February's producer price index came in at 0.7% month-over-month, double what economists expected. That number has become the centerpiece of every argument against cutting rates in 2026. But strip out energy, which surged on Iran-related oil disruptions that no central bank can control, and the print looks far less alarming. Kevin Warsh is right to signal that the Fed has room to cut, and the data, read carefully, supports him.
The fed funds rate sits at 3.50-3.75% after the March 18 hold. The Fed's own long-run neutral estimate just rose to 3.125%, the highest in a decade. That gap is roughly 50 basis points. For a homeowner refinancing a $400,000 mortgage, the difference between a 7.1% rate and a 6.6% rate is about $130 a month. Multiply that across millions of households and small businesses rolling over credit lines, and you see what "near neutral" actually costs in real terms. The Fed is not loose. It is barely restrictive.
The Number Everyone Is Misreading
Year-over-year PPI at 3.4% sounds like embedded inflation. It is not. Energy and services drove the February print, and energy prices are a geopolitical variable, not a monetary one. The Fed cannot drill for oil. What it can influence is the cost of capital for businesses investing in efficiency, and that is where Warsh's AI productivity thesis matters.
Warsh has drawn comparisons to the late 1990s, when the Greenspan Fed held rates relatively low while technology-driven productivity gains kept inflation subdued despite strong growth. The parallel is imperfect. But the direction is right. Corporate capital expenditure on AI infrastructure has been running at record levels for 6 consecutive quarters. If even a fraction of that spend translates into lower unit labor costs over the next 12 to 18 months, the inflation math changes.
I will grant the hawks one thing: declaring victory on inflation prematurely is exactly what the Fed did in 2021, and the cost was enormous. Jeffrey Schmid is correct that inflation has exceeded 2% for nearly 5 years. That track record demands caution.
But caution is not paralysis. The dot plot from March 18 projects one 25bp cut this year. That is not reckless. That is a rounding error on a $28 trillion economy.
What Warsh Actually Changes
The real Warsh story is not rates. It is the balance sheet. His "Sound Money" framework targets shrinking Fed assets from $6.5 trillion to $4 trillion, an aggressive pace of quantitative tightening that would drain liquidity from financial markets far more effectively than holding rates 25 basis points higher. This is the part the rate-cut panic crowd ignores. Warsh is not proposing easy money. He is proposing a different mix: slightly lower short-term rates paired with significantly tighter long-term financial conditions via QT.
That combination is coherent. Lower the overnight rate to relieve pressure on households and small firms. Simultaneously, shrink the balance sheet to prevent asset bubbles and keep long-term yields honest. J.P. Morgan's Michael Feroli has argued that the proposition rates are still restrictive "looks increasingly untenable." If he is right, holding at 3.50-3.75% is not prudence. It is overtightening by inertia.
Yale's William English doubts Warsh can secure votes for a series of cuts. Fine. He does not need a series. One cut in September, which 55 of 82 economists surveyed now see as the earliest plausible window, paired with accelerated QT, would signal discipline without inflicting unnecessary pain.
The market priced out 2 expected cuts between January and late March. That repricing reflects fear of an oil shock, not a sober reading of where monetary policy stands relative to neutral. Warsh's instinct, that productivity gains have shifted the inflation calculus, will be tested by earnings data over the next 2 quarters. I expect the numbers to back him up. A $130-a-month difference on a family's mortgage is not abstract. It is groceries, a car payment, a margin that matters. The Fed should cut once this year, and Warsh should make sure the balance sheet does the rest of the tightening.
