The Politics of Price Measurement: CPI Is the Sole Measure of Retail Inflation, But the Fed Prefers PCE

In February, total consumer prices and prices, excluding food and energy, rose 0.4%, resulting in the last twelve-month increases of 3.2% and 3.8%, respectively. The consumer price report is the sole direct measure of retail inflation, capturing what people buy for consumption.

However, the Fed favors the PCE deflator, which is not a direct measure but rather a blend of CPI prices with administered prices (Medicare and Medicaid). This preference is based on the belief that the PCE captures what people purchase in real time, reflecting the substitution effect of price change.

But this claim is misleading. Detailed spending data, crucial for accurate measurement, is unavailable in real-time. The choice of the PCE over the CPI for inflation measurement is driven by political considerations, as it tends to produce a lower rate.

The Bureau of Economic Analysis (BEA), the government agency responsible for estimating and publishing the PCE deflator, faces significant challenges in data collection for various product categories. For instance, while BEA has a small set of product details updated monthly, such as motor vehicles, prescription drugs, gasoline, and tobacco, other categories suffer from a considerable lag.

For instance, food store sales are supplemented with annual scanner data, resulting in a one-year lag. Similarly, the yearly e-commerce survey report provides additional product details with a one-year lag.

Detailed consumer expenditures for services are also unavailable when the PCE deflator for a given month is estimated and released. The Census's Quarterly Survey of Services (QSS) is released three months after a quarter ends. However, the QSS provides aggregate spending figures for various service industries and offers few details of how much of it is household spending.

Without detailed data, the BEA is forced to rely on imperfect product-to-industry ratios, often based on data from the several-year-old 2017 Economic Census, to estimate household product and service spending. This reliance on outdated data is a significant drawback, as it uses spending patterns from years past to explain how current price inflation impacts people's spending decisions. This practice undermines the accuracy and relevance of the PCE deflator, raising concerns about its effectiveness as a current measure of inflation.

The CPI has been criticized for what it is and isn't (See a recent article by Larry Summers on CPI missing financing costs for consumption). The same should apply to the PCE deflator. The PCE is not what people think it is. Choosing the PCE over the CPI is a convenient way for policymakers to argue that they are close to or hitting their target, even when the only direct measure of consumer price inflation (CPI) runs much higher.