Is Inflation Being Understated?
Measuring prices can be a daunting task. A new method avoids some pitfalls.
- By
- June 16, 2026
- CBR - Economics
Measuring prices can be a daunting task. A new method avoids some pitfalls.
For economists, measuring inflation in the real world is a lot tougher than it might seem at first blush. Sure, reporting the cost of a standard gadget requires little more than tracking its price from one year to the next. But in modern markets, products are constantly changing.
Your cell phone is a prime example. You might have paid roughly the same amount for your latest model as the one you bought a decade ago. But given the difference in features, and how many different kinds of phones are now available, is the comparison really equivalent? Economists typically measure inflation for such evolving products by adjusting for the quality and variety of goods offered.
This method is dauntingly complex and has understated inflation by up to 1 percent, suggests research by Yale’s David Argente, Chicago Booth’s Chang-Tai Hsieh, and University of California at San Diego’s Munseob Lee. They have devised a different way of monitoring inflation that incorporates information about consumer demand for goods rather than only using information about prices.
To track inflation, many economists use the Bureau of Labor Statistics’ Consumer Price Index, which measures the average prices paid by consumers for a representative basket of goods and services. But this leads to two common miscalculations, the researchers write. The first involves inaccuracies in the measurement of the ever-changing quality and variety of products. The second relates to the categorization of new products such as the iPad, which could as easily be dubbed a television as a computer or cell phone. For its part, the Bureau of Economic Analysis’s Personal Consumption Expenditures Price Index (PCEPI)—which measures changes in a broader range of goods and services and is also used to assess inflation—does attempt to adjust for quality changes, but bias remains.
The researchers wondered if they could avoid these miscalculations by using information gleaned from spending data. They considered the evolution of the dairy aisle, where Greek yogurt has taken up more shelf space in recent years. Over the decade between 2007 and 2017, the quality of Greek yogurt rose dramatically as suppliers increased the protein content, added probiotics, reduced sugar, and improved the texture. These improvements did not move prices much but had a dramatic effect on consumer spending. At the outset, regular yogurt varieties accounted for virtually all purchases. Ten years later, buyers were allocating half their dollars to Greek yogurt.
“The key idea is that if you have an accurate measure of the price of one good, you can use changes in the market share to measure the effective price of all the other goods,” says Hsieh.
But, “the empirical challenge is to identify a set of products for which measured prices are reasonably immune to the measurement issues that plague many products,” the researchers write.
From 2007 to 2019, consumers in the United States shifted roughly half of their yogurt spending toward Greek yogurt as that product improved. Yet changes in its price were far less dramatic, which suggests that quality gains may not be apparent in price data alone.
To find other products for which spending changes were informative, they analyzed Bureau of Economic Analysis data on prices and expenditures from January 1959 to December 2019 for about 200 consumer product categories representing about two-thirds of domestic spending. They identified a bundle of 11 product categories, then calculated the weighted average change of the prices of products in the bundle, along with the change in market share, adjusted by the price elasticity of demand. They used an established formula, the constant elasticity of substitution, to combine the products into a single index, which they dub the CES Chosen Price Index.
The index yielded an inflation rate 0.3–1 percentage points higher than the BEA’s PCEPI.
The researchers validated their findings by analyzing a far larger set of products—all 650,000 of them that are tracked in NielsenIQ Consumer Panel Data, housed at Booth’s Kilts Center for Marketing. They then selected about 13,000 barcodes as an informative bundle. An analysis of those indicated that in consumer packaged goods, traditional measures understate the inflation rate by 0.75 percentage points—and confirmed the robustness of the new approach, they write.
When economists seek to correct the biases using the traditional method, they can inadvertently create an even less accurate results, according to the study. By contrast, when the new technique is used to estimate inflation for a set of products where the two types of errors average to zero, it provides a result free of biases. For statistical agencies, the methodology offers a way to improve inflation reporting with readily available price and expenditure data while eliminating the need to detail product attributes and exhaustively track varieties.
David Argente, Chang-Tai Hsieh, and Munseob Lee, “Inferring Prices from Quantities,” Working paper, February 2026.
Your Privacy
We want to demonstrate our commitment to your privacy. Please review Chicago Booth's privacy notice, which provides information explaining how and why we collect particular information when you visit our website.