Bloomberg has attempted to measure the accuracy of policy signals from various Fed officials. Their measure of reliability is based on a survey of 30 Wall Street economists. So it arguably reflects a sentiment survey, rather than a measure of actual performance by Fed officials, which would be tricky to obtain. But from the perspective of Bloomberg, at least, it is objective. Points for that.
Apparently, Yellen is the most reliable, which is not surprising, as she is the most powerful member of the FOMC and tends to communicate the will of the committee, as she reads it, as well as her own tilt relative to that.
Fischer and Dudley are the next most reliable, presumably for similar reasons, although I find it odd that Fischer ranks as highly as Dudley. Maybe the respondents are impressed by pedigree and position, which would not be unusual. In my view, Fischer has been quite unreliable, not just chronically too hawkish, but with poorly-timed deviations as well.
George and Lacker are judged extremely unreliable, which makes perfect sense to me. And Kashkari, who is more dovish, is judged less reliable still.
With Kashkari, who is new, this may be the result mostly of small sample size. With regression to the mean, he may have trouble staying as unreliable as George and Lacker. As I see it, they have a much more robust and replicable process. They can’t always be wrong, but they can always be irrelevant.
What I found odd, and quite typical of the quality of Bloomberg’s Fed watching, is that they attribute no significance to their own research. Here is the money passage in that regard, with emphasis added:
Respondents rated Fed officials on two scales: hawk versus dove, and their reliability in signaling future policy. Economists ranked the chair, historically the Fed’s most powerful member, as the best policy signal, and she was generally followed by relatively dovish permanent voters. Officials have kept rates low for years, which could explain why proponents of easy policy dominated the ranking. The Fed appears to be on the cusp of a more aggressive hiking cycle—so that could be poised to change.
That passage would make sense if there were no such thing as “reliability”, but only the illusion of it created by the perma-doves being right when rates are falling or failing to rise and the hawks being right otherwise.
But if Bloomberg believed that, then what was the point of the story on accuracy? If there is actually such a thing as “reliability”, then the results of Bloomberg’s attempt to measure it favors the dovish take, the exact opposite for what Bloomberg concludes.
It would be like me saying, the survey says I am ugly, but it will be revised as my own sense of handsomeness eventually plays out. Helpful.
I conclude with a caveat, which is not central to my main criticism here but is important for other reasons. It is the Fed’s job to manage aggregate demand properly, to the extent they can. It is not really their job to be “reliable” in their interest rate signal. Their own case for rates signaling went away at least two years ago.