The Odds That a Marriage Failed
[Update January 2015: I belatedly found a mistake in this post's second chart. The GSS variable for "ever divorced or separated" includes only those who are currently married or widowed; it does not include those who list their current status as divorced or separated, and thus it produces numbers that are too low. The chart below is corrected.]
On Tuesday, our update featured Kay Hymowitz's Family Studies writeup of an important new study (PDF) on divorce trends. Most research has concluded that divorce fell off after the 1970s, but the new study's authors say that some key factors have been neglected. From the paper:
In particular, there are reasons to be skeptical about significance of the decline in the divorce rate since 1980. The refined divorce rate does not take into account changes in distributions of age, marriage duration, or age at marriage. Over the past three decades, the population has grown substantially older, the average duration of marriages has grown, and age at marriage has increased. Since older people, those who have been married a long time, and those who marry later in life are at comparatively low risk of divorce, one would anticipate a significant decline in divorce rates simply because of changes in the characteristics of the married population. Because age and duration specific divorce rates from vital statistics are no longer published, it is difficult to estimate how much of the recent change is merely a reflection of change in the demographic composition of the married population.
After correcting for these problems with a new data set, the authors conclude:
1. Controlling for basic changes in the composition of the married population, overall divorce incidence has risen substantially in recent years, especially among older adults.
2. Although divorce rates among younger cohorts are stable or declining, the rapid increase in cohabiting unions -- which are much more unstable than marriages -- means that overall union instability is rising among all age groups.
After reading this, I was curious to see how various birth cohorts stacked up in the marriage-and-divorce arena, so I made some simple graphs with numbers from the General Social Survey. The survey began in 1972, so I started with the 1932-1941 cohort, whose members were 40 or younger at that point, and went up from there.
Here's how likely you were to be never-married at each age, given your birth cohort (the data were a little jittery, so these are three-year rolling averages):
Even more interesting, here are the odds that you'd ever been divorced or separated, assuming you got married to begin with (five-year rolling averages):
This graph is a lot messier, and it shows why the authors of the new study had to dig so deep into the data to find the truth. At first it seems to show two high-divorce cohorts (those born in 1942-1961), a cohort that started that way but lost its zeal for family dissolution in its 40s ('62-'71), and two low-divorce cohorts (the youngest and oldest). But we're forgetting that later cohorts waited longer to get married, meaning that at each age the average ever-married member had had less time to get divorced. And it's especially wrong to compare the folks born just before Millennials with people born in the 1930s: Not only are today's young adults far more hesitant to get married to begin with, but many of them cohabit and break up without ever tying the knot.
There's no way to make Baby Boomers look better, in other words, but things aren't as rosy for young people as they might seem.
Robert VerBruggen is editor of RealClearPolicy. Twitter: @RAVerBruggen