Core PCE Fell to 3.0%. The Measurement Changed, Not the Prices.

On Wednesday the Bureau of Economic Analysis reported that core inflation had fallen to 3.0 percent. Markets cheered. Every outlet I read ran the same line: inflation is cooling, the Fed can stop.

Then I got to the part nobody quotes.

In the same release, the BEA confirmed it had changed how it measures prices in three categories. Not a rounding tweak. A redefinition of what gets counted. And when you strip that out, the picture looks different enough that I think you ought to see it before the Fed meets on the 28th.

This week: the number that moved for the wrong reason, what it means for rates, why Bitcoin has gone quiet, and a warehouse full of boxes that nobody is shipping.

📉 The ruler got shorter, so the room looks bigger

Core inflation did not fall last month. The method used to measure it changed.

Here are the numbers. Core PCE, which is the inflation gauge the Federal Reserve watches most closely, went from 3.3 percent in July to 3.0 percent in August. Headline PCE went from 3.7 percent to 3.4 percent. Both fell by exactly three tenths of a point.

In the same month, the BEA rewrote how it prices three things: portfolio management and investment advisory fees, legal services, and computer software. Sensible housekeeping, by the sound of it. Fee data instead of estimates. Software split into categories that behave differently instead of blended into one.

Economists at Goldman Sachs and JPMorgan had said publicly that the change would knock roughly one to two tenths off core. It knocked off three.

Think of it like stepping on a new bathroom scale and finding you've lost four pounds. Wonderful news, until you notice the old scale weighed your shoes and the new one doesn't. Your trousers still fit the same.

Month to month tells the honest story. August headline prices rose 0.3 percent, against a revised 0.1 percent in July. On a monthly basis, prices sped up.

So is inflation beaten, or did somebody move the goalposts and everyone cheered the goal?

Honestly, a bit of both. The revisions look defensible on their own terms. The problem is the timing. A measurement change landed in the exact month that markets most wanted good news, and the good news got reported as though prices had behaved.

🏛️ What happens on the 28th

The market has decided the Fed will sit still, and it decided that on the strength of two numbers.

Quick recap for anyone who's been living a normal life. On 16 September, under new chair Kevin Warsh, the Fed raised rates to a range of 3.75 to 4.0 percent. The next meeting runs 27 and 28 October.

Two data points since then have pushed the odds toward a pause:

1. Core PCE at 3.0 percent, against a 3.3 percent consensus. The one we've been picking apart above.
2. September payrolls at 29,000, a long way below forecast, with unemployment rising to 4.2 percent. That one is clean. No methodology asterisk, no debate. The jobs market is cooling.
3. The next test is 14 October, when September's consumer price index lands. That's the CPI, a separate inflation measure built by a different agency using a different basket.

That third one is the date to put in your diary, and here's why it carries more weight than usual. CPI is calculated by the Bureau of Labor Statistics, not the BEA. It did not get the same methodology change. So on 14 October we get a reading on the same question from a different ruler.

If CPI also comes in soft, the disinflation story holds and the pause is sound. If CPI comes in hot while PCE came in cool, we'll know which of the two was telling us about prices and which was telling us about spreadsheets.

What should you do with that? Nothing dramatic. Note the date, and treat anyone who sounds certain about the 28th before the 14th as someone guessing in a confident voice.

🪙 Bitcoin had a boring week, which is the interesting part

Bitcoin sat around 84,700 dollars through a week that handed it two major macro shocks, and barely twitched.

A soft inflation print and a badly missed jobs number would once have thrown this market around like a deckchair in a gale. This week it absorbed both and finished roughly where it started.

Boring is a feature, not a bug. A market that stops lurching at every headline is a market with real buyers underneath it rather than a crowd of people guessing at each other.

Does that mean the floor holds forever? Of course not. It means that for now the thing that used to break first isn't breaking. Worth watching, not worth betting the house on.

📦 A warehouse full of boxes nobody is shipping

The hottest story in finance this year is tokenized real-world assets, and 93.9 percent of them are sitting still.

A quick translation first. Tokenizing a real-world asset means taking something that already exists, a Treasury bond or a share of stock, and issuing a digital version of it that can move on a blockchain. The pitch is that it settles faster, trades around the clock, and can be used in places the original can't go.

The enthusiasm is real and it's coming from serious places. JPMorgan, Franklin Templeton and WisdomTree have all rolled out tokenized funds this year. On 17 September the SEC issued what it called an innovation exemption, clearing a path for tokenized US-listed stocks to trade on blockchain venues. Regulators are holding the door open, which is not a sentence I expected to write about this industry.

Now the awkward bit. Dune's latest report puts the whole tokenized market at about 34.5 billion dollars, up more than 140 percent on the year. Of that, only 6.1 percent is being used for anything, roughly 2.11 billion dollars posted into lending. The rest is issued, held, and doing nothing at all.

It's a warehouse the size of an airport, stacked to the roof with sealed boxes, and one forklift moving one box toward the loading bay.

Why so idle? Because for most of these assets there's no reason to move them. The bulk of the growth has been in tokenized Treasury funds paying around four percent. Borrowing against them costs more than that. Nobody takes on extra risk to lose money, so the boxes stay sealed. The one corner that is moving is tokenized private credit, which pays high single digits and therefore has a point: credit products make up 76 percent of all real-world asset deposits in lending.

One more number worth holding onto. Three venues, Morpho, Aave and Kamino, account for 83 percent of all tokenized asset lending between them. Morpho alone holds around a billion dollars of it.

Where that leaves us

Three stories this week, and the same thread runs through all of them. The headline said one thing and the detail said another.

Inflation looked beaten and turned out to have been remeasured. Tokenization looked like a boom and turned out to be a storage business. Bitcoin looked dull and turned out to be the most reassuring chart of the week.

It's the financial equivalent of a weather forecast that promises sunshine while everyone outside is putting their coats on. The forecast isn't lying. It's measuring something slightly different from what you're standing in.

My call for the rest of October: the Fed holds on the 28th, because the jobs number alone gives them cover even if you throw the PCE print out entirely. The bigger question is what the 14th tells us, and whether the pause turns out to be a rest or a finish line.