How We Measure Inflation in the United States

By Thomas N. Jamerson
Vice President – Client Administration

Inflation in the United States of America can be measured in a number of ways. Day to day, a consumer might measure it with his eyes and memory – grabbing a box of cereal at the grocery store and thinking to himself that “it cost half that 2 years ago” – or with her calculator while settling up the checking account month to month. In these ways, and many more, we each are active price indicators during our normal lives. In terms of economic planning and government benefits policy, though, the agencies of the Federal government use, primarily, one of two indexes to measure inflation: the Consumer Price Index (“CPI”), prepared by the Bureau of Labor Statistics (“BLS”), and the Personal Consumption Expenditures (“PCE”), prepared by the Bureau of Economic Analysis (BEA) [while “price index” is not in its name, PCE is a price index]. An accurate measure of inflation is important for everyone – from the cereal purchaser to senior policy officials throughout the federal government and the Federal Reserve because many issues are decided based on the analysis. Many federal agencies, notably but not limited to the Social Security Administration, use CPI to determine benefits schedules. The Federal Reserve’s Federal Open Market Committee, since 2000, uses PCE to guide its hand in setting interest rates.

How the indexes are produced and what’s included (and not):



CPI looks at the change in the out-of-pocket expenditures of all urban households. It measures the average change in price over time of a “market basket” of more than 200 categories of consumer goods and services, arranged into 8 major groups (Food and beverages; Housing; Apparel; Transportation; Medical care; Recreation; Education and communication; and Other goods and services). The market basket includes day-to-day items such as the cereal mentioned above, milk, coffee, rent, utilities, prescription drugs, some insurance products, pet products, baby clothes and, for some, tobacco products, as well goods and services that are purchased less frequently like an automobile, televisions, college tuition, and funeral services. When analyzing the data, the federal government divides the data into three categories of “food”, “energy”, and “all items less food and energy”. CPI further separates data into commodities and services, and divides commodities into durables and nondurables. Please note that “all items less food and energy” separates out the volatile food and energy prices to determine the measure of underlying core inflation.

The CPI is computed for several geographic areas – but its focus is metropolitan based to include either large metropolitan areas or groups of smaller metropolitan areas in the same region and the expenditure patterns of a sample of urban consumers representing over 90% of the population. Rural nonmetropolitan areas, farm households, people in the Armed Forces and those who are institutionalized are not included.

CPI excludes income tax and other direct taxes, interest costs, finance charges, life insurance, investment items such as stocks, bonds, real estate, and business expenses (it considers a home purchase as an excluded investment item), gambling losses, fines, cash gifts to individuals or charities, and child support and alimony payments.


PCE is used as a mechanism to gauge how much earned income of households is being spent on current consumption for various goods and services and includes a measure of consumer spending on goods and services among households throughout the Nation. It measures the change in goods and services consumed by all households, and nonprofit institutions serving households.

Unlike CPI, PCE estimates are not derived from a dedicated sample, but rather are based on statistical reports from various government agencies, administrative and regulatory agency reports, as well as reports from private organizations. The two leading sources of household expenditure data are the Consumer Expenditure Surveys estimates from the BLS and the PCE from the Bureau of Economic Analysis. While the two measures have similar goals, their approaches are vastly different and, along with the numerous other government and private surveys, produce significantly different accounts of consumer spending.

PCE is not limited to urban individuals but includes households throughout the nation, active military personnel on bases in the U.S. and abroad, those who are institutionalized, and U.S. citizen employees working abroad for U.S. companies. Excluded individuals include students, temporary workers, and foreign nationals residing in the United States working for foreign companies.

Comparing the two

While CPI and PCE can track each other, the two methodologies often show different inflation results over the same time period. Since 2000, annual PCE inflation has been around 0.4 percentage point lower than CPI inflation on average. More recently, July 2023 headline CPI and PCE inflation were close, each roughly 0.2 percent, and therefore a difference of nearly zero, but August headline CPI was 0.6 percent while the PCE came in at 0.4 percent, a difference grew of 0.2 percentage point. The difference has many reasons. As stated above, the data points and collection methodologies differ. Each applies a different weight to the data and adjust differently for seasonal issues. And, even among a specific data point, such as airline prices, CPI calculates a fixed basket of routes while PCE uses passenger revenues and miles flown in its analysis.

Conclusion (well, maybe not)

So which is better? Neither. Each has a core methodology and data set that are used for specific purposes. It is, though, important to understand what underlies each when someone, whether it is a talking head on an investment broad(or pod)cast or the neighbor down the street, pontificates about “inflation.” There is broad data being evaluated in a number of ways. It is up to us to parse their words to determine how the speaker wants to influence our response. With understanding, we can control how we respond.