A new semiparametric approach to analysing conditional income distributions

In this paper we explore the application of structured additive distributional regression for the analysis of conditional income distributions in Germany following the reunification. Using a bootstrapped Kolmogorov-Smirnov test we find that conditional personal income distributions can generally be modelled using the three parameter Dagum distribution. Additionally our results hint at an even more pronounced effect of skill-biased technological change than can be observed by standard mean regression.