Nowadays, in most western countries cohabitation is a structural part of relationship formation. Some cohabiters get married eventually, others never tie the knot. An increasingly smaller group does not cohabit prior to marriage. Previous research has shown that cohabiters run higher risks of breaking up and that this is not because of selection effects. However, cohabiters are a heterogeneous group that includes couples that are testing the relationship before getting married, couples that ‘slid into cohabitation’ and committed couples that do not want to get married. Marriage may, therefore, not be the most important factor influencing divorce risk, but rather the quality of the match itself.

We tested this by following cohabiting and married people in the Dutch survey on Family Formation of 2008 for up to 7 years after the survey using registrative data. In a discrete time analysis, also controlling for relationship duration, age at the start of the relationship, birth cohort and selection (education, religion, divorce of parents), divorce risks of cohabiters, married people with and married people without prior cohabitation were compared. Also, the influence of (1) characteristics of the match (age difference and difference in educational level); (2) investments (children and co-owning a house); and (3) relationship satisfaction were tested.

Results show that unmarried cohabiters have higher dissolution risks than married people and that there is no difference between married people with and without prior cohabitation. Investments and relationship satisfaction were important predictors of divorce, both for cohabiters and married people. However, this did not explain the differences found. This implies that even committed and happy cohabiters have less stable relationships than married people. Moving on to marriage may be a sign of more commitment in itself, but cohabiters or cohabiting relationships may also have other characteristics increasing the risk of divorce.