Gender Identity and Relative Income – New Evidence for Germany

Despite the rise in female participation in the labor market, a gender-wise specialization within couples persists. In this field of research, the role of social norms as important drivers of human behavior has received increasing attention as traditional theories on household economics fail to explain the persistence of within-couple gender inequalities. One of the most striking examples of norm-conforming behavior is how couples react to a violation of the male-breadwinner norm as shown by Bertrand et al. (2015). Couples avoid the situation of the wife out-earning her husband which manifests through a discontinuity in the income distribution at the point of equal incomes. The findings that couples engage in norm-conforming behavior have been established for several countries and contexts (see, e.g. Zinovyeva and Tverdostup, 2021; Lippmann et al., 2020). However, studies focusing on the discontinuity at the point of equal incomes find mixed results depending on the data and methodology used. For example, Hederos and Stenberg (2022) and Kuehnle et al. (2021) find that observed discontinuities reduce or even disappear when excluding equal-earning couples and when changing several of the parameters of the discontinuity tests. Furthermore, Roth and Slotwinski (2020) suggest that the observed norm-conforming behavior in relative incomes is the result of systematic misreporting of incomes. In this paper, we recur to a large administrative dataset and analyze the importance of the male breadwinner norm for the case of Germany. Our contribution is twofold. Firstly, to analyze this research question we use administrative tax records, the German Tax Payer Panel. We have access to the universe of the German income taxpayers for the years 2012 to 2019. Apart from the large sample size, this data has the major advantage of not being subject to reporting effects. Interestingly, we do find evidence for a discontinuity in Germany, which, however, varies in size for different subgroups and discontinuity tests used. In addition, in our analysis, we carefully consider recent evolution on the importance of certain modeling parameters and discontinuity tests. Secondly, we stratify the sample in several different ways to gain further insights into the extent of the relative income discontinuity for different subgroups of the population and to identify possible channels, such as couples with or without children, different age groups, and couples living in Western vs. Eastern Germany. Interestingly, we find that the observed discontinuity is largely driven by couples without children and is robust across different age groups.