Revisions to employment data have recently been the focus of an inordinate amount of attention. The latest story on this topic is the preliminary benchmark revisions to Nonfarm Payroll Employment that were announced this morning by the Bureau of Labor Statistics.
Annual benchmark revisions are primarily the result of incorporating data from the Quarterly Census of Employment and Wages (QCEW). As suggested by the word “census” in the title, this data set is a comprehensive summary of national employment, based on data from state unemployment insurance programs. Benchmarking the payroll data introduces significant new, more accurate information to the series on current employment.
To complicate the story further, the actual benchmark revisions will not take place until early 2026. The data released this morning represent preliminary estimates of the revisions. The actual benchmark revision will be based on revised data from the QCEW which will be available at the end of this year. Accordingly, month-to-month reports on nonfarm payroll employment will not be affected by today’s estimates, and there is often a considerable difference between the preliminary and actual benchmark revisions.
With those provisos in mind, we can evaluate the potential impact of the preliminary benchmarked data on recent trends, assuming that the QCEW data revisions between now and the end of the year will be minimal.
The preliminary benchmark revisions announced this morning affect estimates for March 2025. For the entire U.S., the preliminary benchmark suggests a downward revision of 911,000 jobs—amounting to 0.6% of total March employment. The preliminary benchmark revision for Arkansas is a downward revision of 11,300 jobs, or 0.8% of total state employment.
The figure below illustrates the magnitude of the potential revisions, assuming that the cumulative revisions reported for March 2025 were spread across the previous 12 months, with data for March 2024 unaffected. Data for April 2025 and beyond is simply extrapolated from the hypothesized revision to March data.
These hypothetical revised series are intended to be suggestive only. The actual benchmark revisions will not be released or incorporated into current data until the numbers for January 2026 are reported next year. But if the preliminary numbers hold, these hypothetical calculations indicate that employment growth over the past 15 months has been considerably lower than has been previously reported. For July 2024 through July 2025, Arkansas nonfarm payrolls have been reported to expand by 18,000 jobs or 1.3%. With the hypothesized revisions, that growth slows to only 9,800 jobs, amounting to 0.7%. For the same period, U.S. employment growth might be revised down by as much as half—from current estimates of 1.0% to a revised 0.5%.
For more information, see:
- https://www.bls.gov/news.release/prebmk.nr0.htm (U.S. data)
- https://www.bls.gov/news.release/presae.nr0.htm (State and area data)