The American labour market has acquired a strange property

THE AMERICAN labour market has acquired a strange property. In a normal economy, the disappearance of jobs is a sign that something is wrong. In today's America, it may be perfectly healthy.

Until recently, the economy needed to create 150,000 or more jobs a month simply to keep the unemployment rate from rising. That is the "breakeven" level: the pace of hiring required to absorb new entrants to the workforce and prevent the share of jobless people from climbing. The figure is not a magic number; it depends on how fast the population of working-age Americans is growing and how many of them are looking for work. As of mid-2026, thanks to a sharp slowdown in immigration and the steady drain of Baby Boomers into retirement, the breakeven level has collapsed to a range of 15,000 to 87,000 jobs per month, according to calculations by a St Louis Fed economist. Under the low-immigration scenario, the breakeven is about 15,000 jobs per month, so the unemployment rate can hold steady with job growth near that level.

This is not a cyclical slowdown. It is arithmetic. The Congressional Budget Office initially projected net immigration of 175,000 people per month in 2025 and revised that estimate down to 34,000, an 81% drop. The younger cohorts entering the workforce are smaller, a consequence of decades of sub-replacement fertility. Meanwhile, more than 11,400 Americans turn 65 every day. Fewer people are arriving, fewer are being born, and more are leaving. The denominator in the unemployment fraction is shrinking, which means an unchanged unemployment rate now requires the number of unemployed to fall too; an increasing numerator would push the rate up. A shrinking labour force is therefore what allows the unemployment rate to hold steady even with very low job growth.

The practical consequence is that the familiar signals of a strong labour market have become unreliable. Monthly payrolls that fall into negative territory are no longer, by themselves, evidence of a downturn. Indeed, in January and February 2026, total employment growth averaged around 17,000 jobs per month – well below even the most pessimistic breakeven estimates from the past – yet the unemployment rate barely budged. If the market still operated on old rules, that would be alarming. It reflects new rules.

The trouble is that the headline numbers mask a labour market that is growing more fragile, not less. Aggregate statistics conflate two entirely different sets of dynamics. Research from the Dallas Fed finds that the economy is operating in what it calls a "hire-less, fire-more" regime for the unemployed. The rate at which workers are being hired from unemployment has consistently fallen below its pre-pandemic level, while the firing rate has exceeded both the long-term structural trend and pre-pandemic baselines. At the same time, workers who are already employed continue to move between jobs at a robust pace. The headline hiring figures look respectable because they include all job-to-job transitions among the employed, a category that has barely cooled. The people who need jobs most are the ones finding the fewest of them.

The evidence is granular. In June, 27.3% of unemployed Americans had been out of work for 27 weeks or more, near the highest share since late 2021, according to the Bureau of Labour Statistics. That is about 1.9 million people, trapped in a secondary labour market where opportunities are thin and the rungs of the employment ladder are broken. The same data show that 4.7 million people are working part time for economic reasons – they would prefer full-time work but cannot find it. Another 6 million want a job but are not counted as unemployed because they have stopped actively looking. These figures do not make the front page, but they are the real story.

A Fed working paper from May reinforces the discomfort. It finds that low population growth and near-zero breakeven employment make the labour market more prone to reporting negative job gains, even when the economy is performing at potential. That does not, the authors caution, mean the market is systematically more vulnerable to recessions. But it does mean the usual dashboard lights are dimmer. The Fed cannot watch a single jobs number and know what it is looking at, as it once could.

To be sure, the market has real sources of resilience. Job openings were unchanged at 7.6 million in May. Hires held at 5.2 million and layoffs at 1.7 million, levels that are neither alarming nor exuberant. Wage growth, at 3.5% year-over-year, is steady if no longer hot. The Federal Reserve, which cut the federal-funds rate by 75 basis points in the autumn of 2025 amid fears of a labour-market collapse, has shifted its anxiety to inflation. The market did not break. It adjusted.

Yet the adjustment has concentrated almost all of the pain on the wrong people. The dual-market structure means roughly 55% of the population enjoys primary-sector stability – high wages, low turnover, negligible risk of unemployment – while a much smaller secondary sector, just 14% of the population, accounts for 61% of total unemployment and the vast majority of cyclical volatility. The employed are insulated. The unemployed are exposed. That is not a balanced recovery. It is a bifurcation.

The policy implications are uncomfortable. Monetary policy is poorly equipped to deal with a labour market whose headline indicators have stopped tracking underlying fragility. Cutting rates in response to a payroll miss may be unnecessary; holding them steady because unemployment looks fine may be equally wrong. The Fed's dual mandate of price stability and full employment was designed for a labour market where the two could be weighed against each other. In a market where the unemployment rate has lost its informational content, the mandate is harder to execute.

The deeper problem is not cyclical; it is institutional. A shrinking labour force means that future GDP growth will depend more on productivity than on headcount. That is a familiar observation, but the labour-market consequence is less discussed: an economy that cannot absorb workers will see structural unemployment rise even as output expands. Hiring Lab, a labour-market research group, projects that the mismatch between where workers are and where jobs need to be could push the aggregate unemployment rate between 0.5 and 3.5 percentage points higher by 2040. That is not a recession. It is a failure of reallocation.

The better answer is not to prop up employment through blunt demand stimulus. It is to reduce the frictions that prevent workers from reaching the jobs that exist. Credential reform, better matching systems, and upskilling programmes are not sexy policies. They do not produce press releases or rally markets. But they address the actual mechanism of the problem: a workforce that is not being reallocated as fast as the economy is changing. Immigration policy, which has become the single largest wildcard in the breakeven equation, should be treated with the same seriousness. If net immigration falls to the low end of current projections, the structural mismatch will deepen. If it returns to something closer to historical norms, the labour market gains both workers and headroom.

The American labour market is not collapsing. It is changing its rules. The old benchmarks – the monthly jobs target, the headline rate – were built for a different demographic regime. They no longer tell you what they used to. Policymakers who continue to read them as if nothing has changed will mistake fragility for stability. The market needs fewer jobs to look healthy, and that is the kind of good news that should worry you.