BLS Revisions Show 2025 US Job Gains Were Far Lower Than First Reported

U.S. job creation has been overstated in official tallies, according to updated Bureau of Labor Statistics (BLS) benchmark revisions. On Aug. 28, 2026, the BLS released a preliminary annual benchmark revision indicating the economy added 79,000 fewer jobs than initially reported over the 12 months ending March 2026. A bigger reset arrived earlier. In February 2026, the BLS finalized its revision to 2025 employment data. Early estimates put 2025 job growth at about 584,000; the revised count was just 181,000. That implies a downward adjustment of roughly 862,000 to 898,000 jobs, marking one of the largest percentage reductions since 2009. The August 2026 preliminary update also cut total private payrolls by 178,000. Recent monthly reports have pointed in the same direction: the July 2026 jobs report revised May down by 66,000 and June down by 37,000, a combined 103,000 reduction across the two months. Why revisions have been so large Monthly payrolls are built from the Current Employment Statistics survey, which samples businesses. Each year, the BLS benchmarks those estimates against the Quarterly Census of Employment and Wages (QCEW), a broader dataset based on unemployment insurance records. Over the past decade, annual benchmark revisions have averaged about 0.2% of total nonfarm employment. The 2025 revision far exceeded that norm. One factor is lower survey response rates. With fewer businesses responding, the BLS relies more heavily on models to fill gaps. The "birth-death" model, which estimates net job creation from business openings minus closures, has faced increased scrutiny because it assumes relatively steady business formation—an assumption that can weaken when economic conditions deteriorate. Implications for markets and the Fed For the Federal Reserve, weaker employment results complicate the policy picture. If job gains were materially smaller than real-time reports suggested, policymakers may have been assessing economic momentum with an overly optimistic view during key decision windows. The next milestone is the final benchmark revision due in February 2027, when the BLS will incorporate the full QCEW dataset and replace preliminary estimates with final figures.