The Truth About Jobs
The great American job machine is sputtering, but it has not lost any of its underlying power.
Hardly a week goes by without at least one commentator somewhere in America heralding the demise of the middle-class worker. Because of the historic severity of the Great Recession and its aftermath, it is not hard to stoke anxiety. The unease coheres in a conventional wisdom that connects a number of short- and long-term economic trends with today’s tepid conditions, creating a fearful narrative about the future of jobs in America. It is a narrative that mischaracterizes the past and only feeds the anxieties it claims to explain.
In the conventional story, the economy has been unable for decades to produce jobs for all the people who want them. It is also said that the middle class is becoming “hollowed out” as job growth is increasingly confined to occupations that require either very low-level skills or highly sophisticated ones—and that pay accordingly. “Job polarization,” as this pattern is called, has been driven by technological changes that have automated many “middle-skill” jobs and encouraged their offshoring to lower-paid workers in other countries. Increasingly, we are warned that robots and other products of the information technology revolution will usher in mass unemployment in the not-too-distant future. Wages will continue to stagnate or decline despite rising productivity, as they have done for decades.
How worried should we be that such a dark future awaits? In answering that question, it makes sense to focus on the experience of men during their prime working years. Women have seen such strong gains in education and employment over the past few decades as a result of increasing gender equality that it is very difficult to draw many broad conclusions about the underlying condition of the economy from their experience.
At first glance, men have had a hard time of it. Ninety-five percent of men between the ages of 25 and 54 were working in an average week in 1969. By 1983, the employment rate among this segment of the work force had fallen to 86 percent, and it fell again, dramatically, with the onset of the Great Recession. Employment had begun to recover by 2012, but it still stood at just 83 percent.
One reason employment fell is that it became harder for those looking for work to find a job. But while the unemployment rate is abnormally high at the moment, over the longer term it rises and falls with the business cycle. The main reason employment has declined is that more and more men are not looking for work. Between 1969 and 2012, the share of men ages 25 to 54 who were out of the labor force rose from four percent to 11 percent.
Remarkably, if one had forecast the 2012 labor force participation rate (the fraction of men working or looking for work) by simply extending the curve of the 1948-to-2000 decline on a graph, the prediction would have exactly matched the actual 2012 rate. That suggests that whatever is causing the rise in the share of working-age men who are out of the labor force is much more deeply rooted than the past few years of economic ups and downs.
What could be behind this change? People can be out of the labor force for reasons other than despair of finding a job. Many have illnesses or disabilities. Others, even in the 25-to-54 age group, are full-time students. Some are able to retire early. A small number of men have primary responsibility for maintaining their home while a partner works. Others are sustained by unreported sources of income, including off-the-books jobs.
In fact, the vast majority of working-age men who are out of the labor force today tell survey researchers that they do not want to work. When you factor in that preference, the story of employment decline begins to look quite different. By the conventional measure, employment among working-age men declined by seven percentage points between 1969 and 2007, just before the Great Recession, reaching 88 percent. But in my analysis based on “work-interested” men in those years, the drop amounts to just three percentage points—hardly a dire trend. (Counting only work-interested men, 94 percent were employed in 2007.)
To be sure, many of the men who are uninterested in working are only out of the labor force because federal disability benefits have been steadily extended to people who in the past would have had to look for work. Beyond its historic role as a safety net for those with severe conditions, disability has become a welfare program for able-bodied men with low skill levels, since the kinds of jobs they can get don’t command high wages. The number of working-age men drawing benefits has climbed over time, and there has been no overall increase in the incidence of health problems to explain it. Still, the rise in the number of disabled men who say they do not want to work is too small—adding another point to the three-percentage-point decline—to alter the conclusion that the drop in male employment has been modest.
When one looks beyond the core working-age group of men, the long-term employment picture is no more worrisome. Men under 25 have seen a large decline in labor force participation since 1979, but U.S. Department of Education statistics show that this decline is mostly explained by rising high school and postsecondary enrollments. Among 18-to-19-year-old men, school enrollment rose 20 percentage points from 1980 to 2010, while labor force participation declined by 22 points. Among their slightly older peers in the 20-to-24-year-old group, school enrollment gains fully offset the participation decline.
Among men older than 54, the labor force participation rate has actually been rising for the past couple of decades, reversing an old trend. That alarms some observers, who argue that it is another sign of distress—that men are being forced out of retirement or have been unable to retire in the first place. But research by Brookings Institution economist Barry Bosworth suggests that this increase has been concentrated among the best-educated and highest-earning workers, often men who are staying on the job less for a paycheck than for mental stimulation, camaraderie, and other intangible benefits.
If the total supply of jobs has not shrunk that much over the long run, what about the supply of good jobs? In recent years, the work of MIT economist David Autor showing an increase in what he calls “job polarization” has stirred fears that the middle class is being “hollowed out.” Job growth, Autor suggests, has been occurring primarily in occupations that pay either quite well or quite poorly. Solid middle-class occupations—such as clerical, administrative, and production jobs—have seen slower growth or outright declines. A future in which the occupational structure was shaped like an hourglass—fat at the top and bottom and thin in the middle—would condemn us to rising inequality and perhaps diminished economic mobility.
A growing number of routine tasks, Autor argues, can be done by new information technologies or by lower-wage workers in other countries. Increasingly, jobs in the United States will require either abstract skills associated with high levels of education and intelligence or—because nobody has figured out a way to offshore the jobs of short-order cooks and house painters—more basic skills requiring no formal schooling. Jobs in the first group pay well because the demand for abstract thinking outstrips the number of workers who can supply it, while those in the second group pay poorly because so many people can do the work.
Autor’s research, however, has been blown out of proportion by its popularizers even as it has been effectively challenged by other economists. Harry Holzer, of Georgetown University, and Robert Lerman, of the Washington-based Urban Institute, for example, conclude that there has been only a modest decline in “middle-skill” jobs, from 55 percent of the total in 1986 to 48 percent in 2006. “Stories of dramatic polarization . . . seem inconsistent with these facts,” Holzer has written. He and Lerman predict that middle-skill jobs will account for 40 to 45 percent of new hiring in this decade, with particularly strong demand for certain types of workers, such as “technicians, licensed practical nurses, and therapists in health care.” Holzer writes that there will be “substantial opportunities for earnings improvements to many youth and adults for whom a bachelor’s degree might be out of reach.”
The Washington think-tank trio of Lawrence Mishel and Heidi Shierholz of the Economic Policy Institute and John Schmitt of the Center for Economic and Policy Research confirm that the share of “middle-wage” jobs declined only modestly from the late 1980s through 2007. But they see a much steeper decline if the beginning date is stretched back to 1959, from 66 percent to 48 percent. That might be alarming, except that it mostly reflects a net upgrading of jobs. “High-wage” jobs grew from 21 percent of all employment to 34 percent from 1959 to 2007. Employment has shifted much more from middle-paying jobs to higher-paying managerial, professional, and technical positions than from middle- to low-paying jobs.
Indeed, this dynamic held in each decade from the 1960s to the ’90s. From 2000 to 2007, growth was strongest in low-paying jobs, but that was a period dominated by two trends—slow growth in the supply of native-born workers (due to an aging population) and a large increase in the number of immigrants with lower levels of schooling. It is always true that the supply of jobs depends significantly on the supply of labor—an important fact to remember when we evaluate the future of job growth.
While Autor emphasizes the threat of automation, Princeton economists Alan Blinder and Alan Krueger stress the negative effects of offshoring. They estimate that, in principle, a quarter of American jobs are “offshorable,” in that they do not require working in physical proximity to colleagues or customers. However, just because a job is offshorable does not mean it will be eliminated. If the benefits of face-to-face interaction among workers were small, employers would not go to the trouble and expense of bringing workers together in central offices and dense metropolitan areas. Perhaps more important, our history of job upgrading shows that the ill effects of offshoring can be offset. Blinder and Krueger surmise that a quarter of jobs were also offshorable in 1960, and the United States did indeed send many manufacturing positions and other work overseas in the ensuing decades. But the larger story is that the economy adapted to change, and, thanks to continuing domestic job growth, the period brought steady increases in higher-paying occupations.
We have a tendency, when thinking about technology and trade, to zero in on their harmful effects. But they also have a strong upside: lower prices for American consumers for everything from toothbrushes to refrigerators. Evidence from economists Christian Broda and John Romalis, for example, suggests that thanks in part to imports from developing countries, the cost of living over the last 20 years has risen less among lower-income Americans than among richer households.
Consider the impact of Walmart, the often-maligned retail colossus. Jason Furman, the incoming chair of President Barack Obama’s Council of Economic Advisers, noted back in 2005 that Walmart’s “everyday low prices”—the result of relentless efforts to minimize costs, including reliance on imports and foreign labor—had effectively boosted American family incomes by 1.5 to 4.5 percent. In total dollars, that benefit was more than $100 billion, at least 20 times the reduction in wages that Walmart’s critics claimed the company caused.
The poor, Furman found, benefited more from lower prices than others because a bigger share of the things they bought—clothing, groceries, paper products—were goods sold by Walmart. Even if one assumes that the bottom fifth of households bears the entire cost of wage reductions caused by Walmart (a figure that is likely exaggerated), the price-lowering benefits of Walmart for this group are still two and a half times the costs.
In short, if trade and technology reduce demand for labor, the lowered labor costs paid by businesses will translate into lower prices. That can be expected to benefit Americans—including lower-income families—in the aggregate, despite the highly visible costs to those who bear the brunt of the resulting economic dislocations. The dystopic fantasy of an economy based on robots and overseas suppliers with mass involuntary joblessness at home will simply not come to pass.
The widespread feeling that the American economy is failing the middle class owes a great deal to the belief that wages have stagnated or declined. That belief is only half correct. For women, wages have risen smartly. For example, The State of Working America, an annual report by the Economic Policy Institute, indicates that median hourly wages among female workers increased by 24 percent from 1979 to 2007. That number grows to 35 percent (an increase of $4.60 per hour) with adjustments for the value of benefits such as health insurance and to better account for changes in the cost of living. Among men, however, the adjusted increase was only four percent.
During those same years, pay for both women and men badly trailed productivity, or the value of what workers produce per hour, which rose by 60 percent. We would be very right to worry if that disparity between productivity and wages were to continue. Here again, however, a longer-term perspective provides important context.
The hourly compensation of workers has failed to keep pace with productivity since the mid-20th century, but in the 1930s and ’40s pay raced ahead of productivity gains. By 1950, productivity was 65 percent higher than it had been in 1929, but hourly compensation was 115 percent higher. In contrast, pay and productivity rose by the same amount between 1900 and 1929. Workers in 1950 were making about 30 percent more than their productivity should have dictated. Correcting that overpayment required that compensation growth fall behind productivity growth. As of 2010, workers still made 14 percent more than productivity levels suggested they should have, despite the fact that productivity had grown faster than compensation since 1950.
The current Great Correction in the relationship between pay and productivity has surely been frustrating for men, who have borne the brunt of the pay slowdown. Women, who started from a lower base, have fared much better as a group, moving into better-paying jobs thanks to the erosion of discrimination and occupational segregation. But for women and men alike, there is a silver lining to this story. In time, the Great Correction will run its course, bringing productivity growth and compensation back into long-term alignment. At that point, pay and productivity should begin to move in tandem once more, putting Americans’ wages back on an upward trajectory. When will that happen? It would be foolish to attempt a prediction, but the closing of the compensation-productivity gap has proceeded slowly, suggesting that we may have to wait a while for the Great Correction to end.
The U.S. economy has shown an amazing ability over the course of two centuries to create good jobs for Americans and to supply their wants and needs. A clear-eyed reading of long-term trends does not point to a fundamental breakdown in that ability. Even with the decline of manufacturing and the peaks and valleys of recent decades, the economy has been strong and dynamic enough to create jobs for millions of additional female and immigrant workers. There is no reason to think it cannot adapt to today’s challenges and whatever disruptions may lie in store.
Indeed, Americans, by and large, appear to have a healthy attitude toward the vicissitudes of the job market. According to a Kaiser Family Foundation survey conducted earlier this year, only 20 percent of employed workers are “very worried” about the possibility of losing their job. Thirty-six percent say they are “not at all” worried. And while job anxiety has to be taken seriously, it is not always well founded. After all, 20 percent of adults also say they are very worried about being a victim of gun violence, though their real risks are miniscule, and 15 percent fret about being caught up in a terrorist attack.
For most Americans, anxiety about work is a low-grade background concern, not a dark cloud over their everyday existence. The relentless focus in so much public debate on the most negative evidence, and on economic challenges much more than economic strengths, may needlessly raise anxiety levels. It also distracts us from the real problems we face. These include too many workers with limited skills, the plateauing of college graduation rates, distressingly stable economic inequality between white and black Americans, and persistent inequality of opportunity between children born into advantageous and disadvantageous circumstances. Not everyone faces pressing job insecurity, but we can do better by those at risk if we maintain the proper perspective.