There are reasons to be skeptical about the monthly US jobs report
The latest monthly jobs report released by the U.S. BLS (the Current Employment Statistics (CES) report), claims that 517,000 nonfarm payroll jobs were added in January.
While the latest CES jobs report thus suggests that employment growth surged during January, a deeper analysis reveals concerns that create material skepticism surrounding current and past, monthly US nonfarm jobs numbers.
The concerns are more significant than just being related to seasonal adjustments, which many have pointed out are a big factor in January (the non-seasonally adjusted number was -2.5m).
The annual benchmarking process
The most pressing concern is instead in relation to the BLS’ annual benchmarking adjustment. In February of each year (with the release of January’s CES report), the BLS benchmarks its monthly CES jobs report (the most commonly followed indicator of jobs growth), against its Quarterly Census of Employment and Wages (QCEW).
The QCEW is a comprehensive measure of employment that encompasses employers covered by state and federal unemployment insurance laws. This means that the QCEW covers more than 95% of all employers, versus the ~6% of employers that are covered by the monthly CES survey.
As a result of the benchmarking process applied to January’s CES report, monthly jobs numbers across 2022 increased by 311,000.
Sounds good, right? Not so fast. The key thing to note about this annual benchmarking process is that it’s SIGNIFICANTLY LAGGED.
CES numbers weren’t benchmarked to the QCEW report from the December 2022 quarter. Instead, it was benchmarked to the QCEW report from the quarter ending MARCH 2022 — yes, it’s lagged by almost an entire year.
More recent comprehensive employment data paints a far bleaker picture
This is where the significant reasons for skepticism surrounding current employment data come into play, as more recent QCEW data suggests a far weaker employment market.
We are able to know this because while the BLS benchmarks their monthly CES jobs report to QCEW data annually, the underlying QCEW data is released somewhat more promptly — ~5 months after the end of each quarter. This allows insights to be drawn into the employment market from comprehensive QCEW data before the BLS’ annual CES update.
Using this 5 month lagged data, the Philadelphia (Philly) Fed produces the Early Benchmark Revisions of State Payroll Employment (EBR) report, while the BLS produces the quarterly Business Employment Dynamics (BED) report. Both of these reports are currently updated to Q2 2022 QCEW data, as opposed to the CES, which has only been benchmarked to Q1 2022 QCEW data.
What did the Philly Fed’s EBR report have to say? Instead of the now revised 988,000 jobs that the BLS estimates were gained in Q2 2022 in its monthly CES reports, the Philly Fed’s review of the more comprehensive QCEW data, suggests that there were just 10,500 net new jobs created!
Turning now to the BLS’ quarterly Business Employment Dynamics (BED) report, which again, is based on the more comprehensive QCEW dataset, but instead focuses just on private sector establishments.
Here we see that the BED report drew an even worse conclusion — net private job LOSSES of 287,000 from March to June 2022. Over the same period, the monthly CES report recorded total private sector job gains of 951,000.
Household survey supports QCEW data
While this is just one quarter of QCEW data, and it’s possible that the trend may have reversed in Q3 or Q4, its suggestion of a shift in the jobs market is made stronger by the numbers that have been released from the BLS’ household survey of employment.
While the establishment and household surveys generally show a similar trend in terms of job growth over most time periods, the two measures have recently diverged significantly.
When did this divergence begin in earnest? APRIL 2022 — i.e. just as when more comprehensive QCEW data began showing a weak employment market, and which the establishment survey is yet to be benchmarked against. In contrast to the establishment survey, the household survey is NOT benchmarked against QCEW data.
After rising and falling in a broadly similar fashion from January 2019 - March 2022, from April - December 2022, the establishment survey recorded new jobs of 3,132,000, while the household survey rose by just 916,000 — a divergence of 3.4x!
In terms of how each measure counted Q2 2022 employment changes, recall the earlier chart that compared recorded employment changes between QCEW and CES data. Here, one can see that the BIG outlier was the establishment survey. The household survey printed numbers that were much more in-line with the QCEW data.
With the household survey more closely correlated to the Q2 2022 QCEW report than the establishment survey, the weaker job growth recorded by the household survey in in the last 9 months of 2022 may be providing a better indicator of a shifting employment market, and a look into what’s to come for the Q3 and Q4 QCEW reports.
For Q3 and Q4 2022, the household survey recorded average monthly job growth of 264k and 131k respectively, far less than the number recorded by the establishment survey of 423k and 291k respectively.
In terms of January 2023’s change in household survey employment, many may look at the whopping growth of 894,000, and say that even the household survey is now showing robust employment growth — again, not so fast. Just as the BLS adjusts the establishment survey in January, it also conducts adjustments on the household survey. The household survey is adjusted for changes in the BLS’ population estimates. Exclude this technical population adjustment, and what was January’s household survey job growth? Just 84,000.
Serious questions must be asked as to the accuracy of the monthly CES jobs report & lagging QCEW incorporation
We now have a situation whereby the establishment survey contained within the monthly CES jobs report has been benchmarked using QCEW data up to the March 2022 quarter, which resulted in 311,000 additional jobs ADDED to the monthly CES reports across 2022, whilst more recently released Q2 QCEW data points to a significant deterioration in the jobs market.
The shift to a weaker employment market is corroborated by the household survey, where monthly job growth averaged just 102,000 in the last nine months of 2022.
The combination of these factors suggests that serious questions need to be asked about the accuracy of the jobs data that is currently being produced by the CES’ monthly establishment survey.
We will get a further read into the state of the employment market when the Philly Fed releases its next EBR update on 14 March. The BLS releases its next BED report on 26 April. These reports will incorporate Q3 QCEW data, and will help to provide a clearer picture on whether establishment payroll data is being overstated as a result of being benchmarked to older QCEW data up to March 2022, and if so, to what extent.
If the next batch of QCEW data provides further weight to a slowdown in jobs growth, this would create a major issue, as this data would not be incorporated into the monthly CES jobs report until FEBRUARY 2024.
Just as the BLS’ lagging measure of rents in the CPI initially resulted in the CPI being understated, contributing to too lax of a response from the Fed on inflation, we may now be in a situation where the BLS’ lagging incorporation of comprehensive employment data into its monthly jobs report, risks providing an inaccurately strong picture of the jobs market.
With the Fed now placing greater attention on reducing the ‘strength’ of the employment market, this risks more significant overtightening from the Fed and pushing the US economy into a severe recession.
UPDATE: The Philly Fed’s latest EBR report, based on Q3 QCEW data, as well as revisions to prior QCEW data, has resulted in the divergence between the QCEW and the CES’ establishment survey being bridged.
Over the year to Q3 2022, the Philly Fed's EBR report based on QCEW data, estimates that 6,072,000 net new jobs were added, versus net job growth of 5,904,000 in the monthly CES establishment survey.
The two datasets thus no longer suggest a radically different picture of the employment market.
Despite this, employment growth in the household survey yet again materially undershot the growth in payrolls recorded by the establishment survey in February (177k vs 311k respectively).
Thank you for continuing to read and support my work — I hope you enjoyed this latest piece.
In order to help support my independent economics research, which aims to provide not just hedge funds & asset managers with access to institutional grade research & insights, but ordinary people, please like and share this article.
If you haven’t already, make sure to subscribe so that you don’t miss an update — it’s free!
I always thought Fed was always behind because they were incompetent, but maybe they are behind because their data lags behind so much. But then again they should know by now how far behind the data measurements are and they have done nothing to fix or compensate for it. Ok so they are still incompetent. (Somewhat joking)