Recent trends are sending confusing signals about the health of the U.S. labor market. Since late 2025, the employment/population ratio has moved lower even as the unemployment rate has declined. This appears to be without precedent in modern history: Since 1949, for every comparable decline in the employment/population ratio, the unemployment rate rose, and typically by more than the employment/population ratio fell (according to the U.S. Census Bureau and the Bureau of Labor Statistics (BLS)).
This cycle has been different, largely due to an unusually rapid contraction in labor supply. Various structural changes in the labor market are likely contributing, including more retirements among older, potentially higher-income workers. Evidence suggests that overall labor demand remains tepid as well, and these structural forces may themselves be limiting labor costs.
Conflicting signals about U.S. labor markets complicate the Federal Reserve’s pursuit of its dual mandate (maximum employment and price stability). However, it does seem clear that labor markets are not a source of inflationary pressure; the net effects of the structural forces affecting labor markets include moderating wage inflation and subdued unit labor costs. This should help the Fed respond appropriately to broader inflationary pressures as needed.
Confusing labor market signals
At 4.1% in July, the U.S. unemployment rate sits below its late 2025 peak and has drifted lower through 2026, according to the BLS. On the surface, lower unemployment points to a resilient labor market. Yet falling wage growth, labor-cost pressures, tepid payroll gains, and the falling employment/population ratio are all sending a different signal.
Rather than reflecting strengthening labor demand, the decline in unemployment appears increasingly tied to shrinking labor supply. Since December 2025, the labor force participation rate has fallen a full percentage point to its lowest level (outside the COVID pandemic) since 1976, according to the BLS.
To illustrate the magnitude of this drop in the participation rate, consider what the unemployment rate would be under a hypothetical scenario where the labor force grew in line with the population – i.e., assuming the labor force participation rate was unchanged: The unemployment rate would have been roughly 5.6% in July 2026, 1.5 percentage points above the current reading and the highest level since 2021.
While slow-moving demographic factors, including the aging population, tend to reduce labor force participation, declines of this magnitude in such a short period have been very unusual outside of recessions, when cyclically weak labor market conditions leave people on the sidelines.
So what’s different this time? The faster contraction in labor supply partly reflects statistical revisions at the BLS. The January update to incorporate the latest population data resulted in a higher reported share of older Americans, who tend to have lower participation rates than prime-age Americans due to higher rates of retirement as people age. This compositional shift explains just under 40% of the decline in the participation rate this year, leaving the bulk attributable to other factors within age cohorts.
Structural labor market changes are underway
A wave of retirements appears to be one important factor. Starting in the early 1990s, rising life expectancy, Social Security rule changes that increase incentives to work longer, and the shift from defined benefit toward defined contribution retirement plans all contributed to increasing labor force participation among older Americans.
However, more recently that trend has flattened out and even started to reverse. Substantial wealth gains generated by post-pandemic appreciation in housing, equities, and other assets may have enabled some households to retire earlier than previously planned. The post-COVID stagnation in life expectancy as well as return-to-office policies and technological labor reallocations may also be contributing to lower participation rates among older Americans.
This isn’t the only story, however: Participation has also fallen among younger and prime-age workers. After adjusting for the population data update, these groups account for nearly one-third of the total year-to-date decline in labor force participation – but the factors driving these changes are less clear.
One possible explanation is that unemployed workers are finding it increasingly difficult to secure jobs and, in some cases, are leaving the labor force altogether. The job-finding rate, defined as the probability that someone who doesn’t have a job (regardless of whether they were previously looking or not) becomes employed in the following month, has fallen to 5% recently, versus 6%–7% just after the COVID period, and 7%–8% in the 1990s and 2000s, according to the BLS. Such declines are unusual during mature expansions, which normally pull marginal workers into employment rather than push them out of the labor force. More notably, the deterioration in job-finding rates has been concentrated among prime-age and college-educated workers, groups that have historically remained relatively insulated from labor market weakness.
Several cyclical and structural forces could be contributing, including slower hiring, immigration-driven changes in labor supply, policy uncertainty, and potentially the early effects of labor-displacing technologies, including AI. AI-related labor market reallocation seems particularly plausible since the decline in hiring and job-finding rates appears concentrated among younger and more highly educated workers, the group most exposed to early adoption of AI technologies.
Macroeconomic and policy implications
For inflation, the confluence of indicators continues to suggest that labor is not a key source of inflationary pressure. Despite the confusing supply and demand picture in labor markets, measured wage and unit labor cost inflation have continued to ease in 2026. Many highly paid older workers are leaving the labor market and their roles are largely unreplaced, and this contributes to more subdued labor cost trends. Furthermore, the combination of falling prime-age labor force participation rates along with easing wage inflation suggests that tepid labor demand may be contributing to the falling labor supply, not the other way around.
For policymakers, the confusing confluence of labor market signals could argue to avoid overreliance on any one indicator to assess the labor market. For example, Taylor rules that use the unemployment rate as the signal of labor market slack may be overestimating the degree of tightness in the labor market. Such rules now suggest that policy is substantially more accommodative than it should be.
However, as illustrated by the Atlanta Fed’s Taylor Rule Utility, which calculates a wider range of rules, substituting in the employment/population gap presents a very different picture. Indeed, an employment/population-rate-based Taylor rule prescribes policy rates at 100 basis points below similar rules using the unemployment rate.
Overall, while various shocks indicate near-term upside risks to inflation, the current Fed policy rate appears appropriately calibrated to respond to them, especially when labor costs don’t appear to be a primary source of inflationary pressure.
