High Demand, Low Wage

A new federal “do no harm” wage rule draws broad support but doesn’t solve the problem of preparing workers for societally valuable jobs with poverty-level pay. Also, research finds that administrative and clerical workers in smaller cities will be least able to adapt to AI, and an essay on why we need to stop calling some skills “soft.” (Subscribe here.)

Low Bar on Wages

Half of U.S. certificate programs in allied health and medical assisting are projected to fail new wage thresholds from the Education Department, which will soon apply to all of higher education. Federal negotiators decided tough medicine was necessary for training programs in these occupations, which often do not pay a living wage.

The core accountability standard is a rare policy codified by the U.S. Congress rather than the executive branch. Under the “do no harm” test that’s slated to go into effect in July, undergraduate programs must demonstrate that a majority of graduates earn more after four years than the wages of a typical high school graduate in that state.

Just 2% of current degree programs would run afoul of the rule, according to program-level performance data the department published. But 29% of certificate programs are at risk of failing. (Certificates were not covered under the standard Congress passed last year, but those credentials were added during the department’s negotiating process.)

Programs will lose access to federal loans if they fail to meet the wage threshold for two out of three years. They won’t lose Pell Grant eligibility, however. That will only happen if at least half of the college’s federal aid–eligible students are enrolled in failing programs, or half of the institution’s federal aid revenue comes from the programs.

“This is a pretty low bar,” says Preston Cooper, a senior research fellow at the American Enterprise Institute, who was part of the federal committee that reached consensus this month on draft language for the rule, which the department will finalize after a public comment period. “This is the absolute bare minimum.”

Compromise on Pell: Given the low tuition rates at community colleges, relatively few students rely on federal loans. And likely no community college is at risk of losing Pell Grant access due to the rule.

The American Association of Community Colleges is concerned about the potential loss of loan eligibility for certificate programs, including in allied health, says David Baime, the association’s VP of government relations. The primary reason two-year colleges offer those certs is because of their strong value to society, he says, noting that most students enrolled in certificate programs across the sector don’t have to tap loans.

“We surely hope, and expect, that community college allied health programs will not be unduly impacted by the final rule,” Baime says.

Indeed, the federal data set shows that 90% of undergrad certificate programs across public two-year colleges would clear the bar. It’s a different story at for-profit colleges, where just 55% of certificate programs would pass the test. However, Career Education Colleges and Universities, a trade group representing the sector, praised the rule-making process and the proposed framework. CECU called the wage test a strong starting point and applauded its application to all of higher education.

Certificates across the cosmetology training sector will take the biggest hit—93% of those programs are likely to fail. Few experts are defending beauty and cosmetology schools. That’s because a typical graduate of a federal aid–eligible cosmetology program makes roughly $20K four years after earning a credential, while holding $10K to $14K in student loan debt, New America noted last year in an investigative report.

Allied health certificates are trickier. Those jobs are in high demand and of high value to society. Yet the emerging bipartisan consensus appears to be that the federal government shouldn’t back training that leads to poverty wages.

“This is essentially a subsidy for workforce training for industries that are not paying enough to bring their graduates above the level of a high school graduate,” says Cooper, who published an analysis with charts on the federal performance data.

Few experts challenge the high-school-wage test. The rule’s appeal has a lot to do with its simplicity. Who would argue that the feds should subsidize programs with a median payoff that lags the earnings of people who never went to college? 

Michael Itzkowitz played a big role in the push for a “do no harm” test. During the Obama administration, he led the College Scorecard, a major step on the way to the new accountability framework.

Itzkowitz praised the new regulations overall, saying they will better protect students from graduating with large amounts of debt from programs that show little or no economic value. Yet he wishes the rule went further—echoing some consumer advocates on the left.

“The only winners from maintaining Pell Grant funding for underperforming programs are the institutions themselves,” says Itzkowitz, the founder and president of the HEA Group. “Ultimately, taxpayer dollars will still be spent and students’ time will still be wasted on certificates with little value in today’s economy.”

Market Failures and Rural Impacts: Healthcare certificate programs that face the most risk are offered in rural areas, several experts say. The main reason is that the cost of living tends to be lower in those regions, while a statewide median wage gets pulled up by expensive cities.

In Texas, the median for high school graduates is $30K, according to a recent report from researchers at the University of Wisconsin at Madison. Yet that figure reaches roughly $47K around Odessa due to the area’s strong oil and gas industry, compared to $25K for Lufkin in rural East Texas and for greater McAllen, which is in the economically depressed Rio Grande Valley. As a result, the report calls for focusing instead on local income data for measuring college ROI.

(A new study from a researcher at Rice University found that healthcare programs at for-profit career colleges in Texas have substantially higher graduation rates, shorter time to completion, and stronger economic mobility than community colleges, across both certificates and associate degrees.)

The College of Health Care Professions trains about 36% of the medical assistants in Texas. In comments to the Education Department about the proposed rule, the for-profit college said graduates of the medical assistant certificate program at its Austin campus earned an average of 64% more than their peers who completed at CHCP’s campus in McAllen, where jobs pay less and the cost of living is much lower.

In recent years, some colleges have moved away from offering credentials in low-paying “helping professions,” including programs for early childhood educators as well as some healthcare roles.

Likewise, debates about public funding for certificates and other short-term training programs cannot be separated from broader conversations about the job market, wrote Michelle Van Noy, a professor at Rutgers University and an expert on education and employment, in an essay for Work Shift.

The real problem isn’t the quality of certificate programs, she says, but labor market failures that create undervalued, low-wage jobs. Meanwhile, public funding for education and training falls well short of what’s needed in this country, writes Van Noy, particularly for workers in undervalued jobs.

Van Noy cited a landmark 2021 report from the Treasury Department on the economics of childcare supply. The TLDR is that tight margins in the childcare industry mean that there simply isn’t money to pay workers living wages. It’s unclear if healthcare systems can pay medical assistants more. But the industry is facing serious financial strain, which is likely to get worse.

Maybe the federal education system shouldn’t subsidize the preparation of students for roles that don’t pay well. But if someone has a fix for training and supporting desperately needed workers on the frontlines of healthcare, I haven’t seen it.

Who’s Able to Adjust to AI?

AI may change and even threaten the jobs of college-educated workers like no technology before it. But those workers are also generally well positioned to adapt.

That’s not equally true for millions of Americans working in administrative and clerical roles. More than 6M of them are short on the savings, broad skill sets, relative youth, or local opportunities that will likely be needed to adapt.

They’re scattered in smaller cities and towns across the country—places like Española, N.M.; Athens, Ohio; and Tallahassee, Fla.—and 86% are women. They’re mothers and grandmothers rooted in place, not mobile software engineers. 

That’s the thrust of a new report out of the Brookings Institution, along with a National Bureau of Economic Research working paper. 

“Above all, this research suggests that not all exposed and disrupted workers face the same challenges of relocating to good work,” says Mark Muro, a senior fellow at Brookings Metro and a co-author on the report.

The researchers are careful to emphasize that their analysis isn’t a prediction of job loss per se, but rather an important signal about who is both at risk and likely to struggle to adapt if the technology profoundly alters or automates away their work. The goal is to get policymakers and employers focused now on improving those workers’ odds of landing on their feet.

“They can consider what policies can help make workers be more resilient to potential labor market changes before displacement actually happens,” says Sam Manning, the lead researcher on the paper and a senior research fellow at GovAI, a nonprofit focused on the policy implications of AI.

Manning says policymakers could start providing tax incentives or grants for large local employers to invest in on-the-job AI training, incorporating AI literacy across regional workforce development initiatives, or adapting unemployment assistance programs to today’s reality. The federal government also should be investing in better public data on AI adoption and its real-time effects on the labor market, he says, so that we have a clearer picture of whether people are already being impacted and how.  

“Over the next five or 10 years, the policy challenge might look more like there are still lots of jobs, but the skill demands are likely to change, potentially quite rapidly,” Manning says.

The new research found that smaller state capitals and college towns—places where education and government administration are major drivers of the economy—need to be paying particular attention. Those metros tend to be concentrated in the country’s center, not on the coasts.

The Kicker: “Given the exposure profile, the incidence of vulnerability is higher on a share basis in the West and the Midwest,” Muro says. —By Elyse Ashburn

While We’re Talking About AI

A few other recent reports of note on AI adoption and workforce impacts:

+ Anthropic is out with its latest global economic index, which finds that augmentation remains more common than automation on Claude.ai. "The most complex tasks where Claude is used tend also to be those where it struggles most. Rather than displacing highly skilled professionals, this could instead reinforce the value of their complementary expertise in understanding AI's work and assessing its quality," the company writes.

+ The market research firm Forrester is projecting that AI and automation will result in the loss of about 6.1% of U.S. jobs, or about 10.4M, by 2030. Don’t confuse financially driven layoffs with AI layoffs, the report argues: “AI is just the scapegoat, at least today.”

+ Roughly 37% of the time saved by AI is being offset by “rework”—employees spending significant time correcting, clarifying, or rewriting low-quality AI-generated content, according to a new research from Workday. “Measuring success solely by hours saved obscures the real impact of AI on work quality and outcomes. Instead, leaders should evaluate productivity in terms of value created—accounting for both time saved and time lost to rework.”

Open Tabs

Short-Term Credentials
A recent study may be the most comprehensive analysis on the labor market value of nondegree credentials. The working paper from the Brookings Institution used machine learning to identify 54M credentials, mapping them to a standardized taxonomy. Returns vary widely and depend on job relevance, credential type, and worker characteristics. Workers need better information, the researchers conclude, calling for credential registries and ROI tracking systems.

Labor Shortage
The U.S. has a 1.7M annual shortfall between job openings in the skilled trades and the number of qualified graduates of education and training systems, according to an analysis from Lightcast and Area Development magazine. There are roughly 2.4 machinist openings for every graduate and four for each repair worker. The ratio for semiconductor process technicians, furnace operators, and nuclear technicians tops 100 openings per graduate.

Earn and Learn
Degree apprenticeships are growing fast but remain rare, finds a report from New America. The emerging model is concentrated in a small number of occupations, with a third of all degree apprenticeship opportunities being in educational instruction (teaching) and library roles, followed by healthcare (nursing), construction, and manufacturing. Public funding is key to the growth of this model, the report concludes, noting the example of a cash incentive in Alabama.

AI and Frontline Jobs
Low-wage workers in the U.S. are divided in their feelings about AI, but a majority (56%) believe they may have to change careers because of the technology, and roughly half believe they could lose their jobs in the future, according to a survey from HarrisX and Merit America. Seventy percent of respondents said they would stop the development of AI that might eliminate jobs, even among those who believe their role is unlikely to be affected.

Thanks for reading. If this newsletter was forwarded to you, subscribe here. —PF