In the first twenty seconds of television coverage, virtually no one brings up a figure hidden in the monthly jobs data. The banner image does not show it. Seldom does it appear in the first paragraph. However, compared to the headline payroll figure, it may reveal more about the real situation of the American labor market. It’s known as the employment diffusion index, and it has been subtly indicating something important in recent months.
On the surface, the May 2026 report appeared to be truly impressive. According to the Bureau of Labor Statistics, 172,000 new jobs were created, which is more than twice as many as the Dow Jones consensus predicted. Previous months were also raised by revisions, with March settling at 214,000 and April rising by 64,000. It was dubbed the end of the hiring recession by some economists. According to one analyst, it was powerful “from every angle.” That Friday morning, the atmosphere was noticeably lighter than it had been in months as one passed the financial terminals on any major trading floor.
However, that picture is significantly complicated by the diffusion index. Construction, finance, manufacturing, retail, technology, and hospitality are just a few of the industries that move more or less together when job creation is truly healthy. When the index declines, it indicates a more precarious situation: gains accumulating in one or two sectors of the economy while the rest remains stagnant or subtly decline. Nearly all of the net gains in numerous recent reports have been attributed to private healthcare, education, and government. When those sectors are eliminated, the private economy as a whole has occasionally resulted in actual net losses. It’s not a robust labor market. A headline is being held up by a one-legged stool.
May did exhibit a slightly wider range of employment than previous months, with local government contributing 55,000 jobs and leisure and hospitality adding 70,000 jobs, partially due to World Cup-related hiring. However, financial activities are currently down 107,000 from this time last year, having lost 22,000 positions in May alone. Over the last 12 months, employment has decreased in exactly half of all tracked sectors. Typically, that type of split isn’t referred to as “strong.” At best, it’s referred to as complicated.
Then there are the downward revisions, which the BLS quietly makes months after the initial headline was published. When more detailed data becomes available, preliminary figures are adjusted, and these adjustments often go in one direction. Early reports from last summer indicated that job creation was significantly above trend. A different story was conveyed by the revisions. It takes a certain amount of deliberate optimism to treat a single month’s figure as definitive truth because the pattern is repeated so frequently.
Long-term unemployment is another issue that is difficult to ignore. The percentage of Americans without a job for 27 weeks or longer increased to 27.5% in May, a significant increase from 20.4% only a year earlier and significantly higher than pre-pandemic averages. Yes, the unemployment rate remains at 4.3%. However, even when workers completely leave the workforce and the job search takes weeks or months, that rate can remain mathematically stable. Two of the clearer indicators of true economic confidence, the hiring and quitting rates, are both still extremely low. Employees who are employed are staying put because the off-ramps have mostly vanished, not because they are content.
There is a version of this moment where everything is good, where long-term unemployment eventually returns to normal, revisions remain positive, and 172,000 jobs are exactly what they appear to be. That version is conceivable. However, the hidden figure, which measures the actual distribution of job creation throughout the economy, consistently indicates that a more cautious approach is necessary. For a considerable amount of time, a market that is stuck between low hiring and low firing may appear stable. Until it isn’t.
