Puzzling NBER Reports on Blue-Collar Men, Marriage and Manufacturing

Two studies of men, marriage, and manufacturing in the National Bureau of Economic Research (NBER) working paper series stand in some tension with each other. The first study, “When Work Disappears: Manufacturing Decline and the Falling Marriage-Market Value of Young Men,” by David Autor, David Dorn, and Gordon Hanson, finds the loss of manufacturing jobs and wages decreases marriage and fertility, and increases the proportion of unwed mothers and children living below the poverty line. A second study, however, “Male Earnings, Marriageable Men, and Nonmarital Fertility: Evidence from the Fracking Boom,” by Melissa S. Kearney and Riley Wilson, observes an increase in blue-collar jobs and wages and find no corresponding increase in marriage rates. They do, however, find an increase in births, but among both the married and the unmarried.

Here’s the puzzle between the two studies: If the disappearance of good-paying manufacturing jobs leads to a decrease in blue-collar marriage rates, then why doesn’t the creation of good-paying manufacturing jobs lead to a corresponding increase in blue-collar marriage rates?

I discuss a couple of possible answers below. First, however, given the spin on these and related studies recently in conservative circles—that these findings imply an indictment of the market economy—it’s important to underscore that the causal factor in Autor, Dorn and Hanson’s study and in related studies: The identified causal factor is the loss of good-paying manufacturing jobs, not the market. To be sure, these losses in the U.S. occurred in a market economy. Yet unless one argues—implausibly—that underlying patterns of production do not change in non-market economies (due to changing technology or changing national comparative advantages or costs of international trade, etc.), then it is these changes, not the market system per se, that generate the studied pathologies.

But what about the puzzle? Kearney and Wilson provide some comparative leverage on their finding by examining a similar study related to the Appalachian coal boom of the 1970s and 1980s. In this study, both marriage rates and marital births increased. But non-marital births did not increase. They suggest changing social attitudes might account for the different experience in marriage rates (and non-marital births) between the earlier Appalachian experience and t-he more recent fracking experience.

Their claim is certainly plausible if we look only at the significant social changes between the 70s and 80s, on the one hand, and the 90s and 00s on the other hand. But that doesn’t help with the puzzle relative to the findings of Autor, Dorn, and Hanson. Autor, et al., study the impact of the loss of manufacturing jobs and wages during the same time period in which Kearney and Wilson study fracking gains. So why not opposite effects given opposite movements in the underlying cause during the same time period?

I can think of at least two possibilities. One relatively narrow, the other relatively broad.

First the narrow explanation. The difference between the coal boom and the fracking boom isn’t the time period, it’s the transience of the population in the boom areas. To wit, the boom in Appalachia occurred with a regional population that existed (in large part) prior to the boom. That is, the boom affected an intact population with existing social connections and settled social patterns. In short, the girls already knew the boys and the boys already knew the girls.

In contrast, the fracking boom occurred in much more sparsely populated regions. These regions experienced significant numbers of new arrivals as a result of the fracking boom than did Appalachia as a result of the coal boom. Further, expectations were that increased oil production due to fracking would be a transient affair, lasting for a few years then drying up again. In short, in regions of the fracking boom, the girls needed time to get to know the boys, and vice versa. And, also, the expectation was that the boys would likely disappear in a few years when the fracking gains disappeared.

A second explanation for the puzzling asymmetry is broader, and more ominous: Social capital is more easily destroyed than created. Whether caused by economic losses—Autor, Dorn and Hanson’s wage decline relative to Kearney and Wilson’s wage increase—or caused by changed social expectations and legal regimes—Appalachia of the 1970s versus West Texas of the 2000s—the loss of long-standing social habits and expectations that support the creation of intimate communities like marriages cannot be recreated overnight. Marriage rates ratchet down as a result of the impact of deleterious economic and social changes, but marriage rates do not automatically rise with the removal of those deleterious factors.