Comparing apples and oranges

Sticky or not that Sticky? A tale of two divergent inflation measures

Sources: U.S. Bureau of Labor Statistics, U.S. Bureau of Economic Analysis, Haver Analytics, DWS Investment GmbH as of 4/29/24

By contrast, weights for the PCE measure tend to reflect substitution effects relatively quickly, as they are based on what consumers actually buy.^[1] When relative prices change, consumers naturally tend to shift from more expensive goods and stores to cheaper ones. Think of how sensitive your fruit purchases are, if, say, a bad harvest drives up the relative price of oranges compared to all other fruits.

Other differences between the two measures include the fact that CPI is survey based, relying mostly on what consumers in urban areas say they have spent money on. By contrast, the PCE index uses actual spending. It includes rural consumers and more accurately captures costs such as health care, for which many consumers do not pay themselves and may not even know the total cost. 

All this can result in big divergences. For example, airfares pushed up the CPI reading for February, based on prices for a fixed set of flight routes, while PCE readings were much lower.^[2] Rents have an outsized impact on the CPI, they are used to determine the cost of owner-occupied houses, as well as properties actually rented. Costs of shelter account for roughly 30% of the CPI basket but have only a weight of about 15% in the PCE. Prices for shelter proved to be a constant nuisance for economists after the pandemic. Usually, the shelter component in CPI (or housing in PCE) follows the various available house-price and rent indices with a quite robust time lag of 12 months. But when it comes to the economic impact, it is worth keeping in mind that a lot of what the measures concerns housing costs in urban areas that are simply imputed, rather than what consumers – including rural ones - actually spend on housing.

None of which is a reason for the Fed to prematurely declare victory. CPI is widely watched, and hence important in anchoring inflation expectations. Moreover, plenty of questions remain as to what caused inflation to get out of hand in the first place, as the Financial Times Martin Wolf recently pointed out in a remarkable column.^[3] “The reality is that forecasting failures by markets and central banks alike are quite common, because no one really knows what the future will hold, let alone how complex systems, such as whole economies, might react to a shock in the real world,” explains Christian Scherrmann, U.S. economist at DWS. “The best we can hope for is that by paying careful attention to all incoming data, we avoid such sharp swings in assessments as markets have experienced in recent months. The Fed is right to be patient.”