For decades, sociologists have argued that a lack of jobs for young men can reduce their “marriageability.” Call it sexism, culture, biology, or whatever you want, but a man without a job is a far less desirable partner. Therefore, when work dries up for men, women put off marriage or give up on it entirely, leading to an uptick in out-of-wedlock births and the attendant social ills.

This is intuitively plausible, and it’s hard to deny that communities with high male unemployment tend to have high rates of nonmarital childbearing as well. But it’s a difficult hypothesis to prove. Is a lack of work for men really a root cause of family disintegration, or are the two correlated for some other reason?

groundbreaking new study by David Autor, David Dorn, and Gordon Hanson gives us the best evidence to date for this theory. Trade with China has increased markedly since the 1990s, and especially since 2001 (when China became a member of the World Trade Organization), contributing to the decline of American manufacturing jobs—jobs that pay well and are disproportionately held by men. This sets up a natural experiment: If male job loss causes family breakdown, this effect should be easily visible in the areas most exposed to new competition from China, and it should vary depending on how male- or female-dominated the targeted jobs were. That is exactly what the study found.

The first link in the chain here is that men, especially lower-wage men, suffered disproportionately in the areas (more specifically, “commuting zones”) hit by a trade shock. When import penetration rose by a point—roughly the average change over a decade in the U.S.—the male-to-female earnings advantage fell by 2.2 percent at the median and almost 17 percent at the 25th percentile.

In addition to earning less, men in these areas responded with a variety of behaviors, some more adaptive than others. Some joined the military and left the area. Deaths from drug overdoses, poor diet, and smoking increased. It’s possible that others turned to crime and were incarcerated, though the available data make it hard to say for sure. (We know a lot of men disappeared from these areas, but not where they went. The Census tracks prisoners where they are incarcerated, not in the areas they came from.)

Men’s interactions with women also changed. A one-point import shock slightly reduced the percentage of women who were married. It also reduced births, but this effect was concentrated among older and married women, meaning that the percentage of children born to single mothers rose. As the authors write, “a decline in male earnings spurs some women to curtail both motherhood and marriage while spurring others to exercise the option of single-headedness (curtailing marriage but not fertility), thus raising teen and out-of-wedlock birth shares.”

The authors were further able to confirm that it’s male earnings that create this effect because the trade shocks varied in how much they targeted male vs. female employment. Shocks to female employment had the opposite effect, increasing fertility and marriage rates.

These changes, in turn, affected children. A one-unit trade shock increased child poverty by 2.2 points and decreased the percentage of children living with married parents by 0.4 points.

This doesn’t mean we can blame China for the fall of the two-parent family; indeed, the trend began long before Chinese products started making inroads in the U.S. market. The authors estimate that trade with China reduced:

marriage prevalence among young women by 0.75 to 1.25 percentage points; rais[ed] the share of teen and non-marital births by approximately one-half a percentage point each, decreas[ed] the fraction of children living in married two-parent households by roughly a half percentage point, and substantially hik[ed]—by 2.2 percentage points, a 13 percent increase—the fraction of children living in poverty.

If these findings can be extrapolated to other situations, though, the impact could be far more substantial. By one estimate, trade accounted for just 13 percent of the total manufacturing job losses; automation was a much bigger factor. And the male earnings advantage has been eroded from the other side as well, as women have made advances in the labor force.

The idea that married men should be breadwinners still looms large in the collective psyche, in other words, and lower-skilled men are having ever more trouble filling that role.

Robert VerBruggen is managing editor of The American Conservative. Republished with permission from the Institute for Family Studies blog.