This website uses cookies

Read our Privacy policy and Terms of use for more information.

MARKET COMENTARY

The jobs report beat expectations by 109,000. If you're job hunting, you probably didn't feel it.

By Ric @ Jobric · September 2026

Friday morning's jobs report was the biggest upside surprise of the year. Employers added 162,000 jobs in August. Wall Street had penciled in 53,000. The unemployment rate held at 4.1%. Even the revisions, which have been quietly brutal all spring, came back positive for once.

And if you have been sending applications since June, almost none of that describes your week.

That is not a contradiction and it is not spin. The report is a collection of separate measurements that can move in opposite directions at the same time. This month several of them did. Here is what actually happened underneath the number everyone quoted.

The internals were genuinely good this time

Last month the unemployment rate looked stable for an unflattering reason: people gave up and left the labor force, which shrinks the denominator and flatters the rate. August was the opposite. The labor force grew by roughly 683,000 people, labor force participation ticked up to 61.6% from 61.4%, the employment-population ratio rose to 59.1%, and the unemployment rate still held at 4.1%.

More people came back looking and the rate did not deteriorate. That is the healthy version of a flat number, and it is worth saying plainly because it has not been true for most of this year.

The revisions helped too. June was revised up 11,000, from +20,000 to +31,000. July was revised up 44,000, from a loss of 23,000 to a gain of 21,000, which means July was not a job-losing month after all. Combined, that is 55,000 jobs the initial prints had missed, in the direction nobody has gotten used to.

Two categories carried 62% of the month

Now the part that gets left out of the headline.

Food services and drinking places added 59,000 jobs. Local government education added 42,000, which is the annual back-to-school hiring bounce showing up in the data. Together that is 101,000 of the 162,000 total, or about 62% of the month.

Strip those two out and the rest of the American economy added roughly 61,000 jobs. That is not a collapse, but it is a very different picture than "the labor market is running hot." Restaurants staffed up and schools reopened.

Elsewhere it was mixed. Construction added 22,000. Manufacturing added 16,000, which is a rare positive month for a sector that has been flat to negative. Health care added 13,000, still positive but slower than its recent pace. Retail was essentially unchanged, up 1,400.

Information lost 23,000 jobs, the largest decline of any sector, and it is down from 2.86 million a year ago to 2.75 million now. If your search is in tech, media, telecom, or publishing, that trend line is the one that matches your inbox. Federal government shed another 5,000 and is down about 242,000 from a year ago.

So government was the story again this month, just inverted. Local hiring pushed the total positive while federal kept shrinking.

Nothing improved for the people already stuck

Here is the number I would put on the front page if I ran the wire services.

The share of unemployed people who have been out of work 27 weeks or longer rose to 27.0% in August, up from 25.5% in July and 25.6% a year ago. In headcount, that is 1.93 million people.

Average unemployment duration went up too. Mean duration reached 26.3 weeks, from 24.9 in July and 24.5 a year ago. Median duration reached 11.4 weeks, from 10.5 in July and 9.9 a year ago.

Read those two paragraphs together with the headline. Employers added the most jobs since the spring, and the people who were already searching got measurably worse off. Both things happened in the same month. Hiring picked up at the front door while the line at the back of the queue got longer.

That is why "unemployment is only 4.1%" and "I have applied to 200 roles and heard nothing" are both accurate statements. The rate counts how many people are looking. Duration counts how long it takes. Only one of those is about your experience.

Your paycheck is still losing to prices, barely

Average hourly earnings hit $37.75 in August, up 10 cents from July and up 3.1% from a year ago.

The most recent inflation print available is July, since August CPI does not land until mid-September. All-items CPI rose 3.4% year over year in July. Nominal pay up 3.1% against prices up 3.4% means real hourly earnings were slightly negative over the year, on the order of negative 0.3%.

It is a small gap, but it is the wrong side of zero, and it has been for a while. Wage growth that reads fine in a press release still translates to a paycheck that buys marginally less than last August.

One more reason not to over-read this

ADP, which runs its own private-sector payroll count from actual payroll records, reported just 38,000 private jobs for August, its weakest reading since January. The BLS establishment survey implies 127,000 private jobs for the same month.

That is a wide disagreement between two serious measurements of the same thing. It does not mean either is wrong, and the two surveys use different methods and populations, so they diverge regularly. But given that this same report just revised the last two months by 55,000 combined, a single strong print deserves less confidence than the coverage will give it.

Markets read it as strong, and read strong as a problem. Because the Fed has been leaning toward a hike, futures moved to roughly a 58% chance of one at the September 15 and 16 meeting. The S&P 500 closed down 0.38%, the Nasdaq down 0.29%, the Dow down 0.51%, and the 2-year Treasury yield reached its highest level since January 2025. Nobody sold off because the data was bad. They sold off because it was good.

What to actually do with this

Three practical reads.

Watch duration, not the rate. The rate tells you how crowded the market is. Duration tells you how long the process takes. Median 11.4 weeks and mean 26.3 weeks means a search that runs three months is normal right now, not evidence that you are doing something wrong. Plan runway and morale accordingly.

Check your sector against the headline. A 162,000 print built on restaurants and schools says almost nothing about whether information or federal-adjacent roles are hiring. They are not. Your sector's line item is the number that matters, and it is public.

Widen the fit, not the volume. When openings concentrate, sending more applications into the same narrow slice mostly generates more silence. The better move is finding the roles your experience already qualifies you for that you would not have searched by title. That is the problem Jobric was built to solve.

Next Employment Situation release is Friday, October 2, covering September.

Sources

  • BLS Employment Situation, August 2026 (released September 4, 2026): news release. Payrolls, unemployment rate, revisions, sector detail, average hourly earnings, participation, and employment-population ratio all confirmed on the release page and cross-checked against the BLS Public Data API (series CES0000000001, LNS14000000, CES0500000003, LNS11300000, LNS12300000, LNS11000000, LNS13008275, LNS13008276, LNS13008636, LNS13025703, LNS13000000, and sector series).

  • BLS Consumer Price Index, July 2026 (released August 12, 2026): news release. All-items CPI up 3.4% year over year before seasonal adjustment.

  • ADP National Employment Report, August 2026: ADP release, private payrolls up 38,000.

  • Consensus expectations (Dow Jones 53,000, FactSet 65,000): CNBC, FactSet.

  • Market reaction, September 4, 2026: TheStreet, Yahoo Finance.