Upcoming Presentations

Abstract: The unobservability of lifetime income poses a fundamental challenge to estimating the intergenerational elasticity (IGE), forcing researchers to rely on annual income averages. We show that these proxy-based estimators converge to context-specific limits, compromising comparability across studies. We address this by establishing nonparametric identification of the IGE, using partial income data and family characteristics. We then construct a consistent two-step estimator that integrates machine learning with cross-fitting and Neyman-orthogonal moments, ensuring that the resulting IGE estimates are locally robust to first-stage errors. This approach delivers valid inference, overcoming long-standing challenges in the comparability of IGE estimates across studies, time, and place. Using the Panel Study of Income Dynamics, we estimate an IGE of 0.69 for the United States, in line with evidence from long-run parental income averages that suggests the true value exceeds 0.6. This stands in sharp contrast to proxy-based estimators, which yield estimates of 0.38 and 0.51.

*This session is part of the Job Market Seminar Series.

Abstract: Egypt recently introduced universal health insurance (UHI) through a phased governorate-level roll-out. The staggered adoption of UHI is a natural experiment that provides an opportunity to estimate the causal effects of the intervention on outcomes of interest. Using synthetic control methodology (SCM), the paper analyses the impact of implementing UHI in Port-Said and Luxor – the first pilot governorates for Egypt’s UHI system, on out-of-pocket (OOP) health spending and catastrophic health expenditure (CHE). Study findings indicate that the introduction of UHI led to a statistically significant decline in both OOP health spending and the incidence of CHE, when pooled pilot governorates are compared to their synthetic control. Meanwhile, heterogeneous impacts are observed when the effect on each governorate is examined separately, where a decline is only witnessed for the early-adopting governorate.

*This session is part of the Job Market Seminar Series.

Abstract: Most U.S. safety net programs determine eligibility using the same income threshold nationwide, without accounting for regional differences in the cost of living. This paper examines whether indexing eligibility to local costs improves coverage and targeting in Medicaid and the Health Insurance Marketplace—two adjacent programs for low-income individuals without employer-sponsored coverage. I develop and estimate a model in which the regulator sets eligibility rules, insurers compete on both price and quality, and consumers either choose their plans or are assigned to them. I demonstrate that indexing thresholds to the cost of living produces competing effects: it expands coverage by drawing low-income families into insurance in high-cost areas, but also reshapes risk pools across programs, affecting premiums and plan quality. Counterfactual analyses using administrative data from California reveal that adjusting thresholds partially to account for cost-of-living yields greater welfare per public dollar than either uniform thresholds or full indexing. The results highlight how eligibility design interacts with market structure and insurer incentives in segmented health insurance systems. 

*This session is part of the Job Market Seminar Series.

Abstract: This paper examines the relationship between armed conflict and agricultural production in Afghanistan by combining household-level survey data with district-level conflict records. We estimate linear mixed-effects models with district and year random intercepts, focusing on the log of per capita conflict fatalities lagged by one year (Lag1) to capture the short-term post-conflict effect on crop output. Across specifications, the Lag1 coefficient is positive, statistically significant, and robust to the inclusion of household, environmental, and institutional controls. In the fully specified model, the estimated elasticity is approximately 0.033, implying that a doubling of conflict fatalities is associated with a 2.3% increase in agricultural output. While counterintuitive, this finding is consistent with adaptive responses such as shifts in cropping patterns, labour allocation, or input use in the aftermath of conflict. To explore persistence, we extend the fully specified model to lags of up to five years (Lag1–Lag5). These results indicate that the association remains positive for several years but gradually attenuates; magnitudes are smaller at longer lags. A robustness check replacing fatalities with conflict incidents per capita (Lag1) yields comparable results. Our findings suggest that rural households demonstrate short-run adaptive capacity in the face of conflict, but that such adjustments may not be sustainable over longer horizons. The results highlight the importance of timely policy interventions to protect agricultural livelihoods and support post-conflict recovery.

*This session is part of the Job Market Seminar Series.