Peaks vs. Components

We analyze the cross-national distribution of GDP per capita and its evolutionfrom 1970 to 2003. We argue that peaks are not a suitable measure for distinctgrowth regimes, because the number of peaks is not invariant under strictlymonotonic transformations of the data (e.g. original vs. log scale). Instead,we model the distribution as a finite mixture, and determine its number ofcomponents (and hence of distinct growth regimes) from the data by rigorousstatistical testing. We find that the distribution appears to have only twocomponents in 1970-1975, but consists of three components from 1976 onwards...