How Common Are Monthly Payroll Job Losses during Business Expansions?

U.S. economic expansions are typically characterized by regular increases in output and expenditures.Expansion periods are defined and determined by the Business Cycle Dating Committee of the National Bureau of Economic Research (NBER). According to the NBER: “An expansion is a period when the economy is not in a recession. Expansion is the normal state of the economy.” For further details on how the NBER defines an expansion, see this webpage. Regular increases in employment also characterize an expansion. By contrast, employment growth tends to weaken on a persistent basis either concurrently with or shortly after a business cycle peak, and the economy slides into a recession.See Kevin L. Kliesen’s August 2024 blog post, “Employment Trends before and after Business Expansion Peaks.”

But sometimes employment can decline temporarily during an expansion for a variety of reasons. One reason is that nonfarm payroll employment—probably the best measure of aggregate employment—can be revised several times. Data revisions are the bane of policymakers and forecasters. What appears to be a solid gain in employment in one month could show something entirely different in the following month.

Indeed, this was the case earlier this year. In the employment situation (ES) report—also known as the “jobs report”—released on July 3, the Bureau of Labor Statistics (BLS) reported that total nonfarm payrolls increased by 147,000 in June, which modestly exceeded expectations (110,000). One month later, the July ES report showed that nonfarm payrolls rose by only 14,000 in June—a sizable downward revision. Then, in the following month, the August report showed that nonfarm payrolls actually fell by 13,000 in June.

The decline in nonfarm payrolls (number of jobs) in June suggested a possible weakening in the demand for labor, which can precede a downturn in the pace of economic activity. At the same time, though, real gross domestic product (GDP) growth was strong in the second quarter (3.3% at an annual rate), and GDP growth in the third quarter could also be quite strong according to the Atlanta Fed’s GDPNow tracking forecast.

Monthly Job Losses during an Economic Expansion

But are monthly job losses an unusual occurrence during economic expansions? The table below measures the number of monthly job losses during business expansions since 1958. Specifically:

• The number of months when job growth was negative according to the initial (real-time) estimate (column 4).
• The number of months when job growth was negative according to the current vintage estimates—that is, after accounting for all subsequent revisions to the initial estimate (column 5).
• The percentage of months during each expansion when there was a monthly job loss according to both real-time and current vintage estimates (columns 6 and 7, respectively).

The table shows that monthly declines in payroll employment are not unusual during business expansions. Using initial estimates, the number of monthly job declines during an expansion ranges from a high of 20 months (1991-2001 expansion) to a low of one month (current expansion); the average across all expansions is nine months, which is 14% of the average number of the expansion months.

These numbers decline slightly when using current-vintage estimates. As seen in column 5, subsequent revisions have reduced the maximum number of months from 20 to 18, and the smallest number of months to zero, which occurred during the short 1980-81 expansion. Moreover, the average number of expansion months with a negative growth in jobs falls from 14% to 9%.

Two other novelties are worth mentioning. First, the longest expansions—1961-69, 1991-2001 and 2009-20—tend to have the largest number of monthly job losses. A notable exception was the 2001-07 expansion, when job losses occurred in 21% of the months using initial estimates and 25% of the months using current-vintage estimates. However, the bulk of the job losses in the 2001-07 expansion occurred in roughly the first year and a half of the expansion, a period known as the “jobless recovery.”See Michael T. Owyang and Kristie M. Engemann’s 2010 article, “Whatever Happened to the Business Cycle? A Bayesian Perspective of Jobless Recoveries,” in Macroeconomic Dynamics.

Second, the current expansion since May 2020 is slightly less than average in duration (64 months through August 2025), but the percentage of negative monthly job losses is small—one month (2%) using initial estimates and two months (3%) using current-vintage estimates. Admittedly, these numbers could change in the future because of the BLS announcement on Sept. 9 that the preliminary estimate of actual nonfarm payroll employment in March 2025 based on the Quarterly Census Employment and Wages (QCEW) was 911,000 less than the current level of nonfarm employment that is based on the Current Employment Statistics (CES) survey.Accordingly, this means that CES-based nonfarm payrolls have potentially been overstated by about 76,000 per month from April 2024 to March 2025. During that period, there were two months when the CES-based payroll employment estimate increased by less than 76,000 per month: August 2024 (71,000) and October 2024 (44,000). The final March 2025 benchmark estimate will be published in February 2026 with the release of the January 2026 ES report. QCEW employment counts are derived primarily from state unemployment insurance tax records that nearly all employers are required to file with state workforce agencies while the CES surveys about 121,000 businesses and government agencies.

Still, the outlook for the economy remains positive, according to the Philadelphia Fed’s Survey of Professional Forecasters published in August. Private-sector forecasters expect continued positive growth in real GDP and nonfarm employment over the next four quarters. Thus, while the decline in nonfarm payrolls in June is perhaps worrisome—and future revisions may show more because of the final benchmark revision that will be released in February 2026—periodic monthly job declines are nonetheless a regular occurrence during business expansions.

The author thanks St. Louis Fed Research Associate Jack Fuller for research assistance.

Notes

1. Expansion periods are defined and determined by the Business Cycle Dating Committee of the National Bureau of Economic Research (NBER). According to the NBER: “An expansion is a period when the economy is not in a recession. Expansion is the normal state of the economy.” For further details on how the NBER defines an expansion, see this webpage.
2. See Kevin L. Kliesen’s August 2024 blog post, “Employment Trends before and after Business Expansion Peaks.”
3. See Michael T. Owyang and Kristie M. Engemann’s 2010 article, “Whatever Happened to the Business Cycle? A Bayesian Perspective of Jobless Recoveries,” in Macroeconomic Dynamics.
4. Accordingly, this means that CES-based nonfarm payrolls have potentially been overstated by about 76,000 per month from April 2024 to March 2025. During that period, there were two months when the CES-based payroll employment estimate increased by less than 76,000 per month: August 2024 (71,000) and October 2024 (44,000). The final March 2025 benchmark estimate will be published in February 2026 with the release of the January 2026 ES report.