I may have been the happiest person alive yesterday. Why you ask? Well, The Bureau of Labor Statistics had to finally admit that the employment data they published monthly for all of 2023 (likely through the rest of this of this year as well) was wayyyyyy off! How off you ask? Well, just under 70,000 jobs per month were
overstated in the 12 months ended March, 2024. Employment data will be
reduced by 818,000 jobs in the January, 2025 benchmark revisions. This was the second largest downward revision in history. The only one larger was after the great financial crisis, where business bankruptcies and bank failures were well above historical norms and our recession was much longer than typical. This was the largest overstatement during a supposed "booming economy." Remember "Bidenomics?"
Yes, I am going to gloat a little because I have taken some abuse on this subject matter and my opinion on the actual health of the U.S. economy over the last year! I first started pointing out the labor data discrepancies in in July, 2023. I wrote a blog post back on July 11, 2023, that is still available for your review on this site, entitled "Getting Harder and Harder to Believe." It was all about the large discrepancy between government data and private sources of labor data. I then followed up with a series of posts that pointed out more labor and GDP discrepancy entitled "Bidenomics or BS?".
Why do I mention this today? Well, because the window is still open. By that I mean, this revision to the labor data is a revision to a coincident indicator, one that is used by NBER to officially date the beginning and end of recessions. The chart below is busy I know, but it is starting to bring into question the theory of a Fed engineered "soft landing."
The light blue line I call the window. It is the Fed's recession probit model based on the spread between the 10 year treasury and the 3 month treasury yields. It tends to lead recession by 12 months on average. This line is pushed out to September of 2025 which incorporates the typical lead time. That means the economy is still vulnerable to recession until that window closes. The dark blue line is a model used by the St. Louis Fed that measures smoothed recession probabilities today, based on NBER coincident data. It shows that as of June, we are still not in recession. The maroon line is a probit model I built that marries yields and NBER coincident indicators together. As you can see, the maroon line is now above levels reached in both the 1990 and 2001 recessions. That brings into question the soft landing scenario, especially given the really bad news on job creation data yesterday.
I bring this up because if the Fed thinks they have engineered a soft landing and only cut rates in 25 bps increments for the next few meetings, it likely won't be enough. Stay tuned and thanks for letting me gloat a little!