Growth, geography, and the iron law: Understanding divergence across Indian districts

The existing literature on Indian growth finds no evidence of B convergence across states. This represents a puzzle given the relatively free flows of capital, labour and commodities across state borders. We use a new data set to estimate convergence rates across 575 Indian districts and find that the pattern of absolute B- divergence remains. To explain this we develop a model of conditional convergence that includes a gravity indicator of trade and migration costs - specifically the distance from a major metropolitan center - as a conditioning variable. We find strong evidence of conditional convergence with an elasticity close to Barro's "iron law". We also find that geography and public infrastructure variables are important conditioning variables.