Poverty traps, convergence, and the dynamics of household income

We design a new method to detect household poverty traps. We apply the method to a unique panel that follows rural Indian households over thirty years. We find no evidence of poverty traps. Most households had annual income growth of over 2%, and income mobility is high. We then design and apply a method to detect conditional convergence. We find that upper castes are converging to a level of wealth 3 times as high as disadvantaged castes.