Research
Research
Do ‘Women Hold Up Half the Sky’ in Fuel Choices? Revisiting the role of Adolescent Empowerment in a Developing Country Context, (with Somdeep Chatterjee), World Development, Vol 202, 107355, 2026
In this paper, we study the impact of a multidimensional women’s empowerment programme in India, known as SABLA, on the choice of cooking fuel within the household. The programme, launched in 2010 as a pilot project in 205 districts of the country, provided support to women across a multitude of dimensions. We exploit plausibly exogenous geographic and cohort variation generated in the implementation of the SABLA programme using a difference-in-difference identification design. The results of our study, based on this quasi-experimental methodology and using data on 21 states of India from the DLHS-4 survey, suggest that women potentially exposed to the SABLA empowerment programme were more likely to use cleaner and efficient modern cooking fuels post-marriage. This high usage of modern cooking fuels is matched by a fall in the likelihood of using inefficient transitional and traditional cooking fuels. We also show that women potentially exposed to the programme now exhibit improved health behaviors in the context of their children’s as well as their own health. These improved health behaviors, when combined with the empowering effects of the programme in other domains, as evidenced by the literature, serve as potential mechanisms for our effects. The results of our study remain robust to a variety of sensitivity checks and analyses.
Access to Technology and Foundational Math Proficiency among Students: Empirical Evidence from India, (with Valentina Rotondi and Ridhi Kashyap), Humanities & Social Sciences Communications (Nature) 12, 1015, 2025
Featured: Scienmag (Science Magazine)
Digital resources such as laptops have the potential to improve access to educational resources and provide personalized and uninterrupted learning opportunities for students. The impacts of these technologies may be especially salient in contexts where classroom sizes are large and schooling quality poor. Here, we study the impacts of access to laptops on foundational math proficiency of students in one such context, i.e. India, exploiting plausibly exogenous variation in the implementation of the Tamil Nadu Free Laptop Scheme (TFLS). Introduced in 2011, the TFLS was one of the largest and targeted free laptop programmes in the world distributing over 5 million laptops. Using data from the Annual Status of Education Report (ASER) and India Human Development Survey (IHDS) within a double and a triple-difference design, we show positive effects of access to laptops on foundational math proficiency of students, with the largest improvements experienced those in economically disadvantaged households. We further find that these positive effects on foundational math proficiency are complemented by improvements in other education related outcomes of students, such as, more time spent on learning, better comprehension of language, and a reduction in private tuition. We provide evidence that laptops are able to close economic and gender divides in education. Our results remain robust to a variety of falsification and sensitivity checks.
Women's Empowerment and Intimate Partner Violence: Evidence from a Multidimensional Policy in India, (with Somdeep Chatterjee), Economic Development and Cultural Change, Vol 72 (2), 2024
In this paper, we study a multidimensional women’s empowerment program from India to estimate its effects on intimate partner violence faced by women. We exploit plausibly exogenous geographical variation in the implementation of the program and cohort-variation generated by eligibility rules to estimate our effects. Using a reduced form difference-in-differences design, we find evidence of lesser-reported emotional and physical violence for women potentially exposed to the program. Overall, women reported lesser intimate partner violence and controlling behavior on account of the empowerment intervention. We conjecture that improved anthropometric outcomes, higher literacy, and better labor market opportunities serve as potential mechanisms causing this effect.
From Safe Motherhood to Cognitive Ability: Exploring Intrahousehold and Intergenerational Spillovers, (with Somdeep Chatterjee), Economica, Vol 88, 2021
(Earlier Version: GLO Discussion Paper No: 332, Global Labor Organization, Maastricht)
Empirical evidence on Conditional Cash Transfers (CCTs) in the education domain suggest increases in school participation but minimal effects on learning. In this paper, we find strong spillover effects of a CCT program on children’s academic performance, but interestingly the CCT was not conditioned on education. We study a flagship health and safe motherhood program from India and by exploiting plausible exogenous variation in exposure to the program, show positive effects on reading, writing and math test scores for elder children in the household whose parents or uncles/aunts were potentially exposed to the policy. Our results can be viewed as inter-generational effects of CCTs because the estimated benefits accrue to the future generations of the potential recipients of the transfer payments. We identify higher expenditure on private tutoring as potential short term mediating channel for the effects. We also find changes in fertility preferences of women, suggesting a longer term mediating channel applicable in a standard quantity-quality trade-off setting.
When left is ‘right’! The impact of driving-side practice on road fatalities in Africa, (with Vijaya Singh), Transport Policy, Vol 114, 2021
In this paper, we try to understand the impact of a traffic convention, rule of the road, on road fatality rate in Africa. Using the country level panel data on road fatality rates provided by World Health Organization, we try to find whether left side driving rule has any differential causal impact on road fatalities. To address issues related to endogeneity, we instrument ‘Left Side Driving Rule’ with the measure for ‘British Colonization’. Our estimates suggest that the left side driving rule results in lesser road fatalities when compared to the right side driving rule. The illegal usage of ‘Right Hand Drive’ vehicles in right side driving countries and functional superiority of left side driving rule are the potential channels causing this effect. With Africa bearing the highest ‘disease burden’ of traffic fatalities in the world, our research provides a preliminary empirical insight into the role that ‘rule of the road’ traffic convention can play in this regard.
Moving up the energy ladder: does socio-religious status matter? (with Debdatta Pal and Somdeep Chatterjee), Population and Environment, Vol 42, 2021
This paper investigates the role played by socio-religious categories in determining primary cooking fuel choices among Indian households. We study this role in the broader context of climbing up the energy ladder. Our estimates based on a sample of 601,509 households and using multinomial probit regression suggest that socio-religious status along with economic status is critical in the choice of modern eco-friendly fuels. We find that belonging to a marginalized community in Hindu religion significantly dampens a households’ likelihood to move up the ladder when compared with upper caste households. While intra-religion differences among Hindu castes in terms of their probability of using modern fuels are wider, differences among Muslims appear smaller. Also, though Muslims perform worse than Hindu upper castes in terms of probability of using modern fuels they are much better off in comparison with other Hindu castes. Our results remain robust to alternative specifications and several robustness checks.
Innovation and Corruption: Dissecting causal linkage using Patent Application information from India. (with Sanjay Singh), Singapore Economic Review, Vol 67, 2022
Using World Bank Enterprise Survey data on bribery and patent applications, we try to study the causal linkage between firm level innovation and corruption in India. Specifically, we try to understand if corruption impacts innovation at the firm level. Since we find that innovation and corruption are jointly determined, we propose instrumental variables regression approach to identify this causal effect. We instrument bribery by exogenously determined external audit parameter and then use a recursive bivariate probit model combined with industry-fixed effects to reach our results. Our findings suggest that bribery has an adverse impact on innovation. The results of our study are much in contrast to the existing literature, which largely supports a positive relationship between innovation and corruption.
For the love of God? Proselytization, Religious Restrictions and Social Conflicts in India, Revise & Resubmit
I study the social effects of religious restrictions in the context of proselytizing activities which form an important part of some religions. To establish causality, I exploit plausible exogenous variation from the `Freedom of Religion' legislations that were enacted in several Indian states during the period 1964-2005. These legislations primarily restricted conversions that were `forced or fraudulent' or were done through `allurement or inducement', with the judiciary also ruling that the right to propagate religion provided under Indian Constitution does not necessarily include the right to `convert' a person's religion. Using administrative data on riots, I find that the introduction of these legislations resulted in reduced social conflicts. The finding suggests that putting `reasonable' restrictions on religious freedom, i.e., regulating `unjust' conversions in a religiously diverse country like India can have an unintended positive consequence in the form of reduced rioting. The results of the paper remain robust to a variety of sensitivity checks.
Delivering Contraceptives at the Doorstep: Reducing Costs of Access or Increasing Costs of Stigma?, (with Somdeep Chatterjee and Shreya Mishra), Revise & Resubmit
Existing evidence suggests that improved access to contraceptives increases their usage in developing countries. In this paper, we study a unique policy experiment from India where doorstep delivery of contraceptives replaced the practice of acquiring free contraceptives from village health centers. Using a quasi-experimental setup that relies on a difference-in-difference design, we find that this intervention led to no significant changes in fertility outcomes, despite easier access. Instead, we find that women reduce the take-up of home delivered contraceptive methods including condoms and OCPs and substitute them for other concealable modern and traditional methods of contraception. This is primarily on account of reduced privacy for women in procuring contraceptives from ASHA workers in a patrilocal joint family environment. Our study makes an important policy contribution by showing that while doorstep delivery is a disruptive innovation in the healthcare market, its actual welfare consequences remain ambiguous in developing countries.
3G Internet, Intimate Partner Violence, and Women’s Empowerment: Evidence from Nigeria, (with Ridhi Kashyap, Till Koebe, and Valentina Rotondi)
A developing body of research has highlighted the impacts of mobile phones on women’s status in low- and middle-income countries (LMICs), through improved labour market opportunities, contraceptive knowledge and access, and decision-making power within the household. However, this literature often relies on cross-sectional designs and does not distinguish between different type of mobile technology (2G vs 3G). This distinction is important because the spread of the internet can have theoretically ambiguous and complex effects, ranging from exposure to globalized liberal ideas to reinforcing gender stereotypes and misogynistic views. In this paper, we examine how the rollout of 3G mobile networks over time in Africa’s most populous country, Nigeria, affects women’s status within the household looking at the experience of intimate partner violence and husband’s controlling behaviour. We draw on three waves of the Demographic and Health Survey’s women’s and men’s modules, which we link with novel, high-resolution mobile coverage maps, and estimate the impact of 3G using a quasi-experimental design. We find that 3G network expansion results in reductions in the experience of emotional, sexual and physical intimate partner violence among women, however, the results for husband’s controlling behavior are more mixed. Exploring pathways behind these effects, we find support for both ideational and empowerment channels. Under ideational channel, we find reductions in the acceptability of IPV among women, and improved attitudes of men with regards to women’s autonomy in intimate relationships. Under empowerment channel, we observe that women potentially exposed to 3G had higher say in household decision making. They were also likely to be educated for higher number of years and also had higher age at marriage and at first birth. Our results highlight the potential of technological expansion to influence gender dynamics within households.
What's Your Poison? Online Gambling, Alcohol Consumption, and Intimate Partner Violence in India, (with Sarthak Agarwal)
In this paper, we study the causal effects of a policy experiment in India, implemented through legislation that regulated and prohibited online gambling. Using a difference-in-difference design that exploits geographic and temporal dimensions of the ban, we document two key findings. First, we show that the law leads consumers to substitute towards consumption of another `sin' good - alcohol - leading to both higher overall consumption and frequency of consumption of the good. Second, this increased consumption of alcohol coupled with factors such as economic stress and increased exposure to husband produces a negative externality in the form of increased intimate partner violence. The results from our analysis remain robust to a variety of sensitivity checks including the use of alternative identification strategies (PSM-DID and triple difference), accounting for spatial confounders, under-reporting and timing of intimate partner violence, and access to technological resources. Our back of the envelope calculations suggest that the costs associated with the ban are likely to outweigh the benefits to government exchequer. The study highlights the importance of formulating well coordinated policies in the domain of `sin' goods so as to mitigate any potential negative externalities.
When Parents Go Back to School: Reallocation, Sorting, and Child Learning Outcomes, (with Sulagna Bhattacharya and Somdeep Chatterjee)
Parental education is an important input in child education production functions. However, it is not clear if the marginal product of this input is the same if measured as parental educational endowment acquired in the parents' childhood or if it is acquired through adult education initiatives. Structurally, the opportunity costs and behavioral manifestations of these are likely to be different. We study this question exploiting a natural experiment conducted in the form of the largest adult literacy intervention in the world. Using nationally representative data in a quasi-experimental framework, we find that while adult literacy improves women's agency and child private school enrollment indicating a positive shift in parental aspirations, it is also accompanied by a rise in dropouts and worsens learning outcomes. We argue that our findings indicate that adult literacy acts as contemporaneous re-optimization shock, crowding out other parental inputs into child's education owing to tightened income and time constraints while simultaneously triggering endogenous school sorting accompanied by elite exit from public schools. As a policy implication, our results suggest that for parents having school-age children, the returns to time spent with the child, regardless of education level, is likely to be higher as it leverages key complementarities in the education production function.