Publications

This paper estimates the direct and intergenerational effects of one of the world's largest school expansion policies. Starting in 1993-94, the District Primary Education Programme (DPEP) constructed over 100,000 new schools and served over 50 million children in India. Using a fuzzy regression discontinuity design I show that the policy increased enrollment, literacy and years of education for both male and female direct beneficiaries. Further, I find that children whose mothers were DPEP beneficiaries had higher scores on vernacular (18 percent), math (13 percent), and English (15 percent) tests, while Father's DPEP exposure had no effect on children's learning.

We conducted an experiment on a major international online freelancing platform to examine how increased flexibility in daily work hours affects female participation. We post identical job advertisements (for 320 jobs) covering a wide range of tasks (80 distinct tasks) that differ only in flexibility and the wage offered. Comparing the numbers of applicants for these jobs, we find that, while both men and women prefer flexibility, the elasticity of response for women is twice that for men. Flexible jobs attracted 24% more women and 12% more men than inflexible ones. Importantly, these increases did not compromise the quality of the applications. In contrast, there is suggestive evidence that flexible jobs attracted higher-quality female candidates. Our findings have significant implications for understanding gender disparities in labor market outcomes and for shaping equity-focused policies of organizations.

We analyze over 3,500 education-related articles published between 2002 and 2022 in the top 22 economics journals, and document striking geographic disparities. First, 85.7 percent of publications are based on high-income countries, with the United States alone accounting for 58.4 percent. Second, there is a strong positive income-publication gradient: wealthier countries receive more research attention. Third, countries with lower student learning levels are significantly underrepresented, even after accounting for GDP, population, and regional factors. Fourth, studies based on high-income countries are more likely to make external validity claims. We explore whether this bias is offset by development economics journals and find that while low- and middle-income country (LMIC) studies are more common there, even within LMICs, richer countries are disproportionately represented. Our findings suggest that structural features of knowledge production, including researcher location and funding concentration, contribute to persistent visibility gaps, with implications for equity, external validity, and global evidence-informed policymaking.

Obesity rates are rising around the world and so are its health and socio-economic costs. In order to effectively address the problem of obesity, it is important to understand the drivers of the disease. Education is one potential determinant and while there is a large body of literature on the causal impacts of schooling on obesity status, bulk of the existing evidence comes from developed countries. In this paper, we study the education gradients in adult female obesity in two African settings - Uganda and Zimbabwe. In the past, these countries have implemented large national school reform programmes that significantly enhanced educational attainment. We leverage age-specific exposure to these reforms to instrument for women's education and use a two stage least squares (2SLS) methodology to isolate the effects of education on later-life obesity. Despite evidence of positive and statistically significant education-obesity correlations, we cannot reject the null hypothesis that schooling has no overall causal impact on obesity in either study setting. We also find no causal effects on body mass index and being either overweight or obese. The significant cross-sectional associations between education and obesity are evidently driven by correlations between schooling and unobserved factors influencing both schooling and obesity status.

Violent crime profoundly affects several socioeconomic outcomes. But does day-to-day crime also shape human capital accumulation? We answer this question in the Indian context by combining multiple years of district-level crime data with a representative survey on learning achievement of school-aged children. Our empirical strategy leverages the district-year variation in crime to estimate the crime-learning gradient. We find that an increase in violent crime is associated with lower achievement in reading and math, while non-violent crimes have no discernible correlation with learning outcomes. These associations hold similarly for boys and girls. Further, children from socioeconomically disadvantaged households show stronger negative associations. We document evidence that both household-level factors (reduced child mobility, lower adult engagement in work, and poorer mental health) and school-level factors (lower availability of teachers) are possible mechanisms underpinning these findings.

Due to almost-universal enrolment in primary schools, policy focus has shifted towards improving learning outcomes. One important way of doing this is to enhance accountability, especially in the case of public provision of education. In this context, we examine the effectiveness of two different strategies of increasing accountability - one involving only the community, and the other which builds collaboration between the schools and the community. We implement a randomized controlled trial in 400 villages in India, and find: (i) both interventions led to a significant enhancement in children's foundational literacy and numeracy skills, (ii) we observed limited differences between the impacts of the two interventions, and (iii) the community-school intervention exhibited significantly greater effects when parents reported visiting the school, underscoring the vital role of parent-teacher interactions and their shared responsibility in shaping children's learning outcomes. In terms of mechanism, we find that direct learning inputs play a major role in mediating the observed effects of both interventions. Additionally, parent-teacher engagement and children's studying habits outside of the school are potential important channels through which the observed effects operate in the community-school intervention.

We leverage the exogenous variation in education induced by the implementation of a national compulsory schooling law (CSL) in China in 1986 to study the mean and heterogeneous effects of education on mental health. Regression discontinuity (RD) estimates suggest that on average CSL beneficiaries had better mental health and lower probability of being severely depressed. We combine the RD design with novel distributional analysis methods to demonstrate that this average effect is largely driven by improvements in the top half of the mental health distribution (higher scores indicating worse mental health). These findings not only add to the scant evidence on the effect of education on mental health in low- and middle- income contexts, but also suggest that looking beyond average effects might better inform how policies can be targeted to enhance their benefits. In terms of potential mechanisms, we find that CSL beneficiaries experienced better physical health, labor market outcomes and marital outcomes.

Although less than a third of the population in developing countries is covered by health insurance, the number has been on the rise. Many countries have implemented national insurance policies in the past decade. However, there is limited evidence on their impact on child mortality in low- and middle-income contexts. Here we document the child mortality reducing effects of an at-scale national level health insurance policy in India. The Rashtriya Swasthya Bima Yojana (RSBY), was rolled out across India between 2008 and 2013. Leveraging the temporal and spatial variation in program implementation, we demonstrate that it lowered infant mortality by 6% and child under five mortality by 5%. The effects are largely concentrated among urban poor households. In terms of mechanisms, we find that the program effects seem to be driven by increased usage of reproductive health services by mothers. We also demonstrate a rise in usage of complementary health services that were were not covered under the policy (such as child immunizations), which suggests that RSBY had significant positive spillover effects on health care usage.

We use a production function approach to investigate the factors that determine young adults' human capital outcomes in Madagascar and Senegal. Our study relies on unique and comparable panel data spanning over 15 years for both countries. We find that second-grade students' test scores are strong indicators of their French and math skills, as well as their educational attainment in their early twenties. Moreover, we observe that the association between second-grade skills and later-life outcomes is stronger among girls than boys, and in math test scores compared with French test scores. Our results emphasise the critical role of performing well during the early school years, as it predicts long-term outcomes, especially for vulnerable populations such as girls.

We analyze the role of education as a determinant of mental health. To do this, we leverage the age-specific exposure to an educational reform as an instrument for years of education and find that the treated cohorts gained more education. This increase in education had an effect on mental health more than 2 decades later. An extra year of education led to a lower likelihood of reporting any symptoms related to depression ( 11.3% ) and anxiety ( 9.8% ). More educated people also suffered less severe symptoms - depression ( 6.1% ) and anxiety ( 5.6% ). These protective effects are higher among women and rural residents. The effects of education on mental well-being that we document are potentially mediated through better physical health, improved health behavior and knowledge, and an increase in women's empowerment.

This study uses household-level panel data from a nationally representative survey to estimate the effect of agricultural productivity shocks, as proxied by exogenous annual rainfall deviations, on education expenditures and children's work status in rural India. We find that a transitory increase in rainfall significantly reduces education expenditures and increases the likelihood of child labor across multiple work activities. Additionally, households owning land and those with better credit access increase the use of child labor as rainfall increases because labor (and land) markets are incomplete. The effects of productivity shocks are reinforced for marginalized castes and for less educated households, thereby exacerbating inequalities in education.

This paper contributes to the relatively scant literature on the impacts of inequality on the efficacy of public works programmes. We study this in the context of India. In particular, we examine the effect of land inequality on the implementation of the world's largest workfare programme - the National Rural Employment Guarantee Act (NREGA). Our OLS estimates demonstrate that the concentration of land ownership reduces the efficacy of NREGA. An instrumental variable (IV) analysis, where we use the historical land tenure system as an IV for contemporaneous land inequality, further corroborates our findings. This negative relationship is consistent with the hypothesis that public work schemes raise agricultural wages in the private labour market, thereby incentivising big landlords to use their political power to oppose such programmes. We exclude the possibility that the higher provision of public jobs in more equal areas is driven by a higher demand for public jobs or by caste or religious differences. This study suggests that the concentration of land ownership, a proxy for power asymmetries, could hinder effective implementation of development policies.

In this paper we study the long run effects of the 1959-61 Chinese Famine on mental health outcomes. We focus on cohorts that were born during the famine and examine their mental health as adults, when they are roughly 55 years of age. We find that early-life exposure to this famine leads to a large statistically significant negative impact on women's mental health, while there is limited effect on men. This gender differential effect is observed because male fetuses experience a stronger natural selection as compared to female fetuses, which implies that in the longer run, surviving females may exhibit larger detrimental effects of early-life famine exposure. Thus, the observed effects are a composite of two well-established factors, the survival of the fittest and the Fetal Origins hypothesis.

This study examines whether Ugandan women who marry at younger ages fare differently on a wide range of later-life outcomes than women who marry at later ages. Using a nationally representative data set, I identify the plausibly causal impacts of women's marriage age by using age at menarche as an instrumental variable. Results indicate that a one-year delay in marriage for Ugandan women leads to higher educational attainment (0.5-0.75 years), literacy (10 percentage points), and labor force participation (8 percentage points). I also explore intergenerational effects of later marriage and find that the children of mothers who marry later have higher BMI (0.11 kg/m2) and hemoglobin levels (0.18 g/dl), and they are also less likely to be anemic (4 percentage points). Finally, I present evidence suggesting that the observed effects might be mediated through an enhancement of women's agency within their household and positive assortative matching in the marriage market. By pointing to the beneficial consequences of delaying marriage, this research calls for concerted policy action to prevent child marriage.

A nutrition-sensitive food system is one that goes beyond staple grain productivity and places emphasis on the consumption of micronutrient-rich nonstaples through a variety of market and nonmarket interventions. A nutrition-sensitive approach not only considers policies related to macrolevel availability and access to nutritious food, but it also focuses on householdand individual-level determinants of improved nutrition. In addition to agriculture, intrahousehold equity, behavior change, food safety, and access to clean water and sanitation are integral components of the food system. This article provides a detailed review, from an economic perspective, on the multisectoral pathways through which agriculture influences nutrition. A critical challenge is to identify and implement food and nutrition policies that are appropriate to the particular stage of structural transformation in the country of concern.

Background Progress in malaria control has stalled in recent years and innovative surveillance and response approaches are needed to accelerate malaria control and elimination efforts in endemic areas of Africa. Building on a previous China-UK-Tanzania pilot study on malaria control, this study aimed to assess the impact of the 1,7-malaria Reactive Community-Based Testing and Response (1,7-mRCTR) approach implemented over two years in three districts of Tanzania. Methods The 1,7-mRCTR approach provides community-based malaria testing via rapid diagnostic tests and treatment in villages with the highest burden of malaria incidence based on surveillance data from health facilities. We used a difference-in-differences quasi-experimental design with linear probability models and two waves of cross-sectional household surveys to assess the impact of 1,7-mRCTR on malaria prevalence. We conducted sensitivity analyses to assess the robustness of our results, examined how intervention effects varied in subgroups, and explored alternative explanations for the observed results. Results Between October 2019 and September 2021, 244,771 community-based malaria rapid tests were completed in intervention areas, and each intervention village received an average of 3.85 rounds of 1-7mRCTR. Malaria prevalence declined from 27.4% at baseline to 11.7% at endline in the intervention areas and from 26.0% to 16.0% in the control areas. 1,7-mRCTR was associated with a 4.5-percentage-point decrease in malaria prevalence (95% confidence interval: − 0.067, − 0.023), equivalent to a 17% reduction from the baseline. In Rufiji, a district characterized by lower prevalence and where larviciding was additionally provided, 1,7-mRCTR was associated with a 63.9% decline in malaria prevalence. Conclusions The 1,7-mRCTR approach reduced malaria prevalence. Despite implementation interruptions due to the COVID-19 pandemic and supply chain challenges, the study provided novel evidence on the effectiveness of community-based reactive approaches in moderate- to high-endemicity areas and demonstrated the potential of South-South cooperation in tackling global health challenges.

Investing in nutrition sensitive sectors such as education can be an effective strategy for combatting malnutrition. In this paper we analyze the role that men's education plays in determining dietary diversity outcomes using primary data from 3600 households across four districts of India. Dietary diversity scores were calculated to reflect the quality of food intake, for households and women. Men's education level was considered as the primary driver of diet diversity. To establish a causal link between men's education and diet diversity, the education level of parents and siblings were used as instrumental variables. We find that men's education levels are associated with significantly higher diet diversity scores both for the household and for women. The role of men's education continues to be a significant determinant of diet quality after controlling for household and individual- level confounding factors including the education level of the woman. The results are consistent across different definitions of the diet diversity score and reference period. Methodologically we extend the evidence on the education-nutrition pathway from being associational to causal in nature. Results from this study point to the benefits of leveraging men's education as an effective pathway for improving nutritional outcomes within households.

Teachers' perceptions of their students' academic skills are crucial in enabling teaching at the appropriate level and shaping learning processes. We examine how accurate are teachers' perceptions for low achieving students using data collected from around 1,800 primary school teachers across 848 schools in rural India. We find that around 40% of teachers inaccurately perceive that the low performing students in their classroom had already acquired foundational literacy, when they have not. Female teachers, para-teachers, and teachers with lower work burden are more likely to have accurate perceptions of low performing students. Our study highlights the need to work with teachers to ensure more realistic perceptions in order for pedagogical approaches like teaching at the right level to reach its intended impact.

India's mass vaccination efforts have been slow due to high levels of vaccine hesitancy. This study uses data from an online discrete choice experiment with 1371 respondents to rigorously examine the factors shaping vaccine preference in the country. We find that vaccine efficacy, presence of side effects, protection duration, distance to vaccination centre and vaccination rates within social network play a critical role in determining vaccine demand. We apply a non-parametric model to uncover heterogeneity in the effects of these factors. We derive two novel insights from this analysis. First, even though, on average, domestically developed vaccines are preferred, around 30% of the sample favours foreign-developed vaccines. Second, vaccine preference of around 15% of the sample is highly sensitive to the presence of side effects and vaccination uptake among their peer group. These results provide insights for the ongoing policy debate around vaccine adoption in India.

Background: It is widely considered that women have less diverse diets than other household members. However, it has been challenging to establish this empirically since women's diet diversity is measured differently from that of other household members. Objective: In this article, we compare women's dietary diversity with that of their respective households and thereby generate a measure of "dietary gap." Methods: We measure women's "dietary gap" by using the difference of homogenized household and woman dietary scores (using the same scales). This is done using primary data on 3600 households from 4 districts in India. Additionally, we show the robustness of our results to variations in scale and recall periods used to construct the diet diversity scores. Results: Mean difference tests indicate that women consistently consume 0.1 to 0.5 fewer food groups relative to other household members, with the results being statistically significant at the 1% level. The food groups driving this dietary gap are nonstaples like Vitamin A-rich fruits and vegetables, meat/fish/poultry, and dairy. Conclusions: Results point toward the discrimination faced by women in the variety of the food consumed, the importance of considering comparability in creating indices of diet diversity, and the need to collect more detailed information on diets. To our knowledge, this is one of the first studies to examine dietary discrimination faced by women using common scales.

Background: Recent literature, largely from Africa, shows mixed effects of own-production on diet diversity. However, the role of own-production, relative to markets, in influencing food consumption becomes more pronounced as market integration increases. Objective: This paper investigates the relative importance of two factors - production diversity and household market integration - for the intake of a nutritious diet by women and households in rural India. Methods: Data analysis is based on primary data from an extensive agriculture-nutrition survey of 3600 Indian households that was collected in 2017. Dietary diversity scores are constructed for women and households is based on 24-hour and 7-day recall periods. Household market integration is measured as monthly household expenditure on key non-staple food groups. We measure production diversity in two ways - field-level and on-farm production diversity - in order to account for the cereal centric rice-wheat cropping system found in our study locations. The analysis is based on Ordinary Least Squares regressions where we control for a variety of village, household, and individual level covariates that affect food consumption, and village fixed effects. Robustness checks are done by way of using a Poisson regression specifications and 7-day recall period. Results: Conventional measures of field-level production diversity, like the number of crops or food groups grown, have no significant association with diet diversity. In contrast, it is on-farm production diversity (the field-level cultivation of pulses and on-farm livestock management, and kitchen gardens in the longer run) that is significantly associated with improved dietary diversity scores, thus suggesting the importance of non-staples in improving both individual and household dietary diversity. Furthermore, market purchases of non-staples like pulses and dairy products are associated with a significantly higher dietary diversity. Other significant determinants of dietary diversity include women's literacy and awareness of nutrition. These results mostly remain robust to changes in the recall period of the diet diversity measure and the nature of the empirical specification. Conclusions: This study contributes to the scarce empirical evidence related to diets in India. Additionally, our results indicate some key intervention areas - promoting livestock rearing, strengthening households' market integration (for purchase of non-staples) and increasing women's awareness about nutrition. These are more impactful than raising production diversity.

The Multinomial Logit (MNL) model is popular, but a semi-parametric specification of its link/utility function has seldom been used in empirical applications. This is primarily because of the resource intensive nature of semi-parametric estimation. In this paper we propose and implement a parallel computation algorithm to estimate the semi-parametric kernel MNL model. This algorithm reduces model estimation time by a factor of 2-10, depending on the size of the dataset and the available resources for computation. These computational gains make the estimation of this model feasible for large datasets. Additionally, using a Monte Carlo study we show that the kernel MNL outperforms the traditional linear MNL model in terms of fit and predicted choice probabilities. We demonstrate how kernel-based specification can unearth important heterogeneities in the effect of covariates through an empirical exercise. We use data from a nationally representative household survey (N = 157,804) to analyze the factors associated with institutional births (as opposed to home births) in India. Our revealed-preference results indicate that maternal education, household assets, distance to formal health facility, and birth order play an essential role in determining birth location choice. Although the directions of impact are similar across both the linear and the kernel MNL specifications, there are significant differences in the marginal effects of different factors across the two models. These differences, which arise due to the flexibility afforded by the semi-parametric specification, potentially bring additional nuance to policy discussions.