The Relation Between Cohabitation and Divorce

speculation and evidence

“Paradoxically, though, the choice of marriage preceded by cohabitation apparently attracts some couples who are less committed than others to lifelong relationships, since these marriages end in divorce at a far higher rate than those not preceded by cohabitation.” (Margaret Brinig, From Contract to Covenant)

Cohabitation before marriage provides a couple information about their ability to live happily together. That ought to make marital mistakes less likely and so lead to a lower chance of divorce. The evidence, however, goes in exactly the opposite direction.¹ Brinig’s explanation is that cohabitation makes divorce less likely but the sort of people who cohabit are less likely to stay married than the sort who don’t, and the second effect outweighs the first.

It is a possible explanation but I am not sure it is the correct one and she offers no evidence for it. I have some data of my own, although the sample size is a bit small. My first marriage was preceded by cohabitation and lasted about four years. The second was not, and has been going for more than forty.

On the basis of that experience, I offer two alternative explanations:

1. Humans, like some species of birds, pair mate, not exclusively but as an important element in our reproductive strategy. Part of what makes that workable is a link between sexual activity and our emotions, hardwired by evolution. Sleeping with someone, especially on a regular basis, creates emotional bonds. Breaking them can be hard. Those bonds, once created, may result in your marrying someone who, absent those bonds, you would have recognized as insufficiently well suited to you for a permanent relation.

2. Humans have a tendency to heavily discount future benefits in their decisions. This makes evolutionary sense, since we evolved in a risky environment; giving up benefits today in order to get larger benefits ten years from now is probably a mistake if you are likely to starve to death in a famine or get eaten by a predator before the benefits arrive. We deal with the conflict between inclination hardwired in the past and rational calculation in the present by a variety of devices, such as Christmas clubs to commit us to save or awarding status to the wealth from saving.

For many people, cohabitation is much pleasanter than search. Not only does it result in more sex, it provides a range of emotional and practical support. If you are cohabiting with someone sufficiently well suited to you to make cohabitation workable but not to justify marriage, abandoning cohabitation in favor of continued search means giving up a current benefit in exchange for a distant and uncertain future benefit. You may instead continue to cohabit, which means not searching or searching much less. Lack of search means you don’t find a better partner so eventually marry the one you have.

This is, of course, a drastically incomplete account of human mating behavior. For a more expert account than mine of the relation between evolution and human sexual behavior, I recommend the work of David Buss.

When I raised the puzzle of why cohabitation was associated with a higher risk of divorce on my old blog, commenters offered a number of other explanations. Some were variants of Brinig’s, some were not. For example:

The Shotgun Wedding Explanation

Couples that are living together before marriage are almost certainly sleeping together before marriage, couples that are not may not be. Sex sometimes results in pregnancy, pregnancy in marriage. A marriage formed for that reason may be more likely to fail. Couples that are not cohabiting might still sleep together but are less likely to.

The Inertial Explanation

A married couple has committed themselves to a long-term, perhaps permanent, relation one element of which is sexual exclusivity. A cohabiting couple has not. A cohabiting couple that eventually marries may carry the pattern of relationship appropriate to cohabitation over to marriage.

The Burnout Factor

Total years together may matter more than total years since the wedding. If so, a couple that lived together for two years before they married and four after is more comparable to a non-cohabiting couple married six years than one married four. The longer it is since a couple was married the more likely it is that they are now divorced.

Evidence

The Relationship Between Cohabitation and Marital Quality and Stability: Change Across Cohorts? offers statistical evidence on whether cohabitation is associated with a higher probability of divorce because of the people who choose to cohabit, a selection effect, or because of the effect of cohabiting on people, an experience effect. Its central idea is that as cohabitation became more common:

The percentage of marriages preceded by cohabitation rose from about 10% for those marrying between 1965 and 1974 to more than 50% for those marrying between 1990 and 1994

the people who chose cohabitation became less different from those who did not, so the selection effect should have been reduced or eliminated. To test that, the authors compared data on two groups of couples:

those who married between 1964 and 1980 (when cohabitation was less common) and those who married between 1981 and 1997 (when cohabitation was more common).

They found that the association between cohabitation and divorce, while weaker for the later group, was still substantial.

Converting the b coefficient to an odds ratio (e^b= 2.51) reveals that cohabitation was associated with an increase of 151% in the odds of divorce [in the earlier cohort] . … Converting the b coefficient to an odds ratio (e^b= 1.90) indicates that the odds of divorce were 90% higher in the more recent cohort.

As a second test, they calculated the association while controlling for a list of characteristics (race, parental divorce, marriage order, education, family income, and welfare use) that correlated with both cohabitation and divorce and so were plausible drivers of the selection effect. Again they found the association reduced but still substantial.

They interpreted their results as evidence for the experience effect. I interpret them as evidence for both, the selection effect because the association was reduced both by cohabitation becoming more common and by controlling for demographic characteristics, the experience effect because it remained substantial even when cohabitation became common and even after controlling for relevant characteristics.

---

My web page, with the full text of multiple books and articles and much else

Past posts, sorted by topic

A search bar for past posts and much of my other writing

1. The Relationship Between Cohabitation and Marital Quality and Stability: Change Across Cohorts? contains links to literature on the evidence.