Testing Two Views of Famine: Theory and Evidence from South Asia, 1861-1930

Contemporary policy regarding the optimal response to famines (and less extreme agricultural productivity shocks) is divided between the 'food availability decline' (FAD) view of famines, and the 'entitlements' view of famines (commonly associated with the work of Sen, 1981). But empirical work designed to test the relative importance of these views is scarce.  An understanding of the relative importance of these two views is important for informing famine prevention policy, which is expected to be important in coming decades during which climatologists predict an increase in the incidence of drought in many of the world’s poorest countries. 

We construct a general equilibrium trade model that provides a testable theoretical framework nesting both of these views of famine. The relative importance of each view has differing predictions about the responsiveness of incomes and trade flows in food to productivity shocks in open and closed economies. We test these predictions in the environment of colonial South Asia from 1861-1930, a period in which there were tens of famines and lesser scarcities.  In doing so we draw on a newly created dataset to estimate how the elasticity of prices, wages, output and mortality with respect to rainfall shocks (a key determinant of productivity) changes as regions opened to interregional and international trade. An attraction of this empirical setting is the dramatic improvement in domestic market integration that occurred during this time period due to the expansion of South Asia’s railroad network.  We also estimate how trade flows within South Asia respond to rainfall shocks – that is, whether food flows into or out of famine-stricken regions.  Our results inform the understanding the causes of – and optimal policy responses to – famines like those in colonial-era South Asia.