An agent-based computational model of assortative mating and the reversal of gender inequality in education in Europe

André Grow, Katholieke Universiteit Leuven
Jan Van Bavel, Katholieke Universiteit Leuven
Yolien De Hauw, Katholieke Universiteit Leuven

In Europe, higher education was mostly a male domain until the 1970s. Since then, the gender gap has decreased and by now, women largely excel men in terms of educational attainment. To date we know little about how the reversal of gender inequality in education (RGIE) might have affected demographic decisions such as union formation, union dissolution, and fertility. With this paper, we provide one of the first steps towards filling this lacuna. We develop an agent-based computational model that enables us to study the mechanisms that link RGIE to patterns of assortative mating across European countries. Our model builds on the notion that mate search is an adaptive process. In this view, individuals have aspirations for partners with certain characteristics. These aspirations develop and change in response to experiences during partner search. For instance, individuals who fail to find a partner with a desired tertiary degree might relax their aspirations and might become willing to accept partners with lower degrees. We argue that this process is a key mechanism that links RGIE with observed patterns of assortative mating. We use empirical data (e.g., from the European Social Survey) for constructing target patterns of assortative mating and for initializing our model. The central outcomes are patterns of female educational homogamy, hypergamy, hypogamy, and the share of never-married men and women in a cross-temporal and a cross-national perspective. The simulation process is scaled in years. This enables us to initialize exogenous model aspects as to resemble observed data over time. For instance, assuming that the simulation starts with a cohort born between 1948 and 1952, we can initialize key demographic characteristics (e.g., distributions of educational degrees) of this and successive cohorts based on empirical data. This enables us to show how RGIE is linked to patterns of assortative mating across Europe.