The Politics of Drought Relief: Evidence from Southern India

A critical role of government is to provide assistance to households a ff ected by negative shocks, such as natural disasters. This paper studies such government intervention in the form of o ffi cial drought declarations in three Indian states. First, I demonstrate that while states’ targeting of drought relief partly aligns with environmental criteria, it does not strictly adhere to national guidelines. Instead, I find that allocations are disrupted by political motivations: the likelihood of receiving a declaration increases in an area’s electoral competitiveness while areas aligned with the state ruling party are less likely to receive a declaration. I develop a dynamic probabilistic voting model, driven by the state ruling party’s reelection motives, that explains these results. The model also highlights potential endogeneity concerns in empirical specifications connecting political conditions to drought declarations. To causally identify the relationship between electoral incentives and declarations, I thus construct instruments for electoral competition and incumbency utilizing the interaction of a party’s vote share in a prior period with a state-level, leave-one out change in support. I confirm that the associations I report reflect a causal relationship. Finally, I o ff er suggestive evidence that the mistargeting of declarations, relative to national guidelines, results in a misallocation of public resources. We show a causal impact of immigration on innovation and dynamism in US counties. To identify the causal impact of immigration, we use 130 years of detailed data on migrations from foreign countries to US counties to isolate quasi-random variation in the ancestry composition of US counties that results purely from the interaction of two historical forces: (i) changes over time in the relative attractiveness of di ff erent destinations within the US to the average migrant arriving at the time and (ii) the staggered timing of the arrival of migrants from di ff erent origin countries. We then use this plausibly exogenous variation in ancestry composition to predict the total number of migrants flowing into each US county in recent decades. We show four main results. First, immigration has a positive impact