THREE ESSAYS IN DEVELOPMENT ECONOMICS 

 

By 

 

Mukesh Kumar Ray 

A DISSERTATION 

Submitted to 

Michigan State University 

in partial fulfillment of the requirement 

for the degree of 

Agricultural, Food and Resource Economics – Doctor of Philosophy 

2018 

 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

 

 

 

ABSTRACT 

 

THREE ESSAYS IN DEVELOPMENT ECONOMICS 

 

By 

 

Mukesh Kumar Ray 

 
Households experience many types of micro and macro level exogenous shocks.  This 

dissertation includes three essays that focus on household and individual level welfare effects 

and coping strategies to shocks. In the first essay, I examine the role of risk preferences in the 

adoption decision of climate smart technology that helps farmers cope with water shortages due 

to changing climate. In the second essay I estimate the impact of in-utero weather shock on 

child’s nutrition status and whether the effects differ by household income, implying different 

coping strategies used by rich versus poor households. In my final essay, I focus on the impact of 

macro-level political shocks when three large states in India were divided into smaller states and 

estimate the developmental outcomes of this division on both the new and the parent states. 

My first essay is situated in the literature of technology adoption with a focus on role of 

risk preferences in adoption decision. I conducted household surveys and a field experiment to 

elicit the risk preference of farmers from two northern states in India; and used them to explain 

their adoption decision of Laser Land Leveler. This technology is primarily a water saving 

technology and is therefore very useful in this region of India under growing ground water stress. 

Along with reducing water use, this technology also reduces greenhouse gas emission, and is 

therefore termed as a climate smart technology. Our analyses extend the measurement of risk 

preferences beyond expected utility theory by incorporating prospect theory. I use survival 

analysis as my empirical strategy to model time to adoption and find that risk averse farmers and 

farmers who overvalue smaller probabilities tend to adopt this technology sooner than others. 

 
 

 

My second essay examines the impact of rainfall shocks during gestational period that 

pregnant mothers in Malawi experience on the health outcomes of their children less than 5 years 

of age. About 85% of Malawians reside in rural areas and most of them depend on agriculture. 

With almost no irrigation system in place, the reliance on natural rainfall is high and therefore 

shifts in rainfall patterns highly affect their lives. I find that higher rainfall in the in-utero period 

is associated with better child nutrition outcomes, but excessive rainfall negatively affects child’s 

health. We also find that these effects persist until initial years for all children. However, for 

children from richer households the effects seem to fade away over time, but for children born in 

poorer households, these effects persist longer term.  

My third essay focusses on the role of greater autonomy or more homogenous 

jurisdictions on development indicators. In this essay I examine the case of India where the 

splitting of three bigger states lead to the creation of three new states in the year 2000. Since 

states can be considered as the proximate determinants of local institutions driving 

developmental outcomes, a change in their boundaries provides an opportunity to evaluate the 

impact of these shifts on the provision of public goods and services.  I use the comprehensive 

datasets of Demographic Health Survey and apply quasi-experimental methods to show that 

districts in the newly created states on average are doing better than the parent states on 

development indicators after splitting from their parent states. We find that splitting of states has 

induced better governance, service delivery, and significantly improved living standards and 

development outcomes in the newly formed states but negative effect in the parent states post-

split.  The results of this study provide new evidence that institutions matter for development, 

and local control of institutions can have large economic impacts on the smaller state that is 

separating from a larger state. 

 
 

 

ACKNOWLEDGEMENTS 

 
 
First and foremost, I want to thank my advisor, Dr. Mywish Maredia for her guidance and support 

throughout this journey. Thank you for believing in me and motivating me to make progress.  Her 

generosity of time and thoughtful comments were extremely important for my dissertation. Her 

understanding of economics and keen insights in applied work have greatly sharpened my focus. 

My future academic endeavors will deeply benefit from this learning experience.  

I also wish to thank other members of my committee, Dr.  Robert Shupp for introducing 

me to the wide literature, Dr. Songqin Jin, for analyzing my initial research ideas and support on 

the choice of methodology, Dr. Robert Myers for his theoretical insights in the literature and Dr. 

Robert Richardson for his keen attention to detail and constructive criticism. I also thank Ashleigh 

Booth and Nancy Creed for all of the work they have done to keep me on track with department 

and university requirements.  Furthermore, I thank Department of Agricultural, Food and Resource 

Economics,  Asian  Study  Centre,  The  Graduate  School  at  Michigan  State  University  and  the 

Consortium General for International Agricultural Research for their generous financial support 

without which this research would have been impossible. 

In addition, I want to thank my AFRE colleagues Jason Snyder, Joey Goeb, Hamza Haider, 

Joshua Gill, Sarah Kopper, Nathaly Riviera, Ashesh Prasann, Jarrad Farris and Mary Doige who 

have helped me with their comments during different parts of the dissertation- your feedback has 

immensely helped me to improve on my work. Colleagues and friends outside AFRE have also 

helped me broaden my horizon of thinking about my research. In particular, I would like to thank 

Ajith Chakkedath, Vivek Vellanki  and Awa Sanou for the long discussions that allowed me to 

think about my research from different perspectives.  

 

 
iv 

Above all, I would like to thank my wonderful family without which I would not be here. 

I thank my father Jai Narayan Ray for his encouragements during this program and his constant 

belief in me. I also wish to acknowledge Ujjawala and Urja for their constant emotional support 

and  the  joy  that  they  brought  to  my  life  all  these  years.  Finally,  I  want  to  thank  my  partner 

Meenakshi who was always there for me in this entire journey- without you this day would never 

have come.  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
v 

 

 

TABLE OF CONTENTS 

LIST OF TABLES ................................................................................................................................... viii 

LIST OF FIGURES .................................................................................................................................. xii 

CHAPTER 1: INTRODUCTION .............................................................................................................. 1 
REFERENCES ............................................................................................................................................ 4 

CHAPTER 2: RISK PREFERENCES AND CLIMATE SMART TECHNOLOGY ADOPTION: A 
DURATION MODEL APPROACH FOR INDIA ................................................................................... 6 
2.1. Introduction .......................................................................................................................................... 6 
2.2. Literature Review .............................................................................................................................. 10 
2.3. Background on Laser Land Leveler and the Technology Adoption Survey: ............................... 15 
2.3.1 Background on Laser Land Leveler ...................................................................................... 15 
2.3.2 Survey Procedure ..................................................................................................................... 18 
2.3.3 Data Description ...................................................................................................................... 19 
2.4. Field Experiment Design and Procedure ......................................................................................... 20 
2.4.1 Experiment Design ................................................................................................................... 20 
2.4.2 Field Experiment Procedure ................................................................................................... 23 
2.4.3 Estimation of Parameters ........................................................................................................ 27 
2.5. Conceptual Model .............................................................................................................................. 28 
2.6. Econometric Framework and Results .............................................................................................. 32 
2.6.1 Econometric Framework......................................................................................................... 32 
2.6.2 Results ....................................................................................................................................... 34 
2.6.3 Robustness Check .................................................................................................................... 36 
2. 6.4 Further Discussion .................................................................................................................. 38 
2. 7. Conclusion ......................................................................................................................................... 39 
APPENDICES ........................................................................................................................................... 41 
APPENDIX A: FIGURES ........................................................................................................................ 42 
APPENDIX B: TABLES .......................................................................................................................... 45 
REFERENCES .......................................................................................................................................... 54 

CHAPTER 3: EFFECTS OF IN-UTERO RAINFALL SHOCKS ON CHILDREN’S HEALTH IN 
MALAWI ................................................................................................................................................... 60 
3.1. Introduction ........................................................................................................................................ 60 
3.2. Literature Review .............................................................................................................................. 62 
3.3. Conceptual Framework ..................................................................................................................... 67 
3.3.1 Health Production Function .................................................................................................... 67 
3.3.2 Pathways Through Which Rainfall Can Impact Child’s Health ......................................... 69 
3.4.  Data. ................................................................................................................................................... 71 
3.4.1 Socio-Economic and Anthropometric Data ........................................................................... 71 
3.4.2 Rainfall Data ............................................................................................................................ 73 
3.5. Empirical Strategy ............................................................................................................................. 75 
3.5.1 Enumeration Area Fixed Effect: ............................................................................................ 76 
3.5.2 The Household Fixed Effect Model ........................................................................................ 77 
3.6. Results ................................................................................................................................................. 78 
3.7. Conclusion .......................................................................................................................................... 81 
APPENDICES ........................................................................................................................................... 83 
APPENDIX A: FIGURES ........................................................................................................................ 84 

 

 
vi 

APPENDIX B: TABLES .......................................................................................................................... 86 
REFERENCES .......................................................................................................................................... 93 

CHAPTER 4: DO SMALLER STATES LEAD TO MORE DEVELOPMENT? EVIDENCE FROM 
SPLITTING OF LARGE STATES IN INDIA ....................................................................................... 99 
4.1. Introduction ........................................................................................................................................ 99 
4.2. The Case of Formation of New States Through Splitting of Larger States ................................ 101 
4.2.1 Examples from Other Countries .......................................................................................... 101 
4.2.2 Formation of States in Post-independence India: Historical Roots and Current Status . 102 
4.2.3 The Formation of Three New States by Splitting Larger States in 2000: Driving Forces
 .......................................................................................................................................................... 104 
4.3. Concepts and Theories Underlying the Rationale for and Against the Formation of a New State 
and Research Hypotheses ....................................................................................................................... 108 
4.3.1 Pro- and Anti-Split Arguments ............................................................................................. 109 
4.3.2 Research Hypotheses ............................................................................................................. 111 
4.4. Data ................................................................................................................................................... 113 
4.5. Empirical strategy ............................................................................................................................ 114 
4.5.1 Difference - in - Difference Approach .................................................................................. 114 
4.5.2 Geographical Regression Discontinuity Approach ............................................................. 119 
4.6. Results ............................................................................................................................................... 121 
4.7.  Robustness Check ........................................................................................................................... 131 
4.8. Conclusion ........................................................................................................................................ 133 
APPENDICES ......................................................................................................................................... 136 
APPENDIX A- FIGURES ...................................................................................................................... 137 
APPENDIX B: TABLES- SUMMARY OF RESULTS ....................................................................... 140 
APPENDIX C: TABLES – DETAILED RESULTS ............................................................................ 151 
APPENDIX D: TABLE- STATEWISE DETAILED RESULTS ....................................................... 174 
APPENDIX E: TABLES- RESULTS OF PARALLEL TREND TESTS .......................................... 210 
REFERENCES ........................................................................................................................................ 222 
 
  

 

 

 

 

 

 

 

 

 

 

 

 

vii 

 

LIST OF TABLES 

Table 2.1: SUMMARY CHARACTERISTICS ........................................................................... 45 

Table 2.2: PROSPECT THEORY VS EXPECTED UTILITY THEORY ................................... 45 

Table 2.3 : PAYOFF MATRIX FOR THE EXPERIMENTAL GAME ...................................... 47 

Table 2.4: PROBIT MODEL FOR ADOPTION .......................................................................... 49 

Table 2.5: WEIBULL MODEL FOR DURATION OF TIME TO ADOPTION ......................... 51 

Table 2.6: WEIBULL MODEL FOR DURATION OF TIME TO ADOPTION: ROBUSTNESS 
CHECK ......................................................................................................................................... 53 

Table 3. 1: DESCRIPTIVE STATISTICS.................................................................................... 86 

Table 3.2: EFFECTS OF IN-UTERO RAINFALL SHOCKS ON CHILD HEALTH: 
ENUMERATION AREA/VILLAGE FIXED EFFECTS ............................................................ 87 

Table 3. 3: EFFECTS OF IN-UTERO RAINFALL SHOCKS ON CHILD HEALTH: 
HOUSEHOLD FIXED EFFECTS ................................................................................................ 90 

Table 4. 1: ARGUMENTS FOR AND AGAINST SPLITTING A LARGE STATE INTO TWO 
AND THEIR POTENTIAL DEVELOPMENT IMPACTS ON THE PARENT STATE AND 
THE NEWLY CREATED SMALLER STATES....................................................................... 140 

Table 4. 2: CHANGE IN DEVELOPMENTAL INDICATORS FROM YEAR 2001 TO YEAR 
2011 FOR BOTH PARENT AND NEW STATES .................................................................... 141 

Table 4. 3: TABLES OF MEAN COMPARISONS BETWEEN DISTRICTS ON BOTH SIDES 
OF BORDER SEPARATING THE NEWLY FORMED STATE AND PARENT STATE, 1998-
99................................................................................................................................................. 142 

Table 4. 4: SUMMARY OF REGRESSION RESULTS FOR THE EFFECTS OF SPLITTING 
ON SELECTED DEVELOPMENTAL OUTCOMES: EDUCATION AND HOUSEHOLD 
DWELLING CHARACTERISTICS .......................................................................................... 143 

Table 4. 5: SUMMARY OF REGRESSION RESULTS FOR THE EFFECTS OF SPLITTING 
ON SELECTED DEVELOPMENTAL OUTCOMES: DURABLE GOODS AND AMENITIES
..................................................................................................................................................... 144 

Table 4. 6: SUMMARY OF REGRESSION RESULTS FOR THE EFFECTS OF SPLITTING 
ON SELECTED DEVELOPMENTAL OUTCOMES: MAJOR HEALTH INDICATORS ..... 145 

 

 

viii 

Table 4. 7: SUMMARY OF REGRESSION RESULTS FOR THE EFFECTS OF SPLITTING 
ON SELECTED DEVELOPMENTAL OUTCOMES: GOVERNMENT HEALTH PROGRAM 
INDICATORS FOR CHILDREN AND PREGNANT WOMEN .............................................. 146 

Table 4. 8: SUMMARY OF STATE-WISE REGRESSION RESULTS FOR THE EFFECTS OF 
SPLITTING ON SELECTED DEVELOPMENTAL OUTCOMES OF NEW STATES AS 
COMPARED TO THE PARENT STATES ............................................................................... 147 

Table 4. 9: SUMMARY OF STATE-WISE REGRESSION RESULTS FOR THE EFFECTS OF 
SPLITTING ON SELECTED DEVELOPMENTAL OUTCOMES IN COMPARISON TO THE 
ADJACENT STATES USING DID ESTIMATOR ................................................................... 148 

Table 4. 10: SUMMARY OF REGRESSION STATE-WISE RESULTS FOR THE EFFECTS 
OF SPLITTING ON SELECTED HEALTH OUTCOMES OF NEW STATES AS COMPARED 
TO THE PARENT STATES ...................................................................................................... 149 

Table 4. 11: SUMMARY OF REGRESSION STATE-WISE RESULTS FOR THE EFFECTS 
OF SPLITTING ON SELECTED DEVELOPMENTAL OUTCOMES IN COMPARISON TO 
THE ADJACENT STATES USING D-I-D ESTIMATOR........................................................ 150 

TABLE 4. 12: DEVELOPMENTAL INDICATORS- PARENT STATE AS COMPARISON 
AND NEWLY FORMED STATES AS TREATMENT ............................................................ 151 

Table 4. 13: DEVELOPMENTAL INDICATORS-ADJACENT STATES AS COMPARISON 
AND BOTH PARENT AND NEW STATES AS TREATMENTS .......................................... 153 

Table 4. 14: DEVELOPMENTAL INDICATORS- RDD USING DISTRICTS ON THE 
NEWLY FORMED STATES AS TREATMENT AND PARENT STATE AS COMPARISON
..................................................................................................................................................... 155 

Table 4. 15: CONSUMER DURABLES - PARENT STATE AS COMPARISON AND NEWLY 
FORMED STATES AS TREATMENT ..................................................................................... 157 

Table 4. 16: CONSUMER DURABLES - ADJACENT UNDIVIDED STATES AS 
COMPARISON AND BOTH PARENT AND NEW STATES AS TWO TREATMENTS ..... 159 

Table 4. 17: CONSUMER DURABLES - RDD USING DISTRICTS ON THE NEW STATES 
AS TREATMENT AND PARENT STATE AS COMPARISON ............................................. 161 

Table 4. 18:  MAJOR HEALTH INDICATORS- PARENT STATE AS COMPARISON AND 
NEWLY FORMED STATES AS TREATMENT ..................................................................... 162 

Table 4. 19: GOVERNMENT PROVISIONS - PARENT STATE AS COMPARISON AND 
NEWLY FORMED STATES AS TREATMENT ..................................................................... 164 

 

 
ix 

Table 4. 20: MAJOR HEALTH INDICATORS- RDD USING DISTRICTS ON THE NEW 
STATES AS TREATMENT AND PARENT STATE AS COMPARISON ............................. 166 

Table 4. 21: GOVERNMENT PROVISIONS- RDD USING DISTRICTS ON THE NEW 
STATES AS TREATMENT AND PARENT STATE AS COMPARISON ............................. 168 

Table 4. 22: MAJOR HEALTH INDICATORS (CHILD) ADJACENT STATES AS 
COMPARISON AND BOTH PARENT AND NEW STATES AS TWO TREATMENTS ..... 170 

Table 4. 23: GOVERNMENT PROVISIONS (CHILD) - ADJACENT STATES AS 
COMPARISON AND BOTH PARENT AND NEW STATES AS TWO TREATMENTS ..... 172 

Table 4. 24: SELECTED DEVELOPMENTAL INDICATORS- BIHAR AS COMPARISON 
AND JHARKHAND AS TREATMENT ................................................................................... 174 

Table 4. 25: SELECTED DEVELOPMENTAL INDICATORS- MADHYA PRADESH AS 
COMPARISON AND CHHATTISGARH AS TREATMENT ................................................. 176 

Table 4. 26: SELECTED DEVELOPMENTAL INDICATORS- UTTAR PRADESH AS 
COMPARISON AND UTTARAKHAND AS TREATMENT .................................................. 178 

Table 4. 27:  SELECTED DEVELOPMENTAL INDICATORS- BORDERING DISTRICTS IN 
BIHAR AS COMPARISON AND BORDERING DISTRICTS IN JHARKHAND AS 
TREATMENT ............................................................................................................................ 180 

Table 4. 28:  SELECTED DEVELOPMENTAL INDICATORS- BORDERING DISTRICTS IN 
MADHYA PRADESH AS COMPARISON AND BORDERING DISTRICTS IN 
CHHATTISGARH AS TREATMENT ...................................................................................... 182 

Table 4. 29: SELECTED DEVELOPMENTAL INDICATORS- BORDERING DISTRICTS IN 
UTTAR PRADESH AS COMPARISON AND BORDERING DISTRICTS IN 
UTTARAKHAND AS TREATMENT....................................................................................... 184 

Table 4. 30: SELECTED DEVELOPMENTAL INDICATORS- NEARBY STATES AS 
COMPARISON AND BOTH BIHAR AND JHARKHAND AS TREATMENTS ................... 186 

Table 4. 31: SELECTED DEVELOPMENTAL INDICATORS- NEARBY STATES AS 
COMPARISON AND BOTH MADHYA PRADESH AND CHHATTISGARH AS 
TREATMENTS .......................................................................................................................... 188 

Table 4. 32: SELECTED DEVELOPMENTAL INDICATORS- NEARBY STATES AS 
COMPARISON AND BOTH UTTAR PRADESH AND UTTARAKHAND AS 
TREATMENTS .......................................................................................................................... 190 

Table 4. 33: SELECTED HEALTH INDICATORS- BIHAR AS COMPARISON AND 
JHARKHAND AS TREATMENT ............................................................................................. 192 

 

 
x 

Table 4. 34: SELECTED HEALTH INDICATORS- MADHYA PRADESH AS COMPARISON 
AND CHHATTISGARH AS TREATMENT ............................................................................ 194 

Table 4. 35: SELECTED HEALTH INDICATORS- UTTAR PRADESH AS COMPARISON 
AND UTTARAKHAND AS TREATMENT ............................................................................. 196 

Table 4. 36: SELECTED HEALTH INDICATORS- BORDERING DISTRICTS IN BIHAR AS 
COMPARISON AND BORDERING DISTRICTS IN JHARKHAND AS TREATMENT ..... 198 

Table 4. 37: SELECTED HEALTH INDICATORS- BORDERING DISTRICTS IN MADHYA 
PRADESH AS COMPARISON AND BORDERING DISTRICTS IN CHHATTISGARH AS 
TREATMENT ............................................................................................................................ 200 

Table 4. 38: SELECTED HEALTH INDICATORS- BORDERING DISTRICTS IN UTTAR 
PRADESH AS COMPARISON AND BORDERING DISTRICTS IN UTTARAKHAND AS 
TREATMENT ............................................................................................................................ 202 

Table 4. 39: SELECTED HEALTH INDICATORS- NEARBY STATES AS COMPARISON 
AND BOTH BIHAR AND JHARKHAND AS TREATMENTS .............................................. 204 

Table 4. 40: SELECTED HEALTH INDICATORS- NEARBY STATES AS COMPARISON 
AND BOTH MADHYA PRADESH AND CHHATTISGARH AS TREATMENTS .............. 206 

Table 4. 41: SELECTED HEALTH INDICATORS- NEARBY STATES AS COMPARISON 
AND BOTH UTTAR PRADESH AND JHARKHAND AS TREATMENTS .......................... 208 

Table 4. 42: TESTING PARALLEL TREND - ADJACENT STATES AS COMPARISON AND 
BOTH PARENT AND NEW STATES AS TWO TREATMENTS .......................................... 210 

Table 4. 43: TESTING PARALLEL TREND (HEALTH INDICATORS) - ADJACENT 
STATES AS COMPARISON AND BOTH PARENT AND NEW STATES AS TWO 
TREATMENTS .......................................................................................................................... 212 

Table 4. 44: TESTING PARALLEL TREND - PARENT STATE AS COMPARISON AND 
NEWLY FORMED STATES AS TREATMENT ..................................................................... 214 

Table 4. 45: TESTING PARALLEL TREND (HEALTH INDICATORS)- PARENT STATE AS 
COMPARISON AND NEWLY FORMED STATES AS TREATMENT ................................ 216 

Table 4. 46: TESTING PARALLEL TREND - RDD USING DISTRICTS ON THE NEWLY 
FORMED STATES AS TREATMENT AND PARENT STATE AS COMPARISON ............ 218 

Table 4. 47: TESTING PARALLEL TREND (HEALTH INDICATORS)- RDD USING 
DISTRICTS ON THE NEWLY FORMED STATES AS TREATMENT AND PARENT STATE 
AS COMPARISON .................................................................................................................... 220 

 

 
xi 

 

LIST OF FIGURES 

Figure 2. 1: FIRST ROW OF THE RISK GAME ........................................................................ 42 

Figure 2. 2: DISTRIBUTION OF RISK PREFERENCE PARAMETERS ................................. 43 

Figure 2. 3: LOCATION OF TWO STUDY DISTRICTS ........................................................... 44 

Figure 3.1: MALAWI PROVINCES AND DISTRICTS MAP ................................................... 84 

Figure 3.2: ANNUAL AND AVERAGE RAINFALL FOR ALL OF MALAWI ....................... 85 

Figure 4.1: THREE STATE BEFORE AND AFTER SPLITTING. LEFT MAP 
CORRESPONDS TO THE THREE STATES BEFORE SPLITTING AND THE RIGHT MAP 
SHOWS THREE NEW AND THE THREE OLD STATES AFTER SPLITTING. ................. 137 

Figure 4.2: NEWLY CREATED BOUNDARY LINES (IN YELLOW) SHARED BY BOTH 
NEW AND OLD STATES ......................................................................................................... 138 

Figure 4.3: ASPIRANT STATES OF INDIA REPRESENTING ALL THE CURRENT 
DEMANDS FOR NEW STATE FORMATION. ....................................................................... 139 

 

 

 

xii 

 

 

CHAPTER 1: INTRODUCTION 

In most developing countries, households experience exogenous shocks that significantly 

determine their developmental outcomes. Some of these shocks have varying effects on 

households and individuals depending on their characteristics, and thus require micro-level 

coping strategies, while other shocks affect large geographical areas like a district, state or a 

country and have macro-level developmental impacts. In this dissertation I present three essays 

that explore the welfare impacts of such exogenous micro and macro level shocks felt at 

individual and household levels. While the first two essays focus on micro level shocks and 

coping mechanisms, the third essay explores the welfare effects of an exogenous macro level 

shock.  

The first essay focusses on a micro level shock and examines the role of risk preferences in the 

adoption decision of a climate smart technology. This technology helps Indian farmers cope with 

the exogenous shock of climate change and declining ground water table. Both, climate change 

and declining water table are major issues affecting a majority of farmers in northern India, 

where this study was conducted (Bates et al. 2008). On the one hand, changing climate makes 

precipitation less predictable and on the other hand, groundwater aquifers are dwindling due to 

over-exploitation (Hanjra and Qureshi 2010) leading to water crisis for irrigation in this entire 

region. The intensity of the problem is felt heterogeneously among the farmers based on their 

ground water level and availability of other sources of irrigation water like canals. Adoption of 

laser land leveler technology saves irrigation water significantly and is therefore considered as a 

coping mechanism for irrigation water scarcity. As this technology seems to reduce risk, it is 

likely that farmer’s risk preferences will influence its adoption decision. I therefore elicited 

 

 
1 

 

farmers’ risk preferences using lab-in-field type experiments and use them to explain their 

adoption decisions.  

The second essay underscores the impact of a micro level weather shock on child health.  

This essay examines how rainfall shocks during gestational period that Malawian pregnant 

mothers experience affects the health outcomes for children when they are less than 5 years of 

age.  A significant majority of Malawians live in rural areas and depend on agriculture for their 

livelihood, making the rainfall events extremely important for their lives. In the absence of 

modern irrigation technology and social safety nets, any rainfall shocks are likely to have wide 

ranging consequences for these people. As almost all of Malawian agriculture is rain fed, a 

reduced rainfall will directly affect the production of food grain leading to a lower nutrient intake 

of the household members. Furthermore, reduced rainfall can increase the stress on the existing 

water resources, decreasing the quantity and quality of clean water availability for personal 

hygiene, and increase the time to collect water (Mara, 2003; Sobsey, 2002).  A combination of 

less nutritious food and poor sanitation is likely to have a negative effect on a pregnant woman. 

Fetal Origin Hypothesis (FOH) clearly underscores the importance of in utero period in 

determining the nutritional status and shaping the future abilities and health trajectories of a baby 

(Barker, 1999). According to the Fetal Origin Hypothesis, an exogenous rainfall shock that 

induces lower nutrient intake and poor sanitation during the in-utero period is highly likely to 

have an effect on child’s health. This essay is an attempt to quantify such effects for different 

groups of people in society and draws policy implications to mitigate the effects of such shocks 

on vulnerable population.  

The third essay focusses on a macro level shock of political reorganization of large Indian 

states on the developmental outcomes for people residing in those states. In the year 2000, three 

 

 
2 

 

large states of India were split into six smaller states. Since states can be considered as the 

proximate determinants of local institutions driving developmental outcomes, a change in their 

borders provides an opportunity to evaluate the impact of such a macro shock on the provision of 

public goods and services.  Smaller states tend to represent the political aspirations of the 

residents of the states much better as key political decisions are taken geographically closer to 

them. As the formation of the new states were broadly on the lines of linguistic and ethnic 

composition, creation of smaller states leads to more homogenous states. Linguistic 

compatibility and cultural homogeneity are likely to facilitate better management, 

implementation, and allocation of public resources in the provision of basic social and economic 

infrastructure services. The third essay tries to examine these hypotheses and finds empirical 

evidence of developmental outcomes of reorganization of states.  

 

 

 

 

 

 

 

 

 

 

 

 
3 

 

 

 

 

 

 

 

 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

REFERENCES 

 
4 

 

REFERENCES 

 

Bates, B., Kundzewicz, Z., & Wu, S. (2008). Climate change and water. Intergovernmental Panel 
on Climate Change Secretariat. 

Hanjra, M. A., & Qureshi, M. E. (2010). Global water crisis and future food security in an era of 
climate change. Food Policy, 35(5), 365-377. 

Marra, M., Pannell, D. J., & Ghadim, A. A. (2003). The economics of risk, uncertainty and 
learning in the adoption of new agricultural technologies: where are we on the learning curve?. 
Agricultural systems, 75(2-3), 215-234. 

Sobsey, M. D., Water, S., & World Health Organization. (2002). Managing water in the home: 
accelerated health gains from improved water supply (No. WHO/SDE/WSH/02.07). Geneva: 
World Health Organization. 

Mara, D., & Horan, N. J. (Eds.). (2003). Handbook of water and wastewater microbiology. 
Elsevier. 

Barker, D. J. (1999). Fetal origins of cardiovascular disease. Annals of medicine, 31(sup1), 3-6. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
5 

 

 

 

CHAPTER 2: RISK PREFERENCES AND CLIMATE SMART TECHNOLOGY 

ADOPTION: A DURATION MODEL APPROACH FOR INDIA 

2.1. Introduction  
 

Climate change is one of the biggest challenges of the 21st century. It has already caused 

observable negative effects on the environment and it is further expected to have widespread 

impacts on human and natural systems around the world. Sea levels have risen, observed 

temperature have increased and precipitation has become more erratic across the globe leading to 

a myriad of new problems. As agriculture is heavily dependent on weather conditions in most 

parts of the world, changes in temperature and precipitation regimes due to changing climate is 

expected to impact agriculture disproportionately more than other fields of study (IPCC 2014). In 

many regions, agricultural production is already being adversely affected by: rising temperatures, 

increased variability in temperature, changes in levels and frequency of precipitation, greater 

frequency of droughts, and the increasing intensity of extreme weather events. As impact of 

climate change on agriculture intensify, it will become increasingly difficult to grow crops in the 

same ways and in the same places as we have done in the past. A variety of climate smart 

technologies and practices are being developed to help farmers cope with the adverse effects of 

changing climate. These new technologies have three main objectives: sustainably increasing 

agricultural productivity and incomes; adapting and building resilience to climate change; and 

reducing and/or removing greenhouse gas emissions, where possible. In this paper we study the 

adoption of one such climate smart technology--Laser Land Leveler (LLL) in the Indian states of 

Punjab and Haryana.   

 

 
6 

 

LLL uses laser guided beams to level the fields with higher precision compared to 

traditional levelling. The technology helps reduce water use and increases crop yield that helps 

farmers adapt to changing climate (Kaur 2012). Furthermore, LLL saves fuel used for irrigation 

and indirectly reduces carbon emissions to help mitigate climate change. As LLL increases yield 

and reduces irrigation cost it is first order stochastically dominant over traditional leveling. For a 

region preoccupied with increasing yields and with rapidly declining stock of groundwater, this 

technology seems an appropriate choice for most farmers. Unfortunately, even after 15 years 

since its introduction, the technology has not been adopted by a significant population of 

farmers.  Given such promising prospects one must ask why hasn’t this technology yet been fully 

adopted in this region? What are the underlying factors for adoption or non-adoption of this 

technology in this region? These questions are the focus of this study.  

 Technological innovation has always been a major driver of economic development. 

More particularly, advances in agricultural technologies have been considered as a major factor 

in improving living standards of rural population. Given the much-advocated benefits from new 

agricultural technologies, the delay in adoption of proven technologies has always puzzled 

economists.  An extensive literature exists that explain why farmers do not adopt or delay 

adoption of new technology. Much of this literature cites low level of education (Foster and 

Rosenzweig, 1995), lack of information and access to credit (Barrett et al., 2004), learning 

spillovers (Munshi, 2004), tenure insecurity, small farm size, and unreliable supply of 

complementary inputs as the main constraints to technology adoption.  Furthermore, there has 

been significant enquiry on the role of uncertainty about the effectiveness of a new technology as 

a major constraint to adoption (Feder, 1980). In most developing countries, agriculture is 

considered a risky proposition given its dependence on environmental factors that are beyond 

 

 
7 

 

farmers’ control. Moreover, any new agricultural technology can have a wide distribution of 

outcomes, increasing the associated uncertainty. Thus, any new agricultural technology is 

inherently perceived as an uncertain proposition. Consequently, farmers’ perceived uncertainty 

regarding the effectiveness of the technology allows individual subjective risk preferences to 

play a major role in technology adoption (Holden, 2015). 

In this paper we use the time duration from the introduction of LLL in India to farmers’ 

awareness and adoption of LLL to analyze how risk attitudes affect technology adoption 

decisions. Most new technologies promise higher yield but also increase the associated risk, 

often leading to low adoption by risk averse farmers. However, LLL being a climate smart 

technology comes with a scientific assertion that it would increase yield, lower cost of irrigation 

and is considered less risky than the traditional leveling in terms of yield risk. Despite the 

promising scientific claims, ex ante we do not know whether farmers would consider this 

technology as more or less risky than the traditional leveling technology. For farmers this is like 

any other new technology and their subjective beliefs about the success of the scientific claims 

about the technology and their risk preferences play a major role in their adoption decision. It is 

likely that if the farmers consider LLL as risky, then like most other new technologies, risk 

loving farmers are more likely to adopt it sooner. However, if the farmers consider it risk 

reducing then farmers who are risk averse are likely to adopt it sooner. This makes risk 

preferences an important factor in adoption decision.  

Risk preferences have long been recognized as an important factor in explaining 

technology adoption. There are multiple reasons why it is important to account for individual 

risk preferences when studying technology adoption. First, omission of risk preferences is likely 

to bias significant variables like education and wealth, which are correlated with risk (Lybertt,T. 

 

 
8 

 

Just,D.R. 2007).  Second, individual risk preferences have shown to be defining wealth 

accumulation and income growth (McInish, Ramaswami, and Srivastava, 1993).   Third, 

Dohmen (2012) finds intergenerational correlation in risk preferences, which could explain the 

low level of intergenerational income mobility and wealth accumulation. The degree to which 

they play a role in wealth accumulation is less understood from an empirical perspective.  

The common approach to illustrate an individual’s risk attitudes is to use expected utility 

theory (EUT) approach where the curvature of the utility function is solely defined by risk 

aversion. We relax some of these restrictive assumptions inherent in EUT by incorporating 

aspects of prospect theory (PT) where the utility function is jointly defined by risk aversion, loss 

aversion, and non-linear probability weighting measure. While loss aversion measures one’s 

sensitivity to losses as compared to gains, non-linear probability weighting measure captures 

individual’s tendency of overvaluing small probabilities and undervaluing large probabilities.  

We believe that PT might explain individual decision-making better than EUT, as it captures loss 

aversion and non-linear probability weighting measure and is therefore more likely to simulate a 

real-life decision making involving risk.  

 We develop a simple technology adoption model that includes risk preferences. 

Considering uncertainty regarding LLL as a major inhibiting factor for its adoption, we allow 

subjective risk preferences of the farmers to be different from the objective risk. It is likely that 

the subjective risk of the farmers regarding its effectiveness in reducing irrigation cost and 

increasing yield is different from objective risk associated with the technology. The model 

assumes that subjective beliefs of the farmers about the effectiveness of LLL in reducing water 

use for irrigation and their subjective belief of future rainfall will determine their adoption 

decision. As this technology is risk reducing, the model predicts that if the farmers perceive LLL 

 

 
9 

 

as more effective in reducing water cost and increase yield as claimed by the scientists, then 

more risk averse farmers would adopt it sooner. Alternatively, if the farmers considered this 

technology as not very effective in reducing water cost and increasing yield then they would 

delay adoption.  

This paper is organized as follows. Section 2 covers the literature review. Section 3 

provides information on the laser land leveler technology and describes the survey and dataset. 

Section 4 explains the design of the field experiment and presents descriptive analysis. In Section 

5 a conceptual framework is used to describe the role of risk preferences in technology adoption 

decisions. Section 6 provides a general econometric framework to test the predictions and 

describes the empirical results. Section 7 concludes the paper.  

 

2.2. Literature Review  

 

Adoption of technological innovations in agriculture has attracted considerable attention among 

development economists because livelihood of majority of the populations in developing 

countries revolves around agriculture. There is growing consensus among economists believing 

that adoption of new technologies seems to offer an opportunity to increase production and 

income substantially (Foster,A.D,  Rosenweig, M.R, 1996). But the introduction of many new 

technologies has met with only partial success resulting in low rates of adoption. A series of 

studies tried to investigate the reasons for low adoption and farm size was one of first factors 

examined by the empirical literature on adoption. Farm size can have different effects on the rate 

of adoption depending on the characteristics of the technology and institutional setting. The 

literature suggests that large fixed costs reduce the tendency to adopt and slow the rate of 

 

 

10 

 

adoption by smaller farms (Weil,1970). Several studies reviewed by Binswanger (1978) found a 

strong positive relationship between farm size and adoption of tractor power in south Asia. Other 

empirical studies have shown that inadequate farm size also impedes an efficient utilization and 

adoption of certain types of irrigation equipment such as pumps and tube wells (Hodgdon and 

Singh 1966; Dobbs and Foster 1972; Gafsi and Roe 1979).  

Education and labor availability have been considered as important determinants of 

adoption of agricultural technology. Education as defined by formal schooling is hypothesized to 

play an important role in determining allocative ability than worker ability (Welch 1970). This 

hypothesis has been supported by several studies; Ram (1980) found that farm operators' 

contributions to production are positively related to their education, whereas workers' 

contributions are not. Sidhu (1976) found that although farmers' education has some effect on 

yield, it had relatively greater effect on gross sales by farmers in the early stages of the Green 

Revolution in the Punjab. These studies suggest that farmers with better education are early 

adopters of modern technologies and apply modern inputs more efficiently throughout the 

adoption process. Labor availability is another important focus of the technology adoption 

literature. Some new technologies are relatively labor-saving, and others are labor using, which 

means the local labor availability dictates which technologies are adopted and which are not. 

Hicks and Johnson (1974) found that higher rural labor supply leads to greater adoption of labor-

intensive rice varieties in Taiwan, and Harriss (1972) has found that shortages of family labor 

explain non-adoption of HYVs in India.  

Credit constraint among small farmers is another topic studied extensively as a limiting 

factor for technology adoption. Access to capital is required to finance many new agricultural 

technologies, thus differential access to capital explains differential rates of adoption. Lack of 

 

 

11 

 

capital or access to credit becomes a bigger problem when the technology is indivisible like 

tractor and requires a very large initial investment (Lipton 1976; Bhalla & Alagh 1979). 

However, there are many studies which suggest that credit constraints do significantly limit 

adoption of even divisible technologies like HYV or fertilizer where the fixed pecuniary costs 

are not large (Frankel 1969; Wills 1972; Khan 1975).  

There is a general scarcity of empirical studies that have addressed the role of risk and 

uncertainty in adoption. Feder (1985) in this early review of the literature on adoption attributed 

this to lack of adequate methods to measure risk preferences. Since then, there have been a few 

studies using multiple methods to elicit risk preferences of farmers. Even fewer studies of risk 

and adoption have used direct interview techniques to investigate the effect of farmers’ risk 

attitudes and perception of riskiness of enterprises on their allocative decisions (Binswanger, 

1980; O'Mara, 1983; Lindner and Gibbs, 1990; Smale et al., 1994; ; Marra et al, 2003). 

Binswanger et al. (1980) were one of the first to elicit the risk preferences of a sample of 

Indian farmers using several elicitation techniques, including a gambling game with real money. 

Their method measured farmer’s risk aversion, which was used as an explanatory factor in 

regression for adoption. Their results showed mixed results and were inconclusive about the role 

of risk aversion on adoption. Byerlee and Polanco (1986) used farm survey data from Mexico to 

analyze the step wise adoption of components of packages and practices. They found that 

adoption of each innovation was explained primarily by profitability and riskiness. Kebede 

(1992) found that risk aversion played a significant role in the adoption of new technologies 

among Ethiopian farmers. He found that income, farm size, education, family labor and 

experience were negatively related to risk-taking behavior in farming areas where off-farm 

 

 

12 

 

income diversification opportunities existed. However, these factors were found to have the 

opposite effect in areas where farming was the only source of income. 

Marra and Carlson (1990) studied the role of relative risk and risk aversion on the 

allocation of acreage between full-season soybeans and double-cropped wheat/soybeans in the 

southeastern US. They used expected utility framework along with covariance decomposition to 

estimate Arrow-Pratt risk aversion coefficients without assuming a functional form of utility 

function. They found relative riskiness of the two production technologies and risk aversion 

explained the aggregate decision to double crop. Further, Shapiro et al. (1992) used a tobit model 

to explain the effect of risk aversion on adoption of double cropping in the US. They elicited 

farmers’ perception of riskiness of enterprises by asking them to construct frequency 

distributions using the fixed interval approach. They found that adopters were, on average, more 

risk averse than non-adopters. However, differences in their risk perceptions were more 

important than their risk preferences in explaining adoption. 

Smale et al. (1994) conducted a study with Malawian maize growers on the role of risk 

preferences in the adoption of new seed varieties. They found that farmers’ perception of the 

relative riskiness of new seed varieties influenced their adoption and intensity of cultivation 

decisions. The study elicited subjective yield distribution from individual farmers for the 

different varieties of maize to calculate relative riskiness estimates. Ghadim. et.al. 

(2000) conducted a more comprehensive study collecting 3-year series information on farmer’s 

actual and planned adoption behavior for a new crop type, chickpeas in Australia. They 

calculated Arrow-Pratt coefficient of risk aversion based on a set of questions on hedging.  

 

 

13 

 

Most of the studies mentioned above use a limited definition of risk preferences as 

defined under expected utility theory (EUT) framework. Under EUT, the shape of the utility 

function is solely defined by risk aversion. However, there is a growing literature suggesting that 

expected utility theory does not provide a plausible theory of risk aversion for both small-stakes 

and large-stakes gambles and this decision theory should be replaced with an alternative theory 

characterized by loss aversion and non-linear probability weighting measure (Cox  and  Vjollca, 

2001, Rabin 2000). Kahneman and Tversky (1979) in their seminal work developed Prospect 

Theory (PT) as an alternate model to describe decision making under risk, in which value is 

assigned to gains and losses rather than to final assets, and probabilities are replaced by decision 

weights. As PT allows for more flexibility and seems to be closer to actual decision making 

under risk, we use this framework to explain risk preferences of farmers. In choosing PT we do 

not reject EUT completely, as the latter is a special case in PT when loss aversion is same as risk 

aversion and non-liner probability weighting measure does not matter. Using PT, we do a 

comprehensive testing of roles of risk aversion, loss aversion and probability weighted measure 

in technology adoption decision for LLL.  

Recently two papers Liu (2013) and Ward & Singh (2015) comprehensively assess the 

relevance of EUT and PT for adoption of new types of seeds. In a study of adoption of BT cotton 

seeds in China, Liu (2013) found that more risk averse and loss averse farmers adopt the BT 

cotton seed later, while farmers who overweighed smaller probabilities adopted the seeds earlier. 

Ward and Singh (2015) conducted a similar study on the adoption of drought tolerant paddy in 

India and found that risk averse and loss averse farmers are more likely to switch to new rice 

seeds, which outperform other cultivars under moderate and severe drought conditions. Both 

these studies clearly show how loss aversion and probability weighting measure is an important 

 

 

14 

 

parameter, along with risk aversion in defining technology adoption. Building on their work, this 

paper also uses all three risk parameters in explaining technology adoption decision. To our 

knowledge, this is the first study that examines the role of risk preference in explaining the 

adoption of a climate smart technology, which is considered to be and promoted as risk reducing 

among scientific and development community.  

 

2.3. Background on Laser Land Leveler and the Technology Adoption Survey: 

 

2.3.1 Background on Laser Land Leveler 

 

This paper analyzes the adoption of LLL in rice-wheat cropping systems in India. Rice–wheat 

(RW) is the most important cropping system for food security in South Asia with 13.5 M ha of 

land devoted to this farming system and providing food for more than 400 million people. In 

India, the rice-wheat cropping system contributes 26% of total cereal production and 60% of 

total calorie intake (Gupta et al., 2003).  The area under the RW system is static and the 

productivity and sustainability of the system are threatened, because of the inefficiency of current 

production practices, shortage of resources such as water and labor, and demographic changes. 

Pressure is increasing on the limited land, water and increasing variability in the climatic factors 

are increasingly making it difficult to meet the increasing demand for food for the burgeoning 

population.   

To address the dual need of ensuring food security and mitigating the effects of climate 

change, research program on Climate Change, Agriculture and Food Security (CCAFS) of the 

Consultative Group on International Agricultural Research (CGIAR), in partnership with the 

 

 

15 

 

Indian agricultural research centers and other agencies have introduced an array of technologies 

and practices under the rubric of climate-smart agriculture (CSA) in India. LLL is one of the 

technologies promoted by CGIAR, with a claim to equip farmers to adapt to changing weather 

patterns amidst depleting ground water resources. LLL has been around for a while in developed 

countries but was introduced in India in the last decade with a specific aim to reduce irrigation 

water usage.  

Traditionally, farmers in India level their land after ploughing and before sowing to allow 

flood irrigation water to evenly spread across the field. Conventionally, farmers in India use 

plankers drawn by draft animals or by small tractors for land leveling. In the states of Punjab and 

Haryana farmers mostly use iron scrappers or leveling boards connected to 4- wheel tractors. 

Traditional land leveling includes field survey, staking and designing the field, calculation of 

cuts and fills and then using a scraper and a land plan to even the land. Despite all these labor-

intensive efforts, desired accuracy is not achieved. These leveling practices are crude and do not 

achieve a precise land leveling leading to less input use efficiencies and low yield at the cost of 

more water. Laser land leveling solves this problem by meeting the objectives of achieving a 

better crop stand, saving irrigation water and improving the input use efficiencies. LLL is a 

machine equipped with a laser-operated drag bucket that ensures more flat, even surface in less 

time compared to the traditional scraper (Appendix-4). A more even land means irrigation water 

reaches every part of the field with minimal waste from water run-off or water-logging. This 

ensures that farmers use water more efficiently while reducing the labor in irrigation. It also 

reduces potential nutrient loss through improved runoff control, leading to greater efficiency of 

fertilizer use and higher yields.  Use of laser land leveler reduces emission of greenhouse gases 

 

 

16 

 

through decreased water pumping time, decreased cultivation time and better use of fertilizers 

(Aryal et al. 2015).  

Success of LLL in the Indian context has been well documented in a paper by Jat et al. 

(2009) where they show that LLL improves rice-wheat (RW) cropping-system productivity by 

7.4% as compared to traditional land leveling. Total irrigation water savings under LLL versus 

traditional leveling were estimated to be 12–14% in rice and 10–13% in wheat. LLL improved 

per hectare profitability of the RW cropping-system from US$113 to $175 (Jat et al 2009). The 

technology is well suited for smallholder farmers as it does not require any major investment. 

Cost of a new LLL is close to $4,000, which is quite high for small farmers in India, so this 

technology is mostly bought by large farmers who in turn rent it out to other farmers. In our 

sample almost all the farmers using LLL rented it for a nominal charge equivalent to a few days 

of rural wages, which is quite affordable.  The learning curve for farmers to use this technology 

is not very steep as the method of traditional leveling is very similar to LLL. Further, the 

operators of the LLL usually guide farmers about its usage making it easy for them. Given these 

superior characteristics and success of LLL in field trial by agronomists, it is quite puzzling to 

see that a significant proportion of farmers have still not adopted this technology. In conversation 

with experts prior to the field survey and in the pilot survey, we found that uncertainty about the 

technology was one of the main reasons farmers delayed the adoption of what seems likes an 

obvious choice of technology.  

 

 

 

 

17 

 

2.3.2 Survey Procedure 

 

The risk experiments were designed by the authors and conducted as part of larger representative 

adoption surveys. These surveys were conducted in several districts in India to assess the 

adoption of LLL and other natural resource management technologies, namely direct seeded rice 

and zero tillage. However we did not find much adoption of either zero tillage or direct seeded 

rice in these regions and therefore focused only on laser land leveler. The two districts focused in 

this study are—Ludhiana in Punjab with a population of approximately 3.5 million, and Karnal 

in Haryana with a population of 1.5 million as shown in appendix 3.  These districts fall under 

the rice-wheat cropping system and are considered to be most developed agriculturally in their 

respective states and the entire country. Over time there has been a series of interventions by 

various CGIAR institutions (especially, CIMMYT under the CCAFS program) and the state 

governments to introduce climate smart technologies, including the LLL in these two districts.   

Both the districts included in the study have national or state level agricultural 

universities. While Ludhiana has one of the most prominent state agricultural universities of 

India (Punjab Agricultural University), Karnal is home of the Indian Institute for Wheat and 

Barley Research of the Indian Council of Agricultural Research (ICAR). In summary, these two 

districts have always had new agricultural technologies available and historically have led the 

country in terms of technology diffusion and adoption.  

For the broader adoption study, in each study districts 80 villages were randomly selected 

from a list of all wheat growing villages using the probability proportionate to size (PPS) method 

(where size was measured by net sown area in the village as obtained from the last Census data). 

In each of these villages 10 households were selected randomly by the enumerators and a 

 

 

18 

 

detailed questionnaire was administered to collect data on farmer and household characteristics, 

technology specific data for LLL, adoption of other technologies by the household, and farmers’ 

perception on constraints in wheat and rice farming.  Data was collected from September to 

November 2015 using a Computer Assisted Personal Interview (CAPI) method from a total of 

1600 households across the two districts. The risk experiments were conducted on a sub-set of 

villages included in this larger study. The total sample size of households subjected to risk 

experiments include 201 households in Karnal and 231 households in Ludhiana. Agricultural 

data collected in these surveys corresponded to Rabi 2014-15 and Kharif 2015 agricultural 

season. Prior to conducting the experiment, enumerators explained the set of standardized 

instructions and asked questions to confirm whether the farmers understood the experiment. 

Next, before the real experiment, two rounds of practice experiments were conducted without 

cash, but with candies as the payoff outcomes to make sure the farmers understood the rules of 

the game.  

 

2.3.3 Data Description  

 

Figure 1 shows the distribution of risk aversion, loss aversion and non-linear probability 

weighting measure. The lowest value for risk aversion is 0.05 and the highest value is 1.5 with 

majority of the values lying between 0.05 and 1, suggesting that most of the farmers in our 

sample are risk averse. For non-linear probability weighting measure the lowest value was 0.05 

and the highest was 1.45, with most values under 1 suggesting that at an average farmer 

overvalue smaller probabilities of high impact events. Most of the loss aversion values are above 

1 suggesting farmers consider losses differently than gains.  

 

 

19 

 

Table 1 shows the summary statistics for the variables of interest for the households that 

participated in the experiments.   The average farmer is around 42 years old at the time of the 

survey and has 9.43 years of formal education. This is expected as the study districts are in an 

advanced part of India and farmers are likely to be more educated. The farmers are well to do in 

this region with an average land holding of 9.43 acres (3.7 hectares) and the average household 

poverty score on a scale of 0-100 is close to 67, which implies less than 10% probability that a 

typical household included in our survey is living below poverty line. The average household 

size is 5.95 and the average number of working members is 2.44, which is typical for rural India. 

From the total 481 households, 432 households or close to 90% had heard about the Laser Land 

Leveler technology, and of these 432, 67% had adopted it. For explaining time to adoption, we 

use this sample of 432 households who had heard about the technology by the time the survey 

was conducted.   

 

2.4. Field Experiment Design and Procedure 

 

2.4.1 Experiment Design 

 

We use a variant of the Multiple List Price (MPL) method (Miller 1969) to elicit the risk 

preference of Indian farmers. The MPL offers choices between two or more uncertain prospects 

with fixed amount and varying probabilities. In the widely-used Holt and Laury (2002) version 

of MPL, subjects face a single list of binary decisions between two gambles. The payoffs remain 

the same in each decision but the probabilities vary, meaning that any respondent with consistent 

 

 

20 

 

risk preferences should have a “switch” point between preferring gamble A or gamble B. 

Tanaka, Camerer and Nyugen (2009, hereafter TCN) offers a new variant of MPL in which the 

probabilities remain same and the payoff increases in one of the gambles to find at what payoff 

(if they ever switch) the participant switches from a less risky gamble to a high risk-high payoff 

gamble. This variant of the MPL also incorporates PT by eliciting loss aversion and non-linear 

probability weighting measure along with risk aversion. Ex ante, both PT and EUT can act as a 

potential theory explaining farmer’s decision making regarding this new technology. However, 

prior to the study it was not clear which one would explain farmer’s behavior better. We 

therefore use TCN design as it allows us to estimate empirical specifications with the flexibility 

to incorporate both PT and EUT. Further, TCN design has been tested in Vietnam, China (Liu, 

2013) and India (Ward and Singh, 2015) with less educated farmers and it was simple enough to 

understand by farmers in Indian states of Punjab and Haryana.  

 

Following the TCN procedure the following utility function form is assumed. 

 

Where  

     U(x,p;y,q) =
(cid:3)(cid:4)(cid:5)(cid:6)+ (cid:9)(cid:4)(cid:10)(cid:6)(cid:11) (cid:3)(cid:4)(cid:12)(cid:6)− (cid:3)(cid:4)(cid:5)(cid:6)(cid:14)       (cid:15)(cid:16)        (cid:12)>(cid:5)>0 (cid:19)(cid:20) (cid:12)<(cid:5)<0                 (cid:4)1(cid:6)
(cid:2)
(cid:9)(cid:4)(cid:10)(cid:6)(cid:3)(cid:4)(cid:12)(cid:6)+(cid:9)(cid:4)(cid:23)(cid:6)(cid:3)(cid:4)(cid:5)(cid:6)                 (cid:15)(cid:16)             (cid:12)<0<(cid:5)                                            (cid:4)2(cid:6)                     
(cid:3)(cid:4)(cid:12)(cid:6)=(cid:25) (cid:12)(cid:26)                    (cid:16)(cid:19)(cid:20)    (cid:12)>0
and   w(p) =(cid:28)(cid:12)(cid:10) [−(cid:4)−(cid:30)(cid:31)(cid:10)(cid:6) ],   for 0<α ≤ 1                                                                             

−(cid:27)(cid:4)−(cid:12)(cid:26)(cid:6)         (cid:16)(cid:19)(cid:20)     (cid:12)<0                                                                                                 

 

 

 

21 

 

In the above utility function, x and y are the outcomes and p and q are the probabilities 

associated with these outcomes. Parameter σ describes the curvature of the value function above 

zero. For a risk averse individual σ < 1, for a risk neutral individual σ =1 and for a risk loving 

individual σ > 1. Risk aversion decreases in σ, i.e. as σ increases risk aversion decreases. 

Parameter λ, is the loss aversion parameter, that defines the shape of the value function below 

zero comparative to the value function above zero. If λ=1 then there is no kink in the curvature 

of the value function around 0, suggesting that individuals treat losses like gains, however if λ≠1 

indicates a kink.   λ>1 implies a more convex shape of the value function below zero suggesting 

individuals are more averse to loses than to gains. The non-linear probability weighting measure 

α comes from an axiomatically derived weighting function model by Prelec (1998).  α captures 

the degree to which less likely events are disproportionately weighted when valuing risky 

prospects. w(p) shows the probability weighting function. α <1 suggests an inverted S shaped 

w(p), indicating overweighting of low probabilities of larger losses or gains and underweighting 

of higher probabilities. If the estimated results give us α=1 and λ =1, then the above model 

reduces to EUT.  However, our empirical results suggest that prospect theory captures the risk 

behavior of our sampled farmers better than EUT as suggested by table 2, supporting the 

application of prospect theory for the empirical analysis presented in this paper. The average risk 

aversion of sampled farmers is 0.64, the average value of loss aversion is 3.12 and the average 

value of probability weighting measure is 0.7, which are close to the findings in the studies 

conducted by Liu in China and Ward et al. in India working with farmers of similar profile.   

 

 

 

 

22 

 

2.4.2 Field Experiment Procedure 

 

Prior to conducting the experiment, enumerators explained the set of standardized instructions 

and asked questions to confirm whether the farmers understood the experiment. Next, before the 

real experiment, a round of practice experiment was conducted with candies as the payoff 

outcomes. This was conducted to make sure the farmers understood the rules of the game and 

how to note down their choices. A detailed instruction sheet for the enumerators explaining the 

protocol for the risk game has been can be found in appendix 2. Appendix 5 shows some pictures 

of the actual games being played by farmers in the states of Punjab and Haryana.  

To estimate the risk preference parameters (risk aversion coefficient, loss aversion 

coefficient and probability weighting measure), field experiments were conducted in the study 

area with a sample of farmers surveyed to measure the adoption of LLL. Field experiments 

involved playing lottery games with individual farmers (main decision maker of the household) 

who were selected for the adoption survey. Game participants were given three independent 

series of games that had a total of 35 pairwise choices. The first and second series contained 14 

choices each and the third series had 7 choices between two lotteries: A and B.  These two 

options (A and B) differ in the expected value of the lottery, which is a function of the 

probability of winning the noted cash value in a scenario of a random draw of a number between 

1 to 10. Appendix 1 shows all the 35 pairwise options and represents entire game’s payoff 

matrix.   

For each of the 35 games (or rows in Appendix 1), respondents were presented with these 

two options and asked to select either Option A or Option B. For example, for series 1 (i.e., row 

1) the respondents were presented with the options shown in figure 2.1. The figure shows the 

 

 

23 

 

row 1 of series 1 of the lottery was presented to the individuals. Lottery A offers a 30% chance 

of receiving Rs.40 and 70% chance of receiving Rs.10, whereas lottery B offers a 10% chance of 

receiving Rs.68 and 90% chance of receiving Rs.5. This decision to select Option A or Option B 

was repeated for each of the 35 rows in Table 3.  

In all three series (series 1, series 2 and series 3), option A is always less risky compared 

to Option B as suggested by the payoffs and the probabilities. For both series 1 and 2, the 

expected value of lottery A does not change but as we proceed down, the expected value of 

lottery B keeps monotonically increasing and eventually exceeds the expected value of lottery A 

(Table 3). We are interested in finding the “switching row” - row when the individual switches 

from a less risky option A to a more risky option B, for series 1 (rows 1 to 14), series 2 (rows 15-

28), and series 3 (rows 29-35). As we go down the rows, the expected value of Option B keeps 

increasing and therefore a more risk averse individual would choose lottery A for a greater 

number of iterations before shifting to B, as compared to a risk neutral or risk loving individual.  

Following TCN’s procedure, the subject is assumed to be rational, therefore he is allowed 

to switch from lottery A to lottery B only once in each series. The enforcement of monotonic 

switching is a contentious issue with arguments on both sides. It is only rational to assume that if 

a subject chooses a risky option over a less risky option at lower expected value, he/she will 

choose a risky option over a less risky option when the payoff of the risky option increases 

resulting in no switching back. However, it is possible that enforcement causes subjects to 

choose the less risky options for more iterations before switching to the risky option as they 

cannot come back to less risky option after switching. We understand that there is no perfect 

solution to this and we ended up forcing a monotonic switching as it has worked fine with 

subjects of similar educational background in Tanaka et al. (2010), Liu (2013), and Ward and 

 

 

24 

 

Singh (2015). Furthermore, we had a smaller sample size and we were concerned about not 

being able to interpret inconsistent observation from the analysis in case many participants 

switched back to option A after once switching to option B. Also, the option of never switching 

is also available for each of the series. For example, a subject is free to choose lottery A for all 

14 or 7 questions in any/all of the series or he can choose lottery B for all the rows for any/all of 

the series.  

Individuals were told that one of the 35 rounds will be randomly chosen ex post and the 

lottery chosen will be played for actual cash. With this information, the individuals were asked to 

choose between option A and option B for all the 35 rows. Once they completed their selection, 

there was a random draw of 35 numbered plastic chips to decide which game was played for real 

money. Once the game number (1 to 35) was selected, the next step was to select a random 

number between 1 to 10.  For this the TCN method was followed by putting 10 numbered 

wooden chips (each numbered 1 to 10) in an opaque bag and then asking the farmer to draw one 

chip out of it to complete the randomization. For example, if the subject draws plastic chip 

number 1, and he has chosen lottery B for row 1, and a wooden chip number 7 is randomly 

drawn, he would earn Rs.5. However, if he chose lottery A for the same row, and number 7 is 

drawn, then he would earn Rs.10. 

Switching points in each of the three series in Table 3 are used for identifying the 

underlying behavioral parameters. The estimates of risk aversion coefficient (σ - that determines 

curvature of the utility function in the positive domain) and non-linear probability weighting 

measure (α) are simultaneously determined by the switching rounds in series 1 and series 2. 

These two series are carefully designed so that the pair of switching rounds from the two series 

can be used to identify the range for both σ and α, that are consistent with PT.  

 

 

25 

 

Series 3 has both positive and negative payoffs. It has seven choice scenarios, each of 

which comprise of two lotteries like earlier series. In each of the lottery there is a positive and a 

negative payout. The payouts vary across rows and are specified in a way that enables estimation 

of a range of possible loss aversion coefficient for each respondent. 

The loss aversion parameter λ is determined by the switching point in series 3 and series 

1. Notice that λ cannot be uniquely determined from switching point in series 3 alone. Payoffs in 

series 3 are designed to make sure that λ takes similar values across different levels of σ.  In 

calculating the λ, the probability weighting measure α drops out as the probability of getting 

positive or negative payoff are equivalent in each round (p=0.5, q=0.5) and therefore the payoffs 

in series 3 must only correspond to different values of σ.  

Since it would be unethical and impossible to have participating farmers pay from their 

own pocket, in case they lose money in the lottery, Rs.21 was given to each of the participating 

farmer at the beginning of the game. This was the maximum amount a subject can lose in the 

worst-case scenario. This also gave the farmer an ownership over the Rs. 21 and could better 

elicit the loss aversion behavior as now, it was his endowment that he was betting on.  We find 

average of σ as 0.64 suggesting farmers in the sample in general are risk averse. The average of α 

= 0.70 implying farmers seems to overvalue smaller probabilities of high impact gains/losses. 

The average of λ = 3.13, indicating a kink in the values function around 0, with steeper declines 

in prospect value in the loss compared to the inclines under gains. All three risk parameter 

average values are close to values found in other studies done by Liu in China and Ward in 

Eastern India.  

 

 

 

26 

 

2.4.3 Estimation of Parameters  

 

For any participant who switches at row N, we can conclude that he prefers lottery A over B till 

row N-1 and at row N he prefers lottery B over lottery A. So, we can get two sets of inequalities 

from this switching point. Using a combination of switching points from series 1 and series 2, 

yields a range of α and σ that satisfy this pair of inequalities.   

For example, suppose someone switches from lottery A to lottery B in row 7th in series 1. This 

means he preferred lottery A to lottery B till row 6th and he prefers lottery B to lottery A in row 

7th). Then the following inequalities must be satisfied.  

10(cid:26)+exp[−(cid:4)−(cid:30)(cid:31)0.3(cid:6) &(cid:4)40(cid:26)−10(cid:26)(cid:6)     ˃     5(cid:26)+exp [−(cid:4)−(cid:30)(cid:31)0.1(cid:6) &(cid:4)125(cid:26)−5(cid:26)(cid:6)                 (3) 
10(cid:26)+exp[−(cid:4)−(cid:30)(cid:31)0.3(cid:6) &(cid:4)40(cid:26)−10(cid:26)(cid:6)       ˂          5(cid:26)+exp[−(cid:4)−(cid:30)(cid:31)0.1(cid:6) &(cid:4)150(cid:26)−5(cid:26)(cid:6)           (4) 

Solving the above equations, we do not get point estimates but range values for σ and α. 

For example, one of the values for the (σ, α) combinations is 0.36 < σ < 0.44 and 0.36 < α < 

0.44. Following the TCN’s convention we approximate the value of σ and α by taking the 

midpoint of the interval to one decimal place to get the above (σ, α) combinations as (0.4, 0.4). 

Using the midpoint estimates, the (σ, α) combinations that satisfy the above inequalities are 

(0.4,0.4), (0.5,0.5), (0.6.0.6), (0.7,0.7), (0.8,0.8), (0.9,0.9) or (1,1).   

Similarly, if the same person switches from A to B in row 7th in series 2, the following 

inequalities holds true 

30(cid:26)+exp[−(cid:4)−(cid:30)(cid:31)0.9(cid:6) &(cid:4)40(cid:26)−30(cid:26)(cid:6)           ˃            5(cid:26)+exp [−(cid:4)−(cid:30)(cid:31)0.3(cid:6) &(cid:4)65(cid:26)−5(cid:26)(cid:6)          
30(cid:26)+exp[−(cid:4)−(cid:30)(cid:31)0.3(cid:6) &(cid:4)40(cid:26)−30(cid:26)(cid:6)            ˂             5(cid:26)+exp [−(cid:4)−(cid:30)(cid:31)0.3(cid:6) &(cid:4)68(cid:26)−5(cid:26)(cid:6)          

 

 

27 

 

The (σ, α) combinations that satisfy the above inequalities are (0.8, 0.6), (0.7, 0.7), (0.6, 0.8), 

(0.5, 0.9) or (0.4, 1). By intersecting the parameters ranges from series 1 and series 2, we can 

obtain the approximate values of (σ, α) = (0.7, 0.7). Note that λ cannot be uniquely determined 

from switching in series 3. Payoffs in series 3 were designed to make sure that λ takes similar 

values across different levels of σ, which means for each switching point in series 3, we will 

have different values of λ based on the earlier found value of σ for that individual. In this series 

the positive and negative payouts vary from round to round, but are specified in such a way that 

enables estimation of a range of possible loss aversion coefficients. For each of the two lotteries, 

the probabilities of ‘winning’ and ‘losing’ are equivalent(p=q=0.5), so the probability weighting 

function applied to both ‘winning’ and ‘losing’ payouts in each of the lotteries is the same, and 

further drops out when calculating the loss aversion parameter for a particular switching round.  

 

2.5. Conceptual Model  

 

This section presents a simple conceptual framework explaining the role of risk preference 

parameters in technology adoption decision. We have two lotteries; LT, which represents the 

lottery under traditional farming methods and L L , which represents the lottery under laser land 

leveler. As farmer has been using the traditional leveling technology for long, we consider that 

the only source of uncertainty in this case comes from the uncertainty in rainfall growing their 

crops every year as explained in the model below. Suppose there are two states of nature--a good 

monsoon year where the rainfall is adequate and a bad monsoon year where the rainfall is less 

than adequate leading to higher cost of irrigation. Let us consider that the good state of nature 

 

 

28 

 

happens with a probability of q and the bad state of nature occurs with a probability of 1-q.   

Given The above probabilities, the lottery under traditional farming is following.  

 

./=(cid:25)1−0 1ℎ31 (cid:20)(cid:28)(cid:10)(cid:20)(cid:28)4(cid:28)(cid:31)14 3 5(cid:19)(cid:19)6 (cid:5)(cid:28)3(cid:20) (cid:9)(cid:15)1ℎ 3 (cid:10)(cid:20)(cid:19)737(cid:15)(cid:30)(cid:15)1(cid:5) (cid:19)(cid:16) (cid:23)                                (cid:4)5(cid:6)
      1−3 0 1ℎ31 (cid:20)(cid:28)(cid:10)(cid:20)(cid:28)4(cid:28)(cid:31)14 3 736 (cid:5)(cid:28)3(cid:20) (cid:9)(cid:15)1ℎ 3 (cid:10)(cid:20)(cid:19)737(cid:15)(cid:30)(cid:15)1(cid:5) (cid:19)(cid:16) (cid:4)1−(cid:23)(cid:6)             (cid:4)6(cid:6) 

Where 1= normalized revenue per hectare.  

I= cost of all inputs including irrigation cost.  

a= extra input cost as a proportion due to a bad monsoon year, due to increase in irrigation and 

labor cost. a>1. 

For laser land leveler there are two sources of uncertainty. First the uncertainty due to rainfall. 

Second, the uncertainty due to farmer’s belief about the success of the technology. In this case 

we assume farmer believes that the technology is successful with a probability of p and it fails 

with a probability of 1—p.  Given both these sources of uncertainty the farmer decides under the 

following lottery.  

.8=

:::<
9:::;

where 

=−4>0−? 5(cid:19)(cid:19)6 @(cid:19)(cid:31)4(cid:19)(cid:19)(cid:31) 3(cid:31)6  1(cid:28)?ℎ(cid:31)(cid:19)(cid:30)(cid:19)5(cid:5) 
    =−4C0−?  736 @(cid:19)(cid:31)4(cid:19)(cid:19)(cid:31) 3(cid:31)6  1(cid:28)?ℎ(cid:31)(cid:19)(cid:30)(cid:19)5(cid:5) 
 1− 0−?    5(cid:19)(cid:19)6 @(cid:19)(cid:31)4(cid:19)(cid:19)(cid:31) 3(cid:31)6  1(cid:28)?ℎ(cid:31)(cid:19)(cid:30)(cid:19)5(cid:5) 
 1− 3 0−?  736 @(cid:19)(cid:31)4(cid:19)(cid:19)(cid:31) 3(cid:31)6  1(cid:28)?ℎ(cid:31)(cid:19)(cid:30)(cid:19)5(cid:5) 

4A??(cid:28)(cid:28)64 (cid:9)(cid:15)1ℎ (cid:10)(cid:20)(cid:19)737(cid:15)(cid:30)(cid:15)1(cid:5) (cid:19)(cid:16) (cid:10)(cid:23)  −−−−−−−−−− (cid:4)7(cid:6)
4A??(cid:28)(cid:28)64 (cid:9)(cid:15)1ℎ (cid:10)(cid:20)(cid:19)737(cid:15)(cid:30)(cid:15)1(cid:5) (cid:19)(cid:16) (cid:10)(cid:4)1−(cid:23)(cid:6) −−−−−−−   (cid:4)8(cid:6)
(cid:16)3(cid:15)(cid:30)4 (cid:9)(cid:15)1ℎ (cid:10)(cid:20)(cid:19)737(cid:15)(cid:30)(cid:15)1(cid:5) (cid:19)(cid:16) (cid:4)1−(cid:10)(cid:6)(cid:23)−−−−−−−−−− (cid:4)9(cid:6)
(cid:16)3(cid:15)(cid:30)4 (cid:9)(cid:15)1ℎ  (cid:10)(cid:20)(cid:19)737(cid:15)(cid:30)(cid:15)1(cid:5) (cid:19)(cid:16) (cid:4)1−(cid:10)(cid:6)(cid:4)1−(cid:23)(cid:6)−−−−−−−(cid:4)10(cid:6)

 

X= Higher proportion in revenue due to increased yield by the use of LLL. 

 

 

29 

 

4>= saving of irrigation cost as a proportion due to use of LLL in a good monsoon year.  4>< 1.  
4C= saving of irrigation cost as a proportion due to use of LLL in a bad monsoon year.  4C< 1.  
Smaller s means more saving and as this is water saving technology  4C < 4> <1 

c= cost of hiring LLL per hectare.  

 

In conversation with agricultural experts and farmers we got a sense that failure of the 

technology was generally not a source of uncertainty for the farmers in case of LLL. Most 

farmers believe that LLL is a successful technology unlike other less tested technologies. This 

was further substantiated in our data that suggested that only 7 out of 432 farmers ever dis-

adopted LLL after having used it once. The evidence suggests that farmers do not have much 

doubt about the success of LLL. To incorporate this information, we simplified out conceptual 

model allowing only for uncertainty due to rainfall as shown below.  

.8=(cid:25)=−4>0−?  (cid:15)(cid:31) 3 5(cid:19)(cid:19)6 @(cid:19)(cid:31)4(cid:19)(cid:19)(cid:31) (cid:5)(cid:28)3(cid:20) (cid:9)(cid:15)1ℎ 3 (cid:10)(cid:20)(cid:19)737(cid:15)(cid:30)(cid:15)1(cid:5) (cid:19)(cid:16) (cid:23)       −−−−(cid:4)11(cid:6)
=−4C30−?  (cid:15)(cid:31) 3 736 @(cid:19)(cid:31)4(cid:19)(cid:19)(cid:31) (cid:5)(cid:28)3(cid:20) (cid:9)(cid:15)1ℎ 3 (cid:10)(cid:20)(cid:19)737(cid:15)(cid:30)(cid:15)1(cid:5) (cid:19)(cid:16) 1−(cid:23)     −−−(cid:4)12(cid:6) 

In this case, we consider input cost as a function of labor wages, because most pumps here are 

electricity based and electricity for farmers in the state of Punjab is free while it is heavily 

subsidized in Haryana. So, the extra cost of irrigation come from extra labor cost farmers need 

for irrigation. However, this model can be used as a generic case, where irrigation cost is a 

function of multiple other factors.  

Now we plug the lotteries in the functional form of the utility function (as shown above) is as 

follows 

 

 

30 

 

The utility function for traditional lottery is D(cid:4).(cid:6)/ 

D(cid:4).(cid:6)/=(cid:9)(cid:4)(cid:23)(cid:6)  (cid:3)(cid:4)1− 0(cid:6)+(cid:9)(cid:4)1−(cid:23)(cid:6)  (cid:3)(cid:4)1−3 0(cid:6)=exp[− (cid:4)−(cid:30)(cid:31)(cid:23)(cid:6) &(cid:4)1− 0(cid:6)(cid:4)(cid:26)(cid:6)−       

(cid:27) exp[− (cid:4)−ln (cid:4)1−(cid:23)(cid:6)(cid:6) &(cid:4)1−3 0(cid:6)(cid:4)(cid:26)(cid:6) 

 

and utility function for the lottery under laser land leveler D(cid:4).(cid:6)8 
D(cid:4).(cid:6)8=exp[− (cid:4)−(cid:30)(cid:31)(cid:23)(cid:6) &(cid:4)=−4>0−? (cid:6)−(cid:27) exp[− (cid:4)−ln(cid:4)1−(cid:23)(cid:6)(cid:6) & (cid:4)=−4>0−?(cid:6)(cid:4)(cid:26)(cid:6) 

We define the probability of adoption as a function of difference in the two utilities  

As Prob (cid:4).(cid:6)8 =f ( D(cid:4).(cid:6)8−D(cid:4).(cid:6)/ (cid:6) 

Plugging in the functional form of both the utility functions we get  

Pr (cid:4).(cid:6)8= exp[− (cid:4)−(cid:30)(cid:31)(cid:23)(cid:6) &(cid:4)=−4>0−? (cid:6)−(cid:27) exp[− (cid:4)−ln(cid:4)1−(cid:23)(cid:6)(cid:6) & (cid:4)=−4>0−?(cid:6)(cid:4)(cid:26)(cid:6)−  

exp[− (cid:4)−(cid:30)(cid:31)(cid:23)(cid:6) &(cid:4)1− 0(cid:6)(cid:4)(cid:26)(cid:6)+ (cid:27)exp[− (cid:4)−ln (cid:4)1−(cid:23)(cid:6)(cid:6) &(cid:4)1−3 0(cid:6)(cid:4)(cid:26)(cid:6) 
Next, we define Pr (cid:4).(cid:6)8 = F and take derivate of the function wrt risk aversion 

coefficient, loss aversion coefficient and probability weighting measure to know how does 

different risk measures affect adoption.   

Plugging the values of c, a, f(w), p and q from empirical data and the average values of risk 

have following predictions from this model: 

version coefficient (I), loss aversion coefficient ((cid:27)) and probability weighting measure (J) we 
KLK(cid:26) = -ve   i.e. More risk averse farmers are more likely to adopt earlier than risk neutral or risk 

loving farmers. This falls in line with the risk-reducing idea associated with these climate smart 

 

 

31 

 

higher risk aversion means higher likelihood of early technology adoption. 

technologies; a lower I  means higher risk aversion which causes higher adoption. Alternatively, 
 KLK  = -ve i.e. Farmers who overvalues smaller probabilities are more likely to adopt.  A smaller 
J means the farmer overvalues smaller probabilities of big gains/losses. So a farmer who 

overvalues smaller probabilities is more likely to adopt faster than farmers who does not.  

We test these theoretical findings with empirical data in the later sections to see if they hold true 

in the case of LLL technology adoption in the study area. 

 

2.6. Econometric Framework and Results  

 

2.6.1 Econometric Framework  

 

In this section, the main decision variable of interest is the time to adoption. Since, we have the 

retrospective data on the year of release (i.e., 2001), year when the farmers became aware about 

the technology, and the year farmers adopted it, duration model provides a suitable framework 

for modeling adoption probabilities (Kiefer, 1988). Most studies using duration analysis to model 

time to adoption, use the release of the technology as the base year and assume that there is no 

heterogeneity in information dissemination of the technology. Estimating such single duration of 

adoption without accounting for diffusion duration raises econometric issues such as endogeneity 

(Ahsanuzzaman and Maredia, 2018). A farmer who has endogenously better access to 

information, say who is leader of a group, is expected to be aware of the technology earlier than 

 

 

32 

 

other farmers. This, however, does not necessarily indicate that the same farmer is more likely to 

adopt the technology earlier than the other farmers.  

Suppose a technology is released in the year 2010 and a farmer gets to know about it in 

the same year (2010) and waits for 6 years to finally adopt it in 2016. Compare this farmer to 

another farmer who gets to know about the technology in 2014 and waits only 2 years to adopt it 

in 2016. Suppose we conduct our survey in the year 2016 and find both of them to have adopted 

the technology, we would model both of them as same leading to endogenous estimates. While 

estimating adoption duration, the omission of diffusion time might be due to unavailability of the 

information about the diffusion duration. However, we have data on both the duration of 

diffusion (time it took for farmer to know about LLL) and adoption duration (time it took to 

adopt once the farmer got to know about it). We model the later duration as this is the period 

when the farmer is actually exposed to the technology.  

Let t be the time elapsed from the time of first exposure to Laser Land leveler adoption, 

=M(t) be a vector of relevant explanatory variables, and β be a vector of coefficients. Denoting the 
cumulative density function as NM(cid:4)1|=M,β(cid:6)=R(cid:20)(cid:19)7(cid:4)S≤1|=M,β(cid:6) and the density function as f i (t | 
=M , β ).  The hazard function indicating the probability of adopting LLL at period t conditional on 
not having adopted it till time (t-1) is defined by   ℎ(cid:15) (cid:4)1 |=M ,β (cid:6)= (cid:16) (cid:15) (cid:4)1 | =M ,β (cid:6)/[1−
NM(cid:4)1|=M,β(cid:6)&.      

The general form of proportional hazard function   is  

ℎM(cid:4)1|=M(cid:4)1(cid:6),V(cid:6)= ℎW(cid:4)1(cid:6)exp {=MY(cid:4)1(cid:6)V 

Where ℎW  is the baseline hazard and X’s are the explanatory variables. I use a Weibull baseline 

hazard specification to test if the hazard is time dependent. For more intuitively interpretable 

 

 

33 

 

results, the above hazard rate can be parameterized into what is known as the Accelerated Failure 

Time (AFT) model, a simple transformation of the proportional hazards model, which is what we 

AFT model can be expressed as   

use. Under AFT we take exponential of the V to interpret the coefficients. In vector form, the 
log(cid:4)1(cid:6)=VY=+I\ 
variables, and V is the vector of corresponding coefficients. In the case of a Weibull hazard 
function, \ is the error term that follows an extreme value distribution.   

where t is a non-negative random variable denoting adoption time, X is the vector of explanatory 

 

 

2.6.2 Results  

 

Table 4 presents the Probit results of binary adoption decision. As risk aversion goes down in 

risk aversion coefficient. A negative coefficient for risk aversion coefficient suggests higher 

adoption for farmers who are risk averse. Similarly, a negative coefficient for probability 

weighting measure suggest that farmers who overvalue smaller probabilities of high impact 

events are more likely to adopt this technology. However, as most experts believe that this 

technology will be widely adopted in this region and therefore adoption as a binary outcome 

variable does not provide rich information about the role of risk preferences in adoption. If 

majority of the farmers are eventually going to adopt this technology, time to adoption seems to 

provide more interesting information about the role risk preferences play in adoption decision. 

We focus primarily on the time to adoption in the result section.  

 

 

34 

 

The results of the estimates of the duration model for “exposure time” of LLL adoption is 

presented in Table 5. We model the exposure time as the time farmers took after they became 

aware about the technology. State and village fixed effects are controlled for in all specifications. 

The main characteristic of interest is individual risk preference parameters. In the existing 

literature, most studies do not have any control for individual risk preference; therefore, the 

regression result in Column 1 excludes the risk preferences parameter as a comparison.  Column 

2 shows results on time to adoption once the farmer is aware about the technology, and column 3 

show results on time to adoption from release of the technology. It is apparent that the results 

from column 3 underestimate the effect of risk parameters due to missing information on time 

for diffusion. Column 2 results takes into consideration the temporal heterogeneity in diffusion 

and therefore correctly estimates the impact of risk preference on time to adoption.  

To interpret the coefficients, one needs to exponentiate coefficients reported in table 5. 

For example, to interpret the coefficient of σ in Column 2, we need to exponentiate (0.188) = 

1.21. This implies that the risk-averse individual with σ = 0 in the sample is 21% more likely to 

adopt Laser Land Leveler than the risk-neutral individual (σ = 1) at any given time. Similarly, an 

individual who overvalues smaller probability (α=0) is 20% more likely to adopt LLL at any 

given time (Exp[0.185]=1.20). α defines the shape of the probability weighting function and a 

smaller α indicates an individual’s tendency to overweight small probabilities. 

An increase in income from wheat-rice crops as a proportion of total income leads to 

higher probability of adoption. Age at the time of adoption and education both have negative 

coefficient but are not significant. Higher number of plots makes a farmer more likely to adopt 

the technology. This study does not have a detailed social network module like the one in Conley 

and Udry (2010), which can be a concern given that networks might be correlated with risk 

 

 

35 

 

preference and social network is also correlated with technology adoption. We use self-reported 

proxies for network defined as ‘the number of farmers the respondent typically interacts with’ 

and the results clearly show that smaller the social network of the farmer, lower is his probability 

of adoption. We also, control for religious affiliations, which is another major foundation of 

social network. The results show that a Muslim farmer (who is likely to be a member of a 

minority religious group in the study area) is 145% less likely to adopt LLL compared to a Hindu 

farmer.  

Another interesting result is the role of government extension in adoption. If the farmer 

gets the information from government extension his probability of adoption increases by 19% 

(Exp[-.206]=0.81), suggesting the higher trust farmers put in information from public sources 

compared to private sources. Furthermore, the coefficient for “time from release to awareness” is 

negative suggesting if the farmer gets to know about the technology a year later, it is 9% (Exp[-

.0888]=0.91) more likely to reduce his time to adopt. This is expected, because if the farmer 

hears about the technology later, he is likely to get more credible and detailed information about 

the technology as more people around him would have adopted it and their experiences will 

bring him more rich information.  

 

2.6.3 Robustness Check  

 

In this section I estimate the relationship between risk parameters and adoption decision using 

different specification. In column 1 of table 6 I allow standard errors to be correlated at the 

village level and find that risk aversion and non-linear probability weighting to still be 

significant. One might be concerned that I did not consider that some of the farmer might have 

 

 

36 

 

not understood the game even after two rounds of practice games. In our sample 5.6% of the 

farmers always chose either option A or option B for all of their choices in the game. They might 

have chosen so because of their inherent risk preference or because they did not understand the 

game properly. Assuming they did not understand the game, we exclude them from the 

regression and still find very similar results as indicated in column 4 of table 6. Similarly, it is 

possible that some of the farmers did not remember the exact year they heard about LLL or 

misreported it, leading to measurement error. In column 2 of table 6 we define the first time a 

farmer gets to know about a technology as the year when it was first used by someone in the 

village and we get very similar results to our original specification. We found 7 farmers who had 

dis-adopted LLL after using it at least once. At the time of the data collection they identified 

themselves as dis-adopters. The results do not change by excluding these 7 farmers who dis-

adopted LLL as shown in column 3. Credit constraints for adoption decisions are less likely to 

play a role in wealthier farm households, and as a result, the estimate bias should be minimal 

among those households. We restrict the sample to the top one-third wealthier households as 

shown in column 5, and we find that risk aversion is still significant while probability weighting 

does not, suggesting that credit constraint is less likely to play a role in adoption decision.  The 

results from different specifications suggests that σ, λ, and α are not particularly sensitive to 

specification suggesting a robust relationship between risk preference parameters and technology 

adoption.  

 

 

 

 

 

37 

 

2. 6.4 Further Discussion 

 

Concerns on causal identification are typical of any cross-sectional study. In this study we have a 

cross sectional data and we have risk parameters ex post of the adoption decision (Besley and 

Case, 1993). This could be a cause of concern if risk preferences have changed because of 

farmers’ adoption of laser land leveller. The extent to which adoption decisions changes risk 

preferences is beyond the scope of this study and we assume temporal stability of risk 

preferences. Further, lack of any wealth measure prior to adoption may seem problematic, since 

it is often considered a proxy for credit constraint. Lack of credit is associated with bias of the 

parameters of risk preferences. However, as almost all of the farmers (99%) in our sample hired 

this technology and did not buy it makes the credit constraint less of a concern. Cost of hiring 

this technology is equivalent to a couple days of manual labor wages and is quite affordable by 

most farmers in our sample.  

Another limitation of the study stems from duration model’s underlying assumption that 

all the farmers will eventually adopt this technology. Experts and farmer leaders, we have 

consulted compare this technology with the tractor and rotavator that most farmers in these two 

districts have now adopted. Given the pace of adoption and farmers’ response in last few years, it 

is predicted that LLL technology will be also eventually adopted by most farmers. Further, our 

data is cross-sectional that raises the concern that any ex-post measurement of explanatory 

variables could be affected by the adoption decision, and they are therefore endogenous (Besley 

and Case, 1993). Our explanatory variables, however, are unlikely to be endogenous as most of 

them are time invariant. We have data on land sold and bought by farmers in recent past and it 

 

 

38 

 

does not show much sale or purchase suggesting that land size at the time of the survey is a good 

proxy for land at the time a farmer became aware about the technology.  

Another possible limitation of this study could be the absence of uncertainty aversion. Instead of 

risk aversion, uncertainty aversion might play a role in the decision-making process. A 

commonly used definition that distinguishes risk versus Knightian uncertainty is that “risk is 

imperfect knowledge where the probabilities of the possible outcomes are known, and Knightian 

uncertainty exists when these probabilities are unknown” (Hardaker et al., 2004). At the time of 

deciding on the technology, farmers might not have perfect information on the distribution of 

Wheat/Rice yields, thus ambiguity aversion becomes an important factor in decision making. 

Even if one believes that uncertainty aversion ought to play a more prominent role, we find that 

the results are significant across various specifications, and this could suggest that the risk-

aversion measures from the experiment may be a good proxy for uncertainty aversion. 

 

2. 7. Conclusion  

 

Researchers and governments have long promoted the LLL technology as risk reducing. The 

findings of this study confirm this characterization of the technology by showing that once 

becoming aware, farmers who are more risk averse and overvalue smaller probabilities have a 

higher probability of adopting the Laser Land Leveler at any given time. In other words, farmers 

who are more risk averse and who overvalue smaller probabilities adopt this technology sooner 

compared to farmers who are risk neutral/loving and farmers who do not over value smaller 

probabilities.   

 

 

39 

 

Since on average farmers are risk averse, the findings of this study have two important 

implications. First, for a risk reducing technology such as LLL, farmer’s aversion to risk is not a 

limiting variable but an inducing factor in promoting the adoption of this technology. This is 

contrary to the relationship of risk perception and adoption of other types of agricultural 

technologies (Liu 2013). Second, the delay in adoption observed in these two progressive 

districts, is partly explained by the slow rate and speed of diffusion and awareness of this 

technology.  Efforts to promote the diffusion of this and other risk reducing technology using 

public and private extension channels and diverse modes of information delivery should receive 

greater attention to help speed up the adoption of such risk reducing climate smart technologies.  

 

 

 

 

 

 

 

 

 

 

 

 
 

 

 

 

 

 

 

40 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

APPENDICES 

 

41 

 

APPENDIX A: FIGURES 
 

Figure 2. 1: FIRST ROW OF THE RISK GAME 

Option A 

Rupees if you get 
1,2,3  

Rupees if you get 
4,5,6,7,8,9 10 

40 

10 

 

  

  

Option B 

Rupees if you 
get 1  

Rupees if you get 
2,3,4,5,6,7,8,9 10 

68 

5 

  

Q No.  

1 

Series 1 

Q. no 

1 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

42 

 

Figure 2. 2: DISTRIBUTION OF RISK PREFERENCE PARAMETERS 

Curvature of the Value Function

5
1

0
1

5

e
g
a
nt
e
c
r
e
P

0

0

.5

1

1.5

Risk Aversion

Non-linear Probability Weighting Measure

e
g
a
nt
e
c
r
e
P

e
g
a
nt
e
c
r
e
P

0
2

5
1

0
1

5

0

0
2

5
1

0
1

5

0

0

.5

1

1.5

Probability Weighting

Estimated Loss Aversion

0

5

Loss Aversion

10

Source: Authors calculations. 

 

 

 

43 

 

 

 

Figure 2. 3: LOCATION OF TWO STUDY DISTRICTS 

 

Source: Maps of India 

 

 

 

 

 

 

 

 

 

 

 

 

 

44 

 

APPENDIX B: TABLES 
 

Table 2.1: SUMMARY CHARACTERISTICS 

Variables                                                                                                                                                                 Value                           

Risk Aversion  

 
Loss Aversion 

 
Non-Linear Probability Weighting Measure  

 
Age  
 
Education (Years) 

 
Household Poverty Score 

 
Land Owned in Acres 

 
Number of plots   

 
Distance to the nearest Agriculture Extension Centre (KM) 

 
Number of Mobile Phones in the household 

 
Number of Siblings of the head of the household 

 
Total Working Member 

 
Proportion using mobile for info. on Agriculture 

 
Proportion getting advice from Farmer's groups 

 
Self-reported value of total Asset (in 1000 Indian Rupees) 

 
Average distance to Plot(KM)  

 
Self-rated risk attitude (1= Most adventurous, 4= least Adventurous) 

 
Avg Household Size 

 
Avg time from release of LLL to farmers knowing about it(years) 
 
Avg time for farmers to adopt LLL once they know about it (years) 
 

Average time from release of LLL to adoption (years) 
 
Observations  

Note: Standard Deviations are in parenthesis 

 

 

 

 

45 

0.64 
(0.30) 
3.13 
(2.92) 
0.70 
(0.22) 
42.24 
(12.9) 
9.43 
(3.46) 
67.37 
(11.25) 
9.32 
(10.49) 
1.16 
(0.37) 
10.88 
(9.22) 
2.48 
(0.92) 
3.2 
(2.25) 
2.44 
(1.68) 
0.28 
(0.45) 
0.55 
(0.5) 
558.2 
(422) 
1.54 
(2.62) 
2.22 
(0.88) 
5.95 
(2.18) 
9.93 
(1.44) 
3.32 
(1.68) 
13.27 
(1.87) 
432 

 

Table 2.2: PROSPECT THEORY VS EXPECTED UTILITY THEORY 

Prospect Theory Holds True  

Variable 

Description  

Mean 

Std. 
Dev. 

Probability weighting function parameter 

0.704*** 

0.222 

Curvature of the prospect value function (risk aversion) 

0.643*** 

0.307 

Measure of loss aversion 

α 

σ 

λ 

  (Null for σ =1, λ =1, α=1) 
Source: Authors calculations  

 

 

 

 

 

 

3.127*** 

2.918 

  

  

 

 

46 

 

Table 2.3 : PAYOFF MATRIX FOR THE EXPERIMENTAL GAME  

Series 1 

Lottery A 

  

Lottery B 

1  30% chance of winning Rs.40  and 70% chance of winning Rs.10 

10% chance of winning Rs.68 and 90% chance of winning Rs.5 

2  30% chance of winning Rs.40  and 70% chance of winning Rs.11 

10% chance of winning Rs.75 and 90% chance of winning Rs.5 

3  30% chance of winning Rs.40  and 70% chance of winning Rs.12 

10% chance of winning Rs.83 and 90% chance of winning Rs.5 

4  30% chance of winning Rs.40  and 70% chance of winning Rs.13 

10% chance of winning Rs.93 and 90% chance of winning Rs.5 

5  30% chance of winning Rs.40  and 70% chance of winning Rs.14 

10% chance of winning Rs.106 and 90% chance of winning Rs.5 

6  30% chance of winning Rs.40  and 70% chance of winning Rs.15 

10% chance of winning Rs.125 and 90% chance of winning Rs.5 

7  30% chance of winning Rs.40  and 70% chance of winning Rs.16 

10% chance of winning Rs.150 and 90% chance of winning Rs.5 

8  30% chance of winning Rs.40  and 70% chance of winning Rs.17 

10% chance of winning Rs.185 and 90% chance of winning Rs.5 

9  30% chance of winning Rs.40  and 70% chance of winning Rs.18 

10% chance of winning Rs.220 and 90% chance of winning Rs.5 

10  30% chance of winning Rs.40  and 70% chance of winning Rs.19 

10% chance of winning Rs.300 and 90% chance of winning Rs.5 

11  30% chance of winning Rs.40  and 70% chance of winning Rs.20 

10% chance of winning Rs.400 and 90% chance of winning Rs.5 

12  30% chance of winning Rs.40  and 70% chance of winning Rs.21 

10% chance of winning Rs.600 and 90% chance of winning Rs.5 

13  30% chance of winning Rs.40  and 70% chance of winning Rs.22 

10% chance of winning Rs.1000 and 90% chance of winning Rs.5 

14  30% chance of winning Rs.40  and 70% chance of winning Rs.23 

10% chance of winning Rs.1700 and 90% chance of winning Rs.5 

Series 2 

Lottery A 

  

Lottery B 

1  90% chance of winning Rs.40  and  10% chance of winning Rs.30 

70% chance of winning Rs.54 and 30% chance of winning Rs.5 

2  90% chance of winning Rs.40  and  10% chance of winning Rs.30 

70% chance of winning Rs.56 and 30% chance of winning Rs.5 

3  90% chance of winning Rs.40  and  10% chance of winning Rs.30 

70% chance of winning Rs.58 and 30% chance of winning Rs.5 

4  90% chance of winning Rs.40  and  10% chance of winning Rs.30 

70% chance of winning Rs.60 and 30% chance of winning Rs.5 

5  90% chance of winning Rs.40  and  10% chance of winning Rs.30 

70% chance of winning Rs.62 and 30% chance of winning Rs.5 

6  90% chance of winning Rs.40  and  10% chance of winning Rs.30 

70% chance of winning Rs.65 and 30% chance of winning Rs.5 

7  90% chance of winning Rs.40  and  10% chance of winning Rs.30 

70% chance of winning Rs.68 and 30% chance of winning Rs.5 

8  90% chance of winning Rs.40  and  10% chance of winning Rs.30 

70% chance of winning Rs.72 and 30% chance of winning Rs.5 

9  90% chance of winning Rs.40  and  10% chance of winning Rs.30 

70% chance of winning Rs.77 and 30% chance of winning Rs.5 

10  90% chance of winning Rs.40  and  10% chance of winning Rs.30 

70% chance of winning Rs.83 and 30% chance of winning Rs.5 

11  90% chance of winning Rs.40  and  10% chance of winning Rs.30 

70% chance of winning Rs.90 and 30% chance of winning Rs.5 

12  90% chance of winning Rs.40  and  10% chance of winning Rs.30 

70% chance of winning Rs.100 and 30% chance of winning Rs.5 

 

 

 

47 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 2.3 (cont’d) 
 

13  90% chance of winning Rs.40  and  10% chance of winning Rs.30 

70% chance of winning Rs.110 and 30% chance of winning Rs.5 

14  90% chance of winning Rs.40  and  10% chance of winning Rs.31 

70% chance of winning Rs.130 and 30% chance of winning Rs.6 

Series 3 

Lottery A 

1  50% chance of winning Rs.25  and  50% chance of losing Rs.4 

2  50% chance of winning Rs.4  and  50% chance of losing Rs.4 

3  50% chance of winning Rs.1  and  50% chance of losing Rs.4 

4  50% chance of winning Rs.1  and  50% chance of losing Rs.4 

5  50% chance of winning Rs.1  and  50% chance of losing Rs.8 

6  50% chance of winning Rs.1  and  50% chance of losing Rs.8 

7  50% chance of winning Rs.1  and  50% chance of losing Rs.8 

  

 

 

 

 

 

 
  

Lottery B 

50% chance of winning Rs.30 and 50% chance of losing Rs.21 

50% chance of winning Rs.30 and 50% chance of losing Rs.21 

50% chance of winning Rs.30 and 50% chance of losing Rs.21 

50% chance of winning Rs.30 and 50% chance of losing Rs.16 

50% chance of winning Rs.30 and 50% chance of losing Rs.16 

50% chance of winning Rs.30 and 50% chance of losing Rs.14 

50% chance of winning Rs.30 and 50% chance of losing Rs.11 

 

 

48 

 

Table 2.4: PROBIT MODEL FOR ADOPTION 

VARIABLES 

 
Risk Aversion Coefficient 

 
Probability Weighting 

 
Loss Aversion 

 
Age 

 
Education(Years) 

 
Percentage of Income from Wheat-Rice 

 
Household Poverty Score 

 
Land Owned Acres 

 
Number of Wheat-Rice Plots 

 
Distance to Nearest Agricultural Ext. Centre 

 
Number of siblings 

 
Percentage Area Sandy-Loamy 

 
Total Working member 

 
Whether use mobile for info on Agri. 

 
Whether gets info from farmer's group 

 
Religion                                                   1. Muslim 

(Base Hindu) 

                                                                  2. Sikh 

 
                                                                 3. Others 

 
Source of Info- Government Extension 

(Base- Private) 

Average Distance to Plot (KM) 

 
Interaction with no. of Farmers   1. (75-100) 

Base (100+) 

 

 

49 

Coefficient 

 
-0.399* 

(0.216) 

-1.103*** 

(0.280) 

0.006 

(0.023) 

0.012 

(0.010) 

0.042 

(0.031) 

0.016*** 

(0.005) 

-0.003 

(0.009) 

0.003 

(0.018) 

0.576** 

(0.275) 

0.024 

(0.015) 

-0.109** 

(0.054) 

0.149 

(0.264) 

0.005 

(0.028) 

0.330 

(0.201) 

0.271 

(0.235) 

0.390 

(0.350) 

-0.090 

(0.279) 

0.648 

(0.482) 

0.667*** 

(0.242) 

0.000 

(0.033) 

-4.739*** 

(0.480) 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  Table 2.4 (cont’d)                                                                 
 
                                                                    2. (50-75) 

 
                                                                    3. (30-50) 

 
                                                                    4. (20-30) 

 
                                                                    5. (10-20) 

 
                                                                    6. (0-10) 

 
Time from release to Awareness (years) 

 
Whether adopts early(self-reported) 

 
State - Haryana 

(Base Punjab) 

Constant 

 
Observations 

-4.204*** 

(0.491) 

-5.196*** 

(0.535) 

-5.483*** 

(0.623) 

-6.137*** 

(0.664) 

-11.975*** 

(0.654) 

-0.191** 

(0.097) 

0.235 

(0.192) 

-0.525 

(0.570) 

-387.736** 

(195.219) 

432 

Standard errors clustered at Village level> Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 

 

50 

 

Table 2.5: WEIBULL MODEL FOR DURATION OF TIME TO ADOPTION 

VARIABLES 

 
Risk Aversion Coefficient 

 
Probability Weighting 

 
Loss Aversion 

 
Age 

 
Education(Years) 

 
Percentage of Income from Wheat-Rice 

 
Household Poverty Score 

 
Land Owned Acres 

 
Number of Wheat-Rice Plots 

 
Distance to Nearest Agricultural Ext. Centre 

 
Number of siblings 

 
Percentage Area Sandy-Loamy 

 
Total Working member 

 
Whether use mobile for info on Agri. 

 
Whether gets info from farmer's group 

 
Religion                                                   1. Muslim 

(Base Hindu) 

                                                                  2. Sikh 

 
                                                                 3. Others 

 
Source of Info- Government Extension 

(Base- Private) 

Average Distance to Plot (KM) 

 
Interaction with no. of Farmers   1. (75-100) 

Base (100+) 

(1) 

  

 

 

 

 

 

 
-0.00159 

(0.00305) 

-0.0132 

(0.0109) 

-0.00403* 

(0.00216) 

-0.00371 

(0.00314) 

-0.00175 

(0.00239) 

-0.151** 

(0.0679) 

-0.0110 

(0.00895) 

-0.0127 

(0.0188) 

0.0991 

(0.1000) 

-0.00423 

(0.0154) 

-0.105 

(0.0764) 

-0.0585 

(0.116) 

0.821** 

(0.339) 

-0.0951 

(0.187) 

-0.233 

(0.364) 

-0.236** 

(0.117) 

-0.0230** 

(0.00963) 

0.155 

(0.245) 

                                                                    2. (50-75) 

-0.0631 

 
                                                                    3. (30-50) 

 

 

(0.291) 

0.305 

(0.291) 

 

51 

(2) 

  

0.188* 

(0.111) 

0.185* 

(0.107) 

-0.005 

(0.009) 

-0.00132 

(0.00456) 

-0.0132 

(0.0138) 

-0.00441* 

(0.00226) 

-0.00402 

(0.00361) 

-0.00166 

(0.00361) 

-0.150* 

(0.0823) 

-0.00847 

(0.00659) 

-0.0159 

(0.0179) 

0.101 

(0.106) 

-0.00647 

(0.0164) 

-0.122 

(0.0922) 

-0.0209 

(0.128) 

0.896*** 

(0.333) 

-0.140 

(0.160) 

-0.310 

(0.194) 

-0.206* 

(0.118) 

-0.0262** 

(0.0113) 

0.185 

(0.237) 

-0.0188 

(0.269) 

0.363 

(0.291) 

(3) 

  

0.045* 

(0.025) 

0.042* 

(0.023) 

-0.002 

(0.003) 

-0.000127 

(0.000731) 

-0.00255 

(0.00249) 

-0.00104** 

(0.000509) 

-0.000909 

(0.000719) 

-0.000343 

(0.000619) 

-0.0349** 

(0.0147) 

-0.00267 

(0.00217) 

-0.00454 

(0.00433) 

0.0296 

(0.0239) 

-0.000118 

(0.00357) 

-0.0267 

(0.0184) 

0.00178 

(0.0271) 

0.125* 

(0.0754) 

-0.0336 

(0.0422) 

-0.0765 

(0.0819) 

-0.0590** 

(0.0269) 

-0.00526*** 

(0.00156) 

0.0455 

(0.0418) 

-0.0135 

(0.0565) 

0.0749 

(0.0577) 

 

 

 

 

 

 

 

 

Table 2.5 (cont’d)                                                                 
 
                                                                   4. (20-30) 

0.359 

 
                                                                    5. (10-20) 

 
                                                                    6. (0-10) 

 
Time from release to Awareness (years) 

 
Whether adopts early(self-reported) 

 
State - Haryana 

(Base Punjab) 

Constant 

 

 
Observations 

(0.290) 

0.544* 

(0.308) 

5.800*** 

(0.493) 

-0.0888** 

(0.0380) 

-0.0969 

(0.121) 

-1.336*** 

(0.387) 

3.649*** 

(0.742) 

 
432 

0.453* 

(0.273) 

0.621** 

(0.313) 

5.845*** 

(0.721) 

-0.0854* 

(0.0495) 

-0.0453 

(0.155) 

-1.376*** 

(0.307) 

3.323*** 

(0.871) 

 
432 

Note: All regressions include Village fixed effects. Standard errors clustered at Village level  

*** p<0.01, ** p<0.05, * p<0.1 

 

 

0.0966 

(0.0589) 

0.141** 

(0.0642) 

1.345*** 

(0.110) 

0.0609*** 

(0.00958) 

-0.0136 

(0.0293) 

-0.281*** 

(0.0895) 

2.305*** 

(0.169) 

 
432 

 

 

 

52 

 

Table 2.6: WEIBULL MODEL FOR DURATION OF TIME TO ADOPTION: ROBUSTNESS 
CHECK 

 
  

VARIABLES 
  
Risk Aversion 
 
Probability Weighting  

 
Loss Aversion 
 
Age 
 
Education 

 

HH Poverty Score 

 

Income from wheat and Rice 
 

Land Owned Acres 
 
Distance to Ag Ext(KM) 

 
Dummy Info from extension 
 
Year First Known 

 
Dummy for Haryana State 
 
Constant 

 

(1) 

(2) 

Village 
  
0.288*** 
(0.105) 
0.233* 
(0.125) 
0.00127 
(0.0115) 
-0.000870 
(0.00356) 
-0.00126 
(0.0119) 

Cluster 
  
0.249*** 
(0.0846) 
0.218** 
(0.107) 
-0.00825 
(0.0101) 
-0.000728 
(0.00461) 
-0.00653 
(0.0138) 

-0.00450 
(0.00356) 

-0.00410* 
(0.00240) 

-0.00425 
(0.00320) 
-0.000746 
(0.00754) 
-0.306*** 
(0.105) 
-0.0790 
(0.0492) 
-1.013** 
(0.397) 
160.8 
(99.06) 

   (4) 

  (3) 
Excluding-
Disadopters  No Switch 
  
0.204** 
(0.0404) 
0.240** 
(0.0140) 
-0.0107 
(0.251) 
0.000352 
(0.908) 
-0.00542 
(0.633) 

  
0.217* 
(0.110) 
0.259*** 
(0.0985) 
-0.0117 
(0.00953) 
-0.000859 
(0.00299) 
-0.00769 
(0.0113) 

-0.00541 
(0.00371) 

-0.00499 
(0.101) 

-0.00480* 
(0.00247) 
-
0.00890** 
(0.00349) 
-0.0128 
(0.0103) 
-0.441*** 
(0.137) 
-0.226*** 
(0.0841) 
0.545 
(0.631) 
455.5*** 
(169.5) 

-0.00412** 
(0.0466) 

-0.00453** 
(0.0309) 
-0.000240 
(0.979) 
-0.319*** 
(0.00472) 
-0.0842** 
(0.0280) 
-0.925** 
(0.0412) 
171.1** 
(0.0266) 

-0.00520* 
(0.00303) 
-
0.00483** 
(0.00208) 

-0.00360* 
(0.00202) 
0.000934 
(0.00990) 
-0.292** 
(0.122) 
-0.0718* 
(0.0377) 
-0.754* 
(0.424) 
145.9* 
(75.94) 

   (5) 
Top one 
third 
  
0.460* 
(0.241) 
-0.131 
(0.234) 
-0.00535 
(0.0214) 
-0.00335 
(0.0035) 
0.00127 
(0.0175) 
-
0.000704 
(0.0122) 

-0.00210 
(0.0023) 

-0.00114 
(0.0036) 
-0.0163 
(0.0115) 
0.0910 
(0.169) 
-0.122 
(0.0779) 
0.523 
(0.765) 
245.9 
(156.1) 

Observations 

432 

432 

425 

408 

158 

Robust standard errors in parentheses. All other variables used in the main regression used, but not reported in this table. *** 
p<0.01, ** p<0.05, * p<0.1 

 

 

 

 

 

 

 

 

 

 

 

 

53 

 

    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

REFERENCES 

 

54 

 

REFERENCES 

 

Abdulai, A., & Huffman, W. (2014). The adoption and impact of soil and water conservation 
technology: An endogenous switching regression application.Land Economics, 90(1), 26-43. 

Aryal, J. P., Sapkota, T. B., Jat, M. L., & Bishnoi, D. K. (2015). On-farm economic and 
environmental impact of zero-tillage wheat: a case of North-West India. Experimental 
Agriculture, 51(1), 1-16. 

Barrett, C. B., Moser, C. M., McHugh, O. V., & Barison, J. (2004). Better technology, better 
plots, or better farmers? Identifying changes in productivity and risk among Malagasy rice 
farmers. American Journal of Agricultural Economics, 86(4), 869-888. 

Barrett, C. B., Moser, C. M., McHugh, O. V., & Barison, J. (2004). Better technology, better 
plots, or better farmers? Identifying changes in productivity and risk among Malagasy rice 
farmers. American Journal of Agricultural Economics, 86(4), 869-888. 

Battalio, Raymond C., John H. Kagel and Komain Jiranyakul (1990): "Testing between 
alternative models of choice under uncertainty: Some initial results," Journal of Risk and 
Uncertainty, 3(1), 25-50. 

Besley, T., & Case, A. (1993). Modeling technology adoption in developing countries. The 
American economic review, 83(2), 396-402. 

Bhalla, G. S., & Alagh, Y. K. (1979). Performance of Indian agriculture. 

Binswanger, H. P. (1980). Attitudes toward risk: Experimental measurement in rural India. 
American journal of agricultural economics, 62(3), 395-407. 

Binswanger, H. P., Ruttan, V. W., Ben-Zion, U., Janvry, A. D., & Evenson, R. E. (1978). 
Induced innovation; technology, institutions, and development. 

Byerlee, D., & De Polanco, E. H. (1986). Farmers' stepwise adoption of technological packages: 
Evidence from the Mexican Altiplano. American Journal of Agricultural Economics, 68(3), 519-
527. 

Cox, J. C., & Sadiraj, V. (2002). Risk aversion and expected-utility theory: coherence for small-
and large. 

 

 

55 

 

Croppenstedt, Andre, Mulat Demeke and Meloria M. Meschi (2003): "Technology Adoption in 
the Presence of Constraints: the Case of Fertilizer Demand in Ethiopia," Review of Development 
Economics, 7(1), 58-70. 

De Janvry, Aain, Gregory Graff, Elisabeth Sadoulet and David Zilberman (2000): 
"Technological Change in Agriculture and Poverty Reduction," Concept paper for the World 
Development Report on Poverty and Development, 2000/2001, University of California at 
Berkeley/World Bank. 

Dobbs, T., & Foster, P. (1972). Incentives to invest in new agricultural inputs in north India. 
Economic Development and Cultural Change, 21(1), 101-117. 

Dohmen, T., Falk, A., Huffman, D., & Sunde, U. (2012). The Intergenerational Transmission of 
Risk and Trust Attitudes. The Review of Economic Studies, 79(2), 645-677 

Duflo, Esther, Michael Kremer and Jonathan Robinson (2008): "How High Are Rates of Return 
to Fertilizer? Evidence from Field Experiments in Kenya," American Economic Review, 
98(2),482-488. 

Epstein, Larry G. (1999): "A Definition of Uncertainty Aversion," The Review of Economic 
Studies, 66(3), 579-608. 

Feder, G. (1980). Farm size, risk aversion and the adoption of new technology under uncertainty. 
Oxford Economic Papers, 263-283. 

Feder, Gershon, Richard E. Just and David Zilberman (1985): "Adoption of Agricultural 
Innovations in Developing Countries: A Survey," Economic Development & Cultural 
Change,33(2), 255. 

Foster, A. D., & Rosenzweig, M. R. (1995). Learning by doing and learning from others: Human 
capital and technical change in agriculture. Journal of political Economy, 1176-1209. 

Frankel, F. R. (1969). India's new strategy of agricultural development: Political costs of agrarian 
modernization. The Journal of Asian Studies, 28(4), 693-710. 

Gafsi, S., & Roe, T. (1979). Adoption of unlike high-yielding wheat varieties in Tunisia. 
Economic Development and Cultural Change, 28(1), 119-133. 

Ghadim, A. K. A., Pannell, D. J., & Burton, M. P. (2005). Risk, uncertainty, and learning in 
adoption of a crop innovation. Agricultural economics, 33(1), 1-9. 

Griliches, Zvi (1957): "Hybrid Corn: An Exploration in the Economics of Technological 
Change," Econometrica, 25(4), 501-522. 

 

 

56 

 

Gupta, R. K., Naresh, R. K., Hobbs, P. R., Jiaguo, Z., & Ladha, J. K. (2003). Sustainability of 
post-green revolution agriculture: the rice–wheat cropping systems of the Indo-Gangetic Plains 
and China. Improving the productivity and sustainability of rice–wheat systems: Issues and 
impacts, (improvingthepro), 1-25. 

Hanjra, M. A., & Qureshi, M. E. (2010). Global water crisis and future food security in an era of 
climate change. Food Policy, 35(5), 365-377. 

Hardaker, J. B. (Ed.). (2004). Coping with risk in agriculture. Cabi. 

Harris, B. (1972). Innovation Adoption in Indian Agriculture-the High Yielding Varieties 
Programme. Modern Asian Studies, 6(1), 71-98. 

Harrison, Glenn and Elisabet Rutström (2009): "Expected utility theory and prospect theory: one 
wedding and a decent funeral," Experimental Economics, 12(2), 133-158. 

Harrison, Glenn W. and John A. List (2004): "Field Experiments," Journal of Economic 
Literature, 42(4), 1009-1055. 

Heckman, J. J. (1979). Sample selection bias as a specification error. Econometrica: Journal of 
the econometric society, 153-161. 

Hicks, W. W., & Johnson, S. R. (1979). Population growth and the adoption of new technology 
in colonial Taiwanese agriculture. The Journal of Development Studies, 15(4), 289-303. 

Hodgdon, L. L., & Singh, H. (1966). Adoption of agricultural practices in Madhya Pradesh. 

Holden, S. T. (2015). Risk Preferences, Shocks and Technology Adoption: Farmers’ Responses 
to Drought Risk (No. 3/15). Centre for Land Tenure Studies, Norwegian University of Life 
Sciences. 

Holt, Charles. A. and Susan K. Laury (2002): "Risk Aversion and Incentive Effects," American 
Economic Review, 92(5), 1644-1655. 

IPCC, 2014: Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and 
III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Core 
Writing Team, R.K. Pachauri and L.A. Meyer (eds.)]. IPCC, Geneva, Switzerland, 151 pp. 

Jat, M. L., Gathala, M. K., Ladha, J. K., Saharawat, Y. S., Jat, A. S., Kumar, V., ... & Gupta, R. 
(2009). Evaluation of precision land leveling and double zero-till systems in the rice–wheat 
rotation: Water use, productivity, profitability and soil physical properties. Soil and Tillage 
Research, 105(1), 112-121. 

 

 

57 

 

Just, Richard E. and David Zilberman (1983): "Stochastic Structure, Farm Size and Technology 
Adoption in Developing Agriculture," Oxford Economic Papers, 35(2), 307-328. 

Kahneman, Daniel and Amos Tversky (1979): "Prospect Theory: An Analysis of Decision under 
Risk," Econometrica, 47(2), 263-291. 

Kaur, B., Singh, S., Garg, B. R., Singh, J. M., & Singh, J. (2012). Enhancing Water Productivity 
through On-farm Resource Conservation Technology in Punjab Agriculture. Agricultural 
Economics Research Review, 25(1). 

Khan, M. H. (1975). The economics of the green revolution in Pakistan. Praeger Publishers. 

Knight, Frank (1921): Risk, uncertainty and profit. Boston, MA: Schaffner & Marx. 

Knight, John, Sharada Weir and Tassew Woldehanna (2003): "The Role of Education in 
Facilitating Risk-Taking and Innovation in Agriculture," Journal of Development Studies, 
39(6),1-22. 

Lindner, R. K., & Gibbs, M. (1990). A test of Bayesian learning from farmer trials of new wheat 
varieties. Australian Journal of Agricultural and Resource Economics, 34(1), 21-38. 

Lipton, M. (1976). Agricultural finance and rural credit in poor countries. World Development, 
4(7), 543-553. 

Liu, E. M. (2013). Time to change what to sow: Risk preferences and technology adoption 
decisions of cotton farmers in China. Review of Economics and Statistics, 95(4), 1386-1403. 

Marra, M. C., & Carlson, G. A. (1990). The decision to double crop: An application of expected 
utility theory using Stein's theorem. American Journal of Agricultural Economics, 72(2), 337-
345. 

Marra, M., Pannell, D. J., & Ghadim, A. A. (2003). The economics of risk, uncertainty and 
learning in the adoption of new agricultural technologies: where are we on the learning curve?. 
Agricultural systems, 75(2-3), 215-234. 

McInish, T. H., Ramaswami, S. N., & Srivastava, R. K. (1993). Do More Risk‐Averse Investors 

Have Lower Net Worth and Income?. Financial Review, 28(1), 91-106. 

Munshi, K. (2004). Social learning in a heterogeneous population: technology diffusion in the 
Indian Green Revolution. Journal of development Economics,73(1), 185-213. 

O'Mara, Gerald (1983). The microeconomics of technique adoption by smallholding Mexican 
farmers. The Book ofCHAC: Programming Studies for Mexican Agriculture. Baltimore, 
Maryland: Johns Hopkins, 250-289. 

 

 

58 

 

Prelec, D. (1998). The probability weighting function. Econometrica, 497-527. 

Rabin, Mattthew (2000): "Risk Aversion and Expected-utility Theory: A Calibration Theorem," 
Econometrica, 68(5), 1281-1292. 

Ram, R. (1980). Role of Education in Production: A Slightly New Approach. The Quarterly 
Journal of Economics, 95(2), 365-373. 

Shapiro, B. I., Brorsen, B. W., & Doster, D. H. (1992). Adoption of double-cropping soybeans 
and wheat. Southern Journal of Agricultural Economics,24(02), 33-40. 

Sidhu, S. S. (1976). The Productive Value of Education in Agricultural Development (No. 
13937). University of Minnesota, Department of Applied Economics. 

Smale, M., Just, R. E., & Leathers, H. D. (1994). Land allocation in HYV adoption models: An 
investigation of alternative explanations. American Journal of Agricultural Economics, 76(3), 
535-546. 

Suri, T. (2011). Selection and comparative advantage in technology adoption. Econometrica, 
79(1), 159-209. 

Tanaka, T., Camerer, C. F., & Nguyen, Q. (2010). Risk and time preferences: linking 
experimental and household survey data from Vietnam. The American Economic Review, 
100(1), 557-571. 

Tversky, Amos. and Daniel Kahneman (1992): "Advances in Prospect Theory: Cumulative 
Representation of Uncertainty," Journal of Risk & Uncertainty, 5(4), 297-323. 

Ward, P. S., & Singh, V. (2015). Using Field Experiments to Elicit Risk and Ambiguity 
Preferences: Behavioral Factors and the Adoption of New Agricultural Technologies in Rural 
India. The Journal of Development Studies,51(6), 707-724. 

Weil, P. M. (1970). The introduction of the ox plow in central Gambia. African food production 
systems: cases and theory, 229-263. 

Welch, F. (1970). Education in production. Journal of Political Economy, 78(1), 35-59. 

Wills, I. R. (1972). Projections of effects of modern inputs on agricultural income and 
employment in a community development block, Uttar Pradesh, India. American Journal of 
Agricultural Economics, 54(3), 452-460. 

 

 

 

 

 

59 

 

CHAPTER 3: EFFECTS OF IN-UTERO RAINFALL SHOCKS ON CHILDREN’S 

HEALTH IN MALAWI 

 
 
3.1. Introduction 
 

In utero gestation period is a very crucial and sensitive time in the life of an individual. There is a 

growing literature documenting how social, emotional and environmental events during the 

gestational period can have long term consequences throughout the adult life (Barker 1998, 

Bhalotra & Venkatramani 2015, Beach et al 2016). These effects can be biological, and/ or 

socioeconomic in nature.  Due to weak institutions, populations within developing countries face 

shocks related to disease outbreak, food and water scarcity and weather anomalies more 

frequently compared to developed industrialized countries. Also, due to limited resources, 

children born in poor families who are exposed to adverse gestational shocks are less likely to 

recover in their later life (Almond & Currie 2011).  

The gestational effects mentioned above are channeled through the mother. The 

emotional, mental and physical health of the mother thus plays a crucial role in determining the 

biological and socioeconomic growth trajectory of the child. For example, a child born to a 

malnourished mother can limit his/her abilities to cope with health issues later in life (Currie and 

Vogl, 2013). It is important to study the effects of such shocks during gestational period on the 

long -term health consequences among children in the context of developing countries, 

especially because if there is a possibility of mitigation of the adverse effect of such shocks, then 

policy interventions could be designed to minimize the health implications of gestational shocks.  

Many developing countries are increasingly facing health challenges due to 

environmental anomalies and shocks, exacerbated by climate change. Rainfall variability is one 

 

 

60 

 

such factor that has been found to have an impact on general health and well-being (WHO 2008). 

Agriculture in most developing countries is dependent on rain. Thus, rainfall has a major 

influence on food and water availability. Poor rainfall can directly affect the harvest and have 

negative health outcomes for populations directly dependent on agriculture. In addition, climate 

change as an increasing concern is further exacerbating rainfall variability and hydrological 

cycles in many regions of the world (Nan, Bao-hui and Chun-kun, 2011).  

This paper studies the impact of the rainfall shocks during the gestational period on the 

health outcomes of children aged 0-5 years in Malawi.  We use nationally representative datasets 

from the World Bank’s Living Standard Measurement Survey (LSMS) for information on socio 

economic profile of the households and child anthropometrics. We further use the daily rainfall 

data for 1981-2017 for each of these households by connecting them to the high-resolution 

climate data from Climate Hazards Group Infrared Precipitation with Stations (CHIRPS) using 

the corresponding GPS coordinates of the households.  We calculate the total rainfall and rainfall 

anomaly compared to the 37-year average, during the 12 months prior to the birth. The 12 

months includes 9 months of gestational period and 3 months prior to that which is considered 

very important for health outcomes of the child in the literature (Kudamatsu et al., 2010).  

Our identification strategy relies on the hypothesis that temporary rainfall deviations 

from historical averages, are uncorrelated with other latent determinants of health during 

gestation. Under this assumption, we are able to identify the causal impact of rainfall variation 

during the in-utero period on outcomes after birth. Our results indicate that negative rainfall 

shocks during the gestational period leads to lower height for age for 0-59 month old children, 

which is a standard measure of chronic stunting. We also find a significant negative effect of 

flood (excessive rain) defined as one standard deviation above the average rainfall, on the height 

 

 

61 

 

for age measure. The results remain significant in the rural areas but not in the urban areas, 

suggesting an agricultural production pathway for the shock.  

Malawi is one of the poorest countries and ranks 170 out of 187 in human development 

index (Human development Report 2016). It has gone through some major droughts and some 

parts of the country have experienced floods in the past dozen years (Ellis & Manda 2012). With 

significant population dependent on agriculture, these weather shocks are likely to have caused 

major negative impacts on the population especially children. This paper is an attempt to 

quantify the effect of such weather shocks during gestational period on the health of the child for 

providing policy suggestions to mitigate the shocks by positive reinforcements.   

The paper is organized as follows. Section 2 covers the literature review.  Section 3 

discusses the data source. Section 4 discusses the possible pathways for the gestational shock to 

impact the future health of the child.  Section 5 describes the general econometric frameworks 

used to test the predictions and describes the empirical results. Section 6 covers robustness 

check.  Section 7 concludes the paper.  

 

3.2. Literature Review  

 

Until the 1960s placenta was regarded as a “perfect filter, protecting the fetus from harmful 

substances in the mother’s body and letting through helpful ones”. (Landro 2010). During the 

baby boom period in the US, women were told that it was fine to smoke and drink during 

pregnancy. Around half of pregnant mothers in the US reported smoking in 1960; something that 

seems incomprehensible nowadays (Aizer, Stroud, and Buka, 2009).  David J Barker, challenged 

 

 

62 

 

this idea by introducing the fetal origin hypothesis (FOH), which underscored the importance of 

in utero period in shaping the future abilities and health trajectories of a baby. The hypothesis 

argued that the intrauterine environment - and nutrition in particular- “programs” the fetus to 

have particular metabolic characteristics, which can lead to future disease (Barker, 1990). Barker 

argued that individuals starved in utero are more likely to suffer from diseases associated with 

obesity, including cardiovascular problems and diabetes.  

The fetal origin hypothesis uncovers many key ideas that are crucial for understanding the 

impact of in utero and early childhood period in one’s life. First, effects of fetal conditions are 

persistent. Second, the effect of the shock during the in utero period can stay latent till many 

years and start kicking in only when the child grows up as an adult. Typically, diabetes and heart 

disease does not emerge as a problem until middle age. Third, the hypothesized effects reflect a 

biological mechanism, akin to fetal programming, possibly through effects of environment on the 

epigenome, which are just beginning to be understood. Epigenome can be considered analogous 

to a switching mechanism that defines which parts of the genome are going to be expressed, and 

which parts are not. The period in utero may be particularly important for setting these switches 

(Petronis, 2010).  

The large magnitude of impact of in utero period on the life of a person suggests that 

Pareto improvements can be made by reallocating resources from later in life to earlier, at a 

household level, society level or as a policy principle (Almond & Currie 2010). This idea struck 

a chord with economists and their interest in this theory has been ever increasing since early 

1990s Even though, fetal origin hypothesis initially explained the link of mother’s nutritional 

shock to chronic adult health conditions, new evidence by economists suggests that other non-

health outcomes like IQ and wages can be linked to prenatal and early life shocks.  Further, FOH 

 

 

63 

 

literature suggest that it may be possible to counter the effect of negative prenatal shocks by 

improving investments in pregnant women and young children. Such investments can be made 

by government and non-government sectors in the form of education programs, nutrition 

interventions, and clinical support to such women. Three interventions that have been shown to 

be effective are nurse home visiting programs, nutritional supplementation for pregnant women, 

and quality early childhood education programs (Currie 2011). 

There has been a growing literature in economics around the correlation between shocks 

and events experienced by mothers during gestational periods and its later life outcomes for 

children and adults (Almond, 2006).  Within these studies, researchers have studied the link 

between in utero or early infant exposure to nutritional shock, diseases, stress, pollutions and 

environmental shocks and later life outcomes such as adult health, forms of disability, mortality 

rate, cognitive ability, or socio-economic status (Almond, Currie, and Duque, 2017). 

In utero nutritional shocks are probably the most studied type of shock with effects found 

on a range of health and non-health outcomes ranging from birth weight to learning abilities. 

Almond and Mazumder (2011) studied the effect of in utero nutritional shock by estimating the 

effect of pregnant mother’s exposure to Ramadan fasting on birth outcomes and adult disability 

and mortality in Michigan, Iraq and Uganda. They found that, this in utero nutritional shock of 

pregnant mother’s fasting led to higher rate of adult disability, higher mortality as adult and 

lower wealth accumulation. Further, Almond, Mazumder and Van Ewikj (2015) found that the 

effect of the same shock led to lower school attainment and lower test scores for children under 

seven in England. Adhvaryu et al. (2016) found that the of introduction of iodine salt in the food 

of pregnant mother, led to higher labor force participation and positive wage earnings among 

adults 25-55 years old decades later in the US. Fitzsimons and Vera-Hernandez (2013) found a 

 

 

64 

 

significant positive effect of breastfeeding on cognitive development among kids in the UK. 

Likewise, positive nutrition-related shocks can also have substantial effects even when relatively 

mild. For example, Linnemayr and Alderman (2011) examine nutritional supplementation for 

pregnant women and 0 to 3 year-old children in Senegal and find that supplementation during 

pregnancy has a significant effect on the weight-for-age of toddlers, but that post-birth 

supplementation had little impact, suggesting the uniqueness of the in utero period.  

Aizer, Stroud and Buka (2016) examined the effect of stress on pregnant mothers 

characterized by cortisol level on the educational outcomes of the children. They found that 

higher cortisol level during the in utero period, leads to lower educational attainment among 

children at 7 years of age in the US. Exposure to pollution during pregnancy can be another 

major determinant of the child health and learning outcomes. Bharadwaj et al. (2014) found that 

fetal exposure to carbon monoxide leads to a lower math and language test scores among fourth 

graders in the UK.  

Some diseases that have only mild effects in adults are known to have devastating effects 

on a developing fetus (e.g., Rubella, Zika), while others, like pandemic influenza killed millions 

and have also been shown to have effects on fetuses in utero. Almond (2005) uses the 1918 

influenza pandemic as a natural experiment for testing the fetal origins hypothesis. The study 

shows that cohorts in utero during the pandemic displayed reduced educational attainment, 

increased rates of physical disability, lower income, lower socioeconomic status, and higher 

transfer payments compared with other birth cohorts. Schwandt (2017) found that pregnant 

mothers who suffer from seasonal influenza have children with lower birth weight in Denmark.  

Further, these children when they grow up as adults, end up earning less and depend more on 

public welfare compared to children of mothers who did not experience this shock. Similarly, 

 

 

65 

 

diseases like malaria and seasonal influenza can have negative long-term effects if a fetus is 

exposed to these during pregnancy (Venkataramani 2012; Currie and Schwandt 2013). 

Various studies have examined the effect of weather shocks during gestational period, 

early childhood, and on the adult health. Factors such as temperature and rainfall have been 

found to have an effect on pregnant women, fetus in utero, and other lifelong outcomes. Some 

studies focus on the time during the effect of a weather shock during infancy on lifelong 

outcomes. In their study, Maccini and Yang (2009) examined the effect of rainfall variation 

around the time of birth on the health, education, and socioeconomic outcomes of Indonesian 

adults born between 1953 and 1974.  The results showed that higher rainfall in early years of life 

had positive effects on the adult outcomes, especially for women. Higher temperatures during 

early childhood are associated with lower adult height (Agüero, 2014). Aguilar and Vicarelli 

(2011) report that the climatic event of El Niño Southern Oscillation (ENSO), showed effects on 

the early childhood development of children even after four to five years from when they were 

first exposed to the shock. Their analysis showed that children exposed to ENSO during their 

early stages of life had test scores in language development, working-memory, and visual-spatial 

thinking abilities that were 11 to 21 percent lower than same aged children not exposed to the 

shock. Another study showed that Zimbabwean children exposed to drought in early life (12-24 

months of age) showed a decrease in annual growth among these children by 1.5 to 2 cm. Even 

after four years from the drought year, these children remained shorter in heights compared to 

other children their age who had not experienced the drought. (Hoddinott and Kinsey 2006). 

Furthermore, Rocha and Soares (2015) showed evidence that water scarcity during 

gestational period influenced health of a baby at birth. They found that negative rainfall during in 

utero gestation period was strongly associated with higher infant mortality, low birth weight, and 

 

 

66 

 

pre-matured births (i.e., shorter gestational periods). Another study by Linnemayr and Alderman 

(2011) suggests that exposure to moderate low-temperature shocks (i.e., cold waves) during the 

first and second trimesters of pregnancy are associated with lower length at birth. Extreme 

weather shocks during pregnancy also have particularly negative effects on the developing fetus. 

Shocks generated by two powerful tropical storms striking Puerto Rico during the late 1920s and 

early 1930s had long-term consequences on the health of individuals who were exposed to the 

storms as infants. Individuals who were exposed to storms as infants were found more likely to 

report a diagnosis of hypertension, high cholesterol, and diabetes (Orlando Sotomayor 2013). 

Moreover, exposure to a hurricane during pregnancy has been associated with increased 

probability of abnormal conditions of the newborn (Robert S. Scholtea, Gerard J. van den Bergb, 

Maarten Lindeboomd, 2015). We situate our study under this category of studies where we find 

the impact of in-utero weather shock on the health outcomes of 0-5 years of children in Malawi.  

 

3.3. Conceptual Framework  
 

3.3.1 Health Production Function  

 

Following Heckman (2007), we present our simplistic model of health production as a two 

period Constant Elasticity of Substitution (CES) function depicted in equation 1. 

 

h = A [α (0̅>+μ>(cid:6)(cid:6)(cid:26)+(cid:4)1−J(cid:6)(cid:4)0C+μC(cid:6)(cid:26)&>/(cid:26)    

 

 

 

 

(1) 

 

 

 

67 

 

Where h represents health or human capital. In our case h would denote the health of 

children under 5. A denotes the productivity, 0̅> and 0C are the parental investments made in 
gestational period and after birth respectively. Let us suppose  0̅> is fixed and what is under 

consideration is the second period investment. α is the weight parameter that each period in 

childhood gets in the production of adult health.  

Following Currie and Almond (2011), we assume that μ> is an exogenous shock in first 
period (gestational period in our case) and μC is an exogenous shock in second period (0-5 years 

in our model). The parameter σ, denotes the extent to which investments in different periods are 

substitutes or complements, often a key question determining the efficacy of interventions. 

We assume that parents make the investment decisions for their child. Investments are 

costly, and are valued insofar as they improve h. Parents maximize their utility, U = U(C, h), 

which is increasing in both their own consumption (C) and their child’s health (h), given their 

budget constraint. 

 

 ‘= (cid:10)ab+(cid:10)c0>+(cid:10)c0C/(cid:4)1+(cid:20)(cid:6)  

 

 

 

 

 

 

 

 

(2) 

in period 2. If the parents do not have a wealth constraint in period two, they are likely to provide 

If the first period shock μ> is negative, then the effect of this shock can be “faded away” 
either by increasing the second period investment 0C or by giving an exogenous positive shock μC 
higher investment 0C in period 2 to mitigate the effects of negative weather shock on their child’s 
shock will persist till later in life. In such cases, a public investment creating positive μC shocks 

health over time.  However, if the parents face budget constraint, the effect of pre-natal weather 

 

68 

 

 

are likely to help counter the effect of gestational shock. We will test this hypothesis in our 

results to see if the children belonging to wealthier households recover from this shock over time 

and does the shock persists among the children from poorer households.   

 

3.3.2 Pathways Through Which Rainfall Can Impact Child’s Health  

 

Low rainfall is a major problem in many parts of Malawi. In this context we discuss two 

channels of how low rainfall might affect the fetus in the gestational period. First, lack of rainfall 

may directly impact households dependent on agriculture through reduced nutrient intake due to 

lower food production and availability of less variety of food (i.e., low dietary diversity). This 

leads to malnutrition and micronutrient deficiency, potentially including deficits of vitamins A, 

B1, B3, C, and iron (WHO, 2012). Second, lack of inadequate rainfall would lead to water 

scarcity in the traditional water resources like ponds, wells, rivers etc. Scarcity of water, 

decreases the use of water for personal hygiene, increase the travel time to collect water (i.e., 

increases physical stress), increases the need to store water in unsanitary conditions, and 

decreases the quality of the traditional sources of water (Mara, 2003; Pond et al., 2011; Sobsey, 

2002). In combination with poor sanitation, inadequate access to water is the leading risk factor 

for diarrheal diseases. Diarrhea is caused mainly by pathogens that are ingested from unsafe 

water, contaminated food, or hands. It is the second most important factor in the global burden of 

diseases (WHO, 2010) and it increases the susceptibility to and the severity of new infections, 

reinforcing a vicious cycle (Fewtrell et al., 2007).  

 

 

69 

 

Pregnant women and their fetuses are particularly vulnerable to the health problems 

associated with water scarcity (Pond et al., 2011). Biological demands for water and nutrients are 

greatly enhanced during pregnancy. Water deprivation may cause low levels of amniotic fluid in 

the later trimesters, leading to fetal malnutrition and malformation. During gestation, body’s 

need for water increases by 7 to 8 liters, which are roughly shared between the maternal and the 

fetal placental compartments (Barron, 1987). Increased basal metabolism and tissue synthesis 

also raise the demand for nutrients. Requirements of proteins, fats, various vitamins (including 

A, B1, B3, and C), iron, iodine, and zinc are increased. Deficient intake of some of these may 

lead to birth defects, low birth weight, obstetric complications, premature birth, and higher 

perinatal mortality (Steegers-Theunissen, 1995). Water deprivation by itself may also lead to 

dehydration–anorexia, resulting in an additional channel of nutrient stress (Ross and Desai, 

2005). In short, health shocks faced by the mother during pregnancy – related to malnutrition and 

dehydration are reflected on the health of the newborn, which increases its susceptibility to 

infections and is further enhanced by the long-term implications of fetal growth trajectory.  

Excessive rainfall is as bad for the crops as low rainfall. Beyond a certain point, rainfall 

not only starts hurting normal life, but also agricultural productivity and people’s livelihoods. 

There are various mechanisms through which excessive rainfall creates a shock during the 

gestational period. It washes away nutrients and, in some cases, entire crop, resulting in lower 

production. Water is a major carrier of pathogens and floods/excessive rainfall creates conditions 

conducive for spreading harmful pathogens in a tropical country like Malawi. Overall, excessive 

rainfall also has a significant impact on the health of pregnant women and children. 

 

 

 

70 

 

3.4.  Data.  

 

3.4.1 Socio-Economic and Anthropometric Data 

 

We use the Integrated Household Survey (IHS)/LSMS datasets for information on the socio-

economic status and individual level demographic data along with the child level anthropometric 

information. IHS is one of the primary instruments implemented by the Government of Malawi 

through the National Statistical Office to monitor and evaluate the changing conditions of 

Malawian households. The IHS data have, among other insights, provided benchmark poverty 

and vulnerability, while also cover a wide range of information on agriculture.  We use the two 

most recent rounds (third and fourth rounds) of IHS for Malawi. The third round was conducted 

in the year 2010 and covered 12,271 households while the fourth round was conducted in the 

year 2016-17, and covered 12,447 households. Both the third and the fourth round samples 

provide district-level representativeness and a reasonable level of precision for key 

socioeconomic and agricultural indicators. This dataset is not a panel dataset, so there is no 

household in the two rounds that are repeated purposefully.  

Using the repeated cross-sectional microdata on height, weight, age, gender, and odema 

status, we construct the z-scores for children under 5 years of age. We have in total of 13,277 

under 5 years old children across both the rounds with 7,159 children in the third round and 

6,118 children in the fourth. We focus on height-for-age z-score (HAZ) as an objective measure 

of child’s growth and a proxy of child’s health endowment. We restrict all our analyses for 

children who have HAZ scores between -5 and 5, to account for measurement error that might 

have been an outcome of incorrectly measuring child’s anthropometrics (WHO 2006). 

 

 

71 

 

 HAZ score uses an international reference, since the growth in height of healthy children 

under 5 years of age from different ethnic backgrounds and different continents is reasonably 

similar. These international standards were developed based on the scientific evidence that 

children born in any region of the world, if given an optimum start in life, have the same 

potential for growth and develop  within the same range of height and weight for age. The WHO 

child growth standards used worldwide, provides a common basis for the analysis of 

anthropometrics data (WHO 2006). As HAZ measures each child's height in relation to the 

distribution of children of the same age and gender across the globe, it provides an objective 

means of comparison and is used as a measure of malnutrition around the world.  

To find the impact of in-utero rainfall shock on child’s HAZ score, we use the recently 

released GPS coordinates for the households to collect the rainfall information from past 

multiple decades. Then we used this rainfall information and linked them to the corresponding 

child based on the location of the household. A major assumption is that the mother would have 

spent most of the time in the 12 months prior to the birth of a child in the same village where 

child resided at the time of the survey. In the following section we explain the creation of the 

rainfall variables.  

Table 1 presents the descriptive statistics of the sample. There are a total of 13277 

children in sample with an average age of 3.09 years. Almost half of the children are female and 

the average household size is 5.46. More importantly, we find that average height for age Z score 

value is -1.2 signifying the adverse state of child health in Malawi. Further we find that 29% 

percent of the under 5 children are stunted according to the WHO definition of WHO (WHO 

Stunting definition: A child with Height for Age score is less than -2). We find that 16% of the 

households reside in urban areas, which leaves majority of the sample in rural areas with average 

 

 

72 

 

distance from a nearest small town to be 35 kilometers. A mere 8% of the households have 

electricity and 33 percent of sample avails the coupons for agricultural input subsidy; a major 

program of the government of Malawi.  

 

3.4.2 Rainfall Data 

 

We use rainfall data from the Climate Hazards Group Infrared Precipitation with Stations 

(CHIRPS) dataset. This data is built on interpolation techniques that uses high resolution 

estimates based on infrared Cold Cloud Duration (CCD) observations and real station weather 

data and therefore is considered more accurate than just CCD data or station data. CHIRPS 

provides daily data at a very granular level at the resolution of 0.05o x 0.05o, which corresponds 

to approximately a 5km X 5km grid. Higher resolution and precision of this data allows for 

quantifying the impacts of location-specific changing precipitation patterns on human life 

(Peterson et al. 2015). 

We use the daily rainfall data from 1981 till 2017 for every single GPS coordinates that 

corresponds to a household to two variables measuring rainfall fluctuation during an individual's 

gestation period. The first variable is defined by the following equation (Rocha & Saores 2015) 

 

 deM/=(cid:30)(cid:31)  (cid:4)∑

/gh/i>>

(cid:20)Mg

(cid:6)− lnμM   

 

 

 

 

 

 

 

(3) 

 

 

 

73 

 

where (cid:20)Mg  indicates the cumulative monthly rainfall in the location of the household i in month t.  
μM is the average total annual rainfall in the location of household i over the 37 years (1981-
2017). T indicates an individual's month of birth. Thus, deM/ is defined as the deviation between 

the natural logarithm of the total rainfall in the 12 months prior to the individual's birth and the 

natural logarithm of the average yearly rainfall in the location of household i. We consider 12 

months prior to birth, instead of 9 months, because evidence suggests that the nutritional status 

of the mother immediately before conception is also important in determining birth outcomes 

(Kudamatsu et al., 2010). Medical studies, for example, identify a correlation between 

birthweight and the pre-pregnancy weight of the mother (Bloomfield et al., 2006). The variable 

deM/ can be approximately interpreted as the percentage deviation from mean rainfall. For 

instance, a value of 0.05 means that rainfall over the 12 months prior to an individual's birth was 

roughly 5% above average. 

The second shock variable is a binary variable designed to capture excessive rainfall 

event, referred as ‘flood’ variable in this study. Malawi has a sub-tropical climate, which is 

relatively dry and strongly seasonal. Almost all of Malawian agriculture depends on rainfall, so 

given the dry weather it is generally assumed that higher rainfall will lead to better agricultural 

production. However, there is a limit to how much rainfall is good; beyond a certain point 

excessive rain can hurt agriculture, sanitation, transport, and the biological ecosystem. Further 

there are various low-lying areas such as Lower Shire Valley and some localities in Salima and 

Karonga that are more vulnerable to floods during excessive rainfall than higher grounds. 

We define an episode of a flood in the following way. Note that flood here does not 

literally mean a flood, but very heavy rainfall that is likely to disrupt the usual life and prove to 

be counter-productive to agriculture and health.  

 

 

74 

 

 

NM/=j1     (cid:15)(cid:16) ∑

/gi>> 
0 (cid:19)1ℎ(cid:28)(cid:20)(cid:9)(cid:15)4(cid:28)

(cid:20)Mg>(cid:11)(cid:20)̅M−(cid:20)Mkl(cid:14)
m    
 

 

 

 

 

 

(4) 

 

Where (cid:20)Mkl is the historical annual standard deviation of rainfall for the location of household i 
(for the period of 1981–2017). (cid:20)̅M is the average historical yearly rainfall in the area that the 
household i resides in. NM/ =1 if the rainfall over the 12 months prior to an individual's birth was 

more than one standard deviation above the historical average the location of the household i. 

Figure 2 presents the annual average rainfall for the entire country from 1981-2017 and the 

average rainfall for the entire period. The figure shows a zigzag pattern with some years’ rainfall 

below 800 mm while in some years the rainfall goes beyond 1200 mm with the average over the 

entire period as 992 mm per year. However, this figure masks the spatial heterogeneity in 

rainfall, which varies from less than 500 mm in some places in some years to more than 2000 

mm in others.  

 

3.5. Empirical Strategy 

 

The sample for this study is composed of cross sectional data, which is representative at the 

district level. We use two specifications for our analyses, one with enumeration area fixed effects 

and another with household fixed effects. In the first specification, we analyze the impact of 

shocks during gestational period on health outcome at the child level with district fixed effects.  

 

 

75 

 

3.5.1 Enumeration Area Fixed Effect:   

 

There are a total of 1539 enumeration areas and they closely approximate the villages and 

therefore controls for village level unobserved heterogeneity.  

 

nMg= VW+V>deM/+VCNM/+ ɳp+qg+=Mg+\Mg                                                                      (cid:4)5(cid:6) 

 

Where nMg is the health outcome of child i in round t (measured by HAZ), deM/ is the rainfall 
child), NM/ is the indicator of excessive rainfall in the same period, which we defined as one 

variable defined in equation 3 (i.e. log-deviation of rainfall in the 12 months prior to birth of the 

standard deviation above average. We keep both rainfall deviation and flood in the same 

equation to capture both the positive and negative effects of rainfall. As the rainfall increases, the 

effects on nMg is likely to be positive until a certain point, where the effect of flood kicks in and 
therefore keeping the flood dummy will help us capture that effect. ɳp  is the enumeration area 
fixed effect and qg is the round fixed effect. =Mg  includes other control variables at the child 
\Mg is the idiosyncratic error term.  

level, like gender, age and household level variables like education level of household head etc. 

 

The main concern in this specification is the possibility of confounding omitted factors 

correlated both with rainfall and health outcome between the age of 0-5 years. This is clearly the 

case in the cross-section data, since places with harsher climate are also likely to have worse 

socioeconomic conditions. Districts are common across both the rounds and data is 

representative at the district level. Thus we take enumeration area (approximates village) fixed 

 

 

76 

 

effect to control for any effect associated with climatic or socio-economic conditions typical of 

an area. Round fixed-effects, in turn, capture aggregate shocks impacting the entire Malawi and 

secular trends in health outcomes. 

Our identification relies on the assumption that a temporary rainfall deviation from 

historical average or excessive rainfall (i.e., the ‘shock’ variables) – are uncorrelated with any 

latent determinant of health during the first few years of a child. Under this assumption, we are 

able to identify the causal impact of rainfall shocks on early life outcomes. It is difficult to think 

of plausible stories of endogeneity or omitted factors when considering this type of transitory 

variation in rainfall, conditional on all our independent variables. 

A potential problem in our analysis is selection bias due to death of the child during birth 

or between 0-5 years.  Shocks that might have caused the death of the child in fetus, at birth or 

after birth, and before the data collection are not part of our analysis and thus could lead to some 

selection bias. This type of caveat is recurrently mentioned in the birth weight literature (Currie, 

2009 ). Further, these shocks could reduce women's fertility due to health or behavioral 

responses, which would imply that the population of surviving newborns might be different from 

what it would otherwise have been. 

 

3.5.2 The Household Fixed Effect Model 

 

  

In order to control for unobserved heterogeneity at the household level, we use the next 

specification with household fixed effect. Note that we do not have a panel data, so we have to 

 

 

77 

 

restrict this analysis to only those households who have at least two children below the age of 

five.  

nMrg= VW+V>deM/+VCNM/+brg +=Mrg+\Mrg                                                                            (cid:4)6(cid:6) 

 

 

Where nMrg is the health outcome of child i from household y in round t. brg  is the 
household level fixed effect of household y in round t. =Mrg are control variables at the child 

level.  

 

3.6. Results 

 

Tables 2 and 3 present results from the two specifications. In table 2 we present the results from 

equation (5), with enumeration level fixed effects. Column 1 presents the result of all the 

children and shows that there is a positive and statistically significant correlation between 

rainfall deviation and HAZ score. The coefficient suggests that 15% increase (or one standard 

deviation increase in log rainfall-deviation) in rainfall leads to 0.365 increase in height for age 

score, which means better rainfall means taller children for a given age. Coefficient of flood 

dummy is not significant, suggesting no clear relationship between excessive rainfall during 

gestation period and its impact on child’s HAZ.   The data also suggests that HAZ scores are 

better for girls compared to boys, which means that girls under five years of age in Malawi are 

closer to the global standards of height for age than Malawian boys under five. Age of the child 

 

 

78 

 

seems to have a non-linear significant relationship with the HAZ scores; HAZ scores are 

decreasing in age and the decline is at a diminishing rate. This means, as the age increases, the 

average Malawian child is slowly moving away from the global average of height for age 

metrics.   

We further find that children who participated in the under-five government clinic seem 

to have lower height for age as expected. It is likely that only those children visited the clinics 

who were not healthy and therefore have a lower HAZ score. Participation in government’s 

under five nutrition program does not significantly affect HAZ scores. Household head’s age or 

gender are not significant, however, household heads that have completed post graduate are more 

likely to have positive effect on their children’s health, probably due to more information or 

through the higher income channel. Children covered in the second round (Year 2016) have 

significantly higher HAZ scores compared to children in the first round of data in 2010. This 

may be a sign of a developing country like Malawi experiencing higher incomes and better 

healthcare over time. We have also controlled for the birth year and birth month. As close to 

85% of our sample and approximately same percentage of total population of Malawi lives in 

rural areas, in column 2 we restrict our analyses to focus only on rural households. All the 

coefficients have similar magnitude and significance in column 2 as it was in column 1 as rural 

sample contributes heavily to the total sample.  

Columns 3 and 4 test our conceptual framework of whether the effect of such shock 

persists for all or do wealthier parents provide enough positive reinforcements to their children to 

counter the effect of negative weather shocks in the gestation period. As we do not have a clear 

estimate of income of the households, we proxy their income by construction material for their 

houses. To keep it simple, we categorized households based on their type of construction of 

 

 

79 

 

dwelling. We define poor households as those whose houses are made of non-permanent material 

like, mud, thatched roof etc., and we define non-poor as those whose houses are made of more 

permanent structures like concrete etc. Again we restrict our results only to rural areas. In 

column 3 we present the results of children who belong to households that are considered poor 

and live in rural areas. We find that for poor households the magnitude of the effect of weather 

shock on HAZ score is higher than the overall sample. This suggests less coping mechanism for 

the income constrained households to provide positive reinforcements to their children post birth 

to counter the effect of gestational weather shock.   

As the age of the child increases she gets exposed to positive or negative reinforcements, 

based on the income constraints of her household. If the parents are poor, they will not be able to 

provide nutritious food and adequate healthcare to counter the effect of the gestational weather 

shock, exacerbating the negative effect of the shock as time passes by. On the other hand, the 

effect of the negative shock will likely fade away if her parents do not have income constraint 

and are able to provide her the necessary requirements for better health. We test this idea by 

estimating the effect of gestational weather shock on children up to 3 years old and children who 

are older than 3.  

In column 5 and 6, we present the results of children up to three years old. Column 5 

shows the results for rural poor households suggesting that rainfall deviation is significant and 

higher in magnitude for HAZ scores. Same is true of children belonging to not poor households 

as shown in column 6. The magnitudes are very comparable suggesting that in the initial years, 

the effect of gestational weather shock persists among both rich and poor.  

In column 7 and 8 we present the results for children aged above 3 years (3-5 years). 

Column 7 represents children from poor households and we find the effect of gestational weather 

 

 

80 

 

shock still persisting. On the other hand, we have the results of not poor households in column 8 

and we do not find any significant effect of the shock on these children. This suggests that in-

utero shocks seem to persist over time when these in no positive support provided to counter it 

but seems to fade away in the presence of positive reinforcement over time.  

To control for unobserved heterogeneity at the household level, we present in table 3 

results with household level fixed effect. The results are similar to table 2, with rainfall deviation 

positively and significantly affecting HAZ scores, but with a higher magnitude. Interestingly, in 

this specification, we find negative significant effect of excessive rainfall (the variable flood is 

actually +1 SD deviation above normal rainfall).  This is expected, as excessive rainfall can 

cause multiple disruptions in agricultural production and increase the chances of infection to 

negatively affect the health of pregnant women causing a shock to the in-utero baby. The results 

in this specification also holds true across poor and not poor households.  

 

3.7. Conclusion  

 

This paper presents evidence of negative relationship between lower rainfall or excessive rainfall 

during gestational period and the health of the child between the age of 0-5 years. We identify 

the shocks as rainfall fluctuation in the gestational period as compared to the average annual 

rainfall. We also test the hypothesis that whether such effects fade out with time. We found that 

effects of these shocks stay till some age, but after that children who have access to better 

resources are likely to cope up better and the effects fade away. However, for children who 

belong to income constrained households, these effects persist until at least 5 years of age and 

 

 

81 

 

probably much later. The evidence provided by this study is suggestive of the need for public 

policy interventions targeted to expectant mothers during periods of extreme weather shocks, and 

continued support to mother and child after birth, especially for children born in income 

constrained households in the event of rainfall shocks. Short run minimization of impacts during 

episodes of adverse shocks should target maternal health.  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

82 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

APPENDICES 

 

83 

 

APPENDIX A: FIGURES 
 

Figure 3.1: MALAWI PROVINCES AND DISTRICTS MAP 

 

Source: Google Maps  

 

 

 

 

 

 

84 

 

Figure 3.2: ANNUAL AND AVERAGE RAINFALL FOR ALL OF MALAWI 

 

Annual and Average Rainfall in Malawi

m
m
n

 

i
 
l
l

a
f
n
a
R

i

1400

1200

1000

800

600

400

200

0

1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015 2017

Year

Annual Rainfall

Average Rainfall

 

Source : Authors Compilation 

 

 

 

 

 

 

 

 

 

85 

 

APPENDIX B: TABLES 
 

Table 3. 1: DESCRIPTIVE STATISTICS 

Variable 

Obs 

Mean 

Std. Dev. 

Height for Age Z-score 

Weight for Age Z-score 

Weight for Height Z-score 

13,277 

13,277 

-1.20 

-0.53 

13,277 

0.20 

Percentage of children stunted 

13,277 

29 

Age of child in years  

13,277 

3.19 

1.44 

1.10 

1.32 

0.46 

1.30 

Age of HH Head (years) 

13,274 

36.19 

11.35 

Percentage of Male Children  

13,277 

49.06 

Percentage of Male Headed Households 

Household Size 

% HH received coupons for government sponsored Input subsidy 
program 

% Poor Household (As defined by their type of house) 

% HH With electricity 

% Household in Urban Areas 

13,274 

13,277 

13,277 

13,277 

13,277 

13,277 

0.81 

5.46 

33 

45 

8 

16 

Average Distance to nearest road metal (KM) 

13,277 

9.38 

Average Distance to Population Centre (20000+ people) 

13,277 

35.80 

Average Distance to  Agricultural Development and Marketing 
Corporation Centre 

Rainfall log-deviation in the past 12 months 

13,277 

13277 

8.46 

-.03 

.499 

0.39 

1.99 

0.47 

0.50 

0.27 

0.36 

11.36 

22.83 

7.07 

0.15 

 

 

 

 

 

86 

 

Table 3.2: EFFECTS OF IN-UTERO RAINFALL SHOCKS ON CHILD HEALTH: ENUMERATION AREA/VILLAGE FIXED 
EFFECTS 

(3) 

(4) 

(5) 

(6) 

(7) 

(8) 

All Rural 

Rural-Poor 

Rural-Not 
Poor 

Rural-Poor-upto 
3 Years 

Rural-Not 
Poor-upto 3 
Years 

Rural-
Poor-4 
Years and 
Above 
  
0.748* 
(0.409) 
-0.099 
(0.195) 
0.053 
(0.089) 
-0.051 
(0.162) 

 
 

0.156 
(0.192) 

-0.163 
(0.113) 
-0.006 
(0.005) 
0.152 
(0.128) 
-0.044 
(0.153) 
-0.021 
(0.174) 
0.137 
(0.328) 
-1.077 
(1.261) 

Rural-Not 
Poor-4 Years 
and Above 

  
-0.207 
(0.446) 
0.043 
(0.188) 
0.107 
(0.092) 
-0.072 
(0.151) 

 
 

-0.165 
(0.161) 

-0.226** 
(0.113) 
0.005 
(0.004) 
-0.173 
(0.115) 
0.101 
(0.144) 
0.029 
(0.162) 
0.149 
(0.171) 
0.218 
(0.312) 

  
1.240** 
(0.621) 
0.156 
(0.262) 
0.212** 
(0.097) 
-0.381 
(0.463) 
0.025 
(0.108) 

-0.010 
(0.184) 

-0.201 
(0.160) 
-0.002 
(0.005) 
0.088 
(0.128) 
-0.014 
(0.158) 
-0.082 
(0.183) 
0.044 
(0.214) 
0.280 
(0.293) 

(2) 

 

(1) 

 

Total 
Sample 

  
0.365** 
(0.157) 
-0.072 
(0.070) 
0.168*** 
(0.033) 
-0.776*** 
(0.098) 
0.107*** 
(0.014) 

  

 

VARIABLES 
  
Rainfall Deviation 

 
Flood 

 
Gender of the Child(Female) 

 
Age of the child in Years 

 
Age Squared 

 
If the child participate in a nutrition 
program 

 
If the child participate in an under five 
clinic 

 
Household Head Age 

 
Household Head Gender (Female) 

 
Household Head's Education = 2, PSLC 
(Base=Illiterate) 
  JCE 

 
 MSCE 

 
Non-Univ Diploma 

 

 

  
0.317** 
(0.160) 
-0.051 
(0.073) 
0.153*** 
(0.034) 
-0.770*** 
(0.104) 
0.096*** 
(0.015) 

  
0.747*** 
(0.233) 
-0.205* 
(0.114) 
0.161*** 
(0.050) 
-0.790*** 
(0.158) 
0.101*** 
(0.023) 

  
0.095 
(0.267) 
-0.036 
(0.113) 
0.158*** 
(0.056) 
-0.692*** 
(0.159) 
0.080*** 
(0.023) 

  
1.130** 
(0.530) 
-0.372 
(0.241) 
0.301*** 
(0.086) 
-1.535*** 
(0.400) 
0.301*** 
(0.097) 

-0.000 
(0.075) 

-0.002 
(0.072) 

-0.070 
(0.101) 

-0.079 
(0.114) 

-0.029 
(0.158) 

-0.210*** 
(0.046) 
-0.001 
(0.002) 
0.015 
(0.051) 
0.002 
(0.058) 
0.079 
(0.060) 
0.081 
(0.075) 
0.208 
(0.168) 

-0.199*** 
(0.050) 
-0.001 
(0.002) 
0.030 
(0.051) 
0.021 
(0.060) 
0.022 
(0.069) 
0.091 
(0.083) 
0.228 
(0.189) 

-0.126* 
(0.071) 
-0.004 
(0.003) 
0.087 
(0.074) 
-0.023 
(0.101) 
-0.054 
(0.120) 
0.140 
(0.173) 
-1.015*** 
(0.353) 

-0.062 
(0.146) 
-0.001 
(0.006) 
0.013 
(0.113) 
-0.092 
(0.165) 
-0.030 
(0.188) 
0.114 
(0.301) 
-1.184*** 
(0.367) 

-0.195** 
(0.077) 
0.001 
(0.003) 
0.021 
(0.070) 
0.059 
(0.086) 
-0.026 
(0.101) 
0.138 
(0.118) 
0.229 
(0.193) 

 

87 

0.217 
(0.198) 

0.853* 
(0.509) 
0.020* 
(0.012) 
-0.019 
(0.046) 

0.010 
(0.021) 
-0.013 
(0.018) 

1.580*** 
(0.197) 
0.023* 
(0.013) 
0.006 
(0.047) 

0.007 
(0.021) 
-0.012 
(0.019) 

Table 3.2 (cont’d) 
 
 
 
Postgrad degree 

 
Household Size  

 
Credit Availability to the Household  

 
HH Distance in (KMs) to Nearest 
ADMARC Outlet 

 
HH Distance in (KMs) to Nearest Road 

 
HH Distance in (KMs) to Nearest 
Population Center with +20,000 

 
If the household received Coupons  
from govt for agricultural Input 

 
Round 2 (Year=2016) 

 
Birthyear = 2006 
(Base Birthyear=2005) 
Birthyear = 2007 

 
Birthyear = 2008 

 
Birthyear = 2009 

 
Birthyear = 2010 

 
Birthyear = 2011 

 
Birthyear = 2012 

 
Birthyear = 2013 

 

 

0.242 
(0.336) 

-0.352** 
(0.147) 

0.331 
(0.411) 

 
 

0.167 
(0.716) 

0.903*** 
(0.347) 

0.331 
(0.460) 

1.867*** 
(0.166) 
0.031* 
(0.019) 
-0.046 
(0.082) 

-0.011 
(0.030) 
-0.021 
(0.025) 

-0.012 
(0.018) 

-0.010 
(0.065) 
0.860* 
(0.478) 
-0.071 
(0.176) 
-0.079 
(0.215) 
-0.245 
(0.268) 
0.065 
(0.307) 
-0.478 
(0.397) 
-0.343 
(0.389) 
-0.742** 
(0.329) 
-0.489* 

1.925*** 
(0.330) 
0.023 
(0.017) 
0.025 
(0.073) 

-0.019 
(0.032) 
-0.031 
(0.030) 

-0.048** 
(0.020) 

0.037 
(0.068) 
3.642*** 
(0.640) 
-0.179 
(0.199) 
-0.311 
(0.233) 
-0.325 
(0.293) 
-0.469 
(0.344) 
-0.453 
(0.443) 
-3.278*** 
(0.419) 
-3.198*** 
(0.338) 
-3.190*** 

 

88 

 
 
0.015 
(0.032) 
-0.074 
(0.123) 

-0.015 
(0.043) 
-0.011 
(0.046) 

-0.002 
(0.032) 

0.026 
(0.102) 
1.260 
(0.845) 

 
 
 
 
0.100 
(0.201) 
0.526* 
(0.310) 
-0.166 
(0.408) 

 
 
 
 
-1.018*** 

1.786*** 
(0.472) 
0.016 
(0.032) 
0.037 
(0.132) 

-0.016 
(0.054) 
-0.031 
(0.053) 

-0.028 
(0.032) 

0.140 
(0.114) 
5.431*** 
(0.925) 

 
 
 
 
0.291 
(0.262) 
0.528 
(0.390) 
0.415 
(0.506) 

 
 
 
 
-3.689*** 

1.647*** 
(0.300) 
0.057* 
(0.031) 
-0.153 
(0.136) 

0.016 
(0.050) 
-0.058 
(0.047) 

-0.028 
(0.049) 

-0.015 
(0.113) 
-0.348 
(1.094) 
-0.082 
(0.245) 
-0.242 
(0.327) 
-1.642* 
(0.908) 

 
 
 
 
0.904* 
(0.468) 
0.247 
(0.339) 
0.445* 

 
 
0.002 
(0.026) 
-0.030 
(0.110) 

-0.012 
(0.051) 
-0.047 
(0.040) 

-0.071** 
(0.032) 

-0.017 
(0.104) 
-0.767 
(0.998) 
-0.109 
(0.266) 
-0.231 
(0.352) 
0.383 
(1.050) 

 
 
 
 
0.497 
(0.435) 
0.493 
(0.329) 
0.340 

-0.028** 
(0.012) 

-0.030** 
(0.012) 

0.025 
(0.043) 
3.455*** 
(0.387) 
-0.082 
(0.108) 
-0.003 
(0.143) 
-0.060 
(0.177) 
0.101 
(0.204) 
-0.037 
(0.292) 
-3.087*** 
(0.255) 
-3.184*** 
(0.210) 
-3.064*** 

0.043 
(0.043) 
3.160*** 
(0.385) 
-0.123 
(0.122) 
-0.144 
(0.148) 
-0.239 
(0.184) 
-0.149 
(0.211) 
-0.397 
(0.270) 
-2.785*** 
(0.258) 
-2.962*** 
(0.210) 
-2.850*** 

Table 3.2 (cont’d) 
 
 
 
Birthyear = 2014 

 
Birthyear = 2015 

 
Birthyear = 2016 

 
Birthmonth =February 
(Base Birthmonth= January) 
Birthmonth = March 

 
Birthmonth = April 

 
Birthmonth = May 

 
Birthmonth = June 

 
Birthmonth = July 

 
Birthmonth = August 

 
Birthmonth = September 

 
Birthmonth = October 

 
Birthmonth = November 

 
Birthmonth = December 

 
Constant 

 

 

 
Observations 
R-squared 
Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1.  All regressions have enumeration level fixed effect and errors are clustered at enumeration 
level.  

(0.172) 
-2.861*** 
(0.139) 
-2.792*** 
(0.101) 
-2.476*** 
(0.092) 
-0.014 
(0.074) 
-0.034 
(0.070) 
-0.027 
(0.076) 
0.044 
(0.080) 
0.141* 
(0.084) 
-0.045 
(0.080) 
0.005 
(0.075) 
0.045 
(0.075) 
0.045 
(0.084) 
0.123 
(0.083) 
0.260*** 
(0.083) 
0.443 
(0.503) 
13,269 
0.263 

(0.170) 
-2.737*** 
(0.143) 
-2.764*** 
(0.102) 
-2.487*** 
(0.097) 
-0.028 
(0.076) 
-0.047 
(0.073) 
-0.034 
(0.076) 
0.042 
(0.087) 
0.084 
(0.083) 
-0.050 
(0.085) 
0.006 
(0.079) 
0.090 
(0.080) 
0.055 
(0.089) 
0.098 
(0.086) 
0.248*** 
(0.085) 
0.882 
(0.544) 
11,203 
0.262 

(0.270) 
-0.324 
(0.224) 
-0.321* 
(0.166) 

 
 
0.009 
(0.115) 
-0.011 
(0.115) 
-0.034 
(0.120) 
0.121 
(0.118) 
0.077 
(0.120) 
-0.103 
(0.118) 
0.082 
(0.122) 
0.223* 
(0.123) 
0.026 
(0.135) 
0.176 
(0.124) 
0.252** 
(0.127) 
0.533 
(0.687) 
5,632 
0.358 

(0.385) 
-0.554* 
(0.300) 
-0.409* 
(0.212) 

 
 
-0.086 
(0.193) 
-0.139 
(0.195) 
-0.102 
(0.201) 
0.021 
(0.198) 
-0.007 
(0.198) 
-0.211 
(0.193) 
-0.096 
(0.196) 
0.152 
(0.193) 
-0.079 
(0.203) 
0.143 
(0.191) 
0.365* 
(0.209) 
2.576 
(15.538) 
3,288 
0.468 

(0.275) 
-3.178*** 
(0.232) 
-3.231*** 
(0.169) 
-2.998*** 
(0.195) 
0.004 
(0.125) 
-0.095 
(0.112) 
-0.115 
(0.112) 
-0.041 
(0.134) 
0.102 
(0.134) 
0.075 
(0.139) 
-0.094 
(0.125) 
-0.082 
(0.123) 
0.009 
(0.128) 
0.018 
(0.135) 
0.141 
(0.129) 
1.938* 
(1.062) 
5,571 
0.361 

 

89 

(0.248) 

(0.269) 

(0.597) 
-3.402*** 
(0.479) 
-3.309*** 
(0.392) 
-2.818*** 
(0.421) 
0.059 
(0.240) 
0.176 
(0.224) 
0.048 
(0.222) 
0.064 
(0.248) 
0.317 
(0.263) 
0.429 
(0.270) 
0.445* 
(0.248) 
0.190 
(0.245) 
0.203 
(0.233) 
0.173 
(0.257) 
0.618*** 
(0.229) 
-0.462 
(1.634) 
3,016 
0.492 

 
 
 
 
 
 
0.122 
(0.201) 
0.027 
(0.195) 
-0.164 
(0.207) 
0.111 
(0.216) 
0.068 
(0.218) 
-0.250 
(0.201) 
-0.063 
(0.237) 
0.230 
(0.259) 
-0.072 
(0.248) 
0.117 
(0.229) 
0.096 
(0.230) 
0.454 
(1.888) 
2,344 
0.506 

 
 
 
 
 
 
-0.077 
(0.194) 
0.014 
(0.218) 
-0.152 
(0.181) 
0.086 
(0.211) 
0.054 
(0.216) 
-0.077 
(0.203) 
-0.174 
(0.205) 
-0.130 
(0.203) 
0.102 
(0.206) 
0.161 
(0.207) 
-0.247 
(0.217) 
2.285 
(1.670) 
2,555 
0.542 

 

Table 3. 3: EFFECTS OF IN-UTERO RAINFALL SHOCKS ON CHILD HEALTH: HOUSEHOLD FIXED EFFECTS 

  

VARIABLES 

  
Rainfall Deviation 

 
Flood 

 
Gender of the Child (Female) 

 
Age of the child in Years 

 
Age Squared 

 
If the child participated in a nutrition program 

 
If the child participated in an under-five clinic 

 
Birthyear = 2006 

(Base Birthyear=2005) 

Birthyear = 2007 

 
Birthyear = 2008 

 
Birthyear = 2009 

 
Birthyear = 2010 

 
Birthyear = 2011 

 
Birthyear = 2012 

 

 

(1) 

(2) 

(3) 

(4) 

(5) 

All Sample  Poor 

Non-Poor 

Rural 

Rural- Poor 

  

0.587* 

(0.342) 

  

0.903* 

(0.474) 

-0.341** 

-0.320 

(0.149) 

0.117 

(0.073) 

(0.224) 

0.039 

(0.097) 

  

0.237 

(0.452) 

-0.334 

(0.220) 

0.165 

(0.107) 

  

0.528 

(0.357) 

  

1.006** 

(0.444) 

-0.353** 

-0.413* 

(0.154) 

0.101 

(0.078) 

(0.218) 

0.052 

(0.098) 

(6) 
Rural- Not 
Poor 

  

-0.128 

(0.524) 

-0.247 

(0.241) 

0.125 

(0.115) 

-0.775*** 

-1.115*** 

-0.435 

-0.874*** 

-1.170*** 

-0.532* 

(0.194) 

(0.289) 

(0.269) 

(0.205) 

(0.288) 

(0.306) 

0.078*** 

0.115*** 

0.040 

0.083*** 

0.113*** 

0.049 

(0.029) 

-0.208 

(0.185) 

(0.044) 

-0.201 

(0.284) 

(0.041) 

-0.233 

(0.266) 

(0.031) 

-0.225 

(0.191) 

(0.044) 

-0.229 

(0.282) 

-0.290** 

-0.229 

-0.361** 

-0.275** 

-0.242 

(0.120) 

-0.333 

(0.311) 

-0.534 

(0.388) 

(0.169) 

-0.371 

(0.310) 

-0.743 

(0.462) 

(0.163) 

-0.234 

(0.539) 

-0.288 

(0.665) 

(0.126) 

-0.295 

(0.239) 

-0.574 

(0.363) 

(0.167) 

-0.419 

(0.327) 

-0.866* 

(0.482) 

(0.047) 

-0.219 

(0.265) 

-0.295 

(0.185) 

-0.101 

(0.366) 

-0.196 

(0.557) 

-1.001** 

-1.445** 

-0.495 

-1.152*** 

-1.658*** 

-0.490 

(0.452) 

-0.684 

(0.516) 

-0.930 

(0.650) 

(0.599) 

-1.012 

(0.735) 

-1.711* 

(0.917) 

(0.727) 

-0.288 

(0.826) 

-0.164 

(1.003) 

(0.432) 

-0.885* 

(0.503) 

-1.242* 

(0.646) 

(0.609) 

-1.308* 

(0.742) 

(0.608) 

-0.362 

(0.706) 

-2.012** 

-0.359 

(0.927) 

(0.917) 

-3.371*** 

1.055 

-3.546*** 

-3.330*** 

1.388 

-3.496*** 

(0.473) 

(0.900) 

(0.718) 

(0.472) 

(0.884) 

(0.698) 

-3.619*** 

0.529 

-3.539*** 

-3.661*** 

0.853 

-3.556*** 

(0.412) 

(0.740) 

(0.664) 

(0.406) 

(0.737) 

(0.613) 

 

90 

Table 3.3 (cont’d) 
 
 
Birthyear = 2013 

 
Birthyear = 2014 

 
Birthyear = 2015 

 
Birthyear = 2016 

 
Birthmonth =February 

(Base Birthmonth = January) 

Birthmonth = March 

 
Birthmonth = April 

 
Birthmonth = May 

 
Birthmonth = June 

 
Birthmonth = July 

 
Birthmonth = August 

 
Birthmonth = September 

 
Birthmonth = October 

 
Birthmonth = November 

 
Birthmonth = December 

 
Constant 

 

 

-3.413*** 

0.616 

-3.256*** 

-3.475*** 

0.905 

-3.249*** 

(0.383) 

(0.605) 

(0.667) 

(0.366) 

(0.589) 

(0.575) 

-3.723*** 

0.448 

-3.601*** 

-3.907*** 

0.581 

-3.733*** 

(0.480) 

(0.461) 

(0.790) 

(0.477) 

(0.455) 

(0.714) 

-3.604*** 

0.412 

-3.419*** 

-3.862*** 

0.502 

-3.662*** 

(0.529) 

(0.380) 

(0.859) 

(0.534) 

(0.364) 

(0.811) 

-3.669*** 

 

-3.091*** 

-4.034*** 

-3.387*** 

(0.931) 

0.230 

(0.243) 

0.139 

(0.254) 

-0.053 

(0.249) 

0.037 

(0.275) 

0.225 

(0.316) 

0.368 

(0.316) 

0.210 

(0.252) 

0.191 

(0.289) 

0.390 

(0.305) 

0.672* 

(0.359) 

0.607* 

(0.324) 

1.573 

2.587*** 

1.385* 

(0.623) 

0.107 

(0.165) 

-0.086 

(0.182) 

0.053 

(0.190) 

0.056 

(0.189) 

0.077 

(0.210) 

0.092 

(0.214) 

0.164 

(0.187) 

0.276 

(0.202) 

0.155 

(0.191) 

0.421* 

(0.230) 

0.286 

(0.204) 

 

 
0.099 

(0.233) 

-0.213 

(0.249) 

0.174 

(0.261) 

0.187 

(0.266) 

0.015 

(0.295) 

-0.099 

(0.285) 

0.256 

(0.283) 

0.364 

(0.291) 

0.015 

(0.245) 

0.332 

(0.299) 

0.049 

(0.267) 

(0.975) 

0.120 

(0.235) 

0.029 

(0.245) 

-0.139 

(0.234) 

0.019 

(0.264) 

0.128 

(0.278) 

0.352 

(0.293) 

0.111 

(0.237) 

0.074 

(0.274) 

0.175 

(0.301) 

0.622** 

(0.316) 

0.516* 

(0.295) 

1.389 

(0.619) 

0.047 

(0.160) 

-0.148 

(0.176) 

0.004 

(0.182) 

0.026 

(0.179) 

0.049 

(0.194) 

0.087 

(0.202) 

0.116 

(0.180) 

0.191 

(0.191) 

0.093 

(0.184) 

0.411* 

(0.217) 

0.282 

(0.192) 

 
0.070 

(0.230) 

-0.235 

(0.245) 

0.200 

(0.262) 

0.156 

(0.259) 

0.041 

(0.290) 

-0.092 

(0.281) 

0.296 

(0.285) 

0.346 

(0.285) 

0.060 

(0.241) 

0.309 

(0.295) 

0.103 

(0.260) 

2.216*** 

1.149 

 

91 

 

 

 

Table 3.3 (cont’d) 
 

 
Observations 

R-squared 

 

 

 

 

 

 

(0.708) 

(0.743) 

(1.123) 

(0.729) 

(0.735) 

(1.109) 

4,867 

0.714 

2,337 

0.719 

2,530 

0.722 

4,230 

0.713 

2,241 

0.723 

1,989 

0.714 

Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1.  All regressions have household level fixed effect and errors are clustered at enumeration 
area level.  

 

92 

 

   
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

REFERENCES 

 
 

93 

 

REFERENCES 

 

Adhvaryu, A., & Nyshadham, A. (2016). Endowments at birth and parents' investments in 
children. The Economic Journal, 126(593), 781-820. 

Aguilar, A., & Vicarelli, M. (2011). El Nino and Mexican children: medium-term effects of 
early-life weather shocks on cognitive and health outcomes. Cambridge, United States: Harvard 
University, Department of Economics. Manuscript. 

Aguilar, A., & Vicarelli, M. (2011). El Nino and Mexican children: medium-term effects of 
early-life weather shocks on cognitive and health outcomes. Cambridge, United States: Harvard 
University, Department of Economics. Manuscript.  

Aizer, A., Stroud, L., & Buka, S. (2009). Poverty, Stress and the Intergenerationsal Transmission 
of Human Capital. Unpublished Manuscript. 

Aizer, A., Stroud, L., & Buka, S. (2016). Maternal Stress and Child Outcomes. Journal of 
Human Resources, 51(3). 

Aizer, Anna, and Laura Stroud. 2009. “Education, Medical Knowledge and the Evolution of 
Disparities in Health.” NBER Working Paper 15840. 

Almond, D., & Currie, J. (2011). Killing me softly: The fetal origins hypothesis. Journal of 
economic perspectives, 25(3), 153-72. 

Almond, D., & Mazumder, B. (2005). The 1918 influenza pandemic and subsequent health 
outcomes: an analysis of SIPP data. American Economic Review, 95(2), 258-262. 

Almond, D., & Mazumder, B. (2011). Health capital and the prenatal environment: the effect of 
Ramadan observance during pregnancy. American Economic Journal: Applied Economics, 3(4), 
56-85. 

Almond, D., Currie, J., & Duque, V. (2017). Childhood circumstances and adult outcomes: Act 
II (No. w23017). National Bureau of Economic Research. 

Almond, D., Currie, J., & Herrmann, M. (2012). From infant to mother: Early disease 
environment and future maternal health. Labour Economics, 19(4), 475-483. 

Barker, D. J. 1990. “The Fetal and Infant Origins of Adult Disease.” BMJ , 301(6761): 1111 

 

 
 

94 

 

Barker, D.J.P. (1998) In Utero Programming of Chronic Disease. Clinical Science, 95(2): 115– 
28. 

Barker, D.J.P. (1998)  In Utero Programming of Chronic Disease. Clinical Science, 95(2): 115– 
28. 

Beach, B., Ferrie, J., Saavedra, M., and Troesken, W. (2016) Typhoid Fever, Water Quality, and 
Human Capital Formation. The Journal of Economic History, 76(1): 41-75. 

Bhalotra, S. R., & Venkataramani, A. (2015). Shadows of the captain of the men of death: Early 
life health interventions, human capital investments, and institutions. 

Bharadwaj, P., Eberhard, J., and Neilson, C. (2013) Health at Birth, Parental Investments, and 
Academic Outcomes. Journal of Human Resources, R&R. 

Bharadwaj, P., Johnsen, J.V., and Løken, K.V. (2014 ) Smoking Bans, Maternal Smoking, and 
Birth Outcomes. Journal of Public Economics, 115: 72-93. 

Bharadwaj, P., Johnsen, J.V., and Løken, K.V. (2014) Smoking Bans, Maternal Smoking, and 
Birth Outcomes. Journal of Public Economics, 115: 72-93. 

Currie, J., & Almond, D. (2011).  Human capital development before age five. In Handbook of 
labor economics (Vol. 4, pp. 1315-1486). Elsevier. 

Currie, J., & Almond, D. (2011). Human capital development before age five. In Handbook of 
labor economics (Vol. 4, pp. 1315-1486). Elsevier. 

Currie, J., & Hyson, R. (1999). Is the impact of health shocks cushioned by socioeconomic 
status? The case of low birthweight. American Economic Review, 89(2), 245-250. 

Currie, J., & Stabile, M. (2003). Socioeconomic status and child health: Why is the relationship 
stronger for older children?. American Economic Review, 93(5), 1813-1823. 

Currie, J., & Vogl, T. (2013). Early-life health and adult circumstance in developing 
countries. Annu. Rev. Econ., 5(1), 1-36. 

Ellis, F., & Manda, E. (2012). Seasonal food crises and policy responses: A narrative account of 
three food security crises in Malawi. World Development, 40(7), 1407-1417. 

Funk, C., Peterson, P., Landsfeld, M., Pedreros, D., Verdin, J., Shukla, S., ... & Michaelsen, J. 
(2015). The climate hazards infrared precipitation with stations—a new environmental record for 
monitoring extremes. Scientific data, 2, 150066. 

 

 
 

95 

 

Garces, E., Thomas, D., & Currie, J. (2002). Longer-term effects of Head Start. American 
economic review, 92(4), 999-1012. 

Gould, E., Lavy, V., and Paserman, D. (2011) Fifty-five Years after the Magic Carpet Ride: The 
Long-Run Effect of the Early Childhood Environment on Social and Economic Outcomes. 
Review of Economic Studies, 78(3): 938-973. 

Graitcer PL, Gentry EM. Measuring children: one reference for all. Lancet 1981;2:297-9. 

Gutierrez, F. H. (2013). Long-Term Consequences of Early Life Health Shocks: Evidence from 
the 1980s Peruvian Crisis. 

Heckman, J. and Kautz, T. (2014) Fostering and Measuring Skills: Interventions that Improve 
Character and Cognition. In The Myth of Achievement Tests: The GED and the Role of 
Character in American Life, edited by Heckman, J., Humphries, J.E., and Kautz, T. Chicago, IL: 
University of Chicago Press. pp. 341-430. 

Heckman, J.J. (2007) The Economics, Technology, and Neuroscience of Human Capability 
Formation. PNAS, 104(33), 13250–13255. 

Hsin, A. (2012) Is biology destiny? Birth-weight and Differential Parental Treatment. 
Demography, 49(4): 1385-40 

Human Development Report 2016 

Krutikova, S., & Lilleør, H. (2015). Fetal Origins of Personality: Effects of early life 
circumstances on adult personality traits(No. WPS/2015-03). 

Kudamatsu, M. (2010). THas Democratization Reduced Infant Mortality in Sub'Saharan 
Africa. Evidence from Micro Data. U IIES Stockholm. 

Landro, Laura. 2010. “Healthy Living, for Two.” Wall Street Journal, September 9. 

Lavy, V., Schlosser, A., and Shany, A. (2016) Out of Africa: Human Capital Consequences of In 
Utero Conditions. NBER Working paper, 21894. 

Lee, C. (2014) Intergenerational Health Consequences of In Utero Exposure to Maternal Stress: 
Evidence from the 1980 Kwangju Uprising. Social Science & Medicine, 119(C): 284-291. 

Lindo, J.M. (2011) Parental Job Loss and Infant Health. Journal of Health Economics, 30: 869- 
879. 

 

 
 

96 

 

Linnemayr, S., & Alderman, H. (2011). Almost random: Evaluating a large-scale randomized 
nutrition program in the presence of crossover. Journal of development Economics, 96(1), 106-
114. 

Lohmann, S., & Lechtenfeld, T. (2015). The effect of drought on health outcomes and health 
expenditures in rural Vietnam. World development, 72, 432-448. 

M de Onis, C Garza, CG Victora, MK Bhan, KR Norum. The WHO Multicentre Growth 
Reference Study (MGRS): Rationale, planning, and implementation. Food Nutr Bull 2004; 25: 
S1-89 

Maccini, S., & Yang, D. (2009). Under the weather: Health, schooling, and economic 
consequences of early-life rainfall. American Economic Review, 99(3), 1006-26. 

Mazumder, B., Almond, D., & Van Ewijk, R. (2015). Fasting during pregnancy and children’s 
academic performance. Economic Journal. 

Nan, Y., Bao-hui, M., & Chun-kun, L. (2011). Impact analysis of climate change on water 
resources. Procedia Engineering, 24, 643-648. 

Persson, P. and Rossin-Slater, M. (2016) Family Ruptures and Intergenerational Transmission of 
Stress. Forthcoming at the American Economic Review. 

Petronis, Arturas. 2010. “Epigenetics as a Unifying Principle in the Aetiology of Complex Traits 
and Diseases.” Nature, 465(7299): 721–27. 

Rocha, R., & Soares, R. R. (2014). Water scarcity and birth outcomes in the Brazilian 
semiarid. Journal of Development Economics, 112, 72-91. 

Rocha, R., & Soares, R. R. (2015). Water scarcity and birth outcomes in the Brazilian 
semiarid. Journal of Development Economics, 112, 72-91. 

Rosales-Rueda, M. F. (2014). Impact of early life shocks on human capital formation: Evidence 
from el nino floods in ecuador. University of California Irvine working paper. 

Rossin-Slater, M. (2011) The Effects of Maternity Leave on Children's Birth and Infant Health 
Outcomes in the United States. Journal of Health Economics, 30(2): 221-239. 

Rossin-Slater, M., & Wüst, M. (2015). Are different early investments complements or 
substitutes? long-run and intergenerational evidence from denmark. Mimeo, University of 
California Santa Barbara. 

 

 
 

97 

 

Sanders, N.J. (2012) What Doesn't Kill you Makes you Weaker Prenatal Pollution Exposure and 
Educational Outcomes. Journal of Human Resources, 47(3): 826-50. 

Scholte, R. S., Van den Berg, G. J., & Lindeboom, M. (2015). Long-run effects of gestation 
during the Dutch Hunger Winter famine on labor market and hospitalization outcomes. Journal 
of health economics, 39, 17-30. 

Schwandt, H. (2018). The lasting legacy of seasonal influenza: In-utero exposure and labor 
market outcomes. 

Shah, M., & Steinberg, B. M. (2017). Drought of opportunities: Contemporaneous and long-term 
impacts of rainfall shocks on human capital. Journal of Political Economy, 125(2), 527-561. 

Van den Berg, G.J., Lindeboom, M., and Portrait, F. (2006) Economic Conditions Early in Life 
and Individual Mortality. American Economic Review, 96(1): 290-302 

Venkataramani, A.S. (2012) Early Life Exposure to Malaria and Cognition in Adulthood: 
Evidence from Mexico. Journal of Health Economics, 31(5): 767-780. 

Wernerfelt, N., Slusky, D. J., & Zeckhauser, R. (2017). Second Trimester Sunlight and Asthma: 
Evidence from Two Independent Studies. American Journal of Health Economics, 3(2), 227-253. 

WHO Child Growth Standards, 2006 

WHO child growth standards: length/height-for-age, weight-for-age, weight-for-length, weight-
for-height and body mass index-for-age: methods and development. Geneva: WHO; 2006. 

World Health Organization. (2008). Climate change and health. 

 

 

 

 

 

 

 

 

 

 

 
 

98 

 

 

CHAPTER 4: DO SMALLER STATES LEAD TO MORE DEVELOPMENT? 

EVIDENCE FROM SPLITTING OF LARGE STATES IN INDIA 

4.1. Introduction 

 

The advent of the new millennium saw the creation of three new states in India -- Chhattisgarh, 

Uttarakhand and Jharkhand, carved out from the parent states of Madhya Pradesh, Uttar Pradesh 

and Bihar, respectively.  There is a wide literature about the underdevelopment of smaller 

regions within the larger state and how division of states on social and ethnic lines can lead to a 

higher level of homogeneity in the resultant states.  Various sociological, political, economic and 

administrative arguments have been made in favor of such splits (Tillin, 2013). One of the 

rationales for splitting a large state into smaller states is that the earlier neglected (or pockets of 

underdeveloped) regions of the parent state (that retained the original name), will develop faster 

under the newly defined state, due to smaller administrative boundaries for both the parent state 

and the newly formed state. Second, the government and the administration would be more 

responsive to people’s needs due to the democratic framework and therefore there will be more 

development in both the states. An economic argument in favor of such splits is that the newly 

formed states will need new sources of finances, and thus create policies to attract new 

industries, which will bolster growth.  

Preceding arguments have implications for policy decisions made against and in favor of 

forming a new state. However, what is missing in the discussion around the formation of new 

states are the arguments associated with the developmental outcomes of the split. One of the 

questions that needs to be explored is how new states perform in relation to the indicators of 

 

 
 

99 

 

human development once they split from the parent state, and whether the parent state benefits or 

loses from the split.  In this study we try to explore how creation of the three new states have led 

to changes in the quality of lives of its people and the various mechanisms that could have led to 

such changes. There are no studies to our knowledge that have rigorously examined the post-split 

development impacts of the split that happened in India in 2000 and has continued to happen 

since then. Such evidence can support/refute the argument of new state formation on the grounds 

of economic development (underdevelopment) within the smaller states and neutral or positive 

(negative) impacts on the parent states. Looking at the formation of new states with a dimension 

of human development can help theorize if and what might be the outcomes of such a split for 

the affected population in both new and parent state in terms of development indicators. Such 

theorization can also have implications for policy analyses that drive decisions on future 

territorial reorganization (Tillin, 2013). 

As states are the proximate determinants of local institutions driving development 

outcomes, a change in their boundaries provides us an opportunity to evaluate the impact of these 

shifts on the provision of public goods and distribution of development outcomes. However, 

shifting of state borders or creation of new states are rare, and the case of creation of three new 

states in India in November 2000, provides us with a rare opportunity to test these development 

hypotheses. This paper contributes to the literature by providing rigorous evidence of the longer-

term developmental impacts of the creation of the three states on the new and the parent state.  

We consider this split as an exogenous shock and apply difference-in-difference analysis, 

examining socioeconomic outcomes among people living in the new state and the parent state. 

To further the argument, we also use a geographical regression discontinuity analysis to see the 

changes in development indicators on people living in districts on both sides of the newly formed 

 

 
 

100 

 

state borders. As a preview of our results, the analyses confirm the hypothesized positive impact 

of the split on the new states on a wide range of development indicators. However, we find that 

the split had a negative impact on the parent state. This finding is contrary to our expectation and 

has never been considered as an argument against splitting a large state.   

This paper is organized as follows. Section 2 explains the formation of Indian states and 

the history behind it. Section 3 provides the rationale for and against such a division along with 

different hypotheses about potential outcomes of the creation of new states. Section 4 discusses 

the data sources.  Section 5 describes the general econometric frameworks used to test the 

predictions. Section 6 describes the empirical results. Section 7 covers robustness checks where 

we test the assumptions of our empirical strategies, followed by conclusion in Section 8.  

 

4.2. The Case of Formation of New States Through Splitting of Larger States 

 

4.2.1 Examples from Other Countries  

 

I n   m o s t   f e d e r a t i o n s ,   i t   i s   easier to change the status of a subunit – for instance from 

territory to a fully-fledged state than to alter boundaries. Nevertheless, most federal 

systems have created new sub-units by r eorgan izing the boundaries of existing states at 

some stage. This continues to be an active subject of debate, especially in newer 

federations. The most extreme example of federal restructuring is that of Nigeria, which 

created thirty-six states from the three regions it started out when it became independent 

(Suberu 2001). Brazil established the new state of Tocantins, and upgraded Amapa and 

 

 
 

101 

 

Roraima  t o   f u l l   s t a t e h o o d   i n   1 9 8 8   w h e n   a   n e w   c o n s t i t u t i o n   w a s   approved after 

the end of the military regime (Souza 1997). In 1999, Canada created the province of 

Nunavut from its Northwest Territories as a means of accommodating Inuit claims for a 

homeland and for their rights to manage the natural resources. In 1979, Switzerland 

created the new canton of Jura after a referendum.  These are some of the few examples of new 

units with political boundaries formed from splitting larger units or a major re-organization of an earlier 

structure.  

 

4.2.2 Formation of States in Post-independence India: Historical Roots and Current Status 

 

India gained its independence from British occupation in the year 1947 and one of the most 

immediate tasks lying ahead for the first Indian government was to consolidate the country and 

create a national identity. India as we know it today, was a myriad collection of more than 521 

princely states (some historians put it at 565) and they all varied widely in terms of size and 

status.  At one end of the scale were the massive states of Kashmir and Hyderabad, each the size 

of a large European country; at the other end, tiny chiefdoms or jagirs of a dozen or less villages.  

On the eve of Independence, India also went through a very violent division leading to 

the creation of Pakistan. The turmoil of partition made the Union parliament very skeptic of 

creating mini entities within the country that could further ask for separate countries. As a result, 

the Union government monopolized all the rights to reorganize the political structure of the 

country and till date reserves the right to the formation of a new state or altering the boundaries 

of an existing state. Despite the existence of several demands for the formation of numerous 

 

 
 

102 

 

states, the national leadership in the initial years after independence dissuaded from conceding to 

demands for the creation of smaller regional states. While the ruling party Congress had 

promised linguistic provinces before Independence, the country now faced a very critical 

situation resulting from partition. The creation of linguistic provinces, then, had to be deferred 

until such time as India was strong and sure of herself. In 1956 the States Reorganization Act 

was passed which abolished the provinces and princely states in favor of new states which were 

based on language and ethnicity.  Creation of new states and restructuring of existing states 

continued in the sixties, with the prominent splits of the Bombay State into the present day 

Gujarat and Maharashtra (in 1960), and the carving out of the Hindi-speaking Haryana state from 

the southern districts of Punjab state (in 1966). During the next three decades, longstanding 

demands for the smaller states remained in limbo. A number of issues kept policymakers from 

implementing further division of states in India. Reminiscent of ‘partition anxiety’, many fear the 

rise of regional and linguistic fanaticism as threats to national unity and integrity. A global surge 

in ethno-nationalist conflicts over the years served to rekindle these fears of Balkanization of 

India. 

However, in November 2000 the government of India succumbed to the demands and 

three new states were created: Chhattisgarh was created out of eastern Madhya Pradesh; 

Uttarakhand was created out of the mountainous districts of northwest Uttar Pradesh; and 

Jharkhand was created out of the southern districts of Bihar (Figures 1 and 2). After another 14 

years, in 2014, the government gave approval for Telangana to separate from Andhra Pradesh to 

become the 29th state of the union.  

 

 
 

103 

 

Due to lack of data we are not able to study the development impacts of the splits that had 

happened long time ago (i.e., the split of Gujarat-Maharashtra and Punjab-Haryana in the 1960s) 

or very recently (Telangana-Andhra Pradesh in 2014).  Instead, we focus the analysis of this 

paper on the splits that happened in 2000 with the formation of Chhattisgarh, Uttarakhand and 

Jharkhand. It is relatively recent and data are available to study the impacts of this split of three 

larger states that led to the creation of three new states.   

 

4.2.3 The Formation of Three New States by Splitting Larger States in 2000: Driving 

Forces 

 

The formation of the state of Jharkhand, constituting the 18 districts of southern Bihar, was the 

fulfillment of a fifty-year struggle for the creation of a heavily tribal state. The boundaries of the 

new state were less extensive than the originally-conceived Jharkhand, embracing tribal hill 

areas of Madhya Pradesh, Orissa, and West Bengal, in addition to southern Bihar. The newly 

formed state took 35 percent of the population of Bihar--India's second most populous state--but, 

with its coal mines and steel mills, it captured 65 percent of the state's revenue. In the census of 

1991, Bihar had 7.7% tribal population mostly concentrated in the southern districts which 

eventually led to creation of Jharkhand. According to the census of 2011 Bihar was left with 

0.9 % of its population being tribal. On the other hand, 27% of the population of the newly 

formed Jharkhand constituted of tribal population leading to a more homogenous Jharkhand.   

The state of Madhya Pradesh was reorganized with the creation of the state of 

Chhattisgarh, constituting the seven eastern districts of the old state. The division here was 

 

 
 

104 

 

rooted in caste distinctiveness, with upper caste peasant Brahmins and Kurmis and Tribals 

leading the movement for a separate state. Rich in mineral wealth and an important rice-

producer, Chhattisgarh had resented its disproportionately large contribution to state revenues 

but little investment from the state to the region. The new state had a substantial tribal 

population, and was also driven by tribal demands, as was the creation of Jharkhand.   

The formation of the state of Uttarakhand, carved out from the state of Uttar Pradesh 

(U.P.), fulfilled long-voiced demands by the people of the Kumaon and Garhwal hills of 

northwestern U.P. for a separate state based on cultural, social (caste), and economic 

distinctiveness. In Uttarakhand the central justification of demand for statehood was that policies 

formulated in the plains of Uttar Pradesh were often inappropriate for the distinct topography of 

the hills and their natural resource base.  The eleven hill districts and two plains districts that 

formed the new state had long-felt neglected by the U.P. state government (Mawdsley, 2002). 

These three new states that emerged after the spilt in 2000--Uttarakhand, Jharkhand and 

Chhattisgarh, are of the size of a small European country in terms of area and 

population(Jharkhand 32 million ,Portugal 10 million, Netherlands 17 million,  Belgium 11 

million). Therefore, this change can be considered a huge change in the federal structure of India. 

Such massive changes are usually an outcome of multiple processes and covers a wide range of 

political and administrative reasons. The first locates the pressure to create new states within 

statehood movements that have articulated a sense of cultural identity and regional deprivations. 

Second explanation revolves around political motives of the ruling parties that might have 

considered it to be politically beneficial for them. The third rationale suggests that territorial 

reorganization is driven by capitalist expansion and the desire to intensify resource extraction in 

 

 
 

105 

 

the context of economic liberalization. The final explanation is concerned with the administrative 

efficiency gains due to reducing the size of the states to improve governance.  

There are undeniable aspects of all these four explanations that led to the creation of 

these new states; but we argue that some of them explain the process much better than others. 

The political economy argument about capitalist expansion resulting in state creation does not 

hold much ground on the account of multiple factors. First there is considerable diversity in the 

interests among business and industry. Second, there is enough evidence of industrialists being 

ambivalent about the creation of new states as many were not supporting the cause due to the 

fear of increased rent seeking by the new political elites.1 The second argument of political 

mileage by the party ruling in the center which decided on the split also does not gain too much 

traction as it not only badly lost the very next general elections by a massive margin but also 

another subsequent election in 2009 (Sachar, 2009).  

We contend that the important factor which lead to the creation of these 3 new states is a 

continuing part of federal reorganization in which new states are created to better represent the 

subnational diversity. Indian federalism represents a collection of diverse identities reflected in 

the form of different states and following this line of thought the creation of these three states 

reflect the logic of political representation of diversities that were not well accommodated in the 

existing states structure.  In this case particularly, creation of new states gave better 

representation to tribal populations in Jharkhand and Chhattisgarh and to an otherwise socio-

culturally different group in the case of Uttrakhand (Stepan, 2011). Eminent historian 

                                                            
1 For instance, Tillin (2013) reports that the leading industry from Jharkhand - TATA industries 
was of the view that they were better off without the split as the politics was far from them and 
they were ‘kind of neglected which was better for them’; referring to the higher rent seeking due 
to creation of new state. 

 

 
 

106 

 

Ramachandra Guha in his book ‘India after Gandhi’ suggests that creation of Jharkhand and 

Chattisgarh can be considered as an official acknowledgment of the history of tribal suffering.  

Also, the formation of Uttarakhand from the hill districts of Uttar Pradesh fits this narrative as it 

was rich in natural resources and similar to ‘tribal’ suffering, was exploited by powerful external 

interests (Guha, 2007). Guha further argues that Indian republic is still young enough to create 

more states. ‘Regions that have cultural, ecological and historical coherence and are adversely 

affected by their current status as a part of a larger unit, should be granted statehood’ 

(Kumar,2000). Another noted parliamentarian and political scientist Rasheeduddin Khan calls 

for territorial reorganization to divide ‘sprawling huge states’ that perpetuates inequalities. He 

argues for the creation of units that are socio-culturally homogenous and administratively and 

politically manageable (Khan,1997).  

The final explanation for the creation of states is based on administrative efficiency. This 

echoed in the ‘official’ rationale offered by the political leadership that administered the change. 

As this bill was discussed in the parliament, the home minister at that time L. K. Advani 

explained that the states were being created on the grounds of ‘administrative and economic 

viability’ as well as the ‘overwhelming aspirations of the people of the region’ (Tillin, 2013). 

According to professor Ashish Bose at the Institute of Economic Growth, the strategy of 

reducing the size of the states is key to providing better public services (Menon).  We argue that 

this partly explains and justifies the splitting of the states. Our findings of better service delivery 

by public institutions in the new states further substantiates this argument.  

The end of resistance action by the ethnic groups and their mainstreaming into the larger 

political process by the creation of the new states echoes greater political decentralization.  In all 

the three states, the leaders of the political struggle for the statehood ended up becoming the 

 

 
 

107 

 

Chief Ministers (elected head of the state government). While in Jharkhand and Chhattisgarh the 

Chief Ministers came from the local tribal communities, in Uttarakhand there were two chief 

ministers jointly appointed, each one of them representing different regions of the newly formed 

state (Singh 2000). This gave people of the states a sense of representation, as now the masses 

could identify themselves with the ruling political representatives. In this paper we explore 

whether this sense of representation translates into better service delivery by the governments 

leading to better developmental outcomes for the new state, and whether the parent state also 

benefits from the reduced size and efficiencies. 

 

4.3. Concepts and Theories Underlying the Rationale for and Against the Formation of a 

New State and Research Hypotheses 

 

We consider the borders of the states within a federal system as a form of institution, 

underpinned by patterns of social, political and economic power. State boundaries help structure 

political and economic life by determining which groups are in competition with each other over 

distribution of resources as well as determining which resources are at stake. As institutionalist 

theories would suggest, the location of state boundaries help to reinforce and reproduce patterns 

of power. The formation of a new state has its origins in challenging the existing power 

structures. It offers a possibility of creating new types of regional polity in which different 

groups might find prominence, cultural and ethnic groups might find more direct control over 

resources and different relationships might be negotiated between the governments and people. 

 

 
 

108 

 

So, creation of a new state acts as a vehicle for formation of new institutions that are more likely 

to serve its citizenry better. 

 

4.3.1 Pro- and Anti-Split Arguments 

 

Several arguments have been put forward in the literature in favor of the formation of new states 

(from an existing large state). First, the creation of a smaller state unleashes the suppressed 

growth potentials of the hitherto peripheral regions of a large state. It provides an opportunity for 

the earlier neglected (or pockets of underdeveloped) regions of the parent states to develop faster 

under the newly defined states.  

Second, comparatively smaller but compact geographical entities tend to ensure that there 

is better democratic governance, as there is greater awareness among the policy makers about the 

local needs. Smaller states provide gains for the electorates in terms of better representation of 

their preferences in the decisions by the government. Geographically compact states mean key 

decisions will be taken closer to the ground. It makes more sense if the decision on key issues of 

a district in Himalayas (Uttarakhand) is taken in Dehradun (the capital of the new state) which is 

50 km away, compared to in Lucknow (the capital of the parent state) which is 500 kilometers 

away (Kumar, 2010). 

Third, smaller spatial units have linguistic compatibility and cultural homogeneity, which 

allows for better management, implementation and allocation of public resources in provisioning 

basic social and economic infrastructure services. A relatively homogeneous smaller state allows 

for easy communicability, enabling marginal social groups to articulate and raise their voices. 

 

 
 

109 

 

The relationship between social and ethnic homogeneity, and economic growth has been widely 

studied in the economics discipline. Alesina et al. (1994, 2004) studied the effect of distributive 

politics on economic growth and showed that higher inequality leads to lowers growth. They find 

strong evidence of trade-off between economies of scale due to bigger jurisdiction and the cost 

of heterogeneous population that comes with it. Easterly (2006, 2007) shows how social 

cohesion is an important factor for creating more responsive institutions leading to economic 

growth. Furthermore, Easterly (1997) associates low schooling, political instability, 

underdeveloped financial systems, distorted foreign exchange markets, high government deficits, 

and insufficient infrastructure in Sub-Saharan Africa to higher level of ethnic fragmentation and 

the resultant less effective institutions.  

On the other end of the spectrum, there are arguments or concerns about splitting the 

states. First, many fear that such divisions can give rise to regional and linguistic fanaticism, 

which can be threats to national unity and integrity. Second, many believe that bigger states 

ensure cohesion and stability, which is good for development; although, there are myriad forms 

of political violence that go unabated in the big states and can undermine development. Third, 

smaller states (like mineral rich Chhattisgarh and Jharkhand) are often viewed as being much 

more vulnerable to the pressures of the corporations and multi-nationals due to their small scale 

economies and the greed of the newly emergent regional elite (Tillin 2012).  Another argument 

in favor of a larger state (or not splitting it) is that it has many benefits of economies of scale; 

they have more revenue and are able to implement projects and programs that require more 

capital.  

Whether these benefits of not splitting larger states outweigh the benefits of splitting 

them is an empirical question we are interested in exploring in this paper.   

 

 
 

110 

 

4.3.2 Research Hypotheses 

 

Based on the arguments for and against the split outlined above, one can hypothesize that the 

splitting of states can have many effects on both the parent and the new state as shown below in 

table 1.  The table suggests that there could be multiple positive and negative impacts on both the 

states with varying degree of magnitude. As majority of the impacts seem to be positive in this 

case, we hypothesize that the overall impact of state splitting on both states will be positive and 

will be reflected in the developmental indicators. Such development may possibly be credited to 

smaller administrative boundaries for both the parent state and the newly formed state. The 

underlying reasoning is that the government and the administration within these parent and new 

states would be potentially more responsive to people’s needs due to the democratic framework 

and which may consequently lead to more development in both the states. In addition, need for 

new sources of finances to sustain these states may favor formation, adoption and 

implementation of policies that favor and attract new industries and business, leading to 

economic growth and development. A larger state has many benefits of economies of scale; but 

in a heterogeneous large state, it is highly likely that some sections of the populations will be far 

from reaping the benefits of such economies of scale. In case of marginalization of 

underrepresented groups in the existing structure, gaining more political representation in the 

form of creation of new states can lead to better developmental outcomes for such groups. 

In this paper we try to investigate the impact of the creation of the three new states on the 

developmental indicators like literacy rate, provision of electricity, toilets, concrete houses, 

infrastructure, etc. of the newly formed state and the parent state.  We also make an attempt to 

investigate the impact of this exogenous shock on provision of goods and services by 

 

 
 

111 

 

government (like government health facilities), as it is critical for the development of majority of 

population especially located in rural area.  

Table 2 shows summary statistics for several development indicators—i.e., growth in per 

capita income, literacy, percentage households with electricity and percentage household with 

toilets between 2001 and 2011 for all the six states (three parent and three new).  On all these 

indicators, in most cases the new states have fared quite well compared to the parent states. This 

leads to our second hypothesis which suggests that splitting has been more beneficial for new 

states as compared to the parent states.  However, we do not know if the districts that formed 

new states, were inherently different from the districts that were left behind in the parent state 

and that is why they did better after splitting. This might lead to self-selection which makes the 

simple comparison presented in the table below, less meaningful. To deal with the selection bias 

issue and to capture the impact of splitting requires credible identification strategies. This is what 

the present study attempts to do as explained in the empirical strategy section.  

One study that has attempted to attribute the impact of this split is the study by Asher and 

Novosad (2015). They used border regression discontinuity design to show how the new states of 

Jharkhand, Chhatisgarh, and Uttarakhand have done better after their creation in 2000. They 

compared villages across the newly formed border (from 10 up to 50KM) using the border as a 

cutoff point, so that villages in the new state are considered as treatment and villages in parent 

state are considered as control.  They use satellite photos of the earth at night, as a proxy of local 

economic output, and they show that the border regions were indistinguishable from 1992-1999, 

and then sharply diverge, with significantly higher light output in the new states. Although 

nighttime light has been used as a proxy for economic development, in the above study it does 

not give us much information about the details of development indicators or the provision of 

 

 
 

112 

 

public goods.  Night time luminosity is only a function of availability of electricity in the area, 

which is just one of the indicators of development and does not necessarily translate as an 

indicator of overall growth of the region.  As the new states have more coal reserves, they might 

be generating more electricity, which is usually distributed free in many households and might 

show up in luminosity index, but that not necessarily means other indicators of developments are 

doing equally well. As the idea behind splitting of the big states were administrative efficiency 

and better provision of pubic good there is a need for nuanced indicators to show if there has 

been any difference due to the splitting. This paper addresses the limitation of the Asher and 

Novosad (2015) analysis by using micro-data from National family health survey (Demographic 

Health Survey) examines a broad range of indicators to show the broader impact of state 

splitting.  

 

4.4. Data   

 

 This paper uses the data from nationally representative Demographic Health Surveys to capture 

a comprehensive picture of development impacts of splitting of states. The Demographic Health 

Survey is as multi-topic survey with three rounds of data collected in years 1992, 1998 and 2015. 

This dataset was collected with the purpose of documenting the changes in the lives of Indian 

households in a society undergoing rapid transition. This dataset gives a good understanding of 

the effectiveness of the state machinery in provision of public goods like schools, health centers, 

drinking water, electricity, toilets, government housing, etc. Further it has an in-depth health 

module that covers health indicators like infant mortality rate, institutional deliveries in 

government hospitals, provision of health cards, etc. It also covers the details of all the required 

 

 
 

113 

 

vaccine which are primarily provided by government agencies and are a very good indicators of 

the effectiveness of the movement’s service delivery measures. Additionally, it also provides 

information on ownership of consumer durables like television, mobile phone, computer, 

washing machine, refrigerator and motor vehicle that can act as a proxy for economic wellbeing.  

 

4.5. Empirical strategy  

 

The objective of this paper is to test the two hypotheses presented above that-- 1)  The increased 

autonomy and political representation due to splitting of states has positive effect on 

developmental outcomes for both the parent and the new states; but 2) the newer states 

experience relatively more benefits than the parent state We take two quasi-experimental 

approaches to test these hypotheses and to estimate the effects of splitting. First is a difference-

in-difference (DiD) approach with two specifications (described below), and the second approach 

is a (geographical) regression discontinuity (RD) design that compares districts on both sides of 

the newly formed border due to splitting of an existing state.   

 

 

4.5.1 Difference - in - Difference Approach  

 

The quasi experimental method of D-I-D proposed in this paper is in line with the method used 

by Banerjee and Ghatak (2002) to find out the impact of an agricultural land tenancy program – 

“operation Barga” on productivity difference in Bangladesh and West Bengal (Banerjee and 

Ghatak 2002). In their paper, the researchers estimated the impact of the program on West 

 

 
 

114 

 

Bengal (In India) using Bangladesh as a control. Using a similar method, in the present study I 

use the parent state as control.  

The fact that new statehood was granted to Chhattisgarh, Jharkhand and Uttarakhand is 

largely attributed to an exogenous shock. There are around 25 such popular movements in India 

for the creation of new states and almost every single big state has some sort of protest going on 

for a new state. Figure 3 shows all the major mass movements for the creation of new states in 

India. After three years of unstable governments, in the end of year 1999 a new government was 

unexpectedly elected with full majority won by National Democratic Alliances (NDA). The 

victory was associated with a nationwide wave against the congress party government which had 

ruled most of the states since independence. With the victory in general elections NDA wanted to 

make popular appeals and chose these three states to be split. Also, as the decision to split a state 

or alter boundaries of a state are completely controlled by the Union government, these states 

had a limited say in the overall process used, which makes the shock exogenous.  We use two 

model specifications. In the first specification we use the adjacent states as comparison group 

and both the new and parent state as two different treatment groups.  In the second specification, 

we consider the parent state as a comparison group and new state as a treatment group. In both 

these specifications, the outcomes will be compared for before and after the division of the 

parent state occurred. 

In the first specification we use D-I-D with new state and parent state as two treatments 

and the adjacent states as comparison. For this strategy we use the states of Rajasthan, West 

Bengal, Odisha and Himachal Pradesh as control states as they have very similar development 

indicators compared to the three states of Madhya Pradesh, Uttar Pradesh and Bihar. The control 

states also share borders with one or more of the splitting states, and therefore have many areas 

 

 
 

115 

 

that are culturally and linguistically similar. In the late 1980s and 1990s some of these states 

were clubbed together as least developed states with serious developmental challenges.  They 

were referred as BIMAROU, which is an acronym formed from the first letter of the states Bihar, 

Madhya Pradesh, Rajasthan, Odisha and Uttar Pradesh, which meant “sick” in Hindi and 

indicative of the performance of these states. This group of five states together were considered 

to be dragging down the GDP growth rate in India.  

The three control states that we are considering for this strategy -- Odisha, West Bengal 

and Rajasthan also have major movements in support of splitting the states. India’s oldest 

struggle for a new state is for the creation of Gorkhaland which is a part of West Bengal. 

Similarly, there has been a long-standing demand by people of western Odisha (Odisha is a 

control state for us) to have a separate state of Koshal for themselves. Our third control state 

Rajasthan has also, experienced a popular demand for new state for the western part of the state. 

All these demands are at least as strong as the ones that led to the creation of three new states we 

are considering. Our fourth control state, Himachal Pradesh is considered as a control state as it 

has similar hilly regions as Uttarakhand and has comparable agro-climatic zones and cultural 

heritage.  

We measure the impact of splitting of states on developmental indicators of both the 

parent and the new state using a difference-in-difference estimator using the following 

specification.  

(cid:5)Mg= μg+Jl+V>∗ 1(cid:20)(cid:28)31@(cid:28)(cid:31)1>∗(cid:10)(cid:19)41g+VC∗ 1(cid:20)(cid:28)31@(cid:28)(cid:31)1C∗(cid:10)(cid:19)41g +Vt1(cid:20)(cid:28)31@(cid:28)(cid:31)1>+
           Vu 1(cid:20)(cid:28)31@(cid:28)(cid:31)1C+Vv (cid:10)(cid:19)41g+w x=g+ \Mg                                                     --------(1) 

 

 
 

116 

 

The dependent variable (cid:5)Mg is a developmental indicator such as the literacy rate, 
insurance provided by government for the household i in survey round t. (cid:5)Mg also include the 

availability of clean drinking water, toilet, concrete house, clean cooking fuel and health 

indicators of consumer durables and services like, ownership of television, refrigerator, motor 

vehicle, computer, mobile phone, washing machine, fan and the Internet. We do not have clear 

data on income and consumption of the households, so ownership of consumer durables act as a 

good proxy for them. In these states government is the major provider of health services and 

therefore we have included a wide range of health indicators to examine if the service delivery of 

health has changed due to the creation of new states. We include major health indicators like, 

institutional deliveries in government hospitals (whether the child was delivered in an 

institutional setting or at home), rate of death of infant under the age of one year. We also 

include whether the child under the age of 5 is underweight, whether the child is stunted or 

wasted, and does the household own a health card which is provided by the government system. 

Also included is the compensation for sterilization as a public policy program to control 

population and is of high importance given the burgeoning population in these states. 

Furthermore, we include all the important vaccines and pre-natal support to the pregnant mothers 

and their children primarily by the government run health systems. We include tetanus shot and 

iron tablets provided to the pregnant mothers; polio vaccine, measles vaccine, Bacille Calmette-

Guerin (BCG) vaccines and vitamin A provided to infants in our analysis.  

The right-hand variables include fixed effect for each district Jl and a fixed effect for 
each round μg. There are two treatment variables on the right-hand side where 1(cid:20)(cid:28)31@(cid:28)(cid:31)1> 
indicates whether the district falls under the new state and 1(cid:20)(cid:28)31@(cid:28)(cid:31)1C indicates whether the 
district falls in the parent state. The coefficients V> 3(cid:31)6VC would be the difference-in-difference 

 
 

 

117 

 

estimate of impacts of splitting on new and parent state respectively compared to the control 

state. Note that the values of V> 3(cid:31)6 VC will not tell us whether the difference between the new 

and parent state is statistically significant or not. It will only tell us if they are statistically 

significantly different (and the magnitude of the difference) from the comparison districts in the 

adjacent states. This model will help test the first hypothesis on whether the split had any effect 

on both the states (parent and newly formed) 

 In second specification we use D-I-D with new state as treatment and the parent state as 

comparison. In this second specification we measure the differential impact of splitting of states 

on developmental indicators of the new state as compared to the parent state using a difference-

in-difference estimator with repeated cross section data with district fixed effects. This model 

specification helps address the second hypotheses of whether the split has benefitted newly 

formed states more or less than the parent states. Following is the empirical specification used: 

 

  

(cid:5)Mg= μg+Jl+V> 1(cid:20)(cid:28)31@(cid:28)(cid:31)1∗(cid:10)(cid:19)41g+VC 1(cid:20)(cid:28)31@(cid:28)(cid:31)1+Vt(cid:10)(cid:19)41+w x=Mg+ \Mg  --------(2) 

 

The dependent variable (cid:5)Mg is a developmental and health outcome as in the earlier 

specification. The right-hand variables include fixed effect for each district Jl and a fixed effect 
for each round μg. Treatment is a dummy variable that is 0 if the district falls in the parent state 
and 1 if the district falls in the new state. (cid:10)(cid:19)41g is an indicator of whether it is a post splitting 

period so basically it would be 1 for any period after year 2000 and 0 for any period prior to that. 

The district fixed effects control for district-specific factors that are fixed over time, and the 

round fixed effects control for factors that vary over time but are common across all districts—

 

 
 

118 

 

both treatment and control. There are also a series of control variables (the X’s) that vary over 

time and across districts and \Mg is the error term. The coefficient V> would be the difference-in-

difference estimate of impacts of splitting on new state.   

4.5.2 Geographical Regression Discontinuity Approach  

 

We use Regression Discontinuity (RD) design as another identification strategy to test 

hypotheses two. This method uses data only from the districts that fall on both sides of the split 

boundary lines.  The RD method does not require exogeneity of the treatment variable with the 

outcome variable, and therefore solves the problem of endogeneity arising from cherry-picking 

of the states that went through division.  The literature on RD suggests that it can be analyzed 

and tested like randomized experiment if variation in the treatment near the threshold is 

approximately randomized. Which means that all “baseline characteristics”—all those variables 

determined prior to the realization of the assignment variable—should have the same distribution 

just above and just below the cutoff (Lee and Lemieux 2009).  

We compare the developmental indicators among people living in adjacent districts 

across both sides of the newly created border. RD solves the identification challenge by 

assuming that districts around a treatment threshold are similar in all characteristics except for a 

certain exogenous factor that assigns the treatment to some and not to others.  Figure 2 shows 

both the new and the old states with the newly formed boundaries between them in yellow. All 

the districts sharing this boundary are part of our regression discontinuity analysis.  

From existing literature, we know that before the new states were created, the 

socioeconomic differences across the border were minimal; caste and tribal identities differed, 

but public goods, population, economic output and agro-climatic-environments were the same on 

 

 
 

119 

 

both sides of the not-yet-created states (Asher and Novosad 2015). To test and verify this, we 

conduct a balancing test and compare the outcome variables of interest, for districts on both sides 

of the borders in the pre-split era, i.e., for the year 1998-99. Table 3 reports the results of this 

balancing test, which suggests that there is no significant difference in the baseline variables 

across the yet-to be formed state lines prior to the split. This justifies our use of the geographic 

regression discontinuity method.  As the federal government decided which districts will go to 

new states and which ones will stay with the parent states it also takes care of the basic 

requirement of an RD design that the threshold variable cannot be manipulated by the 

beneficiary of the treatment. 

Following Imbens and Lemieux (2008) we use local linear regression and cover all the 

districts on both sides of the newly formed borders. We keep the specification very similar to 

specification (2), but now the treatment only includes the households in the new states which 

reside in the bordering district to the parent state and the control households are those which live 

on the other side of the newly formed border in the parent states.  Following is the specification 

we use for the RD analysis: 

 

(cid:5)Mg= μg+Jl+V> 1(cid:20)(cid:28)31@(cid:28)(cid:31)1_dee∗(cid:10)(cid:19)41g+VC 1(cid:20)(cid:28)31@(cid:28)(cid:31)1zll+Vt(cid:10)(cid:19)41+w x=Mg+ \Mg     

--------(3) 

 

Where (cid:5)Mg is a household level developmental outcome of household i in district D and time t. =g 
is the vector of household level controls.  The term, 1(cid:20)(cid:28)31@(cid:28)(cid:31)1_dee is a dummy variable 
indicating that the households fall in a bordering district across the newly formed state and {Mg is 

 

 
 

120 

 

the household level error term. The coefficient V> would be the RDD estimate of impacts of 

splitting on new state. 

 

4.6. Results 

 

The results for all three models--two specifications of DiD and one of RD for the main variables 

of interest are summarized in Tables 4-7. Detailed results of all the models for all the indicators 

across all the states (three parent and three newly formed states) are included in appendix.  Table 

4 reports the impact of splitting of the states on the newly formed states and the parent states on 

important household level developmental indicators. For most of these indicators related to 

dwelling characteristics, the results indicate a significant positive effect of the split on the new 

states, but an opposite effect on the parent states. The effect on the level of education (as 

measured by the education attained by the head of the household), the effect is positive for the 

new state but only in the RD model. The effect of the split on access to electricity was significant 

and positive for both the parent and the new states; although the effect size is much larger for the 

new states when compared with adjacent states.  The probability of having electricity increased 

by 16% for the households in the new state post-split compared to the adjacent states, but there is 

only a 5% increase in the probability of households in parent states compared to the adjacent 

states. As expected, in the specification where parent states are the comparison and new state is 

the treatment, we find that the probability of having electricity increases by 10.5% for a 

household residing in the new states compared to the parent states. The outcome of the RD 

analysis also suggests that the bordering districts which fall in the new state have done better 

 

 
 

121 

 

compared to the districts on the other side of the border in parent states in terms of electricity. 

The households in the new state are 8% more likely to have electricity compared to their 

comparison households in the other side of the newly formed borders.  

Similarly, the likelihood of getting piped drinking water in their residence is also higher 

by 7 percentage points for the new states as compared to the adjacent states, however for the 

parent state it decreases by close to 10%. When we compare the new state and the parent state, 

we find that the new state is doing better as households in the new state are 2% more likely to 

have piped drinking water compared to the households in parent state. Similarly, when we 

compare the districts across the newly formed borders, using RD specification, we find that 

households in the new state side are 3% more likely to have piped drinking water in their houses 

as compared to their neighboring district households across the border.  

The difference between new state and parent states is more remarkable in the case of 

concrete housing. While the households in new states are 16% more likely to have a concrete 

housing, households in the parent states are 21% less likely to have a concrete housing; 

compared to the adjacent states. This is suggestive of the fact that more economic activity has 

happened in the new states post-split and probably there is a decline in the same in parent state. 

Further, when we compare just the new state and parent state we find that compared to the parent 

state, about 5% more households in the new states lived in a concrete house (Table 4).  

The availability of a toilet and clean cooking fuel in the households follow similar trend 

as concrete housing, with new states outperforming the parent state. It is worth noticing that on 

both these indicators, not only the new states are doing better compared to the adjacent states, but 

the parent states are doing worse compared to the adjacent states after the split.   

 

 
 

122 

 

Government plays a significant role in the development of infrastructure underlying the 

provision of these basic amenities, sometimes more directly than others. For instance, piped 

drinking water, electricity, and clean cooking fuel are only provided by government institutions 

and an increases access to these basic amenities and facilities in the new states makes a case for 

an effective and accountable functioning of the government mechanism.  Further, the households 

in new states are doing better than their counterparts in terms of having concrete housing and 

having the basic necessity of a toilet. In many cases these toilets and concrete houses are 

constructed under the ongoing massive federal programs and household doing better on these 

indicators are indicative of either better government engagement with people or of better 

economic condition of the households. It is worth noticing that the households in same districts 

before splitting of the state had lower access to a concrete house, electricity and clean cooking 

fuel (Appendices). 

Table 5 reports the outcomes for ownership of consumer durables and household goods 

that reflect household income or purchasing power. The results show mostly negative effects for 

the parent state and a slightly mixed picture for new states. Compared to adjacent states, the 

parent state is worse off on most of these indicators except ownership of computer (e.g., access 

to Internet, mobile phone, TV, motorcycle, fan, refrigerator, and washing machine).  In general, 

the new states are doing better compared to the parent states. There was hardly any internet in 

India prior to the year of split, but post-split the households in new states have done better than 

the adjacent state by 3% points, while the households in parent states are lagging behind the 

adjacent states by 1%.  When we compare the parent state and new state households, the latter 

are 4% more likely to have internet in 2015-16 survey. The difference is starker in the RD 

specification where we see that the households in the new states’ bordering districts are 7% more 

 

 
 

123 

 

likely to have internet than their counterparts across the border in the parent state in the post-split 

period (Table 5).  

Household’s ownership of television and refrigerators follow a similar trend with new 

states doing better than old states. A household in the new state is 10% more likely to have both 

TV and refrigerator, while a household in the parent state is 19% less likely to have a TV and 

11% less likely to have a refrigerator compared to the adjacent states, post-split. Using other 

specification, we find household’s ownership of TV is better in new state, but we do not find any 

significant difference in ownership of refrigerators. The results are similar for ownership of fans 

and mobile phones. However, using adjacent state as comparison, we find that the new states are 

doing much worse in terms of ownership of computers and washing machines. In the case of 

washing machines, they are even worse than the parent states (Table 5). 

Moving to the important health indicators in table 6, we find that both new and parent 

states have done better in most aspects compared to the adjacent states, while the parent states 

have done better on some indicators, on other indicators they have faltered. Results suggest that a 

household in a new state is 8% more likely to have government provided health insurance while 

a household in parent state is 19% less likely to have such insurance when compared with 

households in the adjacent states. The trend continues across other specifications, where we find 

that a household in new state is 27% more likely to have the government provided health 

insurance when compared to a similar household in the parent states. In the RD specification, we 

find that households in the bordering districts of the new state are 20% more likely to have the 

insurance compared with a similar household across the border, post-split.  

 

 
 

124 

 

Another important child health variable is the number of deliveries in an institutional 

setting like a hospital or a public health center. As these institutions are more likely to have 

better hygienic facilities leading to significant risk of mother and infant mortality, it is important 

to see how both the new and old states are doing on this parameter. We find that a child born in 

both new and parent state is 9% more likely to be born in a government institution post-split. 

Using other specifications, we find that children in new states are more likely to be born in a 

government institution compared to parent state, post-split. Further, children born in new states 

are more likely to survive their first year in the new states as compared to the adjacent states. 

There is no significant difference between, adjacent states and parent states on child survival. 

There is a marginal increase of 2% in the probability of child survival in the bordering district of 

new states compared to children on the other side of the border.  

On other important child health outcomes such as stunting, wasting and underweight, we 

find mixed results; suggesting new states doing better in some but not on others, and same is true 

with parent states. The probability of a child being underweight in both parent and new state is 

4% less compared to a child from adjacent states, post-split.  Across other specifications, 

children in new states are less likely of be underweight compared to children in the parent state.  

Stunting is associated with an underdeveloped brain, with long-lasting harmful 

consequences, including diminished mental ability and learning capacity, poor school 

performance in childhood, reduced earnings and increased risks of nutrition-related chronic 

diseases, such as diabetes, hypertension, and obesity in future. While India’s economy has been 

growing at impressive rate, the country still has the highest number of stunted children in the 

world (i.e., 46.8 million children) representing one-third of the global total of stunted children 

under the age of five (UNICEF India, 2017). Stunting starts from pre-conception when an 

 

 
 

125 

 

adolescent girl and who later becomes mother is undernourished and anemic; it worsens when 

infants’ diets are poor, and when sanitation and hygiene are inadequate. It is irreversible by the 

age of two and an indicator of long term child health. In our analysis we find that a child born in 

new states is 3% less likely to be stunted compared to children in adjacent states. For parent state 

we do not find any significant difference from adjacent states, suggesting they are doing as good 

as other comparable states. Surprisingly, we find that when we compare the children in new 

states with children in parent states, we find that the likelihood of the child being stunted is 5% 

higher for those in the new states. The effect reverses, when we compare children in the 

neighboring districts across the borders, with children in new states doing better by 6% 

compared to children from parent states.  

Wasting is an outcome of shortage of adequate nutrition in short duration and is 

indicative of a short-term shock, in contrast to stunting, which is regarded as chronic 

malnutrition. Wasting could be an outcome of a temporary agricultural or other shock in the new 

states and does not necessarily mean that the nutritional levels of children in parent and new 

states are different. On wasting parameter, we find that the parent states have done better than the 

adjacent states while the new states are not significantly different from the adjacent states. 

Comparing the children across the borders in RD specification, we find that children in the new 

states have done better.  

Next, we analyze the availability of health cards provided by government, which signals 

the outreach of government’s provision of health services. We find that both parent and new 

states have done better on this front post-split compared to adjacent states; new states have done 

much better than parent states. The government of India runs sterilization programs among 

women to control the burgeoning population and they pay a compensation to the women. The 

 

 
 

126 

 

results on our analysis suggest that women in new states are 3% more likely to get the 

compensation, while a woman from the parent state is 6% less likely to get compensation 

compared to women in adjacent states. Further, the women in the new states bordering districts 

are 3% more likely to get the compensation, suggesting that the health system there is less 

corrupt and more effective.  

In table 7 we look at access to child health services and nutrition programs. These include 

availability of different vaccines, vitamin A tablets, and iron supplements for children and 

pregnant mothers. These services are provided by government health facilities. The results show 

that compared to adjacent states, the probability of getting all four doses of polio, BCG and 

Measles for children is much higher in the new states. The children in parent states have also 

followed the same trend and their probabilities of getting measles vaccine, BCG vaccine and 

vitamin A tablet have also increased compared to adjacent states in post-split period. But the 

likelihood of getting polio shots has decreased in the parent state. This suggests that the health 

systems in both parent and new states have done better ex post to the split. Pregnant mothers in 

both parent and new states follow similar trend and are doing better in terms of getting tetanus 

shots and iron tablets during pregnancy post-split.  

 Tables 8, 9, 10 and 11 present the summarized results at the state level to explain the state-wise 

heterogeneity and similarities in the results. The detailed tables are presented in Appendix D. Table 

8 reports the results of major developmental indicators and consumer durables. Column 1 of table 

8  compares  Jharkhand  to  its  erstwhile  state  Bihar  using  a  difference  in  difference  approach. 

Column  2  compares  the  districts  across  the  newly  formed  borders  using  RD  approach.  Similar 

results for other four states are presented in columns 3-6. The table suggest that a household in 

Jharkhand is 11% more likely to have electricity compared to a household in its parent state Bihar. 

 

 
 

127 

 

Furthermore,  a  household  in  the  border  districts  of  Jharkhand  are  13%  more  likely  to  have 

electricity compared to their neighbors on the other side of the border, post-split. Households in 

Chhattisgarh also experience a surge in electricity provision post-split as shown in columns 3 and 

4. We find an increase in electricity connections in the households in Uttarakhand post-split using 

the D-i-D approach where we compare Uttarakhand to its parent state Uttar Pradesh, but we do not 

find any significant improvements in the bordering districts.   

Next, the availability of piped drinking water and toilets has increased for all three newly 

formed  states  using  either  of  the  specification  as  suggested  by  table  8.  The  improvements  are 

marginal for Jharkhand and Chhattisgarh, but significant for Uttarakhand. Housing shows a mixed 

result  with  improvements  seen  in  Chhattisgarh  using  both  the  specifications,  but  for Jharkhand 

only  the  D-i-D  shows  increase  in  concrete  housing  while  for  Uttarakhand,  D-i-D  specification 

shows  increase  in  the  overall  state  and  RD  shows  decline  in  the  bordering  districts.  For  clean 

cooking  fuel,  there  is  not  a  huge  variation  among  states  as  all  the  three  new  states  show  slight 

improvement.  For consumer durables like television, we find that all three new states have done 

better, but Chhattisgarh and Uttarakhand have done significantly better compared to Jharkhand. 

Both Uttarakhand and Chhattisgarh experienced a huge surge in internet penetration compared to 

their parent states. On the other hand, Jharkhand actually experienced a decline in this indicator. 

For other consumer durables like refrigerator and motorcycle, we find in most cases the new states 

have done slightly better without much heterogeneity in the results.  In general, we find that all 

three newly formed states have done better than their parent states post-split with some amount of 

heterogeneity across various indicators. The overall results do not seem to be driven by skewed 

results in any one state.  

 

 
 

128 

 

Table  9  presents  the  results  for  DiD  estimator  using  the  adjacent  states  as  comparison 

group.  Column  1  shows  the  changes  in  development  indicators  for  the  newly  formed  state 

Jharkhand  as  compared  to  adjacent  states  that  did  not  go  through  any  split;  Odisha  and  West 

Bengal.  Similarly,  columns  2-6  present  state-wise  results  for  two  new  and  three  parent  states 

comparing them with their adjacent states after the split. Columns 1 and 2 suggest that after the 

split a household in Jharkhand is 12% more likely to have electricity as compared to the adjacent 

state,  while  a  household  in  Bihar  does  not  have  any  significant  change  in  the  availability  of 

electricity  after  the  split  as  compared  to  the  nearby  states.  Both  these  states  experienced  an 

improvement  in  piped  drinking  water  but  the  magnitude  of  improvement  for  Jharkhand  is 

marginally  higher  than  Bihar.  In  the  case  of  concrete  housing,  households  in  Jharkhand 

experienced an increase in the possibility of having a concrete house by 11% while a household in 

Bihar there was a decline by 1% compared to the adjacent states. Similarly, we see an increase in 

availability of toilet for  households in Jharkhand and a decline in availability of toilet amongst 

households in Bihar post-split when we compare them to the adjacent states. The trend of positive 

growth  in  Jharkhand  and  a  decline  in  Bihar  persists  for  clean  cooking  fuel,  internet  and  other 

important consumer durables.  

In  columns  3  and  4  we  compare  Chhattisgarh  and  Madhya  Pradesh  with  their  adjacent 

states of Maharashtra, Andhra Pradesh, Gujarat and Rajasthan. The results in columns 3 and 4 are 

similar to columns 1 and 2, where we saw that the newly formed state performs better compared 

to  the  adjacent  states  while  the  parent  states  did  worse.  Chhattisgarh  shows  improvement  in 

electricity,  concrete  housing,  toilets,  clean  cooking  fuel  and  other  consumer  durables  while 

Madhya Pradesh shows a decline in piped drinking water, concrete housing, toilet, clean cooking 

fuel and most consumer durables when compared with the adjacent states post-split. Similar results 

 

 
 

129 

 

exist  for  Uttarakhand  and  Uttar-Pradesh  with  Uttarakhand  doing  better  in  most  developmental 

indicators while Uttar Pradesh condition seems to be worsening compared to the adjacent states of 

Himachal Pradesh, Haryana and Rajasthan. Overall, table 9 shows a very clear image of how the 

new states have done better after split while the parent states done worse, which is very similar to 

the results we saw in tables 4 and 5. Further, there is heterogeneity between the states for different 

outcome variables but no clear winner among the new or old states that seems to be driving the 

results; the results seem similar for all the three new and three parent states.     

Table  10  reports  the  results  of  major  health  indicators.  Column  1  of  table  10  compares 

Jharkhand to its parent state Bihar using a difference in difference approach. Column 2 compares 

the districts on both sides of the newly formed borders using RD approach. Similar results for other 

four states are presented in columns 3-6. Column 1 shows a 5 % increase in the likelihood of a 

child being born in a government facility in all the three new states if we use the entire parent state 

as  comparison,  however,  we  do  not  see  any  significant  difference  in  the  RD  specification  for 

districts  right  across  the  borders.  Infant  mortality  shows  a  decline  for  both  Jharkhand  and 

Chhattisgarh using either of the specification, but there is no significant difference in Uttarakhand.  

Malnutrition indicators of stunting, wasting  and  underweight shows a decline for all three new 

states when compared with the parent state across at least one specification. Similar results persist 

for other vaccines like Polio shots, Measles and BCG where the newly formed states seem to be 

doing better than the parent state in at least one of the specifications and in some case for both the 

specifications. This suggests that the new states in general have done better than the parent states 

in important health indicators as already presented in tb,les 6 and 7. Moreover, all the three states 

have done better in varying degree for different indicators, suggesting heterogeneity in results, but 

the results are not skewed in favor or against any of the states.  

 

 
 

130 

 

 

 Table 11 presents the D-i-D results of important health indicators in comparison to the 

adjacent states. This table suggests a mixed result with new states doing better post-split in most 

health indicators as compared to the adjacent states, but in some cases the parent states have also 

done better than the adjacent states. For example, infant mortality has gone down in Bihar and 

Madhya Pradesh significantly, but not in Uttar Pradesh among the parent states, while it has gone 

down in Jharkhand but not in Chhattisgarh or Uttarakhand. For nutritional indicators of stunting, 

wasting and underweight, Bihar has done much worse on all three indicators, but Jharkhand 

shows some improvement in wasting. Both Madhya Pradesh and Chhattisgarh have done better 

in all three nutritional indicators, suggesting a decline in wasting, stunting, and underweight. 

Uttar Pradesh shows a decline in underweight and stunting, but an increase in wasting, while 

Uttarakhand shows a decline in all three. For measles, BCG, tetanus and Polio vaccines we see a 

general improvement across all the old and new states except Uttar Pradesh.  The results in table 

11 further explain the results in tables 6-7 at a state level and we do not find any particular 

pattern for any states to be consistently skewing the results except Uttar Pradesh’s performance 

on vaccination. 

 

4.7.  Robustness Check  

 

The difference-in-difference model makes the counterfactual assumption that the treatment 

districts would grow at the same rate as the control districts if there were no intervention (i.e., 

division of states). While this assumption is not directly testable, we can test whether the 

treatment districts and the control districts were growing at the same rate in the pre-intervention 

 

 
 

131 

 

period. If we do find that they were growing at the same rate, it would suggest that our 

counterfactual assumption is likely to be correct. This is an indirect test of the validity of the 

difference-in-difference assumption that the observed change in the control districts is what 

would have happened in the treatment districts if there was no intervention. 

To test the parallel trends assumption, we use a similar specification as equation (1) but 

with time dummy variables. The regression model is specified in equation 4. 

 

(cid:5)Mg= μg+Jl+V> 1(cid:20)(cid:28)31@(cid:28)(cid:31)1∗(cid:10)(cid:19)41K|}}r+VC1(cid:20)(cid:28)31@(cid:28)(cid:31)1+Vt(cid:10)(cid:19)41K|}}r+w x=Mg+ \Mg      

(4) 

We have two rounds of data prior to the splitting of states, so we consider the first round 

as baseline and second round as end line. We assume a dummy post (round 2) and interact that 

term with treatment. The hypothesis of parallel trend is rejected if the coefficient V> is 

significantly different from zero.  

We present the regression results of testing parallel trend for the two difference-in-

difference specification models in Appendix E. For the first DiD model specification we report 

the regression results for the outcome variables and some covariates in tables B1 and B2. We 

find that on average the treatment districts in split states and comparison districts in adjacent 

states were growing at very similar rates prior to the split. For DiD specification model 2 , we 

find similar results that for household head education, electricity, piped drinking water, toilet, 

refrigerator, motorcycle, government institutional deliveries, infant deaths, underweight, stunted, 

wasted, tetanus shots, polio vaccines, BCG vaccines, health cards, iron tables and vitamin A to 

 

 
 

132 

 

children there is no significant difference in the trend between treatment and comparison districts 

prior to the split.   

For the regression discontinuity design we had done the balancing test (Appendix E) and 

do not need to do the parallel trend test. However, there is a possibility that districts across the 

newly formed borders might be of different sizes on each side leading to some sort of bias and a 

parallel trend test would make the results more credible. We present the results of this test in 

tables Appendix E. In this case too we find that the treatment and comparison districts were not 

changing at a significantly different rate from each other prior to the split. Based on these tests 

we thus fail to reject the hypothesis that both treatment and comparison districts were growing at 

a similar rate prior to the split. This increases the confidence in the appropriateness of the use of 

the DiD and RD identification strategy used in this analysis.  

 

4.8. Conclusion  

 

In a federal system like India, states play an important role in the development of its population 

and can be considered as a proximate determinant of institutions. In this study we tried to 

examine whether and how more homogenous states leads to more development. Usually it is 

very difficult to compare the effectiveness of institutions from different states as they are created 

by different historical forces. Splitting of states in the year 2000 provided an ideal natural 

experiment to isolate the impact of government institutions and programs on the development 

indicators. Once the border appears, only then we could spot the difference created by formation 

of new states, which earlier was a sub-region within the state. So, creation of this earlier non-

 

 
 

133 

 

existent state boundary line, allows us use quasi-experimental methods and flesh out the role of 

newly formed institutions on the developmental indicators of the population. We used quasi 

experimental methods of difference in difference and regression discontinuity to evaluate the 

impact of splitting of this states on the new and the parent state and we find that institutions do 

matter, as the results of this study clearly show how they have functioned to improve the lives of 

people in the newly formed states once they find more political representation.  The parent states 

are doing significantly better on health indicators (delivery of babies at government facilities, 

access to health card, vaccination, and child nutrition supplements). But on other indicators of 

living standards (e.g., amenities, public services, and consumer durables) by and large they have 

been worse off with the split compared to adjacent states. For example, on 18 out of 29 outcome 

indicators compared in this study, the parent states experienced, on average, a significant 

negative change in outcomes after the split compared to the change over the same time period in 

adjacent states.  

Overall, our findings suggest mixed results. It supports the argument that greater political 

representation and geographical cohesiveness makes the government institutions more 

responsive and effective leading to better service delivery and improving the lives of the regional 

sub groups, at least for the new state resulting from the split. However, such positive impacts are 

not experienced to the same extent by the parent state on all the development indicators.  For the 

parent state, the negative impact of losing control over revenue generating resources that may be 

found in areas carved out of the state may be thwarting any positive impact that could come from 

managing culturally homogeneous compact geographies.  

While on major development indicators like electricity, concrete housing, drinking water 

we find that only the new states have done better while the parent states have done worse, we 

 

 
 

134 

 

also find that on various health indicators like vaccines, child health variables (stunting, wasting, 

underweight) etc. the parent states have also improved, suggesting efficiency. The net aggregate 

welfare implications of the split for the affected population across both the parent and the new 

states thus remain ambiguous. This is an interesting research question that is outside the scope of 

this paper but needs to be addressed in future research. The literature has mostly focused on 

assessing the impact of the split on the new states, rather than on both the parent and the new 

states. The findings of potential negative impact of the split on the parent state points to the 

importance of considering this in future policy decisions on splitting the states not only in India 

but also in other countries with federal system.   

 

 

 

 

 

 

 

 

 

 

 

 
 

135 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

APPENDICES 

 
 

136 

 

APPENDIX A- FIGURES  
 

Figure 4.1: THREE STATE BEFORE AND AFTER SPLITTING. LEFT MAP 
CORRESPONDS TO THE THREE STATES BEFORE SPLITTING AND THE RIGHT MAP 
SHOWS THREE NEW AND THE THREE OLD STATES AFTER SPLITTING.   

 

 Source: Government of India website 

  

 

 

 

 

 
 

137 

 

Figure 4.2: NEWLY CREATED BOUNDARY LINES (IN YELLOW) SHARED BY BOTH 
NEW AND OLD STATES 

 

Source: Government of India website 

 

 

 

 

 
 

138 

 

Figure 4.3: ASPIRANT STATES OF INDIA REPRESENTING ALL THE CURRENT 
DEMANDS FOR NEW STATE FORMATION.  

 

Source: India Redrawn: Reorganization of State, 06Februrary 2012, Outlook magazine  

 

 

 

 
 

139 

 

APPENDIX B: TABLES- SUMMARY OF RESULTS 
 

Table 4. 1: ARGUMENTS FOR AND AGAINST SPLITTING A LARGE STATE INTO TWO 
AND THEIR POTENTIAL DEVELOPMENT IMPACTS ON THE PARENT STATE AND 
THE NEWLY CREATED SMALLER STATES 

 
Arguments for and against 
1.  Division unleashes suppressed growth potential  Neutral 
2.  Compact geographical entities create policies 
Positive 

Potential impact on: 
Parent state 

New state 
Positive 
Positive 

and environment conducive to local needs 

3.  Cultural homogeneity allows for better 

Positive 

Positive 

management and allocation of public resources 
4.  Smaller size means key decisions will be taken 

closer to the ground (efficiency in decision 
making) 

Neutral 

Positive 

5.  Increased vulnerability to the pressures of 

Neutral 

Negative 

corporations and the greed of newly emergent 
regional elite 

6.  Natural resources like coal and minerals etc. are 
mostly found in the areas carved out in the new 
states.  

Source:  Author’s own compilation 

Negative 

Positive 

 

 

 

 

 
 

140 

 

Table 4. 2: CHANGE IN DEVELOPMENTAL INDICATORS FROM YEAR 2001 TO YEAR 
2011 FOR BOTH PARENT AND NEW STATES 

                                                    Parent State 

Indicators       /     Year  

2001 

2011 

% increase 

                                New State 
 

2001 

2011 

% increase 

Per Capita Income 

  

  

  

  

  

  

  

Uttar Pradesh 

9,995 

30,071 

201 

Uttarakhand 

16,232 

85,372 

426 

Bihar 

6,200 

22,582 

264 

Jharkhand 

11,034 

36,554 

231 

Madhya Pradesh 

12,697 

37,979 

199 

Chhattisgarh 

12,170 

48,366 

297 

 

 

Electricity in % Households 

 

  

 

  

 

  

 

  

 

  

 

  

Uttar Pradesh 

31.90 

36.81 

15 

Uttarakhand 

60.33 

87.04 

44 

Bihar 

10.25 

16.36 

60 

Jharkhand 

24.30 

45.78 

88 

Madhya Pradesh 

69.98 

67.11 

-4 

Chhattisgarh 

53.10 

75.26 

42 

 

Literacy rate 

 

  

 

  

Uttar Pradesh 

73.19 

87.2 

Bihar 

63.25 

72.2 

Madhya Pradesh 

90.86 

94 

 

Toilet in % Households 

 

  

 

  

Uttar Pradesh 

19.2 

21.8 

Bihar 

13.9 

17.6 

Madhya Pradesh 

8.9 

13.1 

Source: Planning Commission of India 2012 

 

  

19 

14 

3 

 

  

14 

27 

47 

 

  

 

  

 

  

Uttarakhand 

56.27 

67.7 

Jharkhand 

55.52 

67.2 

Chhattisgarh 

47 

61.8 

 

  

 

  

 

  

Uttarakhand 

31.6 

54.1 

Jharkhand 

6.6 

7.6 

 

  

20 

21 

31 

 

  

71 

15 

Chhattisgarh 

5.2 

14.5 

179 

 

 

 

 
 

141 

 

Table 4. 3: TABLES OF MEAN COMPARISONS BETWEEN DISTRICTS ON BOTH SIDES 
OF BORDER SEPARATING THE NEWLY FORMED STATE AND PARENT STATE, 1998-
99 

N(New states)  Mean(New States) 

Mean 

Difference 

2132 

2132 

2129 

2132 

2132 

2132 

2132 

2132 

2132 

2131 

2132 

2132 

2093 

2132 

756 

756 

756 

756 

756 

756 

616 

616 

601 

756 

756 

756 

756 

756 

756 

675 

756 

0.168 

0.002 

0.089 

0.241 

0.355 

0.261 

0.206 

0.098 

0.472 

0.088 

0.014 

0.142 

0.557 

1.01 

1.87 

1.069 

44.352 

1.932 

7.47 

0.078 

0.518 

0.645 

0.265 

0.668 

0.026 

0.179 

0.372 

0.405 

0.44 

0.193 

0.384 

-0.001 

-0.002 

0.001 

0.006 

-0.063* 

-0.009 

-0.033 

-0.061 

-0.025 

-0.039 

-0.01 

0.081* 

-0.043 

-0.082* 

0.055 

-0.021 

1.03 

-0.2 

.093* 

-0.034 

0.018 

0.012 

0.006 

-0.072 

-0.007 

-0.001 

-0.002 

-0.047 

-0.183 

-0.068 

-0.014 

N(Parent 

states) 

2952 

2952 

2950 

2952 

2952 

2952 

2952 

2952 

2952 

2952 

2952 

2950 

2697 

2952 

1410 

1410 

1408 

1410 

1410 

1410 

960 

975 

930 

1410 

1410 

1410 

1410 

1410 

1410 

1228 

1410 

 

Variables  

Concrete Housing 

Piped Drinking Water 

Clean Cooking Fuel 

Toilet 

Electricity 

Radio 

Television 

Refrigerator 

Bicycle 

Motorcycle 

Car 

Percentage Hindu 

Percentage General Caste 

Education level of HH 

Percentage Rural Households 

% Female headed household 

Age of Household Head 

Women years of Education 

Household Size 

Government Institutional Delivery 

Underweight 

Stunted 

Wasted 

Tetanus Shot During Pregnancy 

Polio Shot 

Measles Vaccine 

BCG Vaccine 

Health Card 

Iron tablets to Pregnant women 
Vitamin A tablets to Pregnant 
women 
DPT Vaccine 

Source: DHS 1998-99 

 

 

Mean( 
Parent 
State) 

0.169 

0 

0.088 

0.247 

0.293 

0.252 

0.173 

0.037 

0.447 

0.049 

0.004 

0.223 

0.514 

0.928 

1.88 

1.048 

45.381 

1.864 

8.562 

0.044 

0.535 

0.633 

0.271 

0.596 

0.019 

0.178 

0.37 

0.357 

0.257 

0.125 

0.37 

 
 

142 

 

Table 4. 4: SUMMARY OF REGRESSION RESULTS FOR THE EFFECTS OF SPLITTING 
ON SELECTED DEVELOPMENTAL OUTCOMES: EDUCATION AND HOUSEHOLD 
DWELLING CHARACTERISTICS 

 

(1) 

(2) 

(3) 

Comparison: 

Adjacent states 

Parent state 

Neighboring districts in parent state 

 

 

 

Estimator: 

Outcome variable 

Impact on: 

HHH Education 

New state 

 

 

 

 

Parent state 

 

Electricity 

New state 

 

 

 

 

Parent state 

 

Piped drinking Water  New state 

 

 

 

Parent state 

 

Concrete house 

New state 

 

 

 

 

Parent state 

 

Toilet 

New state 

 

 

 

 

Parent state 

 

Clean Cooking Fuel 

New state 

 

 

 

Parent state 

 

DiD 

 

-0.000 

(0.083) 

0.030 

(0.043) 

0.161*** 

(0.005) 

0.054*** 

(0.002) 

0.074*** 

(0.004) 

-0.095*** 

(0.002) 

0.158*** 

(0.005) 

-0.208*** 

(0.003) 

0.095*** 

(0.005) 

-0.112*** 

(0.003) 

0.109*** 

(0.005) 

-0.070*** 

(0.003) 

DiD 

 

-0.054 

(0.079) 
-- 

 

RD 

 

0.219** 

(0.111) 
-- 

 

0.105*** 

0.082*** 

(0.005) 
-- 

 

(0.007) 
-- 

 

0.020*** 

0.033*** 

(0.003) 
-- 

 

0.048*** 

(0.005) 
-- 

 

0.010* 

(0.005) 
-- 

 

(0.004) 
-- 

 

0.006 

(0.007) 
-- 

 

0.035*** 

(0.007) 
-- 

 

0.043*** 

 0.062*** 

(0.004) 
-- 

 

(0.006) 
-- 

 

Source: Author’s calculations. 
Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard 
errors are clustered at the village level 
 

 

 

 
 

143 

 

Table 4. 5: SUMMARY OF REGRESSION RESULTS FOR THE EFFECTS OF SPLITTING 
ON SELECTED DEVELOPMENTAL OUTCOMES: DURABLE GOODS AND AMENITIES 

 

 

(1) 

(2) 

(3) 

Comparison: 

Adjacent states 

Parent state 

Neighboring districts in parent state 

 

Estimator: 

DiD 

Outcome variable 

Impact on: 

Internet 
 
 
 

Television 
 
 
 

Refrigerator 
 

 
 

Motorcycle 
 

 

 

Fan 
 
 

 

Mobile Phones 
 

 

 

Computer 
 
 

 

Washing Machine 

 
 

New state 
 
Parent state 
 

New state 
 
Parent state 
 

New state 
 

Parent state 
 

New state 
 

Parent state 

 

New state 
 
Parent state 

 

New state 
 

Parent state 

 

New state 
 
Parent state 

 

New state 
 
Parent state 
 

0.027*** 
(0.003) 
-0.010*** 

(0.002) 

0.100*** 
(0.006) 
-0.19*** 
(0.003) 

0.100*** 

(0.005) 
-0.106*** 
(0.003) 

0.064*** 

(0.006) 
-0.052*** 
(0.003) 

0.047*** 
(0.005) 

-0.10*** 

(0.003) 

0.139*** 

(0.023) 

-0.036** 
(0.014) 

-0.047** 
(0.019) 

-0.017 
(0.012) 

-0.267*** 
(0.021) 
-0.165*** 
(0.013) 

DiD 

 

0.038*** 
(0.003) 
-- 
 

0.091*** 
(0.005) 
-- 
 

0.005 

(0.004) 
-- 
 

-0.010** 

(0.005) 
-- 

 

0.051*** 
(0.005) 
-- 

 

0.025 

(0.023) 
-- 

 

0.000 
(0.016) 
-- 

 

-0.053*** 
(0.019) 
-- 
 

RD 

 

0.074*** 
(0.004) 
-- 
 

0.092*** 
(0.008) 
-- 
 

-0.001 

(0.006) 
-- 
 

0.037*** 

(0.007) 
-- 

 

0.041*** 
(0.008) 
-- 

 

0.178*** 

(0.011) 
-- 

 

0.043*** 
(0.007) 
-- 

 

0.051*** 
(0.009) 
-- 
 

Source: Author’s calculations. 
Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard 
errors are clustered at the village level 
 

 

 

 
 

144 

 

Table 4. 6: SUMMARY OF REGRESSION RESULTS FOR THE EFFECTS OF SPLITTING 
ON SELECTED DEVELOPMENTAL OUTCOMES: MAJOR HEALTH INDICATORS 

 

Comparison: 

 
Outcome variable 
Government Health 
Insurance 
 
 
Deliveries at 
government facilities 

 
Infant Death 
 
 
 
Underweight children 

 
 
Stunted children 
 
 
 
Wasted children 
 

 
 
Health Card 
 
 
 
Compensation for 
Sterilization 
 
 

 

(1) 

(2) 

(3) 

Adjacent states 

Parent state 

Estimator: 
Impact on: 
New state 
 
Parent state 
 
New state 
 
Parent state 
 
New state 
 
Parent state 
 
New state 
 
Parent state 
 
New state 
 
Parent state 
 
New state 
 

Parent state 
 
New state 
 
Parent state 
 
New state 
 
Parent state 
 

DiD 
 
0.081*** 
(0.004) 
-0.194*** 
(0.002) 
0.097*** 
(0.010) 
0.092*** 
(0.005) 
-0.02*** 
(0.005) 
-0.004 
(0.002) 
-0.041*** 
(0.012) 
-0.041*** 
(0.006) 
-0.029** 
(0.013) 
-0.010 
(0.006) 
-0.002 

(0.012) 
-0.07*** 
(0.006) 
0.168*** 
(0.010) 
0.094*** 
(0.005) 
0.032*** 
(0.006) 
-0.064*** 
(0.003) 

DiD 
 
0.274*** 
(0.003) 
-- 
 
0.091*** 
(0.005) 
-- 
 
0.005 
(0.004) 
-- 
 
-0.010** 
(0.005) 
-- 
 
0.051*** 
(0.005) 
-- 
 
0.025 

(0.023) 
-- 
 
0.000 
(0.016) 
-- 
 
-0.053*** 
(0.019) 
-- 
 

Neighboring districts in parent 
state 
RD 
 
0.204*** 
(0.005) 
-- 
 
0.061*** 
(0.012) 
-- 
 
-0.016** 
(0.007) 
-- 
 
-0.026* 
(0.015) 
-- 
 
-0.06*** 
(0.016) 
-- 
 
-0.06*** 

(0.015) 
-- 
 
0.142*** 
(0.013) 
-- 
 
0.026*** 
(0.007) 
-- 
 

Source: Author’s calculations. 
Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard 
errors are clustered at the village level 
 
 
 

 

 

 
 

145 

 

Table 4. 7: SUMMARY OF REGRESSION RESULTS FOR THE EFFECTS OF SPLITTING 
ON SELECTED DEVELOPMENTAL OUTCOMES: GOVERNMENT HEALTH PROGRAM 
INDICATORS FOR CHILDREN AND PREGNANT WOMEN 

 

(1) 

(2) 

(3) 

 

Comparison: 

 
Outcome variable 
Tetanus shot 

 
 
Polio Vaccine 

 
 
Measles vaccine 
 
 
 
BCG vaccine 
 
 
 
Iron tablets 
 
 
 
Vitamin A tablets 
 

Estimator: 
Impact on: 
New state 
 
Parent state 
 
New state 
 
Parent state 
 
New state 
 
Parent state 
 
New state 
 
Parent state 
 
New state 
 
Parent state 
 
New state 
 

Adjacent states 

Parent state 

DiD 
 
0.194*** 
(0.010) 
0.165*** 
(0.005) 
0.072*** 
(0.010) 
-0.022*** 
(0.005) 
0.149*** 
(0.010) 
0.128*** 
(0.005) 
0.196*** 
(0.008) 
0.217*** 
(0.004) 
0.195*** 
(0.010) 
0.160*** 
(0.005) 
-0.016 

DiD 
 
0.029*** 
(0.010) 
-- 
 
0.091*** 
(0.010) 
-- 
 
0.020** 
(0.010) 
-- 
 
-0.019** 
(0.008) 
-- 
 
0.032*** 
(0.010) 
-- 
 
0.028*** 

Neighboring districts in parent 
state 
RD 
 
-0.019 
(0.013) 
-- 
 
0.035*** 
(0.013) 
-- 
 
0.034*** 
(0.013) 
-- 
 
0.067*** 
(0.010) 
-- 
 
-0.018 
(0.013) 
-- 
 
0.043*** 

 
 
DPT vaccine 
 
 
 

Parent state 
 
New state 
 
Parent state 
 
Source: Author’s calculations. 
Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard 
errors are clustered at the village level 
 

 

(0.010) 
0.011** 
(0.005) 
0.197*** 
(0.009) 
0.190*** 
(0.004) 

(0.010) 
-- 
 
0.008 
(0.009) 
-- 
 

(0.013) 
-- 
 
0.031*** 
(0.011) 
-- 
 

 

 
 

146 

 

Table 4. 8: SUMMARY OF STATE-WISE REGRESSION RESULTS FOR THE EFFECTS OF 
SPLITTING ON SELECTED DEVELOPMENTAL OUTCOMES OF NEW STATES AS 
COMPARED TO THE PARENT STATES 

                                       (1)                               (2)                          (3)                            (4)                         (5)                         (6) 

New State 

Jharkhand 

Chhattisgarh 

Uttarakhand 

Comparison 

Bihar 

Estimator 
Outcome 
Variable 

DiD 

 

 

Bordering 
Districts of 
Bihar 

Madhya Pradesh 

Bordering 
Districts of 
MP 

Uttar Pradesh 

Bordering 
Districts of 
UP 

RD 

DiD 

RD 

DiD 

RD 

  

Electricity 

0.11*** 

0.13*** 

0.19*** 

0.08*** 

0.07*** 

 

Piped Drinking 
Water 
 

(0.009) 

0.04*** 

(0.004) 

(0.012) 

(0.006) 

(0.012) 

(0.013) 

0.01** 

0.09*** 

0.02*** 

0.24*** 

(0.004) 

(0.005) 

(0.007) 

(0.008) 

Concrete House 

0.04*** 

-0.00 

0.02** 

0.13*** 

0.08*** 

 

Toilet 

 

Clean Cooking 
Fuel 
 

(0.008) 

0.11*** 

(0.008) 

0.03*** 

(0.007) 

(0.011) 

(0.007) 

(0.012) 

(0.012) 

0.09*** 

0.05*** 

0.17*** 

0.24*** 

(0.010) 

(0.007) 

(0.012) 

(0.012) 

0.05*** 

0.02*** 

0.03*** 

0.01 

(0.008) 

(0.006) 

(0.009) 

(0.011) 

Television 

0.04*** 

-0.01 

0.12*** 

0.21*** 

0.16*** 

 

Internet 

 

(0.008) 

-0.00 

(0.005) 

(0.011) 

(0.008) 

(0.014) 

(0.014) 

-0.02*** 

0.00 

0.07*** 

0.19*** 

(0.006) 

(0.005) 

(0.008) 

(0.008) 

Refrigerator 

0.03*** 

-0.00 

0.00 

0.05*** 

0.10*** 

 

(0.005) 

(0.007) 

(0.006) 

(0.009) 

(0.011) 

Motorbike 

0.05*** 

0.03*** 

-0.02** 

0.12*** 

-0.03** 

  

(0.008) 

(0.010) 

(0.008) 

(0.013) 

(0.013) 

-0.01 

(0.014) 

0.26*** 

(0.011) 

0.06*** 

(0.016) 

-0.02 

(0.014) 

0.06*** 

(0.015) 

0.07*** 

(0.017) 

0.28*** 

(0.009) 

0.06*** 

(0.016) 

-0.01 

(0.017) 

Source: Author’s calculations. 
Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard 
errors are clustered at the village level 
 

 

 

 

 

 

 

 

 
 

147 

 

Table 4. 9: SUMMARY OF STATE-WISE REGRESSION RESULTS FOR THE EFFECTS OF 
SPLITTING ON SELECTED DEVELOPMENTAL OUTCOMES IN COMPARISON TO THE 
ADJACENT STATES USING DID ESTIMATOR 

                                             (1)                    (2)                    (3)                          (4)                          (5)                         (6) 

Outcome Variables  

Jharkhand 

Bihar 

Chhattisgarh 

Madhya 
Pradesh 

Uttarakhand 

Uttar Pradesh 

Electricity 

-0.01 

-0.12*** 

0.09*** 

-0.10*** 

0.17*** 

-0.10*** 

 

(0.009) 

(0.005) 

(0.006) 

(0.004) 

(0.011) 

(0.004) 

Piped Drinking Water 

0.01*** 

-0.02*** 

0.09*** 

-0.00 

0.08*** 

-0.16*** 

 

(0.004) 

(0.002) 

(0.006) 

(0.004) 

(0.009) 

(0.003) 

Concrete House 

0.9*** 

-0.12*** 

0.14*** 

-0.15*** 

0.09*** 

-0.17*** 

 

Toilet 

 

(0.009) 

(0.006) 

(0.008) 

(0.005) 

(0.012) 

(0.005) 

0.19*** 

-0.10*** 

0.05*** 

-0.10*** 

0.06*** 

-0.19*** 

(0.009) 

(0.005) 

(0.008) 

(0.005) 

(0.011) 

(0.004) 

Clean Cooking Fuel 

0.05*** 

-0.03*** 

0.14*** 

-0.11*** 

0.01 

0.00 

 

(0.007) 

(0.005) 

(0.007) 

(0.005) 

(0.011) 

(0.004) 

Television 

0.21*** 

 -0.25*** 

0.03*** 

-0.15*** 

  -0.00 

-0.16*** 

 

Internet 

 

(0.005) 

(0.003) 

(0.004) 

(0.003) 

(0.008) 

(0.003) 

-0.02*** 

- 0.01*** 

0.08*** 

0.07*** 

0.11*** 

-0.08*** 

(0.009) 

(0.006) 

(0.008) 

(0.005) 

(0.013) 

(0.005) 

Refrigerator 

0.07*** 

- 0.09*** 

0.08*** 

-0.08*** 

0.08*** 

-0.19*** 

 

(0.006) 

(0.004) 

(0.007) 

(0.005) 

(0.011) 

(0.004) 

Motorcycle 

0.00 

- 0.05*** 

-0.04*** 

-0.02*** 

0.15*** 

-0.12*** 

  

(0.008) 

(0.005) 

(0.008) 

(0.005) 

(0.013) 

(0.005) 

Source: Author’s calculations. 
Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard 
errors are clustered at the village level 
 

 

 

 

 

 

 

 

 

 

 

 

 
 

148 

 

Table 4. 10: SUMMARY OF REGRESSION STATE-WISE RESULTS FOR THE EFFECTS 
OF SPLITTING ON SELECTED HEALTH OUTCOMES OF NEW STATES AS COMPARED 
TO THE PARENT STATES 

                                          (1)                        (2)                       (3)                                   (4)                      (5)                              
(6) 
New State 

Chhattisgarh 

Jharkhand 

Uttarakhand 

Bordering 
Districts of 
Bihar 

Madhya Pradesh 

Bordering 
Districts of 
MP 

Uttar Pradesh 

RD 

DiD 

RD 

DiD 

Bordering 
Districts of 
UP 

RD 

Comparison 

Bihar 

DiD 

  

Estimator 
Outcome 
Variables 

Deliveries at 
govt. facilities 
 

Infant Death 
 

Underweight 
children 
 

Stunted children 
 

Wasted children 
 

0.05*** 

-0.01 

0.05*** 

-0.02 

0.05* 

(0.016) 

-0.03*** 

(0.008) 

-0.14*** 

(0.019) 

-0.07*** 

(0.019) 

-0.12*** 

(0.017) 

(0.019) 

-0.02* 

(0.010) 

(0.012) 

-0.03*** 

(0.007) 

(0.024) 

-0.02* 

(0.013) 

(0.025) 

0.02 

(0.015) 

-0.16*** 

-0.09*** 

-0.06** 

-0.05 

(0.023) 

(0.015) 

-0.12*** 

-0.09*** 

(0.022) 

(0.022) 

-0.14*** 

-0.03 

(0.020) 

(0.020) 

(0.029) 

-0.03 

(0.041) 

(0.031) 

-0.14*** 

(0.031) 

-0.12*** 

-0.12*** 

(0.039) 

(0.025) 

Health Card 

0.09*** 

0.01 

-0.04*** 

0.12*** 

-0.03 

 

Tetanus shot 
 

Polio Vaccine 
 

Measles vaccine 
 

BCG vaccine 
 

 

 

(0.016) 

0.05*** 

(0.017) 

0.01 

(0.017) 

0.01 

(0.016) 

0.02* 

(0.013) 
 

 

(0.020) 

0.06*** 

(0.020) 

0.01 

(0.020) 

0.01 

(0.019) 

-0.01 

(0.015) 
 

 

(0.011) 

0.06*** 

(0.013) 

0.12*** 

(0.013) 

0.00 

(0.013) 

0.02* 

(0.011) 
 

 

(0.024) 

0.06*** 

(0.025) 

0.12*** 

(0.024) 

0.06*** 

(0.025) 

0.06*** 

(0.019) 
 

 

(0.027) 

-0.02 

(0.027) 

-0.00 

(0.026) 

0.10*** 

(0.027) 

0.14*** 

(0.024) 
 

 

-0.01 

(0.028) 

0.01 

(0.017) 

-0.07** 

(0.036) 

-0.04 

(0.036) 

0.03 

(0.030) 

0.02 

(0.032) 

0.20*** 

(0.032) 

0.08*** 

(0.030) 

0.21*** 

(0.031) 

0.28*** 

(0.027) 
 

 

Source: Author’s calculations. 
Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and 
standard errors are clustered at the village level 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

149 

 

Table 4. 11: SUMMARY OF REGRESSION STATE-WISE RESULTS FOR THE EFFECTS OF 
SPLITTING ON SELECTED DEVELOPMENTAL OUTCOMES IN COMPARISON TO THE 
ADJACENT STATES USING D-I-D ESTIMATOR. 

                                                      (1)                     (2)                  (3)                         (4)                     (5)                       (6) 
Uttar 
Pradesh 

Madhya 
Pradesh 

VARIABLES 

Uttarakhand 

Jharkhand 

Bihar 

Chhattisgarh 

 
Deliveries at govt. facilities 

 
Infant Death 

 
Underweight children 

 
0.05*** 

(0.017) 

 
0.01 

 
0.15*** 

(0.010) 

(0.013) 

-0.03*** 

0.01* 

0.02*** 

(0.008) 

(0.005) 

(0.007) 

 
0.20*** 

(0.008) 

-0.00 

(0.004) 

 
0.18*** 

(0.024) 

0.02 

(0.014) 

 
-0.14*** 

(0.008) 

0.00 

(0.005) 

-0.14*** 

0.01 

-0.13*** 

-0.04*** 

-0.13*** 

-0.08*** 

 
Stunted children 

 
Wasted children 

 
Health Card 

 
Tetanus shot 

 
Polio Vaccine 

 
Measles vaccine 
 

(0.020) 

-0.05** 

(0.021) 

(0.011) 

(0.016) 

-0.00 

-0.09*** 

(0.014) 

(0.023) 

-0.09*** 

-0.03*** 

-0.05*** 

(0.019) 

(0.013) 

(0.020) 

0.10*** 

0.02*** 

0.15*** 

(0.017) 

(0.010) 

(0.013) 

0.19*** 

0.14*** 

0.13*** 

(0.017) 

(0.010) 

(0.014) 

(0.010) 

-0.04** 

(0.014) 

-0.03** 

(0.013) 

0.19*** 

(0.008) 

0.06*** 

(0.009) 

0.10*** 

-0.12*** 

0.13*** 

0.01 

(0.017) 

(0.010) 

(0.014) 

0.12*** 

0.11** 

0.09*** 

(0.017) 

(0.010) 

(0.014) 

(0.029) 

-0.16*** 

(0.030) 

0.04* 

(0.025) 

0.01 

(0.025) 

0.00 

(0.026) 

0.08*** 

(0.025) 

0.08*** 

(0.026) 

-0.10*** 

(0.022) 

 

(0.010) 

-0.02** 

(0.010) 

-0.07*** 

(0.008) 

0.04*** 

(0.008) 

0.02** 

(0.008) 

-0.08*** 

(0.008) 

0.02** 

(0.008) 

0.04*** 

(0.007) 

 

(0.009) 

0.09*** 

(0.009) 

0.11*** 

(0.007) 

 

BCG vaccine 

0.22*** 

0.25*** 

0.09*** 

 
 

(0.014) 

(0.008) 

(0.011) 

 

 

 

Source: Author’s calculations. 
Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and 
standard errors are clustered at the village level 
 

 

 

 

 

 

 

 

 

 

 
 

150 

 

APPENDIX C: TABLES – DETAILED RESULTS 
 

Table 4. 12: DEVELOPMENTAL INDICATORS- PARENT STATE AS COMPARISON AND NEWLY FORMED STATES AS 
TREATMENT 

  

(1) 

(2) 

(3) 

VARIABLES 

  

Districts in New states X Post 

Post 

Treatment 

 

 

 

HHH Education 

Electricity 

Piped 

Drinking 

water 

  

-0.054 

(0.079) 

  

  

0.105*** 

0.020*** 

(0.005) 

(0.003) 

1.289*** 

0.345*** 

0.067*** 

(0.031) 

0.598 

(0.565) 

(0.002) 

(0.001) 

-0.084** 

0.310*** 

(0.037) 

(0.022) 

Age of head of household 

-0.056*** 

0.001*** 

0.001*** 

 

Sex of head of household  
female 

(Base = Male) 
Education of the Household 
head 

 

 

 

(0.001) 

(0.000) 

(0.000) 

-3.790*** 

0.026*** 

(0.035) 

(0.002) 

0.002 

(0.001) 

0.010*** 

0.002*** 

(0.000) 

(0.000) 

Household Size 

-0.026*** 

0.011*** 

-0.001*** 

 

Rural 

(Base=urban) 

(0.004) 

(0.000) 

(0.000) 

-3.101*** 

-0.218*** 

-0.170*** 

(0.031) 

(0.002) 

(0.001) 

Religion of the HH (Muslim) 

-2.573*** 

0.034*** 

0.005*** 

(Base= Hindu) 

(0.037) 

(0.002) 

(0.001) 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(4) 

(5) 

(6) 

(7) 

Concrete 

House 

Toilet 

Clean 

Government Health 

Cooking Fuel 

Insurance 

  

0.048*** 

(0.005) 

  

0.010* 

(0.005) 

  

  

0.043*** 

0.274*** 

(0.004) 

(0.003) 

0.103*** 

0.170*** 

0.126*** 

0.084*** 

(0.002) 

-0.041 

(0.035) 

(0.002) 

-0.019 

(0.035) 

(0.002) 

(0.001) 

0.143*** 

-0.074*** 

(0.031) 

(0.024) 

0.002*** 

0.003*** 

0.001*** 

0.000*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

0.025*** 

0.022*** 

0.027*** 

-0.004*** 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

0.013*** 

0.015*** 

0.013*** 

-0.001*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

0.003*** 

0.004*** 

-0.005*** 

0.002*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

-0.427*** 

-0.449*** 

-0.473*** 

0.017*** 

(0.002) 

(0.002) 

(0.002) 

0.010*** 

0.133*** 

-0.008*** 

(0.002) 

(0.002) 

(0.002) 

(0.001) 

-0.002 

(0.002) 

 

 
 

151 

 

 

 

 

 

 

 

 

Table 4. 12 (cont’d) 
 
 
Religion of the HH (Christian) 
 
 

2.382*** 

-0.016 

-0.004 

(0.169) 

(0.011) 

(0.006) 

Religion of the HH (Sikh) 

0.879*** 

0.100*** 

0.060*** 

 

(0.232) 

Religion of the HH (Others) 

0.838*** 

(0.132) 

(0.015) 

0.001 

(0.009) 

(0.009) 

-0.005 

(0.005) 

 

Caste of the 
household(Scheduled Caste) 

(Base= General Caste) 
Caste of the 
household(Scheduled Tribe) 

 

 

Constant 

Observations 

-1.895*** 

-0.050*** 

-0.014*** 

(0.030) 

(0.002) 

(0.001) 

-2.350*** 

-0.076*** 

-0.026*** 

(0.044) 

(0.003) 

12.555*** 

0.612*** 

(0.326) 

264,833 

(0.021) 

264,833 

(0.002) 

-0.010 

(0.013) 

264,833 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.047*** 

0.096*** 

0.061*** 

-0.038*** 

(0.011) 

(0.010) 

(0.009) 

0.138*** 

0.083*** 

0.144*** 

(0.014) 

(0.014) 

(0.013) 

(0.007) 

-0.011 

(0.010) 

-0.030*** 

0.024*** 

0.054*** 

-0.032*** 

(0.008) 

(0.008) 

(0.007) 

(0.006) 

-0.078*** 

-0.098*** 

-0.069*** 

0.020*** 

(0.002) 

(0.002) 

(0.002) 

(0.001) 

-0.123*** 

-0.127*** 

-0.087*** 

0.018*** 

(0.003) 

(0.003) 

(0.002) 

0.589*** 

0.602*** 

0.366*** 

(0.020) 

264,833 

(0.020) 

264,833 

(0.018) 

264,516 

(0.002) 

-0.024* 

(0.014) 

264,833 

R-squared 
Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village 
level 
 

0.418 

0.236 

0.430 

0.155 

0.344 

0.190 

0.348 

 

 
 

152 

 

Table 4. 13: DEVELOPMENTAL INDICATORS-ADJACENT STATES AS COMPARISON AND BOTH PARENT AND NEW 
STATES AS TREATMENTS 

 

VARIABLES 

  

New States X Post 

Parent State X Post 

Post 

New States 

Parent States 

 

 

 

 

 

Age of head of household 

 

Sex of head of household  female 

(Base = Male) 

Education of the Household head 

 

 

Household Size 

Rural 

(Base=urban) 

(1) 

(2) 

HHH Education 

Electricity 

  

-0.000 

(0.083) 

0.030 

(0.043) 

1.206*** 

(0.028) 

  

0.161*** 

(0.005) 

0.054*** 

(0.002) 

0.295*** 

(0.002) 

-2.043*** 

-0.254*** 

(0.395) 

-0.085 

(0.200) 

-0.061*** 

(0.001) 

-3.612*** 

(0.026) 

 

 

(0.023) 

-0.017 

(0.012) 

0.001*** 

(0.000) 

0.007*** 

(0.002) 

0.007*** 

(0.000) 

-0.053*** 

0.010*** 

-0.001*** 

0.004*** 

0.003*** 

-0.005*** 

(0.003) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

-3.051*** 

-0.136*** 

-0.147*** 

-0.335*** 

-0.371*** 

-0.434*** 

(0.021) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

Religion of the HH (Muslim) 

-2.380*** 

0.006*** 

-0.005*** 

-0.009*** 

0.084*** 

-0.026*** 

(Base= Hindu) 

Religion of the HH (Christian) 

Religion of the HH (Sikh) 

 

Religion of the HH (Others) 

 

(0.029) 

0.861*** 

(0.081) 

0.375*** 

(0.141) 

0.546*** 

(0.002) 

0.005 

(0.005) 

0.055*** 

(0.008) 

0.005 

(0.002) 

0.021*** 

(0.005) 

0.083*** 

(0.009) 

0.022*** 

(0.002) 

0.010* 

(0.005) 

0.081*** 

(0.009) 

0.048*** 

(0.002) 

0.010** 

(0.005) 

0.116*** 

(0.008) 

0.036*** 

(0.001) 

0.004 

(0.004) 

0.022*** 

(0.007) 

0.015*** 

 
 

153 

 

(6) 

(7) 

Clean 

Government Health 

Cooking Fuel 

Insurance 

(3) 

Piped 

Drinking 

water 

  

0.074*** 

(0.004) 

(4) 

Concrete 

House 

  

0.158*** 

(0.005) 

(5) 

Toilet 

  

0.095*** 

(0.005) 

  

0.109*** 

(0.005) 

-0.095*** 

-0.208*** 

-0.112*** 

-0.070*** 

(0.002) 

(0.003) 

(0.003) 

(0.003) 

0.160*** 

0.315*** 

0.282*** 

0.196*** 

(0.001) 

0.023 

(0.018) 

(0.002) 

(0.002) 

(0.002) 

-0.240*** 

-0.276*** 

0.071*** 

(0.026) 

(0.025) 

(0.023) 

-0.031*** 

0.061*** 

-0.081*** 

0.045*** 

(0.009) 

(0.013) 

(0.013) 

(0.012) 

0.001*** 

0.001*** 

0.003*** 

0.001*** 

(0.000) 

-0.000 

(0.001) 

0.004*** 

(0.000) 

(0.000) 

0.012*** 

(0.002) 

0.012*** 

(0.000) 

(0.000) 

0.006*** 

(0.002) 

0.014*** 

(0.000) 

(0.000) 

0.017*** 

(0.002) 

0.014*** 

(0.000) 

  

0.081*** 

(0.004) 

-0.194*** 

(0.002) 

0.279*** 

(0.001) 

0.336*** 

(0.020) 

0.052*** 

(0.010) 

0.000*** 

(0.000) 

-0.003** 

(0.001) 

-0.002*** 

(0.000) 

0.003*** 

(0.000) 

0.052*** 

(0.001) 

0.003* 

(0.001) 

0.031*** 

(0.004) 

0.004 

(0.007) 

-0.022*** 

 

Table 4. 13 (cont’d) 
 
 

(0.063) 

(0.004) 

(0.003) 

(0.004) 

(0.004) 

(0.004) 

Caste of the household(Scheduled Caste) 

-1.830*** 

-0.042*** 

-0.018*** 

-0.084*** 

-0.092*** 

-0.075*** 

(Base= General Caste) 

(0.022) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

Caste of the household(Scheduled Tribe) 

-2.261*** 

-0.088*** 

-0.043*** 

-0.156*** 

-0.165*** 

-0.126*** 

Constant 

Observations 

 

 

(0.031) 

12.554*** 

(0.329) 

499,138 

(0.002) 

0.609*** 

(0.019) 

499,138 

(0.001) 

0.003 

(0.015) 

499,138 

(0.002) 

0.410*** 

(0.022) 

499,138 

(0.002) 

0.587*** 

(0.021) 

499,138 

(0.002) 

0.271*** 

(0.019) 

497,520 

(0.003) 

0.029*** 

(0.001) 

0.020*** 

(0.002) 

-0.068*** 

(0.017) 

499,138 

R-squared 
0.307 
Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village level 
 

0.429 

0.452 

0.161 

0.346 

0.241 

0.394 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

154 

 

Table 4. 14: DEVELOPMENTAL INDICATORS- RDD USING DISTRICTS ON THE NEWLY FORMED STATES AS 
TREATMENT AND PARENT STATE AS COMPARISON 

  

(1) 

(2) 

HHH Education 

Electricity 

 
 

VARIABLES 
  
New State Border Dist. # Post 

Post Split 

 

 

District on the New State Side 

Age of head of household 

 

 

Sex of head of household  female 
(Base = Male) 
Education of the Household head 

Household Size 

 

 

Rural 
(Base=urban) 
Religion of the HH (Muslim) 
(Base= Hindu) 
Religion of the HH (Christian) 

Religion of the HH (Sikh) 

 

Religion of the HH (Others) 

 

 

Caste of the household(Scheduled 
Caste) 
(Base= General Caste) 
Caste of the household(Scheduled 
Tribe) 

Constant 

 

 

  

0.219** 
(0.111) 
0.946*** 
(0.069) 

-1.595*** 

(0.219) 

-0.057*** 

(0.002) 

-3.658*** 

(0.067) 

-0.030*** 

(0.009) 

-2.958*** 

(0.057) 

-2.970*** 

(0.067) 
1.633*** 
(0.277) 
0.882*** 
(0.336) 
0.396 
(0.256) 

-1.737*** 

(0.060) 

-2.106*** 

(0.076) 

11.882*** 

(0.200) 

  

0.082*** 
(0.007) 
0.377*** 
(0.004) 
0.285*** 
(0.014) 
0.001*** 
(0.000) 
0.015*** 
(0.004) 
0.010*** 
(0.000) 
0.009*** 
(0.001) 

-0.220*** 

(0.004) 
0.045*** 
(0.004) 
-0.009 
(0.018) 
0.142*** 
(0.022) 
-0.028* 
(0.016) 

-0.043*** 

(0.004) 

-0.078*** 

(0.005) 
0.230*** 
(0.013) 

(4) 

Concrete 

House 

  

0.006 
(0.007) 
0.142*** 
(0.004) 
-0.08*** 
(0.014) 
0.002*** 
(0.000) 
0.034*** 
(0.004) 
0.015*** 
(0.000) 
0.002*** 
(0.001) 
-0.42*** 
(0.004) 
0.031*** 
(0.004) 
0.035** 
(0.018) 
0.155*** 
(0.022) 
-0.019 
(0.017) 

-0.09*** 
(0.004) 

-0.13*** 
(0.005) 
0.358*** 
(0.013) 

(5) 

Toilet 

  

0.035*** 
(0.007) 
0.221*** 
(0.004) 
0.039*** 
(0.014) 
0.003*** 
(0.000) 
0.031*** 
(0.004) 
0.015*** 
(0.000) 
0.003*** 
(0.001) 
-0.39*** 
(0.004) 
0.171*** 
(0.004) 
0.118*** 
(0.017) 
0.093*** 
(0.021) 
0.016 
(0.016) 

-0.1*** 
(0.004) 

-0.12*** 
(0.005) 
0.218*** 
(0.013) 

(6) 

(7) 

Clean Cooking 

Government 

Fuel 

  

 0.062*** 

(0.006) 
0.139*** 
(0.004) 
0.060*** 
(0.012) 
0.001*** 
(0.000) 
0.035*** 
(0.004) 
0.013*** 
(0.000) 

-0.007*** 

(0.000) 

-0.421*** 

(0.003) 
0.000 
(0.004) 
0.060*** 
(0.015) 
0.131*** 
(0.018) 
0.036*** 
(0.014) 

-0.061*** 

(0.003) 

-0.073*** 

(0.004) 
0.317*** 
(0.011) 

Health Insurance 

  

0.204*** 
(0.005) 
0.102*** 
(0.003) 
0.195*** 
(0.011) 
0.000*** 
(0.000) 

-0.009*** 

(0.003) 

-0.001*** 

(0.000) 
0.004*** 
(0.000) 
0.023*** 
(0.003) 

-0.012*** 

(0.003) 
-0.023* 
(0.014) 
0.007 
(0.016) 
-0.018 
(0.012) 

0.013*** 
(0.003) 

0.020*** 
(0.004) 
0.026*** 
(0.010) 

(3) 

Piped 

Drinking 

water 

  

0.033*** 
(0.004) 
0.050*** 
(0.002) 
0.036*** 
(0.008) 
0.000*** 
(0.000) 
0.005* 
(0.002) 
0.002*** 
(0.000) 
-0.00*** 
(0.000) 
-0.16*** 
(0.002) 
0.001 
(0.002) 
0.014 
(0.010) 
0.011 
(0.012) 
-0.017* 
(0.009) 

-0.01*** 
(0.002) 

-0.01*** 
(0.003) 
0.083*** 
(0.007) 

 
 

155 

 

 

 

 

 

 

 

 

 

 

Table 4. 14 (cont’d) 
 
 
Observations 
R-squared 

71,751 
0.141 

71,751 
0.329 

71,751 
0.219 

71,751 
0.341 

71,751 
0.424 

71,702 
0.392 

71,751 
0.271 

       Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village level 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

156 

 

Table 4. 15: CONSUMER DURABLES - PARENT STATE AS COMPARISON AND NEWLY FORMED STATES AS 
TREATMENT 

  

(1) 

(2) 

(3) 

(4) 

(5) 

(6) 

(7) 

(8) 

VARIABLES 

Internet 

Television 

Refrigerator 

Motorcycle 

Fan 

Mobile 
Phone 

Computer 

Washing 
Machine 

  

  

  

  

  

  

  

  

  

Districts in New states X Post 

0.038*** 

0.091*** 

0.005 

-0.010** 

0.051*** 

0.025 

0.000 

-0.053*** 

 
Post 

 
Treatment 

 
Age of head of household 

 
Sex of head of household  
female 

(0.003) 

(0.005) 

(0.004) 

(0.005) 

(0.005) 

(0.023) 

(0.016) 

(0.019) 

0.078*** 

0.204*** 

0.088*** 

0.219*** 

0.303*** 

(0.001) 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

0.176*** 

0.081** 

0.132*** 

-0.099*** 

-0.350*** 

(0.023) 

(0.038) 

(0.029) 

(0.037) 

(0.039) 

 

 

 

 

 

 

 

 

 

0.001*** 

0.002*** 

0.002*** 

0.002*** 

0.001*** 

 
-0.002*** 

 
0.001*** 

 
0.002*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

0.011*** 

0.014*** 

0.021*** 

-0.018*** 

0.021*** 

-0.018*** 

0.020*** 

0.026*** 

(Base = Male) 

(0.001) 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

(0.001) 

(0.002) 

Education of the Household 
head 

 
Household Size 

 
Rural 

0.006*** 

0.015*** 

0.010*** 

0.015*** 

0.014*** 

0.007*** 

0.007*** 

0.009*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

0.003*** 

0.020*** 

0.006*** 

0.020*** 

0.015*** 

0.026*** 

0.003*** 

0.004*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

-0.106*** 

-0.333*** 

-0.263*** 

-0.128*** 

-0.310*** 

-0.057*** 

-0.102*** 

-0.158*** 

(Base=urban) 

(0.001) 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

(0.001) 

(0.001) 

Religion of the HH (Muslim) 

-0.024*** 

-0.077*** 

0.015*** 

-0.044*** 

0.039*** 

-0.001 

-0.031*** 

0.020*** 

 

 
 

157 

 

Table 4. 15 (cont’d) 
 

(Base= Hindu) 

(0.001) 

(0.003) 

(0.002) 

(0.002) 

(0.003) 

(0.002) 

(0.001) 

(0.002) 

Religion of the HH (Christian) 

0.027*** 

0.029** 

0.062*** 

0.072*** 

0.023* 

0.056*** 

0.041*** 

0.049*** 

 
Religion of the HH (Sikh) 

 
Religion of the HH (Others) 

 
Caste of the 
household(Scheduled Caste) 

(0.007) 

(0.011) 

(0.009) 

(0.011) 

(0.012) 

(0.009) 

(0.007) 

(0.008) 

0.069*** 

0.162*** 

0.244*** 

0.183*** 

0.122*** 

0.054*** 

0.119*** 

0.245*** 

(0.009) 

(0.016) 

(0.012) 

(0.015) 

(0.016) 

(0.013) 

(0.010) 

(0.011) 

0.012** 

-0.006 

0.060*** 

0.034*** 

-0.023** 

-0.009 

0.022*** 

0.049*** 

(0.005) 

(0.009) 

(0.007) 

(0.009) 

(0.009) 

(0.007) 

(0.005) 

(0.006) 

-0.028*** 

-0.071*** 

-0.051*** 

-0.093*** 

-0.094*** 

-0.043*** 

-0.023*** 

-0.036*** 

(Base= General Caste) 

(0.001) 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

(0.001) 

(0.001) 

Caste of the 
household(Scheduled Tribe) 

 
Constant 

 
Observations 

R-squared 

-0.030*** 

-0.134*** 

-0.054*** 

-0.105*** 

-0.179*** 

-0.116*** 

-0.023*** 

-0.023*** 

(0.002) 

(0.003) 

(0.002) 

(0.003) 

(0.003) 

(0.003) 

(0.002) 

(0.002) 

-0.087*** 

0.319*** 

0.094*** 

-0.123*** 

0.501*** 

0.904*** 

0.036*** 

0.113*** 

(0.013) 

(0.022) 

(0.017) 

(0.021) 

(0.023) 

(0.007) 

(0.005) 

(0.006) 

264,833 

264,833 

264,831 

264,826 

264,830 

223,398 

223,398 

223,398 

0.139 

0.327 

0.284 

0.213 

0.351 

0.137 

0.129 

0.237 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village level 
 

 

 

 

 

 
 

158 

 

Table 4. 16: CONSUMER DURABLES - ADJACENT UNDIVIDED STATES AS COMPARISON AND BOTH PARENT AND 
NEW STATES AS TWO TREATMENTS 

  

(1) 

(2) 

(3) 

(4) 

VARIABLES 

Internet 

Television 

Refrigerator 

Motorcycle 

(5) 

Fan 

  

(6) 

Mobile 
Phone 

  

(7) 

Computer 

  

(8) 

Washing 
Machine 

  

  

  

  

  

New States X Post 

0.027*** 

0.100*** 

0.064*** 

0.047*** 

0.139*** 

-0.047** 

-0.267*** 

(0.006) 

(0.005) 

(0.023) 

-0.106*** 

-0.052*** 

-0.10*** 

-0.036** 

(0.003) 

(0.003) 

(0.014) 

(0.019) 

-0.017 

(0.012) 

(0.021) 

-0.165*** 

(0.013) 

0.059*** 

-0.129*** 

-0.275*** 

0.157*** 

-0.19*** 

 

Parent State X Post 

Post 

 

 

New States 

 
Parent States 

 

(0.003) 

-0.010*** 

(0.002) 

0.089*** 

(0.001) 

(0.006) 

-0.19*** 

(0.003) 

0.405*** 

(0.002) 

(0.016) 

-0.005 

(0.008) 

(0.027) 

0.015 

(0.014) 

Age of head of household 

0.001*** 

0.001*** 

 

Sex of head of household  
female 

(Base = Male) 
Education of the 
Household head 

 

Household Size 

 

Rural 

(Base=urban) 
Religion of the HH 
(Muslim) 

(Base= Hindu) 
Religion of the HH 
(Christian) 

 

(0.000) 

(0.000) 

0.003*** 

(0.001) 

-0.003 

(0.002) 

0.006*** 

0.014*** 

(0.000) 

0.003*** 

(0.000) 

-0.089*** 

(0.001) 

-0.030*** 

(0.001) 

0.004 

(0.003) 

(0.000) 

0.023*** 

(0.000) 

-0.26*** 

(0.001) 

-0.09*** 

(0.002) 

0.006 

(0.006) 

 

  

0.100*** 

(0.005) 

(0.003) 

0.193*** 

(0.002) 

(0.023) 

-0.051*** 

(0.012) 

0.003*** 

(0.000) 

0.011*** 

(0.002) 

0.014*** 

(0.000) 

0.009*** 

(0.000) 

0.279*** 

0.409*** 

(0.002) 

(0.002) 

(0.027) 

0.013 

(0.013) 

(0.025) 

0.011 

(0.013) 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.002*** 

0.001*** 

-0.00*** 

0.002*** 

0.002*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

-0.039*** 

(0.002) 

0.000 

(0.002) 

-0.03*** 

0.013*** 

0.017*** 

(0.001) 

(0.001) 

(0.001) 

0.015*** 

0.011*** 

0.007*** 

0.009*** 

0.011*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

0.023*** 

0.014*** 

0.027*** 

0.004*** 

0.004*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

-0.231*** 

-0.116*** 

-0.21*** 

-0.05*** 

-0.103*** 

-0.136*** 

(0.001) 

-0.002 

(0.002) 

0.028*** 

(0.005) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

-0.053*** 

0.016*** 

-0.01*** 

-0.037*** 

0.013*** 

(0.002) 

(0.002) 

(0.002) 

(0.001) 

(0.001) 

0.002 

(0.005) 

0.018*** 

(0.005) 

0.008* 

(0.004) 

0.021*** 

0.035*** 

(0.004) 

(0.004) 

 
 

159 

 

 

 

 

 

 

 

 

 

 

Table 4. 16 (cont’d) 
 
 
Religion of the HH (Sikh) 

 

Religion of the HH 
(Others) 

 

Caste of the 
household(Scheduled 
Caste) 

(Base= General Caste) 
Caste of the 
household(Scheduled 
Tribe) 

Constant 

 

 

 
Observations 

0.042*** 

0.096*** 

0.175*** 

0.144*** 

0.075*** 

0.026*** 

0.082*** 

0.176*** 

(0.006) 

(0.010) 

(0.008) 

(0.009) 

(0.009) 

(0.008) 

(0.007) 

(0.007) 

0.006** 

(0.003) 

0.020*** 

(0.004) 

0.041*** 

(0.004) 

0.022*** 

0.019*** 

(0.004) 

(0.004) 

0.001 

(0.003) 

0.023*** 

0.042*** 

(0.003) 

(0.003) 

-0.028*** 

(0.001) 

-0.06*** 

(0.002) 

-0.072*** 

-0.100*** 

-0.07*** 

-0.04*** 

-0.03*** 

-0.042*** 

(0.001) 

(0.002) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

-0.028*** 

(0.001) 

-0.15*** 

(0.002) 

-0.072*** 

0.213*** 

(0.013) 

(0.023) 

 

 

-0.088*** 

-0.11*** 

-0.18*** 

-0.11*** 

-0.023*** 

-0.026*** 

(0.002) 

0.012 

(0.019) 

 

(0.002) 

(0.002) 

(0.002) 

(0.001) 

(0.002) 

-0.131*** 

0.432*** 

0.887*** 

-0.016** 

-0.032*** 

(0.022) 

(0.021) 

(0.008) 

(0.007) 

(0.007) 

 

 

 

 

 

499,138 

499,138 

499,135 

499,127 

499,132 

391,297 

391,297 

391,297 

R-squared 
Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village level 
 

0.142 

0.312 

0.234 

0.387 

0.143 

0.243 

0.144 

0.358 

 

 

 

 

 

 

 

 

 

 
 

160 

 

Table 4. 17: CONSUMER DURABLES - RDD USING DISTRICTS ON THE NEW STATES AS TREATMENT AND PARENT 
STATE AS COMPARISON 

  

(1) 

(2) 

VARIABLES 
  
New State Border Dist. # Post 

Post Split 

 

District on the New State Side 

 

Age of head of household 

 

 

Sex of head of household  
female 
(Base = Male) 
Education of the Household 
head 

 
Household Size 

 

Rural 
(Base=urban) 
Religion of the HH (Muslim) 
(Base= Hindu) 
Religion of the HH (Christian) 

Religion of the HH (Sikh) 

 

Religion of the HH (Others) 

 

 

Constant 
 
 

Observations 

Internet 

Television 

  

0.074*** 
(0.004) 
0.043*** 
(0.003) 
-0.010 
(0.008) 
0.001*** 
(0.000) 

  

0.092*** 
(0.008) 
0.185*** 
(0.005) 
0.061*** 
(0.015) 
0.002*** 
(0.000) 

0.011*** 
(0.002) 

0.017*** 
(0.005) 

0.005*** 
(0.000) 
0.002*** 
(0.000) 
-0.07*** 
(0.002) 
-0.02*** 
(0.002) 
0.030*** 
(0.010) 
0.021* 
(0.012) 
-0.001 
(0.009) 
-0.002 
(0.007) 

 

0.015*** 
(0.000) 
0.019*** 
(0.001) 

-0.317*** 

(0.004) 

-0.126*** 

(0.005) 
0.020 
(0.019) 
0.181*** 
(0.023) 
-0.015 
(0.018) 
0.136*** 
(0.014) 

 

 
 

 
 
 
 
 
 
 
 
 
 

 
 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

(3) 

(4) 

Refrigerator 

Motorcycle 

  

0.037*** 
(0.007) 
0.183*** 
(0.005) 
0.139*** 
(0.014) 
0.002*** 
(0.000) 

-0.010** 
(0.005) 

0.015*** 
(0.000) 
0.020*** 
(0.001) 

-0.111*** 

(0.004) 

-0.027*** 

(0.005) 
0.107*** 
(0.018) 
0.197*** 
(0.022) 
0.016 
(0.017) 

-0.171*** 

(0.014) 

 

  

-0.001 
(0.006) 
0.094*** 
(0.004) 
0.082*** 
(0.012) 
0.002*** 
(0.000) 

0.026*** 
(0.004) 

0.010*** 
(0.000) 
0.006*** 
(0.000) 

-0.232*** 

(0.003) 
0.024*** 
(0.004) 
0.053*** 
(0.015) 
0.280*** 
(0.018) 
0.029** 
(0.014) 
-0.004 
(0.011) 

 
 
 
 

(5) 

Fan 

  

0.041*** 
(0.008) 
0.317*** 
(0.005) 
0.009 
(0.015) 
0.001*** 
(0.000) 

0.020*** 
(0.005) 

0.015*** 
(0.000) 
0.014*** 
(0.001) 
-0.30*** 
(0.004) 
0.054*** 
(0.005) 
0.006 
(0.019) 
0.169*** 
(0.023) 
-0.033* 
(0.018) 
0.223*** 
(0.014) 

 

(6) 

Mobile 
Phones 

  

0.178*** 
(0.011) 

 
 
 
 

-0.00*** 
(0.000) 

-0.01*** 
(0.004) 

0.008*** 
(0.000) 
0.027*** 
(0.001) 
-0.05*** 
(0.003) 
0.012*** 
(0.004) 
0.047*** 
(0.016) 
0.038* 
(0.021) 
-0.004 
(0.015) 
0.917*** 
(0.011) 

 

(7) 

Computer 

  

0.043*** 
(0.007) 

 
 
 
 

0.001*** 
(0.000) 

0.018*** 
(0.002) 

0.006*** 
(0.000) 
0.002*** 
(0.000) 

-0.079*** 

(0.002) 

-0.025*** 

(0.002) 
0.048*** 
(0.010) 
0.060*** 
(0.013) 
0.027*** 
(0.010) 

-0.021*** 

(0.007) 

 

(8) 

Washing 
Machine 

  

0.051*** 
(0.009) 

 
 
 
 

0.002*** 
(0.000) 

0.024*** 
(0.003) 

0.009*** 
(0.000) 
0.005*** 
(0.000) 
-0.15*** 
(0.003) 
0.030*** 
(0.003) 
0.044*** 
(0.013) 
0.313*** 
(0.017) 
0.031** 
(0.013) 
-0.018** 
(0.009) 

 

71,750 
0.185 

71,751 
0.335 

61,390 
0.149 

61,390 
0.112 

61,390 
0.234 

R-squared 
Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village level 

 

71,751 
0.141 

71,751 
0.304 

71,751 
0.255 

 

 

 
 

161 

 

Table 4. 18:  MAJOR HEALTH INDICATORS- PARENT STATE AS COMPARISON AND NEWLY FORMED STATES AS 
TREATMENT 

  

(1) 

(2) 

(3) 

(4) 

(5) 

(6) 

(7) 

VARIABLES 
  
Districts in New states X Post 

post 

treatment 

 

 

 

Age of head of household 

 

Sex of head of household  
female 
(Base = Male) 
Years of Education Women has 
Received 

 

Rural 
(Base=urban) 
Household Size 

 

Religion of the HH (Muslim) 
(Base= Hindu) 
Religion of the HH (Christian) 

Religion of the HH (Sikh) 

 

 

Religion of the HH (Others) 

 

Caste of the 
household(Scheduled Caste) 
(Base= General Caste) 
Caste of the 
household(Scheduled Tribe) 

Constant 

 

 

 

Govt Institutional 

Delivery 

Infant Death 

Underweight 

Stunted 

Wasted 

Health Card 

Compensation for 

Sterilization 

  

0.004 
(0.009) 
0.426*** 
(0.003) 

-0.232*** 

(0.069) 
-0.000** 
(0.000) 

-0.004 
(0.004) 

0.002*** 
(0.000) 
0.045*** 
(0.003) 

-0.002*** 

(0.000) 

-0.058*** 

(0.004) 

-0.059*** 

(0.021) 
-0.044 
(0.032) 
-0.020 
(0.016) 

0.000 
(0.003) 

-0.042*** 

(0.004) 
0.351*** 
(0.038) 

  

-0.025*** 

(0.005) 

-0.037*** 

(0.002) 
0.018 
(0.037) 
0.001*** 
(0.000) 

-0.006*** 

(0.002) 

-0.002*** 

(0.000) 
0.012*** 
(0.002) 

-0.006*** 

(0.000) 
-0.004* 
(0.002) 
0.004 
(0.012) 
-0.019 
(0.017) 
0.004 
(0.009) 

0.003* 
(0.002) 

0.006*** 
(0.002) 
0.062*** 
(0.021) 

  

0.008 
(0.011) 

-0.117*** 

(0.004) 
0.296*** 
(0.078) 

-0.001*** 

(0.000) 

-0.002 
(0.004) 

  

-0.011 
(0.013) 

-0.142*** 

(0.004) 
-0.016 
(0.082) 

-0.001*** 

(0.000) 

  

0.067*** 
(0.011) 

-0.038*** 

(0.004) 
0.346*** 
(0.070) 
-0.000 
(0.000) 

0.008* 
(0.004) 

-0.007* 
(0.004) 

-0.014*** 

-0.015*** 

-0.002*** 

(0.000) 
0.036*** 
(0.004) 
-0.001 
(0.000) 
0.001 
(0.004) 

-0.069*** 

(0.025) 

-0.160*** 

(0.037) 
-0.006 
(0.019) 

0.055*** 
(0.004) 

0.027*** 
(0.005) 
0.628*** 
(0.049) 

(0.000) 
0.007** 
(0.003) 

-0.002*** 

(0.000) 
0.001 
(0.004) 
-0.003 
(0.022) 
-0.056* 
(0.031) 
0.018 
(0.016) 

0.009*** 
(0.003) 

0.028*** 
(0.004) 
0.106** 
(0.043) 

(0.000) 
0.031*** 
(0.004) 
-0.001** 
(0.000) 

-0.011*** 

(0.004) 
-0.027 
(0.025) 

-0.131*** 

(0.036) 
0.043** 
(0.018) 

0.043*** 
(0.004) 

0.029*** 
(0.005) 
0.384*** 
(0.044) 

 
 

162 

  

0.071*** 
(0.009) 
0.301*** 
(0.003) 

-0.204*** 

(0.069) 
0.000*** 
(0.000) 

0.004 
(0.004) 

0.013*** 
(0.000) 

-0.021*** 

(0.004) 
0.001* 
(0.000) 

-0.058*** 

(0.004) 
0.000 
(0.022) 
0.126*** 
(0.032) 
0.048*** 
(0.016) 

-0.014*** 

(0.003) 

-0.027*** 

(0.004) 
0.533*** 
(0.038) 

  

0.028*** 
(0.005) 
0.123*** 
(0.002) 
0.057 
(0.040) 

-0.001*** 

(0.000) 

-0.002 
(0.002) 

-0.005*** 

(0.000) 
0.010*** 
(0.002) 
0.002*** 
(0.000) 

-0.059*** 

(0.002) 
-0.020 
(0.012) 
-0.022 
(0.018) 

-0.041*** 

(0.009) 

-0.005** 
(0.002) 

0.006** 
(0.002) 
0.049** 
(0.022) 

 

Table 4. 18(cont’d) 
 
 
Observations 
R-squared 
Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered 
at the village level 
 
 

140,417 

137,829 

121,529 

0.022 

140,417 

0.185 

125,390 

0.052 

122,249 

0.063 

0.014 

0.150 

140,417 

0.095 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

163 

 

Table 4. 19: GOVERNMENT PROVISIONS - PARENT STATE AS COMPARISON AND NEWLY FORMED STATES AS 
TREATMENT 

  

VARIABLES 

  

Districts in New states X Post 

post 

treatment 

 

 

 

Age of head of household 

 

Sex of head of household  female 

(Base = Male) 
Years of Education Women has 
Received 

Rural 

(Base=urban) 

Household Size 

 

 

Religion of the HH (Muslim) 

(Base= Hindu) 

Religion of the HH (Christian) 

 

Religion of the HH (Sikh) 

 

Religion of the HH (Others) 

 

Caste of the household(Scheduled 
Caste) 

(Base= General Caste) 
Caste of the household(Scheduled 
Tribe) 

 

(1) 

(2) 

(3) 

(4) 

(5) 

Tetanus Shot 

Polio Vaccine  Measles Vaccine 

BCG Vaccine 

Iron Tablets 

(6) 

Vitamin A 

tablets 

(7) 

DPT Vaccine 

  

0.029*** 

(0.010) 

0.090*** 

(0.004) 

-0.30*** 

(0.073) 

0.001*** 

(0.000) 

0.006 

(0.004) 

0.013*** 

(0.000) 

-0.06*** 

(0.004) 

-0.01*** 

(0.000) 

-0.03*** 

(0.004) 

0.028 

(0.023) 

0.108*** 

(0.034) 

0.018 

(0.017) 

-0.02*** 

(0.003) 

  

0.091*** 

(0.010) 

0.256*** 

(0.004) 

-0.26*** 

(0.070) 

0.000*** 

(0.000) 

0.006 

(0.004) 

0.011*** 

(0.000) 

-0.03*** 

(0.004) 

-0.00*** 

(0.000) 

-0.04*** 

(0.004) 

-0.017 

(0.022) 

0.103*** 

(0.032) 

0.032** 

(0.016) 

-0.02*** 

(0.003) 

  

0.020** 

(0.010) 

0.403*** 

(0.004) 

-0.29*** 

(0.071) 

-0.000 

(0.000) 

0.010*** 

(0.004) 

0.009*** 

(0.000) 

-0.02*** 

(0.004) 

0.000 

(0.000) 

-0.08*** 

(0.004) 

-0.000 

(0.022) 

0.086*** 

(0.033) 

0.022 

(0.016) 

-0.02*** 

(0.003) 

  

-0.019** 

(0.008) 

0.418*** 

(0.003) 

-0.28*** 

(0.061) 

0.000* 

(0.000) 

0.025*** 

(0.003) 

0.011*** 

(0.000) 

-0.03*** 

(0.003) 

0.002*** 

(0.000) 

-0.07*** 

(0.003) 

0.023 

(0.019) 

0.070** 

(0.028) 

0.035** 

(0.014) 

-0.01*** 

(0.003) 

  

0.032*** 

(0.010) 

0.091*** 

(0.004) 

-0.30*** 

(0.072) 

0.001*** 

(0.000) 

-0.000 

(0.004) 

0.018*** 

(0.000) 

-0.05*** 

(0.004) 

-0.00*** 

(0.000) 

-0.04*** 

(0.004) 

0.027 

(0.023) 

  

0.028*** 

(0.010) 

0.448*** 

(0.004) 

-0.44*** 

(0.071) 

0.000** 

(0.000) 

-0.004 

(0.004) 

0.009*** 

(0.000) 

-0.003 

(0.004) 

-0.00*** 

(0.000) 

-0.06*** 

(0.004) 

0.014 

(0.022) 

0.122*** 

0.118*** 

(0.034) 

0.024 

(0.016) 

-0.02*** 

(0.003) 

(0.033) 

0.002 

(0.016) 

-0.008** 

(0.003) 

  

0.008 

(0.009) 

0.336*** 

(0.003) 

-0.161** 

(0.065) 

0.000* 

(0.000) 

0.025*** 

(0.004) 

0.011*** 

(0.000) 

-0.021*** 

(0.003) 

0.002*** 

(0.000) 

-0.081*** 

(0.003) 

0.016 

(0.020) 

0.074** 

(0.030) 

0.022 

(0.015) 

-0.016*** 

(0.003) 

-0.02*** 

-0.05*** 

-0.04*** 

-0.04*** 

-0.02*** 

-0.02*** 

-0.040*** 

 
 

164 

 

Table 4. 19 (cont’d) 
 
 

Constant 

 

Observations 

(0.005) 

0.805*** 

(0.041) 

140,417 

(0.004) 

0.225*** 

(0.039) 

140,417 

(0.004) 

0.490*** 

(0.039) 

140,417 

(0.004) 

0.655*** 

(0.033) 

140,417 

(0.004) 

0.641*** 

(0.040) 

140,417 

(0.004) 

0.294*** 

(0.039) 

130,426 

(0.004) 

0.549*** 

(0.036) 

140,417 

R-squared 
Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village level 
 

0.201 

0.103 

0.057 

0.110 

0.160 

0.217 

0.159 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

165 

 

Table 4. 20: MAJOR HEALTH INDICATORS- RDD USING DISTRICTS ON THE NEW STATES AS TREATMENT AND 
PARENT STATE AS COMPARISON 

  

(1) 

(2) 

(3) 

(4) 

(5) 

(6) 

(7) 

Govt Institutional 

Delivery 

  

0.061*** 
(0.012) 
0.364*** 
(0.007) 
-0.013 
(0.023) 
-0.000 
(0.000) 

-0.021*** 

(0.007) 

0.003*** 
(0.001) 
0.031*** 
(0.006) 

-0.003*** 

(0.001) 

-0.080*** 

(0.007) 
-0.079** 
(0.032) 
-0.022 
(0.047) 

-0.082*** 

(0.030) 

-0.014** 
(0.006) 

-0.040*** 

(0.008) 

VARIABLES 
  
New State Border Dist. # Post 

 

Post 

 
Treatment RDD 

 

Age of head of household 

 

Sex of head of household  
female 
(Base = Male) 
Years of Education Women 
has Received 

 

Rural 
(Base=urban) 
Household Size 

 

Religion of the HH (Muslim) 
(Base= Hindu) 
Religion of the HH (Christian) 

 

Religion of the HH (Sikh) 

 

Religion of the HH (Others) 

 

Caste of the 
household(Scheduled Caste) 
(Base= General Caste) 
Caste of the 
household(Scheduled Tribe) 

 

 

Infant Death 

Underweight 

Stunted 

Wasted 

Health Card 

Compensation for 

Sterilization 

  

-0.016** 
(0.007) 
-0.03*** 
(0.004) 
-0.006 
(0.012) 
0.000*** 
(0.000) 

-0.003 
(0.004) 

-0.00*** 
(0.000) 
0.009** 
(0.003) 
-0.00*** 
(0.000) 
-0.02*** 
(0.004) 
0.011 
(0.017) 
-0.002 
(0.025) 
0.030* 
(0.016) 

0.005 
(0.003) 

0.004 
(0.004) 

  

-0.026* 
(0.015) 

-0.131*** 

(0.009) 

-0.145*** 

(0.027) 

-0.001*** 

(0.000) 

-0.013 
(0.008) 

-0.015*** 

(0.001) 
0.030*** 
(0.007) 
-0.001 
(0.001) 

-0.021*** 

(0.008) 
-0.037 
(0.038) 
-0.129** 
(0.055) 
0.064* 
(0.037) 

0.053*** 
(0.007) 

0.022** 
(0.009) 

  

-0.06*** 
(0.016) 
-0.22*** 
(0.009) 
-0.20*** 
(0.029) 
-0.00*** 
(0.000) 

-0.001 
(0.008) 

-0.02*** 
(0.001) 
0.026*** 
(0.007) 
-0.001 
(0.001) 
-0.016** 
(0.008) 
-0.036 
(0.038) 
-0.22*** 
(0.054) 
-0.073* 
(0.037) 

0.057*** 
(0.007) 

0.011 
(0.009) 

 
 

166 

  

-0.06*** 
(0.015) 
-0.03*** 
(0.008) 
-0.07*** 
(0.025) 
-0.000 
(0.000) 

-0.008 
(0.007) 

-0.00*** 
(0.001) 
0.010 
(0.007) 
-0.001 
(0.001) 
0.001 
(0.007) 
-0.025 
(0.034) 
-0.011 
(0.048) 
0.080** 
(0.033) 

0.007 
(0.007) 

0.022*** 
(0.008) 

  

0.142*** 
(0.013) 
0.236*** 
(0.008) 
-0.07*** 
(0.024) 
0.000 
(0.000) 

-0.016** 
(0.007) 

0.01*** 
(0.001) 
-0.013* 
(0.007) 
0.001 
(0.001) 
-0.07*** 
(0.007) 
-0.053 
(0.033) 
0.116** 
(0.048) 
0.071** 
(0.031) 

-0.04*** 
(0.006) 

-0.011 
(0.008) 

  

0.026*** 
(0.007) 
0.115*** 
(0.004) 

-0.075*** 

(0.014) 

-0.001*** 

(0.000) 

0.004 
(0.004) 

-0.004*** 

(0.000) 
-0.007* 
(0.004) 
0.003*** 
(0.000) 

-0.076*** 

(0.004) 
-0.029 
(0.019) 
0.000 
(0.028) 
-0.026 
(0.018) 

-0.014*** 

(0.004) 

-0.016*** 

(0.005) 

 

Table 4. 20 (cont’d) 
 
 
Constant 

 
Observations 
R-squared 
 
Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village level 
 

0.138*** 
(0.021) 
37,855 
0.151 

0.103*** 
(0.011) 
37,262 
0.012 

0.700*** 
(0.024) 
34,058 
0.054 

0.781*** 
(0.024) 
33,540 
0.073 

0.331*** 
(0.021) 
33,355 
0.016 

0.313*** 
(0.021) 
37,855 
0.144 

0.109*** 
(0.012) 
37,855 
0.085 

 

 

 

 

 
 

167 

 

Table 4. 21: GOVERNMENT PROVISIONS- RDD USING DISTRICTS ON THE NEW STATES AS TREATMENT AND PARENT 
STATE AS COMPARISON 

  

(1) 

(2) 

(3) 

(4) 

(5) 

(6) 

(7) 

VARIABLES 

  
New State Border Dist. # 
Post 

post 

 

 

Treatment_RDD 

 

Age of head of 
household 

 

Sex of head of household  
female 

(Base = Male) 
Years of Education 
Women has Received 

 

Rural 

(Base=urban) 

Household Size 

 

Religion of the HH 
(Muslim) 

(Base= Hindu) 
Religion of the HH 
(Christian) 

 

Religion of the HH 
(Sikh) 

 

 

Tetanus Shot 

Polio Vaccine 

Measles Vaccine 

BCG Vaccine 

Iron Tablets 

Vitamin A tablets 

DPT Vaccine 

  

  

  

  

  

  

  

-0.019 

(0.013) 

0.160*** 

(0.008) 

-0.078*** 

(0.025) 

0.002*** 

(0.000) 

-0.006 

(0.008) 

0.012*** 

(0.001) 

0.035*** 

(0.013) 

0.305*** 

(0.008) 

-0.059** 

(0.024) 

0.000 

(0.000) 

-0.006 

(0.008) 

0.011*** 

(0.001) 

-0.062*** 

-0.025*** 

(0.007) 

-0.009*** 

(0.001) 

(0.007) 

-0.000 

(0.001) 

-0.023*** 

-0.066*** 

(0.007) 

0.020 

(0.034) 

0.135*** 

(0.050) 

(0.007) 

-0.010 

(0.033) 

0.132*** 

(0.049) 

0.034*** 

(0.013) 

0.468*** 

(0.008) 

0.017 

(0.024) 

-0.000 

(0.000) 

0.001 

(0.007) 

0.008*** 

(0.001) 

-0.015** 

(0.007) 

0.001 

(0.001) 

-0.086*** 

(0.007) 

-0.013 

(0.033) 

0.125*** 

(0.048) 

0.067*** 

(0.010) 

0.504*** 

(0.006) 

0.078*** 

(0.020) 

0.000* 

(0.000) 

0.013** 

(0.006) 

0.011*** 

(0.000) 

-0.024*** 

(0.005) 

0.002*** 

(0.001) 

-0.065*** 

(0.006) 

0.044* 

(0.027) 

0.109*** 

(0.039) 

 
 

168 

-0.018 

(0.013) 

0.142*** 

(0.008) 

0.181*** 

(0.025) 

0.001*** 

(0.000) 

-0.007 

(0.008) 

0.018*** 

(0.001) 

-0.050*** 

(0.007) 

-0.008*** 

(0.001) 

-0.056*** 

(0.007) 

0.036 

(0.034) 

0.149*** 

(0.050) 

0.043*** 

(0.013) 

0.454*** 

(0.008) 

0.061** 

(0.025) 

0.000 

(0.000) 

-0.010 

(0.008) 

0.009*** 

(0.001) 

0.006 

(0.007) 

-0.001 

(0.001) 

-0.082*** 

(0.007) 

0.031 

(0.034) 

0.161*** 

(0.048) 

0.031*** 

(0.011) 

0.411*** 

(0.007) 

0.024 

(0.021) 

0.000 

(0.000) 

0.015** 

(0.007) 

0.011*** 

(0.000) 

-0.020*** 

(0.006) 

0.002*** 

(0.001) 

-0.081*** 

(0.006) 

0.011 

(0.029) 

0.097** 

(0.043) 

 

 
Table 4. 21 (cont’d) 
 
 
Religion of the HH 
(Others) 

 

Caste of the 
household(Scheduled 
Caste) 

(Base= General Caste) 
Caste of the 
household(Scheduled 
Tribe) 

 

Constant 

 
Observations 

R-squared 

 
 
 
 

 
 
 
 

 
 
 
 

 
 
 
 

 
 
 
 

 
 
 
 

 
 
 
 

0.078** 

(0.032) 

0.027 

(0.031) 

0.061** 

(0.031) 

0.079*** 

(0.025) 

0.060* 

(0.032) 

0.073** 

(0.031) 

0.072*** 

(0.027) 

-0.046*** 

-0.022*** 

(0.007) 

(0.006) 

-0.032*** 

(0.006) 

-0.018*** 

(0.005) 

-0.025*** 

(0.007) 

-0.024*** 

(0.007) 

-0.033*** 

(0.006) 

0.004 

(0.008) 

0.562*** 

(0.022) 

37,855 

0.065 

-0.050*** 

(0.008) 

0.200*** 

(0.021) 

37,855 

0.117 

-0.017** 

(0.008) 

0.211*** 

(0.021) 

37,855 

0.172 

-0.014** 

(0.006) 

0.325*** 

(0.017) 

37,855 

0.264 

0.011 

(0.008) 

-0.012 

(0.008) 

-0.011 

(0.007) 

0.251*** 

0.140*** 

0.393*** 

(0.022) 

37,855 

0.099 

(0.021) 

35,353 

0.197 

(0.019) 

37,855 

0.186 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village level 
 

 

 

 

 

 

 

 

 

 

 

 
 

169 

 

Table 4. 22: MAJOR HEALTH INDICATORS (CHILD) ADJACENT STATES AS COMPARISON AND BOTH PARENT AND 
NEW STATES AS TWO TREATMENTS 

  

(1) 

(2) 

(3) 

(4) 

(5) 

(6) 

(7) 

Govt Institutional 

Delivery 

Infant Death 

Underweight 

Stunted 

Wasted 

Health Card 

Compensation for 

Sterilization 

  

  

  

  

0.097*** 

-0.02*** 

-0.041*** 

-0.029** 

(0.010) 

0.092*** 

(0.005) 

(0.005) 

-0.004 

(0.002) 

(0.012) 

-0.041*** 

(0.006) 

(0.013) 

-0.010 

(0.006) 

  

-0.002 

(0.012) 

  

0.168*** 

(0.010) 

  

0.032*** 

(0.006) 

-0.07*** 

0.094*** 

-0.064*** 

(0.006) 

(0.005) 

0.345*** 

-0.03*** 

-0.071*** 

-0.13*** 

0.033*** 

0.208*** 

(0.004) 

-0.107** 

(0.051) 

-0.151*** 

(0.023) 

(0.002) 

0.017 

(0.026) 

0.018 

(0.011) 

(0.004) 

0.119** 

(0.057) 

0.010 

(0.026) 

(0.005) 

0.014 

(0.066) 

0.035 

(0.028) 

(0.000) 

0.012*** 

(0.003) 

(0.000) 

-0.003* 

(0.002) 

(0.000) 

-0.001 

(0.004) 

(0.000) 

0.005 

(0.004) 

(0.004) 

0.119** 

(0.058) 

0.022 

(0.024) 

-0.000 

(0.000) 

-0.005* 

(0.003) 

(0.003) 

-0.14*** 

(0.049) 

-0.20*** 

(0.022) 

0.000*** 

(0.000) 

0.006* 

(0.003) 

Age of head of household 

-0.000*** 

0.000*** 

-0.001*** 

-0.00*** 

VARIABLES 

  

New States X Post 

 
Parent State X Post 

Post 

New States 

Parent States 

 

 

 

 

 

Sex of head of household  female 

(Base = Male) 
Years of Education Women has 
Received 

Rural 

(Base=urban) 

Household Size 

 

 

Religion of the HH (Muslim) 

(Base= Hindu) 

Religion of the HH (Christian) 

 

Religion of the HH (Sikh) 

 

(0.003) 

0.192*** 

(0.002) 

-0.042 

(0.032) 

0.001 

(0.014) 

-0.001*** 

(0.000) 

-0.005** 

(0.002) 

(0.000) 

0.025*** 

(0.002) 

0.003*** 

(0.000) 

-0.001*** 

-0.00*** 

-0.014*** 

-0.02*** 

-0.00*** 

0.012*** 

-0.007*** 

(0.000) 

0.022*** 

(0.003) 

-0.001** 

(0.000) 

-0.051*** 

(0.003) 

0.006 

(0.011) 

-0.027 

(0.000) 

0.009*** 

(0.001) 

-0.01*** 

(0.000) 

-0.001 

(0.001) 

-0.009* 

(0.005) 

-0.010 

(0.000) 

0.029*** 

(0.003) 

-0.001*** 

(0.000) 

-0.006* 

(0.003) 

-0.002 

(0.012) 

-0.054*** 

 
 

170 

(0.000) 

(0.000) 

(0.000) 

0.030*** 

0.009*** 

-0.021*** 

(0.003) 

-0.000 

(0.000) 

0.005 

(0.003) 

(0.003) 

-0.00*** 

(0.000) 

-0.000 

(0.003) 

-0.04*** 

0.038*** 

(0.013) 

-0.09*** 

(0.011) 

-0.025 

(0.002) 

0.001* 

(0.000) 

-0.05*** 

-0.056*** 

(0.003) 

0.000 

(0.010) 

0.082*** 

(0.002) 

0.015** 

(0.007) 

-0.011 

 

 

 

 

 

 

 

 

 

 

 

 

 
Table 4. 22 (cont’d) 
 
 

Religion of the HH (Others) 

 

Caste of the household(Scheduled 
Caste) 

(Base= General Caste) 
Caste of the household(Scheduled 
Tribe) 

 

 

Constant 

Observations 

R-squared 

(0.018) 

0.019** 

(0.008) 

0.032*** 

(0.003) 

0.000 

(0.003) 

(0.009) 

-0.000 

(0.004) 

0.003** 

(0.001) 

0.008*** 

(0.002) 

0.355*** 

0.054*** 

(0.033) 

228,754 

0.173 

(0.016) 

225,019 

0.017 

(0.020) 

0.013 

(0.009) 

(0.020) 

0.032*** 

(0.009) 

(0.018) 

-0.003 

(0.008) 

(0.017) 

0.017** 

(0.008) 

0.035*** 

0.046*** 

0.012*** 

-0.01*** 

(0.003) 

(0.003) 

(0.003) 

(0.002) 

0.045*** 

(0.004) 

0.453*** 

(0.037) 

203,899 

0.062 

0.038*** 

0.029*** 

-0.03*** 

(0.004) 

(0.003) 

(0.003) 

0.643*** 

0.136*** 

0.646*** 

(0.039) 

195,895 

0.078 

(0.035) 

194,693 

0.023 

(0.032) 

228,754 

0.165 

(0.011) 

-0.002 

(0.005) 

0.005*** 

(0.002) 

0.002 

(0.002) 

0.051** 

(0.021) 

228,754 

0.101 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village level 
 

 
 

171 

 

Table 4. 23: GOVERNMENT PROVISIONS (CHILD) - ADJACENT STATES AS COMPARISON AND BOTH PARENT AND 
NEW STATES AS TWO TREATMENTS 

(1) 

(2) 

Tetanus Shot 

Polio Vaccine 

  

  

(3) 

Measles 
Vaccine 

  

(4) 

(5) 

BCG Vaccine 

Iron Tablets 

(6) 

Vitamin A 

tablets 

(7) 

DPT Vaccine 

  

  

0.194*** 

0.072*** 

0.149*** 

0.196*** 

0.195*** 

(0.010) 

(0.010) 

(0.010) 

(0.008) 

(0.010) 

0.165*** 

-0.022*** 

0.128*** 

0.217*** 

0.160*** 

(0.005) 

(0.005) 

(0.005) 

(0.004) 

(0.005) 

  

-0.016 

(0.010) 

0.011** 

(0.005) 

  

0.197*** 

(0.009) 

0.190*** 

(0.004) 

-0.076*** 

0.279*** 

0.276*** 

0.205*** 

-0.069*** 

0.437*** 

0.149*** 

(0.004) 

-0.148*** 

(0.052) 

-0.212*** 

(0.024) 

(0.004) 

0.013 

(0.052) 

-0.027 

(0.024) 

(0.000) 

0.001 

(0.003) 

(0.000) 

0.000 

(0.003) 

(0.000) 

0.004 

(0.003) 

(0.004) 

(0.003) 

(0.004) 

(0.004) 

(0.003) 

-0.164*** 

-0.170*** 

-0.257*** 

-0.124** 

-0.175*** 

(0.052) 

(0.043) 

(0.053) 

(0.052) 

(0.046) 

-0.090*** 

-0.197*** 

-0.248*** 

-0.107*** 

-0.188*** 

(0.024) 

(0.019) 

-0.000 

(0.000) 

0.015*** 

(0.003) 

(0.024) 

0.001*** 

(0.000) 

-0.002 

(0.003) 

(0.023) 

0.000* 

(0.000) 

-0.004 

(0.003) 

(0.021) 

-0.000 

(0.000) 

0.013*** 

(0.003) 

  

VARIABLES 

  

New States X Post 

Parent State X Post 

Post 

New States 

Parent States 

 

 

 

 

 

 

Sex of head of household  female 

(Base = Male) 
Years of Education Women has 
Received 

Rural 

(Base=urban) 

Household Size 

 

 

Age of head of household 

0.001*** 

0.000*** 

-0.000*** 

0.013*** 

0.010*** 

0.008*** 

0.010*** 

0.017*** 

0.008*** 

0.010*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

-0.041*** 

-0.018*** 

-0.019*** 

-0.025*** 

-0.033*** 

(0.003) 

(0.003) 

(0.003) 

(0.002) 

(0.003) 

(0.000) 

-0.001 

(0.003) 

(0.000) 

-0.020*** 

(0.002) 

-0.009*** 

-0.002*** 

-0.001** 

0.002*** 

-0.007*** 

-0.003*** 

0.001*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

Religion of the HH (Muslim) 

-0.031*** 

-0.039*** 

-0.077*** 

-0.067*** 

-0.046*** 

-0.051*** 

-0.073*** 

(Base= Hindu) 

Religion of the HH (Christian) 

 

(0.003) 

0.016 

(0.011) 

(0.003) 

-0.032*** 

(0.011) 

(0.003) 

-0.015 

(0.011) 

(0.003) 

0.002 

(0.009) 

Religion of the HH (Sikh) 

0.085*** 

0.051*** 

0.058*** 

0.060*** 

(0.003) 

0.025** 

(0.011) 

0.045** 

(0.003) 

-0.025** 

(0.011) 

(0.003) 

0.012 

(0.010) 

0.083*** 

0.068*** 

 

 
 

172 

 

 

 

 

 

 

 

 

 

 

 

 

Table 4. 23 (cont’d) 
 
 

Religion of the HH (Others) 

 

Caste of the household(Scheduled 
Caste) 

(Base= General Caste) 
Caste of the household(Scheduled 
Tribe) 

 

 

Constant 

Observations 

R-squared 

(0.018) 

0.013 

(0.008) 

(0.018) 

0.002 

(0.008) 

(0.018) 

0.010 

(0.008) 

(0.015) 

0.017*** 

(0.007) 

(0.019) 

0.018** 

(0.008) 

-0.024*** 

-0.015*** 

-0.020*** 

-0.013*** 

-0.008*** 

(0.003) 

(0.003) 

(0.003) 

(0.002) 

(0.003) 

(0.018) 

-0.004 

(0.008) 

-0.005* 

(0.003) 

(0.016) 

0.002 

(0.007) 

-0.018*** 

(0.002) 

-0.017*** 

-0.058*** 

-0.039*** 

-0.028*** 

-0.011*** 

-0.013*** 

-0.036*** 

(0.003) 

(0.003) 

(0.003) 

(0.003) 

(0.003) 

(0.003) 

(0.003) 

0.914*** 

0.217*** 

0.515*** 

0.766*** 

0.750*** 

0.304*** 

0.689*** 

(0.034) 

228,754 

0.074 

(0.034) 

228,754 

0.128 

(0.034) 

228,754 

0.147 

(0.028) 

228,754 

0.190 

(0.034) 

228,754 

0.137 

(0.033) 

214,199 

0.225 

(0.030) 

228,754 

0.144 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village level 
 

 
 

173 

 

APPENDIX D: TABLE- STATEWISE DETAILED RESULTS 
 

Table 4. 24: SELECTED DEVELOPMENTAL INDICATORS- BIHAR AS COMPARISON AND JHARKHAND AS TREATMENT  

(1) 

(2) 

Piped Drinking 

Electricity 

Water 

(3) 

Concrete 
Housing 

(4) 

(5) 

Clean 

(6) 

(7) 

(8) 

(9) 

Toilet 

Cooking Fuel 

Television 

Internet 

Refrigerator  Motorbike 

  

0.04*** 

(0.004) 

0.02*** 

(0.002) 

0.00*** 

(0.000) 

0.00 

(0.002) 

0.00*** 

(0.000) 

-0.00 

(0.000) 

  

  

0.03*** 

0.11*** 

(0.008) 

(0.008) 

0.10*** 

0.16*** 

(0.004) 

(0.004) 

0.00*** 

0.00*** 

(0.000) 

(0.000) 

0.04*** 

0.04*** 

(0.004) 

(0.004) 

0.02*** 

0.02*** 

(0.000) 

(0.000) 

  

0.03*** 

(0.007) 

0.10*** 

(0.003) 

0.00*** 

(0.000) 

0.03*** 

(0.003) 

0.02*** 

(0.000) 

  

  

0.03*** 

-0.01 

(0.008) 

(0.005) 

  

0.03*** 

(0.005) 

0.12*** 

0.06*** 

0.00 

(0.004) 

(0.002) 

(0.002) 

0.00*** 

0.00*** 

0.00*** 

(0.000) 

(0.000) 

(0.000) 

0.03*** 

0.01*** 

0.02*** 

(0.004) 

(0.002) 

(0.002) 

0.02*** 

0.01*** 

0.01*** 

(0.000) 

(0.000) 

(0.000) 

0.00*** 

0.00*** 

-0.01*** 

0.01*** 

0.00*** 

0.00*** 

(0.001) 

(0.001) 

(0.000) 

(0.001) 

(0.000) 

(0.000) 

  

0.05*** 

(0.008) 

0.11*** 

(0.004) 

0.00*** 

(0.000) 

0.01* 

(0.004) 

0.02*** 

(0.000) 

0.01*** 

(0.000) 

(0.002) 

0.00* 

(0.002) 

-0.02*** 

(0.007) 

0.12*** 

(0.023) 

-0.04*** 

(0.004) 

0.01** 

(0.004) 

-0.00 

(0.004) 

0.05*** 

(0.015) 

(0.015) 

0.14*** 

0.14*** 

(0.004) 

(0.003) 

(0.004) 

(0.002) 

(0.003) 

(0.004) 

0.09*** 

-0.01*** 

-0.05*** 

-0.01*** 

0.01*** 

-0.02*** 

(0.003) 

0.02 

(0.012) 

0.19*** 

(0.004) 

(0.002) 

(0.003) 

-0.01 

0.00 

0.01 

(0.015) 

(0.008) 

0.08 

-0.02 

(0.010) 

0.17*** 

(0.004) 

0.04*** 

(0.014) 

0.15*** 

(0.050) 

(0.052) 

(0.041) 

(0.052) 

(0.028) 

(0.033) 

(0.048) 

-0.11*** 

-0.05*** 

-0.06*** 

-0.09*** 

-0.02*** 

-0.04*** 

-0.05*** 

(0.005) 

(0.011) 

(0.011) 

(0.009) 

(0.011) 

(0.006) 

(0.007) 

(0.011) 

-0.25*** 

-0.10*** 

-0.40*** 

-0.43*** 

-0.40*** 

-0.34*** 

-0.10*** 

-0.18*** 

-0.12*** 

Education of the Household head 

0.01*** 

  

0.11*** 

(0.009) 

0.42*** 

(0.004) 

0.00*** 

(0.000) 

0.02*** 

(0.004) 

(0.000) 

0.01*** 

(0.001) 

(0.005) 

0.01*** 

(0.005) 

-0.01 

(0.017) 

0.05 

(0.058) 

-0.01 

(0.013) 

  

VARIABLES 

  

Districts in Jharkhand 

Post 

 

 

Age of head of household 

Sex of head of household female 

 

(Base = Male) 

 

 

Household Size 

Rural 

(Base=urban) 

Religion of the HH (Muslim) 

(Base= Hindu) 

Religion of the HH (Christian) 

Religion of the HH (Sikh) 

 

Religion of the HH (Others) 

 

 

Caste of the household(Scheduled 
Caste) 
(Base= General Caste) 

 

-0.05*** 

-0.00*** 

-0.08*** 

-0.09*** 

-0.05*** 

-0.05*** 

-0.02*** 

-0.01*** 

-0.05*** 

(0.004) 

(0.002) 

(0.004) 

(0.004) 

(0.003) 

(0.004) 

(0.002) 

(0.002) 

(0.003) 

 
 

174 

 

Table 4.24 (cont’d) 
 
 
Caste of the household(Scheduled 
Tribe) 

 
 
 
 

 
 
 
 

 
 
 
 

 
 
 
 

 
 
 
 

 
 
 
 

 
 
 
 

 
 
 
 

 
 
 
 

-0.07*** 

-0.02*** 

-0.10*** 

-0.07*** 

-0.05*** 

-0.07*** 

-0.01*** 

-0.03*** 

-0.06*** 

 

 

Constant 

Observations 

R-squared 

(0.006) 

0.24*** 

(0.008) 

73,156 

0.310 

(0.002) 

0.08*** 

(0.003) 

73,156 

0.106 

(0.005) 

(0.006) 

0.34*** 

0.28*** 

(0.007) 

73,156 

0.316 

(0.007) 

73,156 

0.326 

(0.004) 

0.33*** 

(0.006) 

72,871 

0.331 

(0.006) 

(0.003) 

(0.004) 

(0.005) 

0.22*** 

0.03*** 

0.11*** 

-0.06*** 

(0.007) 

73,156 

0.279 

(0.004) 

73,156 

0.089 

(0.005) 

73,156 

0.166 

(0.007) 

73,156 

0.165 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village 

 

 

 
 

175 

 

Table 4. 25: SELECTED DEVELOPMENTAL INDICATORS- MADHYA PRADESH AS COMPARISON AND CHHATTISGARH 
AS TREATMENT  

  

(1) 

(2) 

Piped Drinking 

VARIABLES 

Electricity 

Water 

(4) 

(5) 

Clean 

(6) 

(7) 

(8) 

(9) 

Toilet 

Cooking Fuel 

Television 

Internet 

Refrigerator  Motorbike 

(3) 

Concrete 
Housing 

  

0.02** 

(0.007) 

0.17*** 

(0.004) 

0.00*** 

(0.000) 

0.00 

(0.005) 

0.01*** 

(0.000) 

0.00*** 

(0.001) 

  

0.02*** 

(0.006) 

0.11*** 

(0.003) 

0.00*** 

(0.000) 

0.01*** 

(0.004) 

0.01*** 

(0.000) 

  

0.05*** 

(0.007) 

0.16*** 

(0.004) 

0.00*** 

(0.000) 

0.01*** 

(0.005) 

0.01*** 

(0.000) 

0.01*** 

(0.001) 

  

0.12*** 

  

0.00 

(0.008) 

(0.005) 

0.29*** 

0.12*** 

(0.004) 

(0.003) 

0.00*** 

0.00*** 

(0.000) 

(0.000) 

  

0.00 

(0.006) 

0.11*** 

(0.003) 

0.00*** 

(0.000) 

  

-0.02** 

(0.008) 

0.28*** 

(0.004) 

0.00*** 

(0.000) 

-0.01* 

(0.005) 

0.00 

0.00 

(0.003) 

(0.004) 

-0.06*** 

(0.005) 

  

0.19*** 

(0.006) 

0.21*** 

(0.003) 

-0.00* 

(0.000) 

0.01 

(0.004) 

0.01*** 

(0.000) 

0.01*** 

(0.000) 

-0.09*** 

(0.003) 

0.02*** 

(0.005) 

-0.04*** 

(0.013) 

0.04 

(0.024) 

0.06*** 

(0.012) 

-0.03*** 

(0.003) 

  

0.09*** 

(0.005) 

0.13*** 

(0.003) 

0.00*** 

(0.000) 

-0.00 

(0.003) 

0.00*** 

(0.000) 

-0.00*** 

(0.000) 

-0.18*** 

(0.002) 

-0.02*** 

(0.005) 

0.02* 

(0.011) 

0.13*** 

(0.021) 

0.03*** 

(0.011) 

-0.02*** 

(0.003) 

  

Districts in Chhattisgarh 

Post 

 

 

Age of head of household 

 

Sex of head of household  
female 
(Base = Male) 
Education of the Household 
head 

 
Household Size 

 

Rural 

(Base=urban) 

Religion of the HH (Muslim) 

(Base= Hindu) 

Religion of the HH (Christian) 

Religion of the HH (Sikh) 

 

Religion of the HH (Others) 

 

 

Caste of the 
household(Scheduled Caste) 
(Base= General Caste) 
Caste of the 
household(Scheduled Tribe) 

 

0.01*** 

0.01*** 

(0.000) 

(0.000) 

-0.00*** 

0.03*** 

0.01*** 

(0.000) 

(0.001) 

(0.000) 

0.01*** 

(0.000) 

0.01*** 

(0.000) 

0.02*** 

(0.000) 

0.03*** 

(0.001) 

-0.39*** 

-0.44*** 

-0.49*** 

-0.29*** 

-0.13*** 

-0.25*** 

-0.15*** 

(0.003) 

0.01 

(0.006) 

0.08*** 

(0.016) 

0.15*** 

(0.029) 

0.11*** 

(0.003) 

0.10*** 

(0.006) 

0.12*** 

(0.016) 

0.15*** 

(0.028) 

0.15*** 

(0.003) 

-0.03*** 

(0.005) 

0.08*** 

(0.013) 

0.19*** 

(0.024) 

0.20*** 

(0.004) 

-0.01* 

(0.007) 

(0.002) 

-0.06*** 

(0.005) 

0.05*** 

0.06*** 

(0.018) 

(0.012) 

0.16*** 

0.12*** 

(0.032) 

(0.021) 

0.13*** 

0.08*** 

(0.003) 

0.01* 

(0.005) 

0.08*** 

(0.013) 

0.26*** 

(0.024) 

0.17*** 

(0.004) 

-0.07*** 

(0.007) 

0.07*** 

(0.017) 

0.21*** 

(0.032) 

0.14*** 

(0.015) 

(0.014) 

(0.012) 

(0.016) 

(0.011) 

(0.012) 

(0.016) 

-0.07*** 

-0.10*** 

-0.06*** 

-0.07*** 

-0.03*** 

-0.05*** 

-0.10*** 

(0.004) 

(0.003) 

(0.003) 

(0.004) 

(0.003) 

(0.003) 

(0.004) 

-0.06*** 

-0.05*** 

-0.16*** 

-0.16*** 

-0.10*** 

-0.19*** 

-0.05*** 

-0.07*** 

-0.14*** 

 
 

176 

 

Table 4. 25 (cont’d) 
 

 

 

 
 
Constant 

Observations 

R-squared 

(0.003) 

(0.003) 

(0.004) 

(0.004) 

(0.003) 

(0.004) 

(0.003) 

(0.003) 

(0.004) 

0.66*** 

0.10*** 

0.36*** 

0.38*** 

0.45*** 

0.25*** 

-0.01** 

0.03*** 

-0.04*** 

(0.006) 

83,314 

0.203 

(0.005) 

83,314 

0.193 

(0.007) 

83,314 

0.351 

(0.007) 

83,314 

0.419 

(0.006) 

83,295 

0.487 

(0.007) 

83,314 

0.321 

(0.005) 

83,314 

0.168 

(0.006) 

83,314 

0.261 

(0.007) 

83,314 

0.235 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

177 

 

 

Table 4. 26: SELECTED DEVELOPMENTAL INDICATORS- UTTAR PRADESH AS COMPARISON AND UTTARAKHAND 
AS TREATMENT  

  

(1) 

(2) 

Piped Drinking 

VARIABLES 

Electricity 

Water 

(3) 

Concrete 
Housing 

(4) 

(5) 

Clean 

(6) 

(7) 

(8) 

(9) 

Toilet 

Cooking Fuel 

Television 

Internet 

Refrigerator  Motorbike 

  

Districts in Uttarakhand 

Post 

 

 

Age of head of household 

 

Sex of head of household  
female 
(Base = Male) 
Education of the Household 
head 

 
Household Size 

 

Rural 

(Base=urban) 

Religion of the HH (Muslim) 

(Base= Hindu) 

Religion of the HH (Christian) 

 

Religion of the HH (Sikh) 

 

Religion of the HH (Others) 

 

Caste of the 
household(Scheduled Caste) 
(Base= General Caste) 

  

0.07*** 

(0.013) 

0.36*** 

(0.003) 

0.00*** 

(0.000) 

0.04*** 

(0.004) 

0.01*** 

(0.000) 

0.01*** 

(0.000) 

  

0.24*** 

(0.008) 

0.07*** 

(0.002) 

0.00*** 

(0.000) 

0.00** 

(0.002) 

0.00*** 

(0.000) 

-0.00*** 

(0.000) 

  

  

0.08*** 

0.24*** 

(0.012) 

(0.012) 

0.08*** 

0.18*** 

(0.003) 

(0.003) 

0.00*** 

0.00*** 

(0.000) 

(0.000) 

0.03*** 

0.02*** 

(0.004) 

(0.003) 

0.01*** 

0.01*** 

(0.000) 

(0.000) 

  

0.01 

(0.011) 

0.16*** 

(0.003) 

0.00*** 

(0.000) 

0.03*** 

(0.003) 

0.01*** 

(0.000) 

  

  

0.16*** 

0.19*** 

(0.014) 

(0.008) 

0.22*** 

0.07*** 

(0.003) 

(0.002) 

0.00*** 

0.00*** 

(0.000) 

(0.000) 

0.03*** 

0.01*** 

(0.004) 

(0.002) 

0.01*** 

0.01*** 

(0.000) 

(0.000) 

0.00*** 

0.00*** 

-0.01*** 

0.02*** 

0.00*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

  

0.10*** 

(0.011) 

0.13*** 

(0.003) 

0.00*** 

(0.000) 

0.02*** 

(0.003) 

0.01*** 

(0.000) 

0.01*** 

(0.000) 

  

-0.03** 

(0.013) 

0.26*** 

(0.003) 

0.00*** 

(0.000) 

-0.01*** 

(0.004) 

0.01*** 

(0.000) 

0.02*** 

(0.000) 

-0.27*** 

-0.20*** 

-0.46*** 

-0.46*** 

-0.50*** 

-0.35*** 

-0.09*** 

-0.31*** 

-0.12*** 

(0.003) 

0.05*** 

(0.004) 

0.08** 

(0.039) 

0.14*** 

(0.019) 

0.04 

(0.027) 

(0.002) 

0.01** 

(0.002) 

-0.05** 

(0.024) 

0.03*** 

(0.011) 

0.01 

(0.017) 

(0.003) 

(0.003) 

0.01*** 

0.17*** 

(0.003) 

0.17*** 

(0.003) 

0.08** 

(0.037) 

(0.036) 

0.14*** 

0.08*** 

(0.018) 

(0.017) 

0.09*** 

0.09*** 

(0.003) 

-0.01** 

(0.003) 

0.08** 

(0.035) 

0.13*** 

(0.017) 

0.18*** 

(0.003) 

(0.002) 

(0.003) 

(0.003) 

-0.11*** 

-0.03*** 

0.01*** 

-0.06*** 

(0.004) 

(0.002) 

0.07* 

0.02 

(0.042) 

(0.023) 

0.17*** 

0.06*** 

(0.020) 

(0.011) 

0.14*** 

-0.00 

(0.003) 

0.16*** 

(0.034) 

0.25*** 

(0.017) 

0.19*** 

(0.004) 

0.10** 

(0.041) 

0.17*** 

(0.020) 

0.14*** 

(0.026) 

(0.025) 

(0.024) 

(0.029) 

(0.016) 

(0.024) 

(0.028) 

-0.06*** 

-0.02*** 

-0.08*** 

-0.10*** 

-0.08*** 

-0.08*** 

-0.04*** 

-0.08*** 

-0.12*** 

(0.003) 

(0.002) 

(0.003) 

(0.003) 

(0.003) 

(0.003) 

(0.002) 

(0.003) 

(0.003) 

 

 
 

178 

 

Table 4. 26 (cont’d) 
 
 
Caste of the 
household(Scheduled Tribe) 

 
Constant 

 

Observations 

R-squared 

-0.12*** 

0.01** 

-0.09*** 

-0.10*** 

-0.05*** 

-0.09*** 

0.00 

-0.03*** 

-0.05*** 

(0.007) 

0.40*** 

(0.006) 

108,363 

0.347 

(0.004) 

0.14*** 

(0.004) 

108,363 

0.221 

(0.006) 

(0.006) 

0.46*** 

0.40*** 

(0.006) 

(0.006) 

108,363 

108,363 

0.369 

0.467 

(0.006) 

0.45*** 

(0.005) 

108,350 

0.431 

(0.007) 

(0.004) 

(0.006) 

(0.007) 

0.25*** 

-0.02*** 

0.10*** 

-0.12*** 

(0.006) 

(0.004) 

108,363 

108,363 

0.310 

0.144 

(0.005) 

108,361 

0.313 

(0.006) 

108,356 

0.202 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

179 

 

Table 4. 27:  SELECTED DEVELOPMENTAL INDICATORS- BORDERING DISTRICTS IN BIHAR AS COMPARISON AND 
BORDERING DISTRICTS IN JHARKHAND AS TREATMENT  

  

(1) 

(2) 

Piped Drinking 

VARIABLES 

Electricity 

Water 

(4) 

(5) 

Clean 

(6) 

(7) 

(8) 

(9) 

Toilet 

Cooking Fuel 

Television 

Internet 

Refrigerator  Motorbike 

  

  

Bordering Districts in Jharkhand 

0.13*** 

  

  

  

-0.01 

-0.02*** 

-0.00 

(3) 

Concrete 
Housing 

  

-0.00 

(0.011) 

0.12*** 

(0.007) 

0.00*** 

(0.000) 

0.04*** 

(0.006) 

0.02*** 

(0.000) 

0.00*** 

(0.001) 

  

0.09*** 

(0.011) 

0.16*** 

(0.007) 

0.00*** 

(0.000) 

0.04*** 

(0.006) 

0.02*** 

(0.000) 

0.00** 

(0.001) 

(0.006) 

0.03*** 

(0.007) 

-0.00 

(0.023) 

0.17 

(0.103) 

-0.08*** 

(0.021) 

(0.006) 

0.11*** 

(0.007) 

0.08*** 

(0.022) 

0.25** 

(0.099) 

-0.04* 

(0.020) 

  

0.01** 

(0.004) 

0.02*** 

(0.003) 

0.00*** 

(0.000) 

-0.00 

(0.002) 

0.00*** 

(0.000) 

-0.00*** 

(0.000) 

-0.10*** 

(0.002) 

-0.00 

(0.002) 

-0.01 

(0.008) 

0.04 

(0.037) 

-0.02** 

(0.008) 

-0.00** 

(0.002) 

(0.012) 

0.42*** 

(0.007) 

0.00*** 

(0.000) 

0.01* 

(0.007) 

0.01*** 

(0.000) 

0.01*** 

(0.001) 

-0.29*** 

(0.007) 

0.04*** 

(0.007) 

-0.02 

(0.025) 

0.23** 

(0.109) 

-0.08*** 

(0.023) 

-0.05*** 

(0.006) 

Post 

 

 

Age of head of household 

 

Sex of head of household  
female 
(Base = Male) 
Education of the Household 
head 

 

 

Household Size 

Rural 

(Base=urban) 

Religion of the HH (Muslim) 

(Base= Hindu) 

Religion of the HH (Christian) 

Religion of the HH (Sikh) 

 

Religion of the HH (Others) 

 

 

Caste of the 
household(Scheduled Caste) 
(Base= General Caste) 
Caste of the 
household(Scheduled Tribe) 

 

-0.01*** 

0.01*** 

0.00*** 

(0.001) 

(0.001) 

(0.000) 

-0.42*** 

-0.47*** 

-0.38*** 

-0.37*** 

-0.08*** 

-0.17*** 

-0.11*** 

  

0.05*** 

(0.008) 

0.11*** 

(0.005) 

0.00*** 

(0.000) 

0.03*** 

(0.005) 

0.01*** 

(0.000) 

(0.005) 

-0.01 

(0.005) 

0.02 

(0.017) 

0.19** 

(0.077) 

(0.011) 

(0.006) 

0.15*** 

0.07*** 

(0.007) 

(0.004) 

0.00*** 

0.00*** 

(0.000) 

(0.000) 

0.01 

(0.007) 

0.01*** 

(0.003) 

0.02*** 

0.00*** 

(0.000) 

(0.000) 

(0.006) 

(0.003) 

-0.05*** 

-0.00 

(0.007) 

(0.003) 

-0.01 

-0.00 

(0.024) 

(0.012) 

0.20* 

-0.03 

(0.106) 

(0.052) 

-0.02* 

-0.05*** 

-0.07*** 

(0.016) 

(0.022) 

(0.011) 

  

0.03*** 

(0.010) 

0.12*** 

(0.006) 

0.00*** 

(0.000) 

-0.01 

(0.006) 

0.01*** 

(0.000) 

0.02*** 

(0.001) 

(0.007) 

0.02*** 

(0.004) 

0.00*** 

(0.000) 

0.01*** 

(0.004) 

0.01*** 

(0.000) 

0.00*** 

(0.001) 

(0.004) 

0.01* 

(0.004) 

0.01 

(0.015) 

0.31*** 

(0.067) 

-0.03* 

(0.014) 

(0.006) 

-0.01** 

(0.006) 

0.08*** 

(0.022) 

0.25*** 

(0.097) 

-0.02 

(0.020) 

-0.10*** 

-0.10*** 

-0.05*** 

-0.08*** 

-0.02*** 

-0.02*** 

-0.07*** 

(0.006) 

(0.006) 

(0.004) 

(0.006) 

(0.003) 

(0.004) 

(0.005) 

-0.11*** 

-0.00 

-0.12*** 

-0.07*** 

-0.05*** 

-0.09*** 

-0.01*** 

-0.02*** 

-0.08*** 

 
 

180 

 

Table 4. 27 (cont’d) 
 
 
 
 
Constant 

 

Observations 

R-squared 

(0.009) 

(0.003) 

(0.008) 

(0.008) 

(0.006) 

(0.008) 

(0.004) 

(0.005) 

(0.008) 

0.30*** 

(0.012) 

31,505 

0.312 

0.08*** 

(0.004) 

31,505 

0.099 

0.36*** 

0.32*** 

0.30*** 

0.25*** 

(0.011) 

31,505 

0.277 

(0.011) 

31,505 

0.310 

(0.009) 

31,465 

0.282 

(0.012) 

31,505 

0.232 

0.01* 

(0.006) 

31,505 

0.072 

0.10*** 

-0.05*** 

(0.007) 

31,505 

0.110 

(0.011) 

31,505 

0.135 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

181 

 

Table 4. 28:  SELECTED DEVELOPMENTAL INDICATORS- BORDERING DISTRICTS IN MADHYA PRADESH AS 
COMPARISON AND BORDERING DISTRICTS IN CHHATTISGARH AS TREATMENT  

  

(1) 

(2) 

Piped Drinking 

VARIABLES 

Electricity 

Water 

(3) 

Concrete 
Housing 

(4) 

(5) 

Clean 

(6) 

(7) 

(8) 

(9) 

Toilet 

Cooking Fuel 

Television 

Internet 

Refrigerator  Motorbike 

  
Bordering Districts in 
Chhattisgarh 

Post 

 

 

  

  

  

  

  

  

  

  

  

0.08*** 

0.02*** 

0.13*** 

0.17*** 

0.03*** 

0.20*** 

0.07*** 

0.05*** 

0.12*** 

(0.012) 

(0.007) 

(0.012) 

(0.013) 

(0.010) 

(0.014) 

(0.008) 

(0.010) 

(0.013) 

0.38*** 

0.02*** 

0.05*** 

0.08*** 

0.04*** 

0.24*** 

0.03*** 

0.05*** 

0.15*** 

(0.010) 

(0.006) 

(0.009) 

(0.010) 

(0.008) 

(0.011) 

(0.007) 

(0.007) 

(0.010) 

Age of head of household 

0.00*** 

0.00*** 

0.00*** 

0.00*** 

0.00*** 

0.00*** 

0.00*** 

0.00*** 

0.00*** 

 

Sex of head of household  
female 

(Base = Male) 

Education of the Household 
head 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

0.01 

(0.009) 

-0.01** 

(0.005) 

0.01 

0.01 

0.02*** 

0.01 

0.01* 

0.01** 

-0.02** 

(0.008) 

(0.009) 

(0.007) 

(0.010) 

(0.006) 

(0.007) 

(0.009) 

0.01*** 

0.00*** 

0.01*** 

0.02*** 

0.01*** 

0.02*** 

0.01*** 

0.01*** 

0.02*** 

 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.001) 

(0.000) 

(0.000) 

(0.001) 

Household Size 

0.01*** 

-0.00*** 

0.00*** 

0.00*** 

-0.00*** 

0.03*** 

0.00*** 

0.01*** 

0.02*** 

 

Rural 

(Base=urban) 

Religion of the HH (Muslim) 

(Base= Hindu) 

Religion of the HH (Christian) 

 

Religion of the HH (Sikh) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

-0.10*** 

-0.21*** 

-0.44*** 

-0.42*** 

-0.45*** 

-0.29*** 

-0.14*** 

-0.27*** 

-0.21*** 

(0.007) 

(0.004) 

(0.007) 

(0.008) 

(0.006) 

(0.009) 

(0.005) 

(0.006) 

(0.008) 

0.07*** 

-0.02 

0.09*** 

0.14*** 

0.06*** 

0.06*** 

-0.02* 

0.02 

0.01 

(0.019) 

0.07** 

(0.030) 

0.19** 

(0.011) 

(0.018) 

(0.019) 

(0.015) 

(0.021) 

(0.013) 

(0.014) 

(0.020) 

0.02 

0.08*** 

0.11*** 

0.08*** 

0.02 

0.05** 

0.04 

0.05 

(0.018) 

0.15*** 

(0.029) 

(0.031) 

(0.024) 

(0.034) 

(0.021) 

(0.023) 

(0.033) 

0.24*** 

0.33*** 

0.30*** 

0.40*** 

0.08 

0.35*** 

0.28*** 

 

(0.084) 

(0.050) 

(0.080) 

(0.086) 

(0.066) 

(0.096) 

(0.057) 

(0.064) 

(0.091) 

 

 
 

182 

 

Table 4. 28 (cont’d) 
 
 
Religion of the HH (Others) 

 
Caste of the 
household(Scheduled Caste) 

(Base= General Caste) 

Caste of the 
household(Scheduled Tribe) 

0.07** 

-0.03 

0.02 

0.00 

0.06** 

0.03 

0.07*** 

0.00 

0.01 

(0.034) 

(0.020) 

(0.032) 

(0.034) 

(0.026) 

(0.038) 

(0.023) 

(0.026) 

(0.036) 

-0.03*** 

-0.02*** 

-0.07*** 

-0.05*** 

-0.05*** 

-0.04*** 

-0.01** 

-0.03*** 

-0.05*** 

(0.008) 

(0.004) 

(0.007) 

(0.008) 

(0.006) 

(0.009) 

(0.005) 

(0.006) 

(0.008) 

-0.08*** 

-0.00 

-0.11*** 

-0.09*** 

-0.06*** 

-0.14*** 

-0.02*** 

-0.04*** 

-0.09*** 

 

 

Constant 

Observations 

R-squared 

(0.006) 

(0.004) 

(0.006) 

(0.007) 

(0.005) 

(0.007) 

(0.004) 

(0.005) 

(0.007) 

0.43*** 

0.18*** 

0.36*** 

0.28*** 

0.38*** 

0.18*** 

0.04*** 

0.12*** 

-0.00 

(0.014) 

17,431 

0.303 

(0.008) 

17,431 

0.170 

(0.014) 

(0.015) 

(0.011) 

(0.016) 

(0.010) 

(0.011) 

(0.015) 

17,431 

17,431 

17,430 

17,431 

17,431 

17,431 

17,431 

0.368 

0.349 

0.400 

0.361 

0.144 

0.251 

0.247 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

183 

 

Table 4. 29: SELECTED DEVELOPMENTAL INDICATORS- BORDERING DISTRICTS IN UTTAR PRADESH AS 
COMPARISON AND BORDERING DISTRICTS IN UTTARAKHAND AS TREATMENT  

  

VARIABLES 

  
Bordering Districts in 
Uttarakhand 

Post 

 

 

Age of head of household 

 

Sex of head of household  
female 
(Base = Male) 
Education of the Household 
head 

 

 

Household Size 

Rural 

(Base=urban) 

Religion of the HH (Muslim) 

(Base= Hindu) 

Religion of the HH (Christian) 

Religion of the HH (Sikh) 

 

Religion of the HH (Others) 

 

 

Caste of the 
household(Scheduled Caste) 
(Base= General Caste) 
Caste of the 
household(Scheduled Tribe) 

 

(1) 

(2) 

Piped Drinking 

Electricity 

Water 

(3) 

Concrete 
Housing 

(4) 

(5) 

Clean 

(6) 

(7) 

(8) 

(9) 

Toilet 

Cooking Fuel 

Television 

Internet 

Refrigerator  Motorbike 

  

  

  

  

  

  

  

  

  

-0.01 

(0.014) 

0.34*** 

(0.007) 

0.00*** 

(0.000) 

0.03*** 

(0.008) 

0.01*** 

(0.000) 

0.01*** 

(0.001) 

-0.22*** 

(0.006) 

0.06*** 

(0.006) 

0.11* 

(0.055) 

0.14*** 

(0.022) 

0.17*** 

(0.040) 

-0.02*** 

(0.007) 

0.26*** 

(0.011) 

0.09*** 

(0.006) 

0.00*** 

(0.000) 

0.03*** 

(0.006) 

0.00*** 

(0.000) 

-0.00 

(0.001) 

-0.20*** 

(0.004) 

-0.02*** 

(0.005) 

0.07* 

(0.044) 

0.01 

(0.017) 

-0.05* 

(0.032) 

-0.04*** 

(0.005) 

0.06*** 

(0.016) 

0.21*** 

(0.008) 

0.00*** 

(0.000) 

0.03*** 

(0.009) 

0.01*** 

(0.000) 

0.00* 

(0.001) 

-0.02 

(0.014) 

0.35*** 

(0.007) 

0.00*** 

(0.000) 

0.05*** 

(0.008) 

0.01*** 

(0.000) 

0.01*** 

(0.001) 

0.06*** 

(0.015) 

0.21*** 

(0.008) 

0.00*** 

(0.000) 

0.05*** 

(0.009) 

0.01*** 

(0.000) 

-0.01*** 

(0.001) 

0.07*** 

0.28*** 

(0.017) 

(0.009) 

0.23*** 

0.02*** 

(0.009) 

(0.004) 

0.00*** 

0.00*** 

(0.000) 

(0.000) 

0.03*** 

(0.010) 

0.00 

(0.005) 

0.01*** 

0.00*** 

(0.000) 

0.02*** 

(0.001) 

(0.000) 

0.00** 

(0.001) 

0.06*** 

(0.016) 

0.20*** 

(0.008) 

0.00*** 

(0.000) 

0.04*** 

(0.009) 

0.02*** 

(0.000) 

0.01*** 

(0.001) 

-0.01 

(0.017) 

0.29*** 

(0.009) 

0.00*** 

(0.000) 

-0.01 

(0.010) 

0.02*** 

(0.000) 

0.02*** 

(0.001) 

-0.41*** 

-0.30*** 

-0.44*** 

-0.29*** 

-0.04*** 

-0.27*** 

-0.07*** 

(0.006) 

0.01 

(0.007) 

0.09 

(0.063) 

0.16*** 

(0.025) 

0.23*** 

(0.006) 

0.21*** 

(0.006) 

0.10* 

(0.056) 

0.10*** 

(0.022) 

0.23*** 

(0.006) 

-0.01** 

(0.007) 

0.12** 

(0.060) 

0.12*** 

(0.023) 

0.33*** 

(0.007) 

(0.003) 

-0.20*** 

-0.03*** 

(0.007) 

0.14** 

(0.065) 

0.16*** 

(0.026) 

0.16*** 

(0.004) 

0.10*** 

(0.034) 

0.02 

(0.013) 

-0.04* 

(0.006) 

0.02** 

(0.007) 

0.22*** 

(0.063) 

0.27*** 

(0.025) 

0.18*** 

(0.046) 

(0.041) 

(0.042) 

(0.047) 

(0.024) 

(0.045) 

(0.007) 

-0.05*** 

(0.008) 

0.22*** 

(0.067) 

0.19*** 

(0.027) 

0.14*** 

(0.049) 

-0.10*** 

-0.14*** 

-0.10*** 

-0.05*** 

-0.04*** 

-0.12*** 

-0.11*** 

(0.008) 

(0.007) 

(0.007) 

(0.008) 

(0.004) 

(0.008) 

(0.008) 

-0.01 

-0.02** 

-0.11*** 

-0.17*** 

-0.11*** 

-0.09*** 

-0.02*** 

-0.08*** 

-0.05*** 

 
 

184 

 

Table 4. 29 (cont’d) 
 
 

Constant 

 

Observations 

R-squared 

(0.013) 

0.47*** 

(0.012) 

21,516 

0.297 

(0.011) 

0.11*** 

(0.010) 

21,516 

0.274 

(0.015) 

0.47*** 

(0.014) 

21,516 

0.324 

(0.014) 

0.37*** 

(0.013) 

21,516 

0.384 

(0.014) 

0.45*** 

(0.013) 

21,509 

0.386 

(0.016) 

(0.008) 

0.27*** 

-0.03*** 

(0.015) 

21,516 

0.277 

(0.008) 

21,516 

0.252 

(0.015) 

-0.01 

(0.014) 

21,516 

0.256 

(0.016) 

-0.16*** 

(0.015) 

21,515 

0.170 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

185 

 

Table 4. 30: SELECTED DEVELOPMENTAL INDICATORS- NEARBY STATES AS COMPARISON AND BOTH BIHAR AND 
JHARKHAND AS TREATMENTS  

  

(1) 

(2) 

Piped Drinking 

VARIABLES 

Electricity 

Water 

  

Jharkhand X Post 

Bihar X Post 

Post 

 

 

 

Age of head of household 

 

Sex of head of household  
female 
(Base = Male) 
Education of the Household 
head 

 
Household Size 

 

Rural 

(Base=urban) 

Religion of the HH (Muslim) 

(Base= Hindu) 
Religion of the HH 
(Christian) 

 

Religion of the HH (Sikh) 

 

Religion of the HH (Others) 

 

Caste of the 
household(Scheduled Caste) 
(Base= General Caste) 

  

-0.01 

(0.009) 

-0.12*** 

(0.005) 

0.55*** 

(0.004) 

0.00*** 

(0.000) 

-0.01** 

(0.003) 

0.01*** 

(0.000) 

0.01*** 

(0.000) 

-0.19*** 

(0.003) 

-0.02*** 

(0.003) 

0.00 

(0.011) 

0.07 

(0.044) 

0.01 

(0.008) 

-0.04*** 

(0.003) 

 

(3) 

Concrete 
Housing 

  

0.09*** 

(0.009) 

(4) 

(5) 

(6) 

(7) 

(8) 

(9) 

Clean Cooking 

Toilet 

  

0.19*** 

(0.009) 

Fuel 

  

0.05*** 

(0.007) 

Television 

Internet 

Refrigerator  Motorbike 

  

  

0.21*** 

-0.02*** 

(0.009) 

(0.005) 

  

0.07*** 

(0.006) 

  

-0.01 

(0.008) 

-0.12*** 

-0.10*** 

-0.03*** 

-0.25*** 

-0.01*** 

-0.09*** 

-0.05*** 

(0.006) 

0.22*** 

(0.004) 

0.00*** 

(0.000) 

0.02*** 

(0.003) 

0.01*** 

(0.000) 

0.00*** 

(0.000) 

(0.006) 

0.26*** 

(0.004) 

0.00*** 

(0.000) 

0.00 

(0.003) 

0.01*** 

(0.000) 

0.00*** 

(0.000) 

(0.005) 

0.12*** 

(0.003) 

0.00*** 

(0.000) 

0.01*** 

(0.003) 

0.01*** 

(0.000) 

-0.01*** 

(0.000) 

(0.006) 

(0.003) 

0.38*** 

0.07*** 

(0.004) 

(0.002) 

0.00*** 

0.00*** 

(0.000) 

(0.000) 

-0.00 

(0.003) 

0.01*** 

(0.002) 

0.01*** 

0.00*** 

(0.000) 

(0.000) 

0.02*** 

0.00*** 

(0.000) 

(0.000) 

(0.004) 

0.10*** 

(0.003) 

0.00*** 

(0.000) 

0.01*** 

(0.002) 

0.01*** 

(0.000) 

0.00*** 

(0.000) 

(0.005) 

0.17*** 

(0.003) 

0.00*** 

(0.000) 

-0.02*** 

(0.003) 

0.01*** 

(0.000) 

0.02*** 

(0.000) 

  

0.01*** 

(0.004) 

-0.02*** 

(0.002) 

0.04*** 

(0.002) 

0.00*** 

(0.000) 

-0.00 

(0.001) 

0.00*** 

(0.000) 

-0.00 

(0.000) 

-0.07*** 

-0.35*** 

-0.33*** 

-0.38*** 

-0.30*** 

-0.09*** 

-0.19*** 

-0.11*** 

(0.001) 

0.00** 

(0.001) 

-0.01 

(0.005) 

0.09*** 

(0.020) 

-0.02*** 

(0.004) 

-0.01*** 

(0.001) 

(0.003) 

-0.04*** 

(0.003) 

0.01 

(0.011) 

0.12*** 

(0.046) 

-0.08*** 

(0.008) 

(0.003) 

0.03*** 

(0.003) 

0.04*** 

(0.011) 

0.16*** 

(0.045) 

0.01* 

(0.008) 

(0.002) 

(0.003) 

(0.002) 

(0.002) 

(0.003) 

-0.07*** 

-0.11*** 

-0.01*** 

-0.02*** 

-0.02*** 

(0.003) 

(0.004) 

(0.002) 

(0.002) 

(0.003) 

-0.02* 

(0.009) 

0.18*** 

(0.037) 

0.01 

-0.00 

0.01 

(0.011) 

(0.006) 

0.07 

-0.02 

(0.008) 

0.19*** 

(0.047) 

(0.024) 

(0.032) 

-0.06*** 

-0.09*** 

-0.02*** 

-0.02*** 

(0.007) 

(0.009) 

(0.004) 

(0.006) 

0.03** 

(0.010) 

0.15*** 

(0.041) 

-0.01 

(0.008) 

-0.11*** 

-0.09*** 

-0.07*** 

-0.06*** 

-0.02*** 

-0.04*** 

-0.07*** 

(0.003) 

(0.003) 

(0.002) 

(0.003) 

(0.001) 

(0.002) 

(0.002) 

 
 

186 

 

Table 4. 30 (cont’d) 
 
 
Caste of the 
household(Scheduled Tribe) 

 

 

Constant 

Observations 

R-squared 

-0.11*** 

-0.02*** 

-0.16*** 

-0.17*** 

-0.09*** 

-0.14*** 

-0.02*** 

-0.06*** 

-0.10*** 

(0.004) 

0.32*** 

(0.005) 

133,064 

0.390 

(0.002) 

0.04*** 

(0.002) 

133,064 

0.086 

(0.004) 

0.34*** 

(0.006) 

133,064 

0.313 

(0.004) 

0.35*** 

(0.006) 

133,064 

0.392 

(0.003) 

0.32*** 

(0.005) 

131,563 

0.338 

(0.004) 

(0.002) 

0.22*** 

0.01*** 

(0.006) 

(0.003) 

133,064 

133,064 

0.322 

0.087 

(0.003) 

0.08*** 

(0.004) 

133,063 

0.210 

(0.004) 

-0.05*** 

(0.005) 

133,060 

0.154 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

187 

 

Table 4. 31: SELECTED DEVELOPMENTAL INDICATORS- NEARBY STATES AS COMPARISON AND BOTH MADHYA 
PRADESH AND CHHATTISGARH AS TREATMENTS 

  

(1) 

(2) 

Piped Drinking 

VARIABLES 

Electricity 

Water 

-0.04*** 

-0.02*** 

-0.08*** 

-0.11*** 

-0.08*** 

-0.07*** 

-0.02*** 

-0.08*** 

-0.11*** 

 
 

188 

(3) 

Concrete 
Housing 

  

0.14*** 

(0.008) 

(4) 

(5) 

Clean 

(6) 

(7) 

(8) 

(9) 

Toilet 

Cooking Fuel 

Television 

Internet 

Refrigerator  Motorbike 

  

0.05*** 

(0.008) 

  

0.14*** 

(0.007) 

  

  

0.03*** 

0.08*** 

(0.008) 

(0.005) 

  

0.08*** 

(0.007) 

  

-0.04*** 

(0.008) 

-0.15*** 

-0.10*** 

-0.11*** 

-0.15*** 

0.07*** 

-0.08*** 

-0.02*** 

(0.005) 

0.33*** 

(0.002) 

0.00*** 

(0.000) 

(0.005) 

0.27*** 

(0.002) 

0.00*** 

(0.000) 

0.00 

0.00 

(0.003) 

(0.003) 

0.01*** 

(0.000) 

0.01*** 

(0.000) 

0.01*** 

(0.000) 

0.00*** 

(0.000) 

(0.005) 

0.22*** 

(0.002) 

0.00*** 

(0.000) 

0.00* 

(0.002) 

0.01*** 

(0.000) 

(0.005) 

(0.003) 

0.45*** 

0.05*** 

(0.003) 

(0.001) 

0.00*** 

0.00*** 

(0.000) 

(0.000) 

-0.02*** 

(0.003) 

0.00** 

(0.002) 

0.01*** 

0.01*** 

(0.000) 

(0.000) 

-0.00*** 

0.03*** 

0.00*** 

(0.000) 

(0.000) 

(0.000) 

(0.005) 

0.19*** 

(0.002) 

0.00*** 

(0.000) 

-0.00 

(0.002) 

0.01*** 

(0.000) 

0.01*** 

(0.000) 

(0.005) 

0.31*** 

(0.003) 

0.00*** 

(0.000) 

-0.07*** 

(0.003) 

0.02*** 

(0.000) 

0.03*** 

(0.000) 

-0.28*** 

-0.37*** 

-0.42*** 

-0.22*** 

-0.08*** 

-0.22*** 

-0.13*** 

(0.002) 

0.01*** 

(0.004) 

(0.002) 

0.07*** 

(0.004) 

(0.002) 

-0.00 

(0.003) 

(0.002) 

(0.001) 

-0.05*** 

-0.04*** 

(0.004) 

(0.002) 

(0.002) 

0.00 

(0.003) 

0.01* 

-0.01 

0.00 

0.00 

0.00 

0.00 

(0.006) 

0.08*** 

(0.016) 

0.10*** 

(0.006) 

0.13*** 

(0.015) 

0.13*** 

(0.005) 

0.12*** 

(0.014) 

0.20*** 

(0.006) 

(0.003) 

0.11*** 

0.06*** 

(0.016) 

(0.009) 

0.15*** 

0.08*** 

(0.005) 

0.23*** 

(0.014) 

0.18*** 

(0.012) 

(0.012) 

(0.011) 

(0.012) 

(0.007) 

(0.011) 

(0.002) 

-0.07*** 

(0.004) 

-0.00 

(0.006) 

0.18*** 

(0.016) 

0.10*** 

(0.012) 

  

0.09*** 

(0.006) 

-0.10*** 

(0.004) 

0.31*** 

(0.002) 

0.00*** 

(0.000) 

0.00 

(0.002) 

0.01*** 

(0.000) 

0.01*** 

(0.000) 

-0.07*** 

(0.002) 

0.00 

(0.003) 

0.00 

(0.005) 

0.04*** 

(0.012) 

0.06*** 

(0.009) 

  

0.09*** 

(0.006) 

-0.00 

(0.004) 

0.14*** 

(0.002) 

0.00*** 

(0.000) 

-0.00 

(0.002) 

0.00*** 

(0.000) 

-0.00*** 

(0.000) 

-0.15*** 

(0.002) 

-0.01*** 

(0.003) 

-0.00 

(0.004) 

0.05*** 

(0.011) 

0.06*** 

(0.009) 

  

Chhattisgarh X Post 

 

Madhya Pradesh X Post 

Post 

 

 

Age of head of household 

 

Sex of head of household  
female 
(Base = Male) 
Education of the Household 
head 

 
Household Size 

 

Rural 

(Base=urban) 

Religion of the HH (Muslim) 

(Base= Hindu) 
Religion of the HH 
(Christian) 

 

Religion of the HH (Sikh) 

 

Religion of the HH (Others) 

 

Caste of the 
household(Scheduled Caste) 

 

 

Table 4. 31 (cont’d) 
 
 
(Base= General Caste) 
Caste of the 
household(Scheduled Tribe) 

 

 

Constant 

Observations 

R-squared 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

(0.001) 

(0.002) 

(0.002) 

-0.09*** 

-0.05*** 

-0.18*** 

-0.18*** 

-0.14*** 

-0.19*** 

-0.03*** 

-0.10*** 

-0.13*** 

(0.002) 

0.61*** 

(0.003) 

221,861 

0.254 

(0.002) 

0.07*** 

(0.003) 

221,861 

0.222 

(0.003) 

0.41*** 

(0.004) 

221,861 

0.380 

(0.003) 

0.36*** 

(0.004) 

221,861 

0.415 

(0.002) 

0.42*** 

(0.004) 

221,485 

0.459 

(0.003) 

(0.002) 

0.21*** 

-0.02*** 

(0.004) 

(0.002) 

221,861 

221,861 

0.339 

0.143 

(0.002) 

-0.00 

(0.004) 

221,861 

0.292 

(0.003) 

-0.04*** 

(0.005) 

221,861 

0.244 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

189 

 

Table 4. 32: SELECTED DEVELOPMENTAL INDICATORS- NEARBY STATES AS COMPARISON AND BOTH UTTAR 
PRADESH AND UTTARAKHAND AS TREATMENTS 

  

(1) 

(2) 

Piped Drinking 

VARIABLES 

Electricity 

Water 

(3) 

Concrete 
Housing 

(4) 

(5) 

Clean 

(6) 

(7) 

(8) 

(9) 

Toilet 

Cooking Fuel 

Television 

Internet 

Refrigerator  Motorbike 

  

  

  

  

  

  

  

  

  

  

Uttarakhand X Post 

0.17*** 

0.08*** 

0.09*** 

0.06*** 

0.01 

-0.00 

0.11*** 

0.08*** 

0.15*** 

 

(0.011) 

(0.009) 

(0.012) 

(0.012) 

(0.011) 

(0.013) 

(0.008) 

(0.011) 

(0.013) 

Uttar Pradesh X Post 

-0.10*** 

-0.16*** 

-0.17*** 

-0.19*** 

-0.00 

-0.16*** 

-0.08*** 

-0.19*** 

-0.12*** 

Post 

 

 

(0.004) 

(0.003) 

(0.004) 

(0.004) 

(0.004) 

(0.005) 

(0.003) 

(0.004) 

(0.005) 

0.25*** 

0.23*** 

0.26*** 

0.36*** 

0.15*** 

0.38*** 

0.16*** 

0.31*** 

0.38*** 

(0.003) 

(0.003) 

(0.004) 

(0.003) 

(0.003) 

(0.004) 

(0.002) 

(0.003) 

(0.004) 

Age of head of household 

0.00*** 

0.00*** 

0.00*** 

0.00*** 

0.00*** 

0.00*** 

0.00*** 

0.00*** 

0.00*** 

 

Sex of head of household  
female 

(Base = Male) 

Education of the 
Household head 

 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

0.04*** 

(0.003) 

0.00** 

(0.002) 

0.04*** 

0.02*** 

0.04*** 

0.03*** 

0.01*** 

0.03*** 

-0.02*** 

(0.003) 

(0.003) 

(0.003) 

(0.003) 

(0.002) 

(0.003) 

(0.003) 

0.01*** 

0.00*** 

0.01*** 

0.01*** 

0.01*** 

0.01*** 

0.01*** 

0.01*** 

0.01*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

Household Size 

0.01*** 

-0.00*** 

0.00*** 

0.00*** 

-0.01*** 

0.02*** 

0.00*** 

0.01*** 

0.02*** 

 

Rural 

(Base=urban) 

Religion of the HH 
(Muslim) 

(Base= Hindu) 

Religion of the HH 
(Christian) 

 

 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

-0.20*** 

-0.20*** 

-0.39*** 

-0.41*** 

-0.50*** 

-0.30*** 

-0.10*** 

-0.27*** 

-0.11*** 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

0.04*** 

-0.00 

-0.00 

0.14*** 

-0.01*** 

-0.11*** 

-0.04*** 

0.01** 

-0.06*** 

(0.003) 

(0.002) 

(0.003) 

(0.003) 

(0.003) 

(0.003) 

(0.002) 

(0.003) 

(0.003) 

0.06** 

-0.02 

0.12*** 

0.10*** 

0.08*** 

0.07** 

-0.01 

0.11*** 

0.04 

(0.027) 

(0.021) 

(0.029) 

(0.028) 

(0.026) 

(0.031) 

(0.019) 

(0.027) 

(0.031) 

 
 

190 

 

Table 4. 32 (cont’d) 
 
 
Religion of the HH (Sikh) 

 

Religion of the HH 
(Others) 

 

Caste of the 
household(Scheduled 
Caste) 

(Base= General Caste) 

Caste of the 
household(Scheduled 
Tribe) 

0.06*** 

0.01 

0.09*** 

0.09*** 

0.11*** 

0.10*** 

0.04*** 

0.18*** 

0.13*** 

(0.009) 

(0.007) 

(0.009) 

(0.009) 

(0.008) 

(0.010) 

(0.006) 

(0.009) 

(0.010) 

0.03* 

0.05*** 

0.06*** 

0.14*** 

0.21*** 

0.14*** 

0.05*** 

0.19*** 

0.11*** 

(0.014) 

(0.011) 

(0.015) 

(0.015) 

(0.014) 

(0.016) 

(0.010) 

(0.014) 

(0.016) 

-0.06*** 

-0.02*** 

-0.09*** 

-0.09*** 

-0.08*** 

-0.07*** 

-0.04*** 

-0.10*** 

-0.13*** 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

-0.11*** 

-0.03*** 

-0.14*** 

-0.16*** 

-0.11*** 

-0.15*** 

-0.02*** 

-0.09*** 

-0.09*** 

Constant 

 

 

Observations 

R-squared 

(0.004) 

(0.003) 

(0.004) 

(0.004) 

(0.004) 

(0.004) 

(0.003) 

(0.004) 

(0.004) 

0.48*** 

0.13*** 

0.46*** 

0.38*** 

0.48*** 

0.25*** 

-0.03*** 

0.05*** 

-0.11*** 

(0.004) 

(0.003) 

(0.005) 

(0.004) 

(0.004) 

(0.005) 

(0.003) 

(0.004) 

(0.005) 

192,885 

192,885 

192,885 

192,885 

192,843 

192,885 

192,885 

192,883 

192,878 

0.358 

0.269 

0.379 

0.459 

0.432 

0.345 

0.176 

0.339 

0.232 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village 

 

 

 

 

 

 

 
 

191 

 

Table 4. 33: SELECTED HEALTH INDICATORS- BIHAR AS COMPARISON AND JHARKHAND AS TREATMENT  

 VARIABLES 

  
Districts in Jharkhand  X 
Post 

post 

 

 

Age of head of household 

 

Sex of head of household  
female 

(Base = Male) 
Years of Education 
Women has Received 

 

Rural 

(Base=urban) 

Household Size 

 

Religion of the HH 
(Muslim) 

(Base= Hindu) 
Religion of the HH 
(Christian) 

 

Religion of the HH (Sikh) 

 

Religion of the HH 
(Others) 

 

 

(1) 

Govt 

Institutional 

Delivery 

(2) 

(3) 

(4) 

(5) 

(6) 

(7) 

(8) 

(9) 

(10) 

Infant 
Death 

Underweight 

Stunted 

Wasted 

Tetanus 

Shot 

Polio 

Vaccine 

Measles 
Vaccine 

BCG 

Vaccine 

Health 
Card 

  

  

  

  

  

  

  

  

  

  

0.05*** 

(0.016) 

0.03*** 

(0.008) 

-0.14*** 

-0.07*** 

-0.12*** 

0.05*** 

0.01 

0.01 

0.02* 

0.09*** 

(0.019) 

(0.019) 

(0.017) 

(0.017) 

(0.017) 

(0.016) 

(0.013) 

(0.016) 

0.43*** 

-0.04*** 

-0.15*** 

-0.15*** 

-0.08*** 

0.12*** 

0.33*** 

0.53*** 

0.57*** 

0.35*** 

(0.007) 

-0.00 

(0.000) 

(0.003) 

0.00*** 

(0.000) 

(0.008) 

(0.008) 

(0.007) 

(0.007) 

(0.007) 

(0.007) 

(0.005) 

(0.007) 

-0.00*** 

-0.00*** 

0.00 

0.00*** 

-0.00 

-0.00** 

0.00 

0.00 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

-0.02*** 

-0.01*** 

(0.006) 

(0.003) 

-0.00 

(0.007) 

-0.00 

-0.01 

0.01 

0.00 

0.00 

0.02*** 

0.01 

(0.007) 

(0.006) 

(0.006) 

(0.006) 

(0.006) 

(0.005) 

(0.006) 

0.00 

-0.00*** 

-0.01*** 

-0.02*** 

-0.00*** 

0.01*** 

0.01*** 

0.01*** 

0.01*** 

0.01*** 

(0.001) 

0.03*** 

(0.007) 

-0.00** 

(0.001) 

-0.09*** 

(0.006) 

-0.06** 

(0.030) 

-0.05 

(0.089) 

-0.03 

(0.020) 

(0.000) 

0.01*** 

(0.004) 

-0.01*** 

(0.000) 

-0.00 

(0.003) 

0.02 

(0.015) 

0.02 

(0.045) 

0.02 

(0.010) 

(0.001) 

0.04*** 

(0.008) 

-0.00** 

(0.001) 

-0.01 

(0.007) 

-0.00 

(0.034) 

0.00 

(0.098) 

0.05** 

(0.023) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.000) 

(0.001) 

0.05*** 

0.01* 

-0.05*** 

-0.01* 

-0.02** 

-0.04*** 

0.01** 

(0.008) 

(0.007) 

(0.008) 

(0.007) 

(0.007) 

(0.006) 

(0.007) 

-0.00* 

-0.00** 

-0.01*** 

-0.00 

-0.00 

0.00 

0.00 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

0.00 

0.01 

-0.02** 

-0.05*** 

-0.07*** 

-0.06*** 

-0.05*** 

(0.007) 

(0.006) 

(0.007) 

(0.006) 

(0.006) 

(0.005) 

(0.006) 

-0.03 

-0.01 

0.02 

-0.02 

0.00 

0.01 

-0.02 

(0.034) 

(0.030) 

(0.032) 

(0.030) 

(0.029) 

(0.024) 

(0.030) 

-0.00 

-0.08 

0.03 

-0.15 

0.01 

0.06 

0.15* 

(0.098) 

(0.086) 

(0.095) 

(0.091) 

(0.088) 

(0.072) 

(0.089) 

0.01 

0.03 

-0.03 

0.02 

-0.02 

0.00 

0.01 

(0.023) 

(0.020) 

(0.021) 

(0.020) 

(0.020) 

(0.016) 

(0.020) 

 
 

192 

 

 
Table 4. 33 (cont’d) 
 
 
Caste of the 
household(Scheduled 
Caste) 

(Base= General Caste) 
Caste of the 
household(Scheduled 
Tribe) 

 

Constant 

 
Observations 

R-squared 

-0.01* 

(0.006) 

0.01*** 

(0.003) 

0.06*** 

(0.007) 

0.06*** 

0.02*** 

-0.02*** 

-0.01 

-0.03*** 

-0.01* 

-0.02*** 

(0.007) 

(0.006) 

(0.006) 

(0.006) 

(0.006) 

(0.005) 

(0.006) 

-0.04*** 

0.00 

(0.008) 

0.05*** 

(0.011) 

43,471 

0.143 

(0.004) 

0.09*** 

(0.006) 

42,711 

0.011 

0.02 

(0.010) 

0.63*** 

(0.013) 

39,312 

0.047 

0.01 

0.04*** 

0.02* 

-0.04*** 

-0.02*** 

-0.03*** 

-0.02** 

(0.010) 

(0.008) 

(0.009) 

(0.009) 

(0.008) 

(0.007) 

(0.008) 

0.68*** 

0.29*** 

0.49*** 

0.11*** 

0.15*** 

0.28*** 

0.26*** 

(0.013) 

(0.011) 

(0.012) 

(0.011) 

(0.011) 

(0.009) 

(0.011) 

39,242 

38,970 

43,471 

43,471 

43,471 

43,471 

43,471 

0.055 

0.019 

0.047 

0.107 

0.195 

0.299 

0.139 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village 

 

 

 

 

 

 

 

 

 

 

 

 
 

193 

 

Table 4. 34: SELECTED HEALTH INDICATORS- MADHYA PRADESH AS COMPARISON AND CHHATTISGARH AS 
TREATMENT  

  

(1) 

(2) 

(3) 

(4) 

(5) 

(6) 

(7) 

(8) 

(9) 

(10) 

VARIABLES 

Govt 

Institutional 

Delivery 

Infant 
Death 

Underweight 

Stunted 

Wasted 

Tetanus 

Shot 

Polio 

Vaccine 

Measles 
Vaccine 

BCG 

Vaccine 

Health 
Card 

  

  

  

  

  

  

  

  

  

  

-0.03*** 

-0.09*** 

-0.09*** 

-0.03 

0.06*** 

0.12*** 

0.00 

0.02* 

-0.04*** 

(0.007) 

(0.015) 

(0.022) 

(0.020) 

-0.04*** 

-0.09*** 

-0.13*** 

-0.01 

(0.013) 

0.02** 

(0.007) 

(0.013) 

(0.013) 

(0.011) 

(0.011) 

0.27*** 

0.40*** 

0.35*** 

0.41*** 

(0.007) 

(0.007) 

(0.006) 

(0.006) 

(0.010) 

(0.011) 

-0.00** 

(0.000) 

-0.00 

0.00*** 

0.00 

-0.00 

-0.00*** 

-0.00** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.004) 

0.00*** 

(0.000) 

0.00 

(0.005) 

(0.008) 

-0.00*** 

(0.000) 

-0.01 

(0.011) 

0.01 

0.01 

-0.01 

-0.01 

-0.01 

0.01 

0.00 

(0.011) 

(0.010) 

(0.010) 

(0.010) 

(0.010) 

(0.008) 

(0.009) 

-0.00*** 

-0.01*** 

-0.01*** 

-0.00*** 

0.01*** 

0.01*** 

0.01*** 

0.01*** 

0.01*** 

  
Districts in Chhattisgarh X 
Post 

post 

 

 

Age of head of household 

 

Sex of head of household  
female 
(Base = Male) 
Years of Education Women 
has Received 

 

Rural 

(Base=urban) 

Household Size 

 

Religion of the HH 
(Muslim) 
(Base= Hindu) 
Religion of the HH 
(Christian) 

 

Religion of the HH (Sikh) 

 

Religion of the HH (Others) 

 

 
 

 

0.05*** 

(0.012) 

0.55*** 

(0.007) 

-0.00** 

(0.000) 

0.03*** 

(0.009) 

0.00*** 

(0.001) 

0.00 

(0.006) 

(0.000) 

0.01 

(0.003) 

-0.00*** 

-0.01*** 

(0.001) 

(0.000) 

0.05*** 

(0.010) 

-0.09*** 

(0.032) 

-0.18*** 

(0.062) 

-0.06** 

(0.032) 

 
 

0.00 

(0.006) 

-0.01 

(0.017) 

-0.05 

(0.034) 

0.01 

(0.017) 

 
 

(0.001) 

0.03*** 

(0.007) 

-0.00** 

(0.001) 

-0.02 

(0.012) 

-0.03 

(0.038) 

-0.09 

(0.072) 

0.04 

(0.038) 

 
 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.000) 

(0.001) 

0.02*** 

0.02*** 

-0.05*** 

-0.05*** 

-0.04*** 

-0.04*** 

-0.06*** 

(0.007) 

-0.00* 

(0.001) 

-0.00 

(0.013) 

(0.007) 

(0.006) 

(0.006) 

(0.006) 

(0.005) 

-0.00 

-0.01*** 

-0.00** 

0.00*** 

0.00*** 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.006) 

0.00** 

(0.001) 

-0.02* 

(0.011) 

0.02* 

-0.00 

0.00 

-0.00 

0.00 

(0.011) 

(0.011) 

(0.011) 

(0.009) 

(0.009) 

-0.12*** 

0.01 

0.02 

(0.040) 

(0.036) 

(0.034) 

-0.04 

0.02 

-0.06* 

(0.034) 

0.09 

-0.01 

0.01 

-0.03 

(0.033) 

(0.027) 

(0.029) 

0.11 

-0.01 

0.02 

(0.068) 

(0.067) 

(0.066) 

(0.065) 

(0.054) 

(0.058) 

-0.03 

0.07* 

0.04 

0.10*** 

0.08*** 

0.05 

(0.041) 

(0.034) 

(0.034) 

(0.033) 

(0.027) 

(0.029) 

 
 

 
 

 
 

 
 

 
 

 
 

-0.03 

(0.075) 

-0.08* 

(0.044) 

 
 

 
 

194 

 

Table 4. 34 (cont’d) 
 
 
Caste of the 
household(Scheduled 
Caste) 
(Base= General Caste) 
Caste of the 
household(Scheduled 
Tribe) 

 

 

Constant 

Observations 

R-squared 

 
 
 
 
 

 
 
 
 
 

 
 
 
 
 

 
 
 
 
 

 
 
 
 
 

 
 
 
 
 

 
 
 
 
 

 
 
 
 
 

 
 
 
 
 

 
 
 
 
 

0.01 

0.00 

0.03*** 

0.06*** 

0.01 

-0.05*** 

-0.03*** 

-0.03*** 

-0.03*** 

-0.03*** 

(0.006) 

(0.003) 

(0.007) 

(0.007) 

(0.007) 

(0.006) 

(0.006) 

(0.006) 

(0.005) 

(0.005) 

-0.07*** 

(0.006) 

0.15*** 

(0.010) 

39,555 

0.244 

0.01** 

(0.003) 

0.11*** 

(0.006) 

39,004 

0.015 

0.05*** 

(0.007) 

0.61*** 

(0.012) 

35,647 

0.050 

0.06*** 

0.03*** 

-0.02*** 

-0.08*** 

-0.06*** 

-0.04*** 

-0.07*** 

(0.007) 

(0.007) 

(0.007) 

(0.006) 

(0.006) 

(0.005) 

(0.006) 

0.63*** 

0.29*** 

0.61*** 

0.19*** 

0.28*** 

0.50*** 

0.40*** 

(0.015) 

32,613 

0.044 

(0.013) 

32,439 

0.015 

(0.011) 

39,555 

0.057 

(0.011) 

39,555 

0.129 

(0.011) 

39,555 

0.148 

(0.009) 

39,555 

0.184 

(0.009) 

39,555 

0.235 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village 

 

 

 

 

 

 

 

 

 

 

 

 
 

195 

 

Table 4. 35: SELECTED HEALTH INDICATORS- UTTAR PRADESH AS COMPARISON AND UTTARAKHAND AS 
TREATMENT  

  

VARIABLES 

(1) 
Govt 

Institutional 

Delivery 

(2) 

Infant 
Death 

(3) 

(4) 

(5) 

(6) 

(7) 

(8) 

(9) 

(10) 

Underweight 

Stunted 

Wasted 

Tetanus 

Shot 

Polio 

Vaccine 

Measles 
Vaccine 

BCG 

Vaccine 

Health 
Card 

  
Districts in Uttarakhand X 
Post 

post 

 

 

Age of head of household 

 

Sex of head of household  
female 

(Base = Male) 

Years of Education Women 
has Received 

 

Rural 

(Base=urban) 

Household Size 

 

Religion of the HH (Muslim) 

(Base= Hindu) 

Religion of the HH 
(Christian) 

 

Religion of the HH (Sikh) 

  

  

  

  

  

  

  

  

  

  

0.05* 

0.02 

-0.05 

-0.14*** 

-0.12*** 

-0.02 

-0.00 

0.10*** 

0.14*** 

-0.03 

(0.025) 

(0.015) 

(0.031) 

(0.031) 

(0.025) 

(0.027) 

(0.026) 

(0.027) 

(0.024) 

(0.027) 

0.37*** 

-0.03*** 

-0.12*** 

-0.14*** 

-0.02*** 

0.10*** 

0.21*** 

0.33*** 

0.35*** 

0.23*** 

(0.005) 

(0.003) 

(0.007) 

(0.007) 

(0.005) 

(0.006) 

(0.005) 

(0.006) 

(0.005) 

(0.006) 

-0.00 

0.00*** 

-0.00*** 

-0.00*** 

-0.00 

0.00*** 

0.00*** 

0.00 

0.00*** 

0.00*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

0.01 

-0.00 

-0.00 

0.02** 

-0.01 

0.00 

0.01 

0.01 

0.01* 

-0.01 

(0.006) 

(0.004) 

(0.007) 

(0.007) 

(0.006) 

(0.007) 

(0.006) 

(0.007) 

(0.006) 

(0.006) 

0.00*** 

-0.00*** 

-0.01*** 

-0.02*** 

-0.00*** 

0.01*** 

0.01*** 

0.01*** 

0.01*** 

0.01*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

0.09*** 

0.02*** 

0.02*** 

0.04*** 

-0.00 

-0.07*** 

-0.02*** 

-0.00 

-0.01** 

-0.02*** 

(0.005) 

(0.003) 

(0.006) 

(0.006) 

(0.005) 

(0.006) 

(0.005) 

(0.005) 

(0.005) 

(0.005) 

-0.00*** 

-0.01*** 

-0.00 

0.00 

-0.00*** 

-0.01*** 

-0.00*** 

-0.00 

0.00*** 

0.00 

(0.001) 

(0.000) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

-0.04*** 

-0.01** 

-0.01** 

-0.00 

-0.00 

-0.04*** 

-0.04*** 

-0.09*** 

-0.08*** 

-0.07*** 

(0.005) 

(0.003) 

(0.006) 

(0.006) 

(0.005) 

(0.006) 

(0.005) 

(0.006) 

(0.005) 

(0.005) 

0.05 

-0.04 

-0.24** 

-0.23** 

-0.01 

0.01 

0.16* 

0.00 

0.18** 

0.26*** 

(0.084) 

(0.052) 

(0.101) 

(0.101) 

(0.082) 

(0.091) 

(0.085) 

(0.089) 

(0.079) 

(0.089) 

-0.01 

-0.02 

-0.18*** 

-0.24*** 

-0.05 

0.15*** 

0.15*** 

0.07* 

0.08** 

0.13*** 

 

(0.040) 

(0.024) 

(0.047) 

(0.047) 

(0.038) 

(0.044) 

(0.041) 

(0.043) 

(0.038) 

(0.042) 

 

 
 

196 

 

 
Table 4. 35 (cont’d) 
 
 
Religion of the HH (Others) 

 

Caste of the 
household(Scheduled Caste) 

(Base= General Caste) 

Caste of the 
household(Scheduled Tribe) 

 

 

Constant 

Observations 

R-squared 

 
 
 
 

 
 
 
 

 
 
 
 

 
 
 
 

 
 
 
 

 
 
 
 

 
 
 
 

 
 
 
 

 
 
 
 

 
 
 
 

0.08 

-0.05* 

-0.02 

-0.05 

-0.10* 

0.12** 

0.04 

0.07 

0.10** 

0.20*** 

(0.053) 

(0.031) 

(0.062) 

(0.062) 

(0.051) 

(0.058) 

(0.054) 

(0.056) 

(0.050) 

(0.056) 

0.02*** 

-0.00 

0.04*** 

0.05*** 

0.00 

-0.01*** 

-0.01*** 

-0.01* 

-0.00 

0.02*** 

(0.005) 

(0.003) 

(0.006) 

(0.006) 

(0.005) 

(0.005) 

(0.005) 

(0.005) 

(0.005) 

(0.005) 

-0.05*** 

0.01* 

0.01 

0.01 

0.01 

-0.02** 

-0.05*** 

-0.04*** 

-0.04*** 

-0.03*** 

(0.008) 

(0.005) 

(0.010) 

(0.010) 

(0.008) 

(0.009) 

(0.008) 

(0.008) 

(0.008) 

(0.008) 

0.03*** 

0.12*** 

0.61*** 

0.70*** 

0.24*** 

0.49*** 

0.12*** 

0.20*** 

0.38*** 

0.34*** 

(0.009) 

(0.005) 

57,391 

56,114 

0.161 

0.015 

(0.011) 

50,431 

0.056 

(0.011) 

(0.009) 

(0.010) 

(0.009) 

(0.009) 

(0.008) 

(0.009) 

50,394 

50,120 

57,391 

57,391 

57,391 

57,391 

57,391 

0.073 

0.019 

0.064 

0.093 

0.136 

0.182 

0.119 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

197 

 

Table 4. 36: SELECTED HEALTH INDICATORS- BORDERING DISTRICTS IN BIHAR AS COMPARISON AND BORDERING 
DISTRICTS IN JHARKHAND AS TREATMENT  

  

VARIABLES 

(1) 
Govt 

Institutional 

Delivery 

(2) 

Infant 
Death 

(3) 

(4) 

(5) 

(6) 

(7) 

(8) 

(9) 

(10) 

Underweight 

Stunted 

Wasted 

Tetanus 

Shot 

Polio 

Vaccine 

Measles 
Vaccine 

BCG 

Vaccine 

Health 
Card 

  
Jharkhand Border Dist. # 
Post 

Post Split 

 

 

Age of head of household 

 

Sex of head of household  
female 

(Base = Male) 

Education of the Household 
head 

 

Household Size 

 

Rural 

(Base=urban) 

Religion of the HH 
(Muslim) 

(Base= Hindu) 

Religion of the HH 
(Christian) 

 

Religion of the HH (Sikh) 

  

  

  

  

  

  

  

  

  

  

-0.01 

-0.02* 

-0.16*** 

-0.12*** 

-0.14*** 

0.06*** 

0.01 

0.01 

-0.01 

0.01 

(0.019) 

(0.010) 

(0.023) 

(0.022) 

(0.020) 

(0.020) 

(0.020) 

(0.019) 

(0.015) 

(0.020) 

0.40*** 

-0.03*** 

-0.16*** 

-0.21*** 

-0.08*** 

0.13*** 

0.33*** 

0.54*** 

0.58*** 

0.25*** 

(0.011) 

(0.006) 

(0.013) 

(0.013) 

(0.012) 

(0.012) 

(0.012) 

(0.011) 

(0.009) 

(0.011) 

-0.00 

0.00** 

-0.00*** 

-0.00*** 

-0.00 

0.00*** 

-0.00 

-0.00 

0.00** 

0.00 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

-0.03*** 

-0.01* 

(0.010) 

(0.005) 

-0.02** 

(0.011) 

-0.02* 

-0.01 

0.01 

-0.02* 

-0.00 

0.01 

-0.02*** 

(0.011) 

(0.010) 

(0.010) 

(0.010) 

(0.009) 

(0.007) 

(0.010) 

0.00*** 

-0.00*** 

-0.02*** 

-0.02*** 

-0.00*** 

0.01*** 

0.01*** 

0.01*** 

0.01*** 

0.01*** 

(0.001) 

(0.000) 

-0.01 

0.01** 

(0.011) 

(0.005) 

(0.001) 

0.03** 

(0.012) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

0.03** 

0.03** 

-0.06*** 

-0.03** 

-0.03** 

-0.05*** 

-0.02 

(0.012) 

(0.011) 

(0.011) 

(0.011) 

(0.010) 

(0.008) 

(0.011) 

-0.00** 

-0.00*** 

-0.00 

-0.00 

-0.00 

-0.01*** 

0.00 

0.00 

0.00 

0.00 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

-0.10*** 

-0.01** 

(0.010) 

(0.005) 

-0.02* 

(0.011) 

-0.00 

0.00 

-0.00 

-0.07*** 

-0.06*** 

-0.04*** 

-0.05*** 

(0.011) 

(0.010) 

(0.011) 

(0.010) 

(0.010) 

(0.008) 

(0.010) 

-0.05 

0.03 

0.02 

0.03 

-0.01 

0.04 

-0.03 

-0.04 

0.03 

-0.08** 

(0.040) 

(0.020) 

(0.046) 

(0.046) 

(0.042) 

(0.042) 

(0.041) 

(0.039) 

(0.031) 

(0.040) 

0.14 

-0.05 

0.43 

0.23 

0.23 

-0.22 

0.28 

0.33 

0.14 

0.29 

 

 

(0.312) 

(0.155) 

(0.334) 

(0.330) 
 
 

198 

(0.299) 

(0.326) 

(0.319) 

(0.304) 

(0.242) 

(0.312) 

 

Table 4. 36 (cont’d) 
 
 
Religion of the HH 
(Others) 

 

Caste of the 
household(Scheduled 
Caste) 

(Base= General Caste) 

Caste of the 
household(Scheduled 
Tribe) 

 

 

Constant 

Observations 

R-squared 

-0.12*** 

0.04** 

0.05 

-0.04 

0.08** 

0.05 

0.02 

0.04 

0.06** 

0.05 

(0.036) 

(0.018) 

(0.043) 

(0.043) 

(0.039) 

(0.038) 

(0.037) 

(0.035) 

(0.028) 

(0.036) 

-0.03*** 

0.00 

0.08*** 

0.07*** 

0.01 

-0.04*** 

-0.03*** 

-0.04*** 

-0.02*** 

-0.04*** 

(0.009) 

(0.004) 

(0.010) 

(0.010) 

(0.009) 

(0.009) 

(0.009) 

(0.009) 

(0.007) 

(0.009) 

-0.06*** 

-0.00 

0.00 

-0.03* 

0.04*** 

0.00 

-0.06*** 

-0.00 

-0.01 

-0.01 

(0.012) 

(0.006) 

(0.014) 

(0.014) 

(0.013) 

(0.013) 

(0.013) 

(0.012) 

(0.010) 

(0.012) 

0.12*** 

0.09*** 

0.67*** 

0.74*** 

0.30*** 

0.51*** 

0.14*** 

0.14*** 

0.28*** 

0.32*** 

(0.017) 

(0.008) 

18,512 

18,246 

0.122 

0.010 

(0.019) 

16,820 

0.047 

(0.019) 

(0.017) 

(0.017) 

(0.017) 

(0.016) 

(0.013) 

(0.017) 

16,807 

16,703 

18,512 

18,512 

18,512 

18,512 

18,512 

0.064 

0.015 

0.051 

0.101 

0.202 

0.321 

0.132 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village 

 

 

 

 

 

 

 

 

 

 
 

199 

 

Table 4. 37: SELECTED HEALTH INDICATORS- BORDERING DISTRICTS IN MADHYA PRADESH AS COMPARISON AND 
BORDERING DISTRICTS IN CHHATTISGARH AS TREATMENT  

  

VARIABLES 

(1) 

Govt 

Institutional 

Delivery 

(2) 

Infant 
Death 

(3) 

(4) 

(5) 

(6) 

(7) 

(8) 

(9) 

(10) 

Underweight 

Stunted 

Wasted 

Tetanus 

Shot 

Polio 

Vaccine 

Measles 
Vaccine 

BCG 

Vaccine 

Health 
Card 

  
Chhattisgarh Border Dist. # 
Post 

Post Split 

 

 

Age of head of household 

 

Sex of head of household  
female 
(Base = Male) 
Education of the Household 
head 

 
Household Size 

 

Rural 

(Base=urban) 

Religion of the HH (Muslim) 

(Base= Hindu) 

Religion of the HH (Christian) 

 

Religion of the HH (Sikh) 

 

Religion of the HH (Others) 

 

Caste of the 
household(Scheduled Caste) 

  

  

  

  

  

  

  

  

  

  

-0.02 

-0.02* 

(0.024) 

(0.013) 

-0.06** 

(0.029) 

-0.03 

-0.12*** 

0.06** 

0.12*** 

0.06*** 

0.06*** 

0.12*** 

(0.041) 

(0.039) 

(0.025) 

(0.024) 

(0.025) 

(0.019) 

(0.022) 

0.53*** 

-0.04*** 

-0.17*** 

-0.26*** 

0.03 

0.21*** 

0.29*** 

0.49*** 

0.44*** 

0.47*** 

(0.020) 

(0.011) 

-0.00 

0.00*** 

(0.000) 

(0.000) 

0.01 

-0.01 

(0.019) 

(0.011) 

(0.024) 

0.00 

(0.000) 

-0.00 

(0.022) 

(0.032) 

(0.030) 

(0.021) 

(0.021) 

(0.021) 

0.00 

-0.00 

0.00*** 

0.00 

-0.00 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.016) 

-0.00** 

(0.000) 

(0.019) 

-0.00 

(0.000) 

0.02 

0.01 

(0.022) 

(0.020) 

-0.04* 

(0.021) 

-0.03 

0.01 

0.02 

-0.00 

(0.020) 

(0.020) 

(0.016) 

(0.018) 

0.01*** 

-0.00*** 

-0.01*** 

-0.01*** 

-0.00 

0.01*** 

0.01*** 

0.00*** 

0.01*** 

0.01*** 

(0.001) 

(0.001) 

0.06*** 

-0.01 

(0.015) 

(0.008) 

-

0.01*** 

-0.01*** 

(0.002) 

(0.001) 

0.07* 

-0.03 

(0.035) 

(0.020) 

-

0.29*** 

-0.00 

(0.075) 

(0.043) 

-0.66** 

-0.05 

(0.277) 

(0.153) 

0.06 

0.11** 

(0.081) 

(0.045) 

(0.001) 

0.04** 

(0.017) 

-0.00 

(0.002) 

-0.08* 

(0.042) 

-0.14 

(0.093) 

-0.24 

(0.306) 

0.17 

(0.105) 

(0.001) 

0.04** 

(0.017) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

-0.01 

-0.09*** 

-0.03* 

-0.02 

0.00 

(0.016) 

(0.016) 

(0.015) 

(0.015) 

(0.012) 

(0.001) 

-0.03** 

(0.014) 

-0.00 

-0.00* 

-0.01*** 

-0.00 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

-0.06 

-0.02 

0.03 

0.01 

0.00** 

(0.002) 

0.01 

0.01*** 

0.00 

(0.002) 

(0.002) 

0.02 

0.02 

(0.041) 

(0.039) 

(0.038) 

(0.037) 

(0.037) 

(0.029) 

(0.033) 

-0.12 

-0.12 

-0.00 

-0.04 

0.08 

0.02 

-0.11 

(0.089) 

(0.083) 

(0.081) 

(0.078) 

(0.078) 

(0.062) 

(0.071) 

-0.06 

-0.29 

(0.291) 

(0.271) 

-0.10 

0.14 

(0.140) 

(0.131) 

0.17 

(0.299) 

0.19** 

(0.088) 

0.36 

0.20 

0.07 

0.10 

(0.288) 

(0.289) 

(0.229) 

(0.260) 

-0.05 

-0.01 

0.06 

0.06 

(0.085) 

(0.085) 

(0.067) 

(0.076) 

0.00 

0.02** 

0.01 

0.02 

-0.00 

-0.05*** 

-0.02 

-0.05*** 

-0.06*** 

-0.08*** 

 

 
 

200 

 

Table 4. 37 (cont’d) 
 
 
(Base= General Caste) 
Caste of the 
household(Scheduled Tribe) 

 

 

Constant 

Observations 

R-squared 

(0.013) 

(0.007) 

(0.016) 

(0.017) 

(0.016) 

(0.014) 

(0.014) 

(0.014) 

(0.011) 

(0.012) 

-

0.07*** 

0.01 

(0.012) 

(0.007) 

0.04 

0.09*** 

(0.024) 

(0.013) 

8,063 

0.251 

7,959 

0.013 

0.06*** 

(0.014) 

0.61*** 

(0.028) 

7,305 

0.063 

0.06*** 

0.01 

-0.01 

-0.05*** 

-0.06*** 

-0.03*** 

-0.08*** 

(0.014) 

(0.013) 

(0.013) 

(0.013) 

(0.013) 

(0.010) 

(0.011) 

0.67*** 

0.37*** 

0.52*** 

0.11*** 

0.25*** 

0.42*** 

0.35*** 

(0.032) 

(0.030) 

(0.026) 

(0.025) 

(0.025) 

(0.020) 

(0.022) 

6,919 

0.056 

6,888 

0.014 

8,063 

0.087 

8,063 

0.169 

8,063 

0.171 

8,063 

0.263 

8,063 

0.260 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

201 

 

Table 4. 38: SELECTED HEALTH INDICATORS- BORDERING DISTRICTS IN UTTAR PRADESH AS COMPARISON AND 
BORDERING DISTRICTS IN UTTARAKHAND AS TREATMENT  

  

VARIABLES 

  
Uttarakhand  Border Dist. # 
Post 

Post Split 

 

 

Age of head of household 

 

Sex of head of household  
female 
(Base = Male) 
Education of the Household 
head 

 
Household Size 

 

Rural 

(Base=urban) 

Religion of the HH (Muslim) 

(Base= Hindu) 

Religion of the HH (Christian) 

 

Religion of the HH (Sikh) 

 

Religion of the HH (Others) 

 

Caste of the 
household(Scheduled Caste) 

(1) 
Govt 

Institutional 

Delivery 

(2) 

Infant 
Death 

(3) 

(4) 

(5) 

(6) 

(7) 

(8) 

(9) 

Underweight 

Stunted 

Wasted 

Tetanus 

Shot 

Polio 

Vaccine 

Measles 
Vaccine 

BCG 

Vaccine 

(10) 

Health 
Card 

  

  

  

  

  

  

  

  

  

  

-0.01 

0.01 

(0.028) 

(0.017) 

-0.07** 

(0.036) 

-0.04 

0.03 

0.20*** 

0.08*** 

0.21*** 

0.28*** 

0.02 

(0.036) 

(0.030) 

(0.032) 

(0.030) 

(0.031) 

(0.027) 

(0.032) 

0.27*** 

-0.03*** 

-0.10*** 

-0.25*** 

0.03** 

0.19*** 

0.28*** 

0.39*** 

0.43*** 

0.15*** 

(0.012) 

-0.00 

(0.000) 

-0.02 

(0.015) 

(0.007) 

0.00** 

(0.000) 

0.02* 

(0.009) 

(0.017) 

-0.00* 

(0.000) 

-0.00 

(0.017) 

(0.017) 

(0.014) 

(0.014) 

(0.013) 

(0.014) 

(0.012) 

(0.014) 

-0.00 

-0.00 

0.00*** 

0.00** 

0.00 

0.00 

0.00*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

0.02 

-0.00 

-0.02 

0.02 

-0.00 

-0.00 

-0.00 

(0.017) 

(0.014) 

(0.016) 

(0.016) 

(0.016) 

(0.014) 

(0.016) 

-0.00 

-0.00*** 

-0.02*** 

-0.02*** 

-0.00*** 

0.01*** 

0.01*** 

0.01*** 

0.01*** 

0.01*** 

(0.001) 

0.07*** 

(0.010) 

-0.00* 

(0.001) 

(0.001) 

0.01 

(0.006) 

-0.01*** 

(0.001) 

(0.001) 

0.03*** 

(0.012) 

-0.00 

(0.002) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

0.02 

0.01 

-0.05*** 

-0.02* 

-0.01 

-0.03*** 

-0.01 

(0.012) 

(0.010) 

(0.011) 

(0.010) 

(0.011) 

(0.009) 

(0.011) 

-0.00 

0.00 

-0.01*** 

-0.00 

-0.00 

0.00*** 

-0.00 

(0.002) 

(0.001) 

(0.002) 

(0.002) 

(0.002) 

(0.001) 

(0.002) 

-0.05*** 

-0.03*** 

-0.03*** 

-0.04*** 

-0.01 

-0.04*** 

-0.05*** 

-0.09*** 

-0.08*** 

-0.08*** 

(0.010) 

(0.006) 

0.08 

-0.06 

(0.089) 

(0.057) 

(0.012) 

-0.26** 

(0.113) 

(0.012) 

(0.010) 

(0.011) 

(0.011) 

(0.011) 

(0.009) 

(0.011) 

-0.22* 

-0.08 

-0.08 

0.15 

0.04 

0.17** 

0.22** 

(0.113) 

(0.093) 

(0.100) 

(0.096) 

(0.098) 

(0.084) 

(0.100) 

0.01 

0.00 

-0.15*** 

-0.25*** 

-0.01 

0.12** 

0.11** 

0.09* 

0.09** 

0.10* 

(0.047) 

(0.028) 

(0.057) 

(0.056) 

(0.047) 

(0.053) 

(0.051) 

(0.052) 

(0.044) 

(0.053) 

0.06 

-0.07 

0.14 

-0.15 

-0.00 

0.18** 

0.17** 

0.23*** 

0.25*** 

0.23** 

(0.081) 

(0.047) 

(0.111) 

(0.105) 

(0.091) 

(0.091) 

(0.087) 

(0.089) 

(0.076) 

(0.091) 

0.06*** 

-0.01 

0.03** 

0.04*** 

0.00 

-0.03** 

-0.01 

-0.00 

0.02* 

0.02* 

 

 
 

202 

 

Table 4. 38 (cont’d) 
 
 
(Base= General Caste) 
Caste of the 
household(Scheduled Tribe) 

 

 

Constant 

Observations 

R-squared 

(0.012) 

(0.007) 

(0.015) 

(0.015) 

(0.012) 

(0.014) 

(0.013) 

(0.014) 

(0.012) 

(0.014) 

-0.03* 

(0.018) 

0.04** 

(0.019) 

10,867 

0.093 

0.02* 

(0.011) 

0.12*** 

(0.011) 

10,652 

0.017 

-0.02 

(0.024) 

0.61*** 

(0.024) 

9,558 

0.056 

-0.02 

-0.01 

0.01 

-0.05** 

-0.00 

-0.01 

-0.00 

(0.024) 

(0.020) 

(0.020) 

(0.019) 

(0.020) 

(0.017) 

(0.020) 

0.80*** 

0.17*** 

0.41*** 

0.11*** 

0.24*** 

0.38*** 

0.36*** 

(0.023) 

(0.020) 

(0.021) 

(0.020) 

(0.021) 

(0.018) 

(0.021) 

9,563 

0.090 

9,515 

0.011 

10,867 

10,867 

10,867 

10,867 

10,867 

0.081 

0.113 

0.140 

0.204 

0.102 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

203 

 

Table 4. 39: SELECTED HEALTH INDICATORS- NEARBY STATES AS COMPARISON AND BOTH BIHAR AND 
JHARKHAND AS TREATMENTS 

  

VARIABLES 

  

Jharkhand X Post 

Bihar X Post 

Post 

 

 

 

Age of head of household 

 

Sex of head of household  
female 

(Base = Male) 

Education of the 
Household head 

 

Household Size 

 

Rural 

(Base=urban) 

Religion of the HH 
(Muslim) 

(Base= Hindu) 

Religion of the HH 
(Christian) 

 

 

(1) 

Govt 

Institutional 

Delivery 

(2) 

Infant 
Death 

(3) 

(4) 

(5) 

(6) 

(7) 

(8) 

(9) 

(10) 

Underweight 

Stunted  Wasted 

Tetanus 

Shot 

Polio 

Vaccine 

Measles 
Vaccine 

BCG 

Vaccine 

Health 
Card 

  

  

  

  

  

  

  

  

  

  

0.05*** 

-0.03*** 

-0.14*** 

-0.05** 

-0.09*** 

0.19*** 

0.10*** 

0.12*** 

0.22*** 

0.10*** 

(0.017) 

(0.008) 

(0.020) 

(0.021) 

(0.019) 

(0.017) 

(0.017) 

(0.017) 

(0.014) 

(0.016) 

0.01 

0.01* 

0.01 

-0.00 

-0.03*** 

0.14*** 

-0.12*** 

0.11*** 

0.25*** 

0.02** 

(0.010) 

(0.005) 

(0.011) 

(0.014) 

(0.013) 

(0.010) 

(0.010) 

(0.010) 

(0.008) 

(0.009) 

0.42*** 

-0.04*** 

-0.15*** 

-0.14*** 

-0.04*** 

-0.02** 

0.45*** 

0.42*** 

0.32*** 

0.33*** 

(0.008) 

(0.004) 

(0.009) 

(0.012) 

(0.011) 

(0.008) 

(0.008) 

(0.007) 

(0.006) 

(0.007) 

-0.00 

0.00*** 

-0.00*** 

-0.00*** 

-0.00 

0.00*** 

-0.00 

-0.00** 

-0.00 

0.00 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

-0.00 

-0.01*** 

-0.00 

-0.00 

-0.01* 

0.00 

-0.01* 

0.00 

0.01*** 

0.00 

(0.005) 

(0.003) 

(0.006) 

(0.006) 

(0.005) 

(0.005) 

(0.005) 

(0.005) 

(0.004) 

(0.005) 

0.00*** 

-0.00*** 

-0.02*** 

-0.02*** 

-0.00*** 

0.01*** 

0.01*** 

0.01*** 

0.01*** 

0.01*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

0.01* 

0.01*** 

0.03*** 

0.03*** 

0.02*** 

-0.03*** 

0.00 

-0.01* 

-0.02*** 

-0.00 

(0.005) 

(0.003) 

(0.006) 

(0.006) 

(0.005) 

(0.005) 

(0.005) 

(0.005) 

(0.004) 

(0.005) 

-0.00*** 

-0.01*** 

-0.00*** 

-0.00* 

-0.00** 

-0.01*** 

-0.00 

-0.00 

0.00* 

0.00** 

(0.001) 

(0.000) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

-0.11*** 

-0.00 

0.01 

0.02*** 

0.01 

-0.02*** 

-0.05*** 

-0.07*** 

-0.06*** 

-0.05*** 

(0.005) 

(0.003) 

(0.006) 

(0.006) 

(0.005) 

(0.005) 

(0.005) 

(0.005) 

(0.004) 

(0.005) 

-0.05** 

0.01 

-0.01 

-0.03 

-0.02 

0.00 

-0.03 

-0.00 

0.00 

-0.02 

(0.021) 

(0.010) 

(0.024) 

(0.024) 

(0.021) 

(0.021) 

(0.021) 

(0.021) 

(0.017) 

(0.019) 

 
 

204 

 

Table 4. 39 (cont’d) 
 
 
Religion of the HH (Sikh) 

 

Religion of the HH 
(Others) 

 

Caste of the 
household(Scheduled 
Caste) 

(Base= General Caste) 

Caste of the 
household(Scheduled 
Tribe) 

 

 

Constant 

Observations 

R-squared 

-0.11 

0.03 

-0.00 

-0.02 

-0.07 

-0.01 

-0.14* 

-0.10 

0.04 

0.11 

(0.081) 

(0.041) 

(0.094) 

(0.093) 

(0.082) 

(0.081) 

(0.082) 

(0.080) 

(0.064) 

(0.074) 

-0.10*** 

0.02** 

0.06*** 

0.04** 

0.04*** 

-0.05*** 

-0.02 

-0.07*** 

-0.02 

-0.01 

(0.014) 

(0.007) 

(0.016) 

(0.016) 

(0.014) 

(0.014) 

(0.014) 

(0.014) 

(0.011) 

(0.013) 

-0.00 

0.01** 

0.04*** 

0.04*** 

0.02*** 

-0.02*** 

-0.01 

-0.02*** 

-0.00 

-0.01*** 

(0.005) 

(0.002) 

(0.005) 

(0.005) 

(0.005) 

(0.005) 

(0.005) 

(0.005) 

(0.004) 

(0.004) 

-0.00 

0.00 

0.03*** 

0.03*** 

0.04*** 

0.02** 

-0.05*** 

-0.02*** 

-0.01** 

-0.00 

(0.006) 

(0.003) 

(0.007) 

(0.007) 

(0.006) 

(0.006) 

(0.006) 

(0.006) 

(0.005) 

(0.006) 

0.14*** 

0.09*** 

0.62*** 

0.65*** 

0.28*** 

0.58*** 

0.13*** 

0.21*** 

0.38*** 

0.37*** 

(0.009) 

(0.004) 

(0.010) 

(0.011) 

(0.009) 

(0.009) 

(0.009) 

(0.009) 

(0.007) 

(0.008) 

64,789 

63,847 

58,155 

56,301 

55,913 

64,789 

64,789 

64,789 

64,789 

64,789 

0.180 

0.014 

0.070 

0.078 

0.023 

0.088 

0.163 

0.195 

0.275 

0.202 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village 

 

 

 

 

 

 

 

 
 

205 

 

Table 4. 40: SELECTED HEALTH INDICATORS- NEARBY STATES AS COMPARISON AND BOTH MADHYA PRADESH 
AND CHHATTISGARH AS TREATMENTS 

  

VARIABLES 

  
Chhattisgarh X Post 

Madhya Pradesh X Post 

 

Post 

 

 

Age of head of household 

 

Sex of head of household  
female 
(Base = Male) 
Education of the Household 
head 

 
Household Size 

 

Rural 
(Base=urban) 
Religion of the HH (Muslim) 
(Base= Hindu) 
Religion of the HH (Christian) 

Religion of the HH (Sikh) 

 

 

Religion of the HH (Others) 

 

Caste of the 
household(Scheduled Caste) 
(Base= General Caste) 
Caste of the 
household(Scheduled Tribe) 

(1) 
Govt 

Institutional 

Delivery 

  

0.15*** 
(0.013) 
0.20*** 
(0.008) 
0.37*** 
(0.005) 
-0.00*** 
(0.000) 

0.04*** 
(0.005) 

-0.00*** 
(0.000) 
0.01*** 
(0.004) 

0.00 

(0.001) 
0.01* 
(0.006) 
0.05*** 
(0.012) 

0.00 

(0.028) 
-0.05** 
(0.026) 

0.04*** 
(0.004) 

(2) 

Infant 
Death 

  

0.02*** 
(0.007) 
-0.00 
(0.004) 
-0.04*** 
(0.002) 
0.00*** 
(0.000) 

0.00 

(0.003) 

-0.00*** 
(0.000) 
0.01*** 
(0.002) 
-0.01*** 
(0.000) 
-0.00 
(0.003) 
-0.02*** 
(0.006) 
-0.02* 
(0.014) 

0.01 

(0.013) 

(3) 

(4) 

(5) 

(6) 

(7) 

(8) 

(9) 

(10) 

Underweight 

Stunted  Wasted 

Tetanus 

Shot 

Polio 

Vaccine 

Measles 
Vaccine 

BCG 

Vaccine 

Health 
Card 

  

-0.13*** 
(0.016) 
-0.04*** 
(0.010) 
-0.05*** 
(0.005) 
-0.00*** 
(0.000) 

-0.00 
(0.006) 

-0.01*** 
(0.000) 
0.02*** 
(0.004) 
-0.00 
(0.001) 
-0.00 
(0.007) 

0.01 

(0.013) 
-0.03 
(0.031) 

0.01 

(0.030) 

  

  

  

-0.09*** 
(0.023) 
-0.00 
(0.014) 
-0.12*** 
(0.006) 
-0.00*** 
(0.000) 

-0.05*** 
(0.020) 
-0.03** 
(0.013) 
0.02*** 
(0.005) 
-0.00 
(0.000) 

0.13*** 
(0.014) 
0.06*** 
(0.009) 
-0.04*** 
(0.005) 
0.00*** 
(0.000) 

  

0.13*** 
(0.014) 

0.01 

(0.009) 
0.26*** 
(0.005) 
0.00** 
(0.000) 

  

0.09*** 
(0.014) 
0.09*** 
(0.009) 
0.30*** 
(0.005) 
-0.00 
(0.000) 

  

0.09*** 
(0.011) 
0.11*** 
(0.007) 
0.24*** 
(0.004) 

0.00 

(0.000) 

  

0.15*** 
(0.013) 
0.19*** 
(0.008) 
0.22*** 
(0.005) 
0.00*** 
(0.000) 

-0.01 
(0.006) 

0.01 

(0.006) 

-0.01 
(0.006) 

-0.01 
(0.006) 

-0.01** 
(0.006) 

0.00 

0.01 

(0.005) 

(0.005) 

-0.01*** 
(0.000) 
0.02*** 
(0.005) 
-0.00 
(0.001) 

0.01 

(0.007) 
-0.01 
(0.014) 
-0.04 
(0.032) 
-0.05 
(0.033) 

-0.00*** 
(0.000) 
0.01*** 
(0.004) 
-0.00*** 
(0.001) 
-0.01 
(0.006) 
0.04*** 
(0.012) 
-0.02 
(0.028) 
-0.02 
(0.029) 

0.01*** 
(0.000) 
-0.04*** 
(0.004) 
-0.01*** 
(0.001) 
-0.02*** 
(0.006) 

0.02 

(0.012) 
0.07** 
(0.029) 
0.07*** 
(0.027) 

0.01*** 
(0.000) 
-0.02*** 
(0.004) 
-0.00*** 
(0.001) 
-0.02*** 
(0.006) 
-0.04*** 
(0.012) 
0.07** 
(0.029) 
0.06** 
(0.027) 

0.01*** 
(0.000) 
-0.03*** 
(0.004) 
-0.00 
(0.001) 
-0.04*** 
(0.006) 
-0.01 
(0.012) 
0.09*** 
(0.029) 

0.04 

(0.027) 

0.01*** 
(0.000) 
-0.02*** 
(0.003) 
0.00*** 
(0.000) 
-0.03*** 
(0.005) 
-0.00 
(0.010) 
0.08*** 
(0.024) 
0.05** 
(0.022) 

0.01*** 
(0.000) 
-0.02*** 
(0.004) 
-0.00 
(0.001) 
-0.00 
(0.006) 
-0.01 
(0.012) 
0.08*** 
(0.027) 

0.02 

(0.025) 

0.00 

(0.002) 

0.03*** 
(0.005) 

0.04*** 
(0.005) 

0.01*** 
(0.004) 

-0.04*** 
(0.004) 

-0.03*** 
(0.004) 

-0.02*** 
(0.004) 

-0.02*** 
(0.003) 

-0.01*** 
(0.004) 

-0.02*** 

0.01*** 

0.06*** 

0.05*** 

0.03*** 

-0.02*** 

-0.07*** 

-0.05*** 

-0.03*** 

-0.03*** 

 

 
 

206 

 

Table 4. 40 (cont’d) 
 
 
Constant 

 

Observations 
R-squared 

(0.004) 
0.14*** 
(0.007) 
93,951 
0.224 

(0.002) 
0.10*** 
(0.003) 
92,678 
0.018 

(0.005) 
0.55*** 
(0.008) 
83,735 
0.064 

(0.005) 
0.60*** 
(0.008) 
78,976 
0.064 

(0.005) 
0.25*** 
(0.008) 
78,491 
0.024 

(0.004) 
0.67*** 
(0.007) 
93,951 
0.061 

(0.005) 
0.20*** 
(0.007) 
93,951 
0.124 

(0.004) 
0.35*** 
(0.007) 
93,951 
0.133 

(0.004) 
0.55*** 
(0.006) 
93,951 
0.162 

(0.004) 
0.40*** 
(0.006) 
93,951 
0.171 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

207 

 

Table 4. 41: SELECTED HEALTH INDICATORS- NEARBY STATES AS COMPARISON AND BOTH UTTAR PRADESH AND 
JHARKHAND AS TREATMENTS 

  

VARIABLES 

  

Uttarakhand X Post 

 

Uttar Pradesh X Post 

Post 

 

 

Age of head of household 

 

Sex of head of household  
female 

(Base = Male) 

Education of the 
Household head 

 

Household Size 

 

Rural 

(Base=urban) 

Religion of the HH 
(Muslim) 

(Base= Hindu) 

Religion of the HH 
(Christian) 

(1) 
Govt 

Institutional 

Delivery 

  

0.18*** 

(2) 

Infant 
Death 

  

0.02 

(3) 

(4) 

(5) 

(6) 

(7) 

(8) 

(9) 

Underweight 

Stunted 

Wasted 

Tetanus 

Shot 

Polio 

Vaccine 

Measles 
Vaccine 

BCG 

Vaccine 

(10) 

Health 
Card 

  

  

  

  

  

  

  

  

-0.13*** 

-0.16*** 

0.04* 

0.00 

0.08*** 

0.08*** 

0.10*** 

0.01 

(0.024) 

(0.014) 

(0.029) 

(0.030) 

(0.025) 

(0.026) 

(0.025) 

(0.026) 

(0.022) 

(0.025) 

-0.14*** 

0.00 

-0.08*** 

-0.02* 

-0.07*** 

0.02** 

-0.08*** 

0.02** 

0.04*** 

0.04*** 

(0.008) 

(0.004) 

(0.009) 

(0.010) 

(0.008) 

(0.008) 

(0.008) 

(0.008) 

(0.007) 

(0.008) 

0.51*** 

-0.04*** 

-0.03*** 

-0.12*** 

0.05*** 

0.08*** 

0.29*** 

0.31*** 

0.32*** 

0.18*** 

(0.007) 

(0.004) 

(0.008) 

(0.008) 

(0.007) 

(0.007) 

(0.007) 

(0.007) 

(0.006) 

(0.007) 

-0.00*** 

0.00*** 

-0.00*** 

-0.00*** 

-0.00 

0.00*** 

0.00*** 

0.00 

0.00*** 

0.00*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

0.01** 

(0.005) 

-0.00 

-0.01 

0.01* 

-0.01** 

0.00 

0.00 

0.00 

0.01 

-0.01 

(0.003) 

(0.006) 

(0.006) 

(0.005) 

(0.005) 

(0.005) 

(0.005) 

(0.005) 

(0.005) 

0.00*** 

-0.00*** 

-0.01*** 

-0.01*** 

-0.00*** 

0.01*** 

0.01*** 

0.01*** 

0.01*** 

0.01*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

0.06*** 

0.01*** 

0.01*** 

0.03*** 

-0.01** 

-0.06*** 

-0.03*** 

-0.02*** 

-0.02*** 

-0.03*** 

(0.004) 

(0.002) 

(0.005) 

(0.005) 

(0.004) 

(0.004) 

(0.004) 

(0.004) 

(0.004) 

(0.004) 

-0.00*** 

-0.01*** 

-0.00 

0.00 

-0.00*** 

-0.01*** 

-0.00*** 

-0.00 

0.00*** 

0.00 

(0.000) 

(0.000) 

(0.001) 

(0.001) 

(0.000) 

(0.001) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

-0.05*** 

-0.00 

-0.01 

0.00 

0.00 

-0.04*** 

-0.04*** 

-0.09*** 

-0.08*** 

-0.07*** 

(0.004) 

(0.002) 

(0.005) 

(0.005) 

(0.004) 

(0.005) 

(0.004) 

(0.005) 

(0.004) 

(0.005) 

0.09 

-0.03 

-0.22*** 

-0.18** 

-0.02 

-0.01 

0.12* 

-0.00 

0.14** 

0.20*** 

 

Religion of the HH (Sikh) 

(0.069) 

(0.040) 

(0.080) 

(0.081) 

(0.068) 

(0.074) 

(0.070) 

(0.072) 

(0.063) 

(0.072) 

-0.01 

-0.01 

-0.06*** 

-0.09*** 

-0.02 

0.10*** 

0.06*** 

0.05** 

0.07*** 

0.08*** 

 

(0.018) 

(0.010) 

(0.020) 

(0.021) 

(0.017) 

(0.019) 

(0.018) 

(0.019) 

(0.016) 

(0.019) 

 

 
 

208 

 

Table 4. 41 (cont’d) 
 
 
Religion of the HH 
(Others) 

 

Caste of the 
household(Scheduled 
Caste) 

(Base= General Caste) 

Caste of the 
household(Scheduled 
Tribe) 

 

Constant 

 
Observations 

R-squared 

-0.04 

-0.02 

-0.03 

-0.05 

-0.06* 

0.10*** 

0.06* 

0.02 

0.08** 

0.10*** 

(0.034) 

(0.019) 

(0.039) 

(0.040) 

(0.034) 

(0.037) 

(0.035) 

(0.036) 

(0.031) 

(0.036) 

0.04*** 

-0.00 

0.04*** 

0.05*** 

0.01*** 

-0.02*** 

-0.01*** 

-0.02*** 

-0.01*** 

0.01*** 

(0.004) 

(0.002) 

(0.004) 

(0.004) 

(0.004) 

(0.004) 

(0.004) 

(0.004) 

(0.004) 

(0.004) 

-0.03*** 

0.01* 

0.04*** 

0.04*** 

0.02*** 

-0.02*** 

-0.05*** 

-0.05*** 

-0.04*** 

-0.03*** 

(0.006) 

(0.003) 

(0.007) 

(0.007) 

(0.006) 

(0.006) 

(0.006) 

(0.006) 

(0.005) 

(0.006) 

0.06*** 

0.11*** 

0.57*** 

0.66*** 

0.24*** 

0.51*** 

0.14*** 

0.25*** 

0.42*** 

0.37*** 

(0.007) 

93,808 

0.193 

(0.004) 

92,018 

0.017 

(0.008) 

83,198 

0.058 

(0.008) 

(0.007) 

(0.007) 

(0.007) 

(0.007) 

(0.006) 

(0.007) 

81,786 

81,305 

93,808 

93,808 

93,808 

93,808 

93,808 

0.075 

0.023 

0.071 

0.123 

0.142 

0.189 

0.129 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

209 

 

APPENDIX E: TABLES- RESULTS OF PARALLEL TREND TESTS 
 

Table 4. 42: TESTING PARALLEL TREND - ADJACENT STATES AS COMPARISON AND BOTH PARENT AND NEW 
STATES AS TWO TREATMENTS 

  

(1) 

(2) 

(3) 

Piped 

HHH Education  Electricity 

Drinking 

VARIABLES 

  

  

  

New States X Post 

-0.188 

-0.084 

 

(0.210) 

(0.081) 

Parent State X Post 

0.294 

-0.051 

Post 

 

 

New States 

 
Parent States 

 

(0.291) 

(0.055) 

-0.138** 

0.105*** 

(0.062) 

(0.003) 

-1.914** 

-0.339*** 

(0.888) 

(0.047) 

(0.667) 

(0.036) 

(4) 

(5) 

Concrete 

house 

  

-0.037 

(0.030) 

-0.029* 

(0.015) 

Toilet 

  

-0.080 

(0.089) 

-0.037 

(0.044) 

(6) 

Clean 

Cooking Fuel 

(7) 

(8) 

(9) 

Television 

Refrigerator 

Motorcycle 

  

-0.139* 

(0.080) 

-0.099 

(0.094) 

  

-0.090 

(0.090) 

-0.076 

(0.075) 

  

-0.044 

(0.088) 

-0.035 

(0.034) 

  

-0.040 

(0.048) 

-0.029 

(0.024) 

0.072*** 

0.073*** 

0.170*** 

0.155*** 

0.046*** 

0.036*** 

(0.003) 

-0.206*** 

(0.044) 

(0.003) 

-0.062 

(0.039) 

(0.002) 

-0.022 

(0.035) 

(0.003) 

(0.002) 

-0.307*** 

-0.068** 

(0.046) 

(0.034) 

(0.002) 

0.039 

(0.035) 

Water 

  

0.015 

(0.012) 

-0.002 

(0.002) 

0.001 

(0.001) 

-0.006 

(0.009) 

0.846 

-0.184*** 

-0.036*** 

0.291*** 

0.186*** 

0.143*** 

0.092*** 

0.161*** 

0.077*** 

Age of HH Head 

-0.050*** 

0.001*** 

 

(0.002) 

(0.000) 

Sex of HH Head=  Female 

-3.096*** 

0.004 

(Base = Male) 

(0.080) 

(0.004) 

Education of the HH Head 

 

 

 

0.008*** 

(0.000) 

Household Size 

0.028*** 

0.010*** 

 

(0.008) 

(0.000) 

(0.007) 

0.000 

(0.000) 

-0.001 

(0.001) 

-0.000 

(0.000) 

0.000 

(0.000) 

(0.033) 

(0.029) 

(0.027) 

(0.034) 

(0.026) 

(0.026) 

0.001*** 

0.002*** 

0.001*** 

0.002*** 

0.001*** 

0.001*** 

(0.000) 

(0.000) 

0.013*** 

0.011*** 

(0.004) 

(0.004) 

(0.000) 

-0.006* 

(0.003) 

(0.000) 

0.001 

(0.004) 

(0.000) 

-0.006* 

(0.003) 

(0.000) 

-0.007** 

(0.003) 

0.008*** 

0.009*** 

0.008*** 

0.010*** 

0.006*** 

0.006*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

0.002*** 

0.002*** 

-0.003*** 

0.014*** 

0.001*** 

0.007*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

Rural 

-3.560*** 

-0.351*** 

0.010*** 

-0.409*** 

-0.545*** 

-0.384*** 

-0.378*** 

-0.165*** 

-0.137*** 

(Base=urban) 
Religion of the HH 
(Muslim) 

(Base= Hindu) 
 

 

(0.059) 

(0.003) 

(0.001) 

(0.003) 

(0.003) 

(0.002) 

(0.003) 

(0.002) 

(0.002) 

-1.991*** 

-0.030*** 

-0.003*** 

-0.053*** 

0.039*** 

-0.054*** 

-0.084*** 

-0.035*** 

-0.043*** 

(0.077) 

(0.004) 

(0.001) 

(0.004) 

(0.003) 

1.140*** 

-0.040*** 

-0.004* 

0.030*** 

0.021** 

(0.003) 

0.011 

(0.004) 

0.003 

(0.003) 

(0.003) 

0.046*** 

0.030*** 

 
 

210 

 

 

 

 

 

 

 

 

 

Table 4. 42 (cont’d) 
 
 
 
Religion of the HH 
(Christian) 

 

(0.217) 

(0.012) 

Religion of the HH (Sikh) 

0.167 

0.090*** 

 

Religion of the HH 
(Others) 

 

Caste of the 
household(SC) 

(Base= General Caste) 
Caste of the 
household(ST) 

 

Constant 

 
Observations 

R-squared 

(0.349) 

(0.019) 

0.594*** 

-0.011 

(0.166) 

(0.009) 

(0.002) 

-0.004 

(0.003) 

-0.001 

(0.002) 

(0.011) 

(0.010) 

(0.009) 

(0.011) 

(0.008) 

(0.009) 

0.088*** 

0.129*** 

0.129*** 

0.114*** 

0.145*** 

0.113*** 

(0.017) 

(0.015) 

(0.014) 

(0.018) 

(0.014) 

(0.014) 

0.048*** 

0.039*** 

0.072*** 

0.082*** 

0.069*** 

0.052*** 

(0.008) 

(0.007) 

(0.007) 

(0.009) 

(0.006) 

(0.007) 

-2.187*** 

-0.111*** 

-0.002** 

-0.094*** 

-0.092*** 

-0.068*** 

-0.090*** 

-0.049*** 

-0.055*** 

(0.067) 

(0.004) 

(0.001) 

(0.003) 

(0.003) 

(0.003) 

(0.003) 

(0.003) 

(0.003) 

-2.160*** 

-0.123*** 

0.002*** 

-0.096*** 

-0.098*** 

-0.077*** 

-0.105*** 

-0.045*** 

-0.051*** 

(0.074) 

(0.004) 

(0.001) 

(0.004) 

(0.003) 

(0.003) 

(0.004) 

10.963*** 

0.774*** 

0.023*** 

0.316*** 

0.549*** 

0.247*** 

0.241*** 

(0.607) 

(0.032) 

107,841 

107,841 

0.111 

0.412 

(0.006) 

107,841 

0.104 

(0.030) 

107,841 

0.350 

(0.027) 

107,841 

0.501 

(0.024) 

106,223 

0.430 

(0.031) 

107,841 

0.326 

(0.003) 

0.019 

(0.024) 

107,838 

0.168 

(0.003) 

0.012 

(0.024) 

107,830 

0.137 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village level 
 

 
 

211 

 

Table 4. 43: TESTING PARALLEL TREND (HEALTH INDICATORS) - ADJACENT STATES AS COMPARISON AND BOTH   
PARENT AND NEW STATES AS TWO TREATMENTS 

(3) 

(4) 

(5) 

(6) 

(7) 

Underweight 

Stunted 

Wasted 

Tetanus Shot  Polio Vaccine 

  

0.000 
(0.028) 
-0.022* 
(0.013) 
-0.013 
(0.009) 
0.199* 
(0.112) 
-0.152** 
(0.064) 

  

-0.026 
(0.048) 
-0.036 
(0.025) 
0.016 
(0.011) 
0.157 
(0.341) 
-0.094 
(0.076) 

-0.001*** 

-0.001*** 

(0.000) 
0.014 
(0.012) 

(0.000) 
-0.004 
(0.013) 

  

0.092 
(0.073) 
-0.032 
(0.024) 
0.002 
(0.010) 
-0.005 
(0.301) 
-0.066 
(0.067) 
-0.000 
(0.000) 
-0.008 
(0.012) 

  

0.069 
(0.072) 
0.060 
(0.050) 
0.116*** 
(0.007) 

-0.335*** 

(0.094) 
-0.006 
(0.053) 
0.001*** 
(0.000) 
0.009 
(0.009) 
0.030*** 
(0.001) 

  

0.050 
(0.048) 
0.089 
(0.088) 

-0.233*** 

(0.006) 
-0.172** 
(0.076) 
0.046 
(0.043) 
0.000* 
(0.000) 
0.027*** 
(0.007) 
0.010*** 
(0.000) 

(8) 

Measles 
Vaccine 

  

-0.094*** 

(0.021) 
-0.026 
(0.020) 
0.030*** 
(0.007) 
-0.204** 
(0.091) 
0.018 
(0.052) 
0.000 
(0.000) 
-0.014 
(0.009) 
0.022*** 
(0.001) 

(9) 

BCG Vaccine 

  

-0.017 
(0.023) 
-0.077 
(0.061) 
0.119*** 
(0.008) 
-0.201** 
(0.098) 
-0.018 
(0.056) 
0.001*** 
(0.000) 
-0.004 
(0.010) 
0.026*** 
(0.001) 

(2) 

Infant 
Death 

  

0.014 
(0.014) 
0.014 
(0.017) 
-0.011** 
(0.005) 
0.011 
(0.059) 
-0.010 
(0.033) 
0.000*** 
(0.000) 
0.013** 
(0.006) 

(1) 
Govt 

Institutional 

Delivery 

  

-0.035 
(0.026) 
-0.034 
(0.027) 
0.025*** 
(0.005) 
-0.048 
(0.067) 
0.167*** 
(0.038) 
0.000*** 
(0.000) 
0.027*** 
(0.007) 
0.012*** 
(0.000) 

-0.133*** 

(0.004) 
0.000 
(0.000) 

-0.046*** 

(0.005) 

-0.035** 
(0.017) 
0.020 
(0.025) 
0.059*** 
(0.012) 

  

VARIABLES 

  
New State * Dummy Post 

 

Old State * Dummy Post 

Dummy Post 

New State 

Old State 

 

 

 

 

Age of HH Head 

 

Sex of Hh head= Female 
(Base = Male) 
Women Education 

 

Rural 
(Base=urban) 
Household Size 

 

Religion of the HH 
(Muslim) 
(Base= Hindu) 
Religion of the HH 
(Christian) 

 

Religion of the HH (Sikh) 

 

Religion of the HH (Others) 

 

 

-0.003*** 

-0.019*** 

-0.016*** 

-0.006*** 

(0.000) 
0.017*** 
(0.004) 

-0.006*** 

(0.000) 

0.000 
(0.004) 

-0.002 
(0.015) 
-0.015 
(0.022) 
-0.001 
(0.011) 

(0.001) 
0.041*** 
(0.008) 
-0.002** 
(0.001) 

0.005 
(0.009) 

-0.052* 
(0.029) 
-0.060 
(0.043) 
-0.009 
(0.022) 

(0.001) 
0.003 
(0.008) 
-0.002** 
(0.001) 

0.004 
(0.009) 

-0.012 
(0.031) 
-0.034 
(0.040) 
0.005 
(0.020) 

(0.001) 
0.032*** 
(0.009) 
-0.000 
(0.001) 

0.003 
(0.010) 

-0.070** 
(0.035) 

-0.129*** 

(0.045) 
-0.008 
(0.023) 

 
 

212 

-0.089*** 

-0.052*** 

-0.031*** 

-0.059*** 

(0.006) 
-0.000 
(0.001) 

-0.014** 
(0.007) 

0.015 
(0.024) 
0.075** 
(0.035) 
0.045*** 
(0.017) 

(0.005) 
0.002*** 
(0.000) 

-0.014** 
(0.006) 

-0.021 
(0.019) 
0.031 
(0.028) 
0.034** 
(0.014) 

(0.006) 
0.001 
(0.001) 

(0.006) 
0.001 
(0.001) 

-0.056*** 

-0.094*** 

(0.007) 

(0.007) 

0.005 
(0.023) 
0.082** 
(0.034) 
0.048*** 
(0.017) 

0.008 
(0.025) 
0.066* 
(0.036) 
0.050*** 
(0.018) 

 

Table 4. 43 (cont’d) 
 
 
Caste of the household(SC) 
(Base= General Caste) 
Caste of the household(ST) 

Constant 

 

 

-0.042*** 

(0.005) 
-0.009** 
(0.004) 
0.123*** 
(0.031) 
45,285 
0.145 

0.004 
(0.005) 
0.013*** 
(0.004) 
0.079*** 
(0.027) 
44,051 
0.024 

0.043*** 
(0.009) 
0.017** 
(0.007) 
0.554*** 
(0.052) 
35,896 
0.071 

Observations 
R-squared 
    Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village level 

0.035*** 
(0.011) 
0.028*** 
(0.008) 
0.723*** 
(0.055) 
28,145 
0.070 

0.051*** 
(0.010) 
0.013* 
(0.007) 
0.199*** 
(0.049) 
27,695 
0.046 

-0.121*** 

-0.036*** 

-0.076*** 

-0.100*** 

(0.008) 

-0.020*** 

(0.006) 
0.704*** 
(0.043) 
45,285 
0.247 

(0.006) 
-0.012** 
(0.005) 
0.147*** 
(0.035) 
45,285 
0.130 

(0.007) 

(0.008) 

-0.029*** 

-0.041*** 

(0.006) 
0.359*** 
(0.042) 
45,285 
0.158 

(0.006) 
0.598*** 
(0.045) 
45,285 
0.213 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

213 

 

Table 4. 44: TESTING PARALLEL TREND - PARENT STATE AS COMPARISON AND NEWLY FORMED STATES AS 
TREATMENT 

  

(1) 

(2) 

(4) 

(5) 

Electricity 

Drinking 

Toilet 

Cooking 

Television 

Refrigerator  Motorcycle 

Concrete 

house 

  

  

(6) 

Clean 

(7) 

(8) 

(9) 

Fuel 

  

  

  

  

HHH 

Education 

  

  

-0.407 

(0.322) 

0.147* 

(0.082) 

-0.037 

(0.776) 

-0.028 

(0.112) 

0.052*** 

(0.004) 

-0.038 

(0.042) 

(0.007) 

0.008*** 

(0.000) 

0.055*** 

0.009*** 

(0.010) 

(0.001) 

(3) 

Piped 

Water 

  

0.010 

(0.012) 

-0.001* 

(0.001) 

(0.006) 

0.000* 

(0.000) 

-0.001 

(0.001) 

-0.000 

(0.000) 

0.000 

(0.000) 

-0.005 

(0.010) 

-0.045 

(0.039) 

-0.048** 

(0.018) 

-0.021* 

(0.012) 

-0.009 

(0.009) 

0.043*** 

0.036*** 

0.072*** 

0.082*** 

0.010*** 

(0.004) 

(0.003) 

(0.003) 

(0.004) 

(0.004) 

0.054*** 

-0.12*** 

-0.20*** 

-0.061** 

-0.081** 

-0.087*** 

(0.035) 

(0.033) 

(0.027) 

(0.041) 

(0.033) 

-0.011 

(0.010) 

0.006* 

(0.004) 

-0.014 

(0.033) 

0.001*** 

0.001*** 

0.001*** 

0.001*** 

0.001*** 

0.001*** 

(0.000) 

(0.000) 

0.025*** 

0.026*** 

(0.006) 

(0.006) 

(0.000) 

0.010** 

(0.005) 

(0.000) 

0.017** 

(0.007) 

(0.000) 

-0.000 

(0.006) 

(0.000) 

0.009 

(0.006) 

0.008*** 

0.008*** 

0.007*** 

0.009*** 

0.005*** 

0.006*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

0.003*** 

0.003*** 

-0.003*** 

0.012*** 

(0.000) 

(0.000) 

(0.000) 

(0.001) 

(0.000) 

0.000 

(0.000) 

(0.000) 

0.006*** 

(0.000) 

VARIABLES 

  
Districts in New states X 
Post 

Dummy Post 

 

 

Dummy Treatment 

 

Age of head of household 

-0.053*** 

0.001*** 

 

(0.002) 

(0.000) 

Sex of HH head-  Female 

-3.298*** 

0.022*** 

(Base = Male) 

(0.133) 

 

 

Education of the HH head 

 

Household Size 

 

Rural 

(Base=urban) 
Religion of the HH 
(Muslim) 

(Base= Hindu) 
Religion of the HH 
(Christian) 

 

Religion of the HH (Sikh) 

 

Religion of the HH (Others) 

1.898*** 

 

(0.399) 

-3.998*** 

-0.462*** 

0.003*** 

-0.473*** 

-0.569*** 

-0.384*** 

-0.406*** 

-0.180*** 

-0.156*** 

(0.098) 

(0.005) 

(0.001) 

(0.004) 

(0.004) 

(0.003) 

(0.005) 

(0.004) 

(0.004) 

-2.148*** 

(0.111) 

-0.005 

(0.006) 

2.038*** 

(0.461) 

0.495 

(0.579) 

0.003 

(0.025) 

0.197*** 

(0.031) 

0.049** 

(0.021) 

-0.002* 

(0.001) 

-0.004 

(0.003) 

-0.005 

(0.004) 

0.004 

(0.003) 

-0.038*** 

0.096*** 

-0.059*** 

-0.078*** 

-0.024*** 

-0.047*** 

(0.005) 

(0.005) 

(0.004) 

(0.006) 

(0.005) 

(0.005) 

0.066*** 

0.094*** 

0.073*** 

0.064*** 

(0.021) 

(0.019) 

(0.016) 

(0.025) 

0.047** 

(0.020) 

0.087*** 

(0.020) 

0.153*** 

0.092*** 

0.219*** 

0.200*** 

0.208*** 

0.189*** 

(0.026) 

(0.024) 

(0.020) 

(0.031) 

(0.025) 

(0.025) 

0.068*** 

0.077*** 

0.185*** 

0.173*** 

0.169*** 

0.168*** 

(0.018) 

(0.017) 

(0.014) 

(0.021) 

(0.017) 

(0.017) 

Caste of the household(SC) 

-2.023*** 

-0.067*** 

-0.00*** 

-0.07*** 

-0.07*** 

-0.06*** 

-0.086*** 

-0.031*** 

-0.050*** 

(Base= General Caste) 

(0.104) 

(0.006) 

(0.001) 

(0.005) 

(0.004) 

(0.004) 

(0.006) 

(0.004) 

(0.005) 

 

 
 

214 

 

 

 

 

 

 

 

 

 

 

 

 

Table 4. 44 (cont’d) 
 
 
Caste of the household(ST) 

 

 

Constant 

Observations 

-2.131*** 

-0.072*** 

0.003*** 

-0.07*** 

-0.07*** 

-0.04*** 

-0.077*** 

-0.022*** 

-0.040*** 

(0.114) 

(0.006) 

11.983*** 

0.685*** 

(0.467) 

41,435 

(0.025) 

41,435 

(0.001) 

-0.002 

(0.004) 

41,435 

(0.005) 

(0.005) 

(0.004) 

(0.006) 

(0.005) 

(0.005) 

0.672*** 

0.753*** 

0.422*** 

0.434*** 

0.223*** 

0.081*** 

(0.021) 

41,435 

(0.020) 

41,435 

(0.016) 

41,118 

(0.025) 

41,435 

(0.020) 

41,433 

(0.020) 

41,428 

R-squared 
0.133 
Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village level 
 

0.117 

0.413 

0.416 

0.117 

0.105 

0.377 

0.501 

0.299 

 

 
 

215 

 

Table 4. 45: TESTING PARALLEL TREND (HEALTH INDICATORS)- PARENT STATE AS COMPARISON AND NEWLY 
FORMED STATES AS TREATMENT 

(2) 

(3) 

(4) 

(5) 

(6) 

(7) 

(8) 

(9) 

VARIABLES 

Delivery 

Institutional 

Infant Death  Under-weight 

Stunted 

Wasted 

Tetanus Shot 

Polio 

Vaccine 

Measles 
Vaccine 

BCG 

Vaccine 

  

  

  

  

  

  

  

  

  

0.000 

(0.015) 

0.003 

(0.005) 

0.020 

(0.053) 

0.000 

(0.000) 

0.010 

(0.010) 

0.021 

(0.029) 

-0.032*** 

(0.010) 

0.227** 

(0.098) 

0.012 

(0.050) 

-0.018 

(0.011) 

0.183* 

(0.100) 

-0.001*** 

-0.001*** 

(0.000) 

(0.000) 

-0.020 

(0.018) 

0.005 

(0.019) 

(0.011) 

0.066 

(0.092) 

-0.000 

(0.000) 

-0.032* 

(0.017) 

0.120 

(0.146) 

0.006 

(0.024) 

-0.039 

(0.037) 

-0.028*** 

0.181*** 

-0.144*** 

(0.008) 

-0.427*** 

(0.082) 

(0.006) 

-0.056 

(0.057) 

-0.070* 

(0.040) 

0.008 

(0.007) 

0.061 

(0.054) 

0.045*** 

(0.008) 

-0.212*** 

-0.201** 

(0.067) 

(0.082) 

0.001*** 

0.000*** 

0.001*** 

0.001*** 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

-0.006 

(0.015) 

0.018* 

(0.010) 

-0.018 

(0.012) 

0.004 

(0.015) 

Women Education 

0.015*** 

-0.003*** 

-0.016*** 

-0.014*** 

-0.005*** 

0.036*** 

0.012*** 

0.025*** 

0.031*** 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

-0.112*** 

0.030*** 

0.058*** 

0.036*** 

  

(1) 

Govt 

  
Districts in New states X 
Post Dummy 

 
Post Dummy 

Treatment 

 

 

-0.005 

(0.013) 

-0.005 

(0.004) 

-0.145*** 

(0.045) 

Age of head of household 

0.000*** 

 

Sex of head of household  
female 

(Base = Male) 

(0.000) 

0.022*** 

(0.008) 

(0.006) 

0.000 

(0.000) 

(0.007) 

-0.005*** 

(0.001) 

-0.020*** 

(0.005) 

-0.017 

(0.028) 

0.004 

(0.006) 

0.007 

(0.033) 

0.030 

(0.045) 

0.009 

(0.024) 

 

Rural 

(Base=urban) 

Household Size 

 

Religion of the HH 
(Muslim) 

(Base= Hindu) 
Religion of the HH 
(Christian) 

 

Religion of the HH (Sikh) 

0.103*** 

(0.038) 

0.010 

(0.021) 

 

Religion of the HH 
(Others) 

 

 

(0.013) 

-0.001 

(0.001) 

-0.000 

(0.012) 

-0.013 

(0.064) 

-0.144* 

(0.086) 

0.056 

(0.046) 

(0.014) 

-0.001 

(0.001) 

0.000 

(0.012) 

-0.046 

(0.070) 

-0.326*** 

(0.089) 

-0.024 

(0.053) 

 
 

216 

(0.001) 

0.013 

(0.013) 

-0.002** 

(0.001) 

0.017 

(0.011) 

-0.075 

(0.065) 

-0.068 

(0.081) 

0.047 

(0.049) 

(0.001) 

(0.001) 

(0.001) 

(0.001) 

-0.158*** 

-0.056*** 

-0.070*** 

-0.094*** 

(0.010) 

-0.000 

(0.001) 

(0.007) 

0.002*** 

(0.001) 

(0.008) 

0.001* 

(0.001) 

(0.010) 

0.001 

(0.001) 

-0.029*** 

-0.024*** 

-0.049*** 

-0.106*** 

(0.010) 

(0.007) 

(0.008) 

(0.010) 

0.081 

(0.052) 

0.109 

(0.070) 

0.084** 

(0.039) 

0.019 

(0.036) 

0.046 

(0.049) 

0.087** 

0.136*** 

(0.043) 

0.107* 

(0.058) 

(0.052) 

0.076 

(0.070) 

0.121*** 

0.092*** 

0.142*** 

(0.027) 

(0.032) 

(0.038) 

 

 
Table 4. 45 (cont’d) 
 
 
Caste of the 
household(SC) 

(Base= General Caste) 
Caste of the 
household(ST) 

 

Constant 

 
Observations 

 
 
 
 

-0.041*** 

(0.007) 

-0.019*** 

(0.005) 

0.279*** 

(0.025) 

22,166 

 
 
 
 

0.007 

(0.008) 

0.012** 

(0.005) 

0.072** 

(0.029) 

21,261 

 
 
 
 

0.036** 

(0.015) 

0.008 

(0.010) 

 
 
 
 

0.047** 

(0.019) 

0.023** 

(0.011) 

0.379*** 

0.625*** 

(0.052) 

17,163 

(0.058) 

14,133 

 
 
 
 

0.044** 

(0.018) 

0.012 

(0.010) 

0.124** 

(0.053) 

13,872 

 
 
 
 

 
 
 
 

 
 
 
 

 
 
 
 

-0.148*** 

-0.038*** 

-0.080*** 

-0.096*** 

(0.013) 

(0.009) 

(0.010) 

(0.013) 

-0.025*** 

(0.008) 

-0.011* 

(0.006) 

-0.025*** 

-0.045*** 

(0.007) 

(0.008) 

0.714*** 

0.191*** 

0.394*** 

0.574*** 

(0.046) 

22,166 

(0.032) 

22,166 

(0.037) 

22,166 

(0.045) 

22,166 

R-squared 
Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village level 
 

0.125 

0.191 

0.127 

0.023 

0.104 

0.050 

0.054 

0.039 

0.157 

 

 

 

 

 
 

217 

 

  Table 4. 46: TESTING PARALLEL TREND - RDD USING DISTRICTS ON THE NEWLY FORMED STATES AS 
TREATMENT AND PARENT STATE AS COMPARISON  

  

(1) 

(2) 

HHH 

Education 

Electricity 

Drinking 

(4) 

(5) 

Concrete 

house 

Toilet 

(6) 

Clean 

Cooking 

Fuel 

(7) 

(8) 

(9) 

Television 

Refrigerator 

Motorcycle 

  

  

  

  

  

  

VARIABLES 

  
New State Border Dist. # 
Dummy Post 

 
Dummy Post 

 

Dummy Treatment 

 

  

  

-0.186 

(0.331) 

-0.267 

(0.202) 

-1.013 

(0.766) 

0.023 

(0.018) 

-0.020* 

(0.011) 

(0.041) 

Age of head of household 

-0.055*** 

0.002*** 

 

(0.005) 

Sex of HH Head=  female 

-3.212*** 

(Base = Male) 

(0.282) 

Education of the HH head 

 

 

 

(0.000) 

0.038** 

(0.015) 

0.010*** 

(0.001) 

Household Size 

0.072*** 

0.009*** 

0.265*** 

0.095*** 

(0.022) 

(0.001) 

-4.209*** 

-0.486*** 

(0.193) 

(0.010) 

-2.512*** 

-0.031** 

(0.241) 

(0.013) 

0.931 

(0.894) 

0.995 

(0.839) 

0.593 

(0.675) 

0.066 

(0.048) 

0.126*** 

(0.045) 

0.038 

(0.036) 

 

Rural 

(Base=urban) 
Religion of the HH 
(Muslim) 

(Base= Hindu) 
Religion of the HH 
(Christian) 

 

Religion of the HH (Sikh) 

 

Religion of the HH 
(Others) 

 

Caste of the 
household(SC) 

 

(3) 

Piped 

Water 

  

0.026 

(0.023) 

0.001 

(0.002) 

(0.007) 

0.000* 

(0.000) 

(0.003) 

-0.000 

(0.000) 

-0.000 

(0.000) 

0.003 

(0.002) 

-0.003 

(0.002) 

0.007 

(0.009) 

-0.007 

(0.008) 

0.003 

(0.006) 

-0.001 

(0.014) 

0.010 

(0.009) 

-0.000 

(0.033) 

-0.086 

(0.074) 

-0.052* 

(0.031) 

-0.008 

(0.017) 

0.061*** 

0.062*** 

0.050*** 

(0.008) 

(0.007) 

(0.011) 

0.095*** 

0.081*** 

0.155*** 

(0.032) 

(0.025) 

(0.040) 

-0.021 

(0.015) 

0.016* 

(0.009) 

0.046 

(0.035) 

0.002*** 

0.002*** 

0.001*** 

0.001*** 

0.001*** 

(0.000) 

(0.000) 

-0.006** 

0.046*** 

0.036*** 

(0.012) 

(0.012) 

(0.000) 

0.012 

(0.009) 

(0.000) 

0.035** 

(0.015) 

(0.000) 

0.005 

(0.013) 

0.009*** 

0.009*** 

0.006*** 

0.010*** 

0.004*** 

(0.000) 

0.002** 

(0.001) 

(0.000) 

(0.000) 

(0.001) 

0.002** 

-0.005*** 

0.011*** 

(0.001) 

(0.001) 

(0.001) 

(0.000) 

0.001 

(0.001) 

-0.006 

(0.016) 

-0.012 

(0.009) 

0.071** 

(0.036) 

0.000** 

(0.000) 

0.007 

(0.013) 

0.006*** 

(0.000) 

0.007*** 

(0.001) 

-0.450*** 

-0.547*** 

-0.339*** 

-0.367*** 

-0.179*** 

-0.145*** 

(0.009) 

(0.008) 

(0.006) 

(0.010) 

(0.009) 

(0.009) 

-0.052*** 

0.110*** 

-0.074*** 

-0.128*** 

-0.033*** 

-0.049*** 

(0.011) 

(0.010) 

(0.008) 

(0.013) 

(0.011) 

(0.011) 

0.086** 

0.150*** 

0.111*** 

(0.037) 

0.063* 

(0.035) 

-0.041 

(0.028) 

(0.029) 

0.096** 

(0.047) 

0.026 

(0.040) 

0.225*** 

0.207*** 

0.263*** 

(0.027) 

(0.044) 

(0.038) 

0.101*** 

(0.022) 

0.024 

(0.035) 

0.041 

(0.031) 

0.134*** 

(0.042) 

0.203*** 

(0.039) 

0.014 

(0.032) 

(0.039) 

0.147*** 

(0.036) 

-0.021 

(0.029) 

-0.082*** 
 
 

218 

-1.662*** 

-0.072*** 

-0.009*** 

-0.075*** 

-0.053*** 

-0.082*** 

-0.037*** 

-0.059*** 

 

 

 

 

 

 

 

 

 

 

 

Table 4. 46 (cont’d) 
 
 
(Base= General Caste) 
Caste of the 
household(ST) 

 

Constant 

 
Observations 

R-squared 

(0.203) 

(0.011) 

(0.002) 

(0.009) 

(0.008) 

(0.007) 

(0.011) 

(0.009) 

(0.010) 

-1.798*** 

-0.053*** 

0.011*** 

-0.070*** 

-0.076*** 

-0.047*** 

-0.101*** 

-0.036*** 

-0.055*** 

(0.223) 

(0.012) 

12.054*** 

0.396*** 

(0.736) 

10,361 

0.118 

(0.040) 

10,361 

0.367 

(0.002) 

-0.004 

(0.007) 

10,361 

0.121 

(0.010) 

(0.009) 

(0.007) 

(0.012) 

(0.010) 

0.406*** 

0.405*** 

0.283*** 

0.195*** 

0.116*** 

(0.032) 

10,361 

0.405 

(0.031) 

10,361 

0.486 

(0.024) 

10,312 

0.400 

(0.039) 

10,361 

0.282 

(0.034) 

10,361 

0.106 

(0.011) 

0.099*** 

(0.035) 

10,360 

0.099 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village level 
 

 

 
 

219 

 

Table 4. 47: TESTING PARALLEL TREND (HEALTH INDICATORS)- RDD USING DISTRICTS ON THE NEWLY FORMED 
STATES AS TREATMENT AND PARENT STATE AS COMPARISON  

  

VARIABLES 

  

(1) 

Govt 

Institutional 

Infant Death 

Delivery 

(2) 

(3) 

(4) 

(5) 

(6) 

(7) 

(8) 

(9) 

Under-
weight 

Stunted 

Wasted 

Tetanus Shot 

Polio 

Vaccine 

Measles 
Vaccine 

BCG 

Vaccine 

  

  

  

  

  

Treatment x Dummy Post 

-0.001 

-0.002 

0.134 

0.029 

0.114 

 

(0.016) 

(0.021) 

(0.140) 

(0.051) 

(0.149) 

  

-0.016 

(0.033) 

  

  

  

0.003 

-0.083** 

0.055* 

(0.024) 

(0.027) 

(0.033) 

Dummy Post 

-0.014 

0.005 

-0.061** 

-0.029 

-0.022 

0.241*** 

-0.145*** 

0.022 

0.066*** 

 

(0.010) 

(0.013) 

(0.025) 

(0.026) 

(0.025) 

Treatment  RDD 

0.059* 

-0.004 

-0.191** 

-0.080 

-0.170* 

 

(0.033) 

(0.042) 

(0.081) 

(0.094) 

(0.091) 

(0.019) 

-0.013 

(0.067) 

(0.014) 

(0.016) 

(0.019) 

0.035 

0.187*** 

0.255*** 

(0.048) 

(0.055) 

(0.066) 

Age of HH Head 

0.000 

-0.000 

-0.000 

-0.001 

0.000 

0.001** 

0.000 

0.001 

0.001** 

 

(0.000) 

(0.000) 

(0.001) 

(0.001) 

(0.001) 

(0.000) 

(0.000) 

(0.000) 

(0.000) 

Sex of HH head= female 

0.035** 

-0.008 

-0.044 

0.023 

-0.016 

0.065** 

-0.005 

0.011 

0.077** 

(Base = Male) 

(0.015) 

(0.019) 

(0.036) 

(0.036) 

(0.034) 

(0.030) 

(0.022) 

(0.025) 

(0.030) 

Women Education 

0.011*** 

-0.00*** 

-0.01*** 

-0.01*** 

-0.002 

0.039*** 

0.013*** 

0.023*** 

0.031*** 

 

(0.001) 

(0.001) 

(0.002) 

(0.002) 

(0.002) 

(0.002) 

(0.001) 

(0.001) 

(0.002) 

Rural 

-0.076*** 

0.020* 

0.074*** 

0.043* 

0.045** 

-0.15*** 

-0.06*** 

-0.05*** 

-0.03*** 

(Base=urban) 

(0.009) 

(0.011) 

(0.023) 

(0.022) 

(0.021) 

Household Size 

-0.002** 

-0.00*** 

-0.002 

0.001 

-0.003 

(0.001) 

(0.001) 

(0.002) 

(0.002) 

(0.002) 

(0.018) 

-0.002 

(0.002) 

(0.013) 

(0.015) 

(0.018) 

0.002** 

0.003** 

0.004** 

(0.001) 

(0.001) 

(0.002) 

-0.033*** 

-0.015 

0.092*** 

0.034 

0.049** 

-0.040** 

-0.04*** 

-0.07*** 

-0.11*** 

(0.010) 

(0.012) 

(0.025) 

(0.024) 

(0.023) 

(0.020) 

(0.014) 

(0.016) 

(0.020) 

0.025 

0.036 

-0.100 

0.026 

-0.143 

0.026 

-0.012 

0.024 

0.230*** 

(0.043) 

(0.056) 

(0.116) 

(0.110) 

(0.109) 

(0.087) 

(0.062) 

(0.071) 

(0.086) 

 
 

220 

 

Religion of the HH 
(Muslim) 

(Base= Hindu) 
Religion of the HH 
(Christian) 

 

 

 

Table 4. 47 (cont’d) 
 
 
Religion of the HH (Sikh) 

 
Religion of the HH 
(Others) 

 

Caste of the 
household(SC) 

(Base= General Caste) 
Caste of the 
household(ST) 

Constant 

 

 

Observations 

R-squared 

0.154*** 

0.074 

-0.281** 

-0.36*** 

-0.110 

0.104 

0.196*** 

0.205** 

0.158 

(0.050) 

(0.062) 

(0.128) 

(0.120) 

(0.114) 

(0.101) 

(0.072) 

(0.082) 

(0.100) 

-0.060** 

0.084** 

0.237*** 

0.011 

0.164** 

0.103* 

0.112*** 

0.116*** 

0.189*** 

(0.027) 

(0.033) 

(0.072) 

(0.072) 

(0.070) 

(0.055) 

(0.039) 

(0.045) 

(0.054) 

-0.038*** 

-0.002 

0.046* 

0.054* 

0.047 

-0.12*** 

-0.05*** 

-0.10*** 

-0.1*** 

(0.011) 

(0.013) 

(0.026) 

(0.030) 

(0.029) 

(0.021) 

(0.015) 

(0.017) 

(0.021) 

-0.007 

0.015 

-0.003 

0.012 

0.013 

-0.036** 

-0.003 

0.000 

-0.022 

(0.009) 

(0.011) 

(0.022) 

(0.021) 

(0.021) 

(0.018) 

(0.012) 

(0.014) 

(0.017) 

0.091*** 

0.116*** 

0.671*** 

0.663*** 

0.488*** 

0.509*** 

0.184*** 

0.076 

0.155** 

(0.031) 

(0.039) 

(0.077) 

(0.074) 

(0.071) 

(0.063) 

(0.045) 

(0.051) 

(0.062) 

5,271 

0.097 

5,096 

0.024 

4,043 

0.069 

3,542 

0.075 

3,471 

0.052 

5,271 

0.221 

5,271 

0.109 

5,271 

0.143 

5,271 

0.172 

Note: * p<0.1; ** p<0.05; *** p<0.01. Standard errors are in parentheses. All regressions have district fixed effect and standard errors are clustered at the village level

 

 
 

221 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

REFERENCES 

 
 

222 

 

REFERENCES 

 

 

Alesina, Alberto and Enrico Spolaore, “On the Number and Size of Nations,” The Quarterly 
Journal of Economics, 1997, 112 (4). 

Asher, S., & Novosad, P. (2015). The Impacts of Local Control over Political Institutions: 
Evidence from State Splitting in India. Unpublished manuscript. 

Banerjee, A. V., Gertler, P. J., & Ghatak, M. (2002). Empowerment and efficiency: tenancy 
reform in West Bengal. Journal of political economy, 110(2), 239-280. 

Bertrand, M., Duflo, E., & Mullainathan, S. (2004). How much should we trust differences-in-
differences estimates?. The Quarterly journal of economics, 119(1), 249-275. 

Easterly, W. (2007). Inequality does cause underdevelopment: Insights from a new 
instrument. Journal of Development Economics, 84(2), 755-776. 

Easterly, W., & Levine, R. (1997). Africa's growth tragedy: policies and ethnic divisions. The 
quarterly journal of economics, 112(4), 1203-1250. 

Easterly, W., Ritzen, J., & Woolcock, M. (2006). Social cohesion, institutions, and 
growth. Economics & Politics, 18(2), 103-120. 

Guha, R. (2007). India after Gandhi: The history of the world's largest democracy.  

Home Minister L.K Advani, speech to Loksabha, 25 July 2000 

Imbens, G. W., & Lemieux, T. (2008). Regression discontinuity designs: A guide to 
practice. Journal of econometrics, 142(2), 615-635. 

Khan, R. (1997). Rethinking Indian Federalism. Inter-University Centre for Humanities and 
Social Sciences, Indian Institute of Advanced Study. 

Kumar, A. (2010). Exploring the demand for new states. Economic & Political Weekly, 45(33), 
15. 

Kumar, P. (2000). The Uttarakhand movement: Construction of a regional identity. Kanishka 
Publishers. 

Lee, D. S., & Lemieux, T. (2009). Regression discontinuity designs in economics (No. w14723). 
National Bureau of Economic Research. 

 

 
 

223 

 

Lee, David and Thomas Lemieux, “Regression discontinuity designs in economics,” Journal of 
Economic Literature, 2010, 48 (2), 281–355 

Lemieux, T., & Milligan, K. (2008). Incentive effects of social assistance: A regression 
discontinuity approach. Journal of Econometrics, 142(2), 807-828. 

Mawdsley, E. (2002). Redrawing the body politic: federalism, regionalism and the creation of 
new states in India. Commonwealth and Comparative Politics,40(3), 34-54. 

N.K. Singh, Uttaranchal, Vananchal and Chhattisgarh celebrate their statehood, India Today  
August 14, 2000 

Papay, J. P., Willett, J. B., & Murnane, R. J. (2011). Extending the regression-discontinuity 
approach to multiple assignment variables. Journal of Econometrics, 161(2), 203-207. 

Sachar Rajindar  Why the NDA lost , The Hindu, May 27, 2004 

Ram, F., & Shekhar, C. (2006). Ranking and mapping of districts based on socio-economic and 
demographic indicators. Mumbai: International Institute for Population Science. 

Simon Johnson, and James Robinson, “The Colonial Origins of Comparative Development: An 
Empirical Investigation,” The American Economic Review, 2009, 91 (5), 1369–1401. 

Sreelatha Menon, 'Smaller state Key to Growth', Business Standard 17 April 2011 

Tillin, L. (2013). Remapping India: New states and their political origins. Hurst Publishers. 

Stepan, A., Linz, J. J., & Yadav, Y. (2011). Crafting state-nations: India and other multinational 
democracies. JHU Press. 

Suberu, R. (2001). Federalism and ethnic conflict in Nigeria. 

UNICEF (2017)Reducing Stunting in Children Under Five Years of Age: A Comprehensive 
Evaluation of UNICEF’s Strategies and Programme Performance India Country Case Study 

 William Easterly, “Public Goods and Ethnic Divisions,” Quarterly Journal of Economics, 1999, 
114 (4), 1243–1284.   

William Easterly, and Janina Matuszeski, “Artificial states,” Journal of the European Economic 
Association, 2011, 9 (2), 246–277. 

Wooldridge, J. M. (2010). Econometric analysis of cross section and panel data. MIT press. 

 

 

 
 

224