Core PCE inflation printed at 2.8% year-over-year in March. The Fed's target is 2%. The gap between those 2 numbers has persisted for nearly 5 years. And Kevin Warsh, the man about to take the chair, wants to cut rates into it. I think that is a policy error with a familiar shape: the Fed declaring victory before the war is over, exactly as it did in 2021 when it called inflation "transitory" while prices were already running away.
Warsh's argument is elegant. AI-driven productivity gains will suppress unit costs, allowing the economy to grow without generating inflationary pressure. Lower rates grease the wheels. Everybody wins. The problem is that this thesis is unfalsifiable in the present tense. It asks you to cut rates now based on productivity gains that might materialize later.
The Greenspan Parallel Cuts Both Ways
Warsh has invoked the late 1990s, when Alan Greenspan held rates relatively low while tech-driven productivity kept inflation quiet. Fair enough. But Greenspan had something Warsh does not: a core inflation rate that was actually falling. In 1997, core PCE was running at 1.9% and trending down. Today it is 2.8% and sticky. The Greenspan Fed was accommodating a disinflationary trend already visible in the data. Warsh is asking the Fed to accommodate a disinflationary trend he believes is coming.
That distinction matters enormously. Greenspan also had no oil shock. Warsh inherits one. The Iran conflict has driven energy prices high enough that Minneapolis Fed President Neel Kashkari cited them on March 3 as a reason to gather more data before cutting. Energy shocks do not stay in the energy column. They bleed into transportation costs, food prices, and services. The February PPI print, 0.7% month-over-month, double expectations, was a preview.
Consider a family earning $85,000 a year. If inflation stays at 2.8% instead of falling to 2%, that 0.8-point gap costs them roughly $680 annually in lost purchasing power. A 25-basis-point rate cut might save them $40 a month on a variable-rate loan. The math does not balance. The inflation cost exceeds the rate relief.
What the Market Already Knows
In January, markets priced 2 rate cuts for 2026. By late March, that expectation collapsed to 1 or 0. The bond market is telling you something. It does not believe the productivity story yet. J.P. Morgan's Michael Feroli has gone further, projecting no cuts in 2026 and a 25-basis-point hike in Q3 2027. When J.P. Morgan is pricing hikes and the incoming Fed chair is signaling cuts, someone is wrong.
Yale's William English, a former Fed official, doubts Warsh can secure votes for a series of cuts even if he wants them. That institutional friction is the best thing the Fed has going for it right now. Kansas City Fed President Jeffrey Schmid has opposed cuts outright, noting the 5-year inflation overshoot. Governor Miran has already dissented once. The FOMC is not a rubber stamp.
I will grant one thing to the productivity optimists: corporate AI capital expenditure has been running at record levels, and if those investments pay off, the inflation calculus genuinely shifts. That is a real possibility. But monetary policy cannot be set on a possibility. It has to be set on what the data says today, and today the data says 2.8%.
The Fed's post-2021 mistake was not that it failed to see inflation coming. The transcripts show several governors flagged the risk. The mistake was that the institution chose to act on a forecast instead of acting on the data in front of it. Warsh is constructing the same trap with better vocabulary. "AI-driven productivity" is 2026's version of "transitory." It might even turn out to be true. But if it doesn't, and you've already cut into a 3% inflation floor with oil prices spiking, you are chasing your own tail with rate hikes 12 months from now. The Fed should hold through 2026. Every meeting. No cuts. Let the productivity thesis prove itself in the numbers before you bet the credibility of the institution on it.
