Have you ever watched a referee blow the whistle, then glance at the replay and quietly wish the call had waited thirty seconds? That is the mood around the next policy meeting. Policymakers vote on borrowing costs in mid-September. Two weeks later, the same government shop that builds the inflation gauge they officially target will rewrite that gauge back to 2021 and drop the August reading in the same bundle. I keep coming back to a simple question. Why lock in a move on a number that is already scheduled to change?
The Calendar Makes The Risk Asymmetric
The timing is not a rumor. It is on the release calendar. Mid-month, the rate-setting committee decides. End of month, the historical PCE inflation series gets a facelift. One official has already flagged the direction of a notable piece of that facelift. A change in how fees paid to market intermediaries get measured could shave a few tenths off twelve-month inflation. He called it a welcome correction. When a sitting voter says the tape is about to look softer, pretending the old tape is carved in stone feels stubborn.
Three voters wanted a quarter-point increase at the prior gathering and lost. That dissent still hangs in the room. Hawks can point at a core reading near 3.3 percent and say the job is unfinished. Fair enough. The level is not the target. Nobody serious claims it is. The argument is about urgency. Is the next six weeks so dangerous that the committee must move before the statistical agency publishes the version of history it is prepared to stand behind?
Raise on the sixteenth and be wrong, and you tightened on a print the government changes fourteen days later. Wait and be wrong, and you can still move at the following meeting.
That is the whole trade. One error is awkward to explain on television. The other error costs six weeks. I have sat through enough policy postmortems to know which kind of mistake lingers in the transcript.
Two Mistakes, Two Price Tags
Think of it as insurance with a cheap premium. If the rewritten series still looks hot, the case for a hike in late October is cleaner. The number will be fresh. Staff will have had time to brief. Markets will have digested the methodology note instead of treating it as a surprise after the fact. If the series looks cooler, the committee avoids the peculiar spectacle of having just raised rates on data that the authors themselves just marked down.
I am not saying policy is already too tight. That is a different essay. This is narrower. When the official inflation history is about to be restated, patience is not dithering. It is housekeeping.
What The Recent Path Actually Shows
Core PCE is too high. Repeat that until it is boring, because it is still true. It is also not speeding up. On figures cited by one governor, three-month annualized core inflation slid from about 4.76 percent early in the year to roughly 3.05 percent through July. He called that a considerable improvement while admitting the level remains above two percent. Both statements can live in the same paragraph.
A trimmed-mean version that knocks off the wildest price swings in both directions has been running closer to 2.3 percent over twelve months. That is not victory. It is a reminder that the headline core basket can look angrier than the middle of the distribution. Underlying pressure is less dramatic than a single 3.3 handle implies, at least if you care about the bulk of the items rather than the tails.
| Measure | Recent Signal | How To Read It |
| Core PCE, 12-month | Still above 3 percent | Too high versus the goal |
| Core PCE, 3-month annualized | Stepped down through summer | Momentum cooling, not gone |
| Trimmed-mean PCE | Near the low twos over a year | Middle of the distribution calmer |
| Nonmarket services | Imputed, noisy | Can distort the official core |
Some of the gap between those rows comes from categories that are a pain to measure. Nonmarket services are imputed rather than rung up at a register. One official has said those prices have bothered him for a long time. Strip that piece out, and the rest of the story looks better than the official core line. The pending methodological tweak aims at one of those awkward corners. Legal services, for instance, have been priced with a consumer-side index that even the agency has described as producing erratic moves that cannot be corroborated. That is not a footnote for specialists. That is part of the number people want to hike on.
Why Measurement Noise Is Not A Side Show
Investors love a clean narrative. Inflation high, raise rates. Inflation falling, cut rates. Reality is messier when a chunk of the basket is constructed rather than observed. Imputed prices can wander. Brokerage-style fees can be booked in ways that do not match how households actually experience costs. When the statistical agency says it is going to fix one of those bookings, the honest response is to wait for the fix.
I have found that markets punish inconsistency more than they punish a six-week delay. A hike that has to be walked back in the minutes, or explained as “we knew the series would change but we moved anyway,” sounds like process failure. Holding still while the books get restated sounds like adults in the room.
- The official target is PCE, not a private nowcast.
- The historical PCE file is scheduled for a rewrite at month end.
- At least one measurement change is expected to lean softer.
- Three-month momentum has already cooled from the winter peak.
- The next scheduled meeting is only six weeks later.
None of those bullets prove inflation is cured. They prove the committee is not staring at a fire that doubles every night. The case for urgency is being assumed more than demonstrated.
What “Sit Tight” Actually Means For Markets
Waiting is not a promise of cuts. It is not a victory lap for anyone who wants easier money tomorrow morning. It is a sequencing choice. First the restated history and the August print. Then the vote. By late October the committee would have a package it does not have in mid-September: a revised back run plus a new monthly observation, published together, defended by the same shop that built them.
For anyone managing duration, that sequence matters. A surprise hike on soon-to-be-revised data can jerk front-end pricing around for reasons that evaporate in a fortnight. A later hike, if it still looks warranted, arrives with less statistical fog. Perhaps the most interesting aspect is how little growth data between the two meetings needs to explode to justify the pause. You are not asking for a recession print. You are asking for the inflation file to finish its homework.
Savers watching cash yields should not over-read a hold either. Policy can stay restrictive while the committee declines to add another quarter point on a moving target. Restriction and increment are cousins, not twins.
The Hawk Case, Stated Fairly
Let me steel-man the other side, because pretending hawks are cartoon characters is lazy. Core is still above three. Services remain sticky in places. Financial conditions can ease on their own if the committee sounds sleepy. Labor markets can reheat. If you believe the last mile is the hardest mile, every unused meeting is a gift to inertia. Dissenters who wanted a hike last time are not inventing the overshoot. They are reading a level that is still too high and asking why the reaction function only works in one direction.
That argument gets stronger, not weaker, if the revision leaves the picture hot. That is the point. If they are right, October will say so with a number the agency just blessed. If they are early, September becomes the meeting people replay in hearings. I would rather be a week late on a solid file than a day early on a draft.
When the government is about to rewrite the inflation series on which the decision rests, waiting one meeting is not indecision. It is the more defensible policy choice.
How Households Should Read A Pause
If you have a floating-rate balance, a hold does not refinance your life. It just avoids an extra twist this month. If you are hunting a mortgage, six weeks of extra data can move quotes more than a theatrical hike that later looks sloppy. If you live on short-term bills, yields can stay attractive even without another step up. The household version of the same logic is dull and useful. Do not rearrange a five-year plan because of a two-week statistical event.
In my experience, people overtrade the press conference and undertrade the footnotes. The footnote this month is the restatement date. Circle it. Then decide whether the committee looks careful or jumpy.
Credibility Is A Stock, Not A Flow
Central banks spend years building the idea that they react to incoming data, not to the loudest guest on a morning show. Incoming data includes known revisions. Ignoring a scheduled rewrite is a choice to privilege the old file. That can be justified in a crisis. This is not that week. Inflation is elevated and cooling in the higher-frequency cuts. Growth has not suddenly caught fire in a way that forces an emergency add-on. The expensive asset here is reputation. Spend it on something that will still look smart after September 30.
There is a human habit of treating any delay as weakness. Markets sometimes reward that habit for a session. They rarely reward it for a quarter. A committee that says, out loud, that it wants the restated series before it adds restriction is describing a process. Process is boring. Process also survives the next revision cycle.
What To Watch Between The Two Dates
- The size of the methodological effect on twelve-month core, not just the color commentary.
- Whether August’s new monthly print confirms the cooling in three-month annualized core.
- How trimmed-mean and market-based measures sit next to the official core after the rewrite.
- Any fresh signal from nonmarket services once the legal-services kink is treated.
- Whether financial conditions ease so far that the October meeting faces a different problem than measurement.
That list is homework, not a treasure map. If items one through four still scream overheating, the hike delayed is a hike delivered with better paperwork. If they do not, the committee will be glad it did not paint itself into a corner on the sixteenth.
A Note On Language And Overconfidence
Policy talk loves words like “data dependent.” Those words mean something only if the data are allowed to finish arriving. A scheduled historical revision is data. Treating it as a nuisance is the opposite of dependence. I have heard enough confident forecasts die on a benchmark revision to keep a little humility in the drawer. Inflation measurement is hard. Fees, imputations, and quality adjustments are not conspiracy. They are the sausage factory. When the factory announces a cleaner recipe on a known date, you taste the new batch before you change the menu price of money.
Is there a chance the rewrite barely moves the twelve-month rate? Sure. A few tenths is not a regime shift by itself. It can still be the difference between a close call and a clear one. Close calls are when process matters most.
The Narrow Claim, Repeated On Purpose
Rates may still need to go up. This piece does not crown a dovish outcome. It says the September meeting is a poor venue for that decision because the inflation history the decision cites is about to be edited. The thirtieth comes first. Then vote. If that sounds too tidy, good. Tidy is what you want when the alternative is explaining a hike against a downwardly revised core print on morning radio.
I will add a personal preference and label it as such. I would rather watch the committee look slightly late than watch it look sloppy. Late can be fixed at the next gathering. Sloppy sticks to the institution. Six weeks is not a long time in a multi-year inflation episode. It is long enough to let the agency publish the series it is willing to defend.
So here is the plain version. Core is high. Momentum has cooled. Measurement quirks are acknowledged by people inside the building. A rewrite is booked. Dissenters can bring the same charts in October with better scaffolding. Nothing in the next fortnight forces a choice that cannot wait for the file to be updated. Sit tight, read the new history, then decide whether another quarter point still earns its keep.
If the new history is hot, hike with a straight face. If it is not, you will be grateful nobody spent credibility on a draft. Either way, the sequence is the strategy. The calendar already wrote it down.
And if you only remember one line after all of this, make it this one. Do not raise rates on a PCE print that is already on the calendar to be rewritten. Wait for the version that survives the edit. Then argue, with numbers someone is prepared to own.
That is not poetry. It is just the order of operations. First the revision. Then the vote. The rest is noise we can live with for six more weeks.
