The War on Work — and How to End It
Editor's note: This piece was adapted from the special issue of City Journal, “The Shape of Work to Come”
Focusing solely on providing material benefits to the jobless will not solve the crisis.
In 1967, 95 percent of “prime-age” men between the ages of 25 and 54 worked. During the Great Recession, though, the share of jobless prime-age males rose above 20 percent. Even today, long after the recession officially ended, more than 15 percent of such men aren’t working. And in some locations, like Kentucky, the numbers are even higher: Fewer than 70 percent of men lacking any college education go to work every day in that state.
The rise of joblessness — especially among men — is the great American domestic crisis of the 21st century. It is a crisis of spirit more than of resources. The jobless are far more prone to self-destructive behavior than are the working poor. Proposed solutions that focus solely on providing material benefits are a false path. Well-meaning social policies — from longer unemployment insurance to more generous disability diagnoses to higher minimum wages — have only worsened the problem; the futility of joblessness won’t be solved with a welfare check. The loss of work for so many also reflects the emergence of a modern labor market with little interest in less-skilled job seekers. American wages were high in the 1960s and 1970s because of steady demand for unionized labor in Detroit and Allentown. Automation and globalization have destroyed many of those jobs, and the process is likely to continue. Technology gurus like Elon Musk believe that future innovations will make the human contribution to other economic sectors, including services, increasingly obsolete as well.
Yet every underemployed American represents a failure of entrepreneurial imagination. We can do better. Our educational system must improve the way it provides skills that bring higher earnings, and we need to experiment with new forms of vocational training. We should encourage entrepreneurial energies, including by making it easier for small businesses to get up and running in low-income areas. And social programs that deter employment should be reformed — and ideally replaced by a simple pro-work subsidy. It’s time to end the war on work.
America began as an oasis of plenty in a world of poverty. Farms from New Hampshire to Georgia offered any free man crossing the Atlantic the chance to exchange hard work for a full belly. In 1820, 78 percent of the American labor force farmed. While droughts and pestilence often threatened disaster, joblessness was no part of then-rural America. If you didn’t work, you starved, and there was always another patch of land to hoe and seed.
Unemployment arrived only when workers moved to cities. A vital strength of urban life is that it can connect people who want to work with people who have capital and ideas. But sometimes, those matches aren’t available. Almost half of America’s workers had left their farms by 1870, setting the stage for the recessions of the 1870s and 1890s. University of Florida economist J. R. Vernon estimates that the unemployment rate hit 8 percent in 1878 and may have exceeded 15 percent in the 1890s. In both downturns, financial crises had led to bank failures and massive firm bankruptcies. In 1894, the Pullman Strike disrupted the nation’s transportation network.
Yet as soon as the banking system recovered, American entrepreneurs resumed hiring cheap, usually unskilled, labor. Nominal wages actually fell over both the 1870s and the 1890s because workers had to accept low pay. With no government safety net, long-term unemployment meant deprivation — or even death.
By 1920, the U.S. had become a majority-urban nation. As urban industry replaced agriculture, the country got wealthier but also more vulnerable to economic dislocation. The Great Depression brought it with terrible force: The unemployment rate exceeded 15 percent in 1931, peaked at 24.9 percent in 1933, and remained above 14 percent as late as 1940. (These figures count those working on federal relief programs as unemployed; exclude these individuals, and the unemployment rate was down to 9.5 percent by the end of the decade.) Depression-era Americans endured long-term joblessness then, but it was fundamentally different from the kind that afflicts us today. The U.S. economy was in disastrous shape throughout the 1930s, with real GDP and industrial output staying below 1929 levels for most of the decade. Whatever the reason — and debates remain lively — American industry recovered from the Depression with painful lethargy. Persistent unemployment mirrored an enduring economic crisis.
The New Deal saw the rise of public programs that worked against employment. Wage controls under the National Recovery Act made it difficult for wages to fall enough to equilibrate the labor market. The Wagner Act strengthened the hand of unions, which kept pay up and employment down. Relief efforts for the unemployed, including federal make-work jobs, eased the pressure on the jobless to find private-sector work.
The carnage of World War II ended both the Nazi regime and the American Depression. The peace ushered in 30 years of remarkable growth and prosperity. America enjoyed technological preeminence and an enormous growth in human capital, thanks to policies like the GI Bill. Women surged into the labor force by the millions, yet demand for male work stayed robust. The empowered postwar unions shifted industrial employment to right-to-work states, as the classic work of Thomas Holmes illustrates, but they didn’t compromise the labor market as a whole.
From 1945 to 1968, only 5 percent of men between the ages of 25 and 54 — prime-age males — were out of work. But during the 1970s, something changed. The mild recession of 1969–70 produced a drop in the employment rate of this group, from 95 percent to 92.5 percent, and there was no rebound. The 1973–74 downturn dragged the employment rate below 90 percent, and, after the 1979–82 slump, it would stay there throughout most of the 1980s. The recessions at the beginning and end of the 1990s caused further deterioration in the rate. Economic recovery failed to restore the earlier employment ratio in both instances.
The greatest fall, though, occurred in the Great Recession. In 2011, more than one in five prime-age men were out of work, a figure comparable to the Great Depression. But while employment came back after the Depression, it hasn’t today. The unemployment rate may be low, but many people have quit the labor force entirely and don’t show up in that number. As of December 2016, 15.2 percent of prime-age men were jobless — a figure worse than at any point between World War II and the Great Recession, except during the depths of the early 1980s recession.
The trend in the female employment ratio is more complicated because of the postwar rise in the number of women in the formal labor market. In 1955, 37 percent of prime-age women worked. By 2000, that number had increased to 75 percent — a historical high. Since then, the number has come down: It stood at 71.7 percent at the end of 2016. Interpreting these figures is tricky, since more women than men voluntarily leave the labor force, often finding meaningful work in the home. The American Time Survey found that non-employed women spend more than six hours a day doing housework and caring for others. Non-employed men spend less than three hours doing such tasks.
Joblessness is disproportionately a condition of the poorly educated. While 72 percent of college graduates over age 25 have jobs, only 41 percent of high-school dropouts are working. The employment-rate gap between the most and least educated groups has widened from about 6 percent in 1977 to almost 15 percent today. The regional variation is also enormous. Kentucky’s 23 percent male jobless rate leads the nation; in Iowa, the rate is under 10 percent.
Why, since 1970, has each new downturn added to the ranks of the permanently unemployed? Social science has not fully answered this question, but the best guess involves a combination of a generous social safety net, deindustrialization, and social change.
Both Franklin Roosevelt and Lyndon Johnson aggressively advanced a stronger safety net for American workers, and other administrations largely supported these efforts. The New Deal gave us Social Security and unemployment insurance, which were expanded in the 1950s. National disability insurance debuted in 1956 and was made far more accessible to people with hard-to-diagnose conditions, like back pain, in 1984. The War on Poverty delivered Medicaid and food stamps. Richard Nixon gave us housing vouchers. During the Great Recession, the federal government temporarily doubled the maximum eligibility time for receiving unemployment insurance.
These various programs make joblessness more bearable, at least materially; they also reduce the incentives to find work. Consider disability insurance. Industrial work is hard, and plenty of workers experience back pain. Before 1984, however, that pain didn’t mean a disability check for American workers. After 1984, though, millions went on the disability rolls. And since disability payments vanish if the disabled person starts earning more than $1,170 per month, the disabled tend to stay disabled.
The economists David Autor and Mark Duggan found that the share of adults aged 25–64 receiving disability insurance increased from 2.2 percent in 1985 to 4.1 percent 20 years later. Disability insurance alone doesn’t entirely explain the rise of long-term joblessness — only one-third or so of jobless males get such benefits. But it has surely played a role.
Other social-welfare programs operate in a similar way. Unemployment insurance stops completely when someone gets a job, which may explain why economist Bruce Meyer found that the unemployed tend to find jobs just as their insurance payments run out. Food-stamp and housing-voucher payments drop 30 percent when a recipient’s income rises past a set threshold by just $1. Elementary economics tells us that paying people to be or stay jobless will increase joblessness.
Scholars Olivier Blanchard and Justin Wolfers have explained Europe’s persistent unemployment, which they called “hysteresis,” by the interaction of adverse economic shocks and extremely generous welfare states. Twenty years ago, the more economically successful European nations, such as Sweden, Germany, and the Netherlands, reorganized their welfare states to emphasize work and witnessed positive results. Others, including France, Italy, and Spain, did not, and they have struggled. In a sense, the euro zone financial crisis of the past half-decade is the legacy of southern European countries that wouldn’t fix their failing welfare systems. The U.S. needs to decide if it wants to follow the path of Germany or of Spain.
Yet these programs didn’t immediately generate a crisis of joblessness in America. Manufacturing workers weren’t going to leave their well-paying union jobs in 1967 because of the existence of food stamps. But over the next half-century, things changed dramatically. As hundreds of studies have documented, wages for the best-educated and most-successful Americans have risen, while those for the least educated and least successful Americans have stagnated.
These developments result from tectonic movements in the economy. Globalization and technological change have steadily eroded — and continue to erode — the demand for American brawn. In 1966, American factories employed millions of industrial workers, making products that were shipped to far poorer places. As technology spread, the world’s lower-wage countries started manufacturing. Asia’s economic tigers initially thrived because of low labor costs, but these increasingly educated countries eventually achieved technological parity with — and sometimes became superior to — many American industries.
Manufacturing’s share of total American output has fallen from 25 percent in 1968 to 12 percent today. The number of manufacturing workers has shrunk from 19.5 million in 1979 to 12.2 million, which represents 8.8 percent of non-farm employment. The fact that manufacturing today is a larger share of GDP than of employment underscores a shift toward technology-intensive production — another response to high U.S. labor costs. For millennia, men were valued for their muscles. Human strength was crucial to feudal farming and to Henry Ford’s assembly line. We still have some jobs that depend on strong backs, as in the building trades. But they are getting rarer because machines can do the work for us.
We’re not moving toward an entirely mechanized economy. Between 1980 and 2000, U.S. service-sector employment rose by 73 percent — a whopping 37 million new jobs. There remains commercial value in a friendly face and the charm of human interaction. But for millions of men, working in the service sector wasn’t a good option.
American joblessness reflects the social unraveling that Charles Murray describes in Coming Apart. A significant portion of the American heartland has moved from a norm of stable marriage and traditional religion to single-parent families and social dysfunction. A study by Raj Chetty and Nathan Hendren calculated mobility across America using income-tax records. Their data show that the share of single-parent families in an area is a particularly strong predictor of low upward mobility. Any parent knows that raising children is tough, even with two adults involved. When only one parent is around, that task gets even harder. Unsurprisingly, many kids from broken families lack the skills needed to get ahead in today’s competitive economy.
During World War II, the Army taught millions of Americans how to behave effectively in a tough organization. Such skills may have helped returning veterans thrive in the industrial America of the 1950s. Yet that very success may also have enabled younger Americans to tolerate joblessness, as they wind up relying for extended periods on their parents’ (or grandparents’) help. Thirty percent of prime-age jobless men currently live with their parents.
The rise of joblessness among the young has been a particularly pernicious effect of the Great Recession. Job loss was extensive among 25–34-year-old men and 35–44-year-old men between 2007 and 2009. The 25–34-year-olds have substantially gone back to work, but the number of employed 35–44-year-olds, which dropped by 2 million at the start of the Great Recession, hasn’t recovered. The dislocated workers in this group seem to have left the labor force permanently.
Unfortunately, policymakers seem intent on making the joblessness crisis worse. The past decade or so has seen a resurgent progressive focus on inequality — and little concern among progressives about the downsides of discouraging work. Advocates of a $15 minimum hourly wage, for example, don’t seem to mind, or believe, that such policies deter firms from hiring less skilled workers. The University of California–San Diego’s Jeffrey Clemens examined states where higher federal minimum wages raised the effective state-level minimum wage during the last decade. He found that the higher minimum “reduced employment among individuals ages 16 to 30 with less than a high school education by 5.6 percentage points,” which accounted for “43 percent of the sustained, 13 percentage point decline in this skill group’s employment rate.”
The decision to prioritize equality over employment is particularly puzzling, given that social scientists have repeatedly found that unemployment is the greater evil. Economists Andrew Clark and Andrew Oswald have documented the huge drop in happiness associated with unemployment — about ten times larger than that associated with a reduction in earnings from the $50,000–$75,000 range to the $35,000–$50,000 bracket. One recent study estimated that unemployment leads to 45,000 suicides worldwide annually. Jobless husbands have a 50 percent higher divorce rate than employed husbands. The impact of lower income on suicide and divorce is much smaller. The negative effects of unemployment are magnified because it so often becomes a semi-permanent state.
Time-use studies help us understand why the unemployed are so miserable. Jobless men don’t do a lot more socializing; they don’t spend much more time with their kids. They do spend an extra 100 minutes daily watching television, and they sleep more. The jobless also are more likely to use illegal drugs. While fewer than 10 percent of full-time workers have used an illegal substance in any given week, 18 percent of the unemployed have done drugs in the last seven days, according to a 2013 study by Alejandro Badel and Brian Greaney.
Joblessness and disability are also particularly associated with America’s deadly opioid epidemic. David Cutler and I examined the rise in opioid deaths between 1992 and 2012. The strongest correlate of those deaths is the share of the population on disability. That connection suggests a combination of the direct influence of being disabled, which generates a demand for painkillers; the availability of the drugs through the health-care system; and the psychological misery of having no economic future
Increasing the benefits received by non-employed persons may make their lives easier in a material sense but won’t help reattach them to the labor force. It won’t give them the sense of pride that comes from economic independence. It won’t give them the reassuring social interactions that come from workplace relationships. When societies sacrifice employment for a notion of income equality, they make the wrong choice.
Politicians, when they do focus on long-term unemployment, too often advance poorly targeted solutions, such as faster growth, more infrastructure investment, and less trade. More robust GDP growth is always a worthy aim, but it seems unlikely to get the chronically jobless back to work. The booms of the 1990s and early 2000s never came close to restoring the high employment rates last seen in the 1970s. Between 1976 and 2015, Nevada’s GDP grew the most and Michigan’s GDP grew the least among American states. Yet the two states had almost identical rises in the share of jobless prime-age men.
Infrastructure spending similarly seems poorly targeted to ease the problem. Contemporary infrastructure projects rely on skilled workers, typically with wages exceeding $25 per hour; most of today’s jobless lack such skills. Further, the current employment in highway, street, and bridge construction in the U.S. is only 316,000. Even if this number rose by 50 percent, it would still mean only a small reduction in the millions of jobless Americans. And the nation needs infrastructure most in areas with the highest population density; joblessness is most common outside metropolitan America.
Finally, while it’s possible that the rise of American joblessness would have been slower if the U.S. had weaker trade ties to lower-wage countries like Mexico and China, American manufacturers have already adapted to a globalized world by mechanizing and outsourcing. We have little reason to be confident that restrictions on trade would bring the old jobs back. Trade wars would have an economic price, too. American exporters would cut back hiring. The cost of imported manufactured goods would rise, and U.S. consumers would pay more, in exchange for — at best — uncertain employment gains
The techno-futurist narrative holds that machines will displace most workers, eventually. Social peace will be maintained only if the armies of the jobless are kept quiet with generous universal-income payments. This vision recalls John Maynard Keynes’s 1930 essay “Economic Possibilities for Our Grandchildren,” which predicts a future world of leisure, in which his grandchildren would be able to satisfy their basic needs with a few hours of labor and then spend the rest of their waking hours edifying themselves with culture and fun.
But for many of us, technological progress has led to longer work hours, not playtime. Entrepreneurs conjured more products that generated more earnings. Almost no Americans today would be happy with the lifestyle of their ancestors in 1930. For many, work also became not only more remunerative but more interesting. No Pennsylvania miner was likely to show up for extra hours (without extra pay) voluntarily. Google employees do it all the time.
Joblessness is not foreordained, because entrepreneurs can always dream up new ways of making labor productive. Ten years ago, millions of Americans wanted inexpensive car service. Uber showed how underemployed workers could earn something providing that service. Prosperous, time-short Americans are desperate for a host of other services — they want not only drivers but also cooks for their dinners and nurses for their elderly parents and much more. There is no shortage of demand for the right kinds of labor, and entrepreneurial insight could multiply the number of new tasks that could be performed by the currently out-of-work. Yet over the last 30 years, entrepreneurial talent has focused far more on delivering new tools for the skilled than on employment for the unlucky. Whereas Henry Ford employed hundreds of thousands of Americans without college degrees, Mark Zuckerberg primarily hires highly educated programmers.
What could change this dynamic? The first step is to improve Americans’ skills. The jobless rate is about 8 percent for prime-age men with a college degree or more but more than 22 percent for men with only a high-school diploma or less. We have levers that can improve educational outcomes, like the very best early-childhood programs and charter schools. Such innovations should be expanded and made better through competition and evaluation.
We should also improve the way that we do vocational education. Many vocational schools, like Boston’s Madison Park High School, have long been troubled. The most ambitious students avoid getting tracked onto a vocational path, and they — and their parents — want schools that focus on college readiness. Consequently, less fortunate or struggling students often get segregated into these vocational centers. The conventional teachers in many vocational programs often lack the know-how for teaching either high-paying blue-collar trades, like plumbing, or cutting-edge fields, like computer programming.
A more effective approach might be to keep students in college-readiness-oriented schools and experiment with out-of-school vocational training. Kids could be taught after school, on weekends, and during the summer by programs specializing in particular occupations. These initiatives can be evaluated swiftly — you can readily determine if a program has produced, say, good carpenters. The superior training programs can then be scaled up and bad ones shut down. Adopting this structure would mean that anyone could potentially compete to run the programs — trade unions, private providers, nonprofits — increasing the chances that some programs will excel. We should also be open to initiatives like Cambridge, Mass.’s “The Possible Project,” which has been training youths, many from poorer backgrounds, to launch themselves in the start-up economy. (I am currently working on a randomized control trial for the project.)
Older workers present the toughest training problem. The extensive literature on retraining adults for new jobs has few success storis. We must keep trying; here, too, the more experimentation, the better.
Along with up-skilling workers, we should lower the regulatory barriers to entrepreneurship. It’s a sad fact that America tends to regulate the entrepreneurship of the poor much more stringently than it does that of the rich. You can begin an Internet company in Silicon Valley with little regulatory oversight; you need more than ten permits to open a grocery store in the Bronx.
One-stop permitting would be a good step, especially in poorer areas. If new businesses had only a single regulatory office to satisfy, the obstacles to entrepreneurship would be less daunting. One-stop permitting would also make it easier to evaluate the regulator on its speed and the number of permits issued. Permitting shops could specialize in the languages and businesses most common in their areas
Occupational licensing is another area crying out for reform. The University of Minnesota’s Morris M. Kleiner has found that the share of American workers who need an occupational license has increased from 5 percent in the 1950s to 29 percent in 2008. States now credential interior designers, tree trimmers, and even florists. In many cases, these requirements are merely means for protecting incumbents from competition. When we license basic service jobs, we make it tougher for the jobless to find something new to do.
American entrepreneurs can solve our joblessness crisis only if the U.S. stops incentivizing joblessness. Consolidating social policies would be a crucial step. Struggling families now receive food stamps, housing vouchers, Temporary Aid to Needy Families, and other assistance — all of which punish work. If the various programs were combined into a single cash benefit, that benefit could be designed so that the tax on earnings never went above 30 percent. We could follow the lead of Norway on unemployment and disability insurance, allowing the disabled to keep, say, 50 percent of their benefit above the $1,170 threshold, while tightening the requirements for being designated as disabled. Unemployment insurance could be structured so that payments were no longer contingent upon staying completely out of work.
Here, the Earned Income Tax Credit (EITC) offers a design model, by providing funds that initially scale up with earnings, especially for lower-income families with children. Economists Nada Eissa and Jeffrey Liebman found that the credit’s introduction in 1986 increased labor-force participation significantly. The EITC was instituted during one of those rare moments in modern U.S. history when policymakers wanted to avoid rewarding joblessness.
We also need to make hiring workers less costly for employers. Temporarily cutting the payroll tax was one of the most constructive policies adopted during the Great Recession. We could enact a permanent payroll-tax reduction. The tax could be gradually phased in for workers once their hourly earnings went beyond a certain threshold. The payroll tax could be eliminated for workers who had been unemployed, at least for an initial period. The costs of reducing the payroll tax could be offset by raising the minimum retirement age for employees who hadn’t paid these taxes for enough years. Reducing mandated benefits, like health care, that employers must provide lower-income earners would help encourage work, too. Ideally, the reform of our health-care system will ensure that workers have health-care options that don’t unduly burden employers.
Making work pay needs one final, major policy initiative: wage support, which would replace the EITC. The EITC had the right overall idea, but it is cumbersome and indirect. Instead, the federal government could simply provide pay to increase the earnings of minimum-wage workers by a fixed amount — say, $3 per hour. Consequently, a worker paid $7.25 would take home $10.25 hourly, with the difference paid for by taxpayers. The subsidy could fall gradually as wages rise, and it could be targeted for specific groups — larger for returning veterans or the long-run jobless — and rise or fall with the level of aggregate unemployment. The phaseout might slightly slow private-sector wage growth, but the cost would be more than offset by the benefits of such a visible push toward employment. Such a program would be expensive, so it should be matched with spending reductions for other social services.
The rise in joblessness is not inexorable. But to solve this crisis, we must educate, reform social services, empower entrepreneurs, and even subsidize employment. That is an ambitious — but necessary — agenda for ending the war on work before it consumes another generation of Americans.
This piece originally appeared on National Review Online
Edward L. Glaeser is the Glimp professor of economics at Harvard University, a senior fellow at the Manhattan Institute, and contributing editor at City Journal. This piece was adapted from the City Journal special issue, "The Shape of Work to Come"
This piece originally appeared in National Review Online