Mid-2026 Job Market Check-In: The First Half Was Weaker Than It Looked

A mid-year read on the US labor market.

I publish JobIntel to help job seekers screen listings before they apply: scoring credibility, detecting duplicates, and flagging ghost jobs. This read is built entirely on public data - the Bureau of Labor Statistics, JOLTS, and Federal Reserve research - written for job seekers as informed consumers and for the journalists and HR leaders who want a citable reference.

In June I wrote that the Q2 job market was stuck, and that being stuck looked completely different depending on which part of it you were standing in. Six months of 2026 data are now in, and I would revise that.

The market is not just stuck. The first half of it was reported to us as stronger than it actually was, and we did not find out until August.

103,000 jobs that were there in the spring and are not there now

On August 7 the BLS released the July Employment Situation. Payrolls came in at -23,000. That number got the coverage.

The revisions were the actual news. May was revised down from 129,000 to 63,000. June was revised down from 57,000 to 20,000. Between the two months, 103,000 jobs that were reported in the spring are not there anymore.

This is not a rounding adjustment. June lost roughly two-thirds of the job growth it was originally credited with. The spring you remember reading about - the one where the market was soft but still adding - was a draft.

Set against that, the trend line is unambiguous. Over the twelve months through July, monthly payroll growth averaged 34,000. That is the real pace of American job creation right now, and it is a fraction of what most people still carry in their heads as normal.

One precision note, because it matters: the BLS characterizes both the -23,000 payroll change and the 4.1% unemployment rate as "changed little," meaning neither moved enough to be statistically meaningful month over month. I am not telling you the economy shed jobs in July. I am telling you the first half was quietly marked down.

The yardstick moved at the same time

Here is the part almost nobody covered, and it is the reason a small payroll number in 2026 does not mean what it meant in 2024.

Breakeven employment growth is the number of jobs the economy has to add each month just to hold unemployment steady. For years the working rule of thumb was 100,000 to 150,000 a month. Below that, unemployment rises.

That number has collapsed. Research from the Federal Reserve Bank of St. Louis now puts 2026 breakeven at a range of 15,000 to 87,000 jobs per month, down from roughly 153,000 estimated in early 2025. The entire range is driven by immigration uncertainty - fewer people entering the labor force means fewer jobs are needed to absorb them. The author's own read is that the evidence points toward the lower end of that range, and that breakeven could fall below 15,000 entirely, which would mean small job gains are fully consistent with a healthy labor market.

So two things changed in the first half of 2026 at once. The reported numbers got revised down, and the benchmark for judging them got revised down further. We spent six months measuring a moving target with a shrinking ruler.

Apply the current benchmark honestly and you get the defensible read: July's -23,000 sits below even the bottom of the published range. It was under breakeven on every published scenario. Not catastrophic. Not fine.

Nobody is moving

The churn data tells the same story from a different angle, and it is the half of the picture job seekers feel most directly.

The June JOLTS release, out August 4, is remarkable for how little moved:

Job openings: 7.4 million, little changed. May was revised down 57,000.
Hires: 5.3 million, unchanged.
Quits: 3.2 million, a rate of 2.0%, unchanged.
Layoffs and discharges: 1.8 million, a rate of 1.1%, unchanged, and changed little in every single industry.

Four headline series, and the word attached to all four is some version of "unchanged." The quits rate has now sat at or below 2 percent for about a year, well under pre-pandemic norms and nowhere near the roughly 3 percent peak of early 2022.

Quits are the cleanest confidence signal in the whole dataset. People quit when they believe something better is available. At 2.0%, they do not believe that.

Indeed Hiring Lab described this market as moving from ongoing deterioration to a new steady state, and that is the right frame. Low hiring, low firing, and very little movement between jobs. It is not a market that is falling apart. It is a market that has stopped circulating.

Where the divergence actually is

Aggregate numbers hide the part that determines whether your search is hard or impossible.

Health care is still the engine, and it is downshifting. July added 22,000 health care jobs, continuing its trend - but the prior twelve-month average was 36,000. The one sector reliably carrying the market is carrying it less.

The financial activities sector is in genuine contraction. Down 14,000 in July, with credit intermediation off 9,000 and insurance carriers off 7,000. It is now down 121,000 from its 2025 peak, per the July release. That is not a soft month. That is a fifteen-month slide.

Retail lost 19,000, concentrated in warehouse clubs, supercenters, and general merchandise.

And two figures from the household survey that describe the experience rather than the economy: people on temporary layoff rose 153,000 to 921,000, and the long-term unemployed hold at 1.8 million, 25.5% of everyone out of work. One in four unemployed people has been looking for more than six months.

Wages, meanwhile, are doing nothing exciting. Average hourly earnings rose two cents in July to $37.62, up 3.2% over the year.

What a stalled market means for the second half

I want to be careful here, because the honest conclusion is not the motivating one.

If you are searching in a market with 34,000 average monthly job growth, a 2.0% quits rate, and one in four unemployed people past the six-month mark, then the difficulty you are experiencing is structural. It is not a signal about you. A lot of job search advice is built on the assumption that effort is the variable. In this market, effort is mostly the constant.

What actually changes your outcome is where you point the effort. Three things follow from the data:

Volume is the worst available strategy right now. When hires are flat and quits are stalled, most postings are not attached to urgent movement. Spraying applications into a market that is not circulating is how you end up with application fatigue and nothing to show for it.

Selectivity beats speed. In a low-churn market the scarce resource is a real, current, actively-hired role. Finding out which listings those are before you invest an evening in one is worth more than three extra applications. That is the entire reason I built the credibility scoring in JobIntel, and it is why spotting a ghost job is a more valuable skill this year than last.

Run it as a pipeline, not a lottery. When the market is slow, your own process is the only variable you control. Tracking your search like a pipeline tells you where you are actually losing - screening, first round, final - which is information no jobs report will ever give you.

The thing to take from mid-year

Every monthly number you use to decide whether to keep going is provisional. May and June both got marked down by a combined 103,000 jobs, months later, quietly, when nobody was covering it. The benchmark you would use to judge those numbers changed too.

You cannot verify the market in time for it to help you. It is revised on a schedule that has nothing to do with your search.

You can verify the listing in front of you: whether the role is real, whether it is current, whether anyone is actually hiring for it. None of that gets revised six weeks later.

The next Employment Situation lands September 4. The preliminary benchmark revision to the establishment survey lands August 28, which is very likely a second revision story this month. I will read both. You should run your search as though neither will change anything about what you do tomorrow, because they will not.

Try JobIntel free at jobintel.com. See credibility scores, skill matches, and salary data for every listing.