Unemployment claims are even with a year ago and are below two years ago.
By Wolf Richter for WOLF STREET.
Every data point about the labor market is getting twisted into “the labor market is cracking” or whatever by the cut-rates-now crowd. Today, we got another the-labor-market-is-cracking data point, that is depicting a decent labor market:
Initial applications for unemployment benefits (“initial claims”) in the week through Saturday were at the same level where they’d been a year ago at the end of August 2024, and below where they’d been at the end of August 2023 as per the seasonally adjusted four-week average released by the Department of Labor today.
At 231,000, the seasonally adjusted four-week average was up a little from early August, but down from the recent peak in mid-June (red).
And in a historic context going back to the 1970s, initial claims are low, despite the substantial growth of nonfarm payrolls over the decades.
They were lower only during the tight labor market of 2018 and 2019, and during the period of the labor shortages coming out of the pandemic — and also for a moment in 1973.
The unemployment claims data here are based on actual unemployment insurance applications filed at state unemployment agencies by workers who’d lost their jobs. This is administrative data, not survey-based data. The state agencies then submit this data weekly to the US Department of Labor, which combines it and releases it. If states don’t make the weekly cutoff for the submission of the data, that data goes into the next week, and if a big state does that, such as California or Texas, then claims fall during that week but jump the next week, which is why the Department of Labor also releases four-week averages of the data, which iron out those meaningless week-to-week squiggles.
The people who filed for unemployment insurance at least a week earlier and are still claiming unemployment insurance because they still haven’t found a job form the “continued unemployment claims” or “insured unemployment.”
Continued unemployment claims dipped in the latest week, and the four-week average also dipped, and at 1.947 million has been roughly level for the past two months.
The level of continued unemployment claims is up from 2024 and 2023, but is still historically low, lower than any time in the decades before 2018 – more in a moment:
Over the past four decades, it’s only during the tight labor market in 2018 and 2019 and in the years since covid, particularly the period of the labor shortages in 2021 and 2022, that the level was lower – despite the much larger nonfarm payrolls.
So people remain on the unemployment insurance rolls a little longer than in 2022-2024 and in 2018-2019, but not nearly as long as they did in the decades before 2018.
These unemployment claims tell us that companies are laying people off at a historically low rate, but the laid-off workers spend more time on the unemployment rolls looking for a job than they did in 2022-2024, and yet, they spend a lot less time on the unemployment rolls than before 2018.
So this is no longer the labor market of 2022 and 2023 where employers were struggling with labor shortages. But these unemployment claims depict an overall decent labor market in the historical context over the past four decades.
Enjoy reading WOLF STREET and want to support it? You can donate. I appreciate it immensely. Click on the mug to find out how:
My favorite trend when it comes to employment:
https://fred.stlouisfed.org/series/LNS12300060
Employment is full. Able bodied working age people are working at a historic clip. There are still some hold out of people who prefer single working parenthood, but there isn’t some magical labor force to dip into.
Which is why we’ll continue to see services inflation rise.
The AI Datacenter and electrical grid infrastructure is going to be nearly impossible to staff.
I think the swinging door will be any reshoring of manufacturing. We don’t have the workforce to build or operate it (or we’ll need people to rotate out of service sectors, leading to even more inflation in those wages). Traditional industries can’t get equity or debt to build facilities that can’t compete on cost, 15-30% tariffs be damned.
So Tariffs will either come from ballooned profits or consumers pocketbooks.
1. this was about unemployment claims (administrative data not survey data), but you knee-jerk-linked data from the month-ago household survey of the jobs report = survey data (the new set will be released tomorrow). So your link has nothing to do with this date here.
2. The employment to population ratio, which you linked, is heavily skewed by the huge wave of boomer retirements: they’re still in the population (the lucky ones) but if they’re retired, they’re no longer working. So your link is meaningless unless you understand how it’s figured.
3. A much better chart is the “Prime age labor force participation rate” which tracks 24-54 year olds and is therefore not skewed by retirements. It is very high, showing a strong labor market:
Wolf I was agreeing with you!
I published the 18-54y/o employment ratio, which is the ratio of working 18-54y/o to total population of 18-54y/o (LNS12300060)
This is basically the same as the the 25-54y/o Labor Force Participation Rate (LNS11300060).
If you overlay the charts on FRED you’ll see they are basically the same. The gap is LNS12300060 accounts for actual current employment. LNS11300060 includes all 18-54y/os that want to be in the workforce, whether they have current employment or now.
“I published the 18-54y/o employment ratio,”
Good lordy, I missed the age bracket. Thanks
Hi Wolf,
Thanks for an excellent update on unemployment. I’m old enough to remember when a rate in the mid 4% range would have been considered a dream.
One data related question piqued by your second chart: initial claims were substantially higher from 1973/4 until about 2017 – a period of 40+ years (and unadjusted for overall employment growth) so the period from 2017 to present – excepting the COVID period – was even lower in proportional terms.
Is it a possibility that the lower initial claims are in anyway driven by the emergence of the so-called “gig” economy? My understanding is that independent contractors wouldn’t be covered by unemployment insurance and so would be ineligible to submit claims. Could this potentially drive down the number of claimants?
Not trying to assign work – a genuinely curious question.
“Is it a possibility that the lower initial claims are in anyway driven by the emergence of the so-called “gig” economy?”
NO, because nonfarm payrolls — so regular employees not gig workers — have grown substantially over the years and decades, from 138 million 10 years ago, to 160 million now.
But the number of self-employed (“gig workers”) has remained relatively stable since covid at 9-9.5 million. This is from the jobs report’s household survey, which is very volatile.
People make up all kinds of stuff about gig workers and the gig-worker economy, but it’s just BS:
I can’t wait for all hell to break lose when continuing claims reach 2M..
Don’t confuse me with the facts…prices aren’t too high, rates are
Yeah but sadly it doesn’t matter; in a way that crowd has won, it’s pretty 100% rate will be cut this month; perhaps they will scream louder it needs to 150 basis points or whatever…either way they got what they want
Then the 10 year (I believe the 10 year is the one that will rise) will shoot up to over 5% hopefully, maybe, should happen,Come on vigilantes!!! Rooting for you!!! And then housing rates go up!!!! Bahahhahahha
The labor market is indicative of where employers feel the economy is headed; they won’t hire unless they have a good feeling in their bellies about the money flow around them. The figures suggest that we are going to continue with a decent economy for some time to come; it’s only a matter of time before a recession hits, perhaps sometime in 2027.
There is no job growth, though.
BLS, ADP, Challenger; all reporting the same thing. Job gains at or near zero. The Fed Beige book also corroborates that data.
Should the Fed cut? Who knows. Rate cuts won’t save the economy either way, so I’d rather they focus on keeping inflation subdued.
On a personal anecdote, I review permit applications in a local jurisdiction in King County, Washington. Development applications this year have been as slow as I have ever seen it. Large projects have not been coming in since 2023, which makes sense given higher interest rates, but it is especially slow this year.
I’d be worried about the labor market, personally.
“BLS, ADP, Challenger; all reporting the same thing. Job gains at or near zero.”
Bullshit.
1. ADP +54,000 in August, and July revised UP to +106,000
2. Challenger doesn’t report job gains or losses, but tracks media-reported global layoff announcements, which is useless because announced layoffs in India don’t have an impact in the US.
3. BLS will report tomorrow for August. For July: +73,000
Hey Wolf, so what are your thoughts on the 6 month yield pricing in rate cuts compared to the data in this article? Are fed rate cuts unsubstantiated at the moment?
The 6-month yield as of today hasn’t changed much from the chart in the article I wrote about it on Friday. Same text from Friday still applies:
“It is now pricing in nearly two rate cuts in its window of about 2-3 months, so one cut at the September 17 meeting, and it’s getting closer to pricing in a cut at the October meeting.”
https://wolfstreet.com/2025/08/30/30-year-treasury-yield-stuck-near-5-on-inflation-fears-but-6-month-yield-drops-sees-rate-cuts-long-end-of-yield-curve-steepens-a-lot-mortgage-rates-still-over-6-5/
Personally, I don’t think the Fed should cut, but it will cut no matter what, given the enormous pressure it is under to cut.
But if inflation data come in even hotter over the next few months, and the labor markets continues to be OK, they have enough ammo for more wait-and-see.
Job growth is small but not zero for the last 3 months (33k/month). Change in 2025 so far is 85k/month. Typical over the last 2 years is 150k/month. Whether it’s a new trend or a blip will depend on the next few months of data. Times when job growth shows but doesn’t go negative exist, but are relatively rare.
Time will tell!
Seems like if you have a decent amount of people retiring, combined with about the same amount of entrants, then you don’t need job growth if there is low unemployment. The larger challenge, outside the scope of this article is skills alignment in a fairly rapidly changing work landscape. The US is trying to pivot at current but really remains to be seen if we have the right skill sets for what is needed, although of course no reason people can’t be trained.
So many gainfully employed and the stock markets at all time highs daily bumping higher even still, exactly why are so many clamoring about interest rate cuts again as my last visit to a fast food restaurant assured me that I won’t be back because somehow with all this Goldilocks affordability remains quietly problematic.
It’s almost like we live in an altered state of reality to what’s really going on. Bizarre times to say the least. Housing markets unaffordable, food prices getting there, and yet all this great news daily. I’m sorry, I’m just not paying11$ for 3 tacos, and 30$ for a haircut.
Because the rich want to get even richer. It’s that simple.
Not sure about you, but where I am, minimum wage is through the roof.
Can’t have higher pay for everybody without having higher expenses…
Usually the same people to claim her for less inflation of the ones clamoring for higher minimum wages, there seems to be a disconnect that people just don’t recognize …
Because the sellers of overpriced assets are starting to seriously sweat and the more they sweat, the more pressure they place on their elected representatives.
It’s a reality bubble for both the elites and the masses…like believing professional sports are competitive outcomes when the actor players get paid, win or lose …
No mention of your favorite recession indicator? It’s a fascinating observation that I’ve passed it along to all my chicken little clients (with a nod to the source, naturally).
For the unaware see the following:
https://wolfstreet.com/2025/05/08/recession-watch-time-to-dig-out-our-favorite-recession-indicator-again/
No, because we’re so far away it’s kind of silly to even bring it up.
Unemployment is a lagging economic indicator and employers will not lay off their workers until they see no other way to maintain their profit margins. Recent college grads have a hard time finding a job, and they can’t apply for unemployment benefits. This week’s JOLTS report showed job turnover is lower than before the pandemic. Young people can’t afford to buy a home and discount stores are seeing an increase in middle class customers. This all points to a weakening economy, or at least an economy that is looking for a direction. Because of their dual mandate, the Fed will have to lower rates even if inflation will pick up again.
Only your imagination is not a “lagging” indictor.
JOLTS was fine. The T in the name is “Turnover,” meaning churn. The data describes CHURN in the labor market. All it said is that that the churn has calmed down: low quits, low layoffs and discharges, and therefore fewer slots left behind that are newly open, and therefore fewer openings to fill. And yet, job openings to fill, at nearly 7.5 million, were at the record high before the pandemic. So the massive and expensive churn from the pandemic is over, that’s all it said:
Einstein said once the rich control the media, you won’t know fact from fiction…the most hated man in the room is the one who tells the Truth, Plato…so as the empire goes doing what empires always do…did you know the bank of England during the Napoleonic wars suspended convertibility and BOE balance sheet went up 130 percent…
When one private equity firm dumps a company to another PE firm routinely besides all the other shenikens we have witnessed the past 25 years…the game has never been more openly rigged…you can buy a trump meme coin at 8 dollars down from 64…what a bargain for you…
Great article. Man I didn’t know jobless claims were so high before the 2000s.