by the
discount rate:

B —C

NPV: —’—’— (11)
[:0 (l'l’f'),

where B t is the beneﬁt in year t; C t is the cost in year t and r is the discount rate. If the

NPV is positive, then the investment is considered proﬁtable.

IR is the return to the money invested; it is the rate of return that makes the present
value of beneﬁts equal to the present value of costs (NPV=O). When the IRR equals or
exceeds the true opportunity cost of capital, project retums are large enough to cover all
operating costs, pay back the principal on the capital invested in the project, and give an
annual average return equal to IR. The investment is considered proﬁtable if and only if

the resulting IR is higher than the opportunity cost of capital.

58

Br -Ct

NPV = —— =
,zo (14.1121?)1

(12)
3.8.2.1.4. Sensitivity Analysis

Sensitivity of the estimated proﬁtability measures to adoption ceiling level and ND
without-control mortality levels is evaluated. For this evaluation, the annual net beneﬁts
estimated at farm level for each of the 20 scenarios of without-control ND rates are used
to estimate the aggregate level proﬁtability. In addition, analysis is completed to evaluate
the degree to which the price of chickens can decrease as a result of technology use

without affecting the overall proﬁtability of the program.

3.8.2.2. Stochastic Analysis of the Returns to ND Control

The procedures for stochastic analysis of the returns to vaccination are similar to those
used in the deterministic analysis. However, in the stochastic analyses, I use the
stochastic annual farm level beneﬁts that were determined, treating ND mortality levels

as random variables, rather than using deterministic annual farm level net beneﬁts.

59

CHAPTER IV: RESULTS OF ANALYSES OF THE IMPACT ND CONTROL

4.1. Beneﬁts of ND Control at Farm Level

4.1.1. Deterministic Incremental Net Beneﬁts at Farm Level

The results of the deterministic analyses for the base scenario (without-control ND
mortality rate of 63 percent, prices per bird of 22 MTN for pullets and cockerels and 33
MTN for adult chickens and a price of 1.3 MTN per egg), suggest that vaccination of
chickens against ND using I-2 is ﬁnancially proﬁtable for the farmers. As seen in Table
5, regardless of the ﬂock size category, the overall incremental net beneﬁts due to
vaccination against ND at farm level are positive. This result is consistent with the
ﬁndings of Asgedom (2007), who used VIPOSIM to analyze the farm level impact of
different interventions in a Village poultry production system, and found that ND
vaccination resulted in higher net returns than housing intervention in the Ethiopian
context. These results are also consistent with the ﬁndings from Udo et al. (2006) and
Woolcock et al. (2004), who found that ND control has a positive effect on bird off-take,

egg production, egg off-take and ﬂock size.

Table 5: Annual Incremental Beneﬁts and Costs of ND Control at Farm Level

 

 

 

Annual Net Annual Costs of
Flock Size Category . Beneﬁts (MTN) Vaccination (MTN)
Category 1 (0-5 Chickens) 269 40
Category 2 (6-10 Chickens) 494 42
Category 3 (l l-15 Chickens) 607 50
Category 4 (16-20 Chickens) 756 ' 62
Category 5 (21—25 Chickens) 789 64
Category 6 (26 or more Chickens) l 179 93
Overall Beneﬁt at Farm Level 481 47

Data Source: Estimations ﬁ'om VIPOSIM simulations

6O

In general, vaccination of chickens using the I-2 vaccine results in a 481 MTN
(equivalent to 19 US dollars) increase in household annual poultry income. According to
Mather et a1. (2008), the median net total income per adult equivalent (AB) in rural
households was about 1,723 MTN in 2005 (corresponding to 6,892 MTN per household
on average). This suggests that vaccination results in an increase of about seven percent
in the total household income. This is a substantial improvement in a rural household’s
income, considering that it requires very low investment. Farmers need to invest about 47
MTN per year per ﬂock, which is less than one percent of the median total household
income. In general, vaccination of chickens results in about 10.3 MTN per one MTN
invested at the household level. This results in about 1030 percent annual rate of return,
which is very high. Additional analysis suggests that in general, farmers’ investments in
chicken vaccination will be proﬁtable as long as the cost of vaccination per bird is less
than or equal to 5.6 MTN which is more than 11 times the actual vaccination cost of 0.5

MTN per bird.

As the results suggest, farmers’ investments in the vaccination of their chickens is clearly
justiﬁed by the returns. However, cash investment in the vaccine may be a constraint for
vaccine adoption by the poorest group of farmers for whom this investment corresponds
to about 14 percent of their household income (Mather et al., 2008). Giving those farmers
opportunities to pay for vaccination in chickens instead of in cash, may be one way to
overcome the problems of lack of cash for vaccine payments, especially because the

vaccinators turn out to be one of the biggest poultry producers in the communities.

These ﬁndings suggest that there are ﬁnancial incentives for the farmers to invest in the

vaccination of their chickens, and that farmers just need opportunities to realize the

61

beneﬁts. The success in increasing the rates Of use of I-2 appears to depend on extension
strategies and distribution of the vaccine to the ﬁnal users, in which DNER and DINAP

play major roles.

Vaccination of chickens is also expected to have a positive effect on gender equality. It is
commonly argued that poultry production plays a very important role in promotion of
gender equality because many women participate in poultry production. According to
Mather et al. (2008) female- and widow-headed households in Mozambique tend to have
fewer economic opportunities and lower asset levels, and both female- and widow-headed
households are more likely to be found in the lower income quintiles. Women are usually
less educated and they usually have fewer skills and fewer assets than men, which limit
their participation in income-eaming activities and formal employment. Chickens provide
women an opportunity to generate income. However, data available for this study are not
sufﬁcient to give conclusive evidence about the effect of vaccination on improvements in
women’s income. The potential for women to earn income from chicken vaccination
programs, not only depends on their participation in the production process, but also on
intra-household arrangements. Women can only beneﬁt from the technology use if they
have the power to make decisions about production, consumption, sales and the use of
income generated. Further research may be performed to evaluate the impact of the

program on gender equality.

4.1.1.1. Sensitivity of Net Beneﬁts at Farm Level to ND Mortality Levels

Performing sensitivity analysis of the adapted VIPOSIM simulations for a range of
without-control mortality rates, as indicated earlier, results suggest that the annual
incremental beneﬁts at the farm level are sensitive to without-control mortality levels. As

62

the overall level of without-control mortality increases, the overall incremental beneﬁts
also tend to increase (see Figure 6). For lower without-control mortality levels, the
beneﬁts of vaccination tend to be lower. This result was expected, since in places where
the ND mortality rate is higher, the vaccine is expected to save more chickens than in
places with a lower ND mortality rate. This result suggests that in the process of
expanding the vaccination program, priority should be given to areas where ND mortality

rate is currently very high.

Annual Net Benefits of Vaccination at Farm Level

400 600
1 I

Annual Net Beneﬁts (MTN)
200
l

 

 

I l I l

T l
O 20 40 60 80 1 00
Without-Control ND Mortality Rates (%)

Figure 6: Sensitivity of the Annual Net Beneﬁts at Farm Level to Without-Control
ND Mortality Levels (Data Source: Author Estimations)

However, as shown in Figure 6, the net beneﬁts do not increase linearly or smoothly as
the ND without-control mortality levels increase. The explanation behind this

relationship between the ND mortality level and the level of annual net beneﬁts has to do

63

with the VIPOSIM model design, which implicitly imposes the level of availability of the
scavenging resource base, by setting the threshold ﬂock size above which all the chickens
are sold or consumed in the model. According to the law of decreasing marginal
productivity, given the level of available scavenging resources, the higher the number of
chickens, the higher is the competition among birds for scavenging resources and the
lower is the productivity. Thus, VIPOSIM forces higher levels of off-take when the ﬂock
reaches those threshold limits. Then, it is necessary to allow some time without off-takes
in order to build up the ﬂock. Controlling ND for higher levels of ND mortality implies
saying a relatively higher number of chickens, and consequently larger ﬂock size than in
situations of lower ND mortality. However, the amount of scavenging resources available

for feeding is the same (same threshold levels) in both situations.

Looking at the household level results in greater detail (Table A 7), regardless of the
overall without-control level of mortality, households with larger ﬂock sizes in general
tend to get relatively higher annual net beneﬁts, as was indicated in Table 5 for the base
scenario. This result was expected, since the vaccine saves more chickens in households
with large ﬂocks than in households with small ﬂocks. In addition, Table A 7 reveals that
the net beneﬁts can even turn out to be negative for households with smaller flock size (0
to 10 chickens) when the without-control mortality levels are very low; however,
regardless of the ﬂock size category, all farmers beneﬁt from vaccination when the
overall without-control ND mortality level in the region is greater than or equal to 40
percent. Since in most cases, chickens are the only asset owned by the poorest farmers,
vaccination of chickens will result in additional income that will help to lift some of those

families above the poverty line. This ﬁnding supports the theory that vaccination has the

64

potential to reduce absolute poverty, particularly for Mozambique where the ND is
endemic and causes severe losses annually, and a large percentage of the rural population
is involved in chicken production.

4.1.1.2. Sensitivity of Incremental Net Beneﬁts at Farm Level to Prices

As previously mentioned, the increase in the supply of chickens as a result of vaccination
may have effects on chicken prices. Results of the analysis of the net beneﬁts as chicken
prices are progressively reduced, ceteris paribus, Show that the net beneﬁts become zero
only if the price becomes nine percent of the original price used in the analysis; the price
can fall by as much as 90 percent without affecting the overall proﬁtability of the
vaccination. This result suggests that the overall proﬁtability of vaccination at the farm
level is not very sensitive to price changes. Only in the extreme case of almost perfectly
inelastic demand for village chickens would the increase on the supply as a result of

vaccination result in negative returns, but this is not likely.

A current and interesting debate in the poultry industry in Mozambique is related to the
arrival of very cheap Brazilian frozen chicken in Mozambique as a result of excess
production in Brazil and possible “dumping”. There is an increasing concern about how
the “dumping” of Brazilian poultry might affect the local poultry industry and

technological interventions in the sector.

Even though this aspect was not directly addressed in this study, it can be said that it may
constitute a big concern to caged poultry and commercially oriented system; however, it
does not constitute a big concern either for village poultry production system in general,
or for adoption of vaccination program, for several reasons. First, village poultry has a

completely different market than imported Brazilian chicken. Improved commercial

65

chicken is mainly available and consumed in urban areas while village chickens are
mostly consumed by the rural population, and usually the producer is also the consumer

of village chicken.

With an underdeveloped transportation infrastructure, Brazilian chickens are unlikely to
reach many rural areas. In the event of reaching rural areas, they will have a minimal
effect on the village poultry market because the sales prices will be higher than in urban
areas due to transport and marketing costs. Also, in rural areas the majority of the farmers
are not commercially oriented; they are more interested in ensuring their subsistence.
Usually they do not have cash available to purchase goods, and they mostly consume

what they produce.

On the other hand, village chickens that are supplied to urban areas are used mainly for
very speciﬁc activities (e.g., cultural ceremonies) for which Brazilian chicken cannot be
used as a substitute. In addition, village chicken is more ﬂavorful than commercial
chickens such that urban families are sometimes willing to pay a premium for village

chicken.

Given that the price of village chickens can fall by as much as 90 percent without
affecting the proﬁtability of vaccination at the farm level, speeding up the adoption of
chicken vaccination in order to increase the supply of village poultry in major
consumption centers could be a valuable strategy to address the problem of Brazilian
chicken “dumping” in the long term. The increase of chicken supply as a result of
vaccination would reduce village chicken price and make it competitive with Brazilian

chicken.

66

4.1.2. Stochastic Analysis of Incremental Net Beneﬁts at Farm Level

As previously mentioned, @Risk software was used to generate the distribution of the
results based on the best ﬁt probability distribution to ND mortality rates data, and then
the distribution of the net beneﬁts was generated. Analysis suggests that the triangular
distribution best ﬁts the mortality data. Results of stochastic analysis treating the ND
rates as a random variable show that there is a 90 percent probability that the values of
the annual net beneﬁts at farm level fall between 248 and 543 MTN (see Figure 7). This
conﬁdence interval covers the overall annual net beneﬁt estimated using deterministic
methodology. In addition, the results suggest that the most likely value of net beneﬁts at

farm level is 483 MTN per year, which is close to 481 MTN; the value estimated using

the deterministic approach.

Annual Net Beneﬁts at Farm Level in MTN

 

 

 

 

 

247.6 543.4
5.0% 'l i
0.12 . ......................................... ., ........... .. . ..... ....... . ........... 5 ....................... , .......... _ ............. i ................... . . .
0.10 ........................................................................................
>4
8
o 0.08:
z
8‘
a 0 061
o
.2
E 0.04 ................................................................................................................................................
D
at
0.02..§.... .. ..................................
0.00
o o o o o o o o o
In a in o n c in o in
'— N N m m q- q- .n m

Figure 7: Distribution of the Overall Annual Net Beneﬁts at Farm Level (Data
Source: Author Estimations)

67

4.2. Proﬁtability of Vaccination Program at District Level

Beneﬁt-cost analysis at the district level was performed. Results of proﬁtability measures
over 20 years, assuming the level of ND mortality base of Chibuto (about 63 percent) and
an adoption ceiling level of 50 percent, suggest that investment in extension and
distribution of I-2 vaccine in the district is highly proﬁtable (see Figure 8, and for more
details refer to Table A 8, in the Appendices). These investments result in a NPV of about

34,662,840 MTN (1,386,514 US dollars) and an IR of 37.39 percent.

Benefits of Extension of |-2 Vaccine in the District

10
I

 
   
  

NPV = 34,662,840 MTN
1,386,514 USD

8
I

IRR = 37.39%

(D-

4
1

Annual Net Beneﬁts in Millions MTN
2
I

4..

l I l I I
2000 2005 2010 2015 2020
Year

0
l

 

 

Figure 8: Beneﬁts of Investments in the Extension of I-2 Vaccine in Chibuto District
(Data Source: Author Estimations)

4.2.1. Sensitivity Analysis
As previously discussed, three types of sensitivity analysis were performed at the district
level of analysis: sensitivity to ND mortality rates, ceiling level of adoption and chicken

price. Results of the analysis suggest that the overall proﬁtability of the vaccination

68

program in Chibuto district is sensitive to the level of without-control mortality due to
ND. The program is proﬁtable when the overall level of mortality due to ND in the
district is at least 10 percent (see Figure 9, and for more details refer to Table A 9 in the
appendices). As the level of ND rates of mortality increases, the value of NPV also
increases. This ﬁnding supports the idea that during the planning process of expansion of
ND program to other sites, the levels of ND occurrence must be taken into consideration,

and places with high levels of without-control mortality should be given priority.

NPV of Extension of l-2 Vaccine in the District

20 30 40
I l l

NPV in Millions MTN

10
I

 

 

I 1 l I I
O 40 60 80 1 00
Without-Control ND Mortality Rates (%)

Figure 9: Sensitivity of NPV to Without-Control ND Mortality Rates in Chibuto
(Data source: author estimations)

However, similar to the analysis at the farm level, the NPV does not increase linearly or

smoothly as ND mortality rates increases. The justiﬁcation of this problem is the same as

69

the one presented at farm level analysis. Since the district level beneﬁts are the

aggregated beneﬁts at the farm level, the results show the. same trend.

Analysis of the effects of changing assumptions about the ceiling level of adoption of 1-2
vaccine suggests that if at least 15% of farmers raising chickens in the district adopt the
technology over 20 years, the vaccination program is proﬁtable. The beneﬁts tend to be
higher as the adoption ceiling levels increase (see ﬁgure 10; for more details see Table A
10, in the Appendices). This suggests that efforts to reach the highest adoption level

possible will maximize the proﬁtability of the program.

NPV of Extension of l-2 Vaccine in the District

NPV in Millions MTN
30
l

20
l

 

 

 

 

 

 

 

 

I I I I

I
20 40 60 80 1 00
Adoption Ceiling Level (%)

Figure 10: Sensitivity of NPV to Adoption Ceiling Levels in Chibuto (Data Source:
Author Estimations)

70

Furthermore, sensitivity analysis of the proﬁtability of the program to reduction in
chicken prices at the district level reveals that the price can fall by as much as 79 percent,

without affecting overall proﬁtability of the program in the district.

4.2.2. Stochastic Analysis

Results of the stochastic analysis treating ND mortality rate as a random variable, and
using triangular distribution (best ﬁt distribution to ND mortality), indicate a 90 percent
probability that the NPV at district level will lie between 15.36 and 39.80 million MTN,
over the 20 years of analysis (See Figure 11). The deterministic value of NPV at district
level also falls within the 90 percent conﬁdence interval. In addition, the results indicate
that the most likely NPV value over the 20 years is 34,823,553 MTN, which is also very

close to 34,662,840 MTN; the value estimated using the deterministic approach.

NPV in Chibuto District in MTN
15.36 39.80

5.0% 90.0%

 

0.12....

P
.—n
O

0.081 ................................................ .1 .............. ......... ..

.o
o
4:.

Relative Frequency
0
O
O\

 

O
O
N

 

 

0.00 ‘
W O W O W O
— F" N N ("'1

Values in Millions

Figure 11: Distribution of NPV of Extension of [-2 Vaccine in Chibuto District (Data
source: Author Estimations)

C
v

W V)
("I V

71

4.3. Proﬁtability of Production and Extension of [-2 in Mozambique

Benefits of Production and Extension of l-2 Vaccine in the 44 Districts

—1

‘—

8
l

   
 

NPV = 410,616,275 MTN
16,424,651 USD

6
i_

IRR = 21.49%

  

4
i

Annual Net Beneﬁts in Millions MTN
|

0
l

 

 

l I I I I
2000 2005 201 O 2015 2020
ear

Figure 12: Beneﬁts of Investments in the Production and Extension of I-2 Vaccine in the
44 Districts (Data Source: Author Estimations)

Assuming a ceiling level of adoption of 50 percent and the base scenario at farm level,
production and extension of 1-2 vaccine across the 44 districts of program
implementation in Mozambique generate positive returns (see Figure 12, and for details
see Table A 11). The NPV is about 410,616,275 MTN (about 16,424,651 US dollars) and

the IR is about 21.49 percent.

4.3.1. Sensitivity Analysis
The results of evaluation of sensitivity of proﬁtability of production and extension of the
I-2 vaccine across the 44 districts of implementation, suggest that it is sensitive to the

level of overall without-control ND mortality in the communities. The program is

72

proﬁtable if the overall level of ND mortality is at least 25 percent in all the districts
where the vaccine is already used (see Figure 13, and for more details, see Table A 12 in
the Appendices). Thus, during the planning process of expansion of the ND program to

other sites, the levels of without-control ND occurrence must be taken into consideration,

and places with high levels of mortality should be given priority.

NPV of Production and Extension of l-2 Vaccine in the 44 Districts

20 30 40
l L L

W in Mllions MIN

10
4

 

 

l l j l 1 T
40 60 80 100
Without-Control ND Mortality Rates (%)

Figure 13: Sensitivity of NPV to Without-control ND Mortality Rates in the 44 the
Districts (Data Source: Author Estimations)

In addition, assuming adoption ceiling levels of at least 10 percent over the 20 years, the
program is proﬁtable (see Figure 14 and other details in Table A 13). As the ceiling level
increases, the beneﬁts of the program also increase. This reinforces the idea that efforts

towards reaching the highest adoption ceiling level possible will maximize proﬁtability

of the program in the long term.

73

NPV of Production and Extension of l-2 Vaccine in the 44 Districts

350 400
l l

NPV in Millions MTN
300 ‘
l

250
l

   

 

200
l
I

 

I l I I l

l
O 20 4O 60 80 100
Adoption Ceiling Level (%)

Figure 14: Sensitivity of NPV to Adoption Ceiling Levels in the 44 Districts (Data
source: Author Estimations)

Additional analysis suggests that the price of chickens can decline by as much as 44
percent without affecting the overall proﬁtability of the program in the areas of

implementation.

The analysis of the vaccination program’s impact in 44 districts shows that the program is
very proﬁtable. Given that the vaccine production facilities are currently operating below
their capacity, expansion of the program to other districts is feasible, and it may result in

higher proﬁtability without major new investments in vaccine production infrastructure.

4.3.2. Stochastic Analysis
Results of analysis of distribution of NPV in the 44 districts of implementation treating

ND mortality rate as random variable and using triangular distribution for the random

74

variable, suggest that there is a 90 percent probability that the NPV of the vaccination
program lies between 97 and 494 million MTN over 20 years (See Figure 15). Similar to
the other two levels of analysis, the 90 percent conﬁdence interval covers the value
estimated using the deterministic approach. In addition, the results suggest that the most
likely NPV value over the 20 years is 413 million MTN. The most likely NPV of
production and extension of I-2 vaccine in the 44 districts is also close to the value

estimated using the deterministic method which is 410,616,275 MTN.

NPV in the 44 Districts in MTN

 

 

 

 

 

O O O O
O O O O O 8
'7 N M v tn

Values in Millions

Figure 15: Distribution of NPV of Production and Extension of [-2 Vaccine in the 44
Districts (Data source: Author Estimations)

75

CHAPTER V: SUMMARY OF FINDINGS AND POLICY IMPLICATIONS

5.1. Findings and Policy Implications

Investments in ND control are ﬁnancially proﬁtable for farmers. Vaccination of chickens
using the I-2 vaccine results in about 481 MTN (equivalent to 19 US dollars) increase in
annual household poultry income. This income increase corresponds to about seven
percent of average annual rural household income, for an investment of less than one
percent of the income. Even if the vaccination costs increase, the program will still be
proﬁtable for the farmers, as long as the costs of vaccination per bird do not exceed 5.6

MTN.

There are ﬁnancial incentives for the farmers to invest in the vaccination of their
chickens, and farmers need to be informed about the beneﬁts. Increasing the rates of use
of I-2 vaccine mostly depends on the strategies of diffusion of information about the
vaccine and distribution of the vaccine to the ﬁnal users, in which DNER and DINAP

play major roles.

The proﬁtability of the program is sensitive to the incidence of without-control ND: the
higher the mortality rate, the higher the beneﬁts of vaccination. When planning expansion
of the program to the other sites, prioritizing areas with high incidence of ND may

constitute a good strategy for maximizing the impact of investments.

Regardless of the size of initial ﬂock, the farmers can beneﬁt from vaccination as long as
the rates of ND mortality in the region reaches at least 40 percent. Chicken vaccination
has a potential to reduce absolute poverty, particularly for Mozambique where ND is
endemic and causes severe losses annually.

76

Looking at the aggregate levels, production and extension of [-2 vaccine has positive
returns. Extension of [-2 vaccine in Chibuto over 20 years results in NPV of about 34.6
millions MTN. On the other hand investments on production and extension of I-2 vaccine
in the 44 districts over 20 years results in NPV of about 410 million MTN. An increase in
the adoption ceiling level results in relatively higher returns. The stochastic analyses
result in estimates close to deterministic ones, and suggest that the results are not highly
sensitive to the uncertainties of the various aspects. Therefore, expansion of the
vaccination program to new areas should be explored and strategies put in place to speed

up the adoption in areas already covered.

Partnerships between the Government of Mozambique, NGOS, and other private entities
in delivering the vaccine to the ﬁnal users are a good strategy to overcome public funding
limitations for coverage of the vaccination program. However, there is a threat to the
sustainability of the program in the long term related to this strategy. If for some reason
the interest of the NGO or private company in participating in the program changes, the
area where they have been operating may no longer have access to the vaccine, unless
other options are developed. Thus there is a need to look for additional strategies as a

complement in order to ensure long-term sustainability of the program.

Purchase of a new freeze dryer or repairing of broken one could increase the efﬁciency of
production, and possibly, result in transportation efﬁciency in getting the vaccine to the
provinces. Higher levels of vaccine production and transportation of larger quantities to
the provinces may enable economies of scale to be realized, reducing the cost incurred

per dose of vaccine. Also, with higher levels of production it may be possible to supply

77

large-scale or regional poultry input markets, provided that the vaccine is commercially

registered and quality guarantees are in place.

Making the vaccine available does not seem to be a sufﬁcient condition for guaranteeing
positive and sustainable results at the village level. Community leaders’ involvement in
the vaccination program is crucial for the success and sustainability of the ND control
program. Getting community leaders and district administrators on board will speed up
the adoption of the vaccine in the district if they raise awareness of the problem of ND

and the importance of vaccination in public community meetings.

The seasonal nature of the vaccinator’s job is a problem for the sustainability of the
vaccination program. Vaccinators can easily quit the vaccination program in favor of
better prospects. To enhance the performance and sustainability of the program,
opportunities for other related income generation activities should be created for the
vaccinators to occupy them during the off-season period, and to lessen the chance that
they will commit themselves to other engagements when needed for the vaccination

campaign.

We cannot conclude from this study about the impact of the vaccination program on
promotion of gender equity. The potential for women to earn income from chicken
vaccination depends not only on their participation in the production process, but also on
intra-household arrangements. Women can only beneﬁt from the technology use if they
have the power to decide about production, consumption, sales and the use of income
generated. While we know that women are active with poultry production, we do not

have sufﬁcient information on control of poultry marketing and ﬁnancial resources.

78

5.2. Further Research

In this research it was found that whenever the levels of ND mortality rates are
appreciable, ND control has positive returns; and it is argued that because village
chickens are one of the mains assets owned by the poorest households, and ND is
endemic in Mozambique, ND control will contribute to lift some households above the
poverty line. However, there is no quantiﬁcation of how much the poverty indicators will
fall as a result of ND control. Further attention should be given to the quantiﬁcation of

impact of ND control on poverty indicators.

While performing the aggregate analysis of impact of I-2 vaccine use in the 44 districts of
implementation, information believed to be related to agro ecology and husbandry
conditions such as ND occurrence and chicken production and utilization parameters
used was heavily based in data from one district because there was no information
available about spatial distribution of the parameters across different agro ecologies
and/or districts. Thus, further research is recommended on geographical distribution and

intensity of ND in Mozambique, as well as spatial distribution of impact of ND control.

This research was not conclusive about the impact of the vaccination program on
promotion of gender equality. There is insufﬁcient information on women’s control of
poultry marketing and ﬁnancial resources. Further research is recommended on impact of

vaccination on gender equality.

Also, no information about elasticity of demand for village chickens is available in

Mozambique. This is needed for greater understanding of the possible price impacts of

79

increasing village chicken production. Thus further research on marketing of village

chickens is recommended.

80

APPENDICES

81

Table A 1: Average Chickens per Household and Total Chickens Vaccinated per
Vaccinator in Each Campaign

 

, Mar-05 Jul-05 Nov-05 Mar-06 I Jul-06 Nov-06 Mar-07 '
Vaccinator l, 2 l 2 l 2 l. 2‘ l 2 I; 2. l 2:

 

AMacie . . . . . . . " . . . 16‘ 415' 141,045
A. Novela 7 33 10 70 17 50 8 421 11 545 14 723 16 679
AMassingue . . . . . . . . . . . ‘ . 14 27
A. Ndove 9 175 12 169 10 105 7 257 9 242 11 159 ll 56
>A.Cuna 7 .. 8 76 13' 80 8 65 . . . . ﬂ _. 10 720 .

‘AChilaule . 8 8. 2 2, 8 .135. 9 305 714- 486 21., 374,

C.Mantchane 5 l6 5 28 8 58 14' 352 141 432 19' 842 13‘ 766‘;
‘E.Mond1ane 11 654 14 992 191,327 141,166' 181,468 21 2,069, 252,370

_F.Chiconela . . . . . . 12 630 16 596 15 581 9 591
Filomena 9 423 11 515 10 434 11 232 12 950, 12 803 10 407‘
F.Mucave1 . . . , . . . _ 8 407 9 296 10, 41117, 422.
‘G.Phacule 11 373 12 362 9 197 8, 450 10, 995 141,057 147 830‘
L.Mazivila 15 375 14 367 15 293 13 375 17 419 21, 487 17 309
L.Saveca 16 49 13 39 . . 14 162' 14 256 19 439, 28 689
M.Amelia 5 9 6 19 19 19 11 135 11 274 19 265 23 278
Miseria ' . . . . . . . . . . . ‘ . . I 14' 920'
1N.Mbiza 11 226 13 241 12 254 11 328 10, 302 12' 524, 15 380
{Paulo . 8 - 80 7 64, 12 112 7 7_ 14 . , g. . 8 42: 9 , 222:
:P.Covela . w . . , . . . ,- 13 140, 17 626 . ; . .19 871,
RRichete ; 9 74 7 41 9 124 ,4“ 40. _10._ 233, 14’ 2391 11; 230.
S.Sutho 19 354 20 383 14 170 15. 642 20 511, 18 703 1,6, 581‘
Saquina . ' . . . . . . . . . . . 15 622‘
Xadreque M. . _ . . . . . . . . _. . i . 1 7 . . 24‘ 1,338;
Zulmira ' ~ . 9; 91, 13 170, 12 136 .

 

‘ 1 - Average number of chickens vaccinated per household per campaign
2 - Total number of chickens vaccinated per campaign
Source: Author Computations

82

Appendix Al: Mathematical Conceptualization of VIPOSIM (extracted from Asgedom
2007, 96-105)

1. Randomization
Random numbers were used as coefﬁcients of standard deviations of average input values
of explanatory variables to determine the different results of a given scenario. The

random coefﬁcients from a normal distribution were generated as:

R=[iRand ( )i]—6 (A1.0l)
12

where R is random coefﬁcient and Rand( )i is the ith random number in a normal
distribution. The value of the explanatory variable is calculated as X = 36' + (R * SD). If

X < 0 , it is set to zero. In addition, for values presented in percentages, if X > 100% , it is
set to 100%. If all SD values are set to zero, the model becomes deterministic instead of
stochastic. It is also possible to enter zero for a single parameter, to exclude the variation

in that parameter.

The standard procedure to deal with random variation of simulation, given the same input
data, is to replicate the simulation a number of times and take the averages of a

parameter. The number of replications required, N(m), is determined using initial

replications:

2
5(m)’m-1,1-a/2

X(m)£ (Al.O2)

 

N(m)=

where N(m) is the number of replications required, given m replications; X(m) is an

estimate of the real mean 11 from m simulations runs (samples); S (m) is an estimate of

83

real standard deviation 0' from m simulations runs; a is the level of signiﬁcance; 8 is

the allowable percentage error of the estimate X(m), a = |X(m)— ,uI/lpltm_1 1—a/2 is

the critical value of the two-tailed t-distribution at a level a of signiﬁcance, given m —1

degrees of freedom.

Using 10 initial sample runs, at 95% level of signiﬁcance and 5% allowable error, the
number of replication runs was calculated to be 50. The initial sample runs were

performed using ﬁeld data from Tigray, Ethiopia.

2. Initial ﬂock
In the model, the ﬂock categories are denoted as i, where i runs from 1 to 5 representing
chicks, pullets, cockerels, cocks and hens, respectively. After entry of initial numbers, the

random number of birds in category i in the initial ﬂock (1F 1‘) is calculated as:

[Pi = CMi + (SDi * R) (A103)
where CMi is the average number of birds in category i entered by the user; SDi is the
standard deviation of the number of birds in category 1 entered by the user; and R the

random coefﬁcient of ﬂock size.

3. Mortality

Mortality is considered to differ among categories of chickens and seasons. The variation
of mortality can be entered by the user for j, representing four'seasons. Mortality is
distinguished between mortality due to diseases, predation or other reasons. The
modeling procedure of mortality considers the different ﬂock categories and year seasons

categories.

84

The predation loss of a ﬂock category is determined from initial number of birds,

predation mortality rate (varied by SD and random number) and season:
P.. = C.. *{PRU +(R. *SD.. D ' (A104)
11 'J 'J J 1.1

where Pij is the number of birds killed by predators in category i in season j, Cl] is the
number of birds present in category i in the beginning of season j; PR1]. is predation rate
(%) of category i in season j; SDij is the standard deviation of PR1]. ; R]. is a random

coefﬁcient in season j. The same random coefﬁcient is used across mortality rates per
season for each category because positive correlation was observed between parameters
of different ﬂock categories: as mortality in one ﬂock category increases, the same
pattern is observed with that of other categories. Mortality from disease and other

unknown reasons is computed using the same procedure as predation mortality.

4. Bird off-take
Farmers try to maintain a bird ﬂock at a certain target number, adjusted to household
resources. Above certain a ﬂock size, birds are consumed or sold. Consumption and sales

vary with seasons and bird categories.

Bird sales: the number of birds sold depends on the number of birds left after mortality.
No sales are allowed below minimum limits which depend on ﬂock categories.

Maximum limits are also set, beyond which all birds are sold. In addition, for ﬂock sizes

85

between the minimum and maximum limits the number of birds sold of category 1 in

season j (Sij ) is calculated using:

S.. =AF.. * SR.. + R. *SD.. (A105)
11 U ’1 J U

where AFij is the number of birds present after mortality in category i in season j; SR1].

is the rate of sale (%) of category 1 in season j; SDij is the standard deviation of sale rate

of category i in season j; and R]. is random coefﬁcient of sale rate in season j.

Bird consumption: The number of birds consumed depends on the number of birds after

mortality, and sale and the same minimum threshold limits as mentioned for bird sales.

Given threshold limits, the number of consumed birds of category 1 in season j [Cij ] is

modeled as:

C.. =SAF.. * CR.. + R. *SD.. (A106)
1] I] 11 J y

where SAFij is the number of birds present after sale of category 1 in season j, CR1]. is
the consumption rate (%) of category 1 in season j; SDij is the standard deviation of
consumption rate of category i in season j; and R]. is a random coefﬁcient for

consumption rate in season j.

86

The total weight of poultry meat consumed in season j (T WBCj) was determined as a

function of number, live weight and carcass percent of each ﬂock category summed over

i as follows:

-= :1: a1:
TWBC] IZCCij LWij CPI]. (A107)
where CCij is the number of birds consumed in category 1 in season j; LWij is the
average live weight of category i in season j; and CPij is carcass percent of category i in

season j.

The difference of the initial number of birds and the birds removed by mortality, sale and

consumption was the net ﬂock size and structure.

5. Egg production
Egg production depends on the average number of hens in the net ﬂock in a speciﬁc
season and the number of eggs laid in a clutch. The number of eggs per clutch is

inﬂuenced by season because of differences in feed availability. The total number of eggs

produced in season j {EPj J is computed as:

EP. =O.5* H0.+Hl. * P.+ SD.*R. (A1.08)
J J J J J J

where H 0]. is the number of hens in the beginning of season j; H1]. is the number of

hens at the end of season j; P]. is the number of eggs per hen in season j; SD]. is the

standard deviation of egg number per hen in season j; R]. is a random coefﬁcient of the

87

number of eggs per hen in season j. The model assumes that all hens become broody and
have to spend some time incubating and rearing their chicks over a period of three

months.

6. Reproduction

Setting eggs, incubation capacity of broody hens and hatchability determine the
reproduction process. The model assumes that for the eggs produced, farmers’ ﬁrst
priority is hatching to maintain the ﬂock. The total number of eggs set for hatching and

the hatchability rate determine the number of chicks born in one season. Number of

incubated eggs in season j {ES}. ) is determined as:

ES.=EP.* SR.+ SD.*R. (A109)
J J J J J

where EPj is the number of eggs produced in season j; SR]. is the rate of incubation in

season j; SD]. is the standard deviation of SR]. ; and R]. is a random coefﬁcient of

SD]. . The number of hatched eggs in season j [EH]. J was calculated as:

EH. =ES.* HR.+(SD.*R.) (A1.10)
J J J J J
where E5}. is the number of eggs incubated in season j; HR]. is the hatchability rate in

season j; SD]. is the standard deviation of HR]. ; and R]. a is random coefﬁcient of

SR..
J

88

7. Egg off-take
The egg off-take rate is deﬁned as the number of eggs consumed or sold as a percentage

of the total number of eggs produced. The egg off-take rate can ﬂuctuate between

seasons. The total number of eggs sold in season j (SE1. 1 was calculated as:

SE. =EP. * SER. + SD. *R.
J J l J i J JD (Am)
where EPj is the number of eggs available (produced) in season j; SERJ. is the egg sale
rate in season j; SD]. is the standard deviation of EPj ; and R]. a is random coefﬁcient
of SD . .
J

The number of eggs consumed in season j [EC]. ] is inﬂuenced by the number of eggs
produced, incubated, lost or broken, and sold:
EC. =EP.—EB.—ES.—SE. (Al.12)

J J J J J
where EPj is the number of eggs produced in season j; EB]. is the number of eggs

broken or lost in season j; ES]. is the number of eggs incubated in season j; and SE]. is

the number of eggs sold in season j.

89

Consumption can also be expressed in terms of egg mass in kg. Egg mass in season j

(EM j ) was calculated from the number of eggs consumed, and egg weight as indicated

below:

EM. = EC. *(EW. ”000] (Al.13)
J J J

where EC]. is the number of eggs consumed in season j; and E W]. is weight of eggs in

grams in season j.

8. Average flock present
Flock size varies within a season due to mortality and off-take. Flock present refers to the

average number of birds available in each season. The average number of birds is the

average number of birds of category i in season j (AvCIj ). This was computed as:

AvC.. =[CO.. +Cl.. J/Z (Al.14)
1J U 1]

where COij is the number of initial birds of category i in season j; and C lij is the

number of birds of category i left at the end of season j.

9. New ﬂock
The ﬂock size and structure change after each time step (season). The number of hens in
a new season depends on the number of hens at the end of the previous season and the

number of pullets becoming hens. The number of pullets joining hens depends on the age

90

 

at ﬁrst egg. The number of hens in the new ﬂock in season j +1 , NH j+l , is calculated

as:
NH. =AFH]. +£21ij *(3/(age—3))j (Al.15)

1+1

where AFHJ. is the number of hens at the end of season j; AFPJ. is the number of

pullets at the end of season j; and age is maturity age.

The number of cocks in the new ﬂock is computed the same way as for hens. The number
of cockerels joining the cocks depends on age of maturity. This age of maturity is
assumed to be equal to the age at ﬁrst egg for the pullets. A11 chicks in a new season
come from chicks newly hatched in the previous season. After one season chicks are

assumed to become growers and will be equally distributed as cockerels and pullets.

10. Manure Production

The model calculates the amount of dry matter in kg of dry matter (DM) manure

produced in season j (M j J as a function of average birds present of different categories

and their respective manure yield and dry matter:

M. = CM.. *A C.. Al.16
J 2( 1J v U j ( )

where CMij is the manure yield (kg of DM) of category i in season j; and AvCij is the

average number of birds of category i in season j.

91

11. Workload

The workload expressed in total labor hours spent on poultry in season j (LHj ] was

determined considering time spent per day per bird and average ﬂock size in a given

season:
LHJ, = ZUMCDJ. * AvCij )*90] (A117)

where MCDJ. is the average number of hours spent per day per bird in season j; and

Ava is the average present number of birds of category in season j.

12. Costs
The model separates costs of labor and other costs. The model has the ability to deﬁne

input values of extra costs per day per bird of various inputs. It then calculates total cost

of production in season j [TC]. J as a result of input cost/day/bird, and average ﬂock size

in this season:

TC.= BC.*90*AvC.. A118
J :l J a] ( )

where BC]. is the cost per bird per day in season j; and. AVCij is the number of birds in

category 1 in season j. Labor cost in season j [LC]. ) was calculated based on labor hours

and labor costs per hour:

92

LC. =Z(LH. *CLHR. *AvC.. *90 (A119)
J I. J J IJ

where LHj is the number of hours spent per day per bird in season j; CLHRJ. is the
cost of labor per hour in season j; and AvCiJ. is the number of birds of category i in
season j.

13. Benefits
The model considers beneﬁts of cash income and opportunity values. The computations

consider various bird categories and seasonal variations in ﬂock size and prices.

Direct beneﬁts include cash values of bird sales and consumptions, and egg sales and

consumptions. The direct beneﬁts in season j [DB]. ] are calculated as:

DB. =2 [S.. +C..]*BP.. *R. + (SE. +EC.]*EP. *R.l (A120)
J 1 1} IJ J J J J J J)

where Sij is the number of birds sold in category i in season j; Cij is the number of
birds consumed in category i in season j; BPij is the price of birds in category i in season
j; R]. is a random coefﬁcient for prices in season j; SE]. is the number of eggs sold in
season j; EC]. is number of eggs consumed in season j; and EPj is the price of eggs in

season j.

93

The indirect beneﬁts in season j [13].) are derived from the cash values of average

present ﬂock and manure production:

IB. :2 AvC.. *(AV/4)*BP.. *R. + M. *MP. (A121)
J 1J H J J J
where AvCij is the average number of present birds in category i in season j; AV is the

animal presence value, the value of having birds in case of urgent cash or social needs;

BPi' is the price of bird category i in season j; R]. is a random coefﬁcient for prices in
season j; M j is the amount (kg DM) of poultry manure produced in season j; and MP].
is the price per kg of DM of manure in season j.

14. Net Return

Net return in season j [NR]. ) is calculated as the difference of the total beneﬁts and total

costs. This is done for every season according to the equation:
NR. = DB. +IB. —TC. (A122)
J J J J

where D3}. is direct beneﬁts in season j; 18}. is indirect beneﬁts in season j; and TC].

is the total cost in season j.

94

The model also calculates effectiveness of labor in terms of net return per labor hour:

LNRJ. = [TB]. — TC]. )/(MCD*90) (A123)
where LNRJ. is the net return per labor hour in season j; TB}. is the total beneﬁts in

season j; T C]. is the total cost in season j; and MCD is the number of labor hours per

day.

95

Table A 2: Estimated Adoption Rates in Chibuto District
Ceiling Level (%)

 

Year 15 20 30 40 50 60 70 80 90
1 1999 0.001 0.008 0.012 0.014 0.015 0.015 0.016 0.016 0.016
2 2000 0.004 0.013 0.018 0.020 0.021 0.022 0.022 0.022 0.022
3 2001 0.009 0.022 0.027 0.029 0.029 0.030 0.030 0.030 0.031
4 2002 0.023 0.035 0.039 0.040 0.041 0.041 0.041 0.041 0.042
5 2003 0.049 0.054 0.056 0.056 0.056 0.056 0.056 0.056 0.056
6 2004 0.085 0.079 0.077 0.077 0.076 0.076 0.076 0.076 0.076
7 2005 0.117 0.107 0.103 0.102 0.102 0.102 0.101 0.101 0.101
8 2006 0.136 0.133 0.133 0.134 0.134 0.134 0.134 0.134 0.134
9 2007 0.144 0.156 0.165 0.169 0.171 0.172 0.173 0.174 0.175
10 2008 0.148 0.172 0.195 0.206 0.213 0.217 0.220 0.222 0.224
11 2009 0.149 0.183 0.221 0.243 0.256 0.266 0.273 0.278 0.282
12 2010 0.150 0.190 0.243 0.277 0.300 0.317 0.330 0.340 0.348
13 2011 0.150 0.194 0.260 0.306 0.341 0.367 0.387 0.404 0.418
14 2012 0.150 0.197 0.273 0.331 0.376 0.413 0.443 0.469 0.490
15 2013 0.150 0.198 0.281 0.350 0.406 0.454 0.494 0.530 0.560
16 2014 0.150 0.199 0.287 0.364 0.430 0.488 0.539 0.584 0.625
17 2015 0.150 0.199 0.292 0.374 0.449 0.516 0.577 0.632 0.682
18 2016 0.150 0.200 0.294 0.382 0.463 0.538 0.607 0.671 0.731
19 2017 0.150 0.200 0.296 0.388 0.473 0.554 0.630 0.702 0.770
20 2018 0.150 0.200 0.298 0.391 0.481 0.567 0.649 0.727 0.802

 

Source: Author Computations

96

 

 

 

 

 

 

99 — .
(Q _
V: .4
N. . ’5
O 4 . __~ -13/
T l l l l
2000 2005 2010 2015 2020
Year
—-1— L15 —*— L20 + L30 -—-—-II-- L40 —-+—- L50
—0— L60 ——-é--— L70 - -- L80 -—9-— L90

 

 

 

Figure A 1: [-2 Adoption Profiles at District Level Given the Adoption Ceiling
Levels (Source: Author Computations)

97

Table A 3: Estimated Adoption Rates in the 44 Districts where [-2 Vaccine is

 

 

Currently Used
Ceiling Level (%)
Year 10 20 30 40 50 60 70 80 90
1 1999 0.013 0.015 0.016 0.016 0.016 0.017 0.017 0.017 0.017
2 2000 0.016 0.018 0.019 0.019 0.019 0.019 0.019 0.019 0.020
3 2001 0.020 0.022 0.022 0.023 0.023 0.023 0.023 0.023 0.023
4 2002 0.025 0.026 0.026 0.027 0.027 0.027 0.027 0.027 0.027
5 2003 0.031 0.031 0.031 0.031 0.031 0.031 0.031 0.031 0.031
6 2004 0.037 0.036 0.036 0.036 0.036 0.036 0.036 0.036 0.036
7 2005 0.043 0.043 0.043 0.042 0.042 0.042 0.042 0.042 0.042
8 2006 0.050 0.050 0.050 0.049 0.049 0.049 0.049 0.049 0.049
9 2007 0.057 0.057 0.057 0.057 0.057 0.057 0.057 0.057 0.057
10 2008 0.064 0.066 0.066 0.066 0.066 0.067 0.067 0.067 0.067
11 2009 0.070 0.075 0.076 0.076 0.077 0.077 0.077 0.077 0.077
12 2010 0.075 0.084 0.086 0.088 0.088 0.089 0.089 0.089 0.089
13 2011 0.080 0.094 0.098 0.100 0.101 0.102 0.102 0.103 0.103
14 2012 0.084 0.103 0.110 0.113 0.115 0.117 0.117 0.118 0.119
15 2013 0.088 0.113 0.123 0.128 0.131 0.133 0.134 0.135 0.136
16 2014 0.090 0.123 0.136 0.143 0.148 0.151 0.153 0.155 0.156
17 2015 0.092 0.132 0.149 0.159 0.166 0.170 0.173 0.176 0.178
18 2016 0.094 0.140 0.163 0.176 0.185 0.191 0.195 0.199 0.202
19 2017 0.096 0.148 0.176 0.193 0.204 0.213 0.219 0.224 0.228
20 2018 0.097 0.155 0.189 0.210 0.225 0.236 0.244 0.251 0.256

 

Source: Author Computations

98

 

 

 

2000 2005 2010 2015 2020
year

 

--+—- L10 -—“-— L20 —-‘-— L30 -—l-—— L40 -—’-— L50
—0— L60 ”—ér-— L70 “- L80 —9— L90

 

 

 

Figure A 2: [-2 Adoption Profiles in the 44 Districts Given the Adoption Ceiling
Levels (Source: Author Computations)

99

Table A 4: Projected of Number of Total Households and Households Raising
Chickens

‘ Number of Households House holds Raising Chickens'

 

Year District All Districts District All Districts
1997_ 28,130 1,367,404 . 19,114 954,756
1998 29,230 1,407,839 19,862 982,989
1999 30,374 1,449,470 20,639 , 1,012,056 .
2000 31,562 1,492,332 21,446 1,041,983
2001 32,796 1,536,461 22,285 1,072,796
2002 34,079 1,581,895 23,157 1,104,519
2003 35,412 1,628,673 24,063 1,137,180
2004 36,798 1,676,834 25,004 ‘ 1,170,808
2005 38,237 1,726,419 ' 25,982 1,205,429 ,
.2006 39,733 1 1,777,471 26,998 1,241,075 Y
‘ 2007 41,287 . 1,830,032 28,055 1,277,774
2008, 42,902 1,884,147 29,152 , 1,315,559 I
2009 44,580 1,939,863 30,292 1,354,461
2010 46,324 1,997,226 31,477 ‘ 1,394,513 ‘
2011 48,136 2,056,285 32,708 1,435,750 .
2012 50,019 2,117,091 33,988 1,478,206
2013 51,976 . 2,179,695 35,317 1,521,918
2014 54,009 2,244,150 36,699 1,566,922
2015 56,121 2,310,511 38,134 1,613,257
2016 58,317 2,378,835 39,626 1,660,962
‘ 2017 60,598 2,449,179, 41,176 1,710,078
,2018 62,968 i 2,521,603 42387 1,760,646

 

Source: Author Computations

100

 

20000

30000 40000 50000 60000
1 1 l J

 

1 1 f l l 1
1995 2000 2005 2010 2015 2020
Year

 

l—O—— Total Households + Households Raising Chickens

 

Figure A 3: Total Number of Households and Number of Households Raising
Chickens at District Level (Source: Author Computations)

81
g-
8

 

1000000

 

1 l l l l l
1995 2000 2005 2010 2015 2020
Year

 

[—0— Total Households + Households Raising Chickens

 

Figure A 4: Total Number of Households and Number of Households Raising
Chickens in the 44 Districts (source: author computations)

101

Table A 5: Estimated Costs of Vaccination Program

 

 

 

 

 

 

 

Cost Description Value (MTN)
1. Costs of Trainingz Materials & Other Investments at District Level
Year 0 $64,139
Year 1 $93,234
Year 2* 441,434
*It is assumed that the value in year 2 is spent every 3 years
for regular training
2. Estimation of Annual Costs of Extension at District Level 180,000
2. 1. Salary (Only a quarter is completely dedicated to the program in a year) I 26,000
Non-Govemamental Institution 384,000
Government Workers (two are need for full time working during a campaign) 120,000
2. 2. Transport of vaccine from SSP (3 times a year) 1,800
Transport of Vaccine from SSP per Campaingn (about 20 liters of fuel) 600
2.3. Vaccine distribution during a earmaign (3 times a year) 7,200
Vaccine Distribution & Monitoring during a campaign (about 80 liter of fuel) 2,400
2.4. Regular Assistance & Monitoring 45,000
3. Vaccine Production, Quality Control & Transporte to Province per Dose 0.197
3.1 . Production & Quality Control 0.096
3.2. Distribution 0.10]

To estimate the total cost in this category, cost per dose is multiplied by the
number of doses produced each year

Source: Author computations

Note: To get the costs in 1 and 2 at national level, the district costs are multiplied by 44.

Table A 6: Number of [-2 Doses Produced

 

Number Number

Year of Doses Year of Doses
1 1,460,500 1 1 7,500,000
2 916,250 12 8,049,045
3 1,124,000 13 8,598,091
4 955,250 14 9,147,136
5 1,422,750 15 9,696,182
6 1,939,750 16 10,245,227
7 2,244,250 17 10,794,273
8 1,896,750 18 11,343,318
9 2,445,795 19 11,892,364
10 2,994,841 20 12,441,409

 

Source: Author Computations

102

Table A 7: Farm Annual Net Benefits per Flock Catgory and ND Mortality Levels
Annual Net Beneﬁts at 5% discount Rate (MTN)

 

ND Mortality Cat 1 Cat 2 Cat 3 Cat 4 Cat 5 Cat 6 Overall
5% (41) (50) (59) 88 346 664 23
10% (41) (50) 95 188 413 907 66
’ 15% (41) (1) 246 400 595 1,053 128
20% (41) 67 288 580 647 1,062 165
25% (41) 222 453 616 717 1,063 233
30% (40) 306 503 670 712 1,108 268
35% (71) 282 607 803 882 1,235 284
40% 77 342 639 790 808 1,162 361
45% 124 369 645 782 807 1,163 389
50% 149 392 633 778 794 1,164 404
55% 181 469 625 776 797 1,188 439
60% 242 481 610 760 794 1,187 466
63% 269 494 608 758 793 1,187 481
65% 261 490 608 755 794 1,188 477
70% 227 510 603 752 791 1,200 467
75% 285 494 596 749 789 1,214 488
80% 371 528 591 745 797 1,230 535
85% 415 546 589 757 810 1,245 562
90% 390 553 599 767 823 1,259 556
95% 498 569 605 773 829 1,264 610

 

Source: Author Computations

103

_T_al_)1e A 8: Proﬁtabiﬂy of the Vaccination Program at District Level
Year Benefits Extension costsOthers Costs Net Benefits ,

 

 

0‘ 1998 564,139 (564,139)
1, 1999 148,070 180,000 593,234 (625,164)
2 2000 216,353 180,000 441,434 (405,082)
3 2001 314,530, 180,000 134,530
4 2002 454,093 180,000 274,093
5 2003 649,450 180,000 441,434 28,016
6 2004 917,323 180,000 _ 737,323
7 2005 1,274,894 180,000 1,094,894 f
8‘ 2006 1,736,276 180,000 7 441,434 1,114,841 .
9, 2007 2,307,600 180,000 . , 2,127,600
10 2008, 2,982,246 , 180,000 7 . 7 2,802,246
11, 2009 3,738,811 180,000: 441,434 3,117,376
12 2010 4,544,033 180,000 , 4,364,033‘
13 2011 5,360,299. _ 180,000 5,180,299
14 2012 6,154,461 180,000 441,434 5,533,027
15, 2013 6,903,989 180,000 6,723,989
16 2014 7,598,601 180,000 , 7,418,601
17 2015 8,238,217 180,000 441,434, 7,616,783
18 2016 8,829,350 180,000 , 8,649,350‘
19 2017 9,381,671 180,000 9,201,671
20 2018, 9,905,564 180,000 9,725,564 .

NPV(MTN)* 34,662,840

NPV(U$D)* $1,386,514

IRR 37.39%

Source: Author Computations
Note: The other costs are the costs of training, material and other investments. A discount
rate of ﬁve percent is used for computation of NPV

104

Table A 9: Sensitivity Analysis of the Benefits of Vaccination to Overall Without-

Control ND Mortality at the District Level
Net Beneﬁtsil’er Level of Overall ND Mortality

 

 

Ye_a.r 5%, 10% 15% 29% zsi
0. 1998 (564,139). (564,139) (564, 139) (564,139)’ (564, 139)
1 ‘ 1999 (766,045) (752,858)‘ (733,953), (722,347) (701,666)
2 2000, (610,930) (591,662) (564,038) (547,080) (516,862)
3, 2001 (164,728) (136,717) (96,558) (71,905) (27,975)
4 2002 (157,952) (117,512) (59,533) (23,941) 39,482
5 2003 (589,901 ) (532,063) (449,141) (398,237) (307,528)
6 2004 (135,461) (53,767) 63,358 135,258 263,380
7 2005 (118,099) (4,561) 158,219 258,145 1 436,209
8 2006 (537,132) (382,505), (160,815) (24,725), 217,780
9, 2007 (67,957), 137,550 432,187 613,058 , 935,359
10 2008 (35,201) 230,388 611,164 844,914 _ 1,261,443
11 2009 (439,901) (106,935) 370,440 663,490 1 1,185,688
12 2010 40,630 445,307 1,025,493 ., 1,381,657 1 2,016,320
13i 2011 80,263 557,634 1,242,042 1,662,185 2,410,856
' 14 2012 (322,612) 225,484 1,011,292 1,493,682 2,353,273
15 2013 155,215 770,061 1,651,569 2,192,708 3,156,985
16 2014 188,941 , 865,647 1,835,844 ' 2,431,426 j 3,492,720
17_ 2015 (221,438) 512,231 1,564,094 2,209,810 1 3,360,438 1
18 2016 248,698 , 1,035,011 2,162,351 2,854,400 , 4,087,592
19‘ 2017 275,515 1,111,016 2,308,877 ’ 3,044,218 4,354,551
20‘ 2018 300,952 1,183,1 10 2,447,861 3,224,265 4,607,771
INPV(MTN)* ($3,160,822) $379,528 $5,455,339 $8,571,268 , $14,123,673
NPV(U$D)* ($126,433) $15,181 $218,214 $342,851 1 $564,947
IRR - 6.00% 15.03%, 18.73% 23.95%

 

Source: Author Computations
Note: A discount rate of ﬁve percent is used for computation of NPV

105

Table A 9 (Cont): Sensitivity Analysis of the Beneﬁts of Vaccination to Overall
Without-Control ND Mortality at the District Level .
1r 1- 3 , ,Net Benefits Per Level of Overall NJ) Mortality, _ _, '

 

 

 

Year 30% 35%,; 40% 45% 50%}

0., 1998 (564,139) (564,139)1 (564,139): (564,139) (564,139)]

1 1999 (690,784) (685,853) (662,241) (653,621) (649,028)

2 2000 (500,962) (493,756) (459,256), (446,661) (439,950);

3 2001 (4,860) 5,616 55,772 V 74,083 , 83,839 :

4 2002 72,853 87,978 , 160,389 , 186,825 , 200,910 I

5 2003, (259,800) (238,169) (134,606) (96,797) (76,652) T

6 2004 330,794 _ 361,347 507,626 ' 561,030 589,485 ;

7., 2005 529,901 572,363 775,662 :' 849,882 ~ 889,429 1

8V __ 2006, 345,378 _; 403,208 V 680,079 IV 781,160 , 835,018 g

9 ,2007_ 1,104,944 1,181,802 , 1,549,779 1. 1,684,120 ; 1,755,701 I

10V 2008 1,480,608 ; 1,579,936 2,055,493 1 2,229,111 , 2,321,619 f
11. H 2009_ 1,460,453 , 1,584,979 . 2,181,181 _, 2,398,843 2,514,819 E

V 12. 2010 2,350,260 V 2,501,606 , 3,226,210 ; 3,490,750 _ 3,631,704 1
f 13, 2011. 2,804,783 ; 2,983,315,, 3,838,084‘ 4,150,145 _,,,-4.3,1<_5,4_19,,,
14,“ 2012 2,805,562 , 3,010,546 ; __3,991,954 1 4,350,248; _ 4,541,157 3
15: 2013 , 3,664,358 3V 3,894,305 L 4,995,235 1 5,397,164,; __ 5,611,323 .3
16 ﬂ, 2014 V 4,051,139 i __4,304,221 . 5,515,916 , 5,958,284 6,193,989
17V , 2015, , 3,965,863 4,240,249,; 5,553,939 , 6,033,543 6,289,089 ,
18 _2016, 4,736,459 L,5,030,533 1 6,438,487 1, 6,952,505V 7,226,387 ;

_- 19; 2017 5,044,008 , 5,356,478 ,1 6,852,507 , _ 7,398,680 ;_ _ 7,689,695 i
20 2018, 5,33SL'729;54665JG48 ’ 7,245,218 . 7,821,890 , 8,129,156 =
NPV(MTN)* , 17,045,175 f18,369,234j 24,708,501 , 27,022,851 , 28,255,998»
NPV(U$D)* 681,807 i_ 734,769 :_ 988,340 ; - 1,080,914 ;‘ . 1,130,240 f
IRR 26.27% 27.25%. 31.55%; 32.99%, 33.73%;

 

Source: Author Computations
Note: A discount rate of ﬁve percent is used for computation of NPV

'106

Table A 9 (Cont): Sensitivity Analysis of the Benefits of Vaccination to Overall

Without-Control ND Mortality at the District Level
Net Benefits Per Level of Overall ND Mortality 1

a

70%.

 

 

Year 55%, 60% 62.90%) 65%}
0, 1998; (564,139); V (564,139) (564,139)T (564,139) (564,139)’
IV 1999 (638,048); (629,733) (625,164) (626,563); (629,534).
2 2000: (423,907), (411,758) (405,082)} (407,125) (411,466);
3. 2001 107,161 124,824 134,530 , 131,560 i 125,248 ?
4 2002 234,581 . 260,082 . V 274,093 V. __269,805 260,694 1
5 2003 (28,495) 7,976 ‘ 28,016 21,883 V' 8,851
6 2004 ,657,504 V 709,018 737,323 . V 728,661 710,254 1
7_ 2005 983,962 . 1,055,555 1,094,894 1,082,855 VV 1,057,274!
8 ,2006; 963,763 1 1,061,266V 1,114,841 1,098,446V' _V 1,063,6061
9- 2007 1,926,809 : 2,056,396 _ 2,127,600 2,105,809 _ 2,059,506 1
10_ 2008.» 2,542,751 _ V 2,710,224 2,802,246 1 2,774,084 - 2,714,243 ,
11V 2009 2,792,051“ 3,002,009 3,117,376 . 3,082,070 : 3,007,049 1
12, 2010 3,968,643 V 4,223,820 4,364,033 4,321,123 4 4,229,945 9
13, 2011 4,713,884 ‘ 5,014,899 5,180,299 _ 5,129,681 V V 5,022,124;
_ 14. V2012 V__4,997,509 1 5,343,121 . 5,533,027 V 5,474,909 - 5,351,417;
15; 2013 6,123,252 V 6,510,956 _V 6,723,989 6,658,794 ,~ 6,520,261 ;
16V 2914. VV6,757,424_V _ 7,184,134 _ 7,418,601 7,346,846 . 7,194,376 1
17. 20151 6,899,951VV 7,362,580 ; 7,616,783VV 7,538,988 _7,373,684 1
18, 20161 V_7,881,082V1 ,V8,37V6,906 , V 8,649,350 8,565,973 1. 8,388,807 1
-19...__2017,__, 8,385,344 _ 8,912,185 V 9,201,671 VV_9_,113,079 ._ VV8,9_24,830_V
20‘ 2018' 8,863,651 i 9,419,913 ,1 9,725,564 9,632,025 1 9,433,264 ,1
1 NPV(MTN)* Vs31,203,738 $33,436,173 $34,662,840 834,287,439 j $33,489,755?
. NPV(U$D)* V $1,248,150 . $1,337,447 $1,386,514§ $1,371,498 _ $1,339,590;
IRR 35.45%» 36.71% 37.39%; 37.18%; 36.74%

Source: Author Computations
Note: A discount rate of ﬁve percent is used for computation of NPV

107

Table A 9 (Cont): Sensitivity Analysis of the Benefits of Vaccination to Overall

Without-Control ND Mortality at the District Level

 

Net Beneﬁts Per Level of Overall ND Mortality

 

 

Year 75%, 80%; 85% 90% 95%
0, 1998 (564,139) (564,139) (564,139) (564,139) (564,139)
1 1999 (623,072) (608,523) (600,433) (602,204) (585,713);
2 V 2000, (402,025). (380,767) (368,946) (371,533) V (347,438):
3 f 2001, 138,974 , 169,878 . 187,063 1 183,302 1 f 218,331 1
4 2002 V 280,510 ' 325,127 . 349,937 1 V 344,507 395,079 1
5 2003' _ 37,192 j 101,004 . 136,487 ;_ 128,722 V‘ V . 201,051 5
6 2004 _ 750,285 a _ 840,417 1 890,536} _ 879,567 . 981,730?
7 2005 1,112,908 r 1,238,173 . _1,307,828 1 1,292,584 Vf V_ 1,434,570 §
8 2006’ 1,139,375 _ 1,309,972 V 1,404,836 : , 1,384,075 1 1,577,445 1
9 20071 2,160,206 2,386,940 i 2,513,018 ‘ 2,485,426 V_ 2,742,424 1
10 2008 2,844,384 1 3,137,405 VV 3,300,343 i 3,264,684 3,596,818 .
11, 2009‘ V 3,170,205 1 3,537,562 . 3,741,836 _' 3,697,130 V 4,113,523 §
12» 2010 4,428,240 1 4,874,713 V 5,122,982 3 5,068,648 1 5,574,719 j
13V 2011 5,256,040 : 5,782,716 6,075,582 6,011,487 j 6,608,467 Vj
14 2012 5,619,988 ﬁ_ 6,224,695 V 6,560,951 ‘ 6,487,360 ‘ V 7,172,786 1
15 2013 . 6,821,541 1 7,499,893 7,877,100 7,794,547 V 8,563,448
16 V 2014 7,525,968 , 8,272,568 , 8,687,727 1 8,596,868 , 9,443,128 .1
17 V 2015 7,733,187 V 8,542,633 V 8,992,738 1 8,894,231; V 9,811,726 3
18 2016' 8,774,107 , V 9,641,634 ‘ 10,124,036 110,018,461 1 11,001,790;
__19V _ 2017_ 9,334,232: 10,256,028 1 10,768,606: 10,656,428 11,701,269;
. 20 2018 9,865,528 1 10,838,799 11,380,001 2, 11,261,558 E 12,364,745 1
g NPV(MTN)* $35,224,556 T$39,130,569 341,302,567 340,827,221 1 $45,254,628 ;
NPV(U$D)* , $1,408,982! $1,565,223, $1,652,103‘l $1,633,089 VV $1,810,185,
[RR 1 37.70%: 39.78%_ 40.91% 40.66% 42.90%

Source: Author Computations
Note: A discount rate of ﬁve percent is used for computation of NPV

108

Table A 10: Sensitivity Analysis of the Beneﬁts of Vaccination to Assumption of
Adoption Ceiling Levels at District Level

 

Net Beneﬁts Per Ceiling Level of Adoption

50%?

 

 

Year 15% 1 20% 30%, 40%;
01 __ l998f__ (564,139) - (564,139); “(564,139) ,_ (564,139) V_(56_4_,139)
11 __ 1999 (759,682), V(697,408)__ (651,832) (634,351) (625,164)
2 2000 (584,010) (487,330) (433,530) (414,685) r (405,082)
3 2001 (79,197) 52,626 107,906 1 125.705 V 134,530
4 2002 75,475 : 211,345 . 254,803 267,800 274,093 _.
5 2003 V (52,684) V V 7,610 V 22,215 ‘ VV 26,170 28,016
6 2004 V 844,893 :_ 770,578 ._ V . 747,807 _ 740,781 * 737,323 .
7' 2005‘ 1,283,014 ; 1,153,843 ‘ 1,113,321 1 1,100,952 V 1,094,894 I
8 2006 1,143,208 1 1,111,353 1,111,504 1 1,113,472 ; 1,114,841 ‘
9 _ 2007. 1,769,829 ? 1,920,776 1 _ 2,043,940 : 2,097,565; 2,127,600 ,
10 2008 V, 1,894,806E 2,232,718 ._ V 2,553,232 . 2,709,541 ‘ 2,802,246]
11. 2009 V 1,553,910 1V 2,046,027 . VVV__2,606,493 2,918,516 V1 V3,117,376 .
12 2010; 2,087,991 * 2,696,912 VV 3,504,948 ; 4,014,183 T . 4,364,033
13V 2011 2,179,629 1 2,875,630 3,914,522 5 4,642,876 1 5,180,299 V
14. V 2012 VV 1,831,623 V? 2,594,386 . 3,836,398 4,786,207 L 5,533,027 ‘
15 20131 V 2,369,448 1V V 3,186,237 4,602,318 j V 5,761,857; 6,723,989 V:
'16, V_ 2014 2,469,341 1 3,332,698 4,897,556 6,247,908 7,418,601
17‘ 2015 2,131,605 3 3,037,503 ‘ 4,731,336 1 6,251,905 1 7,616,783
18 V 2016 2,,680 754 i _. 3,627,326 5,435,749‘ 7,107,632 r . 8,649,350
19: 2017, 2, 792, 667 ‘- 3,779,387 ; _ 5 692, 619 1 7,500,149 1, 9,201, 671 1
201 2018. 2, 908, 952 , 3,936,125 5, 948, 019 7,879,141 1 9, 725,564 ,1
NPV(MTN)* $13,423,804 ; $17,715,360 $24, 492, 751 $29,983,073 $34, 662, 840
NPV(U$D)* . $536,952 ‘ $708,614 $979,710 $1,199,323 ‘ $1,386,514
IRR ‘ 27. 68% V 31.46% _ 34.68% _V V 36.33%, 37.39%

Source: Author computations -
Note: A discount rate of ﬁve percent is used for computation of NPV

109

Table A 10 (Cont.:) Sensitivity Analysis of the Beneﬁts of Vaccination to
Assumption of Adoption CeilJ Levels at District Level

Net Beneﬁts Per Ceiling Level _of Adoption

 

I
I , J

 

 

Year 60% . 70% 80%; 90%,
0. 1998‘ (564,139) (564,139) (564,139) (564,139),E
1 1999 V (619,508); (615,677) (612,911) (610, 819):

2 2000f (399,260)’ V (395,355) (392,553) (390, 444)
3 2001. 139,805 1 143,315 V 145,818 1V 147, 695 .
4 2002; 277, 810 1 V 280,265 . 282,008 . 283,309 j
5 _ 2003 29, 083 V V 29,779 V 30,269 L- V 30,632 ;
6 -2005: _ 735,261 j __ 733,889 _ . 732,9V11V_'V 732,178 '
7 2005: 1,091,291 1 1,088,899 _1,087,l95 TV 1,085,920 V
8 V 2006: _ 1,115,791 V 1,116,478 1 V 1,116,998 1,117,402

9 2007 _ 2,146,808: _ 2,160,149 1 _ 2,169,957 1 , 2,177,470
V 10 2008{ 2,863,622 :- 2,907 ,259V'_ 2,939,879 1 “2,965,185?
11 2009: 3,255,187_V_ 3, 356,319 ~» V 3,433,698 3,494,811;
12 2010.- 4,619,114 f, 4, 813, 314 __ 4,966,100 V 5,089,437 V:
13 2011: 5,592,885 V 5,919,522 6,184,506 j 6,403,772 1
14 201ng 6,135,050 : V_6,630,4_86 ‘ 7,045,273 V 7,397,588 1
15 _ 2013 7,534,167 1 V 8,225,440 , 8,822,072 3 9,342,195 V
_ 16 ._ _ 20141 8,441,810 1 9,343,279 V _ 10,143,326 ; 10,858,061 ;
17 V _ 2015V 8, 846, 862 1 9,960,539 [_ 10,973,324 1 V 11,898,203 1
182016 10 073 3,50 !- 11,391,843er 12,615,801 3 13,754,870V
19 2017' _ 10,803, 839 12,314,221 1. 13,740,092 1 15,088,151 1
20?, 2018 11 ,,490 229,? 13,177,489 ‘ 14,791,894 1 16,337,836 ,;
NPV(MTN)* $38, 759, 991 , $42,409,491 $45, 700, 422 $48,695,702 ‘
NPV(U$D)* $1,550,400 $1,696,380 V $1,828, 017 $1,947,828 .
IRR 38.16% 38.74% 39. 21% 39.60%

Source: Author Computations
Note: A discount rate of ﬁve percent is used for computation of NPV

110

Table A 11: Proﬁtability of the Vaccination Program in the 44 Districts of

 

 

 

 

 

Implementation
'Beneﬁts of ‘Extension VOther rProduction , Distribution Net
Year Adoption (My Cost” Co_sl Co_st Beneﬁts
0 1998 24,822,134 (24,822,134)
1 1999 8,000,981 7,920,000 26,102,314 140,208 1475105 (26,309,051)
212000 9,680,923 7,920,000 19,423,114 87,960 92541.25 (17,842,692)
3 2001‘ 11,701,395 7,920,000 - ‘ 107,904 113524 3,559,967
4 2002 14,126,379 7,920,000 ,, - 91,704 96,480 6,018,195
5 2003 17,029,837 7,920,000 19,423,114 136,584 143,698 (10,593,558)
6 2004 20,496,451 7,920,000 - 186,216 , 195,915 12,194,320
7 2005' 24,622,070 7,920,000 . - 215,448 , 226,669 , 16,259,953
82006 29,513,705 7,920,000 19,423,114 182,088 191,572 1,796,932
9 2007' 35,288,862 7,920,000 - 1 234,796 247,025 26,887,040
102008 42,073,993 7,920,000 - ' 287,505 I 302,479 33,564,010
11 2009, 50,001,843 7,920,000 19,423,114 720,000 757,500 1 21,181,230
12,20101 59,207,474 7,920,000 1 - _ 772,708 812,954 49,701,812
132011 69,822,876 1 7,920,000 _ - ‘ 825,417 , 868,407 60,209,052
141,2012‘ 81,970,181 ,_ 7,920,000 19,423,114 878,125 1 923,861 52,825,082 1
15 2013 95,753,727 7,920,000 - 930,833 979,314 85,923,580 .
16‘2014 111,251,483 7,920,000 - 983,542 1,034,768 101,313,174
17 2015 128,506,620 7,920,000 , 19,423,114 “1,036,250 1,090,222 99,037,034
18 2016 147,520,239 7,920,000 ‘ — 1,088,959 1,145,675 137,365,605
19 2017. 168,246,380 7,920,000 - 1,141,667 1,201,129 157,983,584
20 2018 190,590,314 7,920,000 - 1,194,375 1,256,582 180,219,356
‘NPV(MTN)* $410,616,275
NPV(U$D)* $16,424,651
IRR 21.49%

Source: Author Computations
Note: The other costs are the costs of training, material and other investments. A.discount
rate of ﬁve percent is used for computation of NPV

111

Table A 12: Sensitivity Analysis of the Beneﬁts of Vaccination to Overall Without-
Control ND Mortality in the 44 Implementation Districts

Net Beneﬁts Per Level of Overall ND Mortality

 

 

Year 5% 10% 15% 20% 25%

0 1998 (24,822,134) (24,822,134) (24,822,134) (24,822,134) (24,822,134)

1 1999 (33,921,555) (33,209,013) (32,187,440) (31,560,319) (30,442,826)

2 2000 (27,053,570) (26,191,418) (24,955,349) (24,196,553) (22,844,423)

3 2001 (7,573,281) (6,531,193) (5,037,148) (4,119,986) (2,485,658)

4 2002 (7,422,296) (6,164,246) (4,360,577) (3,253,343) (1,280,318)

5 2003 (26,796,533) (25,279,911) (23,105,525) (21,770,717) (19,392,167)

6 2004 (7,306,951) (5,481,604) (2,864,598) (1,258,074) 1,604,656

7 2005 (7,166,624) (4,973,862) (1,830,093) 99,799 3,538,753

8 2006 (26,283,773) (23,655,378) (19,887,040) (17,573,739) (13,451,573)

9 2007 (6,688,415) (3,545,704) 960,011 3,725,972 8,654,752
10 2008 (6,467,134) (2,720,160) 2,651,887 5,949,670 11,826,126
11 2009 (26,392,837) (21,939,835) (15,555,552) (11,636,380) (4,652,645)
12 2010 (6,630,918) (1,358,092) 6,201,574 10,842,289 19,111,771
13 2011 (6,223,662) (5,463) 8,909,588 14,382,344 24,134,474
14 2012 (25,165,140) (17,865,142) (7,399,112) (974,243) 10,474,496
15 2013 (5,180,946) 3,346,571 15,572,497 23,077,727 36,451,608
16 2014 (4,536,634) 5,371,063 19,575,757 28,295,710 43,834,156
17 2015 (23,230,108) (11,785,725) 4,622,121 14,694,542 32,643,005
18 2016 (2,991,974) 10,145,700 28,981,223 40,543,942 61,148,029
19 2017 (2,093,804) 12,889,673 34,371,529 47,558,774 71,057,673
20 2018 (1,117,085) 15,856,271 40,191,022 55,129,597 81,749262
NPV(MT1~ ($204,261,077) ($146,707,658) ($64,193,151) ($13,539,287) $76,722,932
NPV(U$D; ($8,170,443) ($5,868,306) ($2,567,726) ($541,571) $3,068,917
IRR - - 0.00% 4.09% 9.30%

 

Source: Author Computations
Note: A discount rate of ﬁve percent is used for computation of NPV

112

Table A 12 (Cont): Sensitivity Analysis of the Beneﬁts of Vaccination to Overall
Without-Control ND Mortality in the 44 Implementation Districts

Net Beneﬁts Per Level of Overall ND Mortality

 

 

Year 30% 35% 40% 45% 50%
0 1998 (24,822,134) (24,822,134) (24,822,134) (24,822,134) (24,822,134)
1 1999 (29,854,835) (29,588,351) (28,312,491) (27,846,698) (27,598,512)
2 2000 (22,132,974) (21,810,537) (20,266,788) (19,703,194) (19,402,896)
3 2001 (1,625,725) (1,235,993) 629,946 1,311,166 1,674,138
4 2002 (242,174) 228,326 2,480,959 3,303,354 3,741,548
5 2003 (18,140,648) (17,573,445) (14,857,818) (13,866,392) (13,338,134)
6 2004 3,110,936 3,793,599 7,062,023 8,255,265 8,891,056
7 2005 5,348,223 6,168,296 10,094,603 11,528,026 12,291,792
8 2006 (11,282,617) (10,299,621) (5,593,280) (3,875,081) (2,959,579)
9 2007 11,248,122 12,423,469 18,050,733 20,105,144 21,199,789
10 2008 14,918,135 16,319,469 23,028,711 25,478,132 26,783,249
11 2009 (978,020) 687,363 8,660,803 11,571,759 13,122,795
12 2010 23,462,915 25,434,905 34,876,301 38,323,182 40,159,772
13 2011 29,265,742 31,591,293 42,725,452 46,790,330 48,956,205
14 2012 16,498,466 19,228,601 32,299,805 37,071,862 39,614,541
15 2013 43,488,529 46,677,744 61,946,912 67,521,406 70,491,644
16 2014 52,010,003 55,715,393 73,455,879 79,932,605 83,383,577
17 2015 42,086,928 46,367,025 66,859,065 74,340,333 78,326,553
18 2016 71,989,259 76,902,632 100,426,639 109,014,823 113,590,836
19 2017 83,422,062 89,025,749 115,854,807 125,649,604 130,868,533
20 2018 95,755,702 102,103,586 132,495,673 143,591,266 149,503,295
NPV(MTN) 124,216,104 145,740,569 248,794,351 286,417,396 306,463,959
NPV(U$D) 4,968,644 5,829,623 9,951,774 1 1,456,696 12,258,558
IRR 11.49% 12.41% 16.31% 17.59% 18.25%

 

Source: Author Computations

Note: A discount rate of ﬁve percent is used for computation of NPV

113

Table A 12 (Cont.): Sensitivity Analysis of the Beneﬁts of Vaccination to Overall
Without-Control ND Mortality in the 44 Implementation Districts

 

Net Beneﬁts Per Level of Overall ND Mortality

 

Year 55% 60% 62.90% 65% 70%
0 1998 (24,822,134) (24,822,134) (24,822,134) (24,822,134) (24,822,134)
1 1999 (27,005,241) (26,555,934) (26,309,051) (26,384,606) (26,545,150)
2 2000 (18,685,058) (18,141,412) (17,842,692) (17,934,110) (18,128,363)
3 2001 2,541,794 3,198,902 3,559,967 3,449,469 3,214,674
4 2002 4,789,016 5,582,304 6,018,195 5,884,798 5,601,344
5 2003 (12,075,375) (11,119,040) (10,593,558) (10,754,373) (11,096,086)
6 2004 10,410,863 11,561,871 12,194,320 12,000,770 11,589,497
7 2005 14,117,513 15,500,201 16,259,953 16,027,444 15,533,388
8 2006 (771,144) 886,241 1,796,932 1,518,230 926,021
9 2007 23,816,451 25,798,148 26,887,040 26,553,803 25,845,712
10 2008 29,903,026 32,265,752 33,564,010 33,166,700 32,322,461
'11 2009 16,830,420 19,638,346 21,181,230 20,709,056 19,705,741
12 2010 44,549,993 47,874,874 49,701,812 49,142,709 47,954,677
13 2011 54,133,556 58,054,560 60,209,052 59,549,706 58,148,670
14 2012 45,692,612 50,295,766 52,825,082 52,051,027 50,406,249
15 2013 77,591,762 82,968,951 85,923,580 85,019,365 83,098,012
16 2014 91,632,850 97,880,338 101,313,174 100,262,612 98,030,288
17 2015 87,855,290 95,071,765 99,037,034 97,823,530 95,244,972
'18 2016 124,529,429 132,813,641 137,365,605 135,972,553 133,012,475
19 2017 143,343,965 152,792,083 157,983,584 156,394,813 153,018,853
20 2018 163635525 174338398 180219356 178419588 174595285
NPV(MTN) $354,383,661 $390,675,067 $410,616,275 $404,513,606 $391,546,120
NPV(U$D) $14,175,346 $15,627,003 $16,424,651 $16,180,544 $15,661,845
IRR 19.78% 20.89% 21.49% 21.31% 20.92%

Source: Author Computations
Note: A discount rate of ﬁve percent is used for computation of NPV

114

Table A

Without-Control ND Mortality in the 44 Implementation Districts

12 (Cont): Sensitivity Analysis of the Beneﬁts of Vaccination to Overall

 

Net Beneﬁts Per Level of Overall ND Mortality

95%,

 

 

Year 75% 80% 85%, 90%

_0 , 1998, (24,822,134) _ (24,822,134);__(24,822,134) (24,822,134), (24,822,134)
1, _1999 (26,195,999) (25,409,863)_(24,972,720), (25,068,390)! (24,177,316)‘

2 ,2000 (17,705,902) (16,754,704) (16,225,775); (16,341,532), (15,263,363);
3‘ 2001, 3,725,306 . 4,875,026 . 5,514,345 5,374,429 , 6,677,619 ;
4, 2002 6,217,798 , 7,605,785 8,377,596 1V 8,208,684 9,781,945 5

5 2003, (10,352,929) (8,679,663) (7,749,218)" (7,952,848) (6,056,227)

6 2004 12,483,932 14,497,810 15,617,657 . 15, 372, 576 , 17,655,276 _

7 2005 16,607,859 19,027,099 , 20,372,354 20,077,943 , 22,820,115 '

8 2006 2,213,956 5,113,824 6,726,339 6,373,437 , _ 9,660,392 ;
9, 2007 27,385,666 30,852,972 ; 32,781,019 32,359,062 , 36,289,200 ;
10 2008 34,158,509 . 38,292,487 40,591,248 3 V 40,088,160 1 44,773,961 1;

. 11, 2009,: 21,887,748 1 V 26, 800 677V; , 29,532,585 : 28,934,701] 34,503,431 3
, '12, 2010; V 50,538,404 V; 56, 355 ,832_V; V 59,590,701 . ,, 58,882,743 . 65,476,709 1
13: #2011 61,195,638 ‘1, 68, 056,082V;_ 71,870,9361VV71V,036,047 ,V__,78,8V12,255 i
14: 2012 53,983,308 3 ”62,037,284 j V 66,515,820 1 65,535,683 , 74,664,742 :
15; 2013 87,276,565 ; 96,684,843 1 101,916,458 ; 100,771,509 : 111,435,648 .
16V 2014 102,885,139 5 113,816,149 f 119,894,503 : 118,564,242 . 130,954,374 ;
17* 2015; 100,852,813 113,479,225 120,500,333, 118,963,748, 133,275,594:
18 2016' 139,450,043 , 153,944,638 , 162,004,576 , 160,240,642 , 176,670,043 ‘

, 191 2017 160, 360, 879 , 176,891,920 1 186,084,255 ' 184,072,494 _~ 202,810,176
. 20 2018 182 912367 ; 201,638,810 212051932 ' 209,772,999 ; 230,999,135,»
NPV(MTN)* $419,747,763 $483,245,568 $518,554, 482 $510,827 ,052 VV $582,800,842
,NPV(U$1))* $16, 789,911 . $19,329,823 ; $20,742,179 , $20,433,082 ; $23,312,034
[RR 21 76% 23.60% 24.59%! 24.37% 26.36%»

Source: Author Computations
Note: A discount rate of ﬁve percent is used for computation of NPV

115

Table A 13: Sensitivity Analysis of the Beneﬁts of Vaccination to Assumption of
Adoption CeiliniLevel in the 44 Implementation Districts
Net Benefits Per Ceiling Level of Adoption

 

 

Year 10% 20%: 30% 40% 50%:
0' _1998, (24,822,134) . (24,822,134), (24,822,134): (24,822,134)? (24,822,134)
11 1999 (28,115,910) (26,787,841 )VV (26,502,754) (26,378,570) (26,309,051)
2 2000 (19,437,812) (18,255,695) (18,009,079) (17,902,304): (17,842,692)
3- 2001? 2,304,542 3,240,520 . V 3,431,664 V 3,514,056 V V 3,559,967 j
4 2002 5,220,541 1 5,817,290 , 5,937,640 1‘ 5,989,387 ‘ 6,018,195
5‘ 2003, (10,859,875), (10,660,521) __(10,620,403), (10,603,162) _ (10,593,558)
6. 2004‘ 12,440,545 12,256,737 , 12,219,392 1 12,203,285 1 12,194,320 .
7 2005 16,850,780 _16,412,440 1 16,321,426 16,281,977 16,259,953 .
8 2006 2,358,299 . 1,945,648 ? 1,857,208 V 1,818,577 1,796,932
9 2007' 26,792,773 26,858,773 5, 26,875,269 26,882,753 26,887,040 ,
101 2008 31,910,579 33,077,057 : 33,361,881 ; 33,490,621 ~ 33,564,010
11 2009 16,787,807 19,813,912 V: 20,605,862 20,971,038 21,181,230
* 12 2010 41,133,621 V 46,861,772 : V 48,486,409 ; 49,254,323 VV _ 49,701,812 ,
13 _ 2011. 45,817,331 55,107,509 57,982,717 ,1 V 59,381,673 . 60,209,052
: 14: 2012 30,795,859 1 _44,460,770 V 49,093,952 51,423,417 1 52,825,082 L
15 20131 54,329,298 . 73,080,734 i 80,056,136 3 83,692,339 ; 85,923,580 ,
; 16V -2014; 58,164,161 . _82,576,653V1V V 92,532,930 ; 97,929,033 : _ 101,313,174 ,«
‘ 17V 2015, 42,332,111 , 72,825,557 ; 86,425,596 1 94,105,001 ; 99,037,034 :
_ 18} V 2016V 5511423050-. VV101,98_1,543 , 119,878,189 1 130,420,379 ‘-V 137,365,605 1.
: 19 2017; 68,366,306 * V 111,675,295 134,477,483 . 148,497,011 3 157,983,584
'20 2018: 71,468,058 : 121,250,216 1, 149,493,845 ' 167,614,320 1 180,219,356;
NPV(MTN)* $213,141,867 $318,469,082 $365,480,256 $392,729,343 $410,616,275 ,
NPV(U$D)* $8,525,675 $12,738,763 $14,619,210 $15,709,174 $16,424,651
11111 17.36% 19.96% i 20.80% 21.23% 21.49%

Source: Author Computations
Note: A discount rate of ﬁve percent is used for computation of NPV

116

Table A 13 (Cont.:) Sensitivity Analysis of the Beneﬁts of Vaccination to

Assumption of Adoption Ceilin 11g Level 1n the 44 Implementation Districts

 

ﬂ

Net Beneﬁts Per Ceiling Level of Adoption __ .1

 

 

Year 60% 70%; 80% 90%:
of 1998 (24,822,134), (24,822,134): (24,822,134) (24,822,134)j
1‘ 1999, (26,264,625) (26,233,774) (26,211,107): (26, 193 ,747),
2 2000 (17,804,658)j (17,778,271) , (17,758,898) (17, 744, ,070)|
3 2001 3,589,224 -_ 3,609,509 , 3,624,394 'V 3, 635, 781 ;

4 2002 6,036,538 , 6,049,253 I V 6,058,582 1 V 6,065,713
5 2003 (10,587,449) (10,583,21 1) (10,580,099) (10,577,729)

6 2004 12,188,596 12,184,633 ‘ 12,181,729 12,179,491
7, 2005, 16,245,883 . 16,236,123 . 16,228,965 16,223,462 1
8 20061 1,783,075 1,773,447 1 V 1,766,383 3 1,760,941 3
9 f 2007 26,889,807 L 26,891,738 ; 26,893,189 26,894,262 1
10E 2008‘, _ 33,611,422 , _33,644,_568_ , _VV33,669,083 V; _V 33,687,§75 a
11 V 2009; 21,317,828 . 21,413,709 1 21,484,773 21,539,451 §
12 2010. 49,994,811 ‘ 50,201,520 ' 50,355,242 50,473,904 V
13 2011' ___V60,755,727 _ 61,143,787 61,433,620 1 61,658,159 j
14- 2012 _VV53,761,168VV , 54,430,541 . 54,933,104 j 55,324,084 1
_ 15 _ 2013 87,432,223 V 88,520,270 ‘ 5 89,342,249 V 89,984,846 j
161 2014; 103,633,622 _V105_,323,613 j 106,609,544 : 107,620,5071
17V 2015 102,471,990 1 V_ 105,001,443 ', 106,941,969 EV 108,477,371 :
18 V_ 2016_ 142,285,881 . V 145,953,763 1 148, 793,678 1,151,057,017 1
19, 2017 164,828,702 E 170,000,251 , 174,045,386 ; 177,295,336
201 2018! 189,492,897 ‘ 196,600,789 1 202222 576 1 206,779,435J
NPV(MTN)* $423,286,987 . $432,742,522 $440,074,046 $445,925,285 1
NPV(U$D)* . $16,931,479 $17,309,701 1 $17,602,962 3 $17,837,011 V
IRR 21.66% 21.79%; 21.89%: 21.96%

Source: Author Computations
Note: A discount rate of ﬁve percent is used for computation of NPV

117

REFERENCES

118

 

Alders, R. G., R. Fringe, and B. V. Mata. 2000. Experiences in Control of Newcastle
Disease in Mozambique. Paper presented at the Second IAEA/FAO Research
Coordination Meeting on ‘Improvement of Health and Management of Family
Poultry Production in Africa.’ Morogoro, Tanzania, 4-8 September 2000.

Alders, R. and Spradbrow, PB. 2001. Controlling Newcastle Disease in Village
Chickens: A Field Manual. Australian Centre for International Agricultural
Research. Monograph N0. 82.

Alston, Julian M., Norton, George W. and Pardey, Philip G. 1998. Science Under
Scarcity: Principles and Practice for Agricultural Research Evaluation and
Priority Setting. Cambridge, Massachusetts: CAB International.

Alwang, Jeffrey and Paul Siegel. 2002. Measuring the Impacts of Agricultural Research
on Poverty Reduction. Agricultural Economics 29: 1-14.

Asgedom, H. A. 2007. Village Poultry in Ethiopia: Socio-technical Analysis and
Learning with Farmers. PhD Thesis, Wageningen University.

Bangnol, B. 2001. Socio-economic Aspects of Newcastle Disease Control in Village
Chickens. In: Alders, R. G. and Spradbrow, P. B. ed. SADC Planning Workshop
on Newcastle Disease Control in Village Chickens. Proceedings of an
International Workshop, Maputo, Mozambique, 6-9 March, 2000. ACIAR
Proceedings No 103. pp 69-75.

Batz, Franz-J; Willem Janssen, and Kurt J. Peters. 2003. Predicting Technology
Adoption to Improve Research Priority-Setting. Agricultural Economics 28. 151-
164.

Bellon, M. R., Adato, J. Becerril, and D. Mindek. 2005. Impact of Improved Germplasm
on Poverty Alleviation: The case of Tuxpeﬁo — Derived Materials in Mexico.
Mexico, D. F.: CIMMYT.

Bennett, Richard. 2003. The ‘Direct Costs’ of Livestock Disease: The Development of a
System of Models for the Analysis of 30 Endemic Livestock Diseases in Great
Britain. Journal of Agricultural Economics 54.]: 55-71.

Boughton, D., D. Mather, D. Tschirley, T. Walker, B. Cunguara, and E. Payongayong.
2006. Changes in Rural Household Income Patterns in Mozambique, 1996-2002,
and Implications for Agriculture’s Contribution to Poverty Reduction. Research
Paper 61E. Maputo, Mozambique: Directorate of Economics, Ministry of
Agriculture. Retrieved on 05/23/07 from
http: //www. aec. msu. edu/fSZ/mozambique/wps6le.pdf.

119

Branckaert, R. D. S. and E. F. Guéye. 1999. FAO’s Programme for Support to Family
Poultry Production. In Dolberg, Frands and Poul Henning Petersen (ed). Poultry
as a Tool in Poverty Eradication and Promotion of Gender Equality — Proceedings
of a Workshop. Accessed on 07/07/07 from
http://www.fa0.0rg/DOCREP/OO4/AC1 54E/AC1 54EOO.HTM.

Copland, J.W. and R.G. Alders. 2005. The Australian Village Poultry Development
Programme in Asia and Africa. World Poultry Science Journal 61: 31-38.

Dias, P. T., R. G. Alders, R. Fringe, B. V. Mata. 2001. Laboratory and Field Trials with
Thermostable Live Newcastle Disease Vaccines in Mozambique. In: Alders, R. G.
and Spradbrow, P. B. ed. SADC Planning Workshop on Newcastle Disease
Control in Village Chickens. Proceedings of an International Workshop, Maputo,
Mozambique, 6-9 March, 2000. ACIAR Proceedings No 103. pp 91-96.

Food and Agricultural Organization. 2001. The State of Food and Agriculture 2001.
Rome, Italy. Accessed on 01/05/09 from
http://www.fao.org/docrep/OO3/x9800e/x9800el4.htm.

Gemo, H., C. K. Eicher, and S. Teclemariam. 2005. Mozambique Experience in Building
a National Extension System. East Lansing, Michigan. Michigan State University
Press. '

Gittinger, James Price. 1982. Economic Analysis of Agricultural Projects. 2"d Edition.
Baltimore, Maryland. The Johns Hopkins University Press. (EDI Series of
Economic Development).

Government of Mozambique. 2006. Council of Ministers. Poverty Reduction Strategy
Paper 11. Maputo.

Gueye, EP. 1998. Poultry plays an important role in African village life. World Poultry
14.10: 14-17. ‘

Hoy, M., J. Livernois, C. McKenna, R. Rees, T. Stengos. 2001. Mathematics for
Economics. 2nd Edition. Cambridge, Massachusetts. The MIT Press.

Jensen, H. A., and Dolberg F. 2003: A conceptual framework for using poultry as a tool
in poverty alleviation. Livestock Research for Rural Development 15.5. Retrieved
on 07/13/2008, from http://www.cipav.org.co/1rrd/lrrd15/5/ien5155.htm.

Johnston, Joe, and Robin Cumming. 1991. Control of Newcastle Disease in Village
Chickens with Oral V4 Vaccine. Camberra, Australia. Australian Centre for
lntemational Agricultural Research. Economic Assessment Series 7: 23p.

Joshi, P. K. 2003. Impact Assessment of Agricultural Research: Concepts and
Measurement. ICAR. New Delhi, India.

120

Marasas, C. N., M. Smale, and R. P. Singh. 2004. The Economic Impact in Developing
countries of Leaf Rust Resistance Breeding in CIMMYT - Related Spring Bread
.Wheat. Economics Program Paper 04-01. Mexico, D. F .: CIMMYT..

Maredia, Mywish; Derek Byerlee and Jock Anderson. 2000. Ex-Po'st Evaluation of
Economic Impacts of Agricultural Research Programs: A Tour of Good Practice.
Paper presented to the workshop on “The Future of Impact Assessment in
CGIAR: Needs, Constraints, and Options”, Standing Panel on Impact Assessment
(SPIA) of the Technical Advisory Committee, Rome, May 3-5, Rome.

Masters, W. A., B. Coulibaly, D. Sanogo, M. Sidibe’, and A. Williams. 1996. The
Economic Impact of Agricultural Research: A Practical Guide. Department of
Agricultural Economics, Purdue University, West Lafayette, IN.

Mather, D., B. Cunguara, and D. Boughton. 2008. Household Income and Assests in
Rural Mozambique, 2002-2005: Can Pro-Poor Growth Be Sustained? Research
Paper 66. Maputo, Mozambique: Directorate of Economics, Ministry of
Agriculture. Retrieved on 04/23/2009 from
http://www.aec.msu.edu/f82/mozambique/wps66.pdf.

McDermott, J ., P. Coleman, and T Randolph. 2001. Methods for Assessing the Impact of
Infectious Diseases of Livestock — Their Role in Improving the Control of
Newcastle Disease in Southern Africa. In: Alders, R. G. and Spradbrow, P. B. ed.
SADC Planning Workshop on Newcastle Disease Control in Village Chickens.
Proceedings of an lntemational Workshop, Maputo, Mozambique, 6-9 March,
2000. ACIAR Proceedings No 103. pp 118-126.

McSween, S., T. Walker, V. Salégua, and R. Pitoro. 2006. Economic Impact on Food
Security of Varietal Tolerance to Cassava Brown Streak Disease in Coastal
Mozambique. Research Report No. 1E. Maputo, Mozambique: Directorate of
Training, Documentation, and Technology Transfer, Institute of Agricultural
Research of Mozambique. Retrieved on May 23, 2007 from
http://www.aec.msu.edu/st/mozambique/iiam/rr_1e.pdf.

Ministerio da Agricultura e Desenvolvimento Rural (MADER). 1992. 1992. Relatério
Anual. Direccao Nacional de Pecuaria. Maputo, Mocambique.

Ministerio da Agricultura e Desenvolvimento Rural (MADER). 2004. Relato’rio Anual
2003. Direccao Nacional de Pecuéria.

Ministerio da Agricultura e Desenvolvimento Rural (MADER). 2005. Relato'rio Anual
2004. Direccao Nacional de Pecuaria.

Ministerio de Administracao Estatal (MAE). 2005. Perﬁl do Distrito de Chibuto,
Provincia de Gaza. Perﬁs Distritais. Mozambique. 24p. Acessed on 07/20/2007.
www.undp.orgmz/en/content/download/S38/2470/ﬁle/Chibutopdf.

121

Mom's, ML. and P.W. Heisey. 2003. Estimating the Beneﬁts of Plant Breeding
Research: Methodological Issues and Practical Challenges. Agricultural
Economics 29, 241-252.

Otte, M. J ., and P. Chilonda. 2000. Animal Health Economics: An Introduction. Livestock
Information. Sector Analysis and Policy Branch, Animal Production and Health
Division (AGA), F A0, Rome, Italy. Accessed on 07/30/2007
ftp://ftp.fao.org/docrep/fao/O l 0/ag275e/ag275e.pdf.

Pritchett, J ., D. Thilmany and K. Johnson. 2005. Animal Disease Economic Impacts: A
Survey of Literature and Typology of Research Approaches. International Food
and Management Review 8(1).

Ross, S. A., R. W. Westerﬁeld, J. Jaffe. 2008. Corporate Finance. 8‘h Edition. New York.
MacGraw-Hill Irwing.

Southern Africa Newcastle Disease Control Project (SANDCP). 2003. First Report
Maputo Mozambique. Maputo, Mozambique.

Southern Africa Newcastle Disease Control Project (SANDCP). 2005. Annual Plan
2005-06. GRM lntemational PYT LTD.

Spradbrow, P.B.V1999. Epidemiology of Newcastle Disease and the Economics of its
Control. In Dolberg, Frands and Poul Henning Petersen (ed). Poultry as a Tool in
Poverty Eradication and Promotion of Gender Equality — Proceedings of a
Workshop. Accessed on 07/07/07 from
http: //www. ardaf. org/NR/rdonlyres/1E34B48B- DC4E-4366- AAAD-
6B252FAEDE08/O/ l 99916spradbrow. pdf.

Udo, H. M. J, A. H. Asgedom and TC. Viets. 2006. Modelling the lrnpact of
Interventions on the Dynamics in Village Poultry Systems. Agricultural Systems
88:255-269.

Walker, T.S., D. Tschirley, J. Low, M. Pequenino Tanque, D. Boughton, E.
Payongayong, and M. Weber. 2004. Determinants of Rural Income in
Mozambique in 2001-02. Report No. 57. Maputo, Mozambique: Directorate of
Economics, Ministry of Agriculture. Retrieved on 05/23/2007 from
http://www.aec.msu.edu/fs2/mozambique/wps57e.pdf.

Walker, T.S., R. Pitoro, A. Tomo, I. Sitoe, C. Saléncia, R. Mahanzule, C. Donovan, and
F. Mazuze. 2006. Priority Setting for Public-Sector Agricultural Research in
Mozambique with the National Agricultural Survey Data. Research Report No.
3B. Maputo, Mozambique: Directorate of Training, Documentation, and
Technology Transfer, Institute of Agricultural Research of Mozambique.
Retrieved on 05/23/2007 from
hug/WWW.aec.msu.edu/fs2/mozambique/iiam/rr_3epdf.

122

Woolcock, R. F., M. Harun, R. G. Alders. 2004. The impact of Newcastle disease control
in village chickens on the welfare of rural households in Mozambique. Paper
presented at the Fourth Co-ordination Meeting of the FAO/IAEA Co-ordination
Research Programme on the ‘Assessment of the effectiveness of vaccination
strategies against Newcastle Disease and Gumboro Disease using immunoassay-
based technologies for increasing backyard poultry production in Africa.’ Vienna,
Austria, 24-28 May 2004.

World Bank. 2006. Mozambique Agricultural Development Strategy: Stimulating
Smallholder Agricultural Growth. World Bank, AFTSl. Retrieved on 05/23/ 2007
from

http://siteresources.worldbank.og/MOZAMBIOUEEXTN/Resources/Moz_AG_S
trategypdf.

123

   

111111111111|||11|1111111111111l111H11111111111111H