The Inference

Kevin Warsh held his first press conference as Chairman of the Federal Reserve on June 17, 2026, three weeks into the job. The Committee held the federal funds rate at 3.5 to 3.75 percent and reaffirmed its policy of ample reserves. The written statement was shorter than the one it replaced and, in the chair's words, dispensed with some older language; forward guidance was gone. The chair submitted no projection of his own — the nineteenth voter, who abstained — and announced five task forces to take the conduct of monetary policy back to first principles.

A reader watching only the rate decision would record a hold and move on. The meeting was not, in its substance, about the rate. It was the first installment of something the rest of this series had already described, and it followed the line so closely that the line is the only way to read it. These are testable claims.

The Line to This Room

The June meeting did not arrive without a record. For five months this series followed a single figure and a single argument. The Dissenter set out the chair as the skeptic who distrusted the Federal Reserve's models and the confidence it placed in them. The Nominee tested his public claims against the archive and found the pattern beneath them — an institution that reads each new shock through the framework built for the last one, and tells the public a confidence its own internal record does not support.

After the confirmation hearings, The Pillars formalized the argument into a program with four parts: recognize the productivity transformation that would license lower rates, repair the framework and its communications, reclaim a narrow mandate, retreat from the balance sheet. The Coordination traced the machinery the program would run on — the instruments, the long end of the curve that two forces would have to pull down while the balance sheet pushed it up, and the dollar the doctrine leaned on without naming.

The Game, published the day the chair was sworn in, drew the prediction the rest had been building toward. The cut the program wanted could not be won on the incoming data before a committee disinclined to grant it. So the method would not be persuasion. It would be construction: build a framework, and present the cut as the recalibration of that framework rather than as accommodation argued against the inflation print. The Game closed on a single line — the next data point was being produced now. It arrived on June 17, and the meeting is best read not as a decision but as the first move of the installation the series had predicted.

What the Committee Discussed

A first meeting under a new chair is the natural place to look for what the chairmanship will be. The decision on rates was a hold, and the statement placed it in support of the dual mandate; the chair's opening remarks named both of the mandate's goals, price stability and maximum employment. What the meeting spent its time on, though, was the framework — the chair's framework — and the dual mandate it invoked is not the dual mandate it left in place.

The substance of the meeting was the apparatus of the program. Five task forces, announced from the podium, each mapped to a part of the program The Pillars had named: the balance sheet, communications, the data the Fed relies on, productivity and jobs in an era of transformation, and the inflation framework itself. The charge was to take the conduct of policy back to first principles. Alongside them came the shorter statement, the dropped guidance, the withdrawn dot, and a stated intention to remake the Summary of Economic Projections. The mandate's two goals were named; what went unspoken was the tradeoff between them. This is the framework being put in place.

That tradeoff is the thing the chair refused. The traditional question the dual mandate forces — how much unemployment a central bank should accept to bring inflation down — he set aside, not as a goal he abandons but as a choice he denies having to make.

The dual mandate survives the meeting; the tradeoff at its center does not. What the committee took up in its place was a framework that dissolves the choice — one under which strong growth, low prices, and full employment are presented as mutually compatible, with the supply side doing the work the tradeoff used to do. The first meeting is the first installment of that framework, and the rest of the afternoon is the framework being defended.

The Disagreement

Installing a framework is not the same as agreeing on a rate, and inside the framework discussion the rate path divided the room. Asked what the projections contained, the chair described a committee split down the middle.

The split was not a detail. A reporter put the count at nine participants projecting a rate increase by year-end; another noted the board sounded hawkish and that the two-year yield was reading more tightening, not less. On the data the committee could count — inflation above target for years, a solid economy, a labor market the committee judged stable — the weight of opinion leaned toward holding or tightening, not cutting.

And yet the vote was unanimous. There was one proposal, no alternatives, limited discussion, and a unanimous result.

A unanimous hold over a committee split on the next move is agreement about the present and disagreement about the future, and the two instruments that would let an outside reader see the disagreement were withdrawn at the same meeting: the dropped guidance and the chair's absent dot. The disagreement is real, it leans against a cut, and it has been recorded in a form that reads as consensus. The question the rest of the meeting answers is how a cut gets delivered over it.

The Counted and the Inferred

A cut cannot be carried on the data the committee can count, because that data is running the wrong way — and the chair, asked the decisive question, explained why it should not be carried on the data at all. Pressed on whether the AI build-out was adding more to demand or to supply, he drew the distinction the framework rests on.

The argument the framework builds on that distinction is specific, and on its own terms it holds together. The money behind the build-out is being spent now; the productivity it is meant to produce arrives later. The inflation that spending generates is therefore the upfront cost of an investment whose payoff is disinflationary — which makes the inflation, in this reading, transitory. And the argument carries a policy implication the chair does not need to state for it to operate: raising interest rates to suppress that inflation would suppress the investment producing it, and choke the productivity that would resolve it. On this logic the Fed should not tighten into the build-out. It can ease.

The costs are real, and their scale is not in question. U.S. spending on AI infrastructure reached about $254 billion in 2025, by IDC's count — roughly three-quarters of a global total that had more than doubled in a year. That is the money being spent now, the demand side the chair said the Fed counts, and it is already in the figures. The productivity that would offset it is the supply side he said it infers — and, in his words, it will take longer.

So the framework's case for the cut does not rest on the inflation data, which it concedes runs the other way. It rests on the claim that the inflation is transitory because the productivity is coming. Everything turns on whether the productivity comes, and when — and that is the one quantity the chair placed beyond the committee's sight.

The Open Question

The Federal Reserve has stood at this question before, recently enough that some of the participants are familiar. The reading the chair leans on — investment inflationary while it is being built, disinflationary once it produces — is old in the institution, worked from Marriner Eccles in the 1940s through the supply-side debates of the 1970s. The episode that matters here is the one that most resembles the present: the high-technology investment boom of the late 1990s, when the Committee had to judge, in real time, whether a wave of capital spending was adding more to demand or to supply.

That episode is the one The Assay examined, where Alan Greenspan judged the economy's capacity had expanded and held rates below what his models prescribed. But the fuller record is less a story of patience than of patience won against the institution's grain. The Committee's dominant reflex was to lean against the boom: it tightened in early 1995, took an overheating posture in 1999 and 2000, and worked from a staff forecast that kept underestimating how fast productivity was growing. The working presumption was that a fast-investing economy was an economy nearing its limits, and that the prudent move was to restrain demand before the new supply could arrive.

Against that presumption, the view that the boom was itself disinflationary and should not be met with higher rates was the minority position. Robert McTeer, at the Dallas Fed, argued that high-technology production and its spread through the economy were inherently disinflationary.

His quarrel was not only with the conclusion but with where the institution had trained its attention. The supply side, he argued, had been treated as a fixed quantity while the machinery of policy watched demand.

The minority was substantially right. The disinflation they described arrived; the slowdown the staff kept projecting did not, as Greenspan acknowledged to his own committee.

What matters for the present is not only that the dissenters were vindicated but how. They did not certify the supply gain before it arrived. They read it as it accumulated, in the productivity numbers and the behavior of high-technology prices, and Greenspan's own turn rested on the repeated, documented failure of the pessimistic forecasts to come true. Even then, with the boom years old, he would not call the matter closed. As late as the middle of 2000 he told the Committee he knew of no evidence that could conclusively decide between the staff's pessimistic forecast and a more benign reading, and that the answer might not be available until after the fact — for the data to settle as it came in, not in advance of it.

The Winner's Case

Set against that record, the chair's position reads less as the institution's reflex than as its correction. Asked about the labor market and the supply side, he told the room that strong, productivity-led growth was something to welcome rather than fear.

The instinct is to agree, and the archive gives the agreement teeth. The reading is not merely coherent; it is the historically vindicated dissent. He is taking the side the institution got wrong. The Federal Reserve has a documented bias toward over-tightening into exactly this kind of boom — the 1990s record is the case in point — and the people who resisted it, McTeer among them, were right. A chair who refuses to raise rates into a productivity transformation, and who says the supply side is to be embraced rather than feared, is standing where the minority that proved right stood. That is the strongest his case ever gets, and it is strong.

The turn is in how that minority earned its conclusion. They did not certify the supply gain before it arrived; they read it as it accumulated, in the productivity numbers and the observable behavior of prices, and the eventual convergence of the institution rested on the repeated failure of the pessimistic forecasts to come true. Even with the boom years old, the question of whether investment was adding more to demand or to supply was still being treated as open and empirical. Warsh takes the conclusion — do not tighten, ease — on a supply gain he conceded the Fed only infers and that will take longer, and he assigns a task force to certify it by year-end. He adopts the vindicated answer while skipping the accumulating evidence that earned it, and schedules a finding for what the institution, the last time it faced this, left to the data to settle.

The move has a familiar shape, the one The Nominee identified — the distance between what the institution works out in private and what it says in public. In the 1990s that distance ran hawkish: a chair privately treating the supply question as open while the public statements warned of costs running ahead of productivity. Here it runs the other way. The private concession is that the supply side is still inferred; the public framework is confident enough to schedule a cut around it.

The Delivery Schedule

This is where the deferral that looked like avoidance turns out to be a timetable. Through the afternoon the chair sent the hard questions to the task forces, and the most consequential of them was the question of when the supply side catches up with the demand side. He put it to a reporter directly.

The task force that holds that question — the productivity payoff, the crossover — is charged to conclude, with the rest, by year-end. Read as a deferral, the timing looks evasive: the answer arrives after the summer meetings that might have used it. Read as a delivery schedule, it is precise. The last meeting of the year is in December. A productivity task force reporting at year-end reports in time to furnish, on the eve of that meeting, the supply finding the cut requires — the certification that the productivity is real and arriving, the judgment the data could not yet supply and the one the 1990s Committee left open for the data to settle. The other task forces deliver the rest of the apparatus on the same clock: a new communications framework, a remade SEP, a balance-sheet posture, an inflation framework reviewed from first principles. By year-end the framework is built, and the December decision is taken inside it.

That is the method The Game described, now visible on a calendar. The disagreement of June is not resolved by a change in the incoming data; it is overtaken by the framework arriving on schedule. A cut that the committee would not grant in June as accommodation against an inflation print is presented in December as the recalibration of an integrated framework — licensed by the productivity finding the task forces deliver, not by the data that divided the room. The dots submitted in pencil, the withdrawn forward guidance, the chair's absent projection: each removes a fixed point that a December cut would otherwise have to be reconciled against.

Every route to the cut that runs through the data points the wrong way, which is why the framework, and not the data, has to carry it. The committee leans toward holding or tightening. The balance sheet, as The Test showed, reaches the reserve floor before it reaches the rate the program wants. And the supply pillar is, in the near term and by the chair's own concession, a cost rather than an offset, its payoff still inferred. With the productivity disinflation that The Coordination assigned to the long end of the curve pushing the wrong way for now, the weight of that commitment shifts onto Treasury issuance and the foreign demand The Dollar named as the pillar the doctrine does not — handled off the task forces' books, in the weekly breakfasts with the Treasury Secretary the chair confirmed had widened the central bank's aperture. Given that the data will not deliver the cut, the only route left is the one the program built: the framework delivered by year-end, the cut taken inside it in December.

The Inference