Key Takeaways
- • The U.S. Bureau of Labor Statistics reported a preliminary benchmark revision of -79,000 nonfarm jobs (-0.1%) for March 2026, reflecting a modest downward adjustment to total employment levels based on comprehensive tax records.
- • Private employment was revised down by 178,000 (-0.1%) in the preliminary benchmark, indicating weakness concentrated in the private sector relative to total nonfarm figures.
- • The preliminary revision serves as a measure of cumulative error in Current Employment Statistics (CES) estimates from March 2025 through March 2026, derived from comparison with Quarterly Census of Employment and Wages (QCEW) data sourced from state unemployment insurance tax filings.
- • The 10-year average absolute benchmark revision stands at 0.2% of total nonfarm employment, placing the current -0.1% adjustment within historical norms for annual revisions.
- • The final benchmark revision will be issued in February 2027 coinciding with the January 2027 Employment Situation report, allowing for potential adjustments to the preliminary estimate.
Summary
The Bureau of Labor Statistics released its preliminary benchmark revision for March 2026 employment data on August 28, 2026, revealing a downward adjustment of 79,000 jobs to total nonfarm employment, or -0.1% of the overall level. This revision reflects the difference between the sample-based Current Employment Statistics estimates published throughout the year and the more comprehensive counts derived from actual unemployment insurance tax records filed by employers with state workforce agencies.
The private employment component showed a more pronounced revision, with a downward adjustment of 178,000 positions, also representing -0.1% of private sector employment. These preliminary revisions are derived from the Quarterly Census of Employment and Wages (QCEW), which aggregates nearly all employer filings and provides a definitive census of employment across the U.S. economy. The discrepancy between preliminary estimates and the census counts highlights the inherent statistical sampling error in monthly employment reporting, particularly as these errors compound over a 12-month benchmarking period.
Contextually, the -0.1% preliminary revision falls comfortably within historical precedent. Over the past 10 years, benchmark revisions have averaged 0.2% in absolute terms, suggesting the current adjustment represents a relatively modest correction to employment estimates. Individual industry sectors may experience larger percentage revisions than the aggregate figure, a common pattern attributable to greater statistical sampling error at more granular industry levels compared to aggregate nonfarm totals.
The BLS noted that the preliminary benchmark revision is subject to subsequent refinement. The final benchmark revision will be published in February 2027 alongside the January 2027 Employment Situation news release, providing an opportunity for additional methodological review and potential adjustment before the figures become final. For institutional investors and market participants relying on employment data for macroeconomic assessment, this modest preliminary downward adjustment warrants monitoring, though it does not materially alter the broader trajectory of labor market conditions established through 2026.
The private employment component showed a more pronounced revision, with a downward adjustment of 178,000 positions, also representing -0.1% of private sector employment. These preliminary revisions are derived from the Quarterly Census of Employment and Wages (QCEW), which aggregates nearly all employer filings and provides a definitive census of employment across the U.S. economy. The discrepancy between preliminary estimates and the census counts highlights the inherent statistical sampling error in monthly employment reporting, particularly as these errors compound over a 12-month benchmarking period.
Contextually, the -0.1% preliminary revision falls comfortably within historical precedent. Over the past 10 years, benchmark revisions have averaged 0.2% in absolute terms, suggesting the current adjustment represents a relatively modest correction to employment estimates. Individual industry sectors may experience larger percentage revisions than the aggregate figure, a common pattern attributable to greater statistical sampling error at more granular industry levels compared to aggregate nonfarm totals.
The BLS noted that the preliminary benchmark revision is subject to subsequent refinement. The final benchmark revision will be published in February 2027 alongside the January 2027 Employment Situation news release, providing an opportunity for additional methodological review and potential adjustment before the figures become final. For institutional investors and market participants relying on employment data for macroeconomic assessment, this modest preliminary downward adjustment warrants monitoring, though it does not materially alter the broader trajectory of labor market conditions established through 2026.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.