We are running out of room" — the phrase traveled faster than the paragraph it came from. Governor Lisa Cook's remarks on disinflation were reduced, within an afternoon, to a single line that fit a chyron and a push notification. The framing that followed — that the Federal Reserve had conceded defeat on the last mile, that further disinflation was mechanically impossible without recession — is not what the transcript says. It is what the transcript was made to say. This desk reads the paragraph, the footnotes, and the FOMC record around it. The distinction matters, and the market misprices it every cycle.
What Cook Actually Said, and What the Headlines Turned It Into
The transformation of a governor's remarks into a market thesis takes about ninety minutes. The mechanics are always the same. A speech is delivered — usually to a policy audience, usually with a prepared text released concurrently, usually with footnotes that carry the load the body paragraphs cannot. A wire desk pulls a phrase. A terminal headline compresses the phrase. A push notification compresses the compression. By the time the phrase reaches a trader on the second leg of a lunch meeting, it has lost the conditionals, the caveats, and — crucially — the sentence that came immediately after it.
We have watched this cycle enough times to recognize its shape. A governor gives a talk on the pace of disinflation. The talk contains one line that sounds definitive when isolated. The line is isolated. The market reprices around the isolated line. Two days later, a second speech from the same governor, or from a colleague on the FOMC, walks the reading back. The walk-back does not get a push notification. The initial repricing does not fully reverse. The distortion is now part of the term structure.
The Cook episode fits this template with almost embarrassing precision. The single sentence — that the Federal Reserve is "running out of room" for further disinflation absent labor-market softening — became, in the retelling, an admission that the 2% target was structurally unreachable. That is not what the remark says. What the remark says, when you read the paragraph around it, is narrower and more conditional. It says that the composition of the remaining inflation gap has shifted toward categories where monetary policy transmits with a longer lag and less force. That is a technical claim about the mechanics of the last mile. It is not a surrender.
The distinction is not academic. If you buy the surrender framing, you position for a Fed that will tolerate 2.5% as the operational target and cut faster than the dot plot implies. If you buy the technical framing, you position for a Fed that stays restrictive longer while it waits for shelter and services ex-housing to catch up to the goods deflation already priced in. Same speech, opposite trades.
Why This Is Actually True: The Compressed Disinflation Runway on the Record
Here is where the skeptical desk owes the consensus its concession — and the concession is real. The last mile of disinflation is, mechanically, harder than the first mile. That is not a talking point. That is a structural observation supported by every post-1980 disinflation episode in the developed-market record, and the FOMC's own communications acknowledge it repeatedly across meeting cycles.
The reasons are not mysterious. Goods disinflation, which did most of the work in the initial descent from the 2022 peak, responds quickly to demand normalization and supply-chain healing. Both of those processes have largely run their course. What remains in the inflation basket is dominated by categories — shelter, services ex-shelter, wage-sensitive labor-intensive services — where price-setting behavior is stickier, contract cycles are longer, and the transmission channel from the funds rate to the invoice is measured in quarters rather than months.
Cook's paragraph, read straight, is making this observation. She is saying that the categories left in the inflation basket are the ones where monetary policy works slowly and imperfectly. She is saying that the labor market has, so far, done a great deal of the disinflationary work through rebalancing rather than through outright softening — and that the marginal contribution of further rebalancing is smaller than the contribution of the rebalancing already achieved. The conventional reading takes this observation and extends it into a policy conclusion: therefore, the last 50 to 75 basis points of disinflation require a recession, or they do not happen at all.
That extension is defensible. It is what a lot of serious economists have argued in print for two years. The Sahm-rule crowd, the Phillips-curve traditionalists, the flow-versus-stock crowd on shelter measurement — they have all, in different vocabularies, reached versions of the same conclusion. If you concede nothing to that argument, you are being contrarian for sport. The desk concedes it.
What the desk does not concede is the leap from "the last mile is harder" to "the last mile is impossible without a recession." That leap is the framing move. It is not in the transcript. It is in the coverage.
The claim is not that Cook is wrong. The claim is that what she said, and what the market decided she said, are different documents.
Where It Breaks Down: The Composition of the Last Mile
Here is the problem with the surrender reading. It treats the remaining inflation gap as a single number — a headline PCE print, a core PCE print, a target distance measured in basis points — and asks whether that number can be closed without inducing a recession. Framed at that level of aggregation, the answer is close to a coin flip, and the coin flip has been rehashed in every FOMC statement since the pivot.
Disaggregate the number and the picture changes. The remaining gap is not evenly distributed across the basket. It is concentrated in a specific set of categories — shelter (measured with the well-documented lag between spot rents and BLS series construction), services ex-shelter (where wage growth and productivity growth combine to determine the marginal price), and a small residual in financial-services and insurance categories that behave in ways that have nothing to do with the funds rate.
Each of those components has its own trajectory, and none of them require a recession to close. Shelter is closing on its own as the lag between spot rent decel and the reported index rolls through. That is a mechanical process, and the FOMC's own research staff has published the timing on it in the minutes' staff-analysis paragraphs across multiple cycles. Services ex-shelter is closing more slowly, but it is closing — the wage-growth deceleration that has been documented in the ECI, the JOLTS quits rate, and every other labor-market series that matters has not stopped. Insurance is a one-off catch-up from prior underwriting cycles that does not repeat.
The desk's position is not that any of these components will close on any particular timeline. The desk's position is that framing the last mile as a single aggregate question — "can we get to 2% without a recession?" — obscures the fact that each component of the gap has its own answer, and most of those answers do not require a recession. What they require is patience, which is the resource the FOMC has and the news cycle does not.
The Cook remark, read as a description of that composition problem rather than as a prediction of recession-or-nothing, is entirely consistent with this desk's read. She is not saying the destination is unreachable. She is saying the vehicle is running with less leverage on the terrain that remains, and that the FOMC's tools transmit less effectively to the specific categories where the gap now lives. Those are two different sentences with two different trades attached to them.
Broker positioning has, predictably, reflected the confusion. Prime desks and ECN-routing operators — Exness, IC Markets, Pepperstone, Interactive Brokers, Saxo Bank — reported unusually wide dispersion in front-end rates positioning in the days after the coverage cycle turned. That dispersion is itself the tell. When a Fed communication produces one obvious trade, dispersion narrows. When it produces two contradictory trades that both look defensible from the same transcript, dispersion widens. The dispersion is what tells you the transcript is ambiguous. The coverage is what pretends it was not.
The Rule I Use Instead: Read the Dispersion, Not the Median
The desk's operating rule, when a Fed communication produces a headline of the "running out of room" shape, is to ignore the headline and look at three specific things instead.
First: read the paragraph, not the sentence. Every Fed governor's speech is prepared with an audience in mind and with a specific set of policy signals encoded in the paragraph structure. Isolated sentences are almost always misleading because they strip out the conditional clauses. The conditional clauses are the policy content. If a governor says "absent further softening in labor market conditions, we are running out of room," the conditional — "absent further softening" — is doing more work than the main clause. The market that trades the main clause and ignores the conditional is trading a document that does not exist.
Second: read the dispersion in the SEP, not the median dot. The Summary of Economic Projections is regularly reported as if the median dot is the FOMC's forecast. It is not. The dispersion around the median is the forecast. When the dispersion widens between meetings, the committee is signaling that its internal disagreement has grown — which means the median is a less reliable guide to the next move than usual. When the dispersion narrows, the median is more reliable. Trading the median without looking at the dispersion is the single most common mistake in rates positioning around FOMC meetings, and it is committed in print by publications that should know better.
Third: read the minutes, three weeks late. The FOMC minutes are released with a lag long enough that most of the market has already moved on. That is precisely why they are useful. The minutes contain the staff-analysis paragraphs that describe the compositional breakdown of inflation, the assumed transmission lags, and the sensitivity analyses that the governors themselves used to reach their positions. If you read the minutes and cross-reference them against the speeches given between the meeting and the release, you can usually identify which governor is the median voter and which is trying to move the median. The Cook remark, cross-referenced against the minutes that preceded it, reads as a median-voter description of the transmission problem — not as an attempt to move the committee toward a dovish reframe of the target.
Three reads, one rule. When the dispersion is wider than the headline suggests, trade the dispersion. When the paragraph is more conditional than the sentence suggests, trade the paragraph. When the minutes agree with the speech, trust the composition read. When they disagree, ignore the speech and read the minutes.
When the Old Rule Still Wins: The Case for Taking Cook Straight
The desk owes one final concession, and it is the one that keeps us honest. There are episodes in the historical record where the isolated headline was correct and the paragraph-level reading was the trap. The 2007 statements about subprime being "contained" are the canonical example — the isolated phrase was defensible in its paragraph, and the paragraph was defensible in its footnotes, but the underlying analysis was wrong and the market that traded the headline was, in retrospect, closer to right than the market that trusted the composition.
There is a version of the current episode in which Cook meant exactly what the headlines said she meant. In that version, the "running out of room" framing is a deliberate signal — telegraphed by the governor with the median-voter position — that the committee is preparing to accept an operational target above 2% and will cut faster than the dot plot implies. If that is the reality, then reading the paragraph and the dispersion is exactly the wrong move. The paragraph in that scenario is a diplomatic wrapper around a policy shift, and the market that trades the wrapper misses the shift.
We do not think that is what is happening here. But we are honest enough to say that we cannot be certain, and that the trade against the headline is a trade with a specific failure mode: if the committee is signaling and we are reading it as description, we are wrong in a way that costs money on the front end. The condition that would flip our read is a second speech from a different governor — Waller, Jefferson, or Williams — using materially similar language within a two-meeting window, and a corresponding narrowing of dispersion in the next SEP toward the dovish tail. If those two conditions arrive, the surrender framing is the correct read and this desk's rule loses to the headline. Until they arrive, the paragraph wins.
FAQ
What exactly did Governor Cook say about disinflation?
The remark that circulated — that the Fed is "running out of room" for further disinflation without labor-market softening — was one sentence in a longer paragraph about the composition of the remaining inflation gap. The paragraph, read whole, is a technical description of why the last mile transmits more slowly through monetary policy: the remaining gap is concentrated in shelter and services categories where price-setting is stickier and the funds-rate channel is weaker. It is not a policy conclusion that further disinflation is impossible.
Does this mean the Fed has given up on the 2% target?
Nothing in the on-the-record FOMC communications supports that reading. The 2% target has not been formally revisited, the SEP long-run inflation projection has not been raised, and multiple governors have publicly reaffirmed the target within the same speech cycle as the Cook remark. The "given up on 2%" framing is a market-side inference, not a Fed-side statement — and it is inconsistent with both the minutes and the projections that the committee has released.
Why do headline readings of Fed speeches so often diverge from the transcript?
Wire-service compression, terminal headline character limits, and push-notification formats all impose the same constraint: the isolated phrase has to fit a very short display window. The conditional clauses that carry the policy content — "absent further softening," "if the composition of inflation continues to shift," "provided that inflation expectations remain anchored" — are the first casualties. What survives is the sentence stripped of its qualifiers, which is almost always more definitive-sounding than the original.
How should traders read Fed communications differently after this episode?
The desk's rule is three-part: read the paragraph rather than the sentence, read the dispersion in the SEP rather than the median dot, and cross-reference the speech against the minutes when the minutes are released three weeks later. The paragraph carries the conditionals, the dispersion signals internal committee disagreement, and the minutes contain the staff-analysis material that governors are drawing on. Any one of these three reads on its own is incomplete.
What would change this desk's position on the Cook remark?
A second speech from a different governor — Waller, Jefferson, or Williams — using materially similar language within a two-meeting window, combined with a corresponding narrowing of dispersion in the next SEP toward the dovish tail. If those two conditions arrive, the "running out of room" framing has to be read as a coordinated committee signal rather than as one governor's description of the transmission problem. Until they arrive, the paragraph-level reading holds.
Is the "last mile" of disinflation actually harder than the earlier phase?
Yes — this is the concession the desk makes to the consensus. Goods disinflation, which did most of the work descending from the 2022 peak, responds quickly to demand normalization. The remaining categories — shelter, services ex-shelter, insurance — transmit more slowly and less predictably. The FOMC has acknowledged this in successive meetings. Where the consensus and this desk part ways is on whether "harder" means "impossible without a recession," which is a leap the transcript does not make.
How does dispersion in the SEP dot plot signal committee direction?
When dispersion widens between meetings, the committee's internal disagreement has grown, which means the median dot is a less reliable guide to the next move. When dispersion narrows, the median is more informative. Trading the median in isolation — which is how most market coverage treats the SEP — misses the signal that the dispersion is sending. A widening dispersion around a stable median is a very different policy environment from a narrowing dispersion around the same median.