Grain Exports and the Causes of China's Great Famine, 1959-1961: County-Level Evidence

Abstract

This study quantitatively evaluates the relative importance of different causes of China's Great Famine, especially for the importance of grain exports. We exploit county-level over-time variations in crop specialization patterns to construct Bartik-style measures of export shocks. Using county-level panel data from 1955 to 1963, we regress death rates on the Bartik export measures with county and province-year fixed effects as well as time-varying effects of county-level observables. We use weather shocks to instrument for output and consumption. The regression results suggest that increases in grain exports substantially increase death rates. This effect is larger in counties that are further from railways and with fewer local Chinese Communist Party members. To examine the relative importance of different mechanisms, we also estimate the effects of the procurement policy, the determinants of grain output, and the relationship between death rates and county-level average caloric consumption during the famine period. The counterfactual experiments indicate that the fall in agriculture production, the increase in procurement partly driven by grain exports, and the increasingly progressive and inflexible procurement policy collectively increased the number of excess deaths, where no single factor dominates. In particular, grain exports explain 15 percent of excess deaths, which is one-fourth of the effect of the increase in procurement rates between 1957-1959.

Keywords: famine severity, over-export, county-level data, Bartik-style export shocks, grain procurement, distance to railways, Chinese Communist Party members

JEL Classification: F14, P2, O24, N45