“He testified very strongly to that effect, and I’ll take him at his word.”

So said Jerome Powell last week, when asked during his final FOMC press conference as Fed chair whether he’s confident that Kevin Warsh is in fact committed to preserving institutional independence.

I doubt Powell believes Warsh when he says, as he did on April 21, that he’s “no sock puppet” for Donald Trump. Indeed, Powell’s determination to stay on as a Fed governor rather than retire this month, and his stated rationale for doing so, suggests he harbors serious reservations about the motivation for Warsh’s “regime change” program for the Fed.

Part and parcel of that effort entails looking at different inflation measures which, as Warsh put it, better show “the underlying, generalized change in prices in the economy.”

There’s nothing “wrong” with that, per se. And the metrics Warsh is referring to aren’t voodoo. All the Dallas Fed’s version does, for example, is rank the change in index components from smallest to largest, lops off outliers from both ends and takes the weighted average of what’s left.

Some worry, though, that Warsh isn’t actually committed to those metrics on principle, but rather prefers them currently because they could be used to justify rate cuts.

The figure below shows you core PCE (i.e., what the Fed prefers) and the Dallas Fed metric.

On the trimmed mean measure, we’re more or less at target on the eve of Warsh’s confirmation and first policy meeting as Chair. (“How convenient,” I know.)

By contrast, the latest read on 12-month core PCE was 3.2%, 120bps above target and the fastest in over two years (although I should note that the MoM pace was the slowest since November).

If Warsh sticks to his message — “I prefer looking at trimmed averages [where] we take out all of the tail-risks, all of the one-off items, and we ask ourselves whether the generalized change in prices is having second-order effects on the economy,” as he told Congress — consistently, even when the trimmed averages don’t line up well with Trump’s penchant for lower rates, fine.

But if he deviates — e.g., reverts to core PCE when it paints a more favorable picture for a president inclined to easier monetary policy — he’ll lose any credibility he manages to establish.

The discerning among you can write the punchline yourself by eying the first chart above. Have a look at the spread between core PCE and the Dallas trimmed mean measure going back to the GFC:

The undershoot for one version of Warsh’s preferred measures is a 2020s phenomenon. Traditionally, core PCE (what Warsh wants to deemphasize) is the cooler of the two inflation metrics.

Indeed, as BMO’s Ian Lyngen remarked, “trimmed mean PCE has run hotter than core PCE in 75% of the months since 2000.”

What would Chair Warsh do in the event the longer-term trend reestablishes itself at some point during his tenure?

I worry that’s a rhetorical question, and I’m not alone. As BofA’s US rates team put it late last month, “To avoid optics of cherry-picking, Warsh will need to stick with his preferred metrics even when they are outpacing [core PCE].”