Economic Opportunity and Trump’s Win

There will be an untold number of think pieces about what last night meant. For many people, the assessment is just beginning, because it looked like the nation would narrowly avoid a Trump presidency. But there are others who have been thinking through the consequences of a Trump presidency since he first walked down that golden escalator in 2015.

The conventional wisdom is already congealing around a mistaken set of ideas: that Trump voters are reacting to economic insecurity and duress and that the Republican Party — or, more importantly, conservatism — has to offer them an agenda that looks more like what the center-left prefers or that requires a hands-off approach to entitlement spending.

I’ll be writing up some thoughts here on what’s behind Trumpism and what it implies for policy. In this first installment, let’s look at the claim that it’s economic anxiety or a lack of opportunity that drove Trumpism. A few points:

1. Men’s pay is at an all-time high

Below are estimates I produced from hourly wage data publicly provided by the labor-backed Economic Policy Institute. I modified their estimates using the best cost-of-living adjustment available. (Stop what you’re about to say and read Appendix 2 first.) I then estimated nonwage compensation and added it to wages & salary by comparing the ratio of the former to the latter in the national economic accounts. It makes no sense to exclude fringe benefits from pay; employers presumably would just give their workers cash if that’s what employees preferred. And don’t talk to me about rising health care costs until you read Appendix 1.

In 2015, the hourly pay of the median male worker — the one in the middle of all male workers — was 13 percent higher than in 1973 and 14 percent higher than in 1996. Now, a glass-half-empty person could say that essentially all of that increase came in the five years between 1997 and 2002. She could also say that pay in 2015 was no higher than in 2007 or 2000. But it is not lower than before. Indeed, the 2016 data to date indicates that median hourly pay among men at the end of 2016 is higher than it has ever been. Ever.

And trends in household income look even better because women have done so well. A declinist will immediately suggest that household incomes have risen only because women have bailed out their husbands. But (1) men’s pay hasn’t declined, (2) the share of the population married has fallen, and (3) the rise in female labor force participation has occurred primarily among the wives of high-earning men, not working-class husbands. The decline in marriage is a primarily cultural phenomenon, not an economic one. Indeed, I suspect the conventional wisdom has the causality reversed — the decline in marriage (and declining expectations of men generally) has pulled downward on men’s earnings rather than lower earnings hurting marriage rates.

What has changed for men is their standing relative to women.

2. Men’s pay advantage over women has plummeted

The two periods of male pay stagnation in the first chart are different beasts. Women felt the most recent one too, which probably had a lot to do with the aging of the population (older workers are past their prime earning years). But the pre-1997 stagnation was unique to men. This was a period in which the labor force participation of wives — which began rising in the 1940s, in the thick of the Golden Age of the mid-20th-century — really ramped up. It was also a period of declining unionization.

Those on the center-left tend to view unions romantically as institutions that helped give workers what they deserved. But in practice, unions functioned as a way to deliver a “family wage” to working- and middle-class husbands — often exceeding the objective value these workers provided to their employer — in order to preserve a patriarchal family and society. Unions actively discriminated against women. Even in the 1940s, local union contracts included provisions requiring employers to let go women workers when they got married. The minimum wage was originally a tool to put a floor on how low women could reduce the anticompetitive pay of male union workers; there were sometimes lower minimums for women than for men. (Generally, unions opposed minimum wages for men because they reduced workers’ incentive to organize.)

Male pay actually grew faster than productivity increased between the mid-1930s and the early 1950s — roughly corresponding to the era between the labor-friendly Wagner Act and the union-deterring Taft-Hartley Act. That stagnation during the 1970s, 1980s, and early 1990s? That was a correction in pay to allow productivity to catch up to a level that justified the legacy pay of the male workforce in an era where “breadwinner premiums” were anachronistic and unaffordable.

As the second chart above shows, while men haven’t become worse off in absolute terms, they have become dramatically worse off relative to women. Trumpism is essentially a politics of nostalgia. A March survey from Pew asked respondents to compare life today for “people like you” to fifty years earlier. Three in four Trump supporters said that life was better half a century ago. Obviously a lot of these respondents were women — the point isn’t that men are pigs, it’s that cultural traditionalists of both sexes long for a time when the country was quite different and want to “make America great again.” (And let’s keep in mind that whites’ standing relative to nonwhites has probably also worsened, though less than many believe, as we’ll see.)

3. Falling labor force participation among working-age men is almost entirely accounted for by an increase in the share of nonworking men who say they do not want a job (not even maybe)

Consider non-retired, non-disabled men age 25 to 54 not working or looking for work. In 2006, on the eve of the Great Recession, just 27 percent of these men said they wanted a job (or might, or that it “depends”). That was probably a historical low. In 1993, the last year in which all out-of-the-labor-force working-age men in the Current Population Survey (including the disabled and retired) were asked if they wanted a job, 30 percent said they did. The rise in the number who do not want a job statistically accounts for almost the entire increase in working-age men neither working nor looking for work. Your explanation for falling male work rates that focuses on weak demand for working-class men has not grappled with these facts.

4. Opportunity has not fallen in the United States, and it remains substantially higher for white men than for black men

Question: if you’re a fortysomething white man who was raised in the bottom fifth of parental income in the late 1970s, what are the chances that you remained in the bottom fifth of male earnings? If family background didn’t matter at all, you’d have a 20 percent chance. In actuality, 25 percent of fortysomething white men raised at the bottom made it out as adults. That’s basically Denmark levels of opportunity.

What if you’re a fortysomething black man? Try again. Nope, worse. 48 percent.

Second question: what percentage of fortysomething white men and women were actually raised in the bottom fifth? That’s right, 13 percent. What percent of fortysomething black men and women? Yeah, 51 percent.

These figures come from my own ongoing analyses of the National Longitudinal Survey of Youth. Not only does economic mobility remain much greater for white men than for black men, most of the research on trends in economic mobility in the U.S. indicate that it has been roughly stable (combining all men) over the past fifty years. There is a similar stability in subjective assessment of opportunity. A 1939 Roper poll asked men whether their sons’ “opportunities to succeed” would be better or worse than their own. 58 percent indicated their sons would have more opportunity. In 1997, the same question was posed by Gallup. This time 62 percent said their sons would have more opportunity.

5. Drug deaths are up, but the trend has little relationship to how the economy is doing

The view that Trump voters are basically white men whose days are numbered because of the opioid epidemic took off when Princeton economists Anne Case and Angus Deaton published a paper showing a rise in middle-age deaths among whites in the U.S., one that seemed related to overdoses of pain medication, legal or illicit. Subsequent reanalyses corrected the misinterpretation that mortality among white men was on the rise — it turns out the trend was confined to southern white women — but the trend for white men did look worse than for nonwhites, and it was worse than in other countries.

And drug emergencies and deaths have certainly risen. My chart shows two crude indicators. The first series is from 1978 to 2001 and comes from emergency room reports collected by various federal agencies. The second runs from 1999 to 2014 and is based on mortality data from the National Vital Statistics System.

This is an extremely worrying trend. However, I challenge you to find me a relationship to how the economy is doing in there. Compare it with the pay trends in the charts above. Drug emergencies and deaths have risen over the long-run, but the economy has simply not steadily deteriorated over this period.

It is remarkable how quickly we jump to economic explanations, confronted with evidence of some societal problem. I provided evidence during the primary season that Trump didn’t do especially well among voters who said in exit polls that they prioritized the economy. Other research by Jonathan Rothwell and Pablo Diego-Rosell and by John Sides this year also found a weak case that economics was driving Trump’s support. Trump supporters are less downscale than most people believe, and American sentiment about the economy has been improving and is not especially downbeat. Within the consequential state of Michigan, it turns out that Trump supporters in the primary were not especially concentrated in areas with manufacturing losses.

To be sure, there was some survey evidence that found correlations between subjective reports of one’s financial situation and support for Trump, but these findings must be weighed against the evidence that the economy simply wasn’t the driving factor behind Trump’s support. There is little empirical support for the idea that “it was the economy, stupid.”

That’s not to say that I have great evidence for alternative explanations — it’s much easier to debunk claims that seem wrong than to prove claims that we think are right. But I’ll speculate a bit in my next post this month. To preview, I think that the politics of nostalgia, the declining relative status of men and of whites, the cultural traditionalist-cosmopolitan divide, and a crisis of meaning are more important factors, and those have very different policy implications. They are much more difficult to address with public policy, unfortunately. But conservatives and progressives will have to wrestle with them nonetheless.