Research

              

   Journal Publications


[Abstract] How does religious institutions bear upon the gender-gap in time use for individuals? In this paper we examine how religious mandates of the holy month of Ramadan affects the gendered distribution of time use within Muslim households in India. Using rich data on time use on 56 activities from a nationally representative time use survey we construct a composite measure of gendered life that incorporates gender gaps. We employ a difference-in-difference methodology to test if this gender imbalance in accentuated by Ramadan. We find that contrary to popular belief, Ramadan moderates the gender disparities in intra-household time use of Muslim household. This moderating influence is stronger in districts with a higher Muslim proportion. We find a reduction in absolute gender gap in time use for employment, unpaid domestic work and learning. However at an individual level, men and women in different age groups are differentially affected by Ramadan.

[Abstract]  Expansion in access to public infrastructure can have varied, microlevel impacts. In this paper, we use a discrete and quasi-random change in the access to paved roads through a large-scale rural road construction program in India to study how road access impacts fertility decisions and investments in child health. We find that increased access to paved roads at the district level decreases fertility, improves investments in children, and lowers infant mortality. We also provide evidence that highlights the mechanisms that drive this effect. First, we show that local roads improve access to health care facilities and raise immunization rates, which reduces infant mortality. Then, we demonstrate that last-mile road connectivity has contrasting effects on employment across gender. Overall, the evidence suggests that rural roads can help accelerate demographic transition through their effects on fertility and infant mortality.

[Abstract] Bans on sex-selective abortions are typically implemented to make sex ratios more equitable, but they may have adverse effects on surviving children. We examine the impacts of a ban on prenatal sex selection in India on postnatal health outcomes. We first show that the ban increased the share of female children born to mothers, especially among firstborn female families. Strikingly, we also find that the ban led to a worsening of mortality outcomes for both girls and boys in firstborn female families. In terms of mechanisms, we find that fertility increases in firstborn female families after the ban, pointing to the following channel: firstborn female families are disproportionately affected by the ban and are more likely to use the son-biased fertility stopping rule to achieve a desired number of sons. Children in firstborn female families likely face greater competition for parental resources, which may worsen their health.

[Abstract] Policy measures that seek to address son preference through restrictions on the tools of sex-selective abortions, without addressing the underlying causes, have been found to generate negative welfare consequences for unwanted surviving girls. Unlike these top-down supply-side measures, demand-side measures that focus on increasing the demand for girls by shifting social norms of son preference can mitigate these adverse welfare consequences. We study the impact of an intervention aimed at reducing discrimination against girls, which has both supply-side and demand-side elements. The intervention, implemented in India between 2015-18 included a mass media campaign designed to increase the perception of the value of a female child, while also tightening the policing of illegal sex-selective abortions. We exploit variation in the timing of exposure to the programme across Indian districts as well as quasi-exogenous variation in the sex of the firstborn child to identify the impact of the programme and find that it led to an increased proportion of female births as well as a reduction in the gender gap in mortality in intensively treated families. The main mechanism that explains our results is a relative increase in health investments in daughters, such as breastfeeding and vaccinations.

[Abstract] This paper uses the state-level rollout of beef possession/sale bans in India as a natural experiment to study their impact on women’s health. We hypothesize that beef possession/sale bans lead to increased anemia rates for women in traditionally beef consuming groups as beef is one of the cheapest sources of heme-iron. We compile rich historical data on legislation on beef possession/sale bans by states. Exploiting the intertemporal and spatial variation in these laws we compare hemoglobin levels of women in beef consuming groups with that of non-beef consuming groups. Using a triple difference estimation framework, we find that beef bans reduce women’s hemoglobin, and increase the incidence of severe anemia by around 27% in communities that traditionally eat beef. Our results point out severe negative implications of these bans on public health. 

[Abstract] We examine the effect of disease environment on child mortality and fertility response using evidence from a massive public health intervention in India. In particular, we use the malaria eradication program in India during the 1950's that caused a dramatic fall in malaria incidence in India. Using cohort level variation in exposure to the program we find a significant decline in infant and neonatal mortality in the post intervention period for high malaria endemic districts. Second, we find this has a bearing on the fertility outcomes as measured by the probability of birth. We find that the improvement in disease environment removes a biological constraint for early pregnancy and reduces mother's age of first birth in effect. This finding has significant implications in understanding the dynamics of demographic transition and human capital measure in a population post disease eradication. We reconcile the apparent puzzle of a null effect of malaria eradication on human capital outcomes in the literature and help explain the weak relation between epidemiological transition and growth. Our finding pushes the literature forward in highlighting the role of biological constraint in the relationship between disease environment and fertility response.

[Abstract] This paper examines the unintended consequences of a policy aimed at improving the groundwater level on crop residue burning in India. The Preservation of Subsoil Water Act, 2009 implemented in two Indian states bans the transplantation of paddy before mid-June to preserve groundwater. Theoretically, this leaves a short window of time for clearing the field before the next crop and thus increases the likelihood of farmers adopting time-saving methods like crop residue burning. Using a difference-in-difference framework we find that the ban results in both delay and an increase in crop residue burning in the winter months. 

[Abstract] What explains the low enrolment of women in science and economics? We combine administrative and survey data on a sample of high-income workers in India who have completed the same elite graduate programme to estimate the return to studying different undergraduate degrees - across science, business and economics, and the humanities. We find evidence of a large earnings premium to studying science and economics, yet disproportionately low female enrolment in these subjects. Using data on the subjective expectations of undergraduate students who are in the process of selecting a major, we model major choice as a function of major-specific and job-specific attributes. We identify significant gender differences in the preferences for different attributes as well as in the expectations of future outcomes, especially of grades. Women are willing to pay twice as much as men for course enjoyment and higher grades, even as they expect lower grades in science and economics. This suggests that in addition to pervasive norms about which subjects are better suited for women, women also suffer from a relative confidence gap in their major-specific abilities.

[Abstract] In this paper we examine to what extent social identity of the student and teacher plays a role in the learning outcome of the students. Specifically, we ask if a student benefits by learning from a teacher of the same gender. Past literature has highlighted ‘role model’ effects to explain an interaction effect. We broaden the literature in explaining the gender-interaction effect through gender based sorting across private and public schools in a developing country setting. Our results are driven by two critical differences between male and female members. The gender sorting of teachers originates from the differential opportunity costs of teaching at remote locations for male and female teachers. For students, the difference between male and female members comes from the difference in return to their human capital investment by parents –where for girls, a lower fraction of the return comes to their parental families under patrilineal and patrilocality norms. These factors create a sorting pattern which leads to interaction effect of gender matching. We confirm our theoretical results using information from detailed household and school level learning outcomes from Young Lives Survey, a panel data from Andhra Pradesh, India.

[Abstract] There is a strong political opinion in India in favour of replacing caste based affirmative action with an economic class based one. We contribute to this debate by looking at the interaction of caste and wealth in school choice. We show that too rich and too poor parents behave in the same way irrespective of their caste identities—rich parents sending their children to private schools while poor parents choosing public schools for their children. The caste identity, we find, plays a role for the school choice decision made by the parents belonging to the economic middle class. Among the economic middle class parents, the ones from the privileged castes send their children to private schools, while the children of the parents from the disadvantaged castes are sent to public schools. The result is robust to alternative definitions of privileged and disadvantaged castes. For school quality choice, however, we find a monotonic relationship between wealth and school quality.

[Link to Publication] [Download PDF]

 [Abstract] This paper studies systematic reporting heterogeneity in self-assessed health in India using World Health Survey (WHS)-SAGE survey that has subjective assessments on own health and hypothetical vignettes as well as objective measures like measured anthropometrics and performance tests on a range of health domains. The study implicitly tests and validates the assumption of response consistency in a developing country setting, thus lending support to the use of vignettes. Additionally, we are able to control for unobservable heterogeneities of reporting behavior at the individual level by employing individual fixed-effects estimation using multiple ratings on a set of vignettes by the same person. The study confirms identical pattern of systematic bias by the socioeconomic subgroups as is indicated by vignette technique. It further highlights that substantial amount of reporting heterogeneity remains unexplained after controlling for the usual socioeconomic control variables. The finding has potentially broader implications for research based on self-reported data in a developing country.

[Abstract] The study examines the role of the largest public works program in the world—the National Rural Employment Guarantee Scheme (NREGS) in buffering the negative effects of early childhood exposure to rainfall shocks on long-term health outcomes. Exploiting the spatial and temporal variation in NREGS coverage, the study estimates the extent to which nutritional shocks in early childhood can be offset by access to the program. The study employs a unique identification strategy by integrating detailed administrative records of drought shock and phased roll-out information of NREGS with household-level panel data—the Young Lives survey—conducted over three waves (2002, 2007, and 2009–10) in the state of Andhra Pradesh, India. Using individual fixed effects estimation, the study finds that while the policy does not help correct for long-term past health deficiencies it is useful in buffering recent drought shocks, which vary by policy relevant subgroups.

[ Abstract] This paper exploits the closure of senior secondary schools in urban China from 1966 to 1971 in order to identify the causal intergenerational transmission effects of education. The paper uses the instrumental variable approach to examine the intergenerational causality of educational transmission at the senior secondary schooling level in urban China. The exogenous variation in parental senior secondary educational attainment both over time and across regions allows us account for selection bias and thus identify the causal intergenerational transmission effect in education. We further show that our conclusion is robust to alternative identification strategies and data sets.


[Abstract] India’s efforts on HIV control targeted at high-risk groups, including persons who inject drugs (PWID), have largely remained unevaluated. This paper presents the cost-effectiveness of a WHO-recommended comprehensive harm reduction programme implemented among a cohort of PWIDs in Delhi, India. We estimate the number of infections averted using Bernoulli process model and calculate the cost-effectiveness ratio as the total programme cost per infection averted. The intervention averted 996 HIV infections over 36 months. This was higher during the first phase of implementation when HIV testing and counselling were initiated, than the second phase when the full intervention package, including needle-syringe exchange, condom distribution, harm reduction education etc., was implemented. The project spent an amount of US$ 1,026,361.37 during the entire intervention period, considering only the programme costs. The cost-effectiveness ratio has been estimated at US$ 698.99 per infection averted per year. This paper provides important policy insights for India’s targeted HIV intervention programme. We show that implementation of the WHO-recommended comprehensive package of harm reduction would be cost-effective for the Indian programme targeted to PWIDs, wherein focusing on HIV testing and counselling would be an important first step.

 [Abstract] WHO, UNODC, and UNAIDS recommend a comprehensive package for prevention, treatment, and care of HIV among people who inject drugs (PWID). We describe the uptake of services and the cost of implementing a comprehensive package for HIV prevention, treatment, and care services in Delhi, India. This paper documents the implementation of a comprehensive HIV prevention and care program in a large urban city and provides important insights into the challenges and the facilitators for the program. 


[Abstract] This review presents an analytical framework to discuss the existence and forms of gender discrimination in health. We survey the evidence of gender discrimination in health at the various life stages of an individual. We discuss the underlying drivers of this discrimination and its relationship with economic growth and development. Lastly, the review outlines the welfare implications of gender discrimination in health and offers policy insights. 

[Abstract] In this review we discuss some of the factors that have led to male-biased sex ratios in large parts of the developing world. We discuss the recent efforts to quantify the number of “missing women”.  We explore proximate causes that drive household behaviour and result in male-biased sex-ratios, as well as deeper sociocultural drivers of son preference that leads to discrimination against females. We also examine the welfare consequences of a number of government policies that have been implemented to address this gender imbalance. 



[Abstract] How does gender-specific transport subsidy for urban women affect their participation in the labour force? To answer this question, we study the Government of Delhi's pink pass scheme that made bus travel free for women since October 2019. Using the Time Use Survey of 2019 and employing two alternative difference-in-difference (DID) strategies and a triple difference (DDD) approach, we find weak evidence that the policy increased paid work and employment of women at the intensive margin by 30 to 50 minutes. Strikingly, we find that women from economically marginalized households experience large and statistically significant increases in paid work and employment, both at the extensive and intensive margins. The pink pass scheme increases employment of women from economically marginalized group by 24 percentage points at the extensive margin and by 150 minutes at the intensive margin. Our study has important insights for policies addressing supply-side bottlenecks in improving female participation in paid work. 

[Abstract] How fertility responds to the changes in the disease environment has been an important question for understanding the dynamics of demographic transition. We investigate this question in the context of the COVID-19 pandemic in India. We harmonise district-wise lockdown intensity data in the country with detailed records on fertility and child health outcomes from the latest rounds of the National Family Health Survey (NFHS) which were conducted between 2015-16 and 2019-2021. We leverage the intensity of the nation-wide mobility restrictions in a difference-in-difference framework to estimate the impact of the pandemic on fertility behaviour. Strikingly, we find a significant increase in the probability of higher order births while a decline in the probability of first order births in the country, with the highest reduction in the most restrictive zone. We find the rise in fertility for higher order births is prominent in the economically poorer households from rural locations in states with higher fertility in the baseline. We explore potential mechanisms and a significant decrease in Long-acting reversible contraception (LARC), post pandemic, which aligns with our main finding. Our results have important implications for understanding fertility behaviour and human capital investments in the face of shocks.

[Abstract] It is a well-established fact that the daily lives of individuals in South Asia are a function of the gender in which they were born. In the academic literature, there has rarely been an attempt to quantify the the extent to which daily lives within households are gendered and how they differ on the basis of geographical region, religion or caste. Social scientists documenting gender inequities in 'work' have most often singularly focused on employment and documented the low and declining female employment rates and heterogenieties associated with it. Though less frequently, some economists have also documented gender gaps in domestic work. Such a focus on employment and/or domestic work implicitly suggests a hierarchy of activities in which employment ranks higher than other activities (like care giving activities or voluntary work or self care or cultural pursuits). In a society that values all activities equally and values a non-gendered life within a household, gaps in any activity should be of concern irrespective of the direction of such gaps. In this paper, borrowing from concepts of distance between vectors in an Euclidean space, we use simple measures (gender distance metrics) that measure the extent to which time allocation within households are gendered.

[Abstract] In this paper we examine how traditional institutions like caste interact with socio-economic status to mediate the perception of gender roles and attitudes around female labour force participation. We use third party vignettes to directly test the validity of the hypothesis that lower castes have more egalitarian gender norms and lower acceptance of restrictions on female autonomy. We find that the relationship between conservative gender norms and caste are in turn influenced by the class status of households, measured by land or asset ownership. Lastly, we conduct a simple social experiment to test for 'pluralistic ignorance' and confirm the presence of systematic overestimation of conservative attitude that varies by caste and class identities.

[Abstract] There is mixed evidence in the literature on the effect of rainfall shocks on educational outcomes for children in rural areas, with a limited understanding of how the gender-gap in education evolves in the face of such a shock. We posit that the vulnerability to climatic shocks can vary by the gender institutions of the setting which can have a bearing on the gender-gap in educational outcomes. On one hand, a negative productivity shock can lead to a disproportionate reduction in human capital outcomes for girls, as investments for girls may be more sensitive to income constraints. On the other hand, as the opportunity cost of schooling goes down in the face of a negative shock, it can translate into gains in educational outcomes, which are higher for female children in areas that favour female labour force participation. Leveraging the variation in cropping patterns that guide norms around female labor force participation (FLFP) in rural India, we examine how exposure to contemporaneous and past rainfall shocks affects learning outcomes for girls and boys. We find the widest gaps in outcomes in positive versus negative rainfall shock years for female children in regions that favour FLFP. We provide suggestive evidence that this is driven by increased participation in paid employment and full time domestic work during a positive rainfall shock. 

[Abstract] In this paper, we study the impact of  policy induced teacher transfers in government primary schools in India on  student learning outcomes. We utilize a government policy implemented in an Indian state in 2016 providing us with an exogenous variation in teacher transfers. We leverage the variation in the intensity of the program  to employ a difference-in differences estimation framework using a rich panel data on primary schools and find a significant negative impact of the teacher transfers program on student learning outcomes. In terms of mechanisms, we find that schools with larger transfers lose greater number of teachers, making pupil teacher ratio worse and also losing more qualified teachers. 

Selected Ongoing Projects


Open burning of crop residue (straw) is a common practice in many developing countries and leads to air pollution which has many adverse effects on health. Crop residue burning (CRB) is widely held responsible for the air pollution problem in India, particularly in North India. However, there is scant evidence on the determinants of CRB. In this paper, we look at the political economy of CRB; in particular, whether CRB incidence is related to electoral cycles in India. Using the fact that state elections in India are not perfectly synchronized, we test whether temporal proximity to the year-month of state legislative assembly elections influences the extent of crop residue burning. For our analysis, we put together a unique dataset on CRB using 1-km resolution disaggregated daily data from NASA’s remote sensing products. We aggregate this dataset at the state electoral constituency level and match it with election data to form an all-India dataset. We find that the incidence of CRB significantly increases before elections. Our findings suggest that political incentives are at work behind the continued open burning of crop residue. We also provide evidence that the pre-electoral increase in crop burning is unlikely to be driven by possible increases in crop production before the elections. Thus, the mechanism appears to be a more relaxed implementation of laws before elections for political gain.

        Popular press and other writing  

    Research Grants