Scenes from January s hit jobs report 2 ,Today let me address the second big revelation in the January report: the 2021 and 2022 revisions and their effect on the large discrepancy between the numbers in the survey of households and establishments that began last April.
There were revisions to both the household and employment surveys.
Household revisions are easy to understand. Every year there is an adjustment for population growth. But instead of spreading that over the previous 12 months, the household survey adds it all at once to the January figure.
As a result, 894,000 more people were officially employed in January, according to the Household Survey. But 810,000 of those were caused by adding 954,000 to the population forecast for 2022. Only the remaining 84,000 of that was the change from December to January.
One way to reasonably estimate what the 2022 household survey would look like is to spread those 810,000 new jobs over the past 12 months, which comes out to 67,500/month. The chart below does just that, adding 67,500 to each number from the 2022 household survey and comparing it to the revised establishment numbers:
Scenes from January’s hit jobs report 2: Revisions don’t resolve reporting discrepancies
Second, there has been debate over the past few months about whether the household survey found weaknesses that were entirely absent from the establishment survey. Evidence of this was indicated by an estimate by the Philadelphia Federal Reserve, based on last year’s Q2 QCEW report, that only 11,000 private sector jobs were added during this entire 3-month period. This is important because the QCEW is the gold standard for employment reports, consisting of the actual total of 95% of all businesses through unemployment insurance payments. Subsequently, the Census Bureau itself updated its “Survey of Business Dynamics” to the second quarter of last year, which seasonally adjusted about 70% of the QCEW numbers and reported a total *loss* of -287,000 private sector jobs that year period.
To best illustrate this, here’s a chart from Prof. Menzie Shin of Econbrowser comparing the unrevised business survey reports with the household reports, QCEW earnings normalized to the same number in 2021 and applying his seasonally adjusted estimate, Philadelphia Fed forecasts and result of the survey of business dynamics:
Note that Prof. Shinn’s method of seasonally adjusting resumers is a very different result from both the Philadelphia Federal Reserve and the Census Bureau itself in the BDS.
Scenes from January’s hit jobs report 2:
So at first glance it’s very surprising that the establishment survey’s 2022 benchmark revisions *added* 457,000 jobs to previous results, and while the Q2 2022 numbers were revised down by -59,000, however, they showed 1,047,000 jobs created this quarter, as shown in the chart below:
“QCEW relies on unemployment insurance filings that provide a virtual count of payroll employment. The business survey is compared to the QCEW each year, but the comparison is made with January data, using QCEW data from the first quarter of the previous year. QCEW data from March 2022 has just been included in the business survey, which has increased employment growth in the year from March 2021 to March 2022 by 568,000.’
In other words, while the strong Q1 2022 QCEW data was included in the benchmark revisions, the weak Q2 data was not.
This is seen dramatically if we compare year-on-year changes in wage growth as measured by the QCEW census against the same measure from the establishment survey. This is the best way to measure it because the only drawback of the QCEW is that although it is comprehensive, it is not seasonally adjusted. But that doesn’t affect year-over-year comparisons.
Here’s the monthly year-over-year change in the QCEW survey through its latest Q2 2022 report:
And here’s the same thing for private nonfarm payrolls:
Comparing the QCEW report on a year-over-year basis through the first quarter of 2022 with private payrolls, the former grew by 7.2 million jobs and 5.1%, and the latter by 6.9 million and 5.6%. While the % change is still off, the total is close and equal to the variances of the past. The average monthly difference between the two measures is 0.3% and the median is 0.5%.
For the second quarter on an annual basis, QCEW rose by 5.7 million and 4.0%, while private wages after revisions increased by 6.3 million and 5.3%. That’s a very big difference. The mean difference is 0.8% and the median is 0.9%.
My expectation is that the difference will be resolved in favor of the much more comprehensive final QCEW figures once they are introduced. We will also get preliminary figures for Q3 2022 QCEW numbers later this month and that will tell us if the divergence has continued.
In short, even with the revisions, there’s a wide gap between strong corporate earnings in the last 9 months of 2022 versus the tepid household report. And the discrepancy between the ascertainment sample and the comprehensive QCEW and BDS jobs census data for the second quarter was not accounted for in the January revisions and remains.