We start this story with a question: How low can unemployment go in this country?

It was 3.5% in February 2020, heading into the pandemic. In the pandemic recovery it fell as low as 3.4% (in April 2023) — a more than 50-year low. As of January 2025, the latest month for which we have data, it sits at 4% — still very low in historic terms. (The Bureau of Labor Statistics has been publishing modern unemployment statistics since 1948).

But could unemployment fall even lower than it’s been in recent years? To 2% ? Or 1.5%?

The question is prompted by a story I reported recently, in which I explained why unemployment couldn’t keep falling as rapidly in this most recent five-year period (2020 to 2025) as it had in the comparable period before the pandemic (2015 to 2025 — when it fell from 5.5% to 3.5%). I wrote that that would have meant unemployment falling to 1.5% — which I said “economists will tell you, pretty much can’t happen.”

And then I heard from fellow Marketplace reporter Stephanie Hughes, who asked a simple but probing question: “I know that below 4% is really good. But why can’t we get to 1.5%?”

And right off the bat, I found an economist — Julia Pollak, chief economist at ZipRecruiter — who said that it’s not unthinkable.

“We know from some states and cities in America that very low unemployment rates are possible,” said Pollak. “In 2024, South Dakota had an unemployment rate of just 1.9%.”

But at the national level, unemployment’s only ever fallen as low as 2.5%, for two months in 1953. (The unemployment rate referred to here is the “official” rate, U-3, with unemployed workers defined as those who do not currently have a job and have actively looked for work in the past four weeks.)

One reason the unemployment rate hasn’t gone lower, said former Federal Reserve economist Claudia Sahm, author of the Sahm Rule, now at New Century Advisors, is that certain demographic groups face barriers to full employment, including lack of education and training, and discrimination.

Black unemployment has historically been about twice as high as white unemployment. In South Dakota, unemployment on some Native American reservations is 80% or higher.

“There are mismatches of the skills that workers have, and the geographies of where they are,” said Sahm. “And that is the limit on the national unemployment rate, these pockets of structurally much higher unemployment.”

There is another kind of unemployment that keeps the rate well above 1.5%, said Betsey Stevenson at the University of Michigan: “Something that economists call ‘frictional unemployment.’ The flies in the ointment that prevent workers and jobs from finding each other right away.”

This is the unavoidable churn — in good times and bad — as workers leave one job to find another, or graduate from school and start looking. And we actually want some frictional unemployment. It’s a sign of a healthy labor market, said economist Michael Strain at the American Enterprise Institute.

“Being unemployed for a few weeks and finding the best match you can — you’re more productive, you’re contributing more to the firm, to the economy, it means that you’re earning a higher wage,” said Strain.

“If the unemployment rate got down to 1.5% or something like that,” Strain continued, “I’d be worried that for some reason workers were just scrambling to take the first job they could find. Or, has something happened that has led workers that are relatively harder to employ to exit the workforce entirely, such that only the most employable workers are left.”

We might over time develop better technology to match workers and employers faster, said Strain, which could reduce frictional unemployment a bit.

And that would be a good thing, said economist Heidi Shierholz, president of the Economic Policy Institute. “When unemployment is low, that’s great for workers and the economy. But it actually can get too low.”

Economists tend to agree that anything lower than 2.5% to 3% risks extreme labor shortages developing in the economy.

“Firms who have job openings, they’re just basically poaching workers from other companies, because there’s hardly any unemployed people to hire,” said Shierholz. “That requires big wage increases, and then that translates into big price increases.”

In other words, inflation. The Fed would raise interest rates to fight it, driving unemployment back up again.

And it’s not only a hot labor market that can lead to super-low unemployment, said Betsey Stevenson: “A low unemployment rate could be associated with a very stagnant labor market, where there are no jobs, so there’s no point in looking,” she said.

This is Stevenson’s tech-driven nightmare scenario, where artificial intelligence and robots are able to do the work of most humans — cheaper than we can do it ourselves.

“Human wages are pushed so low that maybe people don’t even want to work,” said Stevenson. “Maybe they can’t even survive while working.”

So, bottom line: Unemployment could fall lower than it has in the last 70 years, maybe as low as 1.5%. But if it ever does, we should be worried that something is seriously wrong with the U.S. economy.