The Federal Open Market Committee raised its benchmark rate a quarter point on Wednesday to a range of 3.75% to 4%, the first increase since 2023, and the coverage settled fast around two numbers: a 12-0 vote and a 4.1% median projection for where the federal funds rate finishes the year. Both are accurate. Neither means quite what it looks like, because the Summary of Economic Projections published alongside the decision carries eighteen interest-rate paths, and the Committee has nineteen participants. The nineteenth belongs to Chair Kevin Warsh, and he did not submit one.
That is not a footnote. The median every desk on Wall Street spent Thursday pricing is calculated from a set that excludes the one participant who writes the statement, chairs the meeting, and takes the questions afterward. Warsh has now withheld his dot at both meetings he has run.
The Projections Moved More Than the Rate Did
The quarter point was the least interesting thing in the release. Buried in the same table is a set of revisions that reframe what the Committee thinks it is dealing with. The median 2026 federal funds projection jumped to 4.1% from 3.8% in June. The 2027 median went to 4.1% from 3.6%, and the longer-run estimate drifted up to 3.2%.
The forecasts underneath those dots moved in an awkward direction. Median PCE inflation for 2026 was revised up to 3.7%, core PCE to 3.4%. Median unemployment was revised down hard, to 4.1% from 4.3%, and GDP growth up to 2.3%. Read together, that is a Committee telling you the economy is stronger than it thought and inflation is worse than it thought, which is the combination that produces a hiking cycle rather than a single insurance move. Sixteen of the eighteen participants who did submit projections see at least one more increase this year.
The statement itself made the same point in fewer words than the Fed has used in years. “Inflation remains elevated,” it read. “Today’s policy action will support a timelier return to the Committee’s 2 percent goal. The Committee will deliver price stability.” That last sentence is a promise, not a forecast, and it is the most hawkish line the Fed has published in this cycle.
Equities did not care. The Dow closed up 316 points on Thursday, the S&P 500 gained 1.1% and the Nasdaq rose 1.7%, a rally CNBC attributed in part to cooling crude prices rather than to anything the Fed said. Markets read a hawkish hike as the end of uncertainty. That is a defensible trade. It is also a trade made against a consensus figure the chair declined to join.
A Median Without the Decisive Vote
Warsh’s abstention is deliberate and he has explained it. CNBC reported in June that he was expected to withhold his dot before his first meeting as chair, and he has been consistent since.
“I, however, have refrained from offering any projections of my own, consistent with my long-held views.”
His objection has intellectual pedigree. Warsh has argued for years that published rate paths and forward guidance harden into commitments, and that a committee which has told the market what it will do in December finds it costly to do something else when November’s data says otherwise. That critique is not frivolous. Plenty of economists share it.
The problem is what he did instead of acting on it. He did not retire the dot plot. He kept publishing it, kept letting the median be computed and reported as the Committee’s collective intention, and removed only his own contribution. The result is the worst available version of both options. The grid still exerts its gravitational pull on rate expectations, still gets printed in every wrap, and still gets treated as the Fed’s plan, while the participant whose preferences matter most is absent from it. A projection exercise that excludes the chair is not a lighter-touch dot plot. It is a less accurate one.
The Vote Tally Nobody Can Check
The abstention did not arrive alone. Warsh has rebuilt the Fed’s communications from the statement outward, and CNBC’s redline of his second statement documented the scale of it: roughly 130 words against the 300-plus that was standard under Jerome Powell, forward guidance gone, and the named voting record gone with it.
That last removal is the one that should bother people, and it has gone almost entirely unremarked. Wednesday’s release says the Committee “approved the following statement for release by a 12-0 vote” and then names nobody. For decades the statement listed every member voting for the action and every member voting against, by name. It does not now. So when you read that this hike was unanimous, you are taking the institution’s word for it, and you cannot tell whether twelve people agreed or whether some number of them simply declined to spend capital dissenting against a chair who had already made up his mind.
Context makes that distinction load-bearing. July’s meeting drew three dissents in favor of hiking, the most this early in a new chair’s tenure in over half a century. Those dissenters got their hike in September. Unanimity that arrives immediately after the dissenters win is not evidence of consensus. It is evidence that the argument is over, which is a different thing, and the old statement format would have let a reader see which.
Less Transparency Is a Choice, and It Is the Wrong One
Here is where we come down. Warsh is right that forward guidance can trap a central bank, and he is entitled to run communications differently than his predecessor. He is not entitled to keep the authority that the Fed’s transparency architecture was built to legitimize while dismantling the parts that let anyone check the work.
Three things were removed in three meetings: the chair’s own projection, the forward guidance, and the named voting record. Each has a respectable individual justification. Together they add up to an institution that publishes a number markets trade on, calls it the Committee’s view, and has made it progressively harder to verify whose view it actually is. JPMorgan’s chief U.S. economist Michael Feroli put the direction plainly, saying the risks clearly tilt toward less transparency in Fed communications. He is being polite.
The fix is not complicated and it does not require Warsh to abandon his own position. If the dot plot is a useful signal, the chair should submit a dot. If it is a misleading artifact that constrains policy, he should take the argument to the Committee and retire it, which the communications task force he has convened could plausibly recommend. Publishing an 18-dot median while the 19th participant sits it out is the option that serves nobody, least of all the people being asked to price mortgages and corporate debt off it. And restoring the named vote costs the Fed nothing at all. An institution that is confident its decisions are unanimous should have no difficulty saying who was in the room agreeing.
What to Watch Next
We have argued before that Warsh was building his case for this hike on the higher of the Fed’s two inflation gauges, and Wednesday’s upward revision to core PCE is the number that vindicated him. He won the argument on the economics. The open question is institutional, and the December meeting is where it gets tested: if sixteen of eighteen participants are projecting another increase and the chair still will not say what he thinks, the market will be trading a consensus that the decisive voter has publicly refused to endorse. That worked this week because the hike was expected. It will work considerably less well the first time the Committee has to surprise someone. For readers who want the mechanics of how the grid became the market’s favorite number in the first place, our earlier explainer on the dot plot covers why it carries the weight it does, and the NPR account of the decision is a clean read on the inflation backdrop that forced it. The full FOMC statement runs shorter than this paragraph, which is rather the point.