Clusters and Loops of the German Phillips Curve

A preliminary regression analysis of different versions of the Phillips Curve on the basis of yearly data of the German economy from 1952 to 2004 leads to the conclusion that the original finding might still be of empirical relevance. A simple plot of seasonal adjusted quarterly data between the change of nominal wage rates and the unemployment rate shows a picture similar to that by which Phillips was inspired to his famous discovery: A long-term tendency of a negative, non-linear relationship coupled with minor deviations from this tendency forming sometimes so called loops. At fist sight, the Phillips Curve of Germany comprises clusters of data points and movements between these clusters. The clusters can be analysed and – together with the rest of data – dissolved into 12 (left or right turning) loops and 9 movements between these loops during the period from 1971Q1 to 2009Q4. In spite of the striking differences of these phenomena, a model with one regression equation is sufficient to explain the loops, the movements between the loops and the long-term tendency of the German Phillips Curve. This empirical finding contradicts several aspects with the ruling dogma of a Phillips Curve that broke down in the ‘70s and with the allegedly better fit of its replacements by augmented and modified Phillips Curves.