BridgeWorks
The Bureau of Labor Statistics released the September employment report this morning. Nonfarm payrolls rose 29,000 — well below forecasts — and the unemployment rate moved up to 4.2 percent from 4.1. Labor force participation ticked up to 61.8 percent.
The revisions underneath that are the part worth your time, and they put us in an awkward position we are going to describe plainly.
We corrected ourselves. The data moved again.
Follow the sequence:
- August 7 — we reported that July payrolls had fallen by 23,000 and called it a contraction.
- September 4 — BLS revised July up to +21,000. We published a correction saying the contraction we reported had not happened, and added update notes to two earlier posts that had leaned on the bad figure.
- Today — BLS revised July back down to −10,000. August was also cut, from +162,000 to +133,000.
So July was a contraction after all, by a smaller margin than we first said. Our original reporting was closer to right than the correction we published four weeks later.
We are not going to pretend that is tidy. The honest reading is that our first post was too confident, our correction was too confident, and the underlying problem was treating any single month's first print as a fact. That is the lesson we wrote into our own process after the September correction, and it has now been demonstrated twice at our expense.
What we would say to a reader: when you see a monthly jobs number, treat it as a noisy estimate from a survey that gets revised twice by design. The trend over several months is the signal. We will hold ourselves to that.
What the trend actually shows
Set the individual months aside and look at the run. May through September, as currently estimated, is a labor market adding jobs in the tens of thousands rather than the hundreds of thousands — well below what it takes to absorb people entering or re-entering the workforce. Unemployment has drifted from 4.1 to 4.2 percent. That is not a collapse. It is a slow, persistent softening, and it is consistent with what our participants have been telling us all year.
Where September's gains landed:
- Health care: +17,000
- Construction: +11,000
- Manufacturing: +9,000
- Financial activities: −7,000
BLS described employment in all major industries as changing little over the month, which is itself the story. Nothing is driving growth.
The wage line
Average hourly earnings rose 0.1 percent in September and 3.0 percent over the year — down from 3.1 percent last month, against consumer prices running around 3.4 percent.
That gap has been open all year and it is the thing most likely to describe a household's experience: employed, and still losing a little ground each month. It is also why gasoline at 27 percent year over year lands as hard as it does.
What it means for the Fed
In mid-September the Fed raised rates for the first time since 2023, with a statement observing that "job gains have kept pace with the workforce." Today's report shows 29,000 jobs, a higher unemployment rate, and 60,000 fewer jobs in July and August than previously believed.
We are not going to predict what the committee does on October 27 and 28. We will say that the labor-market half of its mandate just got materially weaker, and the data it cited in September has partly been revised away. Anyone who told you the direction of rates was settled was overreading.
For a job search
The practical guidance does not move with one report:
- Target the sectors still adding. Health care, construction, and manufacturing were the only meaningful gainers. Our construction program sits in one of them.
- Plan for a longer search. A market adding under 50,000 jobs a month absorbs newcomers slowly.
- Do not time anything to a single print. We have now been wrong twice in that direction, publicly, and we would rather you learn it from our mistake than your own.
