MNI INTERVIEW: New Inflation Baskets To Rebalance In Real Time

Real-time calibration of inflation baskets using consumer transaction data could be key to setting monetary policy in the future, by adjusting for changes in the quantity and even of the quality of items purchased, an economics professor leading a joint U.S.-Census-Bureau project to re-engineer official statistics told MNI.

The failure of traditional inflation data to account for improvements in quality has led to a persistent small overstating of price increases, said Matthew Shapiro, a professor of economics at the University of Michigan who has been leading the RESET project to incorporate scanner and other real-time electronic data into official statistics in collaboration with the Census Bureau and the University of Maryland.

Traditional inflation baskets also fail to capture changes in behaviour, such as when modifications in prices prompt consumers to shift their purchases from one good to another.

"These techniques would help central bankers and other policymakers be informed of what's happening faster and potentially more accurately. That could help them make adjustments. They would also be subsidiary benefits. They would get a better sense of where inflation is coming from across goods and sectors," Shapiro said in an interview in London, where he presented a seminar attended by Official for National Statistics and Bank of England officials.

COVID EXPOSED DATA INADEQUACIES

Statistics bodies have begun to incorporate scanner data from grocery stores and elsewhere, after sharp shifts in consumption patterns during Covid lockdowns highlighted the inadequacy of existing data series, with inflation baskets in the UK and elsewhere adjusted only annually.

"There is progress across the world. Many, many countries are using data like this, at least from grocery store prices. New Zealand and Australia for example. have been doing this in terms of the price measurement,” Shapiro said. But, he added, ”as far as I understand it no one is doing the full simultaneous price and quantity measure.” (See MNI INTERVIEW2: Ex RBA Board Member Argues For Real-Time Data)

A striking example of misleading data came when egg prices in the U.S. surged following a bird flu outbreak, but held steady in PCE data because of mistaken assumptions about spending on eggs as a share of the broader dairy category in which they were included, Shapiro noted.

"It's basic economics when the price is lower individuals will shift to the lower price if one actually gets to measure the consumption basket contemporaneously," he said.

RESET is also looking at ways to more quickly and accurately capture changes in the quality of products, which can be influenced by relative prices as manufacturers adjust recipes or alter suppliers. Whilst hedonic measures of inflation already attempt to capture improvements in quality in areas like computing, discounting price increases to account for higher levels of memory and performance, these are insufficient, said Shapiro, who is looking at how to adjust inflation by incorporating data on the components used to make goods.

"Quality change is pervasive. On average things are getting better, not just computers. Computers are an example when things are getting better faster. What is important to add is that quality change is on average positive but it is not necessarily positive in every instance," he said.

INFLATION OVERSTATED

Looking at grocery store data , Shapiro's research found that inflation was overstated in "the order of a quarter of a percentage point per year for measured quality changes for the products" which "is not huge in any one year" but over time accumulates to significant levels.

One key area for research is whether overestimates of price increases and quality adjustments accelerated during the recent period of elevated inflation, he said.

"The honest answer is we don't know. I think that actually underscores the importance of what we are actually doing. It is of much more concern when Inflation is high and increasing.”