Labor force participation falls to 61.5%, the lowest in 50 years outside COVID, and economists say it’s not just people giving up

Although AI and tech innovation is blamed for that drop, it has more to do with the lack of supply relative to the number of jobs that are available.

Economists have spent the past week arguing about why 720,000 people walked away from the labor force in a single month. Laura Ullrich, director of economics at Indeed Hiring Lab and a former Richmond Fed economist, says that rather than treating June’s slide to a 61.5% labor force participation rate—the lowest reading outside the pandemic since 1976—as a story about discouraged workers giving up, it’s actually about supply: There simply aren’t enough workers left to fill the jobs employers have.

“Historically, you’ve been able to look at jobs numbers like what came out on Friday and say, ‘okay, there was a decline in leisure and hospitality. Well, that means there’s less demand for those workers,'” Ullrich told Fortune. “But I think now, and more commonly as we go forward, it actually could be labor supply driving some of that. There are two reasons why you might not add jobs in a month: One is there’s no demand for workers, the other is there is demand, but there’s not enough supply.”

A shrinking economy in the works

Ullrich co-authored a May report from Indeed Hiring Lab, “The Great Mismatch: How a Shrinking Workforce, AI, and Labor Reallocation Will Define the Next 15 Years,” projecting the labor force would begin shrinking in 2026. The finding was particularly driven by a combination of immigration policy changes and what Ullrich calls “the demographic cliff”—the accelerating retirement of the baby boomer generation. The report estimates the labor force will decline by roughly 3.7%, or 5.9 million workers, between 2025 and 2032, before partially recovering. It also projects the aggregate unemployment rate could climb by 0.5 to 3.5 percentage points by 2040, to nearly 8% in the more severe scenario.

“When we first ran the numbers and saw that we were actually predicting the labor force was going to start declining this year, I was like, ‘oh gosh, I don’t know,'” Ullrich said. Around the same time, then-Fed Chair Jerome Powell told reporters the economy was seeing “very, very low, nonexistent, really” growth in the labor force. “I was like, okay, I think we are here with the demographic changes and the share of baby boomers that are leaving the workforce.”

The Bureau of Labor Statistics’ own 10-year projections already point to declining participation, Ullrich said, and those projections predate the current immigration restrictions.

“I think when their estimates come out this next year, they’ll be even more severe declines, because immigrant workers are both younger than native-born workers, but also have higher labor force participation rates.”

Native vs. foreign-born workers

Ullrich also pointed to BLS data comparing labor force participation between foreign-born and native-born workers, showing foreign-born individuals have a participation rate of 66.3%, compared with 61.6% for native-born workers. Among men, the gap widens: 76.9% for foreign-born men, as compared to 65.8% for native-born men. But the pattern flips among women: Native-born women have higher participation rates than immigrant women, particularly immigrant women with young children, who are less likely to work than their native-born counterparts.

Age compounds the effect: Roughly 70.1% of foreign-born individuals are between 25 and 54, considered the “prime-age” years, as compared with 62.7% of native-born Americans.

“If immigration declines, you have two impacts that are related: Immigrant workers tend to be younger than native-born workers, so by definition you get an older workforce, and labor force participation rates, even within the same age groups, are higher, especially for foreign-born men,” Ullrich said.

The great AI mismatch

AI is accelerating a mismatch in which the available workers aren’t lining up with where the job openings are. That’s because the technology is projected to hit the information sector, financial activities, and professional business services hardest, which are exactly the sectors with the youngest workforces and the most graduates trying to get in. In the report’s more AI-disruptive scenario, the combined unemployment rate across those three sectors rises from 4% in 2025 to 12% by 2032: 21.2% in information, 11.8% in financial activities, and 10.7% in professional and business services.

“People go and study things to go work in finance or computer science because those have historically been really successful paths, and so there continues to be people trying to flow into those sectors when actually AI is impacting those first and most severely,” Ullrich said.

Meanwhile, sectors with rapidly aging workforces, like in government, health care, education, and construction, aren’t attracting enough new entrants, and the report finds AI does little to close that gap. Ullrich cited internal research she’s been doing on health care demographics: In New Mexico, 39.2% of physicians are over age 60. Nursing has a related but different bottleneck, the report finds: 68% of nurses entered the profession directly, and 72% who leave a nursing job stay in the field, which is evidence of how high credentialing and training barriers seal off the profession from workers in other industries, even when there’s plenty of demand.

Low-wage, high-demand fields like home health aide work face the starkest supply gap of all, Ullrich said, as demand rises with an aging population but pay hasn’t kept pace: “It’s a mismatch problem. In theory, AI should make the match easier, but it still creates this friction that we estimate will bring down employment.”

But so far in how much the tech has advanced, there’s still more to be desired.

“I do not believe there are a lot of AI agents doing work that people used to do, yet,” she said. “The investment choices, the purchasing choices of firms of increasing capex spending, that’s offsetting labor. That’s an indirect impact of AI.”

Indeed modeled the labor force through 2040 under two scenarios: one in which AI destroys a significant number of jobs, and one in which it augments human labor and creates new ones.

“No matter what we did with AI in our model, demographics were the bigger story,” Ullrich said.

Ullrich also flagged the upcoming wealth transfer from baby boomers, the wealthiest generation in U.S. history, to Gen X and older millennials as a factor Indeed’s current model doesn’t yet capture.

“Economic theory would tell you that would impact the labor decisions of those people that get those transfers,” she said, suggesting retirement ages could fall further for the generations behind boomers unevenly, concentrated among white-collar workers likely to inherit the most.

That some of the drop in participation may be voluntary rather than distress-driven is also consistent with how financially secure households say they feel. In the Fed’s Survey of Household Economics and Decisionmaking, the share of adults who say they’re at least doing okay financially has held steady at 72%-73% for the past three years—down from a stimulus-inflated 78% in 2021, but not the kind of deterioration that would point to widespread economic desperation behind June’s exodus from the labor force. Still, that’s before the quick technological change that’s been made in the recent few years.

“It’s incredibly interesting to be in this demographic situation at the same time you’re in a period of technological innovation,” she said.

This story was originally featured on Fortune.com