Another seemingly strong jobs headline falls apart on closer scrutiny. The massive divergence between jobs and employment continues.
From September 2020 through early 2022, nonfarm payroll job gains and full time employment changes tracked together.
Starting around March of 2022, a divergence between employment and jobs became very noticeable, and I have been discussing the divergence since then. You seldom see this mentioned in mainstream media.
Payrolls vs Employment Gains Since March 2023
- Nonfarm Payrolls: 2,602,000
- Employment Level: +144,000
- Full Time Employment: -284,000
Payrolls vs Employment Gains Since May 2022
- Nonfarm Payrolls: 5,880,000
- Employment Level: +2,654,000
- Full Time Employment:+270,000
Payrolls are up by 5.88 million since May of 2022, but full time employment up only 270 thousand.
No amount of BLS smoothing can hide this, so hardly anyone discusses it.
Q: What’s going on?
A: People are working multiple part time jobs and or semi-retired boomers are working part time.
Job Report Details
- Nonfarm Payroll: +275,000 to 157,808,000 – Establishment Survey
- Civilian Non-institutional Population: +171,000 to 267,711,000
- Civilian Labor Force: +150,000 to 167,426,000 – Household Survey
- Participation Rate: +0.0 to 62.5% – Household Survey
- Employment: -184,000 to 160,968,000 – Household Survey
- Unemployment: +344,000 to 6,458,000- Household Survey
- Baseline Unemployment Rate: +0.2to 3.9% – Household Survey
- Not in Labor Force: +20,000 to 100,285,000 – Household Survey
- U-6 unemployment: +0.1 to 7.3% – Household Survey
Nonfarm Payroll Change by Sector
Government and Health Services are related to the surge of illegal immigrants and the need to address them. Social assistance jobs rose by 30,000 in January and another 24,000 in February.
Government jobs rose by 52,000 in both January and February.
Those are additional facts you don’t hear about in mainstream media.
Change in Nonfarm Payrolls January 2022 to February 2024
Monthly Revisions
- The change in total nonfarm payroll employment for December was revised down by 43,000, from +333,000 to +290,000
- The change for January was revised down by 124,000, from +353,000 to +229,000.
- With these revisions, employment in December and January combined is 167,000 lower than previously reported.
Part-Time Jobs
- Involuntary Part-Time Work: -46,000 to 4,376,000
- Voluntary Part-Time Work: +153,000 to 22,156,000
- Total Full-Time Work: -187,000 to 132,946,000
- Total Part-Time Work: +51,000 to 27,941,000
The above numbers never total correctly due to the way the BLS makes seasonal adjustments. I list them as reported.
Hours and Wages
This data is frequently revised.
- Average weekly hours of all private employees rose 0.1 hours to 34.3 hours.
- Average weekly hours of all private service-providing employees rose 0.2 hours 33.3 hours.
- Average weekly hours of manufacturers rose 0.1 hour to 39.9 hours marking the fourths consecutive month below 40 hours.
An overall decline or rise of a tenth of an hour does not sound line much, but with employment over 160 million, it’s more significant than it appears at first glance.
Hourly Earnings
This data is also frequently revised. Here are the numbers as reported this month.
Average Hourly Earnings of All Nonfarm Workers rose $0.05 to $34.57. A year ago the average wage was $33.15. That’s a gain of 4.28%.
Average hourly earnings of Production and Nonsupervisory Workers rose $0.07 to $29.71. A year ago the average wage was $28.42. That’s a gain of 4.53%.
Year-over-year wages are keeping up with inflation after underperforming for many months.
Unemployment Rate
The unemployment rate hit a 50-year low in January and April of 3.4 percent. It’s now 3.9 percent, the highest since December of 2021. A
3.9 percent is still a very low number, but it’s now clear the unemployment rate has bottomed this cycle and will generally head higher.
Alternative Measures of Unemployment
Table A-15 is where one can find a better approximation of what the unemployment rate really is.
The official unemployment rate is 3.9%.
U-6 is much higher at 7.3%. Both numbers would be way higher still, were it not for millions dropping out of the labor force over the past few years.
Some of those dropping out of the labor force retired because they wanted to retire. Some dropped out over Covid fears and never returned. Still others took advantage of a strong stock market and retired early.
The rest is disability fraud, forced retirement (need for Social Security income), and discouraged workers.
Birth Death Model
Starting January 2014, I dropped the Birth/Death Model charts from this report.
The birth-death model pertains to the birth and death of corporations not individuals except by implication.
For those who follow the numbers, I retain this caution: Do not subtract the reported Birth-Death number from the reported headline number. That approach is statistically invalid.
The model is wrong at economic turning points and is also heavily revised and thus essentially useless.
Birth-Death Methodology Explained
Every month this subject comes up. I gave a detailed explanation of the model and why the hype is wrong in my December 8, 2023 post How Much Did the Huge 412,000 Birth-Death Adjustment Impact October’s Job Report?
The month does not matter. If you think the model has a big impact, please click on the above link for why it doesn’t.
Household Survey vs. Payroll Survey
- The payroll survey (sometimes called the establishment survey) is the headline jobs number. It is based on employer reporting.
- The household survey is a phone survey conducted by the BLS. It measures employment, unemployment and other factors.
If you work one hour, you are employed. If you don’t have a job and fail to look for one, you are not considered unemployed, rather, you drop out of the labor force.
Looking for job openings on Jooble or Monster or in the want ads does not count as “looking for a job”. You need an actual interview or send out a resume.
These distortions artificially lower the unemployment rate, artificially boost full-time employment, and artificially increase the payroll jobs report every month.
Employment Revisions to December 2023
Note: This chart is a repeat from last month.
Whereas the BLS revised Jobs for December by +333,000, the BLS revised employment for December by -270,000.
Final Thoughts
This report is much worse than headline numbers indicates. I said the same thing for the last three months, and it’s generally been that way for over a year.
In December, I commented “A decline in full time employment of 1.5 million is remarkable. This series is heavily revised so let’s see what January brings.”
In January, I commented “We now have an answer. Other than the headline jobs numbers, this report was terrible. There are 353,000 more jobs, but hours worked took an unusual dive. The continued dependence on government jobs, up another 52,000 in December and 36,000 in January also masks weakness.”
For February, the BLS revised January from 353,000 jobs to 229,000 jobs. The string of generally negative revisions has been huge. And once again we have a huge number of government jobs propping up the economy.
Ignore the Amazing Headline Job Numbers, Note the Revisions
Note: This chart is also a repeat from last month.
Can the Jobs and Employment Numbers Both Be Reasonably Correct?
The answer is yes (discounting measurement error) because they measure different things. A person working three part time jobs counts for three jobs but only a single person employed.
I have repeatedly asked ADP to account for duplicate social security numbers but they won’t. Amusingly, the BLS wants to, but the employees tell me they can’t because “they don’t have access to the data for security reasons.”
This is a simple sort-merge program but alas, we depend on a phone survey for employment numbers.
Notably, discrepancies like these don’t last for years unless there is some truth to the employment numbers because measurement errors are random.
Over the past 21 months, payrolls have increased by 5.88 million while full time employment is up a mere 270,00. Democrats cheer this performance and mainstream media fails to note.
“You seldom see this mentioned in mainstream media.”
Mainstream media is propaganda. After i get the MS #, i read Denninger’s & Mish’s take on what’s under the cover story.
Apparently the left wing mainstream media’s parroted buzz word for Biden’s speech, was “fiery.” Not the word they would have used if Trump gave the exact same speech in the same tone.
evwithin fckn great then ! 10% on Tbills soon then?
on reasons for the jobs/vs employment trends
I agree on people taking multiple part time jobs and retired folks taking part time jobs (count me as one of them).
One piece of confounding data is the increase in average hours worked. Not sure which survey that data comes from.
A few other possibilities:
Gig workers taking part time payroll jobs. This could include people buying and selling stuff on ebay, craigs list etc. (my brother supported himself decently doing this for over 15 years). Uber/Lyft and food delivery drivers, free lancers of various types, etc. I assume these are not counted as Payroll employment. Some might even take more than one part time job as opportunities and other interests wax and wane. I think this could be quite a few folks.
How are other self employed people counted? Also job shoppers, contract consultants etc. They do not seem to be “payroll” jobs, but the definitions could shift over time.
“Employed” people switching from “farm” to “non-farm” payroll jobs. This has been a trend for the last years or so. Also, it is not clear in the agriculture related industries what is considered as “farm” versus “non-farm.” Could these definitions be shifting.
Covid and the recovery from covid could have magnified some shifts of the above.
Interesting post, and a salute to Mish for looking under the hood. I have been critical of some of his posts but think he does an outstanding job with the employment data.
If he is correct about the coming trend, that will be bad news for Biden. In the presidential elections since WW2, the incumbent party’s presidential candidate has lost when U3 rose in the second quarter, and has won if it declined. The only deviation was in 1956, when Eisenhower was re-elected in site of a U3 increase in the spring quarter.
In that year, U3 was trendless, and the Q2 rise was slight. If this spring sees a steady rise between March and June, the indicator will flash a major warning sign. The indicator is just an indicator, but it has an enviable track record including in 1992, when Perot got 19% of the popular vote.
The logic behind it is that unemployment is a proxy for economic conditions on Main Street, which are reflected in sentiment and voting behavior with a lag. What could be relevant this year is that full-time employment levels are flat, and that part-time employment has grown as a share of the total. My guess is that people working part-time aren’t as happy with the economy as those with full-time jobs, but that’s only my guess and nothing more.
So the March through June U3 numbers will be something to watch closely. This isn’t some fluke; not only has the indicator been amazingly reliable, but there is an underlying logic to it. Past that, I agree with Mish that the employment situation is not nearly as rosy as the media portray it to be.
Finally, it doesnt take much of a rise to trigger the indicator. Even small rises have been accurate. Big rises have been very reliable, with 1980 and 2008 being good examples. The big rise in 22 was also accurate, yet that must be examined somewhat searately given the covid economic shutdown. Had an increase of that magnitude not been triggered by the “black swan” event, Trump would have been crushed.
Let U3 go from, say, 4% in March to 4.3% in June, and the handwriting will be on the wall. Let it go from 4% to 4.5%, and Joe will definitely be history. We shall see.
Correction due to wonky keyboard. ” … the big rise in 2020 …”
Unemployment at 3.9% and GDP growth of 3.2% annualized and people are still talking about “rate cuts”.
Ridiculous.
No wonder Gold is going nuts.
Follow the retail trade sleaze meter.
Meanwhile, gold is breaking out of the long-term cup/handle formation. This was expected and launched recently. Now we need the gold miners to follow.
Yes, the handle on the decade-long cup-and-handle formation was completed in November, 2022 at $1,680.
We are now into the “Bang, Zoom!” phase.
I can’t afford to work anymore. But all these nice new social workers need me to attend to and keep my data updated to each new requirement. Without folks like me they would have nothing to do.
Absolutely nothing about this jobs report is “falling apart”. We’re still sub 4% unemployment. The jobs being created are at the margins and have been for the past 18 months or so. Any report that’s positive job growth is good news. Until the labor market really softens (~2.2M continued un-employment claims & we’re sub 1.9M today), then JPowell isn’t going to lower the FFR.
Remember that Uncle Joe changed the definition of a contractor, so all of the uber drivers, and drizzly delivers, etc. have had to go on payroll (another inflationary regulation). I’m betting this is why payroll is rising and jobs are not, its simply continued reclassifications.
So i just looked it up – Uber has about 5 million dirvers. So that would be worldwide, but say half are in the US that would account for about half of the difference alone. From my experience most Uber drivers also work for Lyft, and probably for Doordash, and Drizzley, etc. There you go. Uncle Joe Did create 5 million jobs, just not 5 million new employees.
Its 1984 man!
Very interesting! I have some questions.
When did that go into effect?
Were these people not counted as self-employed before the reg took effect?
What is the effect on the unemployment and parttime v fulltime numbers, if any?
The above questions are NOT rhetorical or ax grinding. I really want to know, because I follow the numbers.
Announced June 2022 and effective January 2024. Since this is probably the last “jobs report” that would be picking up the transaction, it would be interesting to see if “jobs” and actual employment start to track better in the coming months.
Correct. Gig workers were self-employed, and as such they did not received unemployment insurance. Therefore, they were not counted in the payroll survey, but they were counted in the household survey. This means that the reclassification would not change the unemployment rate or participation rate. On the household survey you are working if you are either self-employed as a gig worker, or employed as a gig worker.
So what does this do, it sends the headline numbers up but does not impact anything else. This is exactly what we are seeing. Uncle Joe can crow about the number of “jobs” created, but really no new employment is created.
It also raises labor costs, lowers productivity, and leads to fewer hours worked since Uber (and lyft, drizzly, etc.) does not want to have to also start paying for obamacare, overtime, etc.
So I think this is a major reason for the divergence between the headline number and actual employment. The interesting question, for me is if a driver has to go on payroll for say Uber, can they still drive for Lyft or does Uber restrict that in their labor contract? Or are individuals being employed (and counted in the headline number) 3 or 4 times – an employee of Uber, an employee of Lyft, an employee of Doordash, and an employee of Drizzly??
Does anyone have any thoughts on that?
Did the rule then raise total compensation? When they were self-employed contractors, didn’t they have to pay FICA and income tax? The only extras that occur to me are overtime, which I’d imagine the companies make sure they don’t pay by keeping hours under the threshold, and the UE insurance kitty. What might I be missing?
What do you mean by the “headline number?” I have always viewed U3 as the headline number, but from the context I don’t think you are referring to that.
Once again, these are not rhetorical or ax grinding questions. I feel compelled to repeat that so you don’t think that I am trolling you. I’m a numbers nerd, and you raised something that I had neither noticed nor thought about until seeing your comment. I am genuinely interested.
Headline number os the payroll survey
Lots of other costs. Employers have to not only pay overtime and unny but paid family leave sick time vacation time etc.
Also huge number of regulations osha, labor dept, state regs, local regs, liability insurance. Costs a good 30 percent more to have an employee and as much as 40 percent more if ypu provide insurance
Believe me. Im an employer
I tend to believe you. It makes intuitive sense. Not only that, but I am highly sympathetic to the issues facing businesses. Yet I wonder if you have any links to show how this has affected the companies we are talking about, i.e. Uber, Lyft, Door Dash, et. al.
Have you by chance looked at any of their 10-Ks or Qs? I know this is slippery because of their ability to recapture it through dynamic pricing. If the cost increase is very much, wouldn’t this be seen on the income statement margins?
This is my numbers nerd talking. The time to check most aggressively is when you want to believe something. If you have been able to quantify it through actual, reported results, Id love to know. Trust me, this is not my Inner Communist at play at all.
I have not but its worth a look.
Dimocrats decreed that Employers must cover the cost of Obamacare for 30 hour week employees. The poor devils who voted Dimocrat are now working two 25 hour week jobs to feed themselves.
Finally, MarketWatch has actually lightly covered the discrepancies within these reports. The re-election committee’s work on LBS stats has been unwittingly exposed.
But I expect very little additional exposure. After all, MSM don’t want to be added to any disinformation list.
Don’t blame the BLS. They just count what they count. As I commented, changes in regulations are not taken into account, so yes there are more people on payrolls but likely less people actually working. When you force a big change in the economy, you will bollux up the numbers.
Duplicate social security numbers (meaning stolen or copied, etc.) is eventually going to be a big problem. Only 1 person is going to get benefits, no matter how many are using the same number. The government doesn’t care, because it’s more income without increased cost. Wait till the lawsuits start because an illegal immigrant, who’s been using someone else’s social security number, can’t collect benefits.
It’s worse. The people using stolen SS#’s aren’t paying the required SS tax on it so when the true owner eventually tries to collect they’ll discover they owes those taxes instead. Good luck working that out 15-20 years from now when it finally catches up to them.
50% Discount/Rebate at retail and 50%Gift/Debt Jubilee makes a $500k house cost you only $125k. ‘Nuff said.
Its called making accounting and the money system serve us instead of slavishly forcing us onto all fours.
Maybe we can all work for the government in the future.
Most of us do already, whether or not employed by government. One must assume all government agencies are now politicized and data can no longer be relied upon.
that’s the move that I made.
my probation ends 9 days after Trump takes office in 2025. make it past that and I’m gold…
That’s the plan. A form of neo-feudalism. You’ll own nothing because you can’t afford it. Be grateful for what the overlords provide. Look to Gaza to see your future.
What is the Trumpian phenomenon if not a form of peasant uprising against the elites (pejorative), and oligarchs, who have been dumping on them, and showing them the finger.