Are developing countries becoming equally unequal?

The paper contributes to the sparse literature on cross-country income inequality convergence by testing for convergence in nations’ overall relative income distributions (i.e. sigma convergence).  This is an important departure from previous studies that test for convergence in gini coefficient values (i.e. beta convergence).  If countries converge to different steady state income levels, then dispersion in income may increase despite evidence of beta convergence. Beta convergence is therefore a necessary but not a sufficient condition for sigma convergence.  Until very recently, the lack of high quality data on decile income shares made formally testing for sigma convergence in developing nations impossible.  However, using data from the World Bank’s PovcalNet database, we construct a panel on 1) the decile share of income, and 2) the gini coefficient for more than 100 developing countries over a period of 27 years (1985 to 2011).  Using this data, the paper finds robust evidence of rising inequality from the 1990s until approximately 2003, when the trend abruptly reversed and inequality steadily declined through 2011. Decile income shares also exhibit a statistically significant decline in dispersion over the sample period, providing strong evidence of sigma convergence.  Specifically, we employ a variance ratio test that reveals declining income dispersion for the lower income deciles between 1985 and 2003, and declining income dispersion in all but the lowest two deciles from 2003 to 2011.  In addition, we find statistically significant negative trend components in most of time series formed by the standard deviations of the decile income shares.  These findings indicate that relative income distributions among developing countries are becoming slightly more equitable and noticeably more homogenous.