Is the BLS Cooking the Books?

by Brian S. Wesbury – Chief Economistm & Robert Stein, CFA – Deputy Chief Economist, First Trust Portfolios

Last week Zero Hedge achieved the impossible, they managed to make a report from the Philadelphia Federal Reserve go viral.

The topic in question? How many jobs did the US economy actually gain in Q2 of 2022: 10 thousand or 1 million? Last night Tucker Carlson took up the torch as well, claiming that the Bureau of Labor Statistics (BLS) is cooking the books for the Biden Administration. Given the amount of attention this esoteric (and in spite of the viral nature of the story, it is esoteric) argument is getting, we think it’s worth addressing what is going on.

To begin, we do believe that the US labor market is clearly showing signs of a slowdown and we have been vocal about our expectations of a recession in 2023.  We just doubt that job growth was that weak last Spring.

The Philadelphia Fed isn’t usually a primary source for data on the labor market. Off the top of our heads we can’t ever remember reading a labor market report from them in the past. The two reliable and time-tested sources are BLS surveys of large businesses (the establishment survey) and surveys of individuals (the household survey).

Given that the Philadelphia Fed just started publishing results from its own internal model, meaning there isn’t much of a track record to go on, we’d say it’s probably right to take this report with a grain of salt. Theoretically, everything about this model might be sound, but even sound models often clash with reality.  For example, the ADP model has never lived up to its hype.

Also keep in mind that there are twelve Federal Reserve banks across the US. If all of them were to do similar proprietary reports on the labor market the results would probably have a wide range of outcomes, so this lone report shouldn’t be seen as representative.

So what did the Philadelphia Fed find? Using data on 33 states plus the District of Columbia from the Quarterly Census on Employment and Wages, a lagging report that is often used to make revisions to overall payroll data, they believe the payroll gains that were officially reported in Q2 were significantly overstated. Instead of 1.05 million jobs the US only added 10,500. That’s a downward revision of 99%!

But the report didn’t just include estimates for Q2 2022, it also included estimates for the prior two quarters which have received a lot less attention. Those estimates show a much smaller downward revision to Q1, but they also show a big upward revision to Q4! It’s a lot less sexy than an implied conspiracy but what we might be seeing here is just the result of seasonality not being fully accounted for in a relatively new model.

One of the problems with analyzing employment in 2022 is that there has been a large divergence between the “household survey” and the “establishment survey.” In fact, right now, before any revisions, the household survey shows that employment actually fell 347,000 in Q2.  So, if one wanted to take the Philadelphia Fed report number of 10,500 jobs gained at face value, you could just as easily make the argument that the BLS is cooking the books to make the Biden Administration look bad.

Given the divergence between data from the household and establishment surveys in 2022 we wouldn’t be surprised at all by downward revisions to the employment data when they are released in March of 2023. But we will definitely take the over on 10,500 jobs added in Q2.

Moreover, if there really was such a huge jobs slowdown in Q2 it should be reflected in other data.  But, the ISM employment surveys were relatively strong, while the ADP employment estimates (in spite of their drawbacks) never plummeted.  In addition, initial unemployment claims, the ASA staffing index and surveys of businesses (from the NFIB and elsewhere) reflected strong hiring desires. Add on the fact that the services sector, which was crushed by pandemic restrictions, has been roaring back in 2022 and we just don’t find this downward revision plausible.

A wide range of estimates on the labor market is somewhat expected given that pandemic related measures, from lockdowns to monetary policy, have distorted everything. However, we don’t think the BLS is trying to deceive anyone. When the labor market eventually weakens it won’t be this subtle.

Brian S. Wesbury – Chief Economist

Robert Stein, CFA – Deputy Chief Economist

Click here for a PDF version

Copyright © First Trust Portfolios