The logic behind trimmed mean is straightforward enough. The Fed already strips out noise by watching core inflation, which policymakers treat as a better predictor of where inflation is headed (Brookings, April 2026). Jerome Powell has said as much. Warsh is pressing that idea further by filtering out the most extreme price moves from both ends of the distribution.

The Dallas Fed’s trimmed mean PCE does that by dropping components below the 24th percentile and above the 69th percentile of the price-change distribution (Brookings, April 2026). In February, that meant excluding telephone and related communication equipment, down 50.8% annualized, and moving, storage, and freight services, up 384.6% annualized (Brookings, April 2026). Those are exactly the sort of outliers trimmed-mean measures were built to ignore. Central bankers do not need a monthly price report to become a circus act.

There is real evidence for the appeal. Dallas Fed research cited by Brookings finds that first-release trimmed mean PCE is a better tracker of inflation trends than first-release core PCE, both of which are revised later (Brookings, April 2026). Between 2005 and 2018, first-release trimmed mean averaged 1.78%, very close to the headline PCE average of 1.77%, while first-release core PCE averaged 1.61% (Brookings, April 2026). On those numbers, trimmed mean has earned a place at the table.

It may be especially useful now if recent gains in core goods prices are “primarily tariff-driven and a one-off price level change,” as Dallas Fed researchers put it (Brookings, April 2026). In that case, a measure that trims away the tariff shock can give policymakers a cleaner read on the medium-term trend. That is the strongest version of the Warsh case. If the shock is temporary, the cleaner gauge has a point.

But the same research comes with a warning label. Dallas Fed researchers say trimmed mean may miss “the beginning of broadening inflationary pressures” (Brookings, April 2026). The Cleveland Fed goes a step further, noting that median and trimmed-mean measures can be useful over long periods but can drift from underlying trend over shorter horizons (Brookings, April 2026). That shorter horizon is exactly where monetary policy lives.

There is another wrinkle that matters. Trimmed mean is not automatically the softer reading. Brookings notes that averaged over 2005 to 2018, Dallas Fed trimmed mean inflation ran higher than core PCE and was sometimes above headline inflation, sometimes below it (Brookings, April 2026). It is lower today because the distribution of price changes has shifted. That is a useful distinction. The measure itself is not built to flatter anybody.

Warsh is walking into an institution that is already rethinking its playbook. Brookings notes that the Fed has promised to review its “monetary policy strategy, tools, and communication,” and that the current framework was introduced in August 2020 after the experience of the 2007-2009 financial crisis and recession (Brookings, July 2024; Brookings, September 2024). This is the institutional opening for a measurement debate. It is not just a Warsh talking point in a tie.

The 2020 framework was built for a world in which inflation had run below 2% for years and the effective lower bound was the main problem. The Fed said it would aim for inflation that averages 2% over time, tolerate inflation moderately above 2% for some time after undershoots, and keep rates near zero until both employment and price stability had been met (Brookings, September 2024). That was designed to solve a low-inflation problem. It looks less graceful after the inflation surge of 2021 and 2022.

The Brookings symposium makes clear that reasonable economists still disagree about the right fix. One camp argued the 2020 framework was too tied to the post-crisis, low-rate environment and should become more general and flexible, with clearer room for the Fed to respond when inflation is rising before employment is fully maximized (Brookings, September 2024). Another argued the framework is sturdy enough, and that the Fed should preserve the flexibility it may need if rates hit the zero lower bound again (Brookings, September 2024). Warsh’s trimmed-mean preference fits squarely inside that argument. It is about measurement, but measurement is policy in a necktie.

The important question is what form any shift would actually take. If trimmed mean shows up as a supplement, alongside core PCE, it is one thing. If it becomes the main lens through which officials explain inflation, it is another. And if it starts driving rate decisions, then the framework has changed even if nobody at the podium says so out loud. The Fed has a long history of pretending those distinctions are decorative. They are not.

Warsh would have influence, but not a free hand. The chair sets the agenda, frames the debate, and gives the institution its public voice. The full Federal Open Market Committee sets rates.

That matters more than it sounds. Warsh cannot personally rewrite the Fed’s inflation benchmark any more than he can unilaterally cut the funds rate. He can, however, elevate trimmed mean in speeches, press conferences, and internal discussions until it begins shaping how other members read the data. Brookings notes that the third paper in its framework review studied Fed announcements, minutes, and speeches to see how the 2020 framework affected policy choices (Brookings, September 2024). That is the right frame here too. The written statement is only part of the story.

The hard test is consistency. If Warsh cites trimmed mean when it points toward easier policy and falls back on core PCE when it does not, that is not a measurement debate. It is a credibility problem with charts. If he uses the same measure in good times and bad, then the argument becomes a real framework choice.

The chair’s public remarks will tell the tale. So will the Summary of Economic Projections, where the Fed’s quarterly dot plot lives (Brookings, July 2024). If trimmed mean starts showing up there, or in formal FOMC language, the shift is serious. If it remains mostly in chair-level rhetoric, the market should treat it as a preference, not a regime change.

The case for trimmed mean is strongest in the abstract. The timing is what complicates it. Brookings says that switching to a new metric after inflation has been above the Fed’s 2% PCE target for five years could look like moving the goal posts and could threaten the central bank’s credibility (Brookings, April 2026). That is the whole problem in one sentence.

The Fed’s own framework history reinforces the point. The 2012 inflation target was defined as symmetric, meaning inflation above and below 2% both mattered (Brookings, September 2024). The 2020 framework shifted toward average inflation targeting and said the Fed would aim for inflation moderately above 2% for some time after undershoots, but it made no equivalent promise to make up for overshoots (Brookings, September 2024). That asymmetry made sense when inflation was too low. It looks less tidy after a period of high inflation.

Krishna Guha of Evercore, as cited by Brookings, adds the operational warning. Trimmed-mean and median measures could “systematically and substantially understate the inflation experience for a period” if tariffs or higher oil-related costs are being filtered out (Brookings, April 2026). That does not make the measures wrong in every setting. It does mean they can be misleading exactly when a central bank most needs to know whether inflation pressure is broadening.

The politics make the timing even touchier. Warsh was nominated by a president who has publicly called for rate cuts. Reuters reported in February that investors were already ramping up bets on a steeper yield curve under a Warsh-led Fed, expecting lower short-term rates and a smaller balance sheet of about $6.59 trillion (Reuters, February 2026). TD Securities’ Oscar Munoz said Warsh’s past criticism of the Fed and his notable shift in policy priorities could alienate some members of the policy-setting committee (Reuters, February 2026). That does not disprove his argument. It just means the argument will be judged in a political light whether he likes it or not.

If Warsh gets the job, the debate moves from theory to signals. Reuters reported in February that investors were already positioning for a steeper Treasury curve under a Warsh-led Fed, with higher long-dated yields and subdued short-end yields as the Fed cuts rates while shrinking its balance sheet (Reuters, February 2026). That matters because long-term yields feed into mortgages, corporate bonds, used loans, and equity financing. The borrowing cost story will not stay trapped on Constitution Avenue.

Four things will tell whether trimmed mean is becoming more than a talking point. First, does Warsh present it as a supplement or the main inflation signal? Second, does he use it consistently, or only when it supports easier policy? Third, does it show up in formal FOMC language, or only in chair speeches? Fourth, when core PCE and trimmed mean diverge, which number seems to carry the discussion?

Warsh still has to be confirmed by the Senate, and Powell’s term ends in mid-May (Reuters, February 2026). Until then, the core issue is simple enough. Trimmed mean is a defensible way to think about inflation. The harder question is whether the Fed can adopt it now without looking as if it changed the answer to fit the moment.