Population Adjustment to Asymmetric Labour Market Shocks in India: A Comparison to Europe and the United States at Two Different Regional Levels

Abstract

This paper uses Indian EUS-NSSO data on 32 states/union territories and 570 districts for a bi-annual panel with 5 waves to estimate how regional population reacts to asymmetric shocks. These shocks are measured by non-employment rates, unemployment rates, and wages in fixed-effects regressions which effectively use changes in these indicators over time within regions as identifying information. Because we include region and time effects, we interpret regression-adjusted population changes as proxies for regional migration. Comparing the results with those for the United States (US) and the European Union (EU), the most striking difference is that, in India, we do not find any significant reactions to asymmetric non-employment shocks at the state level, only at the district level, whereas the estimates are statistically significant and of similar size for the state/NUTS-1 (Classification of Territorial Units for Statistics (NUTS, the French abbreviation for "nomenclature d’unités territoriales 21 statistiques")) and district level in both the US and Europe. We find that Indian workers react to asymmetric regional shocks by adjusting up to a third of a regional non-employment shock through migration within 2 years. This is somewhat higher than the response to non-employment shocks in the US and the EU but somewhat lower than the response to unemployment shocks in these economies. In India, the unemployment rate does not seem to be a reliable measure of regional shocks, at least we find no significant effects for it. However, we find a significant population response to regional wage differentials in India at both the state and district level.