The Nominee

This morning, President Trump nominated Kevin Warsh to lead the Federal Reserve.

Warsh is not a typical nominee. He served on the Board of Governors from 2006 to 2011—through the financial crisis, through the bailouts, through the first rounds of quantitative easing. He was inside the room. Then he left, and he's spent the years since arguing that the institution he once served needs "regime change."

Not a new chair. Not new faces. A new intellectual framework.

In July 2025, Warsh sat for an interview on CNBC where he laid out his indictment. Five specific claims about what's broken at the Federal Reserve:

This is not the usual criticism. Politicians complain about rates being too high or too low. Warsh is making a deeper claim: that the Fed's entire analytical apparatus is obsolete. That the models are broken. That the institution cannot see what's in front of it.

Is this just politics? Or is there substance?

These are testable claims. The Federal Reserve has been documenting its internal deliberations for ninety years. Staff memos, committee transcripts, internal forecasts—a paper trail that reveals not just what the Fed decided, but how it thought, what it knew, and what it chose not to say publicly.

The FOMC Insight Engine contains over 280,000 searchable passages spanning 1936 to 2025. We tested each of Warsh's claims against this documentary record. What we found was more damning than we expected—not because the Fed made mistakes, but because the documents reveal a pattern of knowing and ignoring that stretches back decades.

The Disappearance of Money

Warsh's first claim: "The theory that money has something to do with monetary policy is nowhere in the central thinking at the Fed."

This sounds hyperbolic. The Federal Reserve is, after all, a central bank. Surely they think about money?

We searched for how the Fed discusses monetary aggregates in its inflation analysis. What we found was not a gradual drift away from monetarism, but a deliberate abandonment—and a careful effort to obscure the real reason why.

The public story is simple: financial innovation broke the relationship between money supply and inflation. Greenspan told Congress in 1993 that "the historical relationships between money and income, and between money and the price level have largely broken down." Case closed. The aggregates are unreliable. We moved on.

But the internal documents tell a different story.

The Fed didn't just discover that money velocity had become unstable. They chose to prioritize interest rate stability over monetary control—and then framed that choice as a technical necessity.

This is the smoking gun that never made it into the public explanation. The Fed could have continued targeting money supply. But doing so would have required accepting volatile interest rates—and they weren't willing to do that. So they chose to stabilize the price of money rather than the quantity of money.

That's a policy choice, not a technical breakdown.

The warning came even earlier. In 1976, Governor Wallich identified exactly what was happening:

Wallich saw the drift fifty years ago. The interest rate was supposed to be a tool for controlling money. Instead, it became the target itself. The tool became the objective. And the quantity of money—the thing that actually determines the price level over time—disappeared from the analysis.

By 2022, when M2 had exploded by $6 trillion in two years, Chair Powell explained the resulting inflation by pointing to "strong demand" and "supply side blockages." No mention of money. The quantity theory—the oldest idea in monetary economics—had been completely purged from the Fed's operational vocabulary.

Warsh is right. The disappearance wasn't an accident. It was a choice.

The Models They Can't Abandon

Warsh's second claim: "The dominant models at the Fed think inflation comes if the economy runs too hot and workers get paid too much."

He's describing the Phillips Curve—the idea that there's a stable tradeoff between unemployment and inflation. Low unemployment means high wages, and high wages mean inflation. The model dates to 1958. The Fed's version, built around the concept of NAIRU (the Non-Accelerating Inflation Rate of Unemployment), has been the "workhorse" of their forecasting apparatus for decades.

Governor Laurence Meyer made no apologies for this in 1997:

The problem is that the model kept failing. Unemployment would fall, and inflation wouldn't rise. The predicted wage-price spiral never materialized. By the 2010s, the relationship had become so weak that staff began calling it "the flattening of the Phillips Curve."

But here's what's remarkable: they kept using it anyway.

Read that again. The staff isn't saying the Phillips Curve works. They're saying they don't have anything better. The model persists not because it predicts well, but because the institution lacks the intellectual flexibility to replace it.

Some inside the Fed saw this clearly. In 2018, Minneapolis Fed President Neel Kashkari delivered a blunt verdict in a committee meeting:

Broken. Not "imprecise" or "noisy"—broken. The primary instrument the Fed uses to diagnose inflationary pressure doesn't work, and the president of a regional Fed said so explicitly in a committee meeting.

But it gets worse. Some dissenters argued that the model isn't just weak—it's backwards. The causal direction is wrong.

Plosser is saying the Fed has cause and effect reversed. Wages don't cause inflation; inflation causes wage demands. The wage-price spiral is a symptom, not a disease. If Plosser is right, the Fed has been misdiagnosing the economy for decades—seeing inflation pressure where there is none, and missing it where it actually exists.

As late as February 2025, Governor Adriana Kugler was still defending the framework: "Tightness in the labor market means there is little room to expand demand without putting upward pressure on prices." The model that Kashkari called "broken" seven years earlier remains the operating system.

The Arithmetic of Blindness

Warsh's third claim: "We could be at the front end of a productivity boom. And if I were the president, I'd be worried that they might not see it and they might think economic growth is somehow going to be inflationary."

This is the most technical of Warsh's arguments, but also the most devastating. He's saying the Fed's models can't distinguish between two very different kinds of growth: demand-driven growth (which can be inflationary) and supply-driven growth (which is often deflationary).

If the economy grows because consumers are spending more, that's demand. Prices rise. If the economy grows because technology makes production more efficient, that's supply. Prices fall. The two look similar in GDP statistics but have opposite implications for inflation.

Can the Fed tell them apart?

We found a 1999 exchange where Chairman Greenspan himself admitted the models are structurally biased:

This is the Fed Chairman explaining, in plain language, that the models are arithmetically biased toward seeing inflation when productivity accelerates. The coefficients in the model assume productivity grows at a certain rate. When it grows faster—during a technology boom, for example—the model interprets that as excess demand. It sees inflation where there is none.

Greenspan went further in October 1999:

The models don't distinguish supply from demand because, historically, they didn't need to. But what happens when they do need to? What happens during a genuine productivity revolution—like the one Warsh suggests AI might bring?

The staff's answer, from the same period, is revealing:

The possibility that productivity could decouple growth from inflation was explicitly excluded from the Fed's baseline forecast. Not because they had evidence against it, but because the framework couldn't accommodate it. The model became a filter that systematically screened out supply-side explanations.

The result is a structural hawkish bias. During productivity booms, the Fed sees "overheating" where there is actually expansion. Staff forecasts consistently predict inflation that doesn't arrive. Years later, they admit to "one-sided errors"—always in the same direction.

David Stockton, Director of Research and Statistics, confessed in 2003:

Consistent errors. Always in the same direction. For years. This isn't bad luck—it's a broken model producing predictable failures.

The Markets Saw What the Fed Couldn't

Warsh's fourth claim: "Last September, the Fed cut interest rates 100 basis points... And what did the bond market do? Long rates went up 75 basis points. That's a credibility crisis."

When a central bank cuts short-term rates, long-term rates usually fall too. Markets expect lower borrowing costs in the future. But in September 2024, the opposite happened. The Fed cut, and long rates rose. Markets were saying: we don't believe you'll keep inflation controlled.

This phenomenon isn't new. We found the Fed documenting it over thirty years ago:

Parry described this as a "bind"—a trap where the Fed's own actions undermine its goals. Ease policy, and markets price in future inflation. The accommodation you're trying to provide gets negated by rising long-term rates.

Why does this happen? The documents reveal a clear answer: markets expected the Fed to "cave."

This is the Vice Chairman of the Federal Reserve, in an internal meeting, admitting that the public doesn't believe the Fed will hold the line. They expect the institution to prioritize short-term political comfort over long-term price stability. And that expectation gets priced into bond markets immediately.

What's striking is how the Fed has handled this credibility problem institutionally. In internal discussions, they talked about "caving" and "political pressure"—visceral language about resolve and commitment. But in public communications, this got reframed as a technical problem of "anchoring expectations" through "forward guidance."

James Bullard saw the flaw in this approach. In 2011, he warned that forward guidance would face a "time-inconsistency" problem:

Bullard predicted exactly what happened in 2013 (the "Taper Tantrum") and again in 2024. The Fed makes promises about future policy. Markets don't believe the promises because they know the Fed will have incentives to break them. The guidance fails.

The deepest insight came from Philadelphia Fed President David Eastburn in 1980:

Credibility isn't about communication strategies or framework announcements. It's about deeds, not words. Forty-six years later, the Fed is still learning this lesson.

The Framework That Failed

Warsh's fifth claim: "This change in their definition in 2020 is one of the key reasons why we had the great mistake, the great inflation."

He's referring to the Fed's 2020 framework revision, which introduced "Flexible Average Inflation Targeting" (FAIT) and redefined maximum employment as "broad-based and inclusive." Under the new framework, the Fed committed to letting inflation run above 2% to make up for past shortfalls, and to focusing on employment distribution rather than just aggregate levels.

Critics argued at the time that this created an asymmetric reaction function—a bias toward accommodation. The Fed would wait longer to tighten because they wanted to see "realized" inflation, not just forecast inflation. And they would hesitate to raise rates as long as any segment of the labor market remained weak.

We searched for how this played out internally. What we found was a pattern we'd now seen four times: warnings issued, warnings ignored, failure arrives, retrospective admission.

The warning came in August 2019—before the framework was adopted:

Staff warned that "makeup strategies"—the core of FAIT—could make things worse if a supply shock hit while the Fed was trying to compensate for past low inflation. They would be caught between crashing the economy and letting inflation run. This warning was not included in public communications about the new framework.

When the Committee adopted the framework in 2020, Governor Lael Brainard made the asymmetry explicit:

"Take care not to withdraw support preemptively." The Fed was deliberately disabling its preemptive toolkit—the ability to act on forecasts rather than waiting for realized inflation. This was the same institution that had prided itself on "taking away the punch bowl before the party gets going."

Governor Michelle Bowman raised concerns at the time:

She worried the framework would constrain flexibility—that the Committee would feel bound by its own language even when circumstances changed.

Then inflation arrived. And the Fed waited. "Transitory," they said, for eighteen months. By November 2021, Cleveland Fed President Loretta Mester was warning that the word was no longer useful. But the framework had created a procedural and intellectual bias toward inaction.

The retrospective admission came in 2023. Governor Bowman—who had warned three years earlier—delivered the verdict:

A sitting Federal Reserve Governor said, explicitly, that the 2020 framework "contributed to the delay" that produced the inflation. Not external factors. Not supply chains. The framework itself.

Staff had warned in 2019. Bowman had warned in 2020. The inflation arrived in 2021. The admission came in 2023.

The pattern held.

The Pattern

Five claims. Five investigations. Five confirmations.

But the real finding isn't that Warsh is right about any individual critique. It's the pattern that connects them.

The Question

Kevin Warsh says the Federal Reserve needs "regime change." Not just new leadership, but a new analytical framework—one that takes money seriously, that can distinguish supply from demand, that earns credibility through performance rather than pronouncements.

The Fed's own documents suggest he's diagnosing a real institutional pathology. The warnings were there. The dissenters spoke up. The failures were predictable. And the pattern keeps repeating.

Whether Warsh can fix it—whether anyone can fix it—is a different question. Institutions are hard to change. The models are embedded in the staff's training, the forecasting apparatus, the policy rules. The culture filters out heterodox views. The incentives favor consensus over dissent.

But the diagnosis appears sound.

Five claims. Five confirmations. Ninety years of documents.

The question now goes to the Senate.