Abstract:
This dissertation adds to the empirical literature on income inequality and its evolution over the last two decades in Germany, moving from a particularly broad income definition to more specific measures of between-group inequalities in wages. The contribution of this dissertation is broadly characterised by two main themes: First, it emphasises the added value of taking a distributional perspective in an analysis of inequality (Chapters 2 and 3). Second, it underscores the importance of accounting for changing selection on unobservables in an assessment of wage differentials and their evolution over time (Chapters 3 and 4). Setting the scene for the subsequent studies, Chapter 2 explores the distributional changes in net equivalised household incomes between 2005/06 and 2015/16 based on rich survey data. The analysis thereby breaks down the observed changes across the distribution into the individual contributions of eight factors overall, with a particular focus on the employment boom that unfolded over this period. Our results reconcile the seemingly contradictive observation that the record employment levels did not result in a notable decline in disposable income inequality, showing that the employment boom alone did in fact have an equalising effect that was – however – masked by other developments such as immigration and educational expansion. Shifting the focus from the household to the individual level, Chapter 3 turns the attention to selection on unobservables and its role in explaining the German gender wage gap. More specifically, we use German administrative wage data and employ the novel distribution regression approach with selection by Chernozhukov et al. (2025), which constitutes a methodological breakthrough as it allows for non-uniform selection across the entire wage distribution. Our results indeed reveal substantial heterogeneities: For full-time men, we find positive selection at the bottom – possibly related to Germany’s generous social safety net –, whereas it is negative in the rest of the distribution. Full-time women, on the other hand, appear to be generally negatively selected – and increasingly so towards the top –, which may be explained by assortative matching combined with the German tax and transfer system. Our results point to a strong role of selection in explaining the gender wage gap, although it has become less important over time due to slightly converging selection patterns of men and women. Chapter 4 finally considers another particularly salient trend on the German labour market: the ongoing expansion of tertiary education, which has inspired discussions as to whether this trend may be linked to a decline in the average (academic) ability of the highly educated. In order to disentangle potential quality effects from supply effects, Chapter 4 exploits regional variation in educational shares, comparing the wages of college graduates who received their university entry certificate (“Abitur”) in different German states, but who otherwise face the same labour market conditions. The results tend to imply decreasing college wages with higher “Abitur” shares, suggesting that educational expansion is linked to a decline in the average ability of the highly educated. A simple counterfactual thought experiment further shows that this effect has dampened the evolution of the college premium in Germany.
