Imagine stepping on a scale that the manufacturer quietly recalibrated overnight. The number it shows you this morning might be accurate, or it might be measuring something slightly different than yesterday’s number, and you would have no way of knowing which. That is essentially what is happening with today’s inflation report. When the Bureau of Economic Analysis releases the August personal consumption expenditures price index at 8:30 a.m. Eastern, the number will incorporate new measurement methods for several categories of spending, plus revisions stretching back to 2021. The inflation reading will be real, but it will also be, in an important sense, provisional. This is the explainer on why, and how to read the number like an economist.
What the BEA is changing
The PCE price index measures how the prices of everything American households consume change over time. It is the Federal Reserve’s preferred inflation gauge and the basis for its 2% target. To build it, the BEA collects price data across hundreds of categories, from gasoline to haircuts to hospital visits, and weights them by how much households actually spend.
Once a year, as part of its annual update to the national accounts, the BEA improves its methods. This year’s update, which lands with today’s release, changes how the agency calculates inflation for three specific components: portfolio management services (the fees you pay investment managers), computer software and accessories, and legal services. Of the three, economists say the portfolio management change is the one they understand least, and therefore the one that could move the headline number in unpredictable ways.
Why does portfolio management matter for inflation? Because it is a big, fast-growing category. As asset values have risen and more households use professional management, the fees have grown as a share of total spending. How you measure price change in a service whose “price” is a percentage of assets under management is genuinely tricky, and the BEA’s new deflator could shift the reported inflation rate for the category noticeably. As Barron’s reported, the change could complicate the Fed’s read on whether price pressures are actually easing.
The revision problem
Here is the part that makes today’s number genuinely hard to interpret. The annual update does not just change August. It revises history, in this case back to 2021. That means the year-over-year inflation rate, the 3.7% or 3.6% or 3.5% that will dominate headlines, is computed against a revised past. If last year’s prices get revised down, this year’s inflation rate mechanically goes up, even though nothing changed in the real economy.
Fed Governor Christopher Waller has flagged exactly this. He has said publicly that downward revisions to some non-market prices are likely, which could trim the year-over-year rates. Goldman Sachs built those expected revisions into its forecast, estimating the monthly headline gain at 0.33% and the annual rate at 3.58%, with core at 0.27% monthly and 3.17% annually, noticeably below the consensus of 0.4% and 3.7% headline and 0.3% and 3.3% core, according to a preview detailed by ZeroHedge.
Think about what that means. Two careful forecasts can differ by two-tenths of a percentage point on annual inflation, not because they disagree about the economy, but because they disagree about the accounting. At a moment when the Fed is deciding whether to hike again in October, two-tenths is the difference between “inflation is stuck” and “inflation is easing.”
How economists actually read the report
So how do the professionals handle this? They look past the headline in three ways.
First, they check the monthly core number against its recent trend. Core PCE ran at 0.2% in July. If August prints 0.3%, that is an acceleration regardless of the annual rate’s quirks. The month-to-month change is less affected by revisions than the year-over-year figure.
Second, they look at market-based PCE, which strips out the imputed, non-market prices that are most affected by methodology changes. In July, market-based core PCE ran at 3.0% annually, well below the 3.3% headline core rate. If that gap persists or widens in August, it suggests the “true” market inflation is cooler than the official number.
Third, they wait. One month of data distorted by methodology changes is noise. Two or three months establish a trend. The Fed knows this, which is why officials will spend today distinguishing changes in underlying prices from the effects of the BEA’s new methods, as Barron’s noted.
The same annual update also revises GDP history, and today’s third estimate of second-quarter growth, expected to hold at 1.5% annualized, will carry the same provisional quality. The whole statistical picture of the economy is getting repainted at once.
What it means for you
Here is the practical takeaway. When you see today’s inflation number, treat the first decimal place with respect and the second with skepticism. A headline of 3.6% versus 3.8% sounds like a meaningful difference, but if the gap is driven by how the BEA now deflates portfolio management fees, it tells you nothing about the price of groceries, rent, or gas.
What actually matters for your wallet is the trend in the monthly numbers and, more concretely, the prices you pay. The PCE report is the Fed’s dashboard, not yours. Your dashboard is your own spending: what your rent renewal says, what the grocery total looks like, what filling the tank costs. Those are measured without methodology updates.
And for the Fed’s October decision, the honest answer is that today clarifies less than the market wants it to. A clean, hot print would strengthen the case for a hike. A clean, soft print would weaken it. But a muddled print, which is the most likely outcome given the revisions, just pushes the decision to Friday’s jobs report and beyond. Sometimes the most important thing a data release teaches you is humility about data. Today is one of those days.
