The growing disaffection of America’s working class with their traditional political representatives, particularly in the Democratic Party, is frequently (and correctly) attributed to their social conservatism on a range of issues such as crime, immigration, and DEI. But it also represents their disgruntlement with the economic state of working-class America.
Non-college graduates fell far behind their college-educated peers in earnings and income after 1980, and they remain far behind today. Their real wages have risen modestly in the past decade, and some of the gap between their earnings and those of college grads has narrowed, but not enough. Many regions of the country where the working class had traditionally thrived, such as industrial regions in the Midwest, were badly hurt during this period. Though these regions have somewhat recovered, with college grads benefiting from a booming service sector in many areas, the fortunes of older and younger working-class residents have mostly not bounced back.
Why did this happen? Economic research shows that the working class—usually defined as people without a four-year college degree—has been hurt first and foremost by automation, which shifted employer demand towards college grads in the digital age. Rising international trade, which continues to benefit consumers and the U.S. economy overall, also contributed to this gap, especially after China entered the World Trade Organization. The weakening of unions and flattening of the federal minimum wage were factors as well.
By all indications, young working-class men are especially discouraged about their future economic prospects. They are often told to get college degrees, which many find unappealing. We should perhaps not be surprised, therefore, that so many swung so hard towards Donald Trump and his anti-trade and anti-immigrant policies, though some of these could end up bringing more economic damage than benefit to the working class. Many also now fear that their jobs will be displaced by artificial intelligence (AI) over the coming years and decades.
In response to these developments, America needs a powerful labor market agenda for the working class that will benefit them economically over the long run, resting on the themes of “better skills and better jobs.” This agenda should be fiscally and politically pragmatic and consistent with the desires and worldview of workers and their families. It should also address the advent of AI, especially the fear of worker displacement, and help workers benefit from its diffusion into our economy and labor market.
Better Skills
Education and skills remain an important factor behind labor market inequality in the U.S. Though most high school graduates now enroll in college, at least for a while, completion rates are low. Only about half of all Americans manage to obtain a postsecondary credential of any kind, and for those who don’t, earnings are usually quite low.
Furthermore, earnings differ substantially across workers with the same credentials but in different fields. For instance, people who have less than a bachelor’s degree earn much more in high-demand fields like health care, advanced manufacturing, and information technology (IT) than their similarly educated peers in the liberal arts. Certificates in these fields can substantially raise workers’ earnings, especially if they later acquire jobs in the fields for which they have trained.
And in high-demand fields, employers often struggle to fill vacant jobs with qualified applicants. This was especially true during the recovery from the pandemic, in what was often called the Great Resignation, and it remains true in many technical fields as baby boomers retire and immigrants become more reluctant to come to America. This situation means that both employers and workers would benefit from more effective workforce development policies that target these well-paying but hard-to-fill jobs.
To provide better earnings to a wider swath of Americans without college degrees, the country needs a much more robust package of workforce development programs to train workers for high-paying, available jobs and provide needed supports (like career guidance or child care for working parents). These programs come primarily in three forms: (1) workforce programs at community colleges; (2) other private or nonprofit training providers; and (3) work-based learning models like apprenticeships.
Community colleges are the largest providers of workforce programs in the U.S., offering not only degrees but also a wide range of certificates. But many students do not complete their programs for a range of reasons, such as poor academic preparation or a need to work full-time to support their families. Public funding is also too limited, and colleges cannot afford to provide the range of supports and services students need, like better guidance to match them with programs they can successfully complete. Within institutions, workforce programs often get less funding than the liberal arts, and there is too little partnership with employers to align curricula with their skill demands. The growing not-for-credit programs are more responsive to industry, but students in these programs are not eligible for Pell grants and other federal loans. And the for-profit colleges soak up too much federal financial aid while generating credentials with weak market value.
Outside of community colleges, there are a wide range of training programs and providers, and the best ones provide sectoral training. These programs work closely with regional employers in high-demand industries like health care, IT, construction, green energy, and manufacturing to train workers in the skills that employers need the most while also providing strong support for students. Examples of the best programs include Per Scholas, Year Up, the Wisconsin Regional Training Partnership, and Project Quest. But these programs are expensive and hard to scale, requiring funding that goes well beyond the very limited federal support provided through the Workforce Innovation and Opportunity Act. They also screen out many workers with personal barriers or weak skills who would likely not successfully complete their programs.
Work-based learning programs like apprenticeships train workers in the exact skills that employers need while allowing them to “earn as they learn.” Studies indicate that workers gain higher earnings and employers gain higher productivity. But employer take-up is low—many fear the costs and regulations imposed by registered apprenticeships, though some overstate these problems. Some states are finding new ways to encourage more employers to take advantage of these opportunities, like offering tax credits and embedding programs in high schools and community colleges, as we try to learn what works best in this regard.
All three categories of workforce programs need more funding and political support. But they also need accountability and appropriate incentives to ensure that public dollars are well spent and generate better outcomes for both workers and the employers who hire them. New private-public models of funding education, such as income-share agreements and outcome-based loans, can strengthen both the quantity of funding and quality incentives by specifying loan repayment only after worker earnings have risen and meet appropriate minimum standards (often $40,000–$50,000 in annual earnings).
Better Jobs
Too many jobs in America still require few skills and pay low wages. Traditionally, minimum wage statutes and collective bargaining have been our tools for raising pay and skill requirements. But the federal minimum wage has been stuck at $7.25 since 2009, with no evidence that it will go up soon. Minimum wages in blue states are often $15 or higher, perhaps risking job loss in the most extreme cases, while those in red states are mostly stuck at the federal level. Moreover, union membership has fallen to just six percent in the private sector, showing no sign of reversing a downward trend that began 70 years ago.
But there are other approaches to improving job quality. Employers often choose between “high road” compensation, wherein worker performance is high and turnover is low, and the “low road,” wherein they minimize pay and benefits but suffer from weak performance and high turnover. Too many employers opt for the low road, sometimes perhaps not even realizing there are other high-performance options. The higher pay for workers in high-performing establishments is basically a “public good,” which private employers on their own do not often value and will not pay for.
To encourage more high-road employment, government could subsidize and provide technical assistance to well-paying employers, as the Biden administration did through its Good Jobs Challenge at the Commerce Department and its Good Jobs Initiative at the Labor Department. Regional economic development agencies can also play this role, working with employers to create good jobs and ensure the growth of effective skill-building efforts to fill these jobs with productive workers.
We see an example of this approach as the CHIPS Act is implemented around the country. Economic development agencies are partnering with employers and training providers to meet the high technical skill needs for workers in these well-paying jobs.
And What About AI?
AI capabilities are rapidly improving. Programs like ChatGPT or Gemini can already outperform most college graduates in analytical abilities and basic writing or computation as well as generate pictures or music or poetry. AI will almost certainly raise U.S. productivity and, thus, the incomes of many workers. But many also fear that they will be displaced as AI takes over more and more of the tasks they perform on their jobs, and some analysts believe AI is already making it harder for new college grads to gain entry-level employment.
Of course, the Luddites predicted widespread job loss due to automation about 200 years ago, and such fears have flared up in other times, though we have never experienced widespread automation-based joblessness. New job categories replace older ones, and employment grows in industries where automation improves product quality at lower prices, even if the jobs are different than before. On the other hand, some workers do get hurt as automation proceeds, either by job displacement or lower wages.
The trick is for workers to gain skills that enable them to work with the technology rather than be replaced by it. AI, in fact, might help non-college workers perform a wider range of tasks than they do today—and perhaps at higher wages. But AI’s tendency to constantly improve might make it harder for workers to adjust through reskilling, as it overtakes the new skills they learn today just a few years later.
We can minimize these risks through a variety of skill-building efforts. Our schools will need to teach students AI literacy and also AI-complementary skills, such as critical thinking, good judgment, social interactions, and modes of communication that AI will not match for a long time. Employers should be encouraged to retrain workers instead of just laying them off when AI takes over their current job tasks. Perhaps we should subsidize this retraining while taxing worker displacements at modest levels, just as the Unemployment Insurance system more heavily taxes employers who generate higher layoff rates.
Training providers will need to continually engage with employers to make sure their curricula and programs are up to date and not facing imminent obsolescence. Workers will likely need “lifelong learning accounts” to provide for frequent upskilling. Unemployment Insurance programs should be better financed and provide stronger re-employment services to meet the coming upticks in job displacement. A program of “Automation Adjustment Assistance,” modeled after Trade Adjustment Assistance for workers displaced by imports, could be considered as well.
Finally, the federal and state governments should use grants to encourage AI developers to build “human-centered” AI, designed to augment worker skills and tasks rather than just replace them.
Going Forward
A robust agenda for workers should enable millions more to obtain better skills and better jobs without pursuing college degrees. This will require more and newer kinds of funding for such training but also accountability and incentives to make sure that any new funding generates the skills for high-demand and high-paying jobs. Encouraging more employers to provide work-based learning and on-the-job retraining is critical. Training providers will need to better partner with employers and regional economic development agencies to improve job quality. Workers will also need help adapting to the instability that AI might generate with support for lifelong learning and more retraining on the job.
Such a program would likely appeal to young workers without college degrees, especially the young men who are now so disenchanted with their future prospects. We should immediately begin to build that agenda.
Harry J. Holzer is the John LaFarge Professor of Public Policy at Georgetown University and a nonresident fellow in Economic Studies at Brookings. (The opinions expressed here are his own and not those of Georgetown or Brookings.)
All of this could’ve been said in two small paragraphs. There is nothing new presented in this lengthy column today. The problem is, we are already so far behind in achieving any of the objectives that it will be almost impossible to do so. This country is going to be irrelevant by the end of the century if not sooner.
Oh god, we're back to "learn to code". While CS majors can't find a job no less. "Get training" like we are dogs or something. Import a half million economists and college professors is what we need. If economics professors had wages cut by three quarters I'm sure we'd have entire new schools of thought.
LPNs with less than a year education make good money because it's hard to pass the tests, they cull stupid people out, and because the work is very hard. High stress environments that require working as a team pay well, just like being a derrick hand on the floor of a drill rig.
Low wages are a function of supply and demand. We imported ten or twenty million workers, what the heck do you think that does to wages. The folks they imported don't speak English and are unfamiliar with local trades, but because they are bright enough, they can adapt, be productive, and make their employer money, they most certainly didn't go to any trade school.
Most any company can train a worker to be productive at most jobs in a couple of weeks and no one wants any workers over 45 years old. Reality. "retraining" is a make work project for people who can claim to be able to teach. Useless.
The only thing that will bring wages up is when comfortable elites get so scared of pitchforks they figure they better do something, when Trump and Bernie Sanders populism is looked upon as the good old moderate days.