(I write a monthlyish opinion piece for Barron’s. This one was published there in September. My previous pieces are here.)
A lingering puzzle about inflation is why the public still seems so unhappy about it, even though it has, by conventional measures, returned to normal.
One explanation is that people are simply confused, or misled by the media. But another possibility is that what people think of as the cost of living doesn’t match up with the way that economists measure inflation. Maybe people aren’t wrong or confused, they are just paying attention to something different.
Inflation means a rise in the cost of goods and services. But not all your bills are for goods and services. Things like interest payments are also costs. And last fall’s election followed three years of steeply rising interest rates. The average mortgage rate, for example, was over 7%, compared with less than 3% in 2021
That is the explanation for the mismatch between official statistics and public perceptions offered in a fascinating recent paper by former Treasury Secretary Lawrence Summers and several co-authors, titled “The Cost of Money is Part of the Cost of Living.” In one especially dramatic finding, they suggest that if we take interest into account, year-over-year inflation peaked to 18% in 2022, rather than the official 9%, and was still 8% at the end of 2023, when the official rate was 3.3%.
I think the paper overstates its case. But it is still pointing to something real.
Conventional measures of inflation are supposed to reflect the prices of currently produced goods and services, but not asset purchases or financial transactions. But it isn’t always easy to know which payments are which. As a homeowner, you are buying both a place to live for the month, and an asset. In principle, the first should be counted in inflation, the second should not. But your single mortgage check includes both.
Today, the Bureau of Labor Statistics, which produces the country’s main inflation indicator, deals with this problem by imputing “owners equivalent rent.” In effect, we ask how much a homeowner would pay for their home, if they were renting it.
Before 1983, the BLS did things differently. Instead, it counted the full cost of home purchases, but only for houses bought in that period. Today’s measure estimates one month’s rent for all owner-occupied homes; the older method looks at the total cost of those homes purchased this month. Since houses are normally paid for with mortgages, that meant including interest payments that would be paid over many future years as part of this month’s price level. (To be exact, the BLS included future interest payments over half the length of the mortgage.)
Houses are a big part of consumption, so this difference isn’t a small detail. Summers and co-authors are absolutely right that when we compare inflation today to inflation in the 1970s, we aren’t comparing apples to apples. “Inflation” then meant something different than it does today.
An earlier paper by three of the same economists looked at historic inflation using the modern definition. They concluded that, when we measure consistently, the late-1970s inflation was no higher than the inflation during the pandemic.
The new article takes the opposite approach, and applies the pre-1983 definition to the recent inflation. This is a bit odd, given the strong and convincing criticism of the old methodology in the earlier article. Nonetheless, the results are striking. If people were experiencing inflation at double the official rate, no wonder they were upset!
In my opinion, the authors had it right the first time. There are good reasons the BLS abandoned its old approach. By including future interest payments on homes purchased in the current month, the old methodology greatly exaggerates the impact of interest rate changes. You can reasonably say that the mortgage payments you will make a decade from now are part of the price of your house, but they are not in any meaningful sense part of your cost of living today.
That said, it does make sense that interest payments contribute to people’s experience of price increases. But how much? As a back of the envelope guess, we can observe that household interest payments grew from an annualized $600 billion in the last quarter of 2020 to over $1 trillion by the end of 2023. Those payments grew twice as fast as nominal consumer spending. If we add these interest payments to the cost of the consumption basket, then we find that the 2021-2022 increase in inflation was as much as two points greater, and inflation in 2024 remained about half a point higher than by conventional measures.
It seems to me that if you take seriously the idea that financing is part of the cost of goods and services, you can plausibly conclude that people were experiencing an inflation rate of 3% to 3.5% last fall, rather than the 2.5% to 3% percent reported by the BLS. That isn’t trivial. But I’m not sure it’s the sort of difference that elections turn on.
Still, Summers and his co-authors are pointing to something real and important. The cost of money is part of the cost of living. When the Federal Reserve aggressively raised rates over 2022-2023, it may – or may not! – have helped bring down inflation. But it definitely made it harder for families, and businesses, to service their debts. Monetary policymakers would do well to keep that second impact in mind in the future, along with the first.