Data Analysis

The Bureau of Labor Statistics reported that nonfarm payroll employment rose by 29,000 in September 2026 and revised the July and August job counts down by a combined 60,000.

The unemployment rate was 4.2 percent, which the agency said changed little from August. July now reads as a month of job losses, after two revisions moved it from a reported loss of 23,000 to a gain of 21,000 and back to a loss of 10,000.

Nonfarm payroll employment rose by 29,000 in September 2026, the Bureau of Labor Statistics reported on Friday, October 2, and the same release revised the two previous months down by a combined 60,000 jobs. The unemployment rate was 4.2 percent. July 2026, first reported as a loss of 23,000 jobs and then revised to a gain of 21,000, now reads as a loss of 10,000. Average hourly earnings rose 5 cents, or 0.1 percent, to $37.81.

The agency's own language for the month is worth quoting before anything else. Its opening sentence reads: "Both nonfarm payroll employment (+29,000) and the unemployment rate (4.2 percent) changed little in September." BLS uses "changed little" to mean a movement it does not consider meaningfully different from zero given the survey's sampling error. So the 4.2 percent rate, up from 4.1 percent in August, is not a reported increase in unemployment. It is a reading the agency declines to call a change.

What did the September jobs report actually show?

The establishment survey put the September gain at 29,000 jobs. The household survey put the unemployment rate at 4.2 percent and the number of unemployed people at 7.1 million, both described as little changed. The labor force participation rate was 61.8 percent and the employment-population ratio was 59.2 percent, both also little changed.

Two household-survey figures carry more information than the headline rate. The number of long-term unemployed, meaning people jobless for 27 weeks or more, was essentially unchanged at 1.9 million, and that group accounted for 27.1 percent of all unemployed people. The number of people working part time for economic reasons changed little at 4.5 million. When more than a quarter of the unemployed have been out of work over six months, the people losing jobs are having trouble finding new ones, whatever the headline rate does.

On pay, average hourly earnings for all employees on private nonfarm payrolls edged up 5 cents, or 0.1 percent, to $37.81, and have risen 3.0 percent over the past 12 months. The average workweek stayed at 34.4 hours.

Why the revisions matter more than the monthly number

BLS revised July down by 31,000, from a gain of 21,000 to a loss of 10,000, and August down by 29,000, from a gain of 162,000 to a gain of 133,000. That is 60,000 jobs removed from the record, against 29,000 added for the new month.

The arithmetic that follows is the desk's own calculation from the figures BLS published. Last month's report left August employment at 159,075,000 jobs, which is June's level of 158,892,000 plus the 21,000 then reported for July and the 162,000 then reported for August. This month's report puts September at 159,044,000. So after a month that the agency reported as a gain, total payroll employment now sits 31,000 lower than last month's release implied it already was. The new month did not offset the revision to the old ones.

July is the clearest case of why a single print deserves little weight. It was first published on August 7 as a loss of 23,000 jobs. The September 4 release revised it up by 44,000 to a gain of 21,000. This release revised it down by 31,000 to a loss of 10,000. Three vintages, two sign changes, and a 44,000-job spread between the best and worst readings of the same month. Anyone who built an argument on July's figure in early August was arguing about a number that no longer exists.

The reusable frame here is the same one that applied to the Census construction spending release a day earlier: when the revision is larger than the new month's change, the level is the story and the print is noise.

Where did the September jobs come from?

BLS named four industry movements in the summary. Health care added 17,000 jobs. Construction added 11,000. Manufacturing added 9,000. Financial activities, the category that contains real estate, lost 7,000. The agency described the month's other major industries as little changed.

Construction is the line this publication watches, and 11,000 jobs nationally is a small number in an industry employing millions. One month of an industry figure, from a survey the agency itself describes as showing little change, is a data point. It is not a direction.

What does a weak jobs report mean for Florida housing?

Two things, and neither of them is in this release.

Florida's own labor data publishes separately. The most recent state figures, released September 18 and covering August 2026, put Florida's unemployment rate at 4.5 percent. That was down 0.1 percentage point from July's 4.6 percent, a change BLS marked as statistically significant, and up 0.5 percentage point from 4.0 percent in August 2025, also significant. The national rate that month was 4.1 percent, so Florida sat 0.4 point above the country. Florida's September rate is scheduled to publish later in October, and nothing in Friday's national release speaks to it.

The second thing is the cost of money. Freddie Mac's Primary Mortgage Market Survey for the week of October 1 put the 30-year fixed rate at 7.28 percent, up 0.25 point in a week and the sixth consecutive weekly increase. A labor market that the statistical agency describes as changed little, combined with a mortgage rate near its highest level of the year, is a demand environment where buyers are neither gaining income momentum nor getting cheaper credit. That combination shows up in Florida as the price-cut and days-on-market patterns tracked on the Florida housing tracker and across the market hubs, including Miami, Tampa and Jacksonville.

Does one soft month change the 2026 picture?

Not on its own. Using the current vintage of the payroll series, employment has grown by 612,000 jobs over the first nine months of 2026, an average of 68,000 a month, and by 496,000 over the 12 months through September, an average of 41,300 a month. Two of the nine months of 2026 now show declines, February and July. Those averages are the desk's calculation from the published monthly levels, and every one of them is subject to the same revisions that rewrote July twice.

Limitations

This release contains no state or metropolitan detail. Florida figures quoted here come from the separate State Employment and Unemployment release of September 18, 2026, covering August, and are preliminary and subject to revision the following month.

BLS describes payroll employment, the unemployment rate, the participation rate and the employment-population ratio as having "changed little" in September. No claim is made here that any of those measures rose or fell, because the agency does not make that claim.

The industry figures are single-month changes from the establishment survey and carry sampling error that the summary does not quantify per industry. The agency characterized the month's changes outside the four it named as little changed, so no industry figure here should be read as a trend.

Figures were read twice from the BLS summary, and only values identical across both reads are used. The average hourly earnings figure for production and nonsupervisory employees appeared in one read and not the other, so it is omitted. The long-term unemployment figures rest on one read; they are quoted directly and are consistent with the 7.1 million total, which gives a 26.8 percent share against the 27.1 percent stated, a difference explained by rounding.

Payroll levels and the derived monthly changes were cross-checked against the Federal Reserve Bank of St. Louis PAYEMS series, updated at 9:24 a.m. ET on October 2, which matches the BLS release on June, July, August and September to the thousand. The vintage comparison and the monthly averages are the desk's calculations, not BLS publications.

Mortgage rate figures are from Freddie Mac's Primary Mortgage Market Survey and were verified on October 1.

Sources

Bureau of Labor Statistics, Employment Situation Summary, September 2026, released October 2, 2026: https://www.bls.gov/news.release/empsit.nr0.htm

Bureau of Labor Statistics, Employment Situation Summary, August 2026, released September 4, 2026, USDL-26-1435, for the first-reported July and August figures. Read on October 2 at https://www.bls.gov/news.release/pdf/empsit.pdf before that document rolled to the September release.

Bureau of Labor Statistics, State Employment and Unemployment Summary, August 2026, released September 18, 2026: https://www.bls.gov/news.release/laus.nr0.htm

Bureau of Labor Statistics, Unemployment Rates for States, August 2026: https://www.bls.gov/web/laus/laumstrk.htm

Federal Reserve Bank of St. Louis, All Employees, Total Nonfarm (PAYEMS): https://fred.stlouisfed.org/data/PAYEMS

Freddie Mac, Primary Mortgage Market Survey, week of October 1, 2026: https://www.freddiemac.com/pmms

Disclosure: Jon Brooks is a co-founder of Momentum Realty, a Florida real estate brokerage that owns and funds Florida Housing Intelligence. Momentum has a financial interest in Florida real estate transactions, including in markets named here. See our editorial standards.