Temp Help Fell 10,900 in September: What It Signals for Contract Hiring

Last updated: October 18, 2026

By Jennifer Burdick, Recruiting Manager, KORE1

Temporary help services lost 10,900 jobs in September 2026, the second drop in three months, while total payrolls rose just 29,000 and unemployment reached 4.2% (BLS, preliminary). Temp help is still about 15,000 jobs above where it was a year earlier. Our view at KORE1 is that temp dips first and recovers first, which makes the fourth quarter the time to line up contract capacity for the first quarter of 2027.

The streak was already gone before September reported.

That is the part of the October 2 release most people skimmed past. When the August report came out on September 4, the Bureau of Labor Statistics had temp help up 5,200 in July and up 6,800 in August. The September report revised July to a loss of 10,000 and August to a gain of 200. So the picture changed in one morning from eight straight monthly gains to two losses in three months, with a flat month in between. I wrote a piece last month, our contingent workforce statistics breakdown, that described eight months running. It was accurate on the August numbers. It is not accurate now, and I would rather say so here than leave two versions of the same series on our site without comment.

I run recruiting and delivery on our contract desk, and the temp help line is the one number from the jobs report that I actually see move in my own week before it shows up in a release. Not perfectly. But closely enough that I read Table B-1 before I read the headline.

The call that made this concrete came on October 6, from a controller at a cold-storage food distributor in Bakersfield. She had approved two contract AP clerks for year-end close back in August, then pulled the requisition in mid-September when her CFO asked everyone to “hold anything that is not on fire.” By the first week of October the invoice backlog was 1,400 documents deep and the close calendar had not moved. She wanted the two clerks again, by the 20th, at the rate she had already approved. That is a small story. It is also exactly what the temp help series looks like from the inside: a pause that registers as a lost job in September and comes back as a hire a few weeks later.

One disclosure before the numbers. KORE1 earns its fee when a company brings on a contractor or a hire through us, so a post telling you to line up contract help is not a neutral document. I will keep the data and my opinion in separate paragraphs so you can weigh them separately. If you want the business side first, here is how contract staffing with KORE1 works.

What the September Report Actually Says About Temp Help

The Bureau of Labor Statistics series for temporary help services (CES6056132001, seasonally adjusted) fell from 2,497,700 jobs in August to 2,486,800 in September, a loss of 10,900, according to the September 2026 Employment Situation release. Both months are preliminary.

Here is the last fifteen months of the series, from the same release.

Read the table in halves. The back half of 2025 lost jobs in five of six months. Then January through June 2026 put on 56,100 in a row, from 2,451,400 to 2,507,500. Since June the series has given back 20,700 of that, a little over a third.

Thirty-seven percent of a six-month run, gone in a quarter. That is a real slowdown. It is not a collapse.

And the revision deserves its own small table, because it is the thing that changes how much weight you put on the September print.

The first column comes from the August report released September 4. Payroll revisions went the same direction as temp help. BLS says July and August combined are now 60,000 jobs lower than first reported. September will get the same treatment twice more, so the 10,900 figure is a first draft, not a verdict.

Still Up From Last September, by About 15,000

Year over year the series is positive. Temp help stood at 2,471,600 in September 2025 and 2,486,800 in September 2026, a gain of 15,200, or roughly 0.6%. Total nonfarm employment grew about 0.3% over the same twelve months, from 158,548,000 to 159,044,000. So temp help has outgrown payrolls as a whole across the year, even after the summer.

Temp help is still a small slice of the economy. About 1.6% of all payroll jobs. People overread it for that reason and underread it for the same reason, depending on the week.

What the year-over-year number tells a hiring manager is narrower than a forecast. It says the contract labor that companies added this spring has not been cut back to where it was last fall. Some of it has been. Most of it is still working.

Why Temp Help Moves Before the Rest of the Payroll

This section is our opinion, labeled as one, and drawn from what we see on our desk rather than from a study.

Devin Hornick, one of our partners, put it in one line when we were planning coverage of this report: “Temp is the first to dip and the first to come back.” The mechanics behind that are not mysterious. A contract assignment can end with a week of notice and a purchase order that is not renewed. Nobody has to run a reduction in force, call legal, or explain it to a board. So when a CFO says “hold,” contractors are the first line that actually goes down, and they show up in the September count as lost jobs.

The reverse is also true, and faster. A company that is not yet sure enough to open a permanent requisition will approve a contractor for one quarter, sometimes on the same afternoon. The Bakersfield controller did not need a new headcount approval to bring her two AP clerks back. She needed a purchase order she had already signed once. To make the same move for a whole project, see how to pitch a contract team as a project budget.

I am not going to claim temp help predicts recessions. It has dipped plenty of times when nothing bad followed, and the back half of last year is right there in the table. What I will say is that the series tells you about hesitation before payrolls do, because hesitation is cheapest to express through contractors.

What Else the Report, and the Fed, Said

The headline was soft. Total nonfarm payrolls rose 29,000 in September and the unemployment rate rose to 4.2% from 4.1%, both of which BLS described as having “changed little.” Professional and business services, the sector that contains temp help, also changed little in the release text.

Two weeks before that report, the Federal Open Market Committee raised the federal funds target range by a quarter point to 3.75% to 4.00% on September 16, on a 12 to 0 vote. The statement said job gains “have kept pace with the workforce.”

Then Governor Christopher Waller, speaking in Istanbul on October 8, said “the labor market continued to be solid and stable in September” and that “for at least the near term, policy will be focused on the inflation side of our mandate.” He added that further hikes “do not need to come at consecutive meetings.”

Put plainly, the central bank is looking at the same weak September and calling the labor market fine. That matters for your 2027 plan more than the temp number does, because it means borrowing costs are not coming down to rescue a fixed-headcount budget. We covered how the September rate hike changes 2027 headcount math in more detail. The short version is that a variable cost you can end in a week looks better at 4% than it did at 2%.

Book the Contract Seats You Will Need in January Now

Here is the practical consequence, and it comes from our own placement data rather than the report.

Across 1,997 contract placements in KORE1’s job-activity records from 2019 to 2026, the median time from accepted offer to first day was three days, and 86% started within a week. That speed is real, and it is also the reason companies wait too long. If a contractor can start in three days, why book one in October for work that starts in January?

Because the three days only begins once there is a candidate who said yes. Finding that person is the slow part, and it gets slower when everyone who paused in September un-pauses at once. The controller in Bakersfield was early. In a normal year the bulk of those calls arrive in the first two weeks of January, from teams that held through the holidays and then discovered the work had not held with them.

Devin’s note to our marketing team this month was specific about where to focus: “Focus… where companies are still investing and where they simply can’t afford to have projects fail or go unfinished.” On our desk this fall that has meant year-end close and audit support in accounting, ERP and data migration work with a cutover date already on a calendar, and security engineering that a contract or an auditor requires. Those projects do not stop because a jobs report was soft.

A few planning numbers from the same KORE1 records, which cover 2019 to 2026 and skew heavily toward California (79% of the rows):

• The median contract assignment that has ended ran 116 days (n=1,887). A contractor who starts the week of November 16 would, at the median, finish around mid-March. That covers year-end, the close, and most of the first quarter.
• Assignments in the middle half ran 44 to 227 days. A lot of “one quarter” work turns into two.
• Of 552 contract-to-hire placements, 207 converted to a direct hire, or 37.5%. When the permanent requisition does reopen next year, a third of the time the person is already sitting in the seat.

If you are building the budget now, our framework for contractor headcount planning for 2027 walks through which roles to carry as variable cost and which ones to convert. And if you already know you will need people in January, tell our contract desk the role, level, metro, start date, and length, and we will tell you honestly whether it makes sense to book it now or wait for the November report.

The Next Number Lands November 6

BLS has scheduled the October 2026 Employment Situation for Friday, November 6, 2026, at 8:30 a.m. Eastern. That release will carry a first estimate for October and second estimates for August and September.

Watch two things. Whether September’s 10,900 loss gets revised, in either direction, since July moved by 15,200 between reports. And whether October is a third loss or the first gain. A third loss in four months would be the first stretch of steady decline since late 2025. A gain would look like what Devin described.

We will update this post when the revisions come in. For the longer-range picture that this monthly print sits inside, our contract labor market outlook covers supply, rules, and rates for the full year.

Asked on the Monday After the Release

So did temp help really fall, or is this just a revision?

Both, in a sense: September is a new preliminary loss of 10,900, and the same release revised July down to a loss of 10,000 and August to a gain of 200.

The revisions matter as much as the new month. Before October 2, the published series showed eight straight gains. After it, the series shows two losses in three months. Expect September itself to move again on November 6 and once more in December.

Does a drop in temp help jobs mean a recession is coming?

Not on its own, and temp help fell in five of six months in late 2025 without one.

Our read is that it signals hesitation, not contraction. Companies cut contractors first because it is the cheapest way to say “wait.” Whether the wait turns into something worse depends on payrolls, layoffs, and the Fed, and none of those are flashing red in this report.

Up or down from a year ago?

Up, by about 15,200 jobs, from 2,471,600 in September 2025 to 2,486,800 in September 2026.

Should we wait for the October report before booking contractors?

If the work has a deadline in the first quarter, no; waiting five weeks for one data point costs more than the information is worth.

If the work is genuinely optional, waiting is reasonable. The November 6 release will tell you whether September was a blip. Most of the projects I see in October are not optional, though. They are year-end close, a system cutover with a date, or an audit finding, and none of those care what temp help did.

Realistically, how fast can a contractor start?

Three days from accepted offer to first day at the median, across 1,997 KORE1 contract placements from 2019 to 2026.

Eighty-six percent started within a week. The search before the offer is where the time goes, and that part stretches in January when demand returns all at once.

Read the Revision Before You Read the Headline

A soft September is a reason to plan, not a reason to freeze. The temp help series has already shown you what this autumn looks like: companies pausing contractors, payrolls barely moving, and a Federal Reserve that sees no emergency. The companies that come out of the first quarter in good shape tend to be the ones that decided in November which seats they would need, instead of discovering it in January.

We will be back with the October numbers after November 6.