Implications of changes in the gendered educational advantage on family building: an agent-based model of first partnership formation in Sweden (the age-educational attainment trade-off)

Hernan Mondani, Stockholm University
Livia Olah, Stockholm University

We have seen an expansion of female educational advantage in Western Europe in recent decades. There are now more high-educated women than men entering the mating market. This is true also for Sweden. This situation can generate a tension with respect to the traditional pattern of partnership, characterized by female educational hypergamy, and hence leading to the education-specific mating squeeze. In this study we use agent-based modelling (ABM) to explore whether the gender- and educational structure of a society under certain education- and age-specific preferences for mating can lead to imbalances in first partnership formation across groups, and if so the extent of that impact. Specifically, we seek to explore the demographic outcomes of different preference scenarios, examining some of the “escape routes” suggested by van Bavel (2012). Our results suggest that shifts in age preference have a larger impact on partnership patterns than shifts in education preference. Moreover, comparing our simulations with the GGS dataset, we find that different educational level groups may follow different routes. For low-educated cohabiting women, shift in age preference (escape route #3) is the scenario that closest resembles in a qualitative way the patterns observed in the GGS dataset. In contrast, for middle- and highly-educated cohabiting women, we see an increase in educational homogamy (escape route #1). For first marriage, the shift in age preference scenario (escape route #3) works well both for low- and middle-educated women, while highly-educated women seem to follow the path of strengthened educational homogamy (escape route #1). Keywords: first partnership formation, cohabitation, marriage, reversal of gender imbalance in education, education- and age-specific mating, agent-based modelling, Sweden, GGS.