Research
Household Responses to Old Age Pension: Evidence from India, March 2018 [pdf] [presentation slides]
Abstract: This paper estimates the impact of a major expansion of a cash transfer program targeting poor elderly in India on their labor supply and health status, as well as on the nutritional status of co-residing grandchildren. Specifically, the expansion of the Indira Gandhi National Old Age Pension Scheme in 2011 involved lowering the eligibility age from 65 to 60, and increased payment generosity across states. This is a large program with 33 million beneficiaries and an annual expenditure of 150 million USD. I use two different identification strategies--a regression discontinuity design exploiting state age cutoffs for eligibility, and a difference-in-differences strategy exploiting state-year changes in eligibility. I find that the old age pension program decreases the labor supply of the elderly, with males less likely to be employed, and women working less, but has no detectable impact on their health status. The results suggest that cash benefits received by elderly women have a significant positive effect on the nutritional status of children, while cash benefits received by the elderly men have no impact on co-residing children.
Housing and Risk Preferences of Older Individuals [pdf]
Abstract: This paper examines the effects of housing wealth on risk preferences of the elderly. Following earlier literature, I define home equity as the difference between property value and mortgage debt. I estimate the causal relation between housing wealth and risk preferences using the restricted panel dataset of the Health and Retirement Study between 1992 and 2014, merged with the zip code level house price data from the Federal Housing Finance Agency (FHFA). Housing expenditure is the most common committed consumption. If households don't move, they need to make mortgage payments and pay for home insurance and property taxes. These expenses can't be adjusted or postponed. Using the HRS data, I find that around 80 percent of the individuals are homeowners, and housing wealth accounts for most of their aggregate wealth. I hypothesize that a negative shock to housing wealth will increase risk aversion among the elderly.
Crop Diversification and Growth of Maize in Karnataka: An Assessment [pdf]
Abstract: Traditionally, Karnataka is an agriculture-based economy. With the dawn of the secondary and tertiary sectors, the agricultural sector was grossly neglected and consequently its contribution to the State’s economy has been declining constantly over the last few decades. On the other hand, the growing needs of the agricultural sector have been greatly felt with the growth of population, not only in terms of food security but also in providing employment. Crop diversification within the sector has also been noticed on a large scale and as a result, the growth in the production of maize has been the highest with CAGR at 8.5 per cent in the last three decades. Using one-way Least Squares Dummy Variable (LSDV) for 27 districts over 12 years, the present study explored that introduction of new hybrid seed (HYV) as one of the most important factors for significant growth of maize in Karnataka. Further, though the crop is suitable in the drier region, the role of timely rainfall also had a significant impact on the yield level.
Determinants of Child Labor: Evidence from Tamil Nadu, Kerala and Andhra Pradesh [pdf]
Abstract: This paper highlights the determinants of child labor across the three states of Andhra Pradesh, Kerala and Tamil Nadu. This is done with special emphasis on the occupational profile of household members, mother’s educational attainment, gender and head of the household, and the wealth index. The analysis was conducted using the National Family Health Survey data (NFHS-3) of the year 2005-06. The analysis is done using a logit model and a zero inflated negative binomial model. The regression results show that while mother’s education is a significant factor determining child labor in the urban regions, the household head’s education plays an important role in explaining child labor in the rural regions. Also, the effect of the gender of the household head is more significant in deciding if the child should work within the household over them working outside the household. A family having higher number of dependent members is observed to have a higher magnitude of child labor. More number of family members working in the agricultural sector results in higher chances of a child working within, and outside the household. Andhra Pradesh, which records the highest magnitude of child labor as per the 2001 census, is also observed to have higher number of children working outside the household than working within the household. On a relative basis, it was found that the chances of a child working within the household is higher in Tamil Nadu and in Kerala.
Work in progress
Predicting Mental Health Service Demand
Abstract: Mental health service demand has surged dramatically, with healthcare systems struggling to anticipate and plan for therapy needs. Despite loneliness affecting 1 in 6 people globally and serving as a key predictor of mental health service utilization, no systematic approach exists for translating population-level loneliness trends into actionable demand forecasts for therapy services.This research develops a novel predictive model that estimates mental health service demand by detecting and analyzing loneliness indicators in social media data, specifically Reddit discussions. The study adapts proven digital epidemiology methodologies (similar to Google Flu Trends) to create the first population-level forecasting system for therapy service demand. The research employs a two-stage approach: (1) Loneliness Detection Model - Using natural language processing and machine learning techniques, I analyze Reddit posts from mental health-related communities to identify and quantify loneliness indicators with >90% accuracy, validated against established clinical measures (ULS-6 loneliness scale); (2) Demand Forecasting Model - I correlate aggregated loneliness trends with actual therapy utilization data to build predictive models that forecast therapy service demand 1-12 months ahead. I use five years (2019-2024) of Reddit data from loneliness and mental health communities (~1M posts), cross-referenced with national and state-level therapy utilization statistics, economic indicators, and demographic controls.
Effects of Spousal Death on Labor Outcomes
Abstract: This paper examines the impact of spousal death on labor outcomes in India, focusing on how widowhood alters work intensity, and occupational choice. Using panel data from the India Human Development Survey (IHDS, 2004–05 and 2011–12), I identify individuals who transition from marriage to widowhood between survey waves and compare their labor trajectories to those who remain married. The analysis employs individual fixed-effects and difference-in-differences models to control for unobserved heterogeneity and broader time trends. Key outcomes include labor force participation, days worked, and sector of employment. I further investigate heterogeneity by gender, age at widowhood, and household living arrangements to capture how economic necessity, caregiving burdens, and social support structures shape post-bereavement labor supply. By providing micro-level evidence from a large, nationally representative panel, this study contributes to the limited literature on the economic consequences of widowhood in low- and middle-income countries and highlights the intersections of gender, ageing, and labor market vulnerability in India.
Impact of Conflicts on Birth Rates: Evidence from India
Abstract: Armed conflicts and social violence can profoundly affect demographic patterns, yet systematic evidence on how different types of conflicts influence fertility decisions remains limited. India provides a unique natural laboratory with diverse conflicts across states and time periods, from separatist insurgencies in Kashmir and the Northeast to communal violence and Naxalite movements, combined with excellent demographic data collection. This research examines the causal impact of violent conflicts on birth rates across Indian states and districts from 1990-2020. The study exploits exogenous variation in conflict timing, intensity, and geographic distribution to identify how different types of violence affect fertility decisions, distinguishing between immediate disruption effects and longer-term demographic adjustments.
Using a panel regression framework, I analyze the relationship between conflict intensity (measured through deaths, incidents, and duration from ACLED and UCDP datasets) and birth rates (from India's Sample Registration System and Census data). The identification strategy leverages cross-state and within-state temporal variation in conflict exposure, controlling for economic conditions, policy changes, and demographic trends. Multiple conflict types are analyzed separately: separatist insurgencies, communal violence, and left-wing extremism.
This study provides the first systematic analysis of conflict-fertility relationships in India, addressing both immediate demographic disruptions and medium-term recovery patterns. The research identifies which types of conflicts most severely affect birth rates, how effects vary by conflict intensity and duration, and whether demographic recovery follows predictable patterns.
Do cash transfers to grandparents impact child labor and child work in India?
Abstract: This paper examines the impact of cash transfers to grandparents, specifically old-age pensions, on child labour outcomes in intergenerational households in India. Using nationally representative survey data and exploiting variation in pension eligibility and program expansion across states, I identify the causal effects of pension receipt on children’s work and schooling. I find that access to pensions significantly reduces the likelihood of child labour, particularly among younger children and girls, while increasing school enrollment rates. The effects are strongest in rural and low-income households, suggesting that pension income relaxes financial constraints and substitutes for children’s contribution to household earnings. Additional evidence indicates that grandparents’ presence and financial security reduce children’s need to engage in work, while also allowing parents—especially mothers—to adjust their labour supply. These findings highlight the broader intergenerational benefits of social protection programs and underscore the importance of designing transfer schemes that simultaneously address old-age security and child welfare.