DO LEGAL REFORMS IN FAVOR OF WOMEN IMPROVE THEIR ECONOMIC AND
POLITICAL OUTCOMES? EVIDENCE FROM INDIA
By
Fang Xia

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Agricultural, Food and Resource Economics – Doctor of Philosophy
2013

ABSTRACT
DO LEGAL REFORMS IN FAVOR OF WOMEN IMPROVE THEIR ECONOMIC AND
POLITICAL OUTCOMES? EVIDENCE FROM INDIA
By
Fang Xia
Many developing countries suffer from long-standing and large bias against women in
social, economic and political spheres. The dissertation focuses on the case of India not only for
its large gender gaps in key human development indicators but also for its great effort to address
this social problem since its independence. The Amendment to the 1956 Hindu Succession Act
rd

grants daughters equal rights to inherit joint family property with sons. The 73 Constitutional
Amendment requires the village head‟s position to be reserved for women in one-third of
villages. The dissertation consists of three essays that aim to assess the impacts of these two legal
reforms.
Essay one examines whether and to what extent the Amendment to the 1956 Hindu
Succession Act (HSAA) affects women‟s intergenerational transfers of physical and human
capital. Our primary estimation strategy is the difference-in-difference estimator in which we
compare the share of assets received by male and female siblings in the same household before
and after the HSAA came into force. In the case of human capital investment, we compare the
primary education attainment of the young cohorts who were young enough to potentially benefit
from the reform to the old cohorts who were unlikely to benefit from the reform. In light of the
fact that the amendment applies only to Hindus but not to Muslims, we compare the regression
results between Hindus and Muslims for a robustness check. Results suggest that the amendment
increased the share of total assets received by Hindu females who were single before the reform.

They also point towards an increase in the share of gifts transferred to Hindu females who
married after the amendment. In the meanwhile, Hindu girls who were in primary schools and
who were going to enroll in primary schools after the amendment gained more years of primary
education than boys relative to the old cohort.
Essay two explores the impacts of political quotas for women mandated by the 73

rd

Constitutional Amendment on public service delivery, political participation and female
empowerment within the household. As reserved seats are randomly assigned across villages,
OLS regressions of outcome variables on a reservation dummy will yield valid estimates. We do
not find that reservation moved resources in a “pro-female” direction for improving public
service delivery. However, such unexpected effects were weighed against significant and in most
cases persistent impacts on outcomes such as meeting attendance and complaints to local
authorities if problems associated with public service delivery were identified. Reservation in the
current and in the previous period also increased the likelihood of women choosing the best form
of birth control.
Essay three examines whether representation of women persists after political gender
rd

quotas introduced by the 73 Constitutional Amendment lapse. We complement the relevant
literature by using nationally-representative data reflecting population diversity to achieve a
higher degree of external validity. Results show that reservation did not increase the likelihood
of women winning open seats, and failed to encourage women to run for open seats. However,
we note that, after exposed to reservation, villagers had a clear propensity to vote for candidates‟
qualifications rather than their social identities. Moreover, past reservation reduced the
likelihood of women being influenced by their husbands in their voting decisions.

Copyright by
FANG XIA
2013

ACKNOWLEDGMENTS

I would like to use this opportunity to thank many individuals who have contributed to
the completion of this dissertation. First and foremost, I would like to thank Songqing Jin, my
major professor, for his guidance, mentoring and support throughout my years at Michigan State
University. I truly could not have done it without you. I am grateful to my dissertation committee
members, Soren Anderson, Eric Crawford, and Thomas Reardon, whose insightful comments
and suggestions have greatly improved my dissertation and approach to research. This
dissertation would not have been possible without the support of Klaus Deininger, my mentor
and collaborator from the World Bank, who kindly shared data, discussed with me about
research ideas and empirical results, and provided research funding. I would also like to thank
my fellow graduate students, faculty, and staff in the Department of Agricultural, Food and
Resource Economics for their company and help throughout my time at Michigan State
University. Finally, I reserve my special appreciation to my family. I thank my parents and my
husband for their unwavering love and support.

v

LIST OF CONTENTS

LIST OF TABLES ....................................................................................................................... viii
LIST OF FIGURES ........................................................................................................................ x
INTRODUCTION .......................................................................................................................... 1
REFERENCES ............................................................................................................................... 8
CHAPTER 1: DOSE REFORM OF INHERITANCE LAW IMPROVE WOMEN‟S ACCESS
TO CAPITAL? EVIDENCE FROM URBAN INDIA ................................................................... 9
1.1 Introduction ........................................................................................................................... 9
1.2 Background ......................................................................................................................... 12
1.3 Data and descriptive statistics ............................................................................................. 16
1.3.1 Sample composition and construction .......................................................................... 16
1.3.2 Descriptive statistics ..................................................................................................... 18
1.3.3 Estimation strategy ...................................................................................................... 20
1.4 Econometric results ............................................................................................................. 22
1.4.1 Total assets received if the father died ......................................................................... 23
1.4.2 Total assets received if the mother died ....................................................................... 24
1.4.3 Assets received as gifts if the father and the mother are alive ..................................... 25
1.4.4 Educational attainment ................................................................................................. 26
1.5 Conclusion........................................................................................................................... 28
APPENDIX ................................................................................................................................... 43
REFERENCES ............................................................................................................................. 46
CHAPTER 2: DOES POLITICAL RESERVATION EMPOWER WOMEN OUTSIDE AND
WITHIN HOUSEHOLDS? EVIDENCE FROM RURAL INDIA .............................................. 49
2.1 Introduction ......................................................................................................................... 49
2.2 Gender quotas: Rationale, evidence, and approach............................................................. 51
2.2.1 Female political participation in perspective ................................................................ 51
2.2.2 Gender reservations in the Indian context .................................................................... 54
2.3 Conceptual framework and hypotheses............................................................................... 59
2.4 Data sources and estimation strategy .................................................................................. 61
2.5 Econometric results ............................................................................................................. 66
2.5.1 Problems of public service delivery ............................................................................. 66
2.5.2 Political participation .................................................................................................... 68
2.5.3 Female empowerment................................................................................................... 71
2.6 Conclusion and policy implications .................................................................................... 71
APPENDIX ................................................................................................................................... 83
REFERENCES ............................................................................................................................. 87
CHAPTER 3: DOSE POLITICAL RESERVATION AFFECT LONG-TERM ELECTORAL
PARTICIPATION AND OUTCOMES? EVIDENCE FROM RURAL INDIA .......................... 91
vi

3.1 Introduction ......................................................................................................................... 91
3.2 Institutional context and data .............................................................................................. 94
3.3 Hypotheses and estimation strategy .................................................................................... 96
3.4 Econometric results .......................................................................................................... 100
3.4.1 Women‟s likelihood of winning open seats and channels .......................................... 100
3.4.2 Voting patterns ........................................................................................................... 101
3.5 Conclusion......................................................................................................................... 103
APPENDIX ................................................................................................................................. 111
REFERENCES ........................................................................................................................... 114

vii

LIST OF TABLES

Table 1.1 Received assets of generation II………………………………………………………31
Table 1.2 Educational attainment of generation III……………………………………………...33
Table 1.3 Share of assets received from parents if fathers died (generation II)…………………34
Table 1.4 Share of assets received from parents if mothers died (generation II)………………..35
Table 1.5 Share of assets received from parents if fathers and mothers are alive (generation II).36
Table 1.6 Primary educational attainment (generation III)………………………………………37
Table 1.A1 Descriptive statistics of matched individuals………………………………………..44
Table 2.1 Distribution of sample villages across states and reservation status………………….74
Table 2.2 Initial village and household characteristics by reservation status of the village……..75
Table 2.3 Preference for public goods by gender………………………………………………..77
Table 2.4 Distribution of female pradhans across groups and reservation status………………..78
Table 2.5 Problems with delivery of key public services………………………………………..79
Table 2.6 Attendance and nature of participation in panchayat meetings……………………….80
Table 2.7 The likelihood of complaint to public authorities if problems identified……………..81
Table 2.8 Female empowerment…………………………………………………………………82
Table 2.A1 Problems with delivery of key public services (drinking water included)…………84
Table 2.A2 The likelihood of complaint to public authorities if problems identified (drinking
water included)…………………………………………………………………………………...85
Table 3.1 Distribution of reservation status…………………………………………………….105
Table 3.2 Descriptive statistics for key dependent variables…………………………………...106
Table 3.3 Initial village and household characteristics by reservation status of the village……107
Table 3.4 Women running for office in the next open election………………………………...109
viii

Table 3.5 Villager‟s voting pattern in the next open election…………………………………..110
Table 3.A1 Villager‟s voting pattern in the next open election (determinants of vote without
normalization)…………………………………………………………………………………..112

ix

LIST OF FIGURES

Figure 1.1 Coefficients on Female*Father‟s death year and Female*Father‟s death year*Single
before 1994 for the share of assets received from parents by Hindus (generation II)…………...39
Figure 1.2 Coefficients on Female*Father‟s death year and Female*Father‟s death year*Single
before 1994 for the share of assets received from parents by Hindus (generation II)…………...40
Figure 1.3 Coefficients on Female*Married years for the share of gifts received from parents by
Hindus (generation II) ……………………………………………………………………………41
Figure 1.4 Coefficients on Female*Birth year for primary educational attainment for Hindus
(generation III)…………………………………………………………………………………...42

x

INTRODUCTION

Women‟s empowerment and gender equality not only matter for their own sake, as
presented in Millennium Development Goals 3 and 5, but also contribute enormously to
economic development, political choices and welfare of the future generation, as documented by
a growing body of research (Chattopadhyay and Duflo 2004; Qian 2008; Udry 1996). Despite
1

the steady progress in closing the gender gaps in developing countries over the past decades ,
gender disparities persist in mortality, political representation, earnings, asset ownership, and
other areas. For instance, between 1990 and 2008, only 90 out of 147 countries which had a
decrease in maternal mortality rate showed a decline of 40% or more, and 23 countries actually
experienced an increase (WHO, UNIFEC, UNFPA, and World Bank 2010). In the same period
(1995-2009), the share of women parliamentarians increased only from 10% to 17% (World
Bank 2011). In large parts of Africa and South Asia, women possess only temporary rights to
land and are more likely to become victims of land conflicts (Deininger and Castagnini 2006;
Goldstein and Udry 2008).
However, economic growth alone cannot bring about gender equality due to the negative
stereotype created by social norms and cultural customs. For example, in India, the incentive and
the ability of women to use their earnings to influence household decisions depends importantly
on their social background. Only women from the former slave castes with weaker ties to their

1

According to World Bank (2011) female life expectancy increased dramatically in developing
countries (by 20 to 25 years in most regions in the past 50 years) to reach 71 years globally in
2007 (compared with 67 for men), and women now outlive men in every region of the world.
Two-thirds of all countries have reached gender parity in primary education enrollments, while
in over one-third, girls significantly outnumber boys in secondary education. Between 1980 and
2008, the gender gap in participation narrowed from 32 percentage points to 26 percentage
points. By 2008, women represented more than 40 percent of the global labor force.
1

ancestral communities are more able to challenge the norm of male decision-making (Luke and
Munshi 2011). In developed East Asian countries, a married woman is expected to perform most
of home production regardless of her and her husband‟s comparative advantages in terms of age,
education and labor force participation. Japanese and Korean husbands spend only some 40
minutes per day in home production, as compared to 167 minutes for married men in the United
States (Kawaguchi and Lee 2012).
Interventions that aim to empower women within households and increase women‟s
societal voices therefore provide an avenue for redressing the bias. For instance, in Ethiopia, the
Revised Family Code of 2000 requires that the husband and wife jointly choose the family‟s
place of residence and manage family property (Stark 2005). Unilateral divorce laws in the
United States reduced domestic violence by about 30% (Stevenson and Wolfers 2006). The
representation of women in parliament increased from 16% to 22.6% in Mexico, following the
imposition of candidate quotas (World Bank 2011).
The objective of this dissertation is to assess the impacts of two nationwide legal reforms
in India aiming to improve women‟s economic and political status. The first essay (chapter 1) is
entitled “Does Reform of Inheritance Law Improve Women‟s Access to Capital? Evidence from
Urban India”. In this essay, we use a large household survey data collected in 2011 from 7894
Hindu and Muslim households in Karnataka to analyze the impacts of the Amendment to the
1956 Hindu Succession Act. The 1956 Hindu Succession Act granted daughters equal shares in
separate property of deceased Hindus as sons and spouses if the Hindus died without making
wills, but deprived daughters and widows of their rights to be coparceners of joint family
property. Concerns over the ethical injustice and problems associated with increasing dowry,
amendments to the act were proposed by some southern states over the past twenty years and

2

expanded to the entire nation in 2005. These amendments are essentially identical across states,
giving daughters equal rights to inherit joint family property as sons.
Our primary estimation strategy is the difference-in-difference estimator in which we
compare the share of assets received by male and female siblings in the same household before
and after the HSAA came into force. In the case of human capital investment, we compare the
primary education attainment of the young cohorts who were young enough to potentially benefit
from the reform and the old cohorts who were unlikely to benefit from the reform. In light of the
fact that the amendment applies only to Hindus but not to Muslims, we compare the regression
results between Hindus and Muslims for a robustness check.
This essay contributes to the literature in two ways. First, while an extensive literature
concerns inequality caused by intergenerational transfers (Davies 1982; De Nardi 2004), few
focus on inequality between males and females, leaving gaps to be filled. Second, it is widely
acknowledged that access to resources and opportunities could empower females in private and
public spheres, and thereby lead to desirable social and economic outcomes (Hoddinott and
Haddad 1995; Stevenson and Wolfers 2006). Granting equal inheritance rights to daughters
along with sons could also bring about favorable outcomes.
Our results suggest that the HSAA increased the share of total assets received by Hindu
females who were single before the amendment by 0.216. While these females received more
joint family property, the separate property received from their fathers decreased in the long run.
Our results also point towards an increase of 0.147 in the share of gifts from parents received by
Hindu females who married after the amendment. The HSAA materialized gender equality for
those who were single before the reform in terms of physical asset transfers, but failed to fully
eradicate the dowry system. In the meanwhile, Hindu girls who were in primary school and who

3

were going to enroll in primary schools after the amendment gained 0.594 years of more primary
education than boys in the same cohort relative to their older siblings who completed primary
education before the amendment.
The second essay (Chapter 2) is entitled “Does political reservation empower women
outside and within households? Evidence from rural India”. In this essay, we use data from
nationally-representative surveys of 155 villages in rural India conducted in 2000 and 2008 to
rd

estimate the impacts of randomly assigned political quotas for females introduced by the 73
rd

Constitutional Amendment. The 73

Constitutional Amendment mandated far-reaching

decentralization by establishing a three-tier system of district, block, and village-level councils.
The gram panchayat (GP) is the lowest tier of local government at village level. It comprises a
president (pradhan or sarpanch) and council members who are elected from the panchayat‟s
wards. Regular assemblies (gram sabhas) by all voters in the GP are meant to monitor
performance and increase democratic accountability. To prevent decentralization from
reinforcing the power wielded by traditional elites and to counter what was perceived as a legacy
of disenfanchisement and under-representation by females and other disadvantaged groups,
decentralization was combined with reservation of a share of seats for women as well as
scheduled castes and tribes. Most states reserve pradhan positions for women in one-third of
panchayats.
As panchayats subject to reservation are chosen randomly in any given period, OLS
regressions of outcome variables on a reservation dummy will yield valid estimates. We compare
the effectiveness of public service delivery, female‟s ability to participate and voice concerns in
political processes, and female empowerment in terms of reproductive choice and independent
management of financial resources between reserved and non-reserved panchayats.
4

While a large and increasing number of studies explore the reservation policy, evidence
on its impact is mixed and inconclusive (Ban and Rao 2008; Beaman et al. 2009; Chattopadhyay
and Duflo 2004; Rajaraman and Gupta 2008), because most of these studies aggregate effects at
the household level rather than allowing for heterogeneity within the household and focus on
contemporaneous effects during the immediate reservation period from a few states clustered in a
similar geographic region. This essay contributes to the debate by exploring gender-specific
effects and assessing whether some impacts of the reservation policy materialize only in a few
states or in a delayed fashion.
Our results suggest that reserved pradhan positions for women did not move resources in
a “pro-female” direction for improving public service delivery. However, such unexpected
effects were weighed against significant and in most cases persistent impacts on outcomes such
as meeting attendance and complaints to local authorities if problems associated with public
service delivery were identified. Specifically, current reservation increased the likelihood of
meeting attendance, relevant issues raised and participation in discussion for women in states
where gender discrimination is less severe by 6%, 11% and 20%, respectively. Reservation in the
current and in the previous period increased the likelihood of women complaints about problems
associated with public service delivery in states where gender discrimination is more severe by
7% and 18%, respectively. Previous reservation increased the likelihood of women choosing the
best form of birth control in all the states by about 10%.
The third essay (Chapter 3) is entitled “Does Political Reservation Affect Long-Term
Electoral Participation and Outcomes? Evidence from Rural India”. In this essay, I use the same
dataset and estimation strategy as in Chapter 2 but explore whether representation of women
rd

persists after reservation mandated by the 73 Constitutional Amendment lapse. We

5

complement the relevant literature (Beaman et al. 2009; Bhavnani 2009) by using nationallyrepresentative data reflecting population diversity to achieve a higher degree of external validity.
Our results suggest that reservations did not increase the likelihood of women winning
open seats and failed to encourage women to run for open seats. However, we note that, after
exposed to reservations, villagers had a clear propensity to cast their votes based on candidates‟
qualifications rather than their social identities. Specifically, past reservation reduced the
likelihood of men casting their votes based on candidates‟ caste and religion backgrounds by
2.5% and 2.7% respectively, and increased the likelihood of men casting their votes based on
candidates‟ ability to represent local problems to government by 3.9%. Women are estimated to
be less likely to cast their votes based on candidates‟ religion background by 2.4% and more
likely to cast their votes based on candidates‟ knowledge of local problems and ability to
represent local problems to government by 5.5% and 4.4%, respectively, in villages where
leadership was reserved for a woman than in unreserved villages. Moreover, past reservations
empowered women within the household as illustrated by the fact that the share of female
spouses whose voting decisions were manipulated by their husbands was 2.9% lower in reserved
villages than in the unreserved villages.

6

REFERENCES

7

REFERENCES

Ban, R. and V. Rao. 2008. “Tokenism or Agency? The Impact of Women‟s Reservations on
Village Democracies in South India.” Economic Development and Cultural Change 56(3):
501-30.
Beaman, L., R. Chattopadhyay, E. Duflo, R. Pande and P. Topalova. 2009. “Powerful Women:
Does Exposure Reduce Bias?” Quarterly Journal of Economics 124(4): 1497-540.
Bhavnani, R. 2009. “Do Electoral Quotas Work after They Are Withdrawn? Evidence from a
Natural Experiment in India.” American Political Science Review 103(1): 23-35.
Chattopadhyay, R. and E. Duflo. 2004. “Women as policy makers: Evidence from a randomized
policy experiment in India.” Econometrica 72(5): 1409-1443.
Davies, J. B. 1982. “The Relative Impact of Inheritance and Other Factors on Economic
Inequality.” Quarterly Journal of Economics, 9(3): 471-498
De Nardi, M. (2004), “Wealth Inequality and Intergenerational Links.” Review of Economic
Studies, 71(3): 743-768.
Hoddinott, John and Lawrence Haddad. 1995. “Does Female Income Share Influence Household
Expenditures? Evidence from Cote D‟Ivoire.” Oxford Bulletin of Economics and
Statistics 57(1): 77-96.
Kawaguchi, D. and S. Lee. 2012. “Brides for sale: Cross-border marriages and female
immigration.” Harvard Business School Working Paper 12-082.
Luke, N. and K. Munshi. 2011. “Women as agents of change: Female income and mobility in
Inida.” Journal of Development Economics 94(1): 1-17.
Qian, N. 2008. “Missing women and the price of tea in China: The effect of sex-specific earnings
on sex imbalance.” Quarterly Journal of Economics 123(3): 1251-1285.
Rajaraman, I. and M. Gupta. 2008a. “Further evidence on the policy impact of randomised
political reservation.” New Delhi: National Institute of Public Finance and Policy.
Stark, B. 2005. International family law: An introduction. Ashgate Publishing Limited.
Stevenson, B. and J. Wolfers. 2006. “Bargaining in the shadow of the law: Divorce laws and
family distress.” Quarterly Journal of Economics 121(1): 267-288.
Udry, C. 1996. “Gender, agricultural production, and the theory of the household.” Journal of
Political Economy 104(5): 1010-1046.
World Bank. 2011. World Development Report 2012: Gender equality and development. World
Bank, Washington, D.C.

8

CHAPTER 1: DOSE REFORM OF INHERITANCE LAW IMPROVE WOMEN’S
ACCESS TO CAPITAL? EVIDENCE FROM URBAN INDIA

1.1 Introduction
Widespread gender disparity exists in intergenerational transfers of physical and human
capital explicitly controlled by parents. Daughters are found to receive less education, land, and
total inheritance in the Philippines (Quisumbing 1994) and in Ghana (Quisumbing et al. 2004).
The World Bank‟s 2012 development report (World Bank 2011) shows that while remarkable
progress has been achieved in education for most countries, the gender gap remains severe for
the poorest segments of the population. For example, despite similar rates of participation in
school for boys and girls from the top income quintile (fifth) in India, girls lag behind boys by
almost five years in the bottom income quintile. This disparity could generate and widen gender
gaps in other domains, as material wealth and human capital investments are determinants of the
ease with which children can accumulate individual capital (Blinder 1973; Becker and Tomes
1979; Kotlikoff and Summers, 1981; Sheshinski and Weiss, 1982), which plays an important role
in the development of endowments, the distribution of earnings and wealth, the status in the
marriage market, bargaining power within the household, and the quantity as well as the quality
of the next generation. (Thomas 1990; Behrman et al. 1994; Brien and Lillard 1994; Zhang and
Chan 1999).
Legal measures could provide a potential to reduce or eradicate accumulative gender
discrimination created by the interactions of social norms and cultural customs, although
2

sometimes they surrender to the traditional complex and restrict women‟s inheritance rights . In

2

According to World Bank (2001) some customary laws give sons the exclusive right to inherit,
while wives and unmarried daughters have the right to be maintained, and married daughters
have no claim on their deceased father's property. Islamic law grants widows with children an
9

Western Ghana, the 1985 Intestate Succession Law allows wives to legally own the land granted
by husbands as gifts after wives help husbands establish cocoa farms (Quisumbing et al. 2001).
In South Africa, the Communal Property Association Act was passed in 1996 to allow
individuals to acquire land through membership in a communal property association (World
Bank 2001). In India, the amendment to the Hindu Succession Act 1956 which acknowledges
coparcenary ownership of daughters came into force nationally in 2005, following similar
changes in some southern states one or two decades earlier. However, empirical literature of
legal changes in inheritance rights is rare. To our knowledge, only Roy (2008) and Deininger et
al. (2010) quantitatively assess the impact of inheritance law in the context of India.
In this study, we use a large household survey data collected in 2011 from 7894 urban
Hindu and Muslim households to analyze the impact of the amendment of the Hindu Succession
Act in the urban context. The survey contains detailed information on the timing of key life
events, such as birth, death and marriage, and the level of education as well as assets received
from parents by male and female individuals. We estimate the impact of the Hindu Succession
Act amendment by taking advantage of the variation in the timing of death for the parents of
household heads and their spouses, the timing of marriage for household heads, their spouses and
the siblings of household heads and spouses, and the timing of decisions on primary education
for the children of household heads and their spouses. Specifically, our estimation strategy is
difference-in-difference in which we compare the share of total assets received by Hindu males
and females whose parents died before and after the amendment of the act, the share of gifts
received by Hindu males and females who married before and after the amendment of the act,

eighth of property upon their husband's death, while childless widows receive a fourth.
Daughters are entitled to half the amount their brothers inherit. Hindu law gives widows the right
only to maintenance.
10

and primary education years gained by Hindu boys and girls whose education decisions were
made before and after the amendment of the act, after household fixed effects are controlled for.
We rely on one more difference between Hindus and Muslims for a robustness check, given that
the amendment of the act applies to Hindus but not Muslims. Our results suggest that the HSAA
increased the share of total assets received by Hindu females who were single before the
amendment by 0.216. While these females received more joint family property, the separate
property they received from fathers decreased in the long run. Our results also point towards an
increase of 0.147 in the share of gifts from parents received by Hindu females who married after
the amendment. In the meanwhile, Hindu girls who were in primary school and who were going
to enroll in primary schools after the amendment gained 0.594 years of more primary education
than boys in the same cohort relative to their older siblings who completed primary education
before the amendment. The amendment of the act materialized gender equality for those who
were single before the reform in terms of physical asset transfers, but failed to fully eradicate the
dowry system.
The paper contributes to the literature in two ways. First, while an extensive literature
concerns inequality caused by intergenerational transfers, few focus on inequality between males
and females, leaving gaps to be filled. Davies (1982) shows a high income elasticity of bequests
and attributes inherited wealth as a major cause of wealth inequality. De Nardi (2004) finds that
voluntary bequests explain the emergence of large estates, and the introduction of a bequest
motive generates lifetime savings. Second, it is widely acknowledged that access to resources
and opportunities could empower females in private and public spheres, and thereby lead to
desirable social and economic outcomes. Hoddinott and Haddad (1995) show that raising wives‟
share of cash income increases the budget share of food, and reduces the budget shares of

11

alcohol and cigarettes. Stevenson and Wolfers (2006) find declines in females committing
suicide and murdered by their partners following the introduction of unilateral divorce. Granting
equal inheritance rights to females might also bring about favorable outcomes.
The paper is organized as follows: Section 2 provides context by reviewing India‟s Hindu
Succession Act and its amendment. Section 3 discusses the data used and the sample constructed,
reports descriptive statistics on physical capital transfers and human capital investments and
introduces the estimation strategy. Section 4 presents econometric results to quantify the impacts
of the institutional change on total assets received from parents, gifts received from parents, and
educational attainment. Section 5 concludes by drawing out implications for policy and possible
future research.

1.2 Background
3

The Hindu Succession Act 1956 (HAS) governed property rights of Hindus nationally ,
unifying two main schools of Hindu law that prevailed since the twelfth century AD –
4

Mitakshara and Dayabhaga , before state governments enacted legislation to amend it between
1986 and 2005. The Mitakshara system classifies property as separate property and joint family
5

property, while the Dayabhaga system identifies all property as separate property. The 1956

3

The Hindu Succession Act applies to Hindus, Buddhists, Jains and Sikhs but not Muslims,
Christians, Parsis and Jews.
4
Dayabhaga governed Bengal and Assam while Mitakshara dominated in the rest of the country
(Agarwal 1994).
5
According to Roy (2008) the most important distinction between these two schools was in
terms of their classification of property. The Mitakshara system made a distinction between
„joint family property‟ and „separate property‟. Joint family property „consisted principally of
ancestral property (that is, property inherited from the father, paternal grandfather or paternal
great-grandfather), plus any property that was jointly acquired or was acquired separately but
12

HSA granted Hindu daughters equal shares of separate property of deceased Hindus as sons and
6

spouses if the Hindus died without making wills, but deprived daughters and widows of their
rights to be coparceners for joint family property. On the contrary, sons not only enjoyed the
right to inherit parents‟ separate property, but also could receive joint family property, shared
7

only among the fathers plus his male linear descendants, and demand its partition. Therefore,
while daughters in Dayabhaga could possibly receive the same share of property from fathers
dying intestate as sons, they absolutely received a smaller share in Mitakshara as compared to
their brothers. The 1956 HAS seeking gender equality in inheritance failed to do so as
Mitakshara dominates most of India‟s states.
Amendments to the 1956 HSA were proposed by some states in the last twenty years of
the twentieth century (Andhra Pradesh in 1986, Maharashtra in 1989, and Karnataka and Tamil
8

Nadu in 1994) , and expanded the entire nation in 2005, triggered by not only the awareness that
the exclusion of daughters from participating in coparcenary ownership because of sex was
ethically unjust, but also the persistent inflation of dowry and associated violent behaviors. These
amendments are essentially identical across states, giving daughters who married after the reform

merged into the joint property‟ while separate property „included that which was self-acquired (if
acquired without detriment to the ancestral estate) and any property inherited from persons other
than his father, paternal grandfather or paternal great-grandfather‟ (Agarwal 1994). Under
Mitakshara, four generations of male members became joint heirs or coparceners to the joint
family property by birth while women had no such rights. The Dayabhaga system, on the other
hand, treated all property as self-acquired/separate property including the person‟s „notional‟
share of joint family property.
6
All Hindu individuals are entitled to will their separate property to a desired beneficiary.
7

The deceased father‟s notional share of joint property was allocated among all male and female
heirs, normally in equal shares.
8
Kerala abolished joint family property system and granted all family members their separate
share in 1976 (Agarwal 1994). The spirit of the amendment is the same as those in other states,
in favor of the inheritance of daughters.
13

equal rights to inherit joint family property with sons. The change introduced by the Hindu
Succession Act Amendment (HSAA) provides us a natural experiment to explore whether or not
the legislation empowered women in intergenerational transfers of physical and human capital.
Empirical studies are involved in assessing the eventual impact in four respects. First,
several legal measures for eliminating evil social practices such as dowry or caste discrimination
turned out to be fruitless in practice (Anderson 2003), casting doubt on the effectiveness of the
HSAA. Second, the direction of effects among physical capital transfers (joint family property,
separate property and dowry) is unknown. Regarding separate property, parents might either will
separate property away from daughters to maintain the existing allocation, as proposed by the
preference model (Behrman, Pollak, and Taubman 1982), or follow the amendment‟s spirit to
divide separate property equally among children. Regarding dowry, on the one hand, parents
might reduce daughters‟ dowry who married after 1994 because they realized daughters would
inherit more joint family property after the father was deceased, which is consistent with
preferences for inter-sibling equality (Behrman, Pollak, and Taubman 1982). On the other hand,
even potential inheritance might increase women‟s intrahousehold bargaining power and make
women more dependable for their parents, which could stimulate parents to transfer more to
daughters before death, consistent with the model of exchange-motivated bequests (Bernheim,
Schleifer, and Summers 1986). In addition, from the demand side, husbands‟ families may
demand more gifts from their wives‟ parents to take advantage of the improved economic
condition and social status of their wives (relative to their wives‟ male siblings). Third, human
capital could either substitute or complement physical capital. Parents could increase human
capital investment in sons and reduce that in daughters under a household budget constraint, or

14

vice versa. Fourth, whether or not equal property rights translated into females‟ greater
bargaining power within households needs to be empirically examined.
Some recent studies provide partial empirical evidence for these arguments. In a 2005-06
representative sample (the National Family Health Survey), Hindu females who married after the
HSAA came into force are found to enjoy more autonomy within their households, measured by
three self-reported indicators of independent travels. The estimated effect is stronger for females
whose husbands own land and engage in farming as land is the most frequent form of joint
family property (Roy 2008). Using the 2006 wave of the Rural Economic and Demographic
Survey, Deininger et al. (2010) compare within household bequests of land given to sons and
daughters by exploiting the variation in the timing of father‟s death, and find that daughters are
more likely to inherit land in Hindu families. Their results also point towards an increase in age
at marriage and in educational attainment of Hindu girls. While these studies cover rural areas
and focus on inheritance of land which constitutes the main asset and source of livelihood in
rural India, our sample from urban India allows us to explore asset transfers beyond joint family
property. We compare (i) the share of total assets received by females and males in the same
parental household whose parents passed away before and after the HSAA came into force; (ii)
the share of gifts received by females and males in the same parental household who married
before and after the HSAA took into effect; and (iii) educational attainment of girls and boys in
difference cohorts in the same parental household for whom decisions on primary education
were made before and after the implementation of the HSAA.

15

1.3 Data and descriptive statistics
1.3.1 Sample composition and construction
Our data are from the 2011 Urban Property Ownership Records (UPOR) survey
conducted in Karnataka‟s four cities: Davagere, Gulbarga, Mysore and Shimoga. The Household
survey collected detailed information on three generations of people: household heads, their
spouses and the siblings of household heads and spouses (generation II), the parents of
household heads and their spouses (generation I), and the children of household heads, their
spouses and the siblings of household heads and spouses (generation III). We observe basic
characteristics of generation II individuals (i.e., age, education and the year of marriage, as well
as assets received from their parents), the timing of their parent‟s deaths, and the educational
outcomes of their children. While we focus on a sample of 7894 households with 29660
generation II individuals to explore the effects of the HSAA on physical capital transfers, we rely
on a sample of 3112 households with 7948 generation III individuals to assess the effects of the
HSAA on human capital investments. In our sample, 6473 Hindu households with 23779
generation II individuals and 2557 Hindu households with 6122 generation III individuals will be
the sample for our main analysis. A sample of 5881 generation II individuals from 1421 Muslim
families and 1826 generation III individuals from 555 Muslim families are used in the placebo
analysis.
As an approximation of an ideal experiment, the Muslim sample is weighted to be most
similar to the Hindu sample in terms of observed characteristics within the same city. Instead of
claiming that the Hindu sample and the Muslim sample are identical after weighting, we assume
the relevant differences between the two samples are captured by observed characteristics, and
the religion factor only influences generation II‟s asset transfers and generation III‟s educational

16

attainment through the HSAA. Table 1.A1 illustrates the significance of these attributes‟
differences between Hindus and Muslims with and without weights. As expected, weighting
indeed improves the comparability between Hindu and Muslin population. First of all, weighting
significantly decreased the magnitude of difference between the Hindu and the Muslim samples
for some of the attributes. Moreover, weighting also reduced the number of attributes that differ
significantly between the Hindu and Muslim individuals (from 17 to 12 for generation II, and
from 5 to 4 for generation III).
Typical generation II individuals were born in the 1960s and attained between six and
eight years education. They were originally from five children households, formed new families
before 1994 and gave birth to two children. In addition to some 40% non-income individuals
(92% of them are females), they almost equally distributed in three categories of monthly income
– less than 5000 rupees, between 5000 and 9000 rupees, and more than 9000 rupees. Typical
generation III individuals were born in the 1980s. They were from families having three children,
and most of them (about 60%) were single in 2011. While half of generation III individuals did
not earn income (93% of them are domestic workers and students), 10% earned less than 5000
rupees per month, 20% earned between 5000 and 9000 rupees per month, and 15% earned more
than 9000 rupees.
We divide generation II individuals into sub-populations by whether or not one‟s parents
are still alive and in the case they are not, whether or not they died before 1994. 22% of Hindu
males have living fathers, 43% lost fathers before 1994, and 35% lost fathers after 1994. The
situation differs only slightly for Hindu females as 25% of them have living fathers, 40% lost
fathers before 1994, and 35% lost fathers after 1994. The situation is also comparable to the
Muslim population with the corresponding percentages 25%, 38% and 37% for Muslim males

17

and 27%, 36% and 37% for Muslim females. In the meantime, our data show that almost half of
the generation I females are still alive: in Hindu families, 44% of males‟ and 48% of females‟
mothers are still alive, while 25% of males and 23% of females (or 31% of males and 29% of
females) lost mothers before (or after) 1994. The generation II Muslim sample presents a
consistent composition, with the corresponding percentages 51%, 20% and 29% for males and
52%, 19% and 29% for females.

1.3.2 Descriptive statistics
Panel A of table 1.1 presents descriptive statistics on assets received from parents by
generation II Hindu individuals by whether or not one‟s father is still alive, and in the case he is
not, whether he died before or after 1994. Our data shows that Hindu females typically inherited
smaller shares of assets than their male siblings when their fathers died. Although, probably
mainly driven by dowry, females received 6% more assets from parents than men before fathers‟
death, males were compensated after fathers‟ death by inheriting more assets, resulting in 5%
more assets for males whose fathers died before 1994, and 3% more assets for males whose
fathers died after the implementation of the amendment. It is unlikely that evidence is due to the
household‟s demographic structure, suggested by a parallel trend of the share variable
conditional on the number of households‟ generation II members. While the descriptive statistics
for Muslim individuals show an overall similar pattern (panel B of table 1.1), there are two
differences. First, assets received by Muslim males after the death of their fathers were not large
enough to outweigh the total assets received by their female siblings from their parents. Second,
while Hindu females‟ share of assets increased from 0.87 to 0.92 conditional on household
demographics after the amendment took into effect, Muslim females‟ corresponding share

18

decreased from 0.98 to 0.94. However, a more conclusive result should be derived from an
econometric analysis which controls for multiple sources of heterogeneity. Panel C of table 1.1
shows the assets received by generation II Hindu individuals from parents by whether or not
one‟s mother is still alive and if she is not, by whether she died before or after 1994. We find an
opposite pattern as compared to results in panel A – conditional on household demographics
mothers transferred more assets to sons (from 1.08 to 1.13) and fewer assets to daughters (from
0.86 to 0.80) after the HSAA came into force. A similar pattern of asset transfers related to the
timing of mother‟s death is also observed in the Muslim sample (Panel D). Again, the exact
interpretation is not possible without controlling for other determinants.
Table 1.2 reports the level of education for the oldest generation III individuals born in
1970 (24 years old in 1994), and the youngest ones born in 1996 (15 years old in 2011, the year
when the survey was conducted). We focus on this age range for two reasons. First, we intend to
assess the HSAA‟s impact on primary education decisions. Normally primary schools in India
enroll 6-14 years old children, which implies every generation III individual in our sample was
old enough to complete primary education in 2011. Second, we aim to distinguish two treatment
groups, which are constructed based on the potential time length that the HSAA could influence
an individual‟s primary education. The two treatment groups include, individuals less than 5
years old in 1994, and those 6-14 years old in 1994. For the former, the entire primary education
could be affected by the HSAA, while for the latter, only part of the primary education years
were exposed to the HSAA. By comparison, the control group comprises those between 15 and
24 years in 1994 and whose primary education would have been completed before the passage of
the HSAA. Our data show that the share of Hindu girls who completed primary education is
considerably higher for the two treatment groups (80% and 83%) than for the control group

19

(71%). In the meantime, educational attainment of Hindu boys in the two treatment groups
decreased slightly from 83% to 80% and 81%. While our data show a high share of Muslim
individuals in the young cohorts completed primary or higher education than older cohorts (panel
B), the largest improvement is observed with the youngest girls (34% as compared to 9% for the
youngest boys). This could threaten our identification by suggesting some female-favored
educational programs after 1994. However, while the descriptive analysis is informative, it does
not allow us to interpret the casual relationship between HSAA and the outcome variables. We
will rely on econometrics analysis to identify the causal relationships, which we focus in the next
section.

1.3.3 Estimation strategy
While we hypothesize that the HSAA is likely to increase assets inherited by females
after the death of their fathers, how the reform affected total assets received by females from
their parents depends on (i) how it affected inheritance received and (ii) how it affected gifts
(mainly dowry). We examine the two effects separately in our econometric analysis. Our
estimation strategy is to compare the share of assets received by males and females in the same
household before and after the HSAA came into force. Specifically, we define the two equations
as below:
Sij = αj + β1Fij + β2Fij*Dj + β3Mij* Fij*Dj + β4Mij* Fij + β5Mij*Dj + β6Mij + Tij + εij
g

S ij = αj + γ1Fij + γ2Fij*Mij + γ3Mij + Tij + εij

20

(2)

(1)

where Sij is the share of total assets of individual i in household j received as gifts and
9 g

inheritance from parents, normalized by the number of generation II household members. S ij
in equations (2) only includes assets received as gifts, as here we concentrate on individuals
whose fathers and mothers are alive. αj is household fixed effects controlling for time-invariant
household characteristics, Fij, Dj, and Mij are indicator variables for female, whether or not the
father/mother died after 1994, and whether or not the individual was single before 1994, and Tij
is a vector of birth year dummies controlling for time-variant aggregate effects. β2 and β3 in
equation (1) and γ2 in equation (2) are key parameters of interest capturing the impacts of the
amendment on females‟ total assets received after fathers‟/mothers‟ death and on females‟ assets
received as gifts when fathers and mothers are alive. To assure that our estimation identifies the
HSAA‟s impacts rather than captures the long-run trends, we also augment the model by
including a vector of dummy variables for the robustness check. These variables include a set of
dummies for the year when the father/mother died, in place of the indicator variable Dj in
equations (1). By the same token, we also add a vector of indicator variables for the year when
the individual‟s marriage occurred, in place of the indicator variable Mij in equations (2).
To assess the impact of the HSAA on generation III individuals‟ educational attainment,
we estimate the following equation:

9

N

Sij = Aij*(N+1)/( Aij+∑ n=1 Anj). Let Aij indicate total assets received from parents by
individual i in household j, and N denote the number of individual i‟s siblings. Anj is total assets
N

received from parents by sibling n, and ∑ n=1 Anj is the sum of all assets received from parents
by the N siblings.
21

Eij = αj + β1Fij + β2Fij*Gij + β3Gij + εij

(3)

where Eij is the education attainment (years) of individual i in household j, truncated at the
highest grade of primary education, grade nine. αj and Fij are defined similarly as in equations
(1) and (2). Gij is a vector of indicator variables, including whether or not the individual was
born between 1980 and 1988 whose primary education decisions were supposed to be partially
affected by the HSAA, and whether or not the individual was born between 1989 and 1996
whose primary education decisions were supposed to be fully affected by the HSAA. β2 are the
coefficients of interest, measuring the impact of the amendment on educational attainment of
Hindu girls relative to boys. To assure that our estimation identifies the HSAA‟s impacts rather
than captures the long-run trends, a complete set of impacts varying by birth years since 1977 is
captured by Gij in the robustness check.

1.4 Econometric results
The HSAA enforces joint family property to be divided equally among children after the
father‟s death, but allows parents to allocate separate property as they wish. Past field studies
indicate that more than 65% of Indian people die every year without wills (Agarwal 1994),
suggesting that enormous separate property is likely to be distributed evenly among children
after one parent passes away. This could have changed after the implementation of the HSAA. It
is possible that generation I individuals began to will separate property by taking into account the
legal requirement on how the joint family property should be allocated. Since the survey did not
collect information on joint family property and separate property separately, it is impossible to

22

directly trace fathers‟ wills on how separate property is allocated because joint family property is
typically allocated upon the death of fathers. However, it is possible to directly trace mothers‟
wills on how separate property is allocated, because mothers‟ death only influences the
allocation of separate property but not joint family property. If the wills made by mothers who
died after 1994 deviate from those made by mothers who died before 1994, we would detect a
change in total asset transfers due to a change in the allocation of mothers‟ separate property.

1.4.1 Total assets received if the father died
The first two columns of table 1.3 show the results of equation (1) for generation II
Hindu individuals whose fathers passed away. In addition to the result for the base specification
(column 1), the results for the augmented specifications are reported in column 2 (to control for
the individual‟s marriage time effects). Placebo tests using the Muslim sample are reported in
columns 3-6. To make the analysis based on the Muslim sample comparable to that based on the
Hindu sample, we weight the Muslim sample using the propensity score. However, one caveat is
that Muslims cannot be an ideal group for the placebo analysis, since there is a long history of
different cultures and social norms between the two groups. In the assumption that Hindus and
Muslims have common trends, the analysis with a placebo should be less biased than the analysis
based only on the Hindu sample, but the small size of the Muslim sample is likely to lead to a
higher variance.
We find from the base specification (column 1) that, the share of assets received by
Hindu females from parents was 0.301 lower than that received by males, and the share of assets
received by Hindu females whose father died after the amendment took into effect was 0.123
higher than that received by Hindu females whose father died before 1994 relative to their male

23

siblings. To check whether or not the results vary by the individual‟s marriage time, we add to
the base specification the interaction terms of the father‟s death year and whether one was single
before 1994 (column 2). Consistent with the legal provisions, the HSAA mainly made Hindu
females who were single before 1994 better off and received equal share as their male siblings.
In contrast, we find the coefficients on the same variables are not only statistically insignificant,
but also have much smaller magnitude for the Muslim sample (columns 3 and 4), strengthening
the point that our estimation of the impact of the HSAA is indeed caused by the reform itself.
Figure 1.1 plots the trends for Hindu females who were married and single before the
amendment, respectively (from the augment specification controlling for time-variant effects). A
few interesting observations emerge from the graphical presentation. First of all, we find a sharp
increase in assets transferred to Hindu daughters who were single before the amendment appears
in the year when the amendment came into force, and the trend thereafter persists to fluctuate
above zero. On the contrary, the pre- and post- amendment coefficients on Hindu daughters who
married before the amendment fluctuate around zero. Second, the magnitude of the impact on
Hindu daughters who were single before the amendment tends to decline after it peaks in the first
year after the passage of the HSAA, suggesting that fathers willed separate property away from
daughters who are expected to benefit from the passage of the HSAA. This is consistent with the
model of preferences for inter-sibling equality (Behrman, Pollak, and Taubman 1982).

1.4.2 Total assets received if the mother died
Table 1.4 presents the results in the same way as table 1.3. Results for the base
specification (column 1) show that, the share of assets received by Hindu females whose mother
died after the amendment came into force was 0.103 lower than that received by Hindu females

24

whose mother died before 1994 as compared to their male siblings. The combination of the
significant coefficient on the interaction between the female dummy and the dummy variable for
whether or not her mother died after 1994, and the insignificant coefficient on the three-term
interaction (between female dummy, the dummy variable for whether or not she was single
before 1994 and the dummy variable for whether or not her mother died after 1994) in column 2
suggests that, the death of mother after 1994 only made daughters who married before 1994
worse off, but not those who were single before 1994. This is further supported by the graphical
results in figure 1.2 where time-variant effects are controlled for. While the trend for Hindu
females who were single before the reform fluctuates around zero throughout the years, the trend
for Hindu females who were married before the reform fluctuates below zero if their mothers
passed away after the amendment came into force. However, the Muslim sample shows a similar
pattern (columns 3 and 4), suggesting that we cannot attribute the asset allocation behaviors of
Hindu mothers to the HSAA.

1.4.3 Assets received as gifts if the father and the mother are alive
We are also interested in estimating the effects of the HSAA on assets received as gifts.
There are two reasons why this is important. First of all, eradicating the evils of dowry system is
another objective for amending the 1956 HSA. Second, to understand how the HSAA affects
transfers through other channels is an interesting issue in its own rights. To explore this issue, we
have to rely on a subsample that is composed of generation II married individuals whose fathers
and mothers were both alive at the time of survey so the assets they received from parents are
likely to be gifts.

25

The first column of table 1.5 shows the results of equation (2) for generation II Hindu
individuals. Placebo tests using the Muslim sample are reported in column 2 and 3. The results
(column 1) suggest that Hindu females received as many gifts as their brothers even if they
married before 1994, probably because of the practice of dowry. The results also points towards
an increase in the share of gifts received by Hindu females who were single before 1994 as
compared to their sisters who married before 1994 and their brothers who were single before
1994 by 0.144 and 0.167, respectively, among which 0.147 was brought about by the HSAA
exclusively. The much smaller coefficients on the corresponding variables for the Muslim
sample (column 2) again lend support to the validity of our identification strategy. Since we find
in section 4.1 that fathers willed separate property away from daughters who married in postamendment periods, the increasing gifts received by Hindu females could be attributed to the
demand of the husband‟s families rather than parents‟ voluntary behaviors in line with the model
of exchange-motivated bequests (Bernheim, Schleifer, and Summers 1986).
Figure 1.3 plots estimated coefficients on the interaction between the female dummy and
the indicator for the year of marriage based on the Hindu sample. We notice that, despite the
generally declining trend with fluctuations over time, the gifts received by Hindu females
married after the amendment is systematically higher than those married before the amendment,
suggesting that the HSAA failed to achieve its objective of eradicating the dowry system.

1.4.4 Educational attainment
In spite of a clear increase in physical capital transfers to Hindu females after the HSAA
came into force, how the HSAA affects the human capital of daughters relative to sons is less
clear. On the one hand, it is possible that parents could rebalance the overall resource allocation

26

by investing more in their sons‟ education than in their daughters‟ education to compensate for
the fact that their daughters would inherit more after the implementation of HSAA (consistent
with the hypothesis of Behrman, Pollak, and Taubman 1982). If this is the case, HSAA could
widen the gender gap. On the other hand, as suggested by the efficiency hypothesis (Bernheim,
Schleifer, and Summers 1986), parents are likely to increase investment in their daughters so
they can take advantage of the future improved economic opportunities under the new
inheritance law. Meanwhile, the improved bargaining power of wives relative to husbands under
the new inheritance law could alter household‟s education decisions to shift in favor of girls. We
focus on generation III individuals to assess the HSAA‟s impact on human capital investments
because their parents, generation II individuals, were not young enough for their primary
education decisions to be affected by the amendment.
The education regression results are presented in table 1.6. The significant and negative
coefficients on the female dummy in the base specification (column 1) show that girls attained
less primary education than boys in the Hindu sample (0.860 years less) before the reform. For
those whose education decisions were made after the reform, they had 0.594 more years of
primary schooling than their older cohort, but their level of education continues to be lower than
that of the boys in the same cohort. Results in column 2 suggest that the HSAA led to
quantitatively large educational improvements for both young cohorts (0.477 years for those
partially exposed to the reform and 0.701 years for those full exposed to the reform) as compared
to their sisters in the older cohort. The rejection of equal coefficients between the fully exposed
cohort and the partially exposed cohort further supports the hypothesis that the effect of the
HSAA on education increases with the time exposure to the reform. Figure 1.4 compares the

27

educational attainment of those who were born after 1976 with those in the base group

10

(those

born between 1970 and 1976). The graphical results support what we have in table 1.6. A caveat
is due when interpreting these results, as we find a similar trend with bigger magnitudes based on
the Muslim sample (columns 4 and 5 of table 1.6). However, these results provide some evidence
on the impact of the HSAA on educational attainment, given the lack of additional control
groups.
Favorable effects on generation III females‟ educational outcome could be delivered
through two mechanisms. On the one hand, the human capital investment could complement
physical capital transfers if parents changed their attitudes towards daughters. On the other hand,
generation II females‟ growing intrahousehold bargaining could alter household decisions in
favor of daughters. To explore directly which mechanism leads to our finding, we interact
variables in column 1 with an indicator variable for whether or not the maternal grandfather died
after 1994 (in column 3 of table 1.6).

11

The statistically insignificant coefficient on the

interaction, together with the nearly unchanged coefficient on women born between 1980 and
1996 lend support to the first mechanism.

1.5 Conclusion
Interventions for correcting gender discrimination could work through channels of
markets, formal institutions, and informal institutions where different policies may affect
different groups in a society. For formal institutions, changed laws in favor of females can
10

The youngest people in the baseline was 18 years old in 1994, who should be old enough even
considering grade repetition and delayed school entry.
11
The sample size shrinks by about 6%, because for generation III individuals whose mother
already died at the survey time, we have no information on their maternal grandfathers‟ death
status.
28

redistribute resources and power between males and females. The rigorous evaluation turns to be
crucial for recommendation and generalization of a particular policy. The Karnataka amendment
to the inheritance law provides us a natural experiment to examine whether or not its objective
has been achieved, and provide experiences for the entire nation as well as for other developing
countries where females remain disadvantaged in inheritance rights.
A unique dataset containing comprehensive information on three generations of
individuals allows us to explore the mechanism through which parents reallocate physical and
human capital among children following the amendment in inheritance legislation. Regarding
physical capital transfers, three results stand out. First, the HSAA increased the share of assets
received by females who were single before the reform if their fathers passed away after the
amendment. Second, fathers willed away separate property from daughters who were single
before the reform in the long run. Third, females who married after the reform received a higher
share of gifts. While the HSAA eliminated gender inequality for those who were single before
the amendment, it failed to achieve the objective of eradicating the evils of dowry system. In the
meanwhile, an increase in girls‟ educational attainment after the HSAA suggests that, although a
substitution occurred between joint family property and separate property, the amendment has
reinforced rather than undermined human capital investment on daughters, leading to
improvement in women‟s future economic and social status.
As a determinant instead of a simple symptom of gender inequality, property rights
affected by the HSAA and targeting the root cause of the problem could affect other strong
biases against women out of our current study. Physical capital could enable women to access
credit and other inputs, thus facilitating women to achieve the same productivity as men in
agriculture and entrepreneurship. Human capital could allow women to break occupational

29

segregation and to close the gender gap in labor income. Further research could identify indirect
impacts of the HSAA on gender inequality in multiple domains through women‟s strengthened
property rights. In addition, while Roy (2008) finds the positive effect of the HSAA on women‟s
autonomy based on three subjective measures, this paper does not attribute girls‟ increasing
educational attainment to mothers‟ growing bargaining power granted by the HSAA. This
highlights the need to use objective measures of bargaining power including consumption
composition, children‟s nutrition and health, and women‟s time allocation between labor market
and household work in further research.

30

Table 1.1 Received assets of generation II
Total
Sample
Panel A: Hindu
Men
Total assets received from parents (rupees)
Share of assets received from parents among GII
Share of received assets*number of GII members
No. of observations
Women
Total assets received from parents (rupees)
Share of assets received from parents among GII
Share of received assets*number of GII members
No. of observations
Panel B: Weighted Muslim
Men
Total assets received from parents (rupees)
Share of assets received from parents among GII
Share of received assets*number of GII members
No. of observations
Women
Total assets received from parents (rupees)
Share of assets received from parents among GII
Share of received assets*number of GII members
No. of observations

Panel C: Hindu
Men
Total assets received from parents (rupees)
Share of assets received from parents among GII
Share of received assets*number of GII members
No. of observations
Women
Total assets received from parents (rupees)
Share of assets received from parents among GII
Share of received assets*number of GII members
Total assets received from parents (rupees)
No. of observations
Panel D: Weighted Muslim
Men
Share of assets received from parents among GII
Share of received assets*number of GII members

31

Father
alive

Father died
Before
After
1994
1994

208,676
0.28
1.02
13,136

111,298
0.23
0.89
2,905

241,231
0.30
1.07
5,681

230,200
0.27
1.04
4,550

48,660
0.26
0.95
10,643

43,892
0.29
1.10
2,671

51,780
0.25
0.87
4,261

48,509
0.24
0.92
3,711

144,045
0.28
0.97
3,220

84,623
0.22
0.89
795

151,509
0.33
0.99
1,234

168,691
0.26
0.99
1,191

44,496
0.29
1.00
2,661
Total
Sample

39,909
0.29
1.15
727
Mother
alive

50,076
40,704
0.30
0.26
0.98
0.94
953
981
Mother died
Before
After
1994
1994

208,676
0.28
1.02
13,136

117,017
0.24
0.91
5,831

262,760
0.33
1.08
3,256

297,183
0.29
1.13
4,049

48,660
0.26
0.95
144,045
10,643

43,738
0.28
1.08
79,752
5,081

52,951
0.26
0.86
203,158
2,463

53,320
0.21
0.80
18,291
3,099

0.28
0.97

0.22
0.89

0.36
0.99

0.30
1.05

Table 1.1 (cont‟d)
Total
Sample
No. of observations
Women
Total assets received from parents (rupees)
Share of assets received from parents among GII
Share of received assets*number of GII members
No. of observations

32

Mother
alive

3,220

1,628

Mother died
Before
After
1994
1994
648
944

44,496
0.29
1.00
2,661

38,595
0.27
1.09
1,379

58,022
0.34
0.98
501

41,066
0.26
0.90
781

Table 1.2 Educational attainment of generation III
Total
Sample

Father
alive

Father
died

12.05
0.17
11.99
0.20
11.28
0.19
3,510

12.38
0.15
12.30
0.16
11.50
0.16
2,916

11.04
0.20
10.56
0.34
9.38
0.42
594

12.16
0.16
12.01
0.20
11.28
0.19
3,295

11.01
0.24
11.68
0.20
11.34
0.19
215

10.29
0.29
11.82
0.20
11.23
0.17
2,612

10.68
0.26
12.26
0.17
11.38
0.15
2,186

9.13
0.39
9.88
0.34
9.96
0.34
426

10.31
0.29
11.82
0.20
11.24
0.17
2,452

10.13
0.28
11.81
0.28
10.90
0.16
160

8.48
0.46
10.11
0.34
10.14
0.37
1,037

9.48
0.41
10.41
0.30
10.29
0.35
875

6.53
0.55
8.80
0.50
8.82
0.59
162

8.47
0.46
9.98
0.35
10.03
0.39
1,000

10.34
0.56
11.31
0.26
11.78
0.14
37

7.45
0.62
10.19
0.36
10.48
0.28
789

9.28
0.53
10.44
0.35
10.59
0.27
652

4.31
0.76
9.34
0.42
9.63
0.41
137

7.42
0.56
9.95
0.41
10.28
0.31
753

7.64
0.91
12.20
0.02
12.07
0.10
36

Panel A: Hindu
Men
Born 1970-1979: Standard years of schooling
Born 1970-1979: Primary education or below
Born 1980-1988: Standard years of schooling
Born 1980-1988: Primary education or below
Born 1989-1996: Standard years of schooling
Born 1989-1996: Primary education or below
No. of observations
Women
Born 1970-1979: Standard years of schooling
Born 1970-1979: Primary education or below
Born 1980-1988: Standard years of schooling
Born 1980-1988: Primary education or below
Born 1989-1996: Standard years of schooling
Born 1989-1996: Primary education or below
No. of observations
Panel B: Weighted Muslim
Men
Born 1970-1979: Standard years of schooling
Born 1970-1979: Primary education or below
Born 1980-1988: Standard years of schooling
Born 1980-1988: Primary education or below
Born 1989-1996: Standard years of schooling
Born 1989-1996: Primary education or below
No. of observations
Women
Born 1970-1979: Standard years of schooling
Born 1970-1979: Primary education or below
Born 1980-1988: Standard years of schooling
Born 1980-1988: Primary education or below
Born 1989-1996: Standard years of schooling
Born 1989-1996: Primary education or below
No. of observations

33

Mother Mother
alive
died

Table 1.3 Share of assets received from parents if fathers died (generation II)
Hindu
Weighted Muslim
Difference
(1)
(2)
(3)
(4)
(5)
(6)
Female (β1)
-0.301***
-0.330***
-0.069
-0.080
-0.232***
-0.244***
(0.023)
(0.024)
(0.049)
(0.053)
(0.055)
(0.059)
Female*Father‟s death
0.123***
0.045
-0.002
-0.033
0.123
0.085
after 1994 (β2)
(0.034)
(0.037)
(0.070)
(0.074)
(0.078)
(0.083)
Female*Father‟s death after 1994
0.216**
0.103
0.121
*Single before 1994 (β3)
(0.086)
(0.165)
(0.188)
Female*Single before 1994 (β4)
0.115*
0.060
0.018
(0.067)
(0.134)
(0.149)
Father‟s death after 1994*Single
0.076
0.155
-0.098
before 1994 (β5)
(0.059)
(0.128)
(0.140)
Single before 1994 (β6)
-0.266***
-0.124
-0.082
(0.048)
(0.102)
(0.108)
Observations
18,203
18,203
4,359
4,359
22,562
22,562
R-squared
0.099
0.106
0.160
0.164
0.116
0.122
F-Tests:
β1+β2 = 0
100.92***
β1+β2+ β3+β4= 0
0.85
Note: Female is an indicator variable for whether the individual is female. Father‟s death after 1994 is an indicator variable for
whether the father of the child died after the amendment of the act, i.e. after the year 1994. Single before 1994 is an indicator variable
for whether the individual was single before the amendment of the act, i.e. before the year 1994. The last two columns are based on
regressions with the Hindu sample and the weighted Muslim sample. All regressions include household fixed effects and birth year
controls. Robust standard errors to heterogeneity are in brackets. * significant at 10%; ** significant at 5%; *** significant at 1%.

34

Table 1.4 Share of assets received from parents if mothers died (generation II)
Hindu
Weighted Muslim
Difference
(1)
(2)
(3)
(4)
(5)
(6)
Female (β1)
-0.357*** -0.361***
-0.063
-0.070
-0.293*** -0.285***
(0.032)
(0.032)
(0.074)
(0.073)
(0.081)
(0.080)
Female*Mother‟s death after 1994 (β2)
-0.103** -0.148***
-0.160* -0.157*
0.061
0.020
(0.042)
(0.044)
(0.090)
(0.091)
(0.099)
(0.101)
Female*Mother‟s death after 1994
0.209
-0.004
0.206
*Single before 1994 (β3)
(0.136)
(0.334)
(0.367)
Female*Single before 1994 (β4)
-0.005
0.072
-0.100
(0.118)
(0.322)
(0.341)
Mother‟s death after 1994*Single before 1994 (β5)
-0.076
0.117
-0.196
(0.098)
(0.185)
(0.213)
Single before 1994 (β6)
-0.138
-0.001
-0.038
(0.092)
(0.178)
(0.185)
Observations
12,867
12,867
2,874
2,874
15,741
15,741
R-squared
0.133
0.136
0.189
0.191
0.146
0.148
F-test:
β2 + β 3 = 0
0.21
Note: Female is an indicator variable for whether the individual is female. Mother‟s death after 1994 is an indicator variable for
whether the mother of the child died after the amendment of the act, i.e. after the year 1994. Single before 1994 is an indicator
variable for whether the individual was single before the amendment of the act, i.e. before the year 1994. The last two columns are
based on regressions with the Hindu sample and the weighted Muslim sample. All regressions include household fixed effects and
birth year controls. Robust standard errors to heterogeneity are in brackets. * significant at 10%; ** significant at 5%; *** significant
at 1%.

35

Table 1.5 Share of assets received from parents if fathers and mothers are alive (generation II)
Hindu
Weighted Muslim
Difference
(1)
(2)
(3)
Female (γ1)
0.020
0.086
-0.082
(0.045)
(0.086)
(0.086)
Female*Married after 1994 (γ2)
0.147***
0.099
0.074
(0.056)
(0.106)
(0.112)
Married after 1994 (γ3)
-0.003
0.072
-0.061
(0.058)
(0.115)
(0.101)
Observations
4,042
1,036
5,078
R-squared
0.412
0.479
0.438
F-tests:
γ1+ γ2=0
20.09***
γ2+ γ3=0
9.37***
Note: Female is an indicator variable for whether the individual is female. Married after 1994 is
an indicator variable for whether the child got married after the amendment of the act, i.e. after
the year 1994. The two column is based on regressions with the Hindu sample and the weighted
Muslim sample. All regressions include household fixed effects and birth year controls. Robust
standard errors to heterogeneity are in brackets. * significant at 10%; ** significant at 5%; ***
significant at 1%.

36

Table 1.6 Primary educational attainment (generation III)
Female (β1)
Female*Born between 1980-1996 (β2)
Female*Born between 1980-1988 (β21)
Female*Born between 1989-1996 (β22)
Born between 1980-1996 (β3)
Born between 1980-1988 (β31)
Born between 1989-1996 (β32)

Hindu
Weighted Muslim
(1)
(2)
(3)
(4)
(5)
(6)
-0.860*** -0.859*** -0.821*** -0.316 -0.321 -0.126
(0.158)
(0.158)
(0.206) (0.471) (0.473) (0.438)
0.594***
0.512** 0.809*
0.583
(0.167)
(0.227) (0.484)
(0.594)
0.477***
0.729
(0.176)
(0.538)
0.701***
0.860*
(0.178)
(0.486)
0.063
0.064
0.113
0.450
(0.121)
(0.171) (0.428)
(0.531)
0.058
0.117
(0.123)
(0.433)
0.294**
0.378
(0.149)
(0.422)
0.158
-1.105
(0.320)
(1.592)
-0.068
1.055
(0.343)
(1.616)
0.098
-1.035
(0.239)
(0.868)
6,122
6,122
5,747
1,826 1,826 1,753
0.657
0.660
0.651
0.797 0.798 0.800

Difference
(7)
(8)
(9)
-0.544 -0.538 -0.695
(0.527) (0.528) (0.510)
-0.216
-0.071
(0.542)
(0.674)
-0.252
(0.600)
-0.158
(0.548)
-0.050
-0.387
(0.473)
(0.592)
-0.059
(0.478)
-0.085
(0.474)
1.263
(1.733)
-1.124
(1.762)
1.134
(0.958)
7,948 7,948 7,500
0.762 0.764 0.766

Female*Maternal grandfather‟s death
after 1994 (β4)
Female*Born between 1980-1996*Maternal
grandfather‟s death after 1994 (β5)
Born between 1980-1996*Maternal
grandfather‟s death after 1994 (β6)
Observations
R-squared
F-tests:
β1+β2=0
19.45***
β2+β3=0
15.87***
β21=β22
3.56*
Note: Female is an indicator variable for whether the individual is female. The baseline category includes children born between 19701979 (15-24 years old in 1994). Born between 1980-1988 includes children who were 6-14 years old in 1994. Born between 19891996 includes children who were 0-5 years old in 1994 and who were born in the next two years of 1994. The youngest cohort (born
in 1996) was 15 years old in the surveyed year, 2011. Since primary school ages in India are 6-14 years old, all children in the sample
37

Table 1.6 (cont‟d)
should have completed study in primary schools. Maternal grandfather‟s death after 1994 is an indicator variable for whether or not
children‟s maternal grandfather died after 1994. Education years are truncated at grade nine to look at decisions on primary education.
All regressions household fixed effects. Robust standard errors to heterogeneity are in brackets. * significant at 10%; ** significant at
5%; *** significant at 1%.

38

1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011

Coefficients

1
.8
.6
.4
.2
0
-.2
-.4
-.6
-.8
-1
-1.2
-1.4
-1.6

Fathers' death years
Married before 1994

Single before 1994

Figure 1.1 Coefficients on Female*Father‟s death year and Female*Father‟s death year*Single
before 1994 for the share of assets received from parents by Hindus (generation II)
(For interpretation of the references to color in this and all other figures, the reader is referred to
the electronic version of this dissertation)

39

1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011

Coefficients

1
.8
.6
.4
.2
0
-.2
-.4
-.6
-.8
-1
-1.2
-1.4
-1.6

Mothers' death years
Married before 1994

Single before 1994

Figure 1.2 Coefficients on Female*Father‟s death year and Female*Father‟s death year*Single
before 1994 for the share of assets received from parents by Hindus (generation II)
(For interpretation of the references to color in this and all other figures, the reader is referred to
the electronic version of this dissertation)

40

.5
.4
.3

.1
0
-.1
-.2
-.3
-.4
-.5

1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011

Coefficients

.2

Marriage years
Figure 1.3 Coefficients on Female*Married years for the share of gifts received from parents by
Hindus (generation II)
(For interpretation of the references to color in this and all other figures, the reader is referred to
the electronic version of this dissertation)

41

1996

1995

1994

1993

1992

1991

1990

1989

1988

1987

1986

1985

1984

1983

1982

1981

1980

1979

1978

1977

Coefficients

1
.9
.8
.7
.6
.5
.4
.3
.2
.1
0
-.1
-.2
-.3
-.4

Birth years
Figure 1.4 Coefficients on Female*Birth year for primary educational attainment for Hindus
(generation III)
(For interpretation of the references to color in this and all other figures, the reader is referred to
the electronic version of this dissertation)

42

APPENDIX

43

Table 1.A1 Descriptive statistics of matched individuals
Hindu Muslim

Weighted
Muslim
(3)

T-test for equality

(1)
(2)
(1)=(2)
(2)=(3)
Panel A: Generation II
Share of female
0.45
0.45
0.45
0.68
0.50
Year of birth
1962
1966
1961
18.82*** -2.98***
Standard years of schooling
7.96
6.22
7.64
-22.52*** -2.49**
Married in 1994
0.72
0.65
0.74
-11.32***
2.43**
Unmarried in 1994, married in 2011
0.22
0.27
0.21
7.74***
-1.97**
Unmarried in 2011
0.06
0.08
0.05
7.89***
-1.35
Income = < 5000 per month (Rs)
0.16
0.15
0.19
-1.72*
3.03***
Income = 5000 – 9000 per month (Rs)
0.21
0.27
0.21
9.48***
-0.80
Income = > 9000 per month (Rs)
0.19
0.12
0.17
-12.38*** -2.37**
No income
0.44
0.46
0.43
2.87***
-0.18
Father died before 1994
0.42
0.37
0.44
-6.46***
1.96**
Father died after 1994
0.35
0.37
0.35
3.15***
0.46
Father was alive in 2011
0.23
0.26
0.21
3.91***
-3.32***
Mother died before 1994
0.24
0.20
0.27
-7.35***
2.54**
Mother died after 1994
0.30
0.29
0.30
-1.09
0.16
Mother was alive in 2011
0.46
0.51
0.43
7.22***
-2.62***
Number of male GII members
2.53
2.83
2.44
13.29*** -3.08***
Number of female GII members
2.15
2.46
2.10
13.22***
-1.78*
Number of children
2.01
2.30
2.00
16.66***
-0.18
No. of observations
23,779
5,881
5,881
Panel B: Generation III
Share of female
0.43
0.43
0.46
0.41
1.34
Year of birth
1985
1987
1984
7.83***
-2.56**
Married in 1994
0.04
0.03
0.05
-1.31
1.29
Unmarried in 1994, married in 2011
0.36
0.36
0.39
0.09
1.06
Unmarried in 2011
0.60
0.61
0.56
0.40
-1.73*
Income = < 5000 per month (Rs)
0.11
0.12
0.11
1.41
0.15
Income = 5000 – 9000 per month (Rs)
0.20
0.24
0.20
4.06
0.45
Income = > 9000 per month (Rs)
0.14
0.08
0.14
-6.70***
-0.29
No income
0.55
0.56
0.55
0.35
-0.23
Father died before 2011
0.17
0.16
0.21
-0.29
1.98**
Mother died before 2011
0.06
0.04
0.08
-3.46***
1.57
Number of male GIII members
1.82
2.79
1.69
28.78*** -2.69***
Number of female GIII members
1.56
2.22
1.57
17.44***
0.41
No. of observations
6,122
1,826
1,826
Note: Within-city matching. * significant at 10%; ** significant at 5%; *** significant at 1%.

44

REFERENCES

45

REFERENCES

Agarwal, B. 1994. A field of one‟s own: Gender and land rights in South Asia. Cambridge
University Press.
Beaman, L., E. Duflo, R. Pande, and P. Topalova. 2010. “Political reservation and substantive
representation: Evidence from Indian village councils.” Paper presented at the 2010 India
Policy Forum, New Delhi.
Becker, G. S. and N. Tomes. 1979. “An equilibrium theory of the distribution of income and
intergenerational mobility.” Journal of Political Economy 87(6): 1153-1189.
Behrman, J. R., R. A. Pollak and P. Taubman. 1982. “Parental preferences and provision for
pogeny.” Journal of Political Economy 90(1): 52-73.
Behrman, J. R., M. R. Rosenzweig and P. Taubman. 1994. “Endowments and the allocation of
schooling in the family and in the marriage market: The twin experiment.” Journal of
Political Economy 102(6): 1131-1174.
Bernheim, B., A. Shleifer and L. H. Summers. 1986. “Bequests as a means of payment.” Journal
of Political Economy 93(6): 1045-1076.
Blinder, A. 1973. “A model of inherited wealth.” Quarterly Journal of Economics 87(4): 608626.
Brien, M. J. and L. A. Lillard. 1994. “Education, marriage, and first conception in Malaysia.”
Journal of Human Resources 29(4): 1167-1204.
Davies, J. B. 1982. “The relative impact of inheritance and other factors on economic inequality.”
Quarterly Journal of Economics 9(3): 471-498
Deininger, K. and R. Castagnini. 2006. “Incidence and impact of land conflict in Uganda.”
Journal of Economic Behavior and Organization 60(3): 321-345.
Deininger, K., A. Goyal and H. Nagarajan. 2010. “Inheritance law reform and women‟s access to
capital: Evidence from India‟s Hindu Succession Act.” Policy Research Working Paper
no. 5338, World Bank, Washington, DC.
De Nardi, M. (2004), “Wealth inequality and intergenerational links.” Review of Economic
Studies 71(3): 743-768.
FAO (Food and Agriculture Organization). 2011. The state of food and agriculture 201011.Women in Agriculture: Closing the Gender Gap for Development. FAO, Rome.
Goldstein, M. and C. Udry. 2008. “The profits of power: Land rights and agricultural investment
in Ghana.” Journal of Political Economy 116(6): 980-1022.

46

Hoddinott, J. and L. Haddad. 1995. “Does female income share influence household
expenditures? Evidence from Cote D‟Ivoire.” Oxford Bulletin of Economics and
Statistics 57(1): 77-96.
Kotlikoff, L. and L. H. Summers. 1981. “The role of intergenerational transfers in aggregate
capital accumulation.” NBER Working Paper 0445.
Nichols, A. 2007. “Causal inference with observational data.” Stata Journal 7(4): 507-541.
Nichols, A. 2008. “Erratum and discussion of propensity score reweighting." Stata Journal 8(4):
532-539.
Quisumbing, A. R. 1994. “Gender differences in traditional inheritance customs.” Journal of
Development Economics 43(2): 167-195.
Quisumbing, A. R., Payongayong, E., Aidoo, J. B. and Otsuka, K. 2001. “Women‟s land rights
in the transition to individualized ownership: Implications for the management of tree
resources in Western Ghana.” Economic Development and Cultural Change 50(1): 157182.
Quisumbing, A. R., Estudillo, J. P. and Otsuka, K. 2004. Land and schooling: Transferring
wealth across generations. Johns Hopkins University Press for the International Food
Policy Research Institute.
Roy, S. 2009. “Female empowerment through inheritance rights: Evidence from India.” LSE
Working Paper.
Sheshinski, E. and Y. Weiss. 1982. “Inequality sithin and between families.” Journal of Political
Economy 90(1): 105-127.
Stevenson, B. and J. Wolfers. 2006. “Bargaining in the shadow of the law: Divorce laws and
family distress.” Quarterly Journal of Economics 121(1): 267-288.
Thomas, D. 1990. “Intra-household resource allocation: An inferential approach.” The Journal of
Human Resource 25(4): 635-664.
WHO, UNICEF (United Nations Children Fund), UNFPA (United Nations Population Fund),
and World Bank. 2010. Trends in maternal mortality: 1990-2008. WHO, Washington,
D.C.
World Bank. 2001. Engendering development: Through gender equality in rights, resources, and
voice. Oxford University Press.
World Bank. 2011. World Development Report 2012: Gender equality and development. World
Bank, Washington, D.C.
United Nations. 2010. The millennium development goals report 2011. United Nations, New
York, N.Y.

47

Zhang, J. and W. Chan. 1999. “Dowry and wife‟s welfare: A theoretical and empirical analysis.”
Journal of Political Economy 107(4): 786-808.

48

CHAPTER 2: DOES POLITICAL RESERVATION EMPOWER WOMEN OUTSIDE
AND WITHIN HOUSEHOLDS? EVIDENCE FROM RURAL INDIA

2.1 Introduction
Many developing countries suffer from long-standing bias against women in terms of
asset ownership and access to public goods and services. Redressing such bias is important not
only for ethical reasons but also because social exclusion can reduce overall welfare and threaten
social stability. Lack of voice that prevents women from ensuring that their needs are taken into
account in public service delivery and policy design is an important factor reinforcing such bias.
Quotas that aim to increase women‟s political power by reserving a share of the seats contested
in any election for females have therefore attained great popularity. In fact some 100 countries
worldwide are reported to practice gender quotas in some form to overcome gender bias and
long-standing inequalities, such as Rwanda and France.
Despite their political attractiveness and widespread adoption, electoral quotas are
controversial. Supporters argue that empowering members of groups who had historically been
disadvantaged can result in more inclusive processes of policy-making drawing in those
previously excluded. This can change the median voter‟s preferences and thus the outcomes from
political decision-making. To the extent that it improves access to public services (e.g. education
or roads) by those who had earlier been excluded or marginalized, this can be Pareto improving
and ensure better development and use of a society‟s human potential. Critics note that such
measures run a danger of bringing to office individuals who lack necessary qualifications and
may then be easily manipulated by traditional elites. In addition to adverse effects when they are
in force, reservation may also adversely affect behavior by competitively elected leaders who,
because they are prevented from standing for re-election, will have their time horizon truncated

49

and thus be tempted to adopt myopic patterns of behavior that do not promote long-term social
welfare.
Which side of this debate is correct is an empirical question. However, while a large and
increasing number of studies explore the topic, virtually all of them aggregate effects at the
household level rather than allowing for heterogeneity within the household and focus only on
contemporaneous effects during the immediate reservation period from a few states clustered in a
similar geographic region. For a variety of reasons, these effects may be very different from
longer-term national changes in behavior, the change of which is the ultimate goal of the
reservation policy. To assess the extent to which such policies achieve their objective, it is
important to complement these studies with evidence on the extent to which the effects of
reservation persist over time and space. Such information could also help in the understanding of
the channels through which such effects materialize, which is an issue of great policy interest.
India is of particular interest in this respect because, in an effort to overcome longstanding discrimination by gender and caste, far-reaching political decentralization in the early
1990s was combined with regulations mandating that a share of elected leadership positions be
“reserved” for women and other disadvantaged groups. As villages subject to reservation are
chosen randomly in any given period, differences between reserved and non-reserved locations
can be interpreted causally. We draw on a large national sample to assess impacts of female
reservation on the effectiveness of public service delivery, females‟ ability to participate and
voice concerns in political processes, and female empowerment in terms of reproductive choice
and independent management of financial resources. As our data were collected some 15 years
after this policy had taken effect, yet still with about 43% of villages never having been reserved,
we can examine whether reservation has affected outcomes over the long run and, if so, how.

50

Beyond academic interest in such results, recent debates in India about extending female
reservation to the state and national levels imply that such evidence can also inform the country‟s
ongoing policy debate.
The paper is structured as follows. Section 2 provides context by discussing the overall
rationale for gender quotas and the Indian context. Section 3 describes the theoretical argument
and the hypotheses guiding our investigation. Section 4 discusses the data and the empirical
strategy. Section 5 econometrically explores potential gender-differentiated effects of reservation
as well as their persistence over time and space. Section 6 concludes with a set of policy
implications and suggestions for future research.

2.2 Gender quotas: Rationale, evidence, and approach
Concern about a failure of pre-existing and often deep-rooted gender gaps to narrow over
time led many countries to adopt policies that reserve a certain share of political positions for
females. Evidence on the impact of such policies is mixed and inconclusive. In India, high levels
of gender discrimination provided the justification for legislation that reserved, in any election
after 1995, leadership for females in one third of villages, randomly selected. While some note
that leaders coming to power in this way may be less qualified and more prone to being
influenced by powerful people than those elected on unreserved seats, others find that these
policies had significant impact on quality of public goods.

2.2.1 Female political participation in perspective
Women‟s preferences have been shown to differ from those of males (Edlund and Pande
2002) with potentially far-reaching effects on intra-household decision making and democratic

51

voting outcomes (Funk and Gathman 2008). Although standard political economy models with
full pre-commitment imply that policy makers‟ attributes (including gender) will not matter for
outcomes (Downs 1957), these have often been found to be inconsistent with empirical evidence.
Citizen-candidates models formalize the idea that legislators‟ identity can matter for outcomes
(Besley and Coate 1997; Osborne and Slivinski 1996).
Many societies are characterized by under-representation of females in bodies for policy
making at the national or local level (Dahlerup 2006).

12

If such lack of representation is due to

institutional barriers, cultural norms, or political discrimination and if policy makers‟ attributes
matter, actions to increase participation share could affect the nature of political equilibria. This
line of argument provided a justification for affirmative action, including quotas to reserve a
share of political positions for females, to bring about greater equality in opportunity and
outcomes. In some cases, such measures have led to perceptible shifts in the composition of
legislatures. For example, in Rwanda, establishment of a gender quota (of 30%) in 2003 is
widely credited with having contributed to the fact that female parliamentarians now make up the
majority of the legislative assembly (Powley 2007). Quotas have also been introduced in many
European countries, although in some cases their impact may be diminished by applying to
candidates rather than to seats.

13

Across countries, studies suggest that female participation in legislative processes helped
to overcome gender bias in access to specific services and that female legislators tended to
allocate more funds to causes important to women (Paxton and Hughes 2007; Reingold 1992;
12

Globally, less than 20% of members in national parliaments are female, although this varies
from close to parity in Nordic and some African countries such as Rwanda to some 10 percent in
Arab states (International Institute for Democracy and Electoral Assistance (IDEA) 2008).
13
For example in France, although 50% of candidates are required to be women, only 18%
make it to the national assembly (Frechette et al. 2008).
52

Saint-Germain 1989). Effective provision of greater volumes of public goods could then help
previously disadvantaged groups to more quickly eliminate such biases while greater political
participation increases the pool of talent upon which the system can draw or allow greater
deliberation and innovation to eliminate mistakes in the process (Page 2007). Studies suggest
that higher levels of female representation in parliament are associated with lower levels of
corruption (Dollar et al. 2001) as well as more spending per student, higher shares of female
teachers, and greater secondary enrollment by girls. Greater female representation is also found
to improve women‟s access to health care via higher public spending that in turn is noted to
increase the number of doctors and women‟s ability to receive prenatal care (Knack and Sanyal
2000).
Studies exploiting differences in the timing of the introduction of female voting across
U.S. states suggest that expansion of suffrage to women led to increased public spending and
more liberal voting patterns, a phenomenon that persisted over time as females gradually took
advantage of the franchise (Lott and Kenny 1999). Laws extending voting to females were
followed by shifts in legislative behavior and large increases in local health spending (Miller
2008). At the same time, the impact of gender reservation is context specific. Evidence of higher
female participation increasing health spending (Rehavi 2008) contrasts to findings from
mayoral elections during the 2000s where policy makers‟ gender did not appear to have any
effect on the size of government, the composition of municipal spending and employment, or
actual outcomes, such as crime rates (Ferreira and Gyourko 2010). One possible explanation is
that city governments come closer to the Downsian model.

14

In California school boards where

females already hold 45% of elected seats, an anti-incumbent bias was observed, whereby adding
14

Females are found to have higher levels of unobserved political skill, possibly because they
had to overcome gender bias to move up in politics.
53

an additional female increased the likelihood of a male winning the next election and vice versa
(Schwarz 2010).

2.2.2 Gender reservations in the Indian context
India is characterized by large gender gaps in key human development indicators, such as
child mortality and malnutrition, prenatal coverage, contraceptive use, adolescent fertility, and
maternal mortality, which are much worse than what is found in countries with similar or even
lower levels of per capita income.

15

Discrimination against females originates in marriage

practices and caste structures that show little sign of disappearing and may even be reinforced
over time (Anderson 2003). Despite high overall growth since the mid-1990s, some studies
suggest that gender gaps widened in southern states that enjoyed higher levels of income growth
and higher initial levels of female empowerment (e.g. Andhra Pradesh, Karnataka, Tamil Nadu),
suggesting that economic growth alone may not bring about gender equality (Raabe et al. 2009).
Persistently high levels of gender inequality are often believed to be reinforced by deficient
provision of goods that are publicly provided, e.g. maternal health and prenatal care for girl
children (Bhalotra and Rawlings 2011). To the extent that they can bring about better and more
equitable provision of such services, quotas to bring women into positions of power could help to
address and eventually overcome such bias.
rd

In 1992, India‟s 73 Constitutional Amendment mandated far-reaching decentralization
by establishing a three-tier system of district, block, and village-level councils. The gram
panchayat (GP) is the lowest tier of local government at village level. It comprises a president

15

UNDP‟s gender inequality index, based on 2008 data, ranks India 122 out of 138 countries,
below Rwanda (83), Lao PDR (88), Egypt (108), Moldova (38), and China (40).
54

(pradhan or sarpanch) and council members who are elected from the panchayat‟s wards. Its
responsibilities include (i) provision of major public services, such as health care, education,
drinking water, and roads; (ii) setting rates and administering local taxes; (iii) administration,
formulation, and implementation of local development plans; and (iv) selection of beneficiaries
and implementation of social and economic programs established and paid for by the central
government. Regular assemblies (gram sabhas) by all voters in the GP are meant to monitor
performance and increase democratic accountability. To prevent decentralization from
reinforcing the power wielded by traditional elites and to counter what was perceived as a legacy
of disenfanchisement and under-representation by females and other disadvantaged groups,
decentralization was combined with reservation of a share of seats for women as well as
scheduled castes and tribes.

16

A number of influential studies have found that India‟s efforts to increase female
participation in political processes and decision-making had significant impacts. An India-wide
study notes that reservation creates opportunities that many women are able to utilize (Ministry
of Panchayati Raj 2008). In West Bengal and Rajasthan, female policy makers who have come to
power through quotas provide more public goods that benefit and are valued by female voters,
such as water and roads (Chattopadhyay and Duflo 2004). Mandated changes in female
leadership prompted by quotas were also found to increase the quality of political processes and
prompt greater female participation in gram sabha meetings in South India (Besley et al. 2005).
In terms of outcomes, reservation was associated with higher levels of child survival, an effect
16

At village level, most states now reserve one third of council members and pradhan positions
for women. By comparison, the share of positions reserved to ST/SCs equals the population
share of the ST/SC population (Bardhan et al. 2010). Although many studies explored impacts of
caste reservations (Krishnan 2007, Pande 2003), our focus is on those for females. At district
level, it appears that reservation of positions in the legislature for scheduled castes but not tribes
improves access to education facilities, mainly primary schools, for relevant constituencies.
55

that could arise because female leaders, who are more attuned to the needs of child health, helped
to improve access to and use of services such as prenatal care and public birth facilities. This in
turn translates into significantly higher levels of breastfeeding and immunization (Bhalotra and
Clots-Figueras 2010). The impacts of reservation can be felt in other spheres as well. For
example, based on phased introduction of reservation at the state level, it has been argued that
female reservation gave women greater voice, resulting in increased reporting of crimes against
women and greater resistance to violence (Iyer et al. 2010). Reserved seats occupied by lowcaste or tribal (but not high-caste) females are also argued to have resulted in higher levels of
investment in health and early education and greater efforts to implement redistributive land
reforms and inheritance legislation favorable to women (Clots-Figueras 2009).
Studies that find significant and positive impacts of reservation are in contrast to others
arguing that reservation-induced effects may be more ambiguous or even negative. One
possibility may be that, once other factors are controlled for or results are placed in a broader
context, presumed advantages enjoyed by female pradhans‟ vanish or even turn negative so that
their performance would be below that of pradhans elected without restrictions (Ban and Rao
2008; Besley et al. 2004; Rajaraman and Gupta 2008). In some cases, aggregation may be an
issue as effects may be limited to some sub-groups but fail to materialize in the aggregate
(Dongre 2010). A study focusing on selection of beneficiaries from central government schemes
in West Bengal suggests that, in this respect, reservation had few, if any, positive effects on
women. It worsened within-village targeting of such transfers to lower caste groups and failed to
improve other targeting dimensions. This implies that, for disadvantaged women, the net effect
of reservation could, for a number of reasons, actually end up being negative (Bardhan et al.

56

17

2010).

An approach based on individual data and clear hypotheses on how reservation may

affect outcomes can thus help better understand the underlying processes.
Specific studies have also noted that reservation could in principle result in the selection
of leaders with inferior characteristics and, consequently, lower the levels or quality of public
goods, particularly in the case of caste-based reservation (Munshi and Rosenzweig 2008). While
such effects should be less pronounced for gender reservation, in Tamil Nadu women pradhans
on reserved seats are reported to have fared very badly –much worse than SCs (scheduled castes)
and STs (scheduled tribes)– in a test designed to measure their understanding of relevant gram
panchayat procedures (Gajwani and Zhang 2008). Attributing outcomes solely to the leader‟s
gender is also equivalent to assuming full pradhan dominance of the gram panchayat. This may
be at variance with the reality that women occupying reserved seats are often poorly educated
and may be guided by traditional elites (often husbands or family members) who may pull the
strings from behind the scenes (Rajaraman and Gupta 2008). There is indeed evidence
suggesting that female leaders depend on access to social and political networks through
traditional leaders and that this reduces their ability to control events.

18

Finally, as reservation will be in place for a limited time, their impact can be fully
appreciated only by looking beyond the reservation period itself. If members from disadvantaged
groups who entered office through reservation are not accountable and direct public goods or
programs towards cronies or a narrow constituency, effects observed during the reserved period
17

The explanation favored by the study under concern is that female reservation is inconsistent
with traditional models of electoral competition. It concludes that a more complex approach of
capture-cum-clientelism is more appropriate but may itself be is susceptible to being weakened
by election of inexperienced women to reserved positions (Bardhan et al. 2010).
18
In the Tamil Nadu study quoted earlier, some 80% indicated that their official decisions were
to be influenced by their husbands, potentially pointing towards limited autonomy in decisionmaking (Gajwani and Zhang 2008).
57

may be reversed once reservation has expired. This hypothesis receives support from the
ambiguous longer-term effects of reservation on access to and quality of public goods found in
longer-term studies that consider outcomes for more than just one period (Bardhan et al. 2010;
Raabe et al. 2009). On the other hand, long-term effects of reservation can be more positive than
what appears in the short run if the reservation helps to change either the way in which public
goods are provided or the pattern of political participation and the associated political
equilibrium. If the experience of female leaders prompts voters to revise long-standing
prejudices that affect future voting behavior (Beaman et al. 2009), long-term effects of quotas
could be more positive than what emerges in the short term. The same would hold if reservation
allows those who were excluded or lacked voice to more effectively participate in political
decision-making and, through such a shift in political participation, alter the nature of the median
voter. In fact, the desire to prevent local capture in this way was one of the factors motivating
adoption of this policy (Singh 2007).
Beyond the reservation period itself, the literature has identified two channels through
which reservation could affect long-term outcomes. First, it might prompt those who previously
had not participated in political process to change their behavior permanently. Evidence for
persistent effects via greater participation along these lines is available from rural West Bengal
(Beaman et al. 2010) and South India (Besley et al. 2005) as well as urban Mumbai (Bhavnani
2009). This is plausible as previously ignorant voters may require time to learn about how to
access and use information to most effectively hold leaders accountable. Second, it may trigger a
process of learning and revision of prejudices as in cases where exposure to female leaders led to
the revision of stereotypes regarding females‟ leadership qualities (Beaman et al. 2009).

58

2.3 Conceptual framework and hypotheses
We use the framework developed in Chattopadhyay and Duflo (2004), where the final
policy implemented is a mixture of the preferred policy option of the elected official

, and the

policy option μ’, preferred by the local elite (or alternatively the elected official is subject to the
control of the village assembly or the elected council). μ’ is assumed to be more “pro-male” than
the median voter‟s preference, since the local elite tend to be male, and men are also more likely
to attend village meetings than women. Women run for office if the utility gain from the
implementation of the mixed policy (net of the cost of running) is bigger than the utility gain
from the implementation of μ’.
Our argument departs from their model in two respects. First, we assume that μ’ itself is a
combination of the two policy options – the policy option μ’1, preferred by villagers who attend
village meetings with a weight , and the policy option μ’2, preferred by the local elite (or the
elected council) with a weight
denoted by

. The policy option preferred by the elected official is again

and given a weight , so the mixed policy finally implemented by the elected

official is x = αω + (1-α)μ’ = αω + (1-α)(βμ’1 + (1-β)μ’2). We assume that pradhan candidates
have the same ability to impose their preferred policies (the same

across candidates). The same

assumption also applies to meeting participants ( is constant irrespective of the fraction of
females among meeting participants). Second, μ’1 could be influenced by the reservation policy
through its impact on the composition of village meeting participants. In particular, if the
reservation policy motivates more females to participate in village meetings and to voice their
opinions, it tends to move μ’1 in a “pro-female” direction.

59

Following Chattopadhyay and Duflo (2004), the final policy implemented will be
unambiguously in line with female preferences, if (i) no women runs for office without a
reservation system, (ii) women run for office after the reservation policy, and (iii) the most “profemale” outcome implemented by a man with no reservation system is more “pro-male” than the
median voter and what the most “pro-male” woman who will run will implement. We assume
that such “pro-female” outcomes will be reinforced if meeting participants have the ability to
move μ’1 in a “pro-female” direction. The “pro-female” outcomes, if any, could persist over
time, as current pradhan‟s investments may influence future political outcomes, and it takes time
for villagers to accumulate political knowledge.
Meeting attendance is an important channel through which villagers influence the final
policy implemented, but not the only avenue for their political participation. For example,
another way to involve in public affairs is to approach local authorities and to complain about
problems associated with local public services. While the reservation policy is likely to increase
the likelihood of females complaining about problems and attending village meetings, we
hypothesize that outcomes through these two channels are likely to be different across states.
One straightforward hypothesis is that they are bigger and more significant in states where
women are traditionally more severely discriminated. The other hypothesis is that reservation
leads to more complaints about problems in states where women are severely discriminated, but
is more effective for mobilizing women to attend village meetings if they face less severe
discrimination. The second hypothesis is based on the assumption that to complain about
problems is a more basic way to involve in public affairs than to voice concerns in village
meetings. Our data suggest that 82% of female respondents and 93% of male respondents
complained about at least one problem in villages that had no reservation, but the percentages of
60

meeting attendant are only 14% and 42%, respectively. Females may have been used to
participating in political processes in the basic way in states where gender discrimination is less
severe, and therefore we may barely observe further improvement. We further hypothesize that
the two reservation-induced effects persist over time, again as it takes time for villagers to
accumulate political knowledge.
Females‟ involvement in political activities taking place in the public sphere could
improve their social status and translate into autonomy in the domestic sphere. This process
might take time and lead to persistent effects. We test this hypothesis by comparing the
reproductive choice and the ability to access financial resources of females in villages with and
without reservation.

2.4 Data sources and estimation strategy
To explore the issues discussed above, we use data from nationally-representative
surveys of 155 villages in rural India conducted in 2000 and 2008 by the National Council of
Applied Economics (NCAER). Following Filmer et al. (1998), we separate twelve states into
two groups by an average ranking of women‟s status, child mortality, female-male population
ratio, female-male ratio of no treatment for fever, female-male ratio of school enrollment, rural
female labor force participation, and human development index. In addition to its national
coverage, this dataset has a number of desirable features. First, it provides information on
reservation status by panchayat for two rounds of elections after the provisions of the 73
amendment were adopted in a state.

19

rd

Table 2.1, which provides a breakdown of reservation by

19

While this includes three elections for most of the states, Bihar, Orissa, Punjab, Tamil Nadu
and West Bengal held the earliest elections in 1992, before the amendment was adopted.
61

region, illustrates that out of the 155 villages in the sample, 43% had never been reserved while
50% and 7% had been reserved once and twice, respectively. Second, data on a range of
performance measures of the current elected panchayat is obtained individually for all household
members 16 and older. We are thus able to not only explore gender-specific effects but also to
assess whether some impacts of reservation materialize only in a few states or in a delayed
fashion.
We use individuals‟ stated willingness to contribute a modest amount (100 Rs.) to 13
different categories of public services if doing so would allow securing a matching government
contribution of much larger size.

20

This allows us to isolate female preferences from male

preferences. To remove the concern that village-level characteristics could lead to common
preferences on some public services, we compare preferences between females and males within
the same village by controlling for village fixed effects. The first panel of table 2.3 illustrates
that, while males and females in group 1 states value schools and education, electrification, and
natural resource management equally, males value roads and transportation, street lighting,
irrigation, government schemes, employment schemes, social issues, and credit more than
females, but females value drinking water, health facilities, and sanitation and sewage more than
males. The results in the second panel of table 2.3 show that females in group 2 are willing to
contribute more to health facilities and sanitation and sewage than males. However, males in
group 2 care more about drinking water, schools and education, and electrification than in group
1.
20

To obtain an indication of individuals‟ willingness to contribute at the margin (for the current
panchayat period), the question was phrased as follows. “Imagine the Government decides to
contribute an additional Rs. 1 lakh (100,000) to solve a local problem but only if the majority of
village households contribute Rs. 100 each. Imagine that almost enough people are willing to
contribute and your decision to contribute or not will decide the outcome. Would you be willing
to contribute Rs. 100 if the issue was ….?”.
62

A reduction in the share of those who experienced problems with public services,
especially women, could indicate better performance by officials elected under reservation. For
example, female leaders under reservation may be more proactive in trying to anticipate
problems or focus on aspects of the goods that are more valued by females than by males.
Instead of using general ratings or assessments of the quality with which public officials provide
services, we know whether individual respondents were affected by problems with supply of
specific public services, based on 11 categories that are defined using survey data. In case they
had identified problems, we asked whether individuals approached competent authority at local
level to get these issues rectified,

21

a variable that can measure voice and political

empowerment.
To avoid reliance on hypothetical and rather abstract information, we complement the
above indicators with information on actual meeting participation during the last four gram
sabhas held in each village. For each meeting, information on attendance and active participation
(by discussing issues rather than just observing) is complemented with the number of issues
discussed at the meeting that respondents considered to be of relevance for themselves.
Finally we draw on the demographic module that was administered to all females
between the age of 14 and 50 in each household, to construct a number of indicators related to
female empowerment and autonomy, in particular their reproductive choice and ability to access
resources through their own bank account.
Methodologically, as reserved seats are assigned randomly, OLS regressions of outcome
variables on a reservation dummy (whether the pradhan position was reserved for women or not)
will yield unbiased and consistent estimates of policy impact. Village-level control variables are
21

We don‟t consider a complaint to be proper if rather than to a relevant local government
authority, it was directed to an informal or religious group or to a higher level (e.g. the MLA).
63

included for improving the precision of estimates. Letting superscript j index outcome and
subscripts, i, v, s, g and t index individuals, villages, states, group and time, the estimating
equation for identified problems of public service delivery and associated complaints is
j

j

j

j

j

j

Y ivsgt = β sg + β 1Rvsgt+ β 2Rvsgt*Fg+ β 3Rvsgt-1+ β 4Rvsgt-1*Fg
j

j

j

j

+ β 5Fgβ 6Xivsgt+β 7Zvsg+ε ivsgt
j

where Y ivsgt is the outcome variable of interest,

22

(1)

βsg is a state fixed effect, Rvsgt and Rvsgt-1 are

indicator variables for reservation in the current or in the previous period that equal one if the
pradhan position in village v of state s belonging to group g at time t or t-1 was reserved for
females, and zero otherwise, Fg is an indicator variables for female goods, Xivsgt is a vector of
household and individual controls such as wealth and land ownership status, age, sex, caste, and
broad occupation, Zvsg is a vector of initial village characteristics listed in table 2.2, and the βs
are parameters to be estimated. The sum of β1 and β2 as well as the sum of β3 and β4 allow us to
test whether current and previous reservation move current outcomes in a “pro-female” direction,
the latter of which allows us to test for persistence of any reservation-induced effects. The sum
of β2 and β3 allow us to test whether effects differ for males and females. We estimate the
equation separately by gender and group to explore gender-differentiated and state-differentiated
effects.
The estimation equation for meeting attendance and female empowerment is
j

j

j

j

j

j

j

Y ivsgt = γ sg + γ 1Rvsgt+ γ 2Rvsgt-1 + γ 3Xivsgt+ γ 4Zvsg+ε ivsgt
22

(2)

To avoid that results be driven by recall bias, all of the regressions are based on outcome
variables for the current panchayat period only even in cases where similar information (based
on recall) was provided for earlier periods.
64

where γsg is a state fixed effect, and other variables are defined similarly as in equations (1). The
γs are parameters to be estimated. The coefficients γ1 and γ2 can be interpreted as the effect of
j

reservation in the current and previous period on current outcomes Y ivsgt , the latter of which
allows us to test persistence of any reservation-induced effects. Equation (2) is also estimated
separately by gender and group.
It is worth noting that a precondition for our empirical strategy to be valid, i.e. for OLS
estimates to be interpreted as a causal effect, is that allocation of reserved seats across villages is
indeed random. The fact that one round of the panel survey was conducted in 1999 allows us to
test this for the 2000 election. Descriptive statistics for village and household characteristics in
1999 in panel 1 and 2 of table 2.2 reveals little if any bias in this regard. Some 29% (45 out of
155) of the sample villages were reserved in this period. Reserved and unreserved villages are
similarly sized and have similar levels of access to infrastructure and public goods (drinking
23

water, anganwadi , food for work programs), with a significant difference only for 2 out of 21
24

village characteristics (net area irrigated under canal or stream and the number of trained dai ).
Household characteristics such as education, age, assets, income, and its composition, are not
significantly different between reserved and unreserved villages either.

23

As part of the Integrated Child Development Services program, India government started
anganwadi to combat child hunger and malnutrition. They also provides basic health care
in Indian villages, including basic contraceptive counseling and supply, nutrition education and
supplementation, as well as pre-school activities.
24
Trained dai is traditional birth attendant.
65

2.5 Econometric results
2.5.1 Problems of public service delivery
Table 2.4 reports the distribution of female pradhans across groups and reservation status.
While villages in group 2 strictly implemented the reservation policy, 4% of reserved villages in
group 1 elected male leaders. Moreover, the fact that in unreserved villages only 8% of pradhans
were female highlights the importance of the reservation policy for achieving gender equality in
the political sphere. Since some females in our sample ran for office without a reservation
system, and males in a few cases won elections after the reservation policy, under assumptions in
section 3, we expect that the estimates based on the whole sample might be different from those
based on the small sample which excludes unreserved villages with female leaders and reserved
villages with male leaders. The results are reported in panel A (based on the whole sample) and
panel B (based on the small sample) of table 2.5, respectively.
The first two columns of table 2.5 show the results of equation (1) for females and males
in all the states, respectively. The results exploring heterogeneous effects across groups are
reported in columns 3-6. The dependent variable is an indicator that equals one if an individual
was affected by a problem and zero otherwise. We report only coefficients for the key variables
of interest (i.e., current reservation dummy, previous reservation dummy, female goods dummy,
and their interactions) with the standard errors clustered at the village level to control for
heteroskedasticity and within village correlation in errors.
The categories of services in table 2.5 are not exactly the same as those in table 2.3,
because they are extracted from different modules of the questionnaire. Here we do not have
electrification, social issues, employment schemes, and natural resource management which
were asked about in the module of villagers‟ willingness to contribute, but we have additional

66

women‟s issues (poor women needing assistance and pregnant women needing assistance) and
ration shop/card

25

that we treat as a category of female goods because women are more likely to

shoulder the responsibility to buy goods like food grains, sugar and kerosene. In the end, for
group 1, we have five female goods (drinking water, health facilities, sanitation and sewage,
women‟s issues and ration shop/card), three of which are indentified by willingness to contribute
reported in table 2.3. For group 2, we have four female goods (health facilities, sanitation and
sewage, women‟s issues and ration shop/card), two of which are identified by willingness to
contribute reported in table 2.3. As drinking water is a female goods for group1 females but not
group 2 females, it undermines our ability to compare coefficients between group 1 and group 2.
We therefore exclude drinking water from the main regressions, and report the results including
drinking water in appendix tables.
We find from the first two columns of panel A that, reservation in the previous period
reduced the incidence of problems associated with female goods by 5% for both males and
females (the sum of β3 and β4). However, heterogeneous effects reported in columns 3-6 suggest
that reserved leaders failed to allocate resources in line with female preferences. First of all,
reservation had no statistically significant impact on public service delivery for group 1 villagers.
Second, group 2 villagers were less likely to be affected by problems associate with both nonfemale and female goods if leaders were reserved for females in the previous period (both β3 and
the sum of β3 and β4 are statistically significant), and reservation in the current period only

25

A public distribution shop, locally known as “ration shop” is used to distribute rations at a
subsidized price to the poor, operated throughout India by joint assistance of central and state
government. For buying items from this shop one must have a ration card.
67

significantly shifted public resources to non-female goods (β1 is statistically significant but not
the sum of β1 and β2).
To explore whether female leaders in unreserved villages and male leaders in reserved
villages influence estimation results in panel A, we report findings based on the small sample in
panel B. If policies implemented by male and female leaders are dramatically different, and they
are in line with the preferences of villagers in the same gender, we would find more “profemale” outcomes using the small sample. In addition to the statistically insignificant coefficients
for group 1 villagers, the fact that group 2 villagers only found fewer problems associates with
non-female goods is consistent with the heterogeneous effects in panel A, suggesting that
reserved female leaders had problems to implement policies of their interest.
Although our survey refers to actual problems rather than just a subjective perception, a
possible concern with using such data is that the assessment could still be affected by systematic
bias. Complementing the above analysis with an assessment of individuals‟ meeting participation
and likelihood to complain to local authorities provides not only a way to address such concerns
but, insofar as they are measures of respondents‟ voice, also to draw substantive conclusions.

2.5.2 Political participation
While there are many avenues for political participation, relevant legislation designed
village assemblies (gram sabhas) as key mechanism to ensure democratic deliberation and
accountability at local level. The extent to which previously disadvantaged individuals, including
women, participated in such meetings thus allows us to test more directly the impact of
reservation on voice and the ability to effectively articulate issues. Moreover, as individuals will
continue to attend meetings only if they find that doing so is worth their while, attendance can
68

also be interpreted as an indicator of the relevance ascribed to these meetings. Using individuallevel data on meeting attendance in the last four meetings in each village and, for those
attending, the nature of participation points, suggests that reservation-induced effects vary
markedly by respondents‟ gender and time. Overall, reserved leadership led to better political
outcomes in meetings for group 1 states than for group 2 states (see table 2.6), mirroring the
hypothesis that the reservation policy is more effective in advanced domains if females face less
severe discrimination.
We find from estimation based on the sample including all the states (columns 1 and 2)
that current reservation increased the likelihood of females participating in discussion by 16%. In
the long run, issues discussed during meetings were more relevant for both males and females
(6% and 16%, respectively), but females were less likely to participate in discussion (by 28%).
Heterogeneous effects (columns 3-6) suggest that the results reported in the first two
columns are mainly driven by group 1 states. We find that reservation led to a 6% increase in
contemporaneous attendance by group 1 females, but the effect did not persist over time. While
the point estimates 0.1 and 0.2 suggest that females perceived issues discussed were more
relevant and made more voices during meetings under the leadership of reserved pradhans, they
learned to shift meeting agendas in the long run (by 20%) but were less likely to speak up during
the meetings. This might be because they were satisfied with the agendas and raised fewer
questions. In the meanwhile, it is worth noting that males also perceived discussed issues were
more relevant in the long run (by 6%), implying that participants, irrespective of gender,
accumulated political knowledge after exposed to the reservation policy. Combined with the
findings reported by table 2.5, the positive impacts induced by current reservation on group 1

69

females suggest that villagers had little ability to impose policies that they prefer in village
meetings relative to the local elite (or the elected council).
For group 2 females, the fact that only 66 out of 13684 respondents attended meetings not
only result in insignificant and quantitatively small effects on meeting attendance, but also make
it impossible to estimate effects on relevant issues discussed and active participation conditional
on attendance. The significantly negative coefficients on „reserved now‟ in the case of meeting
agendas for group 2 males reflects their negative bias towards reserved leaders. We also find that
previous reservation discouraged males to participate in discussion.
Reporting identified problems to local authorities (see table 2.7) is another avenue for
political participation which provides a precondition for any remedial action. The estimated
results reveal striking differences across gender and also across the states, highlighting the
relevance of using individual level data, as asking a „representative‟ respondent or using data at
the household level may conceal subtle variations in gendered perceptions and thus the genderdifferentiated impacts of certain policies.
We notice clear increases in the likelihood of complaints by females in response to both
female and non-female goods in villages whose leadership was either currently or previously
reserved for women in group 2 states (β1, the sum of β1 and β2, β3, and the sum of β3 and β4 are
all statistically significant). In contrast, reservation-induced effects on the likelihood of
complaining about identified problems in the group 1 states are fairly limited in both periods.
The differentiated effects between the group 1 states and the group 2 states are also evidenced in
male regressions even though only previous reservation had impacts with smaller magnitudes.
These results support the hypothesis that reservation is likely to have bigger effects in states
where gender discrimination was historically more severe in basic domains.
70

It is also important to note that for females in group 2 states, the long-term effects are
larger than the contemporaneous effects. Specifically, the likelihood of their complaining about
identified problems in the villages where the leadership was previously reserved for women are
almost doubled than in those currently reserved (18% vs. 7% for non-female goods measured by
β3 and β1, respectively, and 18% vs. 11% for female goods measured by the sum of β3 and β4
and the sum of β1 and β2, respectively). Interestingly and perhaps more importantly, we also
notice that in group 2 states, reservation in the previous period significantly increased the
likelihood males complaining about identified problems.

2.5.3 Female empowerment
Table 2.8 reports results from regressions for reproductive choice and the female having a
bank account in her own name. Current reservation resulted in a significant increase in the use of
condoms by 10% and a reduction of a similar magnitude in the opinion that female sterilization
is the most appropriate form of birth control (columns 2 and 3). These effects persisted beyond
the period during which reservation was in place with quantitatively larger magnitudes (13% for
the use of condoms in group 1 states, and 19% for both cases in group 2 states), except for the
case of sterilization in group 1 states. However, we do not detect clear changes in terms of
having an own bank account, implying a general empowerment effect did not translate into
positive economic outcomes yet.

2.6 Conclusion and policy implications
Our analysis contributes to the debate on the extent to which reserving political positions
for females can improve political outcomes and autonomy within the household. India‟s adoption
71

of far-reaching policies, allocated randomly across villages, to overcome the country‟s high level
of gender bias and social stratification some time ago makes this particularly suitable for our
analysis. A nation-wide sample with a rich set of information from individual respondents allows
us to contribute to the literature substantively.
Methodologically, we note that differentiating across gender and states can be relevant if
reservation-induced effects on different groups are of different magnitudes or of opposite signs,
e.g., we find pronounced, but strongly gender-differentiated, impacts of female reservation on
the likelihood of complaints to public authorities if problems were identified, and remarkably
state-differentiated, impacts of female reservation on attendance and nature of participation in
rd

village meetings. Information on two elections after the 73 Amendment came into force allows
testing for persistence and long-term impacts of such policies with results confirming that in
many cases the full impact of this policy materializes with some lag.
Substantively, this allows us to complement a focus on outcomes in the short term with
attention to longer-term impacts, in particular individuals‟ participation in political processes,
and females‟ ability to make reproductive decisions. As reservation brings historically excluded
females into political processes, it should not come as a surprise that reservation did not move
resources in a “pro-female” direction for improving public service delivery. In our case, and in
line with the importance of learning, such unexpected effects have to be weighed against
significant and in most cases persistent impacts on outcomes such as meeting attendance and
active participation, and complaints to local authorities if problems of public service were
identified. Our analysis thus allows a more judicious portrayal of the trade-offs which imposition
of reservation, like most other policy measures, is likely to involve, suggesting that in some cases
long-term benefits may more than outweigh the short-term cost for female reservation.
72

There are three areas where follow-up research would be desirable. First, similar analysis
to assess impacts of reservation for certain castes or tribes could help inform the current policy
debate. Second, combining the measures for panchayat performance used here with direct
information on access to different types of outcomes likely to be affected by reservation –from
breastfeeding and child health to running for elected office by females– will help improve
understanding of the impacts of this policy, the mechanisms through which they may materialize,
and the distribution of any benefits it may provide. Finally, it is of great interest to know how the
effects of reservation on women empowerment are translated into asset allocation within
households and the overall welfare at both the individual and household level.

73

Table 2.1 Distribution of sample villages across states and reservation status
Gender
Period when
State
Group
No. of
No. of times reserved
disparity
reserved
(rank)
villages Never Once Twice
Now
Previous
Kerala
1
2.57
13
7
6
0
3
3
Tamil Nadu
4.00
Himachal
1
4.29
4
2
2
0
2
0
Pradesh
Maharashtra
1
4.86
12
6
5
1
4
3
Karnataka
1
5.43
17
5
8
4
7
9
Andra Pradesh
1
6.29
15
12
3
0
1
2
Gujarat
1
7.00
9
3
6
0
2
4
Madhya
1
7.14
14
5
8
1
5
5
Pradesh
Chhattisgarh
1
7.14
7
1
4
2
4
4
Orissa
9.14
West Bengal
9.57
Punjab
10.86
Haryana
2
11.29
9
5
4
0
3
1
Rjastan
2
12.29
23
11
12
0
7
5
Bihar
2
12.43
5
3
2
0
2
0
Jharkhand
12.43
Uttar Pradesh
2
12.86
27
7
18
2
13
9
Total
155
67
78
10
53
45
Note: Villages in Tamil Nadu, Orissa, West Bengal, Punjab and Jharkhand held previous
elections before 1999. The rank of gender disparity is calculated base on Filmer et al. (1998).
Chhattisgarh was formed on November 1, 2000 by partitioning southeastern part of Madhya
Pradesh. Jharkhand was formed on November 15, 2000 by partitioning southern part of Bihar.

74

Table 2.2 Initial village and household characteristics by reservation status of the village
Total
Village reserved?
Test for
mean
Yes
No
equality
Initial Village Characteristics
Population in 1999
4,026
3,842
4,101
Share of population Hindu (%)
90.92
93.76
89.76
Share of population Muslim (%)
6.35
4.78
7.00
Net area sown (acres)
1,248
1,272
1,238
Net area irrigated under canal/stream (acres)
332
546
244
**
Access to bank
0.74
0.73
0.74
Distance to mandi (km)
15.35
18.25
14.17
Distance to retail market (km)
9.18
10.25
8.74
Distance to nearest town (km)
14.15
13.53
14.40
Distance to nearest pucca road (km)
2.67
1.36
3.21
Share of village with street light
0.45
0.50
0.43
Share of village with PCO
0.33
0.25
0.36
Share of village with trained dai
0.57
0.61
0.55
*
Share of village with male health workers
0.45
0.36
0.49
Share of village with female health workers
0.61
0.57
0.62
Share of villages with Anganwadi
0.79
0.82
0.78
Share of scheme building drinking water
0.25
0.14
0.30
Share of scheme building road
0.20
0.14
0.23
Share og villages with food for work programs
0.47
0.50
0.46
Daily male agricultural casual wage (Rs./day)
57.06
55.65
57.63
Daily female agricultural casual wage (Rs./day)
47.81
47.15
48.07
No. of observations
155
45
110
Initial Household Characteristics
Members < 14
2.09
1.97
2.15
Members 14-60
4.05
4.09
4.03
Members > 60
0.42
0.39
0.43
Head's years of education
4.60
4.61
4.60
Head completed primary school
0.48
0.48
0.48
Head completed middle school
0.30
0.29
0.30
Head‟s age
49.24
48.59
49.53
Female head
0.05
0.06
0.05
Area of cropland (acres)
2.69
2.70
2.69
Value of physical assets (Rs.)
44,067
40,318
45,723
Value of total assets (Rs.)
53,103
49,420
54,730
Household total income (Rs.)
14,829
14,895
14,800
Share of crop income
0.54
0.56
0.53
Share of agriculture
0.66
0.72
0.63
Share of off-farm income
0.33
0.27
0.36
Household total expenditure (Rs.)
38,316
37,169
38,823
Household consumption (Rs.)
9,335
9,198
9,396
No. of observations
2,990
916
2,074
Note: Columns (1)-(3) report means. The column (4) reports the |t|-statistics for tests based on
75

Table 2.2 (cont‟d)
regressions with state fixed effects. Standard errors are clustered by village for household
initial characteristics. * significant at 10%; ** significant at 5%; *** significant at 1%.

76

Table 2.3 Preference for public goods by gender
Female
Male
Difference
Test for equality
Group1
Drinking water
0.53
0.51
0.02
**
Health facilities
0.37
0.33
0.05
***
Sanitation and sewage
0.30
0.28
0.02
**
Road and transportation
0.28
0.34
-0.06
***
School and education
0.26
0.27
-0.01
Street lighting
0.23
0.25
-0.02
***
Electrification
0.19
0.19
0.00
Irrigation
0.16
0.27
-0.11
***
Government schemes
0.15
0.17
-0.01
**
Employment schemes
0.12
0.14
-0.01
***
Social issues
0.12
0.15
-0.03
***
Natural resource management
0.09
0.09
0.00
Credit
0.07
0.12
-0.05
***
No. of observations
4,681
4,808
Group 2
Drinking water
0.51
0.55
-0.04
***
Sanitation and sewage
0.43
0.39
0.04
***
Health facilities
0.41
0.35
0.06
***
Road and transportation
0.37
0.42
-0.05
***
School and education
0.30
0.33
-0.03
***
Street lighting
0.22
0.25
-0.03
***
Electrification
0.18
0.20
-0.02
**
Irrigation
0.12
0.15
-0.04
***
Social issues
0.08
0.11
-0.03
***
Government schemes
0.08
0.09
-0.02
**
Employment schemes
0.06
0.07
-0.01
**
Credit
0.05
0.07
-0.02
***
Natural resource management
0.05
0.05
0.00
No. of observations
3,393
3,546
Note: Columns (1) and (2) report means. The column (4) reports the |t|-statistics for tests of
differences of means based on regressions with village fixed effects. Standard errors are
clustered by household. * significant at 10%; ** significant at 5%; *** significant at 1%.

77

Table 2.4 Distribution of female pradhans across groups and reservation status
Reserved now
Reserved in past
Yes
No
Yes
No
Group 1 (%)
96.43
9.52
96.67
8.20
Group 2 (%)
100.00
7.69
100.00
8.16

78

Table 2.5 Problems with delivery of key public services
All States
Group 1
Group 2
Female
Male
Female
Male
Female
Male
Panel A
Reserved now (β1)
-0.0195
-0.0208
0.0156
0.00473
-0.0785*** -0.0645**
(0.0147) (0.0163)
(0.0151) (0.0172)
(0.0283)
(0.0318)
Reserved now
-0.00280 0.00418
-0.0425
-0.0279
0.0379
0.0256
*Female goods (β2)
(0.0257) (0.0255)
(0.0344) (0.0333)
(0.0303)
(0.0313)
Reserved in past (β3) -0.00901 -0.0195
0.0165
0.00424
-0.0884** -0.0915**
(0.0157) (0.0166)
(0.0189) (0.0198)
(0.0351)
(0.0373)
Reserved in past
-0.0414
-0.0342
-0.0367
-0.0159
-0.000862 -0.0177
*Female goods (β4)
(0.0286) (0.0278)
(0.0356) (0.0340)
(0.0333)
(0.0358)
Female goods
0.132*** 0.0944***
0.104*** 0.0527**
0.159*** 0.142***
(0.0171) (0.0181)
(0.0211) (0.0214)
(0.0252)
(0.0266)
Observations
88,626
91,943
51,141
52,538
37,485
39,405
R-squared
0.042
0.042
0.032
0.031
0.048
0.039
F-tests:
β1+β2=0
0.89
0.49
0.73
0.51
1.67
1.47
β3+β4=0
3.73*
4.21**
0.39
0.12
4.89**
6.67**
Panel B
Reserved now (β1)
-0.0178
-0.0165
0.0209
0.0125
-0.0768*** -0.0611**
(0.0151) (0.0170)
(0.0174) (0.0208)
(0.0246)
(0.0295)
Reserved now
0.0142
0.0170
-0.0270
-0.0189
0.0572**
0.0462
*Female goods (β2)
(0.0246) (0.0241)
(0.0341) (0.0319)
(0.0274)
(0.0292)
Reserved in past (β3) -0.00955 -0.0230
0.0168
0.00664
-0.0292
-0.0379
(0.0150) (0.0164)
(0.0194) (0.0219)
(0.0350)
(0.0388)
Reserved in past
-0.0288
-0.0204
-0.0177 0.000223
-0.00255
-0.0101
*Female goods (β4)
(0.0275) (0.0262)
(0.0359) (0.0335)
(0.0311)
(0.0359)
Female goods
0.120*** 0.0847*** 0.0981*** 0.0519**
0.140*** 0.120***
(0.0172) (0.0175)
(0.0243) (0.0239)
(0.0213)
(0.0222)
Observations
76,871
79,796
44,679
45,922
32,192
33,874
R-squared
0.040
0.040
0.032
0.030
0.047
0.042
F-tests:
β1+β2=0
0.03
0.00
0.04
0.03
0.49
0.22
β3+β4=0
2.42
2.93*
0.00
0.04
0.64
1.16
Note: Dependent variable is a zero-one indicator whether a problem existed during the current
panchayat period. Drinking water is excluded from regressions, because it is a female goods for
group 1 but not group 2. Regressions are clustered at village level and information on education,
age, gender, marriage status, household size as well as dummies for caste, religion, land
ownership, and overall wealth are included in all regressions but not reported. Village-level
characteristics in table 2.2 are also included but not reported.

79

Table 2.6 Attendance and nature of participation in panchayat meetings
All States
Group 1
Group 2
Female
Male
Female
Male
Female
Male
Attended meeting
Reserved now
0.0226
-0.000290
0.0632* 0.0300
-0.000203 -0.00810
(0.0194)
(0.0226)
(0.0337) (0.0349)
(0.00318) (0.0149)
Reserved in past
-0.0215
-0.0145
-0.0285
0.0393
0.00185
-0.0205
(0.0214)
(0.0254)
(0.0396) (0.0367)
(0.00456) (0.0317)
Observations
30,698
32,005
17,209
17,807
13,489
14,198
R-squared
0.239
0.470
0.238
0.494
0.013
0.096
Relevant issues discussed
Reserved now
0.0678
-0.0494
0.109*
-0.0243
-0.260**
(0.0567)
(0.0391)
(0.0623) (0.0375)
(0.102)
Reserved in past
0.159***
0.0621*
0.202*** 0.0639*
-0.0685
(0.0585)
(0.0364)
(0.0492) (0.0383)
(0.109)
Observations
1,909
7,044
1,843
6,099
945
R-squared
0.222
0.223
0.228
0.241
0.223
Participated discussion
Reserved now
0.162**
-0.0409
0.200** -0.0681
-0.0546
(0.0680)
(0.0386)
(0.0867) (0.0441)
(0.0689)
Reserved in past
-0.281*** -0.0512
-0.260*** -0.0451
-0.190***
(0.0738)
(0.0450)
(0.0856) (0.0455)
(0.0658)
Observations
1,909
7,044
1,843
6,099
945
R-squared
0.185
0.157
0.183
0.183
0.166
Note: In each panel, the dependent variable is a zero-one indicator of meeting attendance in the
first panel and, conditional on having attended the meeting, whether issues relevant to the
respondent were discussed and whether s/he participated discussing. Regressions are clustered at
village level and information on education, age, gender, marriage status, household size as well
as dummies for caste, religion, land ownership, and overall wealth are included in all regressions
but not reported. Village-level characteristics in table 2.2 are also included but not reported.

80

Table 2.7 The likelihood of complaint to public authorities if problems identified
All States
Group 1
Group 2
Female
Male
Female
Male
Female
Male
Reserved now (β1)
0.0182
-0.00539
0.0466
0.00337
0.0741*
0.00176
(0.0262) (0.0214)
(0.0356) (0.0281)
(0.0432) (0.0262)
Reserved now
0.00412 0.00970
-0.0273
-0.0191
0.0334*
0.0348
*Female goods (β2)
(0.0214) (0.0203)
(0.0362) (0.0317)
(0.0185) (0.0217)
Reserved in past (β3)
0.0431
0.0101
0.00616 -0.00996
0.176*** 0.109***
(0.0301) (0.0251)
(0.0356) (0.0316)
(0.0539) (0.0288)
Reserved in past
0.0220
0.0353
0.0272
0.0428
0.00109
0.0199
*Female goods (β4)
(0.0258) (0.0259)
(0.0391) (0.0388)
(0.0190) (0.0215)
Female goods
0.0212 0.0320**
0.0217
0.0361*
0.0233** 0.0305**
(0.0132) (0.0129)
(0.0221) (0.0216)
(0.0106) (0.0124)
Observations
31,522
37,678
15,986
18,713
15,536
18,965
R-squared
0.100
0.035
0.094
0.040
0.115
0.053
F-tests:
β1+β2=0
0.52
0.04
0.17
0.22
5.72**
1.28
β3+β4=0
3.39*
2.78*
0.53
0.90
12.10*** 18.11***
Note: Dependent variable is a zero-one indicator whether, if a problem was indentified, it was
reported to local authorities. Drinking water is excluded from regressions, because it is a female
goods for group 1 but not group 2. Regressions are clustered at village level and information on
education, age, gender, marriage status, household size as well as dummies for caste, religion,
land ownership, and overall wealth are included in all regressions but not reported. Village-level
characteristics in table 2 are also included but not reported. Village-level characteristics in table
2.2 are also included but not reported.

81

Table 2.8 Female empowerment
All States

Observations
R-squared
Sterilization best form of birth control
Reserved now
Reserved in past
Observations
R-squared
Female has bank account in own name
Reserved now

Group 2

0.104*
(0.0531)
0.125**
(0.0539)
3,237
0.239

0.100**
(0.0436)
0.192**
(0.0739)
2,530
0.168

-0.0508
(0.0354)
-0.0790**
(0.0396)
5,767
0.255

Reserved in past

Group 1

0.0487
(0.0372)
0.112**
(0.0440)
5,767
0.175

Use condom
Reserved now

-0.101**
(0.0485)
-0.0681
(0.0609)
3,237
0.263

-0.0974**
(0.0429)
-0.187***
(0.0579)
2,530
0.179

0.0151
0.0221
-0.00777
(0.0108)
(0.0158)
(0.0139)
Reserved in past
-0.00485
0.0139
-0.0308
(0.0112)
(0.0162)
(0.0216)
Observations
5,767
3,237
2,530
R-squared
0.106
0.119
0.139
Note: Sample includes all women aged between 14 and 50 who are resident in a household.
Regressions are clustered at village level and information on education, age, gender, marriage
status, household size as well as dummies for caste, religion, land ownership, and overall wealth
are included in all regressions but not reported. Village-level characteristics in table 2 are also
included but not reported. Village-level characteristics in table 2.2 are also included but not
reported.

82

APPENDIX

83

Table 2.A1 Problems with delivery of key public services (drinking water included)
All States
Group 1
Group 2
Female
Male
Female
Male
Female
Male
Panel A
Reserved now (β1)
-0.0161
-0.0164
0.0136
0.00146
-0.0682** -0.0515*
(0.0141) (0.0155)
(0.0156) (0.0179)
(0.0264)
(0.0296)
Reserved now
-0.00963 -0.00155
-0.0426
-0.0222
0.0311
0.0159
*Female goods (β2)
(0.0221) (0.0215)
(0.0321) (0.0313)
(0.0284)
(0.0290)
Reserved in past (β3) -0.0180 -0.0290*
0.00706 -0.00613
-0.0785** -0.0811**
(0.0146) (0.0157)
(0.0176) (0.0202)
(0.0327)
(0.0344)
Reserved in past
-0.00120 0.00319
0.00330
0.0166
-0.00541
-0.0227
*Female goods (β4)
(0.0240) (0.0234)
(0.0319) (0.0312)
(0.0305)
(0.0312)
Female goods
0.134*** 0.0986*** 0.143*** 0.0962***
0.119*** 0.104***
(0.0155) (0.0156)
(0.0205) (0.0206)
(0.0235)
(0.0244)
Observations
96,674
100,299
55,780
57,310
40,894
42,989
R-squared
0.045
0.043
0.045
0.038
0.036
0.029
F-tests:
β1+β2=0
1.50
0.71
1.05
0.50
1.34
1.21
β3+β4=0
0.65
1.10
0.13
0.12
4.37**
6.17**
Panel B
Reserved now (β1)
-0.0128 -0.00995
0.0184
0.0101
-0.0620** -0.0430
(0.0146) (0.0163)
(0.0183) (0.0215)
(0.0245)
(0.0288)
Reserved now
0.000318 0.00343
-0.0356
-0.0228
0.0463*
0.0317
*Female goods (β2)
(0.0230) (0.0220)
(0.0341) (0.0321)
(0.0260)
(0.0273)
Reserved in past (β3) -0.0169 -0.0302*
0.00631 -0.00241
-0.0187
-0.0280
(0.0142) (0.0157)
(0.0181) (0.0218)
(0.0340)
(0.0363)
Reserved in past
0.00758
0.0109
0.0138
0.0221
-0.0125
-0.0206
*Female goods (β4)
(0.0254) (0.0243)
(0.0338) (0.0322)
(0.0274)
(0.0307)
Female goods
0.127*** 0.0950*** 0.141*** 0.101***
0.105*** 0.0868***
(0.0169) (0.0162)
(0.0241) (0.0229)
(0.0207)
(0.0210)
Observations
83,848
87,046
48,730
50,092
35,118
36,954
R-squared
0.043
0.042
0.047
0.040
0.034
0.031
F-tests:
β1+β2=0
0.36
0.09
0.34
0.15
0.29
0.12
β3+β4=0
0.16
0.60
0.46
0.38
0.63
1.25
Note: Dependent variable is a zero-one indicator whether a problem existed during the current
panchayat period. Regressions are clustered at village level and information on education, age,
gender, marriage status, household size as well as dummies for caste, religion, land ownership,
and overall wealth are included in all regressions but not reported. Village-level characteristics in
table 2 are also included but not reported.

84

Table 2.A2 The likelihood of complaint to public authorities if problems identified (drinking
water included)
All States
Group 1
Group 2
Female
Male
Female
Male
Female
Male
Reserved now (β1)
0.0255
0.00187
0.0408 -0.000988
0.101** 0.0188
(0.0272) (0.0208)
(0.0365) (0.0281)
(0.0466) (0.0265)
Reserved now
-0.00527 -0.00515
-0.0198
-0.0153
0.0116
0.0138
*Female goods (β2)
(0.0215) (0.0203)
(0.0358) (0.0322)
(0.0168) (0.0196)
Reserved in past (β3)
0.0360 -0.000238
-0.0161
-0.0203
0.186*** 0.105***
(0.0327) (0.0263)
(0.0366) (0.0309)
(0.0571) (0.0263)
Reserved in past
0.0300 0.0538**
0.0467
0.0569
-0.00163 0.0198
*Female goods (β4)
(0.0272) (0.0258)
(0.0372) (0.0365)
(0.0171) (0.0165)
Female goods
0.0267** 0.0350*** 0.0505*** 0.0765*** 0.000710 -0.00418
(0.0124) (0.0132)
(0.0189) (0.0201)
(0.0110) (0.0125)
Observations
36,814
43,636
18,891
21,886
17,923
21,750
R-squared
0.102
0.037
0.097
0.049
0.116
0.053
F-tests:
β1+β2=0
0.46
0.03
0.25
0.37
5.76**
0.96
β3+β4=0
3.68*
4.57**
0.56
1.84
12.40*** 18.66***
Note: Dependent variable is a zero-one indicator whether, if a problem was indentified, it was
reported to local authorities. Regressions are clustered at village level and information on
education, age, gender, marriage status, household size as well as dummies for caste, religion,
land ownership, and overall wealth are included in all regressions but not reported. Village-level
characteristics in table 2 are also included but not reported. Village-level characteristics in table
2 are also included but not reported.

85

REFERENCES

86

REFERENCES

Anderson, S. 2003. “Why Dowry Payments Declined with Modernization in Europe but Are
Rising in India.” Journal of Political Economy 111(2): 269-310.
Bagnoli, M. and B. L. Lipman. 1989. “Provision of Public-Goods – Fully Implementing the Core
Through Private Contributions.” Review of Economic Studies 56(4): 583-601.
Bagnoli, M. and B. L. Lipman. 1992. “Private Provision of Public-Goods Can be Efficient.”
Public Choice 74(1): 59-78.
Ban, R. and V. Rao. 2008. “Tokenism or Agency? The Impact of Women‟s Reservations on
Village Democracies in South India.” Economic Development and Cultural Change 56(3):
501-30.
Bardhan, P., D. Mookherjee and M. Parra Torrado. 2010. “Impact of Political Reservations in
West Bengal Local Governments on Anti-Poverty Targeting.” Journal of Globalization
and Development 1(1): 1-34.
Beaman, L., R. Chattopadhyay, E. Duflo, R. Pande and P. Topalova. 2009. “Powerful Women:
Does Exposure Reduce Bias?” Quarterly Journal of Economics 124(4): 1497-540.
Beaman, L., E. Duflo, R. Pande, and P. Topalova. 2010. “Political reservation and substantive
representation: Evidence from Indian village councils.” Paper presented at the 2010 India
Policy Forum. New Delhi: National Council of Applied Economic Research.
Besley, T. and S. Coate. 1997. “An economic model of representative democracy.” Quarterly
Journal of Economics 112(1): 85-114.
Besley, T., R. Pande, L. Rahman and V. Rao. 2004. “The Politics of Public Good Provision:
Evidence from India Local Governments.” Journal of European Economic Association
2(2-3): 416-26.
Besley, T., R. Pande, and V. Rao. 2005. “Political Selection and the Quality of Government:
Evidence from South India.” Center Discussion Paper No.921, Economic Growth Center,
Yale University.
Bhalotra, S. and I. Clots-Figueras. 2010. “Health and the political agency of women.” Bristol and
Madrid: University of Bristol and Carlos Madrid III Working Paper.
Bhalotra, S. and S. Rawlings. 2011. “Intergenerational persistence in health in developing
countries: The penality of gender inequality.” Journal of Public Economics 95(3-4): 286299.
Bhavnani, R. R. 2009. “Do Electoral Quotas Work after They Are Withdrawn? Evidence from a
Natural Experiment in India.” American Political Science Review 103(1): 23-35.

87

Chandrashekhar, S. 2008. “Willingness to contribute to public projects – evidence from India.”
Working Paper. Mumbai: Indira Ghandi Institute of Development Research.
Chattopadhyay, R. and E. Duflo. 2004. “Women as Policy Makers: Evidence from a
Randomized Policy Experiment in India.” Econometrica 72(5): 1409-43.
Clots-Figueras, I. 2009. “Are female leaders good for education? Evidence from India.” Madrid:
University Working Paper.
Dahlerup, D. 2006. Women, Quotas and Politics. London and New York: Routledge.
Dollar, D., R. Fisman and R. Gatti. 2001. “Are Women Really the “Fairer” Sex? Corruption and
Women in Government.” Journal of Economic Behavior and Organization 46(4): 423429.
Dongre, A. 2010. “Female political leadership and the prevalence of water borne diseases:
Evidence from a natural experiment in India.” Santa Cruz, CA: University of California
Santa Cruz.
Downs, A. 1957. An Economic Theory of Democracy. Boston, MA: Addison Wesley.
Edlund, L. and R. Pande. 2002. “Why have Women Become Left-Wing? The Political Gender
Gap and the Decline in Marriage.” Quarterly Journal of Economics 117(3): 917-61.
Ferreira, F. and J. Gyourko. 2010. “Does gender matter for political leadership: The case of U.S.
mayors.” Philadelphia, PA: Wharton School, University of Pennsylvania.
Filmer, D., E. M. King and L. Pritchett. 1998. “Gender Disparity in South Asia.” Policy
Research Working Paper no. 1867, World Bank, Washington, DC.
Frechette, G. R., F. Maniquet and M. Morelli. 2008. “Incumbents‟ interests and gender quotas.”
American Journal of Political Science 52(4): 891-909.
Funk, P. and C. Gathman. 2008. “Gender gaps in policy making: Evidence from direct
democracy in Switzerland.” Barcelona: University Pompeu Fabra.
Gajwani, K. and X. Zhang. 2008. “Gender, caste, and public goods provision in Indian village
governments.” Discussion paper 00807. Washington DE: IFPRI, Development strategy
and governance division.
International Institute for Democracy and Electoral Assistance (IDEA). 2008. “Global database
of quotas for women.” www.quotaproject.org.
Iyer, L., A. Mani, P. Mishra, and P. Topalova. 2010. “Political representation and crime:
Evidence from India.” Draft paper. Washington DC: International Monetary Fund.
Knack, S. and P. Sanyal. 2000. “Making a Difference: Do Women in Government Affect Health
and Education Outcomes.” Mimeo, World Bank.
Krishnan, N. 2007. “Political Reservation and Rural Public Good Provision in India.” Discussion
Paper Series No.175, Department of Economics, Boston University.
88

Krook, M. L. 2009. Quotas for women in politics: Gender and candidate selection worldwide.
New York and London: Oxford University Press.
Lott, J. R. and L. W. Kenny. 1999. “Did Women's Suffrage Change the Size and Scope of
Government?” Journal of Political Economy 107(6): 1163-98.
Ministry of Panchayati Raj. 2008. Study on elected women representatives in Panchayati Raj
institutions. New Delhi: Government Printer.
Messer, K. D. and H. Zarghamee. 2007. “New hope for the voluntary contributions mechanism:
The effects of context.” Journal of Public Economics 91(9): 1783-99.
Miller, G. 2008. “Women‟s suffrage, political responsiveness, and child survival in American
history.” Quarterly Journal of Economics 123(3): 1287-327.
Munshi, K. and M. R. Rosenzweig. 2008. “The efficiency of parochial politics: Caste,
commitment, and competence in Indian local governments.” Center Discussion Paper 964.
New Haven, CT: Economic Growth Center Yale University.
Osborne, M. J. and A. Slivinski. 1996. “A model of political competition with citizen-candidates.”
Quarterly Journal of Economics 111(1): 65-96.
Page, S. E. 2007. The Difference: How the Power of Diversity Creates Better Groups,Firms,
Schools, and Societies. Princeton: Princeton University Press.
Pande, R. 2003. “Can Mandated Political Representation Increase Policy Influence for
Disadvantaged Minorities? Theory and Evidence from India.” American Economic
Review 9 (4): 1132-51.
Paxton, P. and M. M. Hughes. 2007. Women, Politics, and Power: A Global Perspective. Los
Angeles: Pine Forge Press.
Powley, E. 2007. “Rwanda: The impact of women legislators on policy outcomes affecting
children and families.” The state of the world‟s children 2007: Background paper.
Geneva: UNICEF.
Raabe, K., M. Sekher, and R. Birner. 2009. “The Effects of Political Reservations for Women on
Local Governance and Rural Service Provision.” IFPRI Discussion Paper 00878.
Washington D.C.: International Food Policy Research Institute.
Rajaraman, I. and M. Gupta. 2008a. “Further evidence on the policy impact of randomised
political reservation.” New Delhi: National Institute of Public Finance and Policy.
Rehavi, M. M. 2008. “Sex and politics: Do female legislators affect state spending?”. University
of Berkeley working paper.
Reingold, B. 1992. “Concepts of Representation among Female and Male State Legislators.”
Legislative Studies Quarterly 17(4): 509-37.

89

Saint-Germain, M. A. 1989. “Does Their Difference Make a Difference? The Impact of Women
on Public Policy in Arizona Legislature.” Conference Papers No.150. Atlanta: 85th
Annual Meeting of the American Political Science Association.
Schwarz, J. A. 2010. “Election behavior with high-female representation: A regression
discontinuity analysis.” Pittsburgh: University of Pittsburgh, Department of Economics.
Singh, N. 2007. “Fiscal federalism and decentralization in Inida.” Working Paper. Santa Cruz,
CA: University of California.

90

CHAPTER 3: DOSE POLITICAL RESERVATION AFFECT LONG-TERM
ELECTORAL PARTICIPATION AND OUTCOMES? EVIDENCE FROM RURAL
INDIA

3.1 Introduction
Gender quotas introduced in some 100 countries worldwide have brought a substantially
increasing number of women into the political arena.

26

Whether appropriate representation of

women persists after quotas lapse deserves an exploration, because quotas are at the cost of
preventing nontarget groups from running for office and the ultimate goal of quotas is to achieve
the equilibrium where historically excluded or marginalized groups could participate in political
processes without particular supports.
Despite the importance of understanding whether electoral quotas achieve the long-term
goal of increasing representation of targeted groups in leadership positions, there are very few
studies that answer this question empirically. Bhavnani (2009) gives two main challenges faced
by empirical studies of electoral quotas. First, quotas are rarely withdrawn, which makes it
practically impossible to evaluate the impact of quotas after quotas are withdrawn. Second,
quotas are typically not implemented in a random fashion, which prevents researchers from
drawing conclusions about causal relationships between quotas and outcomes. Taking advantage
of the gender quotas introduced by law in Italy in 1993 and repealed in 1995 (only municipal
council elections which took place between April 1993 and September 1995 were affected by the
reform), De Paola et al. (2010) show that women‟s representation after the reform increased
26

The form of gender quotas varies widely across countries. For example, Argentina is the first
country in the world to legislate a candidate gender quota of 30% in 1991 (Franceschet and
Piscopo 2008). The 2003 Rwandan Constitution established a gender quota providing for 30%
reserved seats for women (Powley 2007). The French National Assembly passed a law in 2000
that requires 50% of candidates chosen by parties are women (Frechette et al. 2008). Despite no
legal mandate for the share of female legislators, all political parties in Sweden commit to some
equal gender representation (Krook 2009).
91

significantly more in municipalities that were affected by the reform than in municipalities that
were not affected. The fact that the reservation policy in India is implemented randomly opens
up a better opportunity for researchers to study the impacts of the reservation policy on a variety
of outcomes. Indeed, there have been many studies on the effects of the reservation policy
implemented at different levels of leadership positions. However, most of studies tend to focus
on the effectiveness of reservation in terms of quantity and quality of public good provisions and
citizens‟, especially female citizens‟ perceptions about performances of reserved leaders (Ban
and Rao 2008; Beaman et al. 2010; Chattopadhyay and Duflo 2004). The two studies that
specifically examined the impacts of female reservation on electoral outcomes yield mixed
results and rely on data from rather narrowly focused areas (one city or one state).
Bhavnani (2009) finds that gender quotas in the previous election increase the probability
of a woman winning office by approximately five times in urban Mumbai. Beaman et al. (2009)
suggest that, in rural West Bengal, women are only more likely to contest and win open seats
when continuously exposed to female reservations. Relative to never reserved wards, the
proportion of female candidates and elected leaders increased by 3.3% and 5.2%, respectively, in
wards that were twice required to have a female chief councilor. In this paper, we try to
complement the existing literature by using nationally-representative data reflecting population
diversity to achieve a higher degree of external validity (Duflo 2006). Specifically, we draw on a
large national sample from surveys conducted in rural India in 2000 and 2008 by the National
Council of Applied Economics Research (NCAER) to assess whether, after gender quotas are
withdrawn, (i) women are more likely to run for and win office; (ii) women are more likely to
participate in voting; and (iii) villagers vote for leaders based on their qualifications (technical
qualifications, knowledge of local or national problems, honesty and fairness, or ability to

92

represent local problems to government) rather than their social influences (the candidate‟s social
power, caste, religion, or party affiliation), spousal influence, or monetary inducement.
Three sets of results are relevant. First, we do not find evidence that reservation in the
past increases the likelihood of women running for or winning open seats. Reasons might be
because one time reservation may not last long enough (i) to motivate a female villager to enter
the political life, or (ii) to reduce voters‟ long-standing biases against female leadership. This is
consistent with the findings of Beaman et al. (2009) that reservation increases the likelihood of
female candidates winning elections only in villages where reserved seats were sustained for two
27

consecutive election cycles . Second, we do not find that female villagers are more likely to
participate in voting in villages that reserved positions for females in the past, perhaps because
the participation rate of voting has already been high. Third, reservation seems to have
significant impacts on voting quality as both men and women in places where leadership was
reserved for women in the past have a clear propensity to cast their votes based on candidate‟s
qualifications rather than candidate‟s social identities relative to villages where there was no
reservation in the past. Moreover, past reservation empowers women to make voting decisions
on their own. The combination of these findings may reinforce the importance of longer-term
exposure to female leadership.
The paper is structured as follows: Section 2 discusses the Indian context of gender
quotas and describes the data we use. Section 3 provides the hypotheses guiding our

27

We also suspect that accumulated effects are more significant than immediate effects, as
documented by Beaman et al. (2009). However, of all the villages that have open elections, in
only 5% of villages (7 out of 147) was the leadership reserved for women in two consecutive
previous periods, which makes it difficult to estimate accumulated effects, leaving us to identify
past-period effects induced by female reservation.
93

investigation and the estimation strategy. Section 4 reports econometric results. Section 5
concludes with a set of policy implications and suggestions for future research.

3.2 Institutional context and data
rd

In 1992, India‟s 73 Constitutional Amendment mandated far-reaching decentralization
by establishing a three-tier system of district, block and village-level councils. The gram
panchayat (GP) is the lowest tier of local government at village level. It comprises a president
(pradhan or sarpanch) and council members who are elected from the panchayat‟s wards. Its
responsibilities include (i) provision of major public services such as health, education, drinking
water, and roads; (ii) setting rates and administering local taxes; (iii) administration, formulation
and implementation of local development plans; and (iv) selection of beneficiaries and
implementation of social and economic programs established and paid for by the central
government. Regular assemblies (gram sabhas) by all voters in the GP are meant to monitor
performance and increase democratic accountability. To prevent decentralization from
reinforcing the power wielded by traditional elites and to counter what was perceived as a legacy
of disenfanchisement and under-representation by females and other disadvantaged groups, this
was combined with reservation of a share of seats for women as well as scheduled castes and
28

tribes.

28

At village level, most states now reserve one third of council members and pradhan positions
for women. By comparison, the share of positions reserved to ST/SCs equals the population
share of the ST/SC population (Bardhan et al. 2010). Although many studies explored impacts of
caste reservations (Krishnan 2007, Pande 2003), our focus is on those for females. At district
level, it appears that reservation of positions in the legislature for scheduled castes but not tribes
improves access to education facilities, mainly primary schools, for relevant constituencies.
94

As seats reserved for women are assigned randomly in each election cycle, they allow us
to explore causal relationships between quotas and electoral outcomes. We use data from
nationally-representative surveys of 155 villages in rural India conducted in 2000 and 2008 by
the National Council of Applied Economics (NCAER). In addition to its national coverage, this
dataset has a number of desirable features. First, it provides information on reservation status by
rd

panchayat for two rounds of elections after the provisions of the 73 amendment were adopted
29

in a state.

Table 3.1, which provides a breakdown of reservation by panchayat, illustrates that

out of the 155 villages in the sample, 35 villages had open elections in the current period but not
in the previous period (the treatment group), and 67 villages had open elections in both periods
(the control group). A comparison between treatment and control villages allows us to examine
whether past reservation affects outcomes in current open elections. Second, data are available
for all candidates and elected pradhans of current panchayat. Third, data on a range of voting
patterns are obtained individually for all household members 18 years and older. We are thus
able to not only examine the main question –whether past reservation increases the likelihood of
women winning office in the next open election– but also to identify some channels through
which the results materialize.
Table 3.2 provides descriptive statistics for key dependent variables. Overall women won
elections in 8.8% of villages (8.6% of treatment villages and 9.0% of control villages), but only
6.8% (6.5% in treatment villages and 7.2% in control villages) of candidates were women. The
bottom panel suggests the high participation rate of voting: 82.8% of female respondents (82.3%
in treatment villages and 83.0% in control villages) and 84.6% of male respondents (84.0% in

29

While this includes three elections for most of the states, Bihar, Orissa, Punjab, Tamil Nadu
and West Bengal held the earliest elections in 1992, before the amendment was adopted.
95

treatment villages and 84.9% in control villages) cast their votes in current open elections. While
we test whether reservation could further motivate women to participate in voting, it is more
important to explore the basis on which villagers made their voting decisions, which allows us to
examine whether reservation has improved the quality of elections. Since respondents often
identified multiple determinants of vote, we construct for each determinant a normalized
indicator that varies between 0 (none of the listed factors determined one‟s vote) and 1 (only one
of the listed factors determined one‟s vote) to serves as our main outcome variable.

30

Descriptive statistics suggest considerable improvement in the quality of election in villages that
were reserved in the last election cycle relative to those not reserved. For instances, reservation
increases the percentage of villagers who reported knowledge of local problems, honesty and
fairness, and ability to represent local problems to government as determinants by between 1.5%
and 5.2%, and decreases the percentage of female villagers who voted as spouse told by 2.6%.
Of course, a caveat is due for achieving solid conclusions, as we have not reported any statistics
for testing equality of means. Econometric analysis in the fourth section will remove this
concern.

3.3 Hypotheses and estimation strategy
We follow the framework developed in Chattopadhyay and Duflo (2004). Citizens run for office
with the cost δ (the cost for women, δw, is greater than the cost of men, δm). The final policy
implemented is determined by the policy option

, preferred by the elected official with a weight

, and the policy option μ’, preferred by the local elite. μ’ is assumed to be more “pro-male” than

30

For example, if the individual reported to vote for two reasons, the value of each reason turns
from one to a half.
96

the median voter‟s preference,

, since the local elite tend to be male. Citizens run for office if

the utility gain from the implementation of the mixed polity (net of the cost of running for office)
is bigger than the utility gain from the implementation of μ’. They propose that there is no
equilibrium where a woman runs for office in the absence of reservation, if δw -.5*δm > μ’-m and
δw > m-(1-α) μ’.

31

Therefore, any policy that could reduce the cost for women to run for office

is likely to increase women‟s likelihood to run for office.
The underrepresentation of females in bodies of policy making in the absence of
reservation could be due to gender differences in ambition (Gupta et al., 2005; Niederle and
Vesterlund 2007), traditional sex-role socialization (Fox and Lawless 2004; Gneezy et al. 2009),
or political discrimination (Sanbonmatsu 2006; Niven 2010). For example, laboratory
experiments suggest that women are more risk averse, more malleable in terms of social
preferences, and less competitive inclined than men (Croson and Gneezy 2009). In a survey
conducted in six Indian states, Chhibber (2002) finds that only women who have an identity that
is independent of the household are likely to avail the opportunity brought by the reservation
policy to contest elections for local bodies. Beaman et al. (2009) show that villagers who have
never been required to have a female leader prefer male leaders and perceive hypothetical female
leaders as less effective than their male counterparts, when stated performance is identical.

31

According to Chattopadhyay and Duflo (2004) the first condition is the condition under which
no woman runs unopposed. The intuition is that when the cost of running is high for women,
only women with strong pro-women preferences will want to run. But if the cost of running is
low for a man, a man can then enter and win for sure. If the second condition is satisfied, no
woman agrees to run against a man: The two candidates must have equal chances of winning,
and thus the outcome they will implement must be symmetric around the median voter. Under
this condition, the distance between the outcomes implemented by the two most extreme
candidates symmetric among the median voter is too small to compensate even the most extreme
woman‟s cost of running.
97

There are two competing arguments on whether the reservation policy leads to more
female candidates and female leaders in the future after reservation was withdrawn. There are
reasons why reservation could increase female participation in future elections. First, reservation
is likely to motivate women to directly participate in public and political spheres and also
encourages their families to accept their decisions to engage in political life (Bhavnani 2009).
Second, exposure to female leaders would improve voters‟ perception about female leader ability
to lead and alters their voting behaviors to gradually accept female candidates (Beaman et al.
2009). We thereby hypothesize that women would like to run for office in open elections, if
reservation in the previous period reduces their cost of running (δw).
In the meanwhile, reservation draws in women who did not participate in voting before to
participate in the voting process, which moves the median voter‟s preference,

, in a more “pro-

female” direction. We also expect a higher likelihood of women winning open elections, if
reservation increases the candidate pool of women and weakens the stereotype about the
capabilities of women in politics.
However, we may not observe what are hypothesized, since there are also a number of
counterarguments that go against electing female leaders, as summarized in Beaman et al.
(2009). For example, one reason is that voters may dislike the reservation policy which restricts
their choices, and therefore women leaders (Thernstrom and Thernstrom 1997). Another reason
is that voters may perceive gender quotas as violating social norms and potentially reducing the
value of traditionally male activities (Goldin 1990). It is also possible that if a voter was exposed
to an underperformance of a female leader, then the voter has even more negative voting attitude
toward women leaders in future election (Coate and Loury 1993). For all these reasons, the

98

reservation policy could increase women‟s cost of running and reinforce the stereotype view
against female leaders.
Finally, we hypothesize that past reservation could improve the quality of current open
elections by reinforcing villagers‟ equal minds in general and reducing judgments based on
candidates‟ identities not only of gender but also of wealth, caste, religion, and party affiliation
that might translate into election-related violence.

32

Methodologically, as reserved seats are assigned randomly, OLS regressions of outcome
variables on a past reservation dummy will yield unbiased and consistent estimates of the policy
impact. Village-level control variables are included for improving the precision of estimates.
Letting superscript j index outcome and subscripts, i, v, s, and t index individuals, villages, states,
and time, the equations to estimate include
j

j

j

j

Y ivst = β s + β 1Rvst-1+ε ivsgt
j

j

j

j

j

(1)
j

Y ivst = β s + β 1Rvsgt-1 + β 2Xivst+β 3Zvs+ε ivst

(2)

j

where Y ivst is the outcome variable of interest, βs is a state fixed effect, Rvst-1 is an indicator
variable for reservation in the previous period that equal one if the pradhan position in village v
of state s at t-1 was reserved for females and zero otherwise, Xivst is a vector of household and
individual controls such as wealth and land ownership status, age, sex, caste, and broad
occupation, Zvs is a vector of initial village characteristics listed in table 3.3, and the βs are
parameters to be estimated. The coefficients β1 can be interpreted as the effect of reservation in

32

For example, Kumar (2001) documents that, during the 1998 panchayat elections of Bihar, a
most peaceful state devoid of any kind of violence in its day-to-day life, 94 people were killed in
the election-related violence, of which 29 people were killed in police firing alone.
99

j

the previous period on Y which allow us to test for any reservation-induced effects after
reservation is withdrawn. We estimate equation (2) separately by gender to explore genderdifferentiated effects.
It is worth noting that a precondition for our empirical strategy to be valid, i.e. for OLS
estimates to be interpreted as a causal effect, is that allocation of reserved seats across villages is
indeed random. The fact that one round of the panel survey was conducted in 1999 allows us to
test this for the 2000 election. Descriptive statistics for village and household characteristics in
1999 in panel 1 and 2 of table 3.3 reveals little if any bias in this regard. Some 29% (45 out of
155) of the sample villages were reserved in this period. Reserved and unreserved villages are
similarly sized and have similar levels of access to infrastructure and public goods (drinking
33

water, anganwadi , food for work programs), with a significant difference only for 2 out of 21
34

village characteristics (net area irrigated under canal or stream and the number of trained dai ).
Household characteristics such as education, age, assets, income, and its composition, are not
significantly different between reserved and unreserved villages either.

3.4 Econometric results
3.4.1 Women’s likelihood of winning open seats and channels
Regarding the main question, the first column of table 3.4 reports the result from the
linear probability model of equation (1). The dependent variable being an indicator that equals

33

As part of the Integrated Child Development Services program, India government started
anganwadi to combat child hunger and malnutrition. They also provides basic health care
in Indian villages, including basic contraceptive counseling and supply, nutrition education and
supplementation, as well as pre-school activities.
34
Trained dai is traditional birth attendant.
100

one if a woman won office in the open election and zero otherwise. We do not find that past
reservation increases women‟s chances of winning open elections, suggested by the insignificant
coefficient on the „past reservation‟ indicator even at the 10% level. We explore potential
channel through which this non result emerges in the rest of this section.
Reservation could increase the likelihood of women winning office by bringing more
women into the candidate pool. However, this proposition is rejected by the fact that the
coefficients on the „past reservation‟ indicator are not statistically significant in the case of
„number of female candidates‟. The fact that elections at village level are typically not party
based (Aziz 2000; Kerbart and Sivakumar 2005; Suri 2001) excludes the reason that political
parties discriminated against women by systematically granting them fewer tickets relative to
men. Bhavnani (2009) attributes the dearth of female candidates to individual hurdles and
familial hurdles: (i) women are less likely than men to view themselves as qualified to run for
office (Fox and Lawless 2004); and (ii) lack of autonomy within the household of male
dominance prevents women from stepping out of the household and involving in political
activities taking place in the public sphere (Chhibber 2002). And it may require long-term
exposure to motivate females to start to engage in political activities and for their families and
other voters to accept them as qualified and effective leaders.

3.4.2 Voting patterns
If we view candidates as the supply side of the electoral system, voters comprise the
demand side. Failures in the supply side to produce female pradhans might be compensated by
mobilizing female villagers to cast their votes which alters median voters‟ preferences and
increases the likelihood of “pro-female” candidates being elected, and by boosting votes based

101

on candidate‟s qualifications rather than their identities which brings about more qualified and
responsible pradhans irrespective of their gender.
Table 3.5 presents results from the linear probability model of equation (2). Although the
lack of significance of coefficients on the „past reservation‟ indicator suggests that women, and
even men, did not become more likely to cast their votes in past-reserved villages, we note that
reservation significantly alters the determinants of their voting decisions from religion and caste
to ability to represent local problems to government and knowledge of local problems.
Specifically, reservation reduces the probability of male villagers casting their votes based on
candidate‟s caste and religion backgrounds by 2.5% and 2.7% respectively, and increases the
probability of male villagers casting their votes based on candidate‟s ability to represent local
problems to government by 3.9%. Female villagers are estimated to be less likely to cast their
votes based on candidate‟s religion background by 2.4% and more likely to cast their votes based
on his/her knowledge of local problems and ability to represent local problems to government by
5.5% and 4.4%, respectively, in villages where leadership was reserved for a woman than in
unreserved villages. Results in table 3.4 also suggest that reservation empowers women within
the household by decreasing the likelihood of women whose voting decisions were manipulated
by their husbands by 2.9%.
We also report results for non-normalized determinants of vote in appendix table 3.A1 for
a robustness check. Consistent with what we find from the main regressions, reservation reduces
the probability of villagers casting their votes based on candidate‟s caste and religion
backgrounds, and increases the probability of villagers casting their votes based on candidate‟s
qualifications.

102

3.5 Conclusion
Our analysis was motivated by the question whether political quotas can increase women
in leadership positions after they are withdrawn. Related to this, we also explored the channels
through which the results come about and whether such policies have positive effects on the
demand side of the electoral systems. India‟s setting where far-reaching policies to politically
empower women and overcome gender bias and social stratification have been adopted some
time ago and are allocated randomly across villages is ideal for such analysis. A nation-wide
sample with a rich set of information at village- and individual-level allows us to contribute to
the literature by enhancing both internal and external validity.
A few interesting results stand out. First, unlike Bhavnani (2009), we do not find
reservation of leadership for women in the past boosts the number of female candidates and
increases the likelihood of women winning open election in the next round of election. Instead,
our results are more in line with those of Beaman et al. (2009) that reservation in the immediate
past election cycle does not lead to more female candidates or female winners of the elections.
This may suggest that the effect of reservation on political empowerment of female residents are
more immediate in the urban area than in rural area. It may also suggest that the traditional
prejudice against women in leadership positions may be more severe in the rural area than in
urban area. In order to achieve the long-term goal of increasing female representation in local
leadership positions in rural India, it is advisable to implement the reservation policy in repeated
manners. Second, there is clear evidence that reservation has considerable impacts on villagers‟
voting behaviors. Reservation motivates villagers irrespective of their gender to cast their votes
based on candidate‟s qualifications, in particular knowledge of local problems and ability to
report these problems to government, rather than on candidate‟s social power, caste, religion,

103

party affiliation or monetary inducement. Third, female reservation also has a significant
empowerment effect whereby women vote for candidates they suppose to be qualified rather
than obey the husband. Regarding complex dynamics of reservation-induced effects within and
outside the household, such a positive effect witnesses the potential of women being politically
active in the public arena in the absence of affirmative policies. The combination of results tends
to suggest that the low number of elected female leaders in previously reserved villages is
probably mainly due to few motivated female villagers to run for leaders position to begin with.
There are two areas where follow-up research would be desirable. First, similar analysis
to assess impacts of the reservation policy for certain castes or tribes could help inform the
current policy debate. Second, we focus on effects in the immediate period after quotas lapse,
hindered by the limited sample size of our dataset. For a variety of reasons, these may be very
different from longer-term and accumulated changes in behavior. To achieve a judicious
assessment of such policies, it will be important to complement what we find in this paper with
evidence on the extent to which reservation-induced effects are reversed and amplified over time.
They could also help in the understanding of the channels through which such effects materialize
an issue of great policy interest.

104

Table 3.1 Distribution of reservation status
Previous Panchayat
Open Reserved Total
67
35
102
Current
Open
43
10
53
Panchayat Reserved
110
45
155
Total

105

Table 3.2 Descriptive statistics for key dependent variables
Total Reserved in the past Open in the past
mean
Open now
Open now
Pradhan and candidates
% of female pradhans
8.82
8.57
8.96
Number of total candidates
3.84
4.00
3.76
Number of female candidates
0.26
0.26
0.27
No. of observations
102
35
67
Female voters
% of voters out of respondents
82.77
82.33
83.00
% of voting for technical qualifications
3.91
3.25
4.26
% of voting for local knowledge
25.60
28.99
23.82
% of voting for national knowledge
5.58
4.57
6.12
% of voting for honesty and fairness
20.51
21.48
20.01
% of voting for ability to represent problems 11.27
13.32
10.20
% of voting for candidate‟s social power
13.31
12.03
13.97
% of voting for caste
5.17
4.25
5.65
% of voting for religion
4.46
3.61
4.91
% of voting for political parties
0.76
0.60
0.85
% of voting as spouse told
7.07
5.34
7.98
% of voting for monetary inducement
0.02
0.02
0.02
Male voters
% of voters out of respondents
84.62
84.02
84.94
% of voting for technical qualifications
5.43
5.09
5.61
% of voting for local knowledge
26.52
29.06
25.18
% of voting for national knowledge
5.89
5.62
6.03
% of voting for honesty and fairness
21.66
23.32
20.78
% of voting for ability to represent problems 12.22
13.67
11.45
% of voting for candidate‟s social power
15.80
14.15
16.68
% of voting for caste
4.90
3.89
5.44
% of voting for religion
4.16
2.82
4.87
% of voting for political parties
1.29
1.03
1.43
% of voting as spouse told
0.37
0.20
0.45
% of voting for monetary inducement
0.02
0.02
0.02
No. of observations
7,631
2,648
4,983

106

Table 3.3 Initial village and household characteristics by reservation status of the village
Total
Village reserved?
Test for
mean
Yes
No
equality
Initial Village Characteristics
Population in 1999
4,026
3,842
4,101
Share of population Hindu (%)
90.92
93.76
89.76
Share of population Muslim (%)
6.35
4.78
7.00
Net area sown (acres)
1,248
1,272
1,238
Net area irrigated under canal/stream (acres)
332
546
244
**
Access to bank
0.74
0.73
0.74
Distance to mandi (km)
15.35
18.25
14.17
Distance to retail market (km)
9.18
10.25
8.74
Distance to nearest town (km)
14.15
13.53
14.40
Distance to nearest pucca road (km)
2.67
1.36
3.21
Share of village with street light
0.45
0.50
0.43
Share of village with PCO
0.33
0.25
0.36
Share of village with trained dai
0.57
0.61
0.55
*
Share of village with male health workers
0.45
0.36
0.49
Share of village with female health workers
0.61
0.57
0.62
Share of villages with Anganwadi
0.79
0.82
0.78
Share of scheme building drinking water
0.25
0.14
0.30
Share of scheme building road
0.20
0.14
0.23
Share og villages with food for work programs
0.47
0.50
0.46
Daily male agricultural casual wage (Rs./day)
57.06
55.65
57.63
Daily female agricultural casual wage (Rs./day)
47.81
47.15
48.07
No. of observations
155
45
110
Initial Household Characteristics
Members < 14
2.09
1.97
2.15
Members 14-60
4.05
4.09
4.03
Members > 60
0.42
0.39
0.43
Head's years of education
4.60
4.61
4.60
Head completed primary school
0.48
0.48
0.48
Head completed middle school
0.30
0.29
0.30
Head's age
49.24
48.59
49.53
Female head
0.05
0.06
0.05
Area of cropland (acres)
2.69
2.70
2.69
Value of physical assets (Rs.)
44,067
40,318
45,723
Value of total assets (Rs.)
53,103
49,420
54,730
Household total income (Rs.)
14,829
14,895
14,800
Share of crop income
0.54
0.56
0.53
Share of agriculture
0.66
0.72
0.63
Share of off-farm income
0.33
0.27
0.36
Household total expenditure (Rs.)
38,316
37,169
38,823
Household consumption (Rs.)
9,335
9,198
9,396
No. of observations
2,990
916
2,074
Note: Columns (1)-(3) report means. The column (4) reports the |t|-statistics for tests based on
107

Table 3.3 (cont‟d)
regressions with state fixed effects. Standard errors are clustered by village for household
initial characteristics. * significant at 10%; ** significant at 5%; *** significant at 1%.

108

Table 3.4 Women running for office in the next open election
Female
Number of total Number of female
pradhan
candidates
candidates
Past reservation
0.0271
0.361
0.0456
(0.0503)
(0.471)
(0.135)
Observations
102
102
102
R-squared
0.195
0.590
0.199
Note: Regressions are with state fixed effects. Robust standard errors are in parentheses.

109

Table 3.5 Villager‟s voting pattern in the next open election
Voted
Voted for
Voted for
Voted for
technical
knowledge of
knowledge of
qualifications
local problems national problems
Female
Past reservation
-0.0315
0.00481
0.0550*
-0.0186
(0.0581)
(0.00850)
(0.0295)
(0.0114)
Observations
3,743
3,098
3,098
3,098
R-squared
0.330
0.095
0.140
0.285
Male
Past reservation
-0.0550
0.0139
0.0267
-0.0126
(0.0589)
(0.0127)
(0.0261)
(0.0113)
Observations
3,888
3,290
3,290
3,290
R-squared
0.344
0.083
0.165
0.228
Voted for social
Voted for caste Voted for religion Voted for political
power
party
Female
Past reservation
Observations
R-squared
Male
Past reservation

-0.0172
(0.0213)
3,098
0.054

-0.0223
(0.0147)
3,098
0.076

-0.0240**
(0.0117)
3,098
0.130

-0.00280
(0.00351)
3,098
0.047

Voted for
honesty and
fairness

Voted for ability
to represent local
problems to govt.

-0.00293
(0.0213)
3,098
0.102

0.0436**
(0.0193)
3,098
0.115

0.00319
(0.0209)
3,290
0.125
Voted as
spouse told

0.0394**
(0.0194)
3,290
0.126
Voted for
monetary
inducement

-0.0286*
(0.0170)
3,098
0.127

6.38e-05
(0.000510)
3,098
0.438

-0.0117
-0.0247*
-0.0272***
-0.00281
0.000767
0.000408
(0.0216)
(0.0132)
(0.00920)
(0.00451)
(0.00175)
(0.000748)
Observations
3,290
3,290
3,290
3,290
3,290
3,290
R-squared
0.086
0.066
0.161
0.147
0.030
0.514
Note: Voted is a zero-one indicator for whether the individual voted. Determinants of vote are zero-one indicators for whether the
individual voted for this reason and are normalized by the individual. Regressions are clustered at village level and information on
education, age, gender, marriage status, household size as well as dummies for caste, religion, land ownership, and overall wealth are
included in all regressions but not reported. Village-level characteristics in table 3 are also included but not reported.

110

APPENDIX

111

Table 3.A1 Villager‟s voting pattern in the next open election (determinants of vote without normalization)
Voted for
Voted for
Voted for
Voted for
Voted for ability
technical
knowledge of
knowledge of
honesty and
to represent local
qualifications
local problems national problems
fairness
problems to govt.
Female
Past reservation
0.0129
0.147***
-0.0513
0.0967*
0.154**
(0.0393)
(0.0338)
(0.0407)
(0.0539)
(0.0674)
Observations
3,098
3,098
3,098
3,098
3,098
R-squared
0.126
0.143
0.457
0.077
0.191
Male
Past reservation
0.0245
0.0824***
-0.0507
0.0950**
0.119
(0.0522)
(0.0309)
(0.0475)
(0.0474)
(0.0732)
Observations
3,290
3,290
3,290
3,290
3,290
R-squared
0.105
0.106
0.391
0.119
0.214
Voted for social
Voted for caste Voted for religion
Voted for
Voted as spouse
power
political party
told
Female
Past reservation
Observations
R-squared
Male
Past reservation

-0.0261
(0.0623)
3,098
0.084

-0.0897**
(0.0442)
3,098
0.090

-0.0916**
(0.0397)
3,098
0.226

-0.00369
(0.0145)
3,098
0.078

-0.0358
(0.0481)
3,098
0.169

Voted for
monetary
inducement
0.000255
(0.00204)
3,098
0.438

-0.0422
-0.106**
-0.123***
0.0115
0.00138
0.00163
(0.0630)
(0.0425)
(0.0429)
(0.0237)
(0.00820)
(0.00299)
Observations
3,290
3,290
3,290
3,290
3,290
3,290
R-squared
0.163
0.081
0.260
0.259
0.043
0.514
Note: Voted is a zero-one indicator for whether the individual voted. Determinants of vote are zero-one indicators for whether the
individual voted for this reason. Regressions are clustered at village level and information on education, age, gender, marriage status,
household size as well as dummies for caste, religion, land ownership, and overall wealth are included in all regressions but not
reported. Village-level characteristics in table 3 are also included but not reported.

112

REFERENCES

113

REFERENCES

Aziz, A. 2000. “Democratic decentralisation: Experience of Karnataka.” Economic and Political
Weekly 35(39): 3521-26.
Ban, R. and V. Rao. 2008. “Tokenism or Agency? The Impact of Women's Reservations on
Village Democracies in South India.” Economic Development and Cultural Change 56(3):
501-30.
Bardhan, P., D. Mookherjee and M. P. Torrado. 2010. “Impact of Political Reservations in West
Bengal Local Governments on Anti-Poverty Targeting.” Journal of Globalization and
Development 1(1): 1-34.
Beaman, L., R. Chattopadhyay, E. Duflo, R. Pande and P. Topalova. 2009. “Powerful Women:
Does Exposure Reduce Bias?” Quarterly Journal of Economics 124(4): 1497-540.
Beaman, L., E. Duflo, R. Pande, and P. Topalova. 2010. “Political reservation and substantive
representation: Evidence from Indian village councils.” Paper presented at the 2010 India
Policy Forum. New Delhi: National Council of Applied Economic Research.
Bhavnani, R. 2009. “Do Electoral Quotas Work after They Are Withdrawn? Evidence from a
Natural Experiment in India.” American Political Science Review 103(1): 23-35.
Chattopadhyay, R. and E. Duflo. 2004. “Women as Policy Makers: Evidence from a
Randomized Policy Experiment in India.” Econometrica 72(5): 1409-43.
Chhibber, P. 2002. “Why Are Some Women Politically Active? The Household, Public Space,
and Political Participation in India.” International Journal of Comparative Sociology
43(3-5): 409-29.
Coate, S. and G. Loury. 1993. “Will Affirmative-Action Policies Eliminate Negative
Stereotypes?” American Economic Review 83(5): 1220-1240.
Croson, R. and U. Gneezy. 2009. “Gender differences in preferences.” Journal of Economic
Literature 47(2): 448-474.
De Paola, M., V. Scoppa and R. Lombardo. 2010. “Can gender quotas break down negative
stereotypes? Evidence from changes in electoral rules.” Journal of Public Economics
94(5-6): 344-353
Duflo, E. 2006. “Field Experiments in Development Economics.” Prepared for the World
Congress of the Econometric Society. Cambridge, MA: Department of Economics and
Abdul Latif Jameel Poverty Action Lab, MIT.
Fox, Richard L., and Jennifer L. Lawless. 2004. “Entering the Arena? Gender and the Decision
to Run for Office.” American Journal of Political Science 48(2): 264-80.

114

Franceschet, S. and J. M. Piscopo. 2008. “Gender quotas and women‟s substantive representation:
Lessons from Argentina.” Politics & Gender 4(3): 393-425.
Frechette, G. R., F. Maniquet and M. Morelli. 2008. “Incumbents‟ interests and gender quotas.”
American Journal of Political Science 52(4): 891-909.
Goldin, C. 1990. “A Pollution Theory of Discrimination.” Mimeo, Harvard.
Gneezy, U., K. Leonard and J. List. 2009. “Gender differences in competition: Evidence from a
matrilineal and a patriarchal society.” Econometrica 77(5) 1637-1664.
Gupta, N., A. Poulsen and M. Villeval. 2005. “Male and Female Competitive BehaviorExperimental Evidence.” IZA Discussion Paper No. 1833.
Kerbart, E. and S. Sivakumar. 2005. “Panchayat elections in Rajasthan: A view from the field”
Economic and Political Weekly 40(8): 723-25.
Krishnan, N. 2007. “Political Reservation and Rural Public Good Provision in India.” Discussion
Paper Series No.175, Department of Economics, Boston University.
Krook, M. L. 2009. Quotas for women in politics: Gender and candidate selection worldwide.
New York and London: Oxford University Press.
Kumar, G. 2001. “Panchayat elections: Overcoming state‟s resistance.” Economic and Political
Weekly 36(20): 1681-84.
Niederle, M. and L. Vesterlund. 2007. “Do women shy away from competition? Do men
compete too much?” Quarterly Journal of Economics 122(3): 1067-1101.
Niven, D. 2010. “Party Elites and Women Candidates: The Shape of Bias.” In Krook, M. L., eds.,
Women, gender and politics: A reader. Oxford: Oxford University Press.
Pande, R. 2003. “Can Mandated Political Representation Increase Policy Influence for
Disadvantaged Minorities? Theory and Evidence from India.” American Economic
Review 93(4): 1132-51.
Powley, E. 2007. “Rwanda: The impact of women legislators on policy outcomes affecting
children and families.” The state of the world's children 2007: Background paper. Geneva:
UNICEF.
Sanbonmatsu, K. 2006. “Do Parties Know That „Women Win‟? Party Leader Beliefs about
Women‟s Electoral Chances.” Politics & Gender 2(4): 431-50.
Suri, K. C. 2001. “Setback for TDP in panchayat elections.” Economic and Political Weekly
36(41): 3892-95.
Thernstrom, S. and A. Thernstrom. 1997. America In Black And White: One Nation, Indivisible.
New York, NY: Simon and Schuster.

115