Abstract

Over the last decade, research based on administrative data became more and more popular, especially in labor economics. Large sample sizes and the accuracy of the contained information are seen as advantages of this type of data. During the same time most big household surveys have started to increase their samples to ensure coverage also of specific subpopulations. Compared to administrative data, survey data contain much more information and cover more topics but at the cost of small sample sizes. However, this raises the general question of comparability of thus data sources.
In this project, we inter alia analyze the impact of the choice of the dataset on the estimation of trends in wage inequality and wage mobility in Germany. In particular, we compare results based on administrative data from the Sample of Integrated Labor Market Biographies (SIAB) with results based on the SOEP. In the context of wage analyses, SIAB data provides the opportunity to analyze specific subgroups due to the large sample size, but contains only wage observations up to the social security contribution limit. On the other hand, SOEP offers uncensored data over the full wage distribution but for a significant smaller number of individuals. In addition the SOEP is confronted with item non-response which may bias results.
The aim of the first paper is, to analyze if wage inequality and mobility show similar trends in both datasets in general and for selected sub-groups. A special focus is on the population above the social security contribution limit given that international literature argues that most part of increasing inequality can be attributed to top income earners. Our findings can be seen as basis for guidelines which dataset – subject to the research question – is better suited for the empirical analyses.