This study investigates the sources of cross-country income disparities in time series dynamics for both aggregate and sectoral levels. We first introduce a new approach for testing income disparity persistence. The method proposed by Pesaran, Pierse and Lee (1993) is then used to decompose the persistence into components due to various shocks. By using the quarterly data of the U.S., U.K., and Canada from 1961:i to 1996:iv, we find the persistence of aggregate income disparities is caused by the divergence in the non-agricultural sectors. In addition, macroeconomic shocks are quantitatively not as important as sector-specific shocks in driving the persistence.