Measuring Inflation: CPI and PCE

By Ben Crabb, Regional Economist

Inflation has been one of the big economic stories of the past two years, with prices rising at the fastest rates in 40 years. With the price of food, energy, housing and service all up substantially, inflation’s impact has been felt by just about everyone.


To understand how inflation is measured, this article will explore two of the most commonly used inflation indexes: the Consumer Price Index (CPI) developed by the U.S. Bureau of Labor Statistics (BLS); and the Personal Consumption Expenditures (PCE) price index from the U.S. Bureau of Economic Analysis (BEA).


While the two inflation measures closely track each other, methodological differences result in slightly differing estimates. Ultimately, the PCE price index has some advantages — including better capturing changes in consumer behavior and more accurate weighting of spending categories — over the CPI that make it the preferred index for the policymakers at the Federal Reserve.1


Main Takeaways:

  1. Two of the main indicators of consumer price inflation are the Consumer Price Index from the BLS and the Personal Consumption Expenditures price index from the Bureau of Economic Analysis. The PCE is the Federal Reserve’s preferred measure, while the CPI is what we usually hear about in the media.
  2. The CPI is based on households’ direct out-of-pocket expenditures. The PCE tracks expenditures by businesses on behalf of households.
  3. The PCE is re-weighted more frequently than the CPI (weighting is assigning the degree of influence a variable will have against the whole), so it does a better job of capturing substitution effects (i.e. consumers shifting to less expensive items when others become relatively more expensive). As a result, the CPI tends to not adjust as well and can therefore inflate inflation’s actual impact.
  4. While nominal wages (not adjusted for inflation) in Utah have risen appreciably in recent years, real wages (nominal minus inflation) are not up as much because inflation has countered the nominal gains. Inflation has impacted renters harder than homeowners due to large increases in rental costs. All renters are subject to current prices. Not all homeowners are subject to current prices, as many are locked into past prices.

CPI and PCE tend to move together

To be clear, despite their differences, the CPI and PCE price indexes tend to move together and reflect similar price-level trends. Both the CPI and PCE indexes are essentially trying to do the same thing: measure price changes for a basket of goods and services designed to represent the overall consumption pattern of American households. The figure below shows that over the last ten years, before the onset of high inflation in 2021, the yearly inflation rate by either measure had generally remained around the Federal Reserve’s 2% goal. However, with the onset of elevating inflation in 2021, the two series diverged noticeably.

The divergence has to do with how the two indices are measured and how often they are updated. The PCE price index is able to respond more fluidly to changing consumer behaviors, while the CPI is more tethered to past activity.

 

What is the Consumer Price Index?


The CPI is calculated by the U.S. Bureau of Labor Statistics, an agency of the U.S. Department of Labor. Over time, it measures the average change in prices paid out-of-pocket by consumers for the aforementioned representative basket of goods and services. In assembling the CPI index, price changes for goods and services are weighted (or assigned differing influence) according to their prevalence in consumer spending. For example, prices for shelter account for about 34% of the index because this category makes up a large part of household spending.2 Weight determinations come primarily from the Consumer Expenditure (CE) Survey, a household survey conducted by the U.S. Census Bureau for the BLS with the goal of measuring consumer out-of-pocket purchasing expenditures.3


While the Consumer Expenditure Survey determines the influence or weight of the different goods and services categories tracked by the CPI, the actual prices paid for goods and services are collected by BLS personnel at outlets used in the CE Survey (brick and mortar stores, e-commerce websites, and some secondary sources). Housing prices in the CPI are based on rental housing unit prices collected via personal visits or phone calls. Additionally, homeowner (not renter) information is transformed into “homeowners’ equivalent rent,” an estimate of the implicit rent that owner occupants would have to pay if they were renting their homes.3


The CPI is a so-called “fixed weight” index, where the weights assigned to each of the goods and services are held constant between annual updates. This means that from month-to-month the CPI does not account for ongoing changes called substitution effects, where consumers tend to substitute less expensive items for those that have become relatively more expensive. Because the CPI tends to ignore substitution effects, it tends to overstate inflation by not allowing consumer adjustment to be incorporated.4


Note: Annual re-weighting has been introduced to the CPI as of 2023. From 2002 through 2022, re-weighting happened every two years. Before 2002, re-weighting happened once every ten years.5 The rapid price changes of recent years helped spur the BLS to adopt the annual re-weighting interval for the CPI.

What is the Personal Consumption Expenditures price index?


The PCE price index is another measure of overall consumer price levels. It is produced by the U.S. Bureau of Economic Analysis (BEA), an agency in the Department of Commerce. The PCE price index is constructed primarily from surveys of businesses, in contrast to the CPI which relies primarily on a survey of households. The PCE captures payments made out-of-pocket by households plus payments made on behalf of households, such as medical care services paid for by employer-provided health insurance or by Medicare and Medicaid.6 This is one of the main differences with the CPI, which tracks only out-of-pocket expenditures paid directly by households.7


Another primary difference is that the weights assigned within the PCE price index to the goods and services tracked are adjusted every month. More frequent re-weighting means the PCE more accurately adjusts to changes in consumer behavior due to the previously mentioned substitution effect (i.e. choosing less expensive options when other options increase in price).8 PCE re-weighting is performed with knowledge of price levels in both trailing and future time periods, therefore PCE inflation measurements are subject to revision several times in subsequent months. In contrast, monthly CPI releases are not subject to future revision.


Due to this more frequent re-weighting, the PCE price index tends to provide a more accurate measure of inflation. As such, the Federal Reserve prefers it to the CPI as its measure of inflation. Hence the Federal Reserve's  2% inflation goal is relative to price changes measured by the PCE price index.


Among the principal source data for PCE estimates are the Census Bureau’s annual and monthly retail trade surveys, Service Annual Survey, Quarterly Services Survey, the Census Bureau’s Economic Censuses and the BEA’s International Transactions Accounts. The PCE price index also uses some prices from the CPI in its calculation.6


The differing source data result in differing weights among the major expenditure categories in the CPI and PCE price index. While there are slight differences in the categories used, the table below provides an overview of how the different CPI categories are weighted in the two price indexes as of the latest CPI re-weighting in December 2022:


Table 1: Comparison of PCE vs CPI weights in major expenditure categories, December 2022.

CPI Expenditure category

PCE Weights [%]

CPI Weights [%]

PCE less CPI

All items

100

100

0.0

Food and beverages

13.6

14.4

-2.2

Housing

23.7

44.4

-20.7

Shelter

16.3

34.4

-18.1

Other housing

7.4

10.0

-2.6

Apparel

3.8

2.5

1.3

Transportation

10.3

16.7

-6.4

Medical Care

22.3

8.1

14.2

Recreation

6.8

5.4

1.4

Education and communication

6.4

5.8

0.6

Other goods and services

13.1

2.7

10.4

Sources:

CPI weights: https://www.bls.gov/cpi/tables/relative-importance/2022.htm;

PCE weights: Author’s calculation of weights using BEA Table 2.4.5U-M, Personal Consumption Expenditures by Type of Product, from www.bea.gov 
 

What do these metrics say about inflation in Utah?


No timely state-specific inflation measures are done by any federal or state agency. The BEA publishes only a national PCE price index monthly. There are no additional state- or region-specific measurements published monthly. However, the BLS publishes geographic variants of the CPI that are meant to hone in on regionalized price trends. But again, no state-specific measures.


These geographic CPI variants exist for Census Regions and Divisions, which are broad multi-state regions of the country (e.g. West Region, Mountain Division), and for some very large metro areas. The Salt Lake City metropolitan area is not one of the metro areas for which a CPI variety is calculated, so the most applicable regional inflation measures that speak toward Utah are the “Mountain Division” and “West Region” variants of the CPI.


Both the Mountain Division and the West Region cover a large geographic expanse. The Mountain Division covers eight states (Utah, Arizona, Colorado, Idaho, Montana, Nevada, New Mexico and Wyoming), and the even larger West Region covers 13 states: all eight Mountain states, plus California, Oregon, Washington, Alaska and Hawaii.

Plotting these CPI geographic variants over time shows that West Region inflation closely mirrored the national trend over the last few years, while the Mountain Division has had noticeably higher inflation rates since mid-2021. One factor here is that the West Region includes the west coast, which accounts for a large portion of the United States total. The Mountain Region, on the other hand, is small enough to not have its internals heavily influence the United States average.


Another reason for higher inflation in the Mountain states is the contribution of housing to the CPI. In the latest weightings, rent of shelter (covering rent as well as owners’ equivalent rent) makes up about 34% of the CPI2, and Mountain states have seen housing costs climb higher and faster than the national pace in recent years. For example, the price level of the “Shelter” component of the Mountain CPI has increased 24% since January 20209, while nationally it has increased 15%.10 Looking at a Utah-specific dataset, the House Price Index for Utah from the U.S. Federal Housing Finance Agency has gone up 50% since January 2020 compared to 39% for the nation.


Note the declining housing prices in Utah, and the plateaued national housing prices, since about mid 2022. A cooling housing market should help contribute to reduced inflation in CPI measurements in future months since housing price changes show up in inflation measures with a substantial lag.

 

Both the PCE and the CPI estimate similar levels of shelter expenditures, but shelter only accounts for about 16% of the PCE price index, compared to 34% of the CPI. The reason for this difference is that total consumer expenditures (i.e. the denominator used in calculating 16% vs 34%) estimated in the PCE price index are larger than total expenditures reported in the Consumer Expenditure Survey, which is the basis for CPI weights.11

 

Inflation and wages


With a labor shortage affecting much of the nation, wages have increased significantly, making their own contribution to the recent high inflation. Nominally (before adjusting for inflation), average private sector wages in Utah have increased 29% since 2018.12 After deflating wages by the various metrics discussed — whether CPI, PCE, or Mountain CPI — the story on wage growth in Utah is quite different.


When wages are deflated using the PCE price index, real wages (purchasing power) are up about 10% over 2018 levels. When the CPI is the wage deflator, real wages are up about 7% since 2018. And when the Mountain Division CPI is used, real wages are only up about 2.5% since 2018. 


So which real-wage story is most relevant for Utah wage earners? The inflation impact will vary based on individual circumstances. For renters, the Mountain Region’s CPI measurement is probably the most relevant, as housing prices have climbed faster in Mountain states than nationally. For homeowners with fixed rate mortgages, one of the national-level inflation metrics is probably a more appropriate wage deflator.


An even better measure for long term homeowners would be one not plotted here, the “All items less shelter” version of the Mountain Division CPI which strips out the impact of shelter from the overall measure of inflation.13 The “All items less shelter” year-over-year CPI inflation rate for the Mountain Division was 4.5% for the year ending in February 2023, compared to 6.7% when shelter is included.

On a positive note, regardless of the inflation measure used, average wages for private sector workers in Utah from 2018-2023 have kept pace with inflation, with real wages as of February 2023 above their 2018 levels. But, homeowners are likely benefitting from this more than the renting community.


References

  1. Board of Governors of the Federal Reserve System. February 17 2000. Monetary policy report to the Congress pursuant to the Full Employment and Balanced Growth Act of 1978. https://www.federalreserve.gov/boarddocs/hh/2000/February/FullReport.pdf. Retrieved 3/28/23.
  2. Consumer Price Index: Table 1 (2021 Weights). Relative importance of components in the Consumer Price Index: U.S. city average, December 2022. https://www.bls.gov/cpi/tables/relative-importance/2022.htm. Retrieved 3/27/23.
  3. BLS. Retrieved 3/24/23. Handbook of Methods: Consumer Price Index: Data Sources. https://www.bls.gov/opub/hom/cpi/data.htm 
  4. Federal Reserve Bank of Cleveland. Consumer Price Index Details. Retrieved on 3/23/23 from: https://www.clevelandfed.org/-/media/project/clevelandfedtenant/clevelandfedsite/center-for-inflation-research/consumer-price-data/slideshow-cpi.pdf. Linked from “Consumer Price Data” webpage: https://www.clevelandfed.org/center-for-inflation-research/consumer-price-data.
  5. BLS. 10 January 2023. Commissioner’s Corner: Weight (wait) up! Increasing the Relevance of Consumer Price Index Weights. https://www.bls.gov/blog/2023/weight-wait-up-increasing-the-relevance-of-consumer-price-index-weights.htm. Retrieved 3/4/23.
  6. BEA. December 2022. NIPA Handbook, Chapter 5: Personal Consumption Expenditures.  Retrieved on 3/23/23 from: https://www.bea.gov/resources/methodologies/nipa-handbook/pdf/chapter-05.pdf.
  7. BLS 2011. Focus on Prices and Spending: Difference between the Consumer Price Index and the Personal Consumption Expenditures Price Index. U.S. Bureau of Labor Statistics May 2011: Volume 2, Number 3.
  8. Ballard, James. 08 September 2022. Making sense of inflation measures. St Louis Federal Reserve publications: https://www.stlouisfed.org/publications/regional-economist/2022/sep/making-sense-inflation-measure. Retrieved 3/27/23.
  9. BLS 2023. Data series CUUR0480SAH: Shelter in Mountain, all urban consumers, not seasonally adjusted. https://data.bls.gov/timeseries/CUUR0480SAH1. Retrieved 3/27/23.
  10. BLS 2023. Data series CUUR0000SAH: Shelter in U.S. city average, all urban consumers, not seasonally adjusted. https://data.bls.gov/timeseries/CUUR0000SAH1. Retrieved 3/27/23.
  11. Clinton McCully, Brian Moyer, Kenneth Stewart, “Comparing the Consumer Price Index and the Personal Consumption Expenditures Price Index,” Survey of Current Business 87 (November 2007): 26-33. https://apps.bea.gov/scb/issues/2007/scb-2007-november.pdf
  12. BLS 2023. Data series SMU49000000500000003: Utah Average Hourly Earnings of All Employees, In Dollars, Total Private. https://data.bls.gov/timeseries/SMU49000000500000003. Retrieved 3/27/23.
  13. BLS 2023. Data series CUUR0480SA0L2: All items less shelter in Mountain, all urban consumers, not seasonally adjusted. https://data.bls.gov/timeseries/CUUR0000SA0L2. Retrieved 3/27/23.