Imagine a barber who charges 1% of your net worth for every haircut. Your stocks jump 20% in a year, so the same haircut now costs 20% more. The scissors, the chair and the ten minutes are all the same. For years, something close to this happened inside America’s favorite inflation gauge, and the PCE methodology change released on Sept. 30 changed how it is counted.
First, a quick map. PCE is the Personal Consumption Expenditures price index. It is the number the Federal Reserve uses for its 2% inflation goal. “Core” PCE leaves out food and energy, because those prices jump around too much to show the trend.
Here is where the barber comes in. Most fund managers charge a percentage of the money they look after. When markets rise, their bill rises on its own. The old method mostly treated that bigger bill as a higher price, so a stock rally could push measured inflation up. By RBC’s estimate, this line made up about 1.7% of core PCE and grew about 14% a year under the old method since late 2023.
The Bureau of Economic Analysis (BEA), which builds PCE, changed three things in its yearly update:
- Fund management fees. Instead of starting from a price, BEA now estimates how much service people got by looking at jobs and hours worked in the industry, using Labor Department data.
- Software. The old price series also picked up some hardware. It is now a mix of consumer and producer price data.
- Legal services. The old price series had mostly stopped being published, so BEA swapped in a new one. This change pushes inflation slightly up, not down.
All of it was applied backward to 2021.
Then the August numbers came out. Core PCE rose 3.0% from a year earlier, against a forecast of 3.3%. Headline PCE was 3.4%, against 3.7%. That looks like a big win. But on the same day, July was rewritten too. The revisions cut July’s core rate by 0.36 percentage point, from 3.3% to 3.0%.
So core inflation did not fall from July to August. It sat at 3.0% on the new ruler. The forecasts missed because they were built on the old one.
One more detail is worth noticing. Before the release, banks guessed the PCE methodology change would trim core inflation by roughly 0.2 point. The real revision to July was bigger. Economists also disagree about which of the three changes did most of the work. That gap is a good reason to be careful with any single month right now.
Bond traders seemed to agree. Treasury yields dipped after the report, then turned higher, and the 10-year yield closed near 5.29%. A real cooling in prices would usually pull long-term yields down. A new way of measuring does not change how much rent, gas or groceries cost.
This is not the first time this year that the measuring stick itself became the story. In June, the debate was about trimmed mean PCE, another way of reading the same prices. Both point to one habit worth keeping: when a number changes, check whether prices moved or the method did.
What to watch next is the September PCE report at the end of October. It will be the first month where forecasts and the result use the same ruler. If core stays near 3.0%, that is real progress. If it drifts back up, August was mostly a change in arithmetic.
My take: New ruler, same fever.
Not advice. Just how I see it.