Warsh at Sintra: The Chair Who Refuses to Guide

Kevin Warsh's message at Sintra was that anyone waiting for the Fed to loosen policy "would be disappointed." The interesting part is not the hawkishness. It is that he delivered it without a dot, without forward guidance, and without a domestic venue — a deliberate rewiring of how the Fed talks to markets, using the ECB's stage to do it.

The communication is the policy. Warsh has done two things since taking the chair that would count as procedural footnotes under Powell and are, under him, the substance. He has stopped submitting a personal projection to the Summary of Economic Projections. And he has shortened the FOMC statement while cutting the volume of forward guidance around it. Neither move changes a rate. Both change how the market forms expectations about rates.

The Powell-era loop ran like this. The Fed guided; markets positioned into the guidance; positioning became a fact the Fed then had to either validate with a decision or fight with a correction. Guidance produced its own gravitational pull, and by the last two years of that regime, the FOMC was frequently managing the consequences of its own communications rather than the economy. Warsh is breaking that loop by refusing to feed it.

The Sintra choice was not incidental. President Trump has been publicly calling for cuts for months. Warsh, his own appointee, used a European central-banking forum, sitting between Lagarde, Bailey and Macklem, to say the opposite in front of cameras. This is not scandal. Central bank independence is designed for exactly this friction. But the venue is load-bearing.

A domestic press conference is politically exposed in a way an international panel is not. At Sintra, Warsh is on a stage with peers who are also fighting inflation, framed as one of four central bankers doing the same job. The image absorbs the political charge. If he had said the same words at a Washington podium, the next 48 hours of coverage would be about the president's response. Instead, the coverage is about monetary policy. That is a communications choice, and it worked.

The hawkish frame has soft edges, and they matter. Warsh did not stop at "disappointed." He also said inflationary risks have "eased" since the last FOMC meeting, and he opened the door to AI as a deflationary force — the first sitting Fed Chair to entertain that thesis as policy-relevant rather than as a Silicon Valley pitch.

A chair who genuinely believed cuts were off the table for the foreseeable future does not volunteer either qualification. The market seems to have read it that way.

That is not the yield of a market that believes cuts are permanently off. It is the yield of a market that believes cuts are not imminent, which is a much narrower claim than the headline suggests. If Warsh had truly closed the door, the 10-year would have moved harder. It did not, and that tells you what the professional bid actually heard.

The AI comment is the underreported story. Everything else Warsh said was about near-term rates — the sort of thing central bankers say at every forum. The AI-as-disinflation remark is about the structural inflation regime, which is a different order of question.

If the Fed formally begins to model AI adoption as disinflationary, three things move. The estimate of the neutral rate, the rate that is neither stimulating nor restraining the economy, drifts lower. The terminal rate for the current cycle drifts lower with it. And the pace of future normalisation gets faster once the disinflation shows up in the data, because the Fed will trust that the trend is structural rather than cyclical.

The catch is timing. AI as a disinflationary force depends on adoption, not capability. The models are capable now; the price effects only show up when enterprises actually deploy them into the workflows where labour cost sits. That curve bends when it bends. Warsh's "more open-minded" language is the Fed telling markets it is watching for the bend — not that it has seen it.

The Fed is now paying attention to when the AI adoption curve bends, and that is a change in what monetary policy considers within its field of view.

Bailey's warning shares the stage for a reason. On the same panel, Andrew Bailey flagged rising leverage in equity ETFs and hedge funds while equity indexes sit at second-quarter records. Read the two messages together and the shape of the risk becomes clearer than either delivers alone.

The rate-cut door is being held closed. Leverage is accumulating against suppressed volatility. Indexes are at highs. If the inflation data shifts sharply in either direction, the unwind path is narrow. That is not a prediction of a disorderly move — it is a description of the conditions under which one becomes more likely. Bailey named it because someone on that panel needed to, and Warsh's brief made it awkward for him to do so.

The counter-case, honestly. The strongest argument against reading Warsh as structurally hawkish is that his own communication reforms cut both ways. Less forward guidance means less credible commitment to any specific stance, including the current one. A chair with no published dot can reverse course more easily than a chair whose projection is a matter of record. The hawkish rhetoric may be posture precisely because the machinery to enforce it has been dismantled.

I think that argument is real but incomplete. Warsh has not removed the Fed's ability to commit; he has removed its habit of committing in advance of the data. Those are different. The first is a loss of policy tools; the second is a return to a mode of central banking that predates the forward-guidance era and worked fine. Whether it works now depends on whether markets can hold positions without a schedule of Fed communications to lean on. That is the experiment Warsh is running.

The result at Sintra is a Fed that talks less, promises less, and asks markets to do more of their own work. Whether you read that as discipline or as vacuum depends on how much you trusted the previous regime's guidance to mean what it said. The market, at 4.474% on the 10-year, is reserving judgment. So am I.

Footnotes