How can we accurately measure changes in living standards over time in the presence of price inflation? In this post, I discuss a novel and simple methodology that uses the cross-sectional relationship between income and household-level inflation to construct accurate measures of changes in living standards that account for the dependence of consumption preferences on income. Applying this method to data from the U.S. suggests potentially substantial mismeasurements in our available proxies of average growth in consumer welfare in the U.S.
These insights and results are based on my recent research paper coauthored with Xavier Jaravel of the London School of Economics, which is forthcoming at the Quarterly Journal of Economics.
Background
Economic theory shows that we can approximate changes in living standards using straightforward formulas known as price and quantity indices. These indices combine observed data on changes in the quantities of what we consume and the prices of those items, and are widely used by statistical offices around the world to construct measures of overall inflation in the cost of living and growth in living standards. However, this approach hinges on a crucial assumption—that the composition of what we want to buy does not shift when our incomes do (see, for example, Diewert 1993). Unfortunately, this simplifying assumption, known as homotheticity or income invariance, does not match up with a host of real-world evidence on the dependence of consumption patterns on consumer income, going as far back as the work of Engel (1857). Recent evidence on inflation inequality—the strong relationship between income