Skip to main content

Male backlash, bargaining, or exposure reduction?: women’s working status and physical spousal violence in India


Labor force participation of women is expected to decrease the risk of spousal violence by enhancing their bargaining power or diminishing their contacts with abusive partners. The opposite effect is predicted when female employment induces male backlash. I identify the effect of female employment on spousal violence by exploiting the exogenous variations in rural women’s working status driven by rainfall shocks and the rice–wheat dichotomy. The instrumental variable regression result indicates that female employment significantly reduces the incidence of spousal violence. This result is mainly driven by the exposure reduction effect that dominates male backlash. There is, however, no evidence on the bargaining effect.

This is a preview of subscription content, access via your institution.

Fig. 1
Fig. 2
Fig. 3


  1. Aizer (2010) investigates the exposure reduction effect of a favorable labor market condition for women, but her study uses data in the United States.

  2. Although the two explanations—backlash and extraction—predict the same positive effect, the extraction effect is intrinsically different from the backlash effect. The backlash effect characterizes violence as expressive, whereas the extraction effect defines violence as instrumental. In this study, however, it is not possible to differentiate between the two effects.

  3. A few exceptions are the studies by Aizer (2010) and Luke and Munshi (2011). Aizer (2010) measures an exogenous change in female economic status by the ratio of wages in female dominated industries to wages in male dominated industries. Luke and Munshi (2011) instrument female and household income with land elevation and rainfall shocks. The main difference between their studies and the current study is that, while their studies identify the bargaining or backlash effects of female status, this study provides evidence on the exposure reduction effect.

  4. The list and definitions of the variables are provided in the notes below Table 1.

  5. If the rice–wheat ratio, which is defined by a rice area/(rice area + wheat area), in a state is greater than 0.5, it is defined as a rice state.

  6. The datasets include the Rural Economic and Demographic Survey 1998–1999 by the National Council of Applied Economic Research, Districtwise Area and Production of Principal Crops in India by Directorate of Economics and Statistics, and a rain-gauge-based 0.5° daily grid precipitation product by the APHRODITE project.

  7. Household level covariates cover variables such as woman age, man age, number of children, woman education, man education, number of men in the household, number of women in the household, a low-caste dummy, and wealth index. The wealth index is constructed using household asset data and principal components analysis. Assets include a number of consumer items such as a telephone, bicycle or car as well as availability of drinking water and sanitation facilities, etc. Each asset is assigned a score generated through principal components analysis and the scores are summed by each household (Kishor and Johnson 2004).

  8. Additional state level covariates are crop variables such as total crop area per capita in hectares, total crop area per capita interacted with rainfall shocks. Also included are some village level variables such as arable land per capita in 1,000 ha and the distance from the nearest town in 1,000 km.

  9. For example, studies in meteorology suggest that, other things being equal, rainfall decreases violence rates though cooling down effects (Simister and Van de Vliert 2005; Van de Vliert et al. 1999).

  10. In the data, only 0.42% of working women in capitals and large cities are engaged in the agricultural sector.

  11. Unfortunately, NFHS-2 does not provide data on male working status or household income. It does contain information on husbands’ usual occupations—answers to the question “What kind of work does (did) your (last) husband mainly do?”—But this information does not necessarily translate into the husbands’ working status during the past 12 months.

  12. The falsification test using households in capitals or large cities, unfortunately, does not directly disprove this concern. It is because male employment or household labor income in the non-agricultural economy are not expected to be affected by the instrument. Therefore, omitting the two variables would not serve as a potential source of bias in the control experiment, whereas it might do so in the main sample based on agricultural households.

  13. A similar finding has been made by Rosenzweig and Schultz (1982). In their study, while female employment is positively correlated with district normal rainfall, the rainfall has an insignificant effect on male employment.

  14. If the income contribution of a woman is relatively low, a change in female working status or work days might not necessarily translate into a significant change in the household labor income. Therefore, a significant change in female working status, but an insignificant change in the household labor income driven by the instrument can be because the female income contribution is relatively low.

  15. Women’s contribution to the household income is a categorical variable that takes 1 if the contribution is almost none, 2 if it is less than half, 3 if it is about half, 4 if it is more than half, and 5 if it is all.

  16. Non-agricultural occupations include professional jobs, technicians, management, clerical work, sales, and services.

  17. Although NFHS-2 does not provide a district level identifier, the districts can be identified based on Census of India 1981 and 1991. The district identifier is provided by Alessandro Tarozzi.

  18. Seven rice states include Andhra Pradesh, Assam, Bihar, Madhya Pradesh, Maharashtra, Orissa, and West Bengal. The other five states – Karnataka, Kerala, Manipur, Meghalaya, and Tamil Nadu – are dropped due to lack of crop area information at the district level.

  19. Similarly, the first stage regression result shows that more rain fall shocks in a district with higher rice-wheat ratio improve the probability of female labor force participation. This result is not reported for brevity.

  20. In this analysis, I use the categorical variable for female income contribution explained in footnote 15, and additionally include women who do not work by assigning 0 to them. The instrument for the female contribution is again the interaction between rainfall shocks and the rice–wheat dichotomy at the state level. The assumption behind the identification strategy is similar to the case of female working status, in that the degree of female income contribution exogenously increases when there are more rainfall shocks in the rice-growing area.


  • Aizer A (2010) The gender wage gap and domestic violence. Am Econ Rev 100(4):1847–1859

    Article  Google Scholar 

  • Bardhan P (1974) On life and death questions. Econ Polit Wkly 9(32-34):1293–1304

    Google Scholar 

  • Behrman JR, Foster AD, Rosenzweig MR, et al. (1999) Women’s schooling, home teaching, and economic growth. J Polit Econ 107(4):682–714

    Article  Google Scholar 

  • Bloch F, Rao V (2002) Terror as a bargaining instrument: a case study of dowry violence in rural India. Am Econ Rev 92(4):1029–1043

    Article  Google Scholar 

  • Boserup E (1970) Women’s role in economic development. Allen and Unwin, London

    Google Scholar 

  • Duflo E (2003) Grandmothers and granddaughters: old age pension and intra-household allocation in South Africa. World Bank Econ Rev 17(1):1–25

    Article  Google Scholar 

  • Dugan L, Nagin DS, Rosenfeld R (1999) Explaining the decline in intimate partner homicide: the effects of changing domesticity, women’s status, and domestic violence resources. Homicide Stud 3(3):187–214

    Article  Google Scholar 

  • Dugan L, Nagin DS, Rosenfeld R (2003) Exposure reduction or retaliation? The effects of domestic violence resources on intimate-partner homicide. Law Soc Rev 37(1):169–198

    Article  Google Scholar 

  • Farmer A, Tiefenthaler J (1996) Domestic violence: the value of services as signals. Am Econ Rev 86(2):274–279

    Google Scholar 

  • Farmer A, Tiefenthaler J (1997) An economic analysis of domestic violence. Rev Soc Econ 55(3):337–358

    Article  Google Scholar 

  • Goetz AM, Gupta (1996) Who takes the credit? Gender, power, and control over loan use in rural credit programs in Bangladesh. World Dev 24(1):45–63

    Article  Google Scholar 

  • Hornung CA, McCullough BC, Sugimoto T (1981) Status relationships in marriage: risk factors in spouse abuse. J Marriage Fam 43(3):675–692

    Article  Google Scholar 

  • India Meteorological Department (2005) A high resolution daily gridded rainfall for the Indian Region. CD-ROM

  • Indian National Science Academy, Chinese Academy of Sciences (2001) Growing populations, changing landscapes: studies from India, China, and the United States. National Academies Press, Washington

    Google Scholar 

  • Koenig MA, Ahmed S, Hossain MB, et al. (2003) Women’s status and domestic violence in rural Bangladesh: individual- and community-level effects. Demography 40(2):269–288

    Article  Google Scholar 

  • Kishor S, Johnson K (2004) Profiling domestic violence- a multi-country study. ORC MACRO, Calverton

    Google Scholar 

  • Luke N, Munshi K (2011) Women as agents of change: female income and mobility in India. J Dev Econ 94(1):1–17

    Google Scholar 

  • Lundberg S, Pollak RA (1994) Noncooperative bargaining models of marriage. Am Econ Rev 84(2):132–137

    Google Scholar 

  • Macmillan R, Gartner R (1999) When she brings home the bacon: labor force participation and the risk of spousal violence against women. J Marriage Fam 61(4):947–958

    Article  Google Scholar 

  • Mbiti I (2008) Monsoon wedding?: The effect of female labor demand on marriage markets in India. Mimeo Southern Methodist University

  • Miller BD (1981) The endangered sex: neglect of female children in rural North India. Cornell University Press, Ithaca

    Google Scholar 

  • Molm (1989) Punishment power: a balancing process in power-dependence relations. Am J Sociol 94(6):1392–1418

    Article  Google Scholar 

  • Pitt MM, Khandker SR (1998) The impact of group-based credit programs on poor households in Bangladesh: does the gender of the participant matter? J Polit Econ 106(5):958–996

    Article  Google Scholar 

  • Qian N (2008) Missing women and the price of tea in China: the effect of sex-specific earnings on sex imbalance. Q J Econ 123(3):1251–1285

    Article  Google Scholar 

  • Rosenzweig MR (1990) Population growth and human capital investments: theory and evidence. J Polit Econ 98(5):38–70

    Article  Google Scholar 

  • Rosenzweig MR, Schultz TP (1982) Market opportunities, genetic endowments, and intrafamily resource distribution: child survival in rural India. Am Econ Rev 72(4):803–815

    Google Scholar 

  • Senauer B, Garcia M, Jacinto E (1988) Determinants of the intrahousehold allocation of food in the rural Philippines. Am J Agric Econ 70(1):170–180

    Article  Google Scholar 

  • Simister J, Van de Vliert E (2005) Is there more violence in very hot weather? Tests over time in Pakistan and across countries worldwide. Pak J Meteorol 2(4):55–70

    Google Scholar 

  • Srinivasan S, Bedi AS (2007) Domestic violence and dowry: evidence from a South Indian Village. World Dev 35(5):857–880

    Article  Google Scholar 

  • Tauchen HV, Witte AD, Long SK (1991) Domestic violence: a non-random affair. Int Econ Rev 32(2):491–511

    Article  Google Scholar 

  • Van de Vliert E, Schwartz SH, Huismans SE et al. (1999) Temperature, cultural masculinity, and domestic political violence: a cross-national study. J Cross Cult Psychol 30(3):291–314

    Article  Google Scholar 

  • Vanneman R, Barnes D (2000) Indian District Data, 1961-1991: machine-readable data file and codebook. Internet address: Center on Population, Gender, and Social Inequality, College Park, Maryland

Download references


I am grateful to Andrew Foster, Mark Pitt, and Nancy Qian for their insight and guidance. I also thank Richard Blundell, Michael Conlin, John Giles, Todd Elder, Robert Pollak and Jeffrey Wooldridge for their comments and encouragement. Special thanks to Doug Park, Delia Furtado, Isaac Mbiti, Muna Miky, Nolan Noble, and Alessandro Tarozzi. I also appreciate valuable comments by Deborah Cobb-Clark and the two anonymous referees. All remaining errors are mine.

Author information

Authors and Affiliations


Corresponding author

Correspondence to Yoo-Mi Chin.

Additional information

Responsible editor: Deborah A. Cobb-Clark


Appendix 1

Table 7 Spousal violence experience of women in the past 12 months by work status

Appendix 2

Rural economic and demographic survey 1998–1999

Rural Economic and Demographic Survey (REDS) 1998–1999 is collected by the National Council of Applied Economic Research in India and covers about 7,500 rural households over 250 villages in 16 states of India. Out of 18 states included in the main sample, REDS has data on 14 states. The two states that are not included in REDS are Manipur and Meghalaya. The dataset contains detailed information on household agricultural labor and production, as well as demographic characteristics.

REDS is used in this study to show the rice–wheat dichotomy in female employment (Fig. 1). Further, it provides information on work days for agricultural laborers by gender and household agricultural labor income, both of which are not included in NFHS-2, but important for testing exclusion restrictions. For males, reported work days of 1,747 male agricultural laborers aged over 16 in landless households over 13 states are used for the regression result in column 6 of Table 2. The regression equation is defined by Eq. 2, except for the dependent variable being male work days. For females, reported work days of 267 female agricultural laborers aged over 16 in landless households over 10 states are used for the result in column 7 of Table 2.

In addition, household labor income is obtained by adding the wages of all of the household members who earn agricultural wages in a landless household. Again, the regression equation is the same as defined by Eq. 2, except for the dependent variable being household labor income. After dropping some households with missing variables on demographic variables or village level covariates, I find 606 households over 13 states. The estimation result is reported in column 5 of Table 2.

District wise area and production of principal crops in India by directorate of economics and statistics

District wise Area and Production of Principal Crops in India provides district-level crop data. The Directorate of Economics and Statistics, for the first time in 2001, published this brochure that includes comprehensive data on district wise area and production of principal crops in India for 1997–1998 and 1998–1999. I use 136 district-level crop data for seven rice states—Andhra Pradesh, Assam, Bihar, Madhya Pradesh, Maharashtra, Orissa, and West Bengal. The other five states in the rice-growing area are dropped, because their districts do not match the districts of NFHS-2.


Rainfall data for the district level analysis is a rain-gauge-based 0.5° daily grid precipitation product developed by the Asian Precipitation Highly Resolved Observational Data Integration Towards Evaluation of Water Resources (APHRODITE). The dataset contains interpolated rainfalls for a series of 0.5 × 0.5° grids covering Asia. I map all rainfall data into each district in India. Then, the district level rainfalls are calculated by averaging all the gridded rainfall information within one district. Rainfall shocks are defined by the difference of yearly rainfalls from the normal rainfall (30-year average) at the district level, scaled by standard deviation. Again, I use 136 district level rainfall shocks over seven rice states.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Chin, YM. Male backlash, bargaining, or exposure reduction?: women’s working status and physical spousal violence in India. J Popul Econ 25, 175–200 (2012).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:


  • Violence
  • Female employment
  • Exposure reduction

JEL Classification

  • J12
  • J16
  • J43