The San Francisco Fed tackles that question in a new study of the labor force. The study is shockingly flawed, yet worth discussing.
Please consider Breakeven Employment Growth by the San Francisco Fed.
Defining breakeven employment growth
The unemployment rate has stayed very close to its long-run level of 3.8% since early 2022, while monthly job gains have consistently outpaced past trends for breakeven employment growth. Before the pandemic, monthly payroll growth consistent with maintaining unemployment at its long-run value was generally considered to range from 70,000 to 100,000 (see Bidder et al. 2016). Over the six months prior to May 2024 however, monthly job growth averaged 242,000 without significant changes in the unemployment rate. This extended period of elevated job growth has raised questions about the need to reassess the level of breakeven employment growth. In other words, could the pandemic have led to longer-term structural changes that affected normal labor market patterns?
Simply stated, the size of monthly job growth needed to avoid a change in the unemployment rate depends only on the size of labor force growth. Say, for example, the labor force is adding 100,000 individuals per month and the unemployment rate is currently at 4%. The economy would need to create 96,000 jobs to break even with labor force growth and avoid any change in the rate of unemployment. A faster pace of job growth will push the unemployment rate down because people are finding jobs faster than the labor force is growing. However, if job growth falls short of that number, the unemployment rate will begin to rise.
Fact Check False!
This statement is false in theory and practice by definition: “The size of monthly job growth needed to avoid a change in the unemployment rate depends only on the size of labor force growth.“
We are discussing employment and unemployment, not jobs. At least we should be.
The Unemployment Rate = Unemployed / Working Force.
Working Force = Employed + The Unemployed
The Unemployed = Those not employed but are actively seeking employment
There is no number of jobs the economy can add to guarantee a steady unemployment rate no matter what the Labor Force is or does.
For example, let’s assume the economy adds 200,000 jobs.
Q: Does the Labor force necessarily rise as a result?
A: No
Q: Does the addition of 200,000 jobs change employment in any way?
A: No
Q: Does the addition of 200,000 jobs change the unemployment rate in any way?
A: No
The Labor Force, the Unemployment Rate and Employment are all from the Household Survey and directly related.
The San Francisco Fed mixes in monthly nonfarm payrolls, part of the Establishment Survey, without mention.
The entire discussion is therefore silly, but we can learn some things from it, so let’s continue.
Unemployment Rate
The unemployment rate is creeping up despite allegedly strong job growth, with emphasis on allegedly.
The author discusses long- and short-term impacts of rising immigration. Despite crucial flaws in the SF Fed discussion, this chart is worth discussing.
Long- and Short-Run Breakeven Employment Growth
Our baseline calculations assume a long-run unemployment rate of 3.8%. The results are shown in Figure 2 for long-run employment growth and Figure 3 for short-run employment growth. Long-run breakeven employment growth is estimated to range between 70,000 and 90,000 jobs per month in 2024, even after taking into account recent higher-than-expected labor force growth and projections for higher immigration scenarios from Census and the CBO (green and red lines in Figure 2, respectively).
Short-run breakeven employment growth is estimated to be higher than long-run growth under each scenario, as shown in Figure 3. Under the baseline scenario, short-run breakeven employment growth is estimated to be around 140,000 jobs per month in the first quarter of 2024 (dark blue line). It is somewhat more elevated under the high immigration Census scenario at 151,000 jobs per month (not shown) and significantly higher at 230,000 jobs per month under the CBO high immigration scenario (red line), reflecting the recent surge in immigration that is projected to largely continue in the near term. Under the baseline projections, short-run breakeven growth will converge on the long-run breakeven growth rate (gray line) by the end of 2025. However, this return to the long-run trend stretches further out to 2027 for the CBO high immigration scenario.
I suggest the SF Fed stop mixing employment with jobs or at least state what it believes the relationship is.
The discrepancy between the two is enormous.
Jobs Much Weaker than Expected
The above chart is from my July 5 report Jobs Much Weaker than Expected, the Unemployment Rate Ticks Up
Here are some amusing details.
Jobs vs Employment
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.
Job Stats vs One Year Ago
- Nonfarm Payrolls (Blue): +2,611,000
- Employment (Red): +195,000
- Full Time Employment (Yellow): -1,551,000
The SF Fed says the Census scenario is 151,000 jobs per month but significantly higher under the CBO high immigration scenario at 230,000 jobs per month.
In the last year, jobs (which the SF Fed uses in its analysis) are up by 2.6 million. That’s 217,000 a month on average.
Meanwhile, employment is up a mere 195,000. That’s only 16 thousand a month.
Is it any wonder the unemployment rate is up 0.5 percentage points from 3.6 percent to 4.1 percent in the past year despite a rise in employment of 2.6 million?
Report Conclusion
The unemployment rate has been very near its long-run value since early 2022, while monthly job gains over the six months prior to May 2024 have consistently outpaced estimates of the breakeven employment growth that holds unemployment constant. This pattern has raised questions about whether the breakeven pace of employment growth has risen above past estimates. The analysis in this Economic Letter distinguishes between short-run breakeven employment growth, which depends on near-term labor market volatility, and long-run breakeven employment growth, which reflects more durable determinants of labor force growth. We find that the long-run breakeven pace of employment growth has changed little from prior estimates of just under 100,000 per month, after accounting for short-run dynamics of the business cycle and immigration. However, the short-run breakeven pace remains well above the long-run value for the time being, which explains the stability of the unemployment rate in the face of persistently high job growth.
Mish Conclusion
The report is shockingly flawed for failing to discuss the points I raised. If the Fed believes there is some stable long-term relationship between jobs and employment then it should state those assumptions.
Moreover, the authors seems oblivious to the fact that unemployment rate is not exactly stable right now.
Failure to understand (or discuss if it does) the difference between employment and jobs is at the heart of this discussion.
It’s perhaps understandable for the CBO to make such mistakes (at least more understandable), but there is no excuse for the Fed.
Questions Not Answered
- Immigration: Has immigration changed the relationship between jobs and employment?
- Covid: Did Covid or work-at-home impact the relationship between jobs and employment?
- Inflation: How does a sustained rise in inflation impact the relationship between jobs and employment?
- Demographics: How does the fact over 20 million boomers are of retirement age but still working change things?
Those questions were not even discussed.
Job Revisions
Jobs are not in the equation, but if the Fed insists on putting them in the equation, then they also need to discuss revisions.
For a discussion of revisions and the very unstable nature of the reports lately, please see How Much Faith Do You Have in BLS Job Reports?
Negative revisions in nearly all the economic strongly suggest recession. That’s another topic the San Francisco Fed did not touch.
Data Says a Recession Has Already Started, Let’s Now Discuss When
The Fed models and CBO models never factor in recession. But recession is now my base case.
For discission, please see Weak Data Says a Recession Has Already Started, Let’s Now Discuss When
The recession indicator discussed above is based off the unemployment rate.
It might behoove the Fed to check that out in addition to reading up on the essential difference between jobs and employment.
I find this all fairly confusing – according to usdebtclock.org the workforce/labor force #’s have been stagnant for at least a 1/4 century
I remember looking at the way they calculate inflation in a college course called math for Economists. The teacher showed us the equation, which was the 1990 model, and after showing us all of the ways they play with it- he asked by a show of hands how many of us thought it was accurate- no one raised their hands.
Look at US Labor Participation on the St Louis Federal Reserve website.
Its at 62.6% and nearly same as it was in July 2023, and compare it to 63.3% in February 2020.
Mish is right that less jobs need to be created if there is less labor participation.
However, how many working under the table and not being counted by the labor participation rate ?
do they work the gig economy (and under the table) and make enough money while not being counted in the labor participation rate ?
Perhaps the actual labor participation rate is much higher and is closer to the peak rate set of 67% in 2000, when the gig economy was not as prevalent.
Pro Hamas won in France. Gabriel Attal, France first gay PM, might resign. He and Macron nominated his ex lover Stephae Sejourne as minister for Europe and foreign affair, kicking him upstairs. Mbappe and Dembele were ineffective. Conclusion : Iran is behind the Pro Hamas protests in the US, France and the UK creating chaos before the election for Biden and Macron.
The most likely answer is that jobs have not actually risen by 240,000 per month for the past several months.
P.S. Mish, I think near the end you wrote “employment” when you meant “jobs”: “Is it any wonder the unemployment rate is up 0.5 percentage points from 3.6 percent to 4.1 percent in the past year despite a rise in employment of 2.6 million?” As you wrote above, Household survey Employment is only up 191k, it’s the establishment “jobs” estimate that is up 2.6M.
One thing I really want to add: the relevant part of the Fed’s legal mandate is “maximum employment” NOT maximum “jobs” nor “minimum U-3 unemployment”.
It’s clear from the Household Survey that the economy is not increasing employment as rapidly as it should be. This should be triggering serious thought and creative approaches at the Fed, especially since interest rates cannot currently be lowered without reigniting inflation. But they instead are fixated on Estabkishment payrolls, U-3 and JOLTS data. Normally those track actual Emoloyment but now they do not, and the Fed is oblivious.
also missing is any acknowledgement of F-T jobs being permanently replaced by P-T jobs.
literally millions of FT jobs (w/ benefits) have evaporated, and those FT jobs are not coming back anytime soon – payrolls are now being structured for permanent PT capacity & lower costs (eliminating FT jobs w/ benefits).
the qualitative/downstream value of FT jobs VS. PT jobs is massive… enormous… impacting macro/social/economic behavior & assumptions for years+decades to come.
family formation, home ownership, rate of savings, debt capacity, physical mobility, inter-generational dependence, investment preferences, standard-of-living… you can burn all the historical data & assumptions as the definition of “job” shifts away from Full-Time and is replaced by Part-Time.
Exactly. Eberstadt has good analysis of the millions of people out of the labor force. Because the composition of the jobs market does not provide enough for a good 10% of people to have a home and a family
The 2 causes are automation and globalization of the labor market. Lots of argument about the contribution of each
Yes, it creates a nation of renters and boarding houses. I wonder what jobs were converted from full time to part time.
I know where I live in the Florida panhandle that the service workers work 2 part time jobs to make an equivalent of a full time job.
They are careful not to reveal in an interview for a service job like a restaurant job that they have another part time job like cleaning vacation rental condos.
Hiring managers won’t hire them if they know they have another job as they want them to work 24 hrs per week, and then be “on call” or readily available for another 16 to 24 hours if someone calls out sick, etc.
N-gDp falls too low after JULY this year. And foreign demand for Treasuries exceeded over $700B over the last year. Time to buy TLT
“The study is shockingly flawed…”
If that is true, the study is shockingly fraud.