BLS Preliminary Benchmark Revision Trims 79,000 From March 2026 Payrolls
The Bureau of Labor Statistics released its preliminary estimate of the Current Employment Statistics (CES) national benchmark revision on August 28, 2026, showing total nonfarm payroll employment for March 2026 revised down by 79,000, or -0.1 percent. The figure is calculated only for March 2026 across major industry sectors, per the BLS release table structure, and the underlying data has been made available for download from the BLS website. BLS
The revision is modest in percentage terms but lands against a labor market that has already been printing soft headline numbers. The July 2026 Employment Situation Summary (2026 M07) reported a preliminary change in total nonfarm payroll employment of -23,000 for the month. The same release revised May 2026 payroll growth down by 66,000, from +129,000 to +63,000, and trimmed June 2026 down by an additional 37,000. BLS
Taken together, the trajectory is clear: payroll gains have been sequentially revised lower across the spring and summer, and the March 2026 benchmark confirms that the level of employment entering this period was slightly lower than initially reported.
The CES benchmark revision process compares the survey-based payroll estimates, which are derived from the monthly sample of establishments, against administrative payroll tax records from the Quarterly Census of Employment and Wages (QCEW) collected by state unemployment insurance programs. The preliminary estimate released in August provides an early read; the final revision is typically incorporated with the January employment report the following year.
The March 2026 preliminary revision of -79,000 is small relative to recent precedent. The final benchmark revision for March 2025, published in February 2026, revised US employment growth down by 862,000 jobs for the 12 months through March 2025, equivalent to roughly 76,000 fewer jobs per month. That final figure came in below the BLS's own August 2025 preliminary estimate of -911,000, meaning the preliminary overestimated the downward adjustment by about 49,000. Reuters
The September 2025 preliminary estimate that preceded the final March 2025 revision had implied nonfarm payroll gains averaging approximately 71,000 per month over the benchmark period. Reuters
For state-level data, the average absolute percentage revision across all states for total nonfarm payroll employment was 0.9 percent as of September 2025, providing a useful dispersion benchmark: the national -0.1 percent revision for March 2026 sits well within the typical range of state-level adjustments. BLS
The magnitude matters. A -0.1 percent revision to the national payroll level is, by historical standards, negligible. The CES benchmark revisions have varied widely: the March 2025 cycle wiped out 862,000 jobs, while the March 2026 preliminary suggests a far smaller miscalibration. The sample-based estimates appear to have tracked administrative records more closely over the most recent benchmark period.
But the small benchmark revision does not mean the labor market is strong. It means the BLS's survey-based estimates were approximately correct about the level of employment in March 2026. What those estimates showed, and what the subsequent monthly releases have confirmed, is a labor market that has decelerated meaningfully. Nonfarm payrolls grew by 130,000 in January 2026, per Reuters reporting of BLS data. By July, the preliminary reading turned negative at -23,000, and the spring months were revised down by a combined 103,000. Reuters
For market participants, the benchmark revision itself is unlikely to shift expectations materially. A -79,000 adjustment to the March 2026 level translates to roughly 6,600 fewer jobs per month over a 12-month averaging window, well within the noise band of monthly survey error. The more actionable signal is in the monthly payroll trajectory: consecutive downward revisions to May and June, followed by a negative July print, describe a labor market that has moved from softening to contracting on the headline payroll measure.
The final March 2026 benchmark will be published with the January 2027 Employment Situation report. Until then, the -79,000 preliminary figure stands as the BLS's best estimate of the correction needed to align survey-based payroll counts with administrative records.
