Last month, I reported on a relatively new recession probability indicator (the "markov switching" series recently introduced to the Fed FRED/Blytic) that was giving a pretty clear, though preliminary, indication of probable recession.
While I noted that the series was highly revised, I pointed out that even taking into account the revisions, the series was giving a recession signal since using just the "maximum" reported values (values that had been all been revised lower). The reporting 20% probability was very unusual, and typically associated with oncoming trouble.
In the latest release of the data, we find that not only has the September (there is a lag) value come in at a relatively low level of 2.94% probability of recession, but the August number has now been revised down from 19.6% to 3.8%. It's important to note, though, that the point of my prior post was to highlight just the "maximum" reported values. While the latest release revises down the 19.6% and reports an additional low probability for the latest month, it makes no difference. The fact remains that this series has not given such a significant overestimate of recession without there being a probable recession ahead.
Clearly, there could always be a first time -- this is just estimated data. But the prior 19.6% reported figure clearly argues for following this series very closely in the coming months.