Cohabitation, Marriage, and the Happiness Gap

Married couples have generally been found to be consistently happier than cohabiting couples, but has the happiness gap closed? Brienna Perelli-Harris, Stefanie Hoherz, Trude Lappegård, and Ann Evans addressed this question in impressive detail in the highly technical journal Demography, concluding: “Marriage does not lead to higher subjective well-being; instead, cohabitation is a symptom of economic and emotional strain.”

Let me explain both the evidence supporting that conclusion and the reasons why it seems overstated. First, let’s take a look at their evidence that cohabitants are less happy than married people that has nothing to do with marriage per se:

1. First, they found that the happiness gap is not universal: among men in the UK and Norway and among women in Germany, cohabitants and marrieds report the same levels of subjective well-being. (Four countries analyzed separately for men and women yielded eight combinations where marrieds might be happier; they were, in fact, happier only in five).
2. After accounting for the fact that cohabitants have other factors working against their happiness—e.g., they generally have lower incomes, more previous partners and previous children, and worse health—the happiness gap was significant only for women in the UK and Norway.
3. After additionally controlling for relationship quality, only Norwegian married women experienced greater happiness than their cohabiting counterparts.

Perelli-Harris and her colleagues point out that cohabitation shares many characteristics with marriage that would augment well-being: intimacy, emotional and social support, and joint residence. Their data affirms that during midlife, or 38 to 50 years of age, people are happier when they have partners. They then scrutinize whether marriage, per se, augments happiness (i.e., whether there was a “marital happiness premium” enjoyed by couples whose unions had legal status). They purposively chose four countries that varied in their combinations of family policy and cultural orientations toward marriage to get insight into what it was about the context that might condition the marital happiness premium.

Their analysis “explained away” differences in happiness according to legal union status for every group besides Norwegian women. However, it is worth looking at how differences were “explained away” in the study, as well as what questions remain. 

Explaining Away the Differences

When the authors concluded that “cohabitation is a symptom of economic and emotional strain," the variables measuring economic and emotional strain included partnership and fertility behavior. That is, after controlling out the effects of hardly incidental life events—like whether people were in their first co-residential union and whether they had children with previous partners—married people’s happiness advantage seemed smaller.

The logic behind including these controls is that people with past experiences that might reduce their happiness are more likely to be cohabiting. While that makes sense, it opens the door for the consequences of choosing a lower commitment track to be taken out of the happiness equation. It amounts to saying, “If I’m cohabiting with my second partner because I’m wary of committing for life, the happiness hit from my previous union disruption shouldn’t be charged against my cohabitation status.” 

The authors acknowledge this point, writing: “Unlike selection in childhood, selection in adulthood could be confounded with partnership formation and happiness.” 

I agree, and if the “emotional strain” associated with previous unions is fully discounted when comparing marriage and cohabitation, it doesn’t quite follow that cohabitation is only a symptom of the strain it could have contributed to.

Additionally, in the study, Australian women’s marital happiness premium went away after the authors controlled for partnership and fertility behavior—even though it had persisted when controlling for childhood characteristics. 

Similarly, the authors included men’s own socioeconomic status, which also made the marital happiness premium for both Australian and German men disappear. While it makes sense to account for the fact that lower-earning men are less likely to marry, doing so amounts to statistically preventing any earnings boost that comes from marriage to contribute to marital happiness. This is important because marriage helps attach men to the labor force; it isn’t just that jobs make men more marriageable. 

Additionally, for women in the UK, the marital happiness premium didn’t disappear until the authors controlled for relationship satisfaction. “This result may imply that the quality of the relationship matters more than whether it is legally recognized,” the authors write, “or that only women with high-quality relationships and suitable marriage partners marry.” 

But it is important to consider that this could also imply that part of the reason married people are happier is because marriage increases their relationship satisfaction. Again, to quote the authors, “Married people may also have greater trust in the long-term prospects of their relationship given that marriage is usually intended for life.” Concluding that marriage does not lead to greater happiness after controlling for relationship satisfaction is a bit like concluding that the frequency of sex does not contribute to happiness after controlling for sexual satisfaction. In other words, satisfaction may be the “real” determinant of happiness, but that doesn’t mean that things that contribute to satisfaction are unimportant.

More Questions

This study raises at least two remaining questions that are important to address. First, if marriage really does not contribute to happiness, why did groups that were less likely to marry (e.g., those with low income or divorced parents) experience a greater marital happiness premium in some countries? The authors answered this one, saying: 

People with a lower propensity to marry may benefit more from marriage because it is recognized not only legally but also by family, friends, and the community, who may then provide greater social support.

I fully agree; I just don’t find this argument consistent with the overall conclusion that “marriage does not lead to higher subjective well-being."

Second, why did Norwegian women experience a marital happiness premium? This one stumped the authors a bit because they expected that the combination of generous family policies, irrespective of legal union status, and the lack of social stigma associated with cohabitation would help close the happiness gap. They correctly argued that a statistically significant happiness gap between marrieds and cohabitants after the whole battery of controls does not prove that marriage has a causal effect on happiness among Norwegian women. There’s lots of stuff like personality that, if they could have measured it, might have made the marital happiness premium “go away.” Nonetheless, they found a marital happiness premium where they least expected it.

Overall, Perelli-Harris and her colleagues convinced me that context does not shape the happiness gap between cohabiting and married people in predictable ways, but for the reasons addressed above, they have not proven that marriage does not play a causal role in determining happiness.

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.