Recession talk is swirling densely all over the place, so let’s have a look.
By Wolf Richter for WOLF STREET.
What are we looking for? The National Bureau of Economic Research (NBER), which calls the official recessions in the US, has always defined recessions as broad economic downturns that include downturns in the labor market, such as declines in employment and significantly rising unemployment.
Weekly data for unemployment insurance benefits are the earliest indicators of systemic job losses. Among the data sets, weekly continued claims for unemployment benefits, also called “insured unemployment,” is our favorite Recession Indicator. It essentially counts the number of people receiving unemployment benefits after their initial claim.
The prior three business cycle recessions – not counting the Pandemic which was a lockdown, not a business cycle recession – came after Insured Unemployment had surged to:
2.64 million in December 2008, beg. of Great Recession
2.56 million in March 2001, beg. of 2001 Recession
2.49 million in July 1990, beg. of 1990 Recession.
The levels that entail a recession have risen as total employment has risen. This growth of employment over the years causes the Recession Indicator line to be slanted upward (black slanted line in the chart below).
Today, about 2.7 million insured unemployment would indicate the beginning of a recession.
But insured unemployment dipped by 29,000 in the latest week, to 1.88 million, according to the Labor Department today. The four-week average edged up by 8,750 to 1.87 million, essentially unchanged since October. These levels are still historically low, and far below the Recession Indicator of 2.7 million. The purple columns indicate recessions.
Watch for 1. surge to 2. high levels. This takes two factors: a surge, and high levels. Note how insured unemployment began to surge before a recession. It’s this kind of surge to much higher levels that we need to watch out for.
There was a surge in 2023, but from record low levels during the labor shortages, and since the level it surged to was still so low, that surge did not entail a recession; it just “rebalanced” the labor market.
This measure is a combination of two factors: Essentially, how many people got laid off, and how long it takes them to find a new job. If it takes people longer to find a job, even if there are few new people getting laid off, then that measure slowly rises, which was the case in 2024. But it has been roughly unchanged for the past six months.
It surges when there are suddenly lots of layoffs that the labor market cannot absorb, and the number of people receiving unemployment benefit rises sharply to much higher levels. That’s the recession indicator.
Layoffs are still historically low. Weekly initial claims for unemployment insurance benefits – people initially filing for unemployment insurance benefits after they got laid off – dipped by 13,000 in the latest week, to 228,000 seasonally adjusted, according to the Department of Labor today.
The four-week average, which irons out the week-to-week squiggles, edged up 1,000 to 227,000 and is historically low. The purple columns indicate recessions.
Note how those claims begin to surge before a business cycle recession (the 2020 recession wasn’t a business cycle recession but a lockdown).
The fact that insured unemployment (first chart) has risen from the labor-shortage lows in 2022, despite the low number of new layoffs (initial claims, second chart) indicates that it now takes more time to find a new job than it did back in 2022. And the low number of initial claims shows that few people are getting laid off. So companies are not cutting staff, but they’re slower in hiring than they were in 2022, though historically they’re still absorbing people at a fast clip.
These are signs of a reasonably “balanced” labor market, where supply of labor and demand for labor are roughly balanced, and that the labor shortages and the huge churn in the labor force in 2021 and 2022 are gone.
Negative GDP growth for other reasons. When a historic spike in imports, which are subtracted from GDP, pushes GDP growth into the negative, it’s not an indicator of a recession. A spike in imports is not a sign of weak demand, on the contrary. In Q1 2025, this situation happened, same as in Q1 2022, but consumer spending growth was OK and business investment was very strong, and so this was not a recessionary indication. But GDP is released quarterly and way behind.
Leading indicator. This data on unemployment insurance benefits is weekly and more immediate than monthly or quarterly data, and its surge to a high level warns of a recession in advance, which makes the Recession Indicator a leading indicator.
Enjoy reading WOLF STREET and want to support it? You can donate. I appreciate it immensely. Click on the mug to find out how:
Why let facts get in the way.
Because, just like at work, nobody likes facts. They get in the way of the truth.
I think the data hasn’t picked up the true impact of what is incoming yet.
Can’t speak for everyone, but our sales in the truck equipment world have fallen off a cliff since all the tariff talk began. I still believe tariffs have their place, but this implementation has been shockingly, stunningly bad. We are hearing the same sales story from our suppliers and customers. A contact at large dealer summed it up in a recent chat: “The used market is decent, and people are still picking through tariff-free dealer stock. New chassis orders are near zero.”
Unit orders here vs 5-year average: Jan: -1.6%, Feb: -31.0%, Mar: -40.7%, Apr: -49.8%. May is looking even worse so far.
We are about halfway through our cost-control list. All the painless cuts have been done. The next steps get ugly.
Despite killing all overtime and pausing hiring, production is exceeding monthly orders by about 25%, so the backlog is falling like a rock. Stock product is piling up to the ceiling. At this rate, we’ll be issuing pink slips to half the shop in June.
Sounds ominous.
There are many stories of surging orders at smaller manufacturers, they’re trying to staff up, they’re running more shifts, they’re trying to figure out how to produce more… even the anti-tariff WSJ has published some of those stories.
I for one only rely on anecdotal evidence
I’d love to see how employment rates changed along these timelines. Thanks Wolf!
?? – don’t imports add to either inventory or consumption – isn’t it a wash
Andy Apiecionek
No it’s not a wash, for the umpteenth time, LOL:
To simplify, we’ll exclude transportation and warehousing activity:
If you buy a $20 import, you sent $20 overseas, and the $20 circulates overseas and is gone from US, pays wages overseas and investments overseas, while US savings is reduced by $20 = future US spending is reduced by $20.
If you buy a $20 domestic product: GDP = +20 consumer spending -$0 imports = +$20 added to GDP.
If you buy a $20 domestic product, the $20 stays in the US and circulates in the US, pays US wages, investments in new equipment and factories, etc., and further adds to GDP as it circulates.
If you buy a $20 import: GDP = +20 consumer spending -$20 imports = $0 added to GDP. And the $20 is gone.
That’s the choice. That’s why large amounts of imports instead of domestic production are so devastating to the economy and to GDP. I don’t know why this “it’s a wash” ignorant BS keeps circulating.
I agree with you. I work in the solar and battery storage industry. We were busy as hell the last three years and we are still finishing jobs out from 2024. These jobs are starting to wind down and I’m not seeing many new jobs or bidding opportunities come through. Solar and BESS is notoriously up and down, so it’s unclear whether this is the beginning of a long-term trend, but if so, I agree that the pain has not really started yet. Give it three to six more months of no work coming in and people are going to start being laid off.
Yes some industries like Wind installs and solar installs will fall dramatically as subsidies and mandates are withdrawn. Moves maybe needed to where the jobs will be located but those jobs are available. Oil price is down which is huge for the consumer just need USA made stuff to buy
You were busy as hell because it was being subsidized by the last administration
Truckers that were equipped to meet the recent surge (front-running of the tariffs) might not need much new equipment for a while. How long depends on what the new equilibrium level is for imports.
On the other hand, if domestic manufacturing picks up, then more components and subassemblies might need to be moved around the country. That could increase demand for trucking, rail and barge transport, as well as intracoastal shipping?
I heard commentary a few years ago that employers are less willing to officially layoff employees because of the Covid era employment shortages, especially amongst lower pay workers. Any indications that hours worked are going down while headline unemployment stays incredibly low?
The employment situation summary that came out last week also reported steady hours worked.
Prices will not go up according to this site. Tariffs will only affect corporate profits. Not sure why even Trump says dolls and toys will be more expensive. Contradictory narratives.
You’re trying to stick BS into my mouth. That’s one of the seven deadly sins of commenting here.
What I have always said: Tariffs are NOT a sales tax, but are tax on corporate profits, and whether or not companies can pass them on without crushing their sales depends on market conditions.
What I have always said: New vehicle prices are very high after the huge price spikes during the pandemic, and profit margins of automakers and dealers are very high, and they’re maxed out, prices have hit a ceiling, and automakers have had to CUT PRICES (incentives) to sell what they have. That’s an example of “market conditions” that make it hard/impossible to pass on tariffs without crushing your sales.
I’ve been thinking about this (always dangerous). For that to be true implies that your profit margin is large enough to absorb the tariffs. Companies that have China exposure have to be rocking a 70% margin to absorb the current tariff as a tax on profits.
There are very, very few companies that operate at that margin. I know of a few pharma companies, but that’s it.
1. No profit margins don’t “have to be” large enough to absorb anything. Companies go out of business no problem if they lose money selling stuff, which happens all the time.
2. Profits are gigantic, after ballooning because of huge price increases since 2020:
So if you’re a US based manufacturer with inputs from China, your choices are 1) raise prices or 2) go out of business (or both)? I’m not talking about resellers sourcing cheap junk from Guangzhou, but there are plenty of bona fide small US manufacturers in that position.
Look at those corporate profits go thru the roof. No doubt they will try the tariff excuse to jack up even more where possible. Americans will keep buying, they no longer care about bargains. We’re all conditioned now for higher prices. Spend like the government.
‘Americans will keep buying, they no longer care about bargains. We’re all conditioned now for higher prices.”
That’s is exactly what is no longer true. Which is why durable goods prices prices dropped sharply, including used cars, since the end of 2022. The free money is gone. Americans are once again resisting price increases.
IMHO, it’s more likely than not that Trump will sign sufficient new trade deals with the important countries like MX, CA & CH before any ugly spike in unemployment happens. CA will come first, then MX & finally CH before the end of July. The Fed is boxed in. It makes way more sense for them to do everything they can to maintain a wait & see approach as long as possible before they move in either direction, with to the downside, of course, being the most likely scenario.
Excellent review of what it’s going to take in terms of unemployment claims to find the next recession. At least for now, there appears to be ample economic energy to overcome the downside of tariffs at least through June in my opinion.
I’d like to know all 7 deadly sins to see if I can weave them into a single comment. Solid creative writing assignment.
Definitely ask a question that was answered in the first paragraph. Bring in BS. Leave a link without even a synopsis. Quote ZH as if it’s true. Put BS in Wolf’s mouth. Say the same stupid thing every time you comment. Ask, angrily, why your comments are getting deleted.
I think that’s the seven.
Recessions are a normal component in the cycle of a healthy capitalist economy. Think of them as a time to pause and regroup after years of frantic effort.
Nice work, Wolf.
Is the “feeling” different than the numbers? If so, is that because: Cost of basics went up more than wages. Looming layoffs in NGOs. Slow down within self employed sector and fewer hours per week for employed people. ‘Quiet leaving’ needs to be added to unemployed numbers. Trump syndrome. Tarrif uncertainty. My student lian is due. Other social factors.
Maybe other measures are also needed: Happiness index. Ratio of total employed on total population.
We’ll all be about as happy as the average Russian soon.
Free money during the pandemic allowed people to leave poor fit jobs and find better fit jobs.
That’s usually what happens during a recession, except people are forced to leave jobs and find better fit jobs. (The pandemic was a massive recession for service workers, and there were stories about people dramatically improving their employment situation by leaving the service industry for other jobs during lockdowns.)
It’s worth wondering whether the Great Resignation of the pandemic served the purpose of a recession by reallocating labor to better fit uses.
If so, that could explain how we ended up with historically low numbers of people drawing unemployment.
Good point chicks:
Went through exactly that during my decades in the work force, both hourly wage slave and contractor, including contracting manual and white collar labor…
Worked from 1956 to 2019, so pretty much through all the recessions of that era.
Starting with minimum wage, if even that, ending with what is now being called a living wage in SF area.
Yep agree with Wolf at present I don’t see any sign of recession with economic activity ongoing . Our Trane factory here in East Texas completed a major renovation to increase production of AC and Heat Pumps . Their parking lot and train unloading stays packed .
Nothing can stop a trane. Unless the train carrying the trane breaks down. Then the trane is stopped on the train carrying it.
:,(
I have not been trained to read these comments. To the untraned eye, it’s confusing.
The Pope put is in
I appreciate reading your article and found it highly engaging. However, one crucial aspect you haven’t addressed is the gig workforce. The gig economy in the U.S. is rapidly expanding, with over 63 million gig workers in 2025, accounting for 38% of the total workforce. This number is projected to reach nearly 100 million by 2027. Gig workers play a significant role in the economy, contributing over $1.45 trillion to the U.S. GDP. Many of them juggle multiple jobs, with 56% reporting they hold two or more gigs.
Despite this growth, gig workers, freelancers, and independent contractors are typically excluded from traditional unemployment benefits, as they do not contribute to state unemployment insurance funds. Given the increasing demand for gig work, the omission of this critical issue weakens the overall conclusion of your article.
you’re completely wrong because you’re double-counting gig workers.
Data: Of the 163 million total employment, including gig work, 152 million are employed with regular wages/salaries. And only 9 million are exclusively self-employed (gig work, no salary). But of the 152 million with regular wages/salaries, many also have side jobs (consulting, coding, driving, construction work on weekends, etc.) that count as gig work. These are the multiple jobholders. (BLS jobs report last Friday, household survey data).
Business owners who earn a regular salary from their business, such as me, and are included in the 152 million, are not eligible for unemployment compensation though we pay into the system. The idea is that we cannot lay off ourselves. So if our business can no longer pay us and lays us off, we’re just SOL, and there are millions of us, but that has always been the case, nothing new.
There are some other categories in that 152 million that are not eligible for unemployment insurance. But that’s not new either.
So you may be a teacher and drive for Uber in the evenings and weekends. Or a Google engineer working on some projects on the side (gig work). You may be a construction worker with side jobs on the weekends. So you’re a multiple jobholder with regular wages/salary and gig work.
If you lose your gig work, but keep your job with regular wages/salary, you cannot apply for unemployment benefits. But you’re still employed too. If you lose both jobs, you can apply for unemployment benefits based on your salaried job. If you have two full-time jobs working from home, and you lose one of them, you cannot apply for unemployment benefits, but you still have one full-time job and are working. Working from home created a lot of multiple job holders, LOL.
And so every time I post this, year after year, we have the same discussion, oh wait, the gig workers… and it never makes any difference because of reality.
Hi Bob,
I asked Google and its AI replied:
“Gig economy work, while growing, constitutes a small portion of total hours worked in the US. Estimates suggest that gig work represents approximately 0.1% of total hours worked. This is further broken down, with a substantial portion attributed to services like Uber, according to one estimate. While 36% of US workers engage in some form of gig work, this doesn’t necessarily mean they are working full-time gig-related hours.”
Speaking of questions, here are my internsl and external favorites.
Internal: “You’ve been wrong so many times in the past, what makes you think you’re right this time?”
External: “Discount for cash?”
Can’t wait to hear all of the anecdotal evidence
This reminds me of an expression: “The plural of ‘anecdote’ is not ‘data’.”
Pithy!
Good one P, except anecdotal information became declared data by and as the opinion of statisticians and others working to convince the world that the aggregation — and ”proper” massaging of such anecdotal information was acceptable as data.
This change, initially mostly based on the work of medical epidemiology regarding asbestos began to be accepted in 1977, and probably saved my life and others who were working with the stuff at that time.
You like this over the Sahm rule Wolf?
Oops mean that for you. Sahm rule triggered twice last year
My recession indicator needs both: a “surge” (a fast increase) AND a “high level” to get triggered. See the four paragraphs under first chart.
The Sahm rule failed in 2023/4, predicting a recession when there was none. Last year, Sahm already said her rule doesn’t apply anymore. So that was that.
What the Sahm rule failed to include is the second factor, the “high level.” It only says that there needs to be a fast increase, such as in 2022/3. But there was no recession in 2022/3/4 because the levels of unemployment were very low. If she had included the “high level” in her rule (and come up with a historically relevant level), her formula would not have given false recession calls.
So read those four paragraphs under the first chart. They’re key.
Aside from that, the Sahm rule relies on a three-month moving average of monthly data, which is much slower to react than the four-week moving average of my indicator. So I’ll get a few weeks extra notice LOL
No chance for a recession, Bitcoin is at an all time high.
I have a question. Perhaps Wolff or others can answer.
2025 is different in terms of employment when compared to the last recession which was from 2008.
The difference is that now we have gig jobs. If a person is out of work and they do a gig job for Uber or food delivery, that counts them as employed. So how do we take that in to account when trying to decide if we are in the beginning stages of a recession?
It’s not that different this time. There is nothing new about gig jobs.
Data: Of the 163 million total employment, including gig work, 152 million are employed with wages/salaries. And only 9 million are exclusively self-employed (gig work, no salary). But of the 152 million with regular wages/salaries, many also have side jobs (consulting, coding, driving, construction work on weekends, etc.) that count as gig work. These are the multiple jobholders. (BLS jobs report last Friday, household survey data).
Business owners who earn a regular salary from their business, such as me, and are included in the 152 million, are not eligible for unemployment compensation though we pay into the system. The idea is that we cannot lay off ourselves. So if our business can no longer pay us and lays us off, we’re just SOL, and there are millions of us, but that has always been the case, nothing new. (But their employees are eligible for unemployment insurance).
There are some other categories in that 152 million that are not eligible for unemployment insurance. But that’s not new either.
So you may be a teacher and drive for Uber in the evenings and weekends. Or a Google engineer working on some projects on the side (gig work). You may be a construction worker with side jobs on the weekends. So you’re a multiple jobholder with regular wages/salary and gig work.
If you lose your gig work, but keep your job with regular wages/salary, you cannot apply for unemployment benefits. But you’re still employed too. If you lose both jobs, you can apply for unemployment benefits based on your salaried job. If you have two full-time jobs working from home, and you lose one of them, you cannot apply for unemployment benefits, but you still have one full-time job and are working. Working from home created a lot of multiple job holders, LOL.
And so every time I post this, year after year, we have the same discussion, oh wait, the gig workers… and it never makes any difference because of reality.
Wolf: “It’s not that different this time. There is nothing new about gig jobs.”
Far damn shore!!! Regular work mid 1950 era: delivering 125 newspapers EVERY morning; gig work: mowing lawns and other yard maintenance, washing cars, etc., at a time when almost everyone had something else going such as office then farming, etc…
Regular work late ’60s: Lab helper M-F; gig work: handyman whatever, yard maintenance and rehab, etc.
Etc., right up to retirement, because I enjoyed it, especially the gigs!
Thanks Wolf. Common sense is so damn rare these days. I really appreciate your insights.
Not for nothing, but what kind of employee gets laid off also matters.
Is it a $30K McDonald’s line chef, or a $300K Google software engineer? The impact is probably a factor of how much taxes they paid.
Hey Wolf, I don’t know if they will be able to actually get full price, but Ford just raised the prices on three models from Mexico…..
This is what I said above about Ford’s price increases:
Automakers might want to raise prices but demand will plunge if they do because they’re already having trouble selling what they built, and are having to throw huge incentives on their vehicles to move them. They’re loaded with new-vehicle inventory! If Ford hikes prices on its Mustang Mach-E, sales will collapse. It will just sit on these vehicles. And it will stop making them (in Mexico). The free money is gone. That stuff doesn’t work anymore.
Ford is screwed. But GM is screwed even more; it imports vehicles from China (Envision), South Korea, Mexico, and Canada. GM said in an earnings warning a few days ago that the tariffs are going to come out of its profits to the tune of $4 – $5 billion this year. Ford also issued an earnings warning because it will have to eat the tariffs or watch its sales collapse, but the number was smaller (a hit of about $1.5 billion) because it imports a smaller share of its vehicles than GM.
There are many models made in the USA with far more US content, and therefor far fewer tariffs, from Tesla, Honda, Toyota, Hyundai-Kia, etc. and if Ford jacks up the price of the Mach-E (made in Mexico), people will just buy an EV made in the USA. That’s the purpose of tariffs.
New vehicle inventory in the US is large, and automakers have to CUT PRICES (throw in incentives) to sell it and to keep their plants running. So price increases are going halt sales. It’s as simple as that. The free money is gone, prices are VERY HIGH already, and there is no more room left to hike, and all automakers know this.
Long time friend is a auto insurance agent told me that insurance on most Buicks will be rising considerable due parts sourced from China. The offshoring chickens are coming home to roust here as well.
Wolf, it’s probably not a significant impact but you aren’t factoring people who reached the end of unemployment benefits. I was laid off last March so I would have shown up on graph 2 at that time. I received unemployment for 6 months where I showed up on graph 1. After that I disappeared from the data but was still unemployed for another 7 months before starting a new job. I’m sure it’s not this extreme but if 200k are laid off each month and are unemployed for years but lose benefits after 6 months you would only see the total unemployment in graph 1 at 1.2 million but there would be an extra 200k each month that lost benefits. Obviously the numbers aren’t that big but could be a small factor. This is especially true if the layoffs happen over time over a few years instead of a dramatic spike.
I’m sorry to disagree with you, Wolf, but unemployment statistics aren’t always a leading indicator of a recession. It depends on the type of recession. In a financial crisis like the one in 2009, unemployment is the consequence, not the canary in the mine, of the widespread decline in the value of financial assets.
Unemployment can be the trigger for a recession, due to the drop in consumption that leads to a self-reinforcing drop in activity. But there are other triggers, such as a debt crisis.
All the best from Spain.
Esmolante – one might instead say that the factors you mentioned culminated to the point where they led to a surge/elevated levels of unemployment? And that a surge/high level of unemployment has proven to be a leading indicator (correlated to) the onset of a recession.