Print Post
  • With most second and higher-order unions being nonmarital in Finland, repartnering in the aggregate did not produce higher fertility. Tweet This
  • Where cohabitation in young adulthood is common, fertility is more likely to remain low. Tweet This
  • “Marriage and cohabitation are far from indistinguishable in a country often described as a second demographic transition forerunner,” a new study of Finland concludes. Tweet This

People are frequently surprised when they find out that I have five children, but less so upon learning that my second husband did not have any before we had three together. His childlessness at midlife was uncommon, and either he was going to maintain unusually low fertility, or I was going to have unusually high fertility. We couldn’t both have 1-2 children. 

As odd as it may sound, this kind of scenario has given hope to population scientists who would like to see fertility increase back to around two children per woman, a level that would make labor supply more stable and old age support more tenable. The transition to below replacement fertility has happened at the same time as (and likely because of) the global “retreat from marriage,” a phrase that captures the many ways that people spend fewer of their adult years married. Fewer enter marriage at all, and those who do marry typically do so later in life, often divorce, and either do not remarry or do not remarry quickly. The retreat from marriage takes many adult years away from the normative context for childbearing (yes, marriage still warrants this description), and population scientists have wondered whether this fertility-dampening cloud might have a silver lining—or whether remarriage might cause people to have more children. 

To date, the research literature has shown that it does not. Even though there are people like me who have three more children than they would have had if their first marriage had remained intact, the more common experience is for divorce to suppress childbearing.1 The conclusion from multiple studies across time and space is that divorce is not an engine for fertility.

Nonetheless, the retreat from marriage itself provides a reason for a new revisiting of the question in Finland. Linus Andersson, Marika Jalovaara, Caroline Uggla, and Jan Saarela point out that previous research may have missed some of the “silver lining” by focusing on whether remarriage is an engine for fertility, instead of considering repartnering more generally. Their work, published in Demography in November, counted every marriage and cohabitation across the reproductive years to determine whether more unions produced more children.

This approach would be appropriate in any country with a large share of nonmarital unions, but it also requires something else that Finland has—namely register data. Unlike more typical vital registration systems that produce certificates of events (like birth, marriage, and death), registers record a person’s residences and life events in one place (i.e., the record belongs to the person rather than the event). As a result, Andersson and his colleagues were able to count all the births, marriages, and cohabitations2 occurring for Finnish men and women born between 1969 and 1972 who reached the end of typical childbearing years quite recently.3

When counting fertility in second and higher-order cohabitations as well as marriages, the research team reached the same answer that had come from previous work: "less is more." Finns with fewer partnerships had more children.

Andersson and his colleagues explain that most repartnering followed the dissolution of first cohabitations, and that most second unions were also cohabitations. This means that the people most likely to repartner were also unlikely to marry the second time around, or as the authors put it, “repartnering behavior is dominated by sequences of nonmarital unions.” In the minority of cases where second unions were marriages (like mine), repartnering did in fact lead to more lifetime children. But with most second and higher-order unions being nonmarital, repartnering in the aggregate did not produce higher fertility. Less remained more overall. The authors explain how first unions that do not transition to marriage are “foundational events,” which tend to reduce future fertility. This means that where cohabitation in young adulthood is common, fertility is more likely to remain low.

The authors’ conclusions go beyond their primary question of whether repartnering can be an engine for fertility. They also address the question of whether the “partnership transition” from a marriage-dominant model to one with a plurality of union forms will ever result in a society where marriage becomes indistinguishable from other unions. Based on their fertility analysis, the authors conclude that “Marriage and cohabitation are far from indistinguishable in a country often described as a second demographic transition forerunner.” 

Laurie DeRose is a senior fellow at the Institute for Family Studies, Assistant Professor of Sociology at the Catholic University of America, and Director of Research for the World Family Map Project.

1. Union dissolution reduces both intended and unintended births.

2. Statistics Finland considers different-sex individuals (other than close relatives and those with an age gap of more than 20 years) to be living in a union if they shared the same dwelling for more than 90 days. Individuals with an age difference of more than 20 years are also counted as a cohabiting couple if they have a child together. See here and here for work confirming the accuracy of this approach.

3. In the absence of register data, the most viable approach for tackling this question would be to survey women over 50, asking them to report how many coresidential unions they had ever been in and how many children they had given birth to. The approach Andersson et al. used avoids two problems inherent in this alternative methodology: 1) counts made retrospectively are more likely to be in error than counts recorded at the times of the events, and 2) generalizing from women willing to fill out a survey to an entire population requires many assumptions that having national data on the entire population does not.