The latest US jobs report, released by the Bureau of Labor Statistics (BLS) covers January 2026 data; delayed slightly due to a partial government shutdown.
It showed nonfarm payrolls increased by 130,000, beating economists’ expectations, which had ranged around 55,000 to 70,000 like the Dow Jones consensus at 55,000, Reuters/LSEG around 70,000. Nonfarm payrolls: +130,000 vs. expectations of ~55,000–70,000; prior month December revised to +48,000.
Unemployment rate: Edged down to 4.3% from 4.4% in December; better than forecasts expecting it to hold steady at 4.4%. Private sector added 172,000 jobs, while government especially federal saw declines of around 42,000.
Job gains were led by health care and social assistance; combined over 120,000, construction (+33,000), and business and professional services (+34,000). Sectors like retail, leisure/hospitality, federal government, and financial activities saw little change or losses.
Register for Tekedia Mini-MBA edition 19 (Feb 9 – May 2, 2026).
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab.
Wages rose solidly (average hourly earnings up ~0.4% monthly, ~3.7% annually). A broader unemployment measure (U-6, including underemployed) fell to 8.0%. This marked a solid start to 2026 after a very weak 2025, providing some relief to labor market concerns.
However, the report included significant annual benchmark revisions that painted a much softer picture for prior years:2025 job growth was revised down sharply to only +181,000 total from an initial ~584,000 estimate, averaging ~15,000 per month—the weakest non-recession year in decades.
Revisions also lowered prior estimates for 2024. These revisions (based on more complete data like unemployment insurance records) suggest the labor market was closer to stalling in 2025, with growth heavily concentrated in health care and limited elsewhere, some blue-collar sectors even saw net losses over parts of the period.
Markets and analysts viewed the January beat as a positive surprise, potentially signaling stabilization or early recovery momentum into 2026—possibly influenced by factors like warmer weather boosting construction or policy effects—but tempered by the weak revisions and ongoing questions about sustainability beyond essential sectors.
Benchmark revisions in the US jobs report refer to the annual process by which the Bureau of Labor Statistics (BLS) adjusts its monthly estimates of nonfarm payroll employment from the Current Employment Statistics (CES) survey to align them with more comprehensive and accurate “universe” counts of employment.
The CES survey provides the headline nonfarm payroll numbers like the +130,000 jobs added in January 2026. It surveys about 121,000 businesses and government agencies each month for timely data.
However, this survey has limitations: response rates aren’t 100%, it can miss newly formed businesses (births) or not immediately detect closures (deaths), and it uses models like the birth-death model to estimate net changes from these unobserved events.
To improve accuracy, the BLS annually “benchmarks” these estimates using near-complete administrative data, primarily from the Quarterly Census of Employment and Wages (QCEW). The QCEW draws from state unemployment insurance (UI) tax records, covering nearly all nonfarm employees with minor supplements from other sources.
These records are more exhaustive but less timely—they become available with a lag, so benchmarking focuses on a specific month typically March of the prior year. The difference is the benchmark revision e.g., -898,000 jobs for March 2025, or -0.6%. Wedge back and forward — To create a smooth, continuous time series.
Adjust monthly estimates between the previous March benchmark and the new one using a linear “wedge-back” procedure (spreading the revision proportionally across intervening months). Extend adjustments forward to later months up to about 9-10 months after March on a not seasonally adjusted basis, with seasonal factors recalculated.
This release also includes final monthly revisions from additional survey responses and seasonal factor recalculations. Historical data often back 1-2 years, with seasonal adjustments sometimes farther get revised. In the February 11, 2026 release: The March 2025 benchmark level was revised downward by about 898,000 jobs.
This led to 2025 total job growth being slashed from an initial ~584,000 to just +181,000 averaging ~15,000 per month—the weakest non-recession year in decades. Some 2024 estimates were also lowered; cumulatively over a million fewer jobs than previously reported across recent years.
These revisions are normal and transparent—BLS has done them since the 1930s, with preliminary indications released in summer and fall and finals in February. Historically, average absolute revisions are small ~0.2% over the prior decade, but they’ve been larger recently, possibly due to factors like shifts in business dynamics, immigration patterns affecting the birth-death model, or economic turning points where models over- or under-estimate.
In short, benchmark revisions make the data more accurate over time by trading initial timeliness for better completeness. The January 2026 report’s strong beat (+130K jobs, unemployment to 4.3%) stood out against these downward revisions to prior years, highlighting a potential stabilization in early 2026 despite the softer underlying picture for 2025.