The January jobs report delivered a headline that looked like good news: 130,000 new nonfarm payroll jobs added, above most forecasts. But buried beneath that number is a revision that recasts the labor market story. Still, experts say hiring is happening, it’s just focused and deliberate.

As part of its annual benchmarking process, the Bureau of Labor Statistics revised total nonfarm job growth for 2025 from 584,000 down to 181,000. That means the economy added roughly a third of the jobs previously reported last year, averaging just 15,000 per month.

The short version of the revision methodology is a growing gap between what BLS’s monthly employer survey captures and what comprehensive tax records show, driven in part by declining survey response rates and a statistical model that overestimated job creation at new businesses.

What January actually shows

The January gain of 130,000 jobs was led by healthcare, which added 82,000 positions, according to BLS data. Social assistance contributed 42,000 jobs, primarily in individual and family services. Construction added 33,000, driven largely by nonresidential specialty trade contractors.

On the other side of the ledger, federal government employment fell by 34,000 and financial activities lost 22,000 jobs in January, with insurance carriers and related activities accounting for about half of that decline.

The unemployment rate changed little at 4.3%, up from 4% a year earlier. Average hourly earnings for all employees on private nonfarm payrolls rose 0.4% to $37.17, with year-over-year growth at 3.7%.

A note on why the revision matters

Every year, BLS re-anchors its monthly survey estimates to comprehensive employment counts derived from unemployment insurance tax records that nearly all employers are required to file. The monthly jobs report is a sample of about 119,000 businesses. The tax records are a near-complete count covering roughly 97% of nonfarm payroll jobs. The gap between the two is the benchmark revision.

BLS has pointed to two primary drivers: response error, meaning businesses reported less employment to the tax records than they reported to the survey, and nonresponse error, meaning businesses that didn’t respond to the survey at all turned out to have weaker employment than those that did.

The survey’s response rate has declined from roughly 60% before the pandemic to about 43%, according to BLS data. When fewer than half of surveyed employers respond and the statistical model estimating job creation at new businesses runs consistently hot, the picture looks healthier than reality until the tax records come in a year later and correct it.

Read the full article here.