BLS Releases Preliminary March 2026 Benchmark Revision: Downward Adjustment of 79,000 Jobs, Contrary to Some Pre-Release Expectations
BLS preliminary benchmark revision for March 2026 nonfarm payrolls is -79k, a downward adjustment that contrasts with some expectations of an upward revision; final figures due in 2027. Context includes prior large negative revisions and QCEW methodology.
The U.S. Bureau of Labor Statistics (BLS) released its preliminary estimate of the annual benchmark revision to nonfarm payroll employment for March 2026 on August 28, 2026, showing a downward adjustment of 79,000 jobs (-0.1 percent). The revision for total private employment was larger at -178,000 (-0.1 percent). This follows the final benchmark revision for March 2025, which was a substantial downward adjustment of 862,000 jobs (-0.5 percent).
The benchmark process aligns the sample-based Current Employment Statistics (CES) survey with more comprehensive counts from the Quarterly Census of Employment and Wages (QCEW), derived from state unemployment insurance records. The preliminary revision serves as an early indicator of the error in CES estimates from March 2025 to March 2026 and will be finalized in February 2027 alongside the January 2027 Employment Situation report.
Pre-release analyses, including from Goldman Sachs (as referenced in financial commentary) and independent economists, had suggested a possible modest upward revision—the first since 2022—based on nine months of QCEW data pointing to an undercount in CES figures amid slowing immigration. However, the actual preliminary figure came in negative, aligning more closely with the recent pattern of downward revisions. Historical data shows preliminary estimates have often understated the final revision by around 100k on average in recent years, and QCEW data itself has seen upward revisions in most quarters since 2019.
Industry breakdowns reveal variation: Construction saw a +62,000 revision, while sectors like Wholesale Trade (-86,200) and Retail Trade (-154,600) contributed to the net decline. Government employment received a +99,000 upward adjustment. These revisions reflect a combination of sampling errors, birth-death model adjustments, and other reporting factors, with past large negative revisions partly attributed to challenges in capturing unauthorized workers in UI records (less relevant this cycle due to immigration slowdowns).
BLS: Preliminary downward revision of 79k signals continued caution on recent payroll growth estimates; final 2027 update could shift modestly higher but unlikely to reverse broader slowdown narrative in labor data.
Sources (5)
- [1]Current Employment Statistics Preliminary Benchmark (National) Summary(https://www.bls.gov/news.release/prebmk.nr0.htm)
- [2]Table 1. National Current Employment Statistics March 2026 Preliminary Benchmark Revisions(https://www.bls.gov/news.release/prebmk.t01.htm)
- [3]How to interpret the annual jobs revision(https://budgetlab.yale.edu/research/how-interpret-annual-jobs-revision)
- [4]The Preliminary Benchmark Revision Estimate(https://macromostly.substack.com/p/the-preliminary-benchmark-revision)
- [5]CES Benchmark Announcement (March 2025 final)(https://www.bls.gov/web/empsit/cesbmk.htm)