Who partners with whom in a diverse society? Union formation among 50 origin groups in the Netherlands

Authors – Kasimir Dederichs, Frank Van Tubergen.

Abstract :

The social integration of minority groups has been studied among Sociologists for a long time. This research has overwhelmingly taken a bilateral perspective by asking questions about structural relationships between minority and majority members. Our approach extends this approach to a multigroup perspective examining the social distance between all major groups, i.e., not only between the majority and minority groups but also between different minority groups. Focusing on partnership formation as a classic indicator of social distance we ask: “Who is partnering with whom?” and “Which individual and group-level characteristics can explain variation in the prevalence of partnership formation across origin groups?” Switching to such a multigroup perspective is not only relevant in light of the rising levels of diversity that make unions between members of different minorities the fastest growing type of union. It also offers an opportunity to test longstanding theories about interethnic union formation based on a larger and more diverse set of origin group combinations.

To address these questions, we focus on the Netherlands, where more than 20% of the population has origins in a diverse set of countries abroad. The Dutch register data contains information on all cohabitations and marriages formed between 1995 and 2021. These rich data not only include at least 500 union-forming individuals of each sex for the largest 50 origin groups but also allow us to model the full opportunity structure on the partnership market each year. To apply the multi-group perspective to the Dutch partnership market, we expand upon an innovative recently developed approach and take two steps: First, for each union-forming individual, we construct a choice set consisting of their actual partner and 20 randomly sampled 20 counterfactual partners. Among each origin group-sex combination (e.g., Turkish women or Surinamese men), we then run conditional logit models to estimate the probability of everybody in their choice set being the actual vis-à-vis a counterfactual partner, given their origin group. This results in a 50×50 matrix containing 2,500 predicted probabilities of mutual union formation. In the second step, we show that much of this variation can be explained by the religious composition of the country of origin, union formation norms, educational differences, and residential segregation. The implications of our findings for traditional integration theories are discussed.