The Philadelphia Fed Just Revised Jobs Lower by 1.2 Million for Q2

Early Benchmark Revisions of State Payroll Employment

Please consider Early Benchmarks for All 50 States and the District of Columbia for the second quarter of 2022.

Estimates by the Federal Reserve Bank of Philadelphia indicate that the employment changes from March through June 2022 were significantly different in 33 states and the District of Columbia compared with current state estimates from the Bureau of Labor Statistics’ (BLS) Current Employment Statistics (CES). Early benchmark estimates indicated higher changes in four states, lower changes in 29 states and the District of Columbia, and lesser changes in the remaining 17 states.  

In the aggregate, 10,500 net new jobs were added during the period rather than the 1,121,500 jobs estimated by the sum of the states; the U.S. CES estimated net growth of 1,047,000 jobs for the period. Payroll jobs in the nation remained essentially flat from March through June 2022 after adjusting for QCEW data.  

Early Benchmark Methodology

Preliminary (not-yet-benchmarked) state employment estimates from the Bureau of Labor Statistics (BLS) continue to be subject to significant revisions around turning points in the economy. These large revisions occur primarily because the preliminary state estimates are based on a small sample of firms, while subsequent annual benchmark revisions incorporate other BLS data based on a full count from nearly all firms.

The Federal Reserve Bank of Philadelphia has developed early benchmark estimates of monthly state payroll employment on a quarterly basis to predict the subsequent annual benchmark revisions by the Bureau of Labor Statistics (BLS). Our process enhances the monthly Current Employment Survey (CES) payroll employment data with the more comprehensive Quarterly Census of Employment and Wages (QCEW) payroll employment data. The CES provides a timely estimate of monthly state employment data, but the QCEW follows about five months later with a more complete picture, covering more than 95 percent of all employers. 

Synopsis

• The BLS Current Employment Survey (CES), the monthly jobs report, is timely but inaccurate, and grossly inaccurate at turns. 
• Nonetheless, economic cheerleaders tout the monthly data even when the household survey shows major discrepancies. 

For months on end, we heard an endless stream of “Jobs are too strong for there to be a recession,” despite the known fact that jobs are a lagging indicator, and despite the Household Report differs strongly.

My long-held position has been the household survey is more likely to be accurate at turns. 

Fall & Winter Double Issue

With that background please consider the Philadelphia Fed Fall & Economic Insights Winter Double Issue for 2022 Q3 and Q4, emphasis mine.

 Using QCEW Data to Revise Estimates In March of each year, the BLS releases revised estimates of monthly nonfarm payroll employment for states and metropolitan statistical areas (MSAs) as part of its Current Employment Statistics (CES) program. For its annual revisions of CES state estimates, the BLS incorporates more comprehensive data from the Quarterly Census of Employment and Wages (QCEW) program, which is also released by the BLS. The BLS also introduces new seasonal adjustment factors and other corrections to make the data revisions more accurate. For our purposes, the most significant monthly revisions affect the prior seven quarters of data. The QCEW data make a significant contribution to the annual revisions. Whereas the QCEW data cover more than 95 percent of all employers, the CES sample represents just 6 percent of the QCEW total. Therefore, the CES state estimates that result from the annual revision process reflect the broad universe of firms (as well as new seasonal factors) and thus more accurately depict a state’s job growth trend than does the original CES sample alone. 

The CES program relies on a monthly nationwide survey of about 131,000 businesses and government agencies representing about 670,000 establishments. These samples are used to estimate total employment not only of states and MSAs but also of industrial sectors within states and MSAs. In contrast to the CES sample of 670,000 establishments, the QCEW program reported employment counts for nearly 11 million establishments covered by state and federal unemployment insurance (UI) laws in the first quarter of 2021. The QCEW data for October, November, and December 2021 were released on June 8, 2022.

The monthly jobs report by the BLS samples a mere 6 percent of jobs. Yet people put as much faith in these reports as they do the Bible.

Looking Ahead 

If payroll job growth did shift to a markedly slower pace during the second quarter of the year as interest rates were raised to counter high inflation, our Early Benchmark process should note larger downward revisions in December 2022. Not until February 2024—with the incorporation of the March 2023 benchmarks—will the CES estimates offer a full accounting of U.S. employment for the bulk of 2022. Unfortunately, our Early Benchmarks lag the moments when critical policy deliberations are made, but they do offer earlier confirmation of apparent shifts in recent payroll job trends. And pervasive, persistent, and deep downward revisions may presage the NBER’s declaration of a recession.  

Hoot of the Day

The BLS will finally have a good accounting of what’s happening right now with jobs in February of 2024, about 14 months too late. 

With that observation, consider what I have been saying for many months.

Another Strong Jobs Report? Phooey, and I Can Prove It

Please consider my December 2, 2022 post Another Strong Jobs Report? Phooey, and I Can Prove It

Initial Thoughts

• The discrepancy between jobs and employment continues for the eighth month.
• Lost in the unemployment and jobs headline noise are huge divergences between jobs and employment dating back to March.

Payrolls vs Employment Since March 2022

• Nonfarm Payrolls: +2,692,000
• Employment Level: +12,000
• Full Time Employment: -398,000

Employment fell by 138,000 in November.

Full time employment is down 398,000 since March and down by 480,000 since May!

Typical Headlines

• The StreetInsider headline says “U.S. job juggernaut rolled on in November; nonfarm payrolls up 263,000”
• The Wall Street Journal headline says “U.S. Economy Added 263,000 Jobs in November”

Both reports and mainstream in general miss the big picture captured in my lead chart.

The internal details have been weak for 8 month and I have been talking about the discrepancy for six of them. 

Q&A What’s Going On?

Q: Hey Mish, What’s Going On?
A: People are taking on second part time jobs to make ends meet. But overall employment (the total number of people working is stagnant.

As I have been saying for many months, don’t watch the unemployment rate, watch employment levels.

But hooray! The media reports of a “strong jobs juggernaut” continue unabated.

For March, the BLS said full-time employment was 132,718,000. For December, the BLS said 132,320,00, a decline of 398,000. This discrepancy has lasted 8 straight months. 

Everything points to part time jobs to fueling the job gains (assuming there were gains of any kind).

The StreetInsider forgot to look inside the jobs report to come up with a headline of “U.S. job juggernaut rolled on in November.“

Just Noise

“The BLS provides data used to attribute the 2 million diff.” 

And how timely is that data?

Mish: “Does job data noise last 8 months? I think not. I have tracked job divergences for 10 years and never saw one go beyond 3 months to any degree. GDP resolved in favor of GDI [Gross Domestic Income]. What is the likelihood this goes the other way?“

Questions Answered

1: Apparently not
2: Right now, something approaching 100 percent

Splitting the difference or averaging is reasonable to a point. It’s what the NBER does, sort of. 

Q: Why, do I say “sort of” for the NBER? 

A: The NBER waits so long to declare recessions that major discrepancies have already been revised away. By the time the NBER declares recession, the GDP-GDI discrepancy will be resolved and the Employment-Jobs discrepancy resolved as well.

In this case, once the BEA revised GDI (income) lower, the odds of jobs revisions shot up.

By the way, that jobs revisions is a Philly Fed estimate. The BLS will keep reporting nonsense until its next revision in March. 

“No Recession” Idea Based On GDI Was Just Revised Out the Window

I discussed the GDI-GDP discrepancy on September 29 in “No Recession” Idea Based On GDI Was Just Revised Out the Window

As predicted in this corner, the idea that Gross Domestic Income was too strong for the US to be in recession was just revised out the window.

Given Income was revised lower, what should one have expected from jobs?

With that GDI revision, I thought it was reasonably clear recession had started or soon would.

However, a collapse in exports in favor of domestic demand in Q3 may have sunk that idea. 

Regardless, the 4th quarter is not over yet and weakness everywhere is pronounced.

Barring GDP revisions, perhaps a recession started in November. Previously, I penciled in May. Apologies offered if turns out to be later.

Housing Supporting Evidence 

Please note Existing Home Sales Decline 9th Month, Down Another 5.9 Percent

Except in recessions, nothing like that has ever happened.

Industrial Production Supporting Evidence

In case you missed it, please note The Fed’s Industrial Production Report is a Disaster, Cyclical Data Signals Recession

Clean Sweep for Cyclicals

• Manufacturing Durable Goods: -3.79 Percent
• Motor Vehicles and Parts: -2.84 Percent
• Consumer Durable Goods: -2.12 Percent
• Manufacturing: -0.61

Retail Sales Evidence

• Food Service: +0.9 percent
• Food Stores: +0.8 percent
• Gas Stations: -0.1 Percent
• General Merchandise: -0.1 Percent
• Excluding Motor Vehicles and Gas: -0.2 Percent
• Excluding Motor Vehicles: -0.2 Percent
• Nonstore (Think Amazon): -0.9 Percent
• Motor Vehicles: -2.3 Percent
• Department Stores: -2.9 Percent 

Retail Sales: With Food and Shelter Soaring, Who Can Afford Anything Else?

But hey Mish, the household survey is nothing but noise. Jobs are too strong for there to be a recession. 

Thanks, I’ll make a note based on 6% of the data with the rest of the data coming in 14 months from now.

This post originated at MishTalk.Com

Please Subscribe!

Like these reports? I hope so, and if you do, please Subscribe to MishTalk Email Alerts.

Subscribers get an email alert of each post as they happen. Read the ones you like and you can unsubscribe at any time.

If you have subscribed and do not get email alerts, please check your spam folder.

Mish