Distortions in Cross-Sectional Convergence Analysis when the Aggregate Business Cycle is Incomplete

One of the most important drawbacks of the existing literature on convergence is that it largely ignores the effect of aggregate fluctuations on the evolution of income disparities. To the extent that regional income disparities follow a distinct cyclical pattern in the short-run, moving either pro- or counter-cyclically, the period of analysis should be chosen with great care. Failing to do so might in fact lead to an overestimation of the tendency towards either convergence or divergence, depending on the type of short-run cyclical pattern followed by the disparities and on which cycle phases are over-represented within the period being analyzed. In this paper, we use the distribution dynamics approach to show that the distortion introduced when the period of analysis contains incomplete business cycles could be quite sizeable and then analyze convergence among 48 conterminous US states over a appropriately chosen period (1989-2007) that includes only complete cycles.