The overall aim of this project is to examine spatial income inequality and to disentangle its determinants over time in Germany on a small regional scale. While the taxpayer data provide reliable income details, it does not cover variables concerning the social disaggregation that are relevant for income prediction. However, the German Microcensus does contain these variables. Therefore, we discuss performant prediction models for income information as well as potentials of statistically matching data from the taxpayer data with data from the Microcensus obtained from a simulation study to improve the income models.