This has been a big week for inflation numbers. We got two major reports from the Bureau of Labor Statistics, on consumer prices and producer prices. We also got reports on consumer inflation expectations and on business inflation expectations — which I’ve become increasingly interested in. Across these, the news ranged from decent to great. We still don’t know for sure if we’ll manage to get inflation under control without a recession, but the odds are looking better.
But did you get that message clearly from news reports? My guess is that you didn’t, not because the media got the facts wrong but because much of the coverage — certainly the coverage I saw — was unduly shaped by one number. That, of course, is “core inflation,” or annual inflation excluding food and energy, a measure that was useful in the past but has become misleading in the post-Covid era.
At any rate, many of the reports I saw said something like this: “Headline inflation was down in May, but core inflation remained elevated,” conveying the impression that we may not be making much progress.
Indeed, the consumer price data looks like this:
If you look at price increases over the past year, overall inflation has come down a lot, but much of that reflects falling gasoline prices, so the traditional measure of core inflation, which doesn’t include those gas price cuts, has barely fallen at all.
But everyone knowledgeable who is following these numbers knows that this comparison isn’t telling us much about what’s actually happening.
To see why, it’s helpful to know something about the history of why we typically talk about two different measures of inflation.
Way back in 1975, the economist Robert Gordon argued that it was important to distinguish between inflation driven by goods with highly volatile prices — like, yes, food and energy — and what he called “hard-core” inflation, driven by goods and services whose prices tend to change more sluggishly. Inflation caused by, say, a spike in oil prices tends to be easy come, easy go, but inflation driven by, say, rising wages tends to have a lot of inertia and be hard to bring down. Or to put it another way, a measure that excluded more volatile prices could help extract the signal from the noise.
As a practical matter, the Federal Reserve ended up focusing on a measure of “core” inflation — the “hard” part of the name got lost along the way — that simply excluded food and energy, which historically were the main sources of large but temporary short-term fluctuations in inflation. And this focus was extremely useful in the aftermath of the financial crisis. There was a brief surge in inflation during 2010-11, which had some people, mainly on the political right, screaming that the Fed was “debasing” the dollar. The Fed, however, kept calm and carried on, because core inflation remained subdued — and the Fed was right.
The general idea of estimating a measure of core inflation and using it to guide policy, then, was, and is, a good one. The problem is that the traditional measure of core inflation no longer does a good job of extracting the signal from the noise. If anything, it adds noise.
One reason is that large temporary shocks are now coming from sources other than food and energy, notably supply chain disruptions that led, for example, to huge swings in the price of used cars.
Even more important, traditional core inflation is strongly affected by the price of shelter, which is about 40 percent of the core. The main components of shelter inflation, in turn, are the average rent paid by tenants and “owner’s equivalent rent,” an estimate of what homeowners would be paying if they were renters — a measure basically derived from average rents.
But here’s the thing: Most tenants have fairly long leases, so the average rent tenants pay lags far behind the rents paid by new tenants, which more closely reflect the current state of the economy. Ordinarily this isn’t a big issue, but there was a huge surge in rents between 2021 and early 2022, probably driven by the rise in remote work.
This surge is now well behind us. We know this from a number of private estimates of market rents, for example, from Zillow, and we also now have some official numbers. Researchers at the Bureau of Labor Statistics calculate both average rental rates (ATRR) and new tenant rates (NTRR); NTRR is actually falling, but ATRR is still rising, although much less than it was:
What this tells us is that the standard measure of core inflation is strongly affected by data that is very out of date. The latest numbers are driven in large part by surging shelter costs:
But market rents aren’t actually rising at an 8 percent annual rate; they’re flat or falling.
One more thing: In the past, it may have made sense to look at changes over the last year, but in an economy going through as much turmoil as we’ve seen recently, that’s just too long a lag. Monthly data is too noisy, so many economists are now focusing on either three- or six-month changes. My sense is that even three-month data is too noisy, so six months is better, but in any case, we don’t want to focus on annual rates of change.
So what should we be focusing on? My preferred measure these days is “super core,” which excludes both used car prices and shelter. Here’s what it looks like on a six-month basis:
OK, so this measure shows a clear picture of disinflation, which probably isn’t the picture you get from most recent reporting. To be fair, other measures are less clear, and my main purpose today isn’t to argue that we’re winning the war on inflation (although I do think we are).
My point instead is that news organizations should stop playing up estimates of annual inflation excluding food and energy. Once upon a time, this was a useful number, but at this point it’s a relic, a legacy of a bygone age. And putting that fossil statistic in a story’s lede ends up misleading readers rather than informing them.
