The “success sequence” — the idea that the ticket out of poverty is to graduate high school, work full-time, and have kids only after reaching age 21 and getting married — has taken a beating from the left lately.
Some of the criticisms are legitimate. Some people do wind up poor despite following the three cardinal rules, for instance, and this happens for blacks more than for whites. Further, dropping out of high school or failing to work full-time are not always just bad personal decisions; sometimes they’re a result of, say, low IQ, a traumatic home life, or a bad economy.
But a new report from Wendy Wang and W. Bradford Wilcox of the Institute for Family Studies (IFS), cosponsored by the American Enterprise Institute, seeks to rejuvenate this idea. They look at a group of Millennials whom the National Longitudinal Survey of Youth has been following since 1997, and who were last interviewed in 2013 and 2014, when they were 28–34 years old. Wang and Wilcox find that the success sequence does matter, even after accounting for other differences among these young adults. (Full disclosure: I’m a contributor to Family Studies, which is run by IFS and overseen by Wilcox.)
The report’s headline numbers are straightforward comparisons of people who follow the sequence with people who don’t, and the results are about what you’d expect. A whopping 97 percent of Millennials who followed the sequence are not poor; 86 percent of them are at least middle class.
Racial minorities, as well as individuals with lower-income parents, are also more likely to thrive if they follow the success sequence. For example, just 39 percent of blacks who have a child out of wedlock are middle-income or better, compared with 76 percent of those who got married first and 59 percent of those who were unmarried and childless when last interviewed. Similarly, among those who grew up in lower-income households, there was only a 9 percent poverty rate for those who followed the success sequence, compared with a 58 percent poverty rate for those who missed all three parts of it.
But these differences say little about whether the success sequence causes better outcomes. It may be that the success sequence just tends to go hand-in-hand with other advantages, such as rich parents or high intelligence. To find out, ideally we’d conduct an experiment: randomly assign certain people to have kids out of wedlock or drop out of high school, and see what happens. That’s not possible, so unfortunately we’re stuck with inferior ways of trying to make sure we’re not comparing apples and oranges.
The authors deploy the usual statistical technique, a “regression” in which various personal characteristics are “controlled.” The results are less dramatic than the original comparisons — but only a little. Even with race, ethnicity, age, sex, the results of an academic/intelligence test (the AFQT, or Armed Forces Qualification Test), and parental income taken into account, getting married before having kids instead of the reverse reduces one’s odds of living in poverty by 60 percent. Getting a high-school degree instead of dropping out cuts those odds 38 percent, and working full-time drops them by 66 percent.
The flip side is that accounting for those other factors reveals their own power. Growing up in a higher-income family instead of a lower-income one cuts one’s odds of living in poverty by 47 percent — again, after controlling for all the other variables. A high AFQT score reduces one’s odds of poverty by 46 percent. Being black instead of white increases them by 85 percent.
These statistical exercises always should be taken with a grain of salt, because researchers have to make a lot of subjective decisions that can alter the results, and we can never be confident we’ve controlled all the variables that could confound the results. With this particular study, my own biggest concern is that AFQT scores and parental income are not measured in enough detail. The former is collapsed into two categories, “Above Median” and “Below Median,” meaning that a super-genius is treated the same as someone who is just barely above average in smarts, while parental income is divided up into “Low,” “Middle,” and “High.” This approach fails to capture the full range of effects that one’s intelligence and one’s circumstances growing up can have on one’s chances.
But at any rate, the report provides empirical support for an idea that, deep down, no one really rejects. How many of the liberals criticizing the success sequence would be okay with it if their own child dropped out of school, had a kid at 17, and left the labor force to play video games?
So it’s worth asking what the trends are for these three key indicators and what we might do to change them.
Unmarried childbearing is the step where we see the clearest disintegration. It’s hardly news that the percentage of kids born out of wedlock has skyrocketed over the past half-century, but Wang and Wilcox report that an outright majority, 55 percent, of Millennial parents in their study weren’t married before their first kid was born.
I have a significant quibble with that number, though: Nowadays, I don’t think we can judge a cohort’s childbearing decisions based on a survey conducted when they were 28 to 34 years old, because at that point a lot of the decisions haven’t been made yet. Roughly 30 percent of first births are to women in that age range, meaning they won’t necessarily be counted (e.g., if someone is interviewed at 29 and has a baby at 32), and another 9 percent are to women age 35 and older, meaning they definitely can’t be counted. And of course, older mothers are much more likely to be married; few teen moms have tied the knot, but the vast majority of women giving birth in their 30s have.
To put it differently, imagine if the survey was conducted when Millennials were teenagers. The parents in the sample would be overwhelmingly unmarried — about 90 percent — but this alarming statistic wouldn’t say much about what the rest of the group would do. The same problem occurs, albeit to a lesser extent, when you look at 28- to 34-year-olds.
If I had to guess, I’d say 45 percent of Millennial parents, not 55 percent, will not get married first when all is said and done. There’s no denying that’s worse than the numbers we saw decades ago, but nationwide statistics are probably a better source for such data than a survey of people as young as 28.
The new report doesn’t dwell on trends in work, but in Men without Work, Nicholas Eberstadt finds that increasing numbers of men are dropping out of the labor force — and after rushing into the work force over the preceding several decades, women started leaving around the turn of the century, too.
Wang and Wilcox say that only half of Millennials either have followed the success sequence or are ‘on track’ to.
It’s not all bad news, though, as Annie Lowrey of The Atlantic and Ron Haskins of the Brookings Institution pointed out at an event debuting the report yesterday. High-school-degree attainment among 25- to 29-year-olds has risen from about 85 percent in 1980 to above 90 percent today, and teen childbearing has fallen more than half since 1990.
But all told, Wang and Wilcox say that only half of Millennials either have followed the success sequence or are “on track” to (i.e., they are unmarried and childless thus far). That number falls to 24 percent for blacks, and to 31 percent among individuals who grew up in lower-income households.
The report’s final chapter runs down a list of policy implications — mostly proposals that those of us who follow such things are already familiar with. These include vocational training and apprenticeships for kids who are not academically inclined, wage subsidies to encourage work, and campaigns to encourage marriage.
The government has a limited ability to change our decision-making, especially when it comes to decisions involving marriage and kids. But Wang and Wilcox have provided solid evidence that those decisions do, in fact, matter when it comes to whether individuals thrive in modern America.
— Robert VerBruggen is a deputy managing editor of National Review.