The Ridiculous Error Rates of the Monthly New Jobs Reports

The BLS released it latest jobs numbers today; the expectations and the initial reports are extremely consistent in their wild inaccuracy

Photo by Markus Winkler on Unsplash

This morning, another positive jobs report came out from the BLS. They’ve reported 303,000 new jobs created in March, the highest initial report since September. Unemployment also went down a tick to 3.8%. The CNBC crowd is doing cartwheels this morning and every one is touting a great economy.

For a while, I’ve been flummoxed by the reports and the seemingly consistent revisions downward as the BLS adjusts their reports, issuing a first and then second revision for each month subsequently. In other words, in this March report they also revise the figures from February and January and do so each time.

This morning I decided to get to the bottom of the change in jobs numbers from those revisions. I went to the BLS site and searched each month to find the initial figures reported and the subsequent second revision for each month from February 2023 to January 2024 (the last month for which we have the second revision numbers currently). The results are shocking:

The BLS revises its own figures to the tune of anywhere from an 8.5% to 50% error rate within two months. And almost always downward.

There is therefore very little value to the initial reporting of new jobs created, nor the expectations set by Wall Street before the report. Yet invariably the new report and its comparison to expectations become breathless headlines. For example, Business Insider writes this morning, “the latest jobs report is the Fed's holy grail: strong job market with non-inflationary growth."

But here’s what the error rate for Wall St’s expectations look like:

So, the expectations from Wall Street for the monthly jobs report are almost always lower that the report and have an error rate as high as 100%; likewise, the BLS’ own initial new jobs report has an error rate as high as 50%.

And yet, each month all major news outlets and political types flail wildly back and forth with their predictions, analyses and reactions to wildly inaccurate data that has little relation to reality. It’s the silliest game going. Don’t believe a word of it.