There is a town in western Japan named Nagi that’s famous for making babies. Its fertility rate in 2021 was 2.68 lifetime births per woman, compared with 1.3 for Japan as a whole, according to an article in The Wall Street Journal that my Opinion colleague Jessica Grose recently cited. Delegations from elsewhere in Japan and abroad have come to Nagi to learn its secret formula. Is it the free medical care for all children? The affordable child care? The cash gifts to new mothers?
I’ve been considering another theory. Maybe people in Nagi are having babies because other people in Nagi are having babies. That would be what economists call a “peer effect.” We are social animals and we take our cues from family, friends and sometimes even passers-by. Peer effects could help explain the decline in fertility as well: It could be that in most of the world, people are having fewer babies partly because other people are having fewer babies. That would explain why so many towns and countries are putting in Nagi-like pronatalist measures but not getting Nagi-like results.
Researchers have claimed to find peer effects on obesity, smoking and drinking, so it’s plausible that they influence fertility. We all know of siblings, in-laws and friends who have their first children around the same time. George Akerlof, a Nobel laureate economist from the University of California, Berkeley, wrote in a 1997 paper in the journal Econometrica that “social decisions — such as the demand for education, the practice of discrimination, the decision to marry, divorce and bear children, and the decision whether or not to commit crimes — are not simple choices based primarily on individual considerations.”
I came across two problems as I looked into peer effects, though. One is that it’s not clear what to do about them. Let’s say there are two equilibria, one low-fertility and one high-fertility, and both are the result of peer effects. How would a government planner or a church or a nonprofit or anyone else flip a society from the low-fertility equilibrium to the high-fertility equilibrium?
Another problem is whether peer effects are real. Other explanations for high or low fertility are also plausible, after all. Maybe friends behave the same way because they’re alike, not because one influences the other. Or maybe they’re all influenced by some outside factor rather than each other. Dozens of such factors have been suggested: Working parents don’t get the support they need. Young people can’t afford a dwelling that’s big enough for children, or they’re saddled with student debt. The opportunity cost of having children is higher for people with higher education and better earnings prospects. Or maybe women simply can’t find suitable mates.
In studying peer effects, “The scope for spurious correlation in peer analysis is wide,” Joshua Angrist, an economist at the Massachusetts Institute of Technology who is also a Nobel laureate, warned in 2014 in an article in the journal Labour Economics titled “The Perils of Peer Effects.”
Scholars who aren’t careful can get caught going around in circles. If there is a peer effect, the same women causing it are also being affected by it, making causality almost impossible to isolate. As Angrist put it, there needs to be “a clear separation between the subjects of a peer effects investigation and the peers who provide the mechanism for causal effects on these subjects.” (Here’s a link to an early working paper version.)
Even scholars who avoid that circularity can run into trouble determining the direction of causality. One solution is the instrumental variables approach, of which Angrist is a master, as I wrote in 2021 after he shared the Nobel. To study the relationship between military service during the Vietnam War and subsequent earnings, he looked at draftees’ lottery numbers. Men with low numbers were more likely to serve, and there was no risk that people who drew low numbers were systematically different from those who drew high ones.
Angrist wrote me an email on Thursday to explain why he’s cautious about asserting peer effects: “People who live, work, and study near one another — peers — tend to be similar. But there are all sorts of reasons for that, few, if any, causal. For instance, my neighbors on both sides and I all drive late-model German luxury cars. But I don’t think my (or their) peer pressure is to blame for this. In fact, they got theirs before I even moved in.”
I read several papers on peer effects on fertility with Angrist’s caveats in mind. One, by Jason Fletcher and Olga Yakusheva, looked at American teenagers and found that a 10 percentage point increase in pregnancies of classmates is associated with a 2 to 5 percentage point greater likelihood of a teenager herself becoming pregnant.
Disentangling causality is “a really hard problem,” Fletcher, an economist at the University of Wisconsin’s La Follette School of Public Affairs, told me. He and Yakusheva, who is at the University of Michigan School of Nursing, tried to identify causality by using a trick similar to the one Angrist used to study the impact of military service on veterans’ earnings. They needed a factor that was correlated with a girl’s likelihood of becoming pregnant, but with no risk that the factor could be caused by the girl (reverse causality). They found two such factors, called instrumental variables: whether her classmates began to menstruate early and whether the classmates were themselves children of teenage mothers. Those factors made the classmates more likely to become pregnant, which in turn influenced the fertility outcomes of the girls being studied.
Amalia Miller at the University of Virginia is one of four scholars who studied fertility in workplaces in Denmark. If two women in an office got pregnant around the same time, it was hard to tell if A influenced B or B influenced A. They came up with the idea of looking at the fertility of A’s sister, realizing that if A’s sister got pregnant, A herself was more likely to get pregnant. That was their instrumental variable. Miller and her research partners found that for more-educated women, a colleague of the same education level having a baby increased their own chance of having a baby. Less-educated women were less likely to have a baby when a colleague of the same education level had one. The negative peer effect for less-educated women “could come from a desire to distinguish oneself from one’s peers,” among other factors, they speculated.
The third person I interviewed is Nie Peng, a professor at Xi’an Jiaotong University in Xi’an, China. He and two co-authors studied Chinese women ages 18 to 49. They found that an increase in the fertility of peers reduced the probability that women would want only one child and increased the probability that they would want three, four or more children. The research “provides support for the role of social norms in the fertility choices of reproductive-aged Chinese women,” they wrote.
Nie told me he read Angrist’s paper years ago and kept its caveats in mind as he designed his team’s research. In any case, he said, peer effects aren’t the whole story. Even though Chinese authorities desperately want to raise the nation’s birthrate, it’s still hard for young families to raise children, he said. He said he and his wife have one child, a 5-year-old girl. When he picks her up from kindergarten, he said, there’s a long line — of grandparents, because all the other parents have to work.