The "make-up" of a regression coefficient: Gender gaps in the european labor market

We provide a comprehensive picture of the relationship between labor market outcomes and age by gender in the 28 European countries covered by the European Statistics on Income and Living Conditions. The analysis is based on a somewhat unconventional approach that refers to concentration curves in the Gini regression framework. It allows identification of ranges in the explanatory variables where local slopes change sign and/or size, i.e. the components that "make up" a regression coefficient. Gender is a crucial factor differentiating participation among workers, although employment-age profiles do not substantially differ. Relevant differences in age profiles concern working-hours patterns: some countries are characterized by an almost specular behavior in men and women; other countries instead show similar patterns. Generally, earnings increase with age for both men and women. However, local regression coefficients are not monotonic over the entire age range and can even be locally negative in some countries. © 2014 International Association for Research in Income and Wealth.