Mr. Steven Kopits asserted that the Philadelphia Fed’s early preliminary benchmark supported a recession in 2022H1, to wit:

You, Menzie, held the Est Survey was more likely right. You wrote: So: (1) I put more weight on the establishment series, and (2) the gap between the two series is more likely due to increasing, and biased, measurement error in the household series, rather than, for instance, primarily increases in multiple-job holders.  https://econbrowser.com/archives/2022/12/the-household-establishment-job-creation-conundrum

Dead wrong, as it turned. And predictably so.

You were wrong because you did not consider the statistics more holistically. That’s the learning point for your students. Cross check your indicators if you have dials which are telling you different things. If jobs are increasingly rapidly, then GDP should also be up. If jobs are increasing rapidly, then mobility and gasoline consumption should also be up, because so many people need to drive to work in this country. Finally, if productivity is imploding when jobs are up, you really need to take a pause and put together some sort of narrative as to why that might be happening. It suggests something anomalous in the data which requires closer inspection.

Had you done that, Menzie, you might have concluded as did the Philly Fed…

What remains of that hypothesis? Well, today, the Philly Fed released an update. Putting together the official nationwide nonfarm payroll series published by the BLS, the sum of states from CES and the Philadelphia Fed early benchmark, we have the following picture.

Figure 1: Nonfarm payroll employment, FRED series PAYEMS (bold black), CES sum of states (tan), sum of states early benchmark by Philadelphia Fed (red), civilian employment over age 16 adjusted to NFP concept (teal), Quarterly Census of Employment and Wages total covered employment, adjusted by Census X-13 by author (pink), all in thousands, bseasonally adjusted. Light blue shading denotes hypothesized (by Mr. Steven Kopits) 2022H1 recession.  Source:  PAYEMS from BLS via FRED, civilian employment adjusted to NFP concept from BLS, QCEW from BLS, sum of states data from Philadelphia Fed. 

I would suggest that the various vintages of these data confirm the following propositions: (1) the CES series is more reliable for business cycle tracking than the aggregate employment series derived from the CPS; (2) the series are all revised over time (including QCEW).

If you see a recession in 2022H1, then contact me. Earlier vintages of these series have been displayed on Econbrowser repeatedly; yet I have not yet seen Mr. Kopits admit that perhaps he might have been hasty in his assertion that labor market data indisputably signalled a recession in 2022H1.