I’m attending two weddings this weekend, from which I hope two long and happy marriages will result. Having no insights of my own to add to the topic of marriage and happiness, I took another look at the subjective well-being literature on the subject. As I noted in a Catallaxy post eighteen months ago, this area of research is surprisingly controversial, with one prominent happiness researcher denying that marriage brings most people any lasting additional happiness.
One point that is not in dispute is that, at any given time, married people are happier than single people. I had a look at the most recent Australian survey to ask about happiness, the 2005 Australian Survey of Social Attitudes, which finds the same relativities persistently found across time and around the world.
78% of married people rate themselves as 7 or above on 0-10 scale, compared to 63% of the never married. People in de facto relationships were similar to married people in the proportion in the normal 7+ range, but married people were considerably more likely to be in the very happy 9-10 range, 32% compared to 21% among de facto couples (even though you would think that married people would have had longer to grow bored of each other). Separated people are the least happy – 53% at 7 or above, but they get over it, as divorced people are about as happy as the never-married singles.
One reason for some initial doubt that marriage has the widely-assumed happiness benefits is that average happiness has been stable over time. Though it has fluctuated a little between surveys over sixty years, it has fluctuated without trend. Clive Hamilton and many others have seized on this as evidence that greater income does not make you happy. But if marriage makes people happier, or conversely non-marital states make you less happy, shouldn’t the declining share of the adult population who are married have led to a declining average level of happiness?
The differences aren’t trivial – in the 2005 Australian Survey of Social Attitudes singles on average rate themselves as 6.7 on the 0-10 scale, while married people rate themselves as 7.6. The divorced are on 6.9 and the separated on 6.2. Between 1976 and 2001, the proportion of people who are married dropped by 12 percentage points, enough you would think to all other things being equal cause a sustained drop in average happiness levels, if indeed marriage is what takes you from below 7 to 7.6.
Richard Lucas is the marriage sceptic, and he has an explanation as to why average happiness is not showing a sustained decline. In an article in the Journal of Happiness Studies last year he defended his thesis that marriage does not bring happiness. He has two basic arguments. The first is that there is a selection effect, ie people who are happier to begin with are more likely to get married. The second is that there is an adaptation effect – that happiness provides a temporary boost, but after a few years of marriage the couple returns to their baseline happiness (one of the major theories found in the subjective well-being literature is that most people have a baseline or setpoint level of happiness, and that while events can push people up or down, they tend to revert to their natural, personality-determined level). He uses a panel study, which tracked the same people over the years before and after marriage, to substantiate his case.
On Lucas’s account, then, the kind of figures I cited from the Australian Survey of Social Attitudes are largely a statistical illusion, the result of measuring the happiness of already-happy people who got married plus the added happiness of people who have married in the last few years. He does, however, grant one exception – people who marry at later ages do get lasting benefits.
On the other side of the debate are Alois Stutzer and Bruno Frey (a recent journal article on this is on Stutzer’s site). They find similar effects to Lucas – which is just as well, since they are using the same dataset – but note high standard deviations in life satisfaction among married couples.
Contrary to gender equity theory, couples with a high division of labour, essentially a euphemism for the male breadwinner model of marriage, are happier than those in which both work. (Though to avoid bringing the feminist terror down on themselves, these two male authors suggest that this may be because women end up doing all the housework whether they are in paid work or not.) Narrow gaps in education levels also help – presumably giving more shared experiences and conversation topics. So for some couples, adaptation is not complete.
So on the Stutzer and Frey analysis, there are three explanations for the greater happiness of married people evident in the cross-sectional data: selection, temporary high levels of happiness among the newly married, and non-complete adaptation by some couples. Given the nature of the data, nobody can look at one important counter-factual: how happy the people who did get married would have been if they had remained single. The fact that Lucas found that people who married late had lasting boosts to their happiness is perhaps suggestive here – that they were below their happiness setpoint and marriage took them up to it.
This is all interesting (to me, anyway), but I wonder how much we can extrapolate from the dataset being used – the German Socio-Economic Panel – to the effects of well-being and marriage in other countries. One difficulty is that while every country has marriage, it doesn’t mean the same thing everywhere. There are varying degrees of social pressure to get married, and to stay married. In some countries, married women are more likely to work than in others, and the average number of children per couple varies too.
Another potential issue is that though in this post I have followed common practice in using life satisfaction and happiness interchangeably, they are not quite the same. The German question is about life satisfaction, which I take as prompting an overall assessment of one’s life rather than simply asking respondents how they feel, as a happiness question does. Somebody could be satisfied partly because he or she has married – meeting parental expectations and providing a stable family for kids to grow up in – without necessarily rating themselves highly for happiness.
Eventually Australia’s HILDA survey will provide some local answers about how individuals perceive their well-being as they move between being single and being married, and between being married and separated or divorced. When I get time, I will look at some of the older surveys to see if I can find effects consistent with the Lucas argument – essentially, that single people are on average happier now than in the past because more of the already-happy group is postponing marriage or not getting married at all.
But I think what we can say is that marriage produces short to medium term gains, but after that it is a gamble. Some couples do achieve lasting happiness or life satisfaction benefits, but others end up joining one of the unhappiest groups in society, the separated.