Why Has the CPI Inflation Calculation Changed Over Time?

Consumer prices are increasing at the fastest clip in about 40 years, climbing 8.3 percent in April compared with a year earlier.

As popular anger about the rising costs mounts, a chorus of critics have been arguing that the sky-high inflation figure is actually being undercounted.

In YouTube videos, on conservative talk shows and in posts by financial analysts, the critics argue that over the past several decades economists have tweaked one of the government’s standard measures of inflation, the Consumer Price Index, in ways that understate how quickly prices are rising. Those lower inflation figures give the government some economic breathing room, they claim, saving money on expenses like Social Security.

“The bottom line is these are not accurate numbers,” Tucker Carlson, the Fox News host, said during a segment on inflation late last year. He added, “Do the math and you will see that the actual number, the rise in inflation, is not even close to the 7 percent that Washington is claiming.”

But experts on inflation say the changes to calculations over the years have made the reported rate a more accurate snapshot of how much prices are rising for shoppers. The rate under a different methodology might be higher, they say, but the effect would be small, and the alternative number would do a poorer job of reflecting the costs consumers were grappling with. Inflation affects different people differently, but that does not mean that the overall numbers are incorrect.

“You have to understand the concept: What are people currently paying for consumption?” said Alan Detmeister, who was formerly head of the prices and wages section at the Federal Reserve and is now at the bank UBS. “It is trying to get at out-of-pocket expenses.”

Here are two major changes made to inflation since the 1980s and why economists adopted them.

Change No. 1: Inflation doesn’t include house prices

People who are skeptical about America’s inflation measures often cite a change to how home costs are measured in the Consumer Price Index, a closely watched metric produced by the Bureau of Labor Statistics.

In 1983, the government switched from using home prices — which also included mortgage payments and maintenance costs — to using rental prices to gauge the cost of housing.

The cost of housing for people who own their property is now measured using what is called “owners’ equivalent rent”: how much their house would cost to rent if they did not own it.

The idea is that homes are an investment. House prices appreciate, and you may eventually sell for a profit a property that you have purchased. Rent, however, represents consumption. It does not leave you with an asset that you can sell down the road.

Critics often argue that by leaving home prices out of the equation, the inflation metric underestimates the cost of living at moments when home prices are increasing markedly and when it costs first-time buyers more to get a foothold in the market. Some even claim that if the government used the old methodology, its reported inflation rate would be much higher today than it was during the 1980s.

It is true that inflation isn’t perfectly comparable over time because of the change in how housing was measured, said Omair Sharif, founder of the research firm Inflation Insights. But the change would not be enough to make today’s inflation higher than the nearly 15 percent it hit 40 years ago.

“Yes, inflation today would be higher, but by roughly 1.25 percentage points, not the 4 to 5 percentage points people say,” said Mr. Sharif, who last year pulled home price, mortgage costs and home repair data from the 1970s, applied the relevant weights, and did the math on the old numbers to see how much the change in methodology changed inflation.

“It wasn’t a mind-blowing number like a lot of people think it is,” he said.

Another estimate — using calculations used in a paper for The Quarterly Journal of Economics and updated for the newsletter Full Stack Economics — found that including home prices and interest rates instead of rent would have pushed the inflation rate to 11.5 percent in February, the latest date available, up 3.6 percentage points from the official figure that month. That’s more than Mr. Sharif’s estimate but still less than the 1980s.

Others argue that the C.P.I.’s rent measure understates the cost of other types of shelter, pointing out that real-time rent trackers tend to capture rising prices much more quickly. But that is for a simple reason: They track new rents, while the C.P.I. tracks a sample of existing rents, including for people who renew their leases.

“This divergence means that at the moment, the C.P.I. does not do a good job telling the story of how costly it is for an individual or household to secure housing in a new city,” said Jeff Tucker, a senior economist at the real estate website Zillow. Yet the point is to better reflect what prices look like for all consumers, not just ones looking for a new home, he said.

Change No. 2: Economists swap expensive products for cheaper ones

Economists once collected a basket of items — like eggs, milk, shampoo and other items — and simply tracked how much they cost over time, updating the basket only rarely. But that measure was criticized for potentially overestimating inflation because it ignored that consumers adjust their spending both over time and as prices increase.

Economists began to update the basket more regularly about 20 years ago, and the weights are now reset every two years to reflect what people actually spend their money on.

They also tried to account for substitutions. Imagine that the price of cupcakes went up one month. Instead of paying more, a consumer might buy cookies instead — a decent but cheaper dessert alternative — and their monthly costs wouldn’t go up.

They might also buy a container with fewer cupcakes, switch to a cheaper brand or shop at a discount store where cupcakes are cheaper. To factor in that behavior, the government tweaked how it calculates inflation in some categories in 1999, correcting the problem in the eyes of many economists.

Critics sometimes raise a separate point: that product swaps are made between entirely different categories, like using chicken when the price of steak increases. Those larger substitutions are not included in the normal C.P.I. calculation, but are included for a measure called the Chained Consumer Price Index. While the C.P.I. showed prices rose 8.3 percent in April from a year earlier, the Chained C.P.I. was a little more muted, at just 7.8 percent.

Think those changes aren’t enough? There will certainly be more. The Labor Department is still constantly instituting changes to try to make C.P.I. a more accurate reflection of reality.

“It’s a good long-run method,” Mr. Detmeister at UBS said. “Over the course of a couple of months, even over the course of a year, it may differ from what is happening on the ground.”