CPI-U Explained: Basket, Calculation, and Limitations

The Consumer Price Index for All Urban Consumers, known as the CPI-U, is the Bureau of Labor Statistics’ monthly measure of the average change in prices paid by urban households for a fixed basket of goods and services. It is the most widely cited inflation gauge in the United States and covers more than 90 percent of the population. As of March 2026, the all-items CPI-U stood at 326.785 on its 1982–84 base, a 2.4 percent increase over the prior twelve months.¹ That single number feeds tax brackets, bond principal, benefit formulas, and millions of private contracts.

Who the Index Covers

The “U” stands for urban, and the population is deliberately broad: professionals, the self-employed, the unemployed, retirees, and people below the poverty line, so long as they live in a metropolitan area. The BLS puts this group at over 90 percent of the U.S. population.²

Several groups sit outside the sample. Rural non-metropolitan residents, farm households, active-duty military personnel, and people in institutional settings like prisons or psychiatric hospitals are excluded.¹ If you fall into one of those groups, the CPI-U is not measuring your spending patterns.

What’s in the Basket

The BLS tracks prices for roughly 80,000 goods and services each month, drawn from eight major spending categories.³ Housing dominates, covering rent, the imputed cost of owning a home, fuel, and furnishings. Food and beverages come next, spanning groceries and restaurant meals. Transportation includes new and used vehicles, gasoline, airline tickets, and insurance. The remaining categories are medical care, apparel, recreation, education and communication, and a catch-all for other goods and services like haircuts and funeral costs.

The specific items in the basket are chosen from detailed Consumer Expenditure Surveys that record what households actually buy, so the mix shifts over time as spending habits change. Each category is weighted by its share of average household spending, which is why a jump in housing prices moves the index far more than the same percentage jump in apparel.

How Prices Are Collected

BLS data collectors contact roughly 23,000 retail establishments and 6,000 housing units across 75 urban areas every month.³ They record prices at supermarkets, department stores, hospitals, gas stations, and other outlets. Most prices are gathered monthly; some smaller metro areas rotate on a bimonthly cycle.

Adjusting for Quality Changes

A laptop today is not the same product a laptop was five years ago, and the BLS has to separate real price increases from improvements in what the money buys. It uses hedonic regression models that estimate the dollar value of individual features like screen resolution, processor speed, or storage capacity. When a new model replaces an older one in the sample, the BLS strips out the portion of the price change attributable to better features so the index reflects pure inflation.⁴

For fast-moving categories like smartphones, the BLS also uses directed substitution. If a phone in the sample is two or more generations old, analysts randomly decide whether to keep pricing it or swap in a newer model, then quality-adjust the transition. Smartphone models get re-estimated twice a year.⁴

Owners’ Equivalent Rent

Most homeowners do not pay rent, which creates an obvious measurement problem. The BLS handles it with owners’ equivalent rent (OER): the amount a homeowner would hypothetically pay to rent the same home, unfurnished and without utilities.⁵ Rather than surveying owner-occupied homes directly, the BLS derives OER from actual rental units, adjusting for differences in structure type, quality, and the aging of the housing stock. Because shelter is the single largest component of the CPI-U, OER has an outsized effect on the headline number, and it can lag behind what buyers and renters feel in real time.

How the Number Is Calculated

All CPI-U values sit on a base period of 1982–84, which is set to 100.⁶ A current reading of 326.785 means urban consumer prices have risen roughly 227 percent since that base.

The math has two layers. At the lower level, the BLS combines individual price observations within narrow categories, say, ice cream in the Philadelphia metro area, using a geometric mean formula. That formula allows for modest substitution within a category: if one brand of ice cream spikes and shoppers switch to a cheaper brand, the geometric mean picks that up rather than treating consumption as fixed.⁷

At the upper level, those basic indexes are aggregated into the overall CPI-U using expenditure weights drawn from the Consumer Expenditure Survey. The geometric mean is not applied across these broader categories, so the standard CPI-U does not capture substitution between them, such as buying more chicken when beef gets expensive.⁷ That gap is what led to the Chained CPI-U.

Seasonally Adjusted vs. Unadjusted

The BLS publishes CPI-U data in two forms. Seasonal adjustment removes predictable patterns that recur each year, like higher gasoline prices in summer or post-holiday clothing sales. For short-term economic analysis, the seasonally adjusted version is preferred because it strips out noise.⁸

For contracts, the unadjusted version is standard. Collective bargaining agreements, pension escalation clauses, and most private contracts tie payments to the unadjusted index. The BLS advises against using seasonally adjusted data for escalation because seasonal factors are revised each year and could retroactively change the basis of a contract.⁸ Treasury Inflation-Protected Securities also use the non-seasonally adjusted CPI-U to adjust bond principal.⁹

CPI-U, CPI-W, and Chained CPI-U

Three related indexes get confused constantly. All three use the same underlying price data, but they cover different populations or apply different weighting rules.

The CPI-W tracks prices for urban wage earners and clerical workers, roughly 30 percent of the population and a subset of the CPI-U group.² It weights categories according to how those households spend, giving somewhat more weight to items like apparel.¹⁰ Annual Social Security Cost-of-Living Adjustments are calculated from the CPI-W, not the CPI-U. The Social Security Administration compares the average CPI-W in the third quarter of the current year to the same quarter in the prior year; if it rose, benefits rise by that percentage.¹¹

The Chained CPI-U (C-CPI-U) covers the same population as the CPI-U but updates its expenditure weights monthly with contemporaneous spending data, so it captures cross-category substitution as it happens.¹² In practice it usually rises a bit more slowly than the standard CPI-U. Since the Tax Cuts and Jobs Act of 2017, the IRS has used the C-CPI-U rather than the traditional CPI to adjust income tax brackets, standard deductions, and other inflation-sensitive thresholds.¹³ Because the Chained index grows more slowly, bracket thresholds rise less each year than they would have under the old formula, which pushes more income into higher brackets over time.¹⁴

Core vs. Headline, and What the Fed Actually Watches

The “headline” CPI-U includes all eight categories. The “core” CPI strips out food and energy because those prices swing sharply on weather, supply shocks, and geopolitics in ways monetary policy cannot directly control.¹⁵ Core gives a cleaner read on the underlying trend.

The Federal Reserve’s 2 percent inflation target is not set against the CPI. When the Fed adopted the target in 2012, it defined it in terms of the Personal Consumption Expenditures (PCE) price index, which covers a wider population, weights healthcare more heavily because it includes employer-paid insurance and government programs, and updates its weights monthly.¹⁶ PCE inflation tends to run lower than CPI inflation. The Federal Open Market Committee still monitors CPI closely, but the number it is targeting is a different one.¹⁷

Where CPI-U Shows Up in Your Finances

Treasury Inflation-Protected Securities adjust their principal based on changes in the non-seasonally adjusted CPI-U. When the index rises, principal grows; when it falls, principal shrinks, though it cannot dip below par at maturity.¹⁸
Federal income tax brackets, standard deductions, and credit phase-outs are adjusted annually using the Chained CPI-U.¹³
Social Security COLAs use the CPI-W, a related but narrower index.¹¹
Poverty thresholds and eligibility cutoffs for various federal assistance programs are updated with CPI data.
Commercial leases, alimony agreements, and union contracts frequently include escalation clauses tied to the unadjusted CPI-U.⁸
Businesses and economists use the CPI-U as a deflator to convert nominal dollar figures into inflation-adjusted terms.

Known Limitations

No single number can perfectly capture the price experience of 330 million people, and the CPI-U has documented blind spots. The best-known critique came from the Boskin Commission in 1996, which estimated the index overstated inflation by roughly 1.1 percentage points per year, identifying substitution bias, new product bias, quality bias, and outlet bias from the growing share of discount retailers.¹⁹

The BLS has addressed several of those concerns since. The geometric mean formula, introduced in 1999, handles within-category substitution. Hedonic models better isolate quality improvements. The Chained CPI-U tackles cross-category substitution directly. Critics on the other side argue the index still understates what people feel, particularly because OER can lag real-time housing costs and because healthcare spending hits retirees harder than the general urban population the CPI-U represents.

When New Data Comes Out

The BLS publishes the CPI report monthly, typically between the 10th and 14th, at 8:30 a.m. Eastern time, covering the prior month’s data.²⁰ The full release, detailed tables, and historical series are available at bls.gov/cpi. Financial markets react within seconds, which is one reason the release is among the most closely watched on the economic calendar.

1
U.S. Bureau of Labor Statistics. Consumer Price Index Summary
2
U.S. Bureau of Labor Statistics. Consumer Price Indexes Overview
3
U.S. Bureau of Labor Statistics. Handbook of Methods – Consumer Price Index Design
4
U.S. Bureau of Labor Statistics. Hedonic Price Adjustment Techniques
5
U.S. Bureau of Labor Statistics. Owners’ Equivalent Rent and Rent
6
U.S. Bureau of Labor Statistics. Consumer Price Index for All Urban Consumers
7
U.S. Bureau of Labor Statistics. Incorporating a Geometric Mean Formula Into the CPI
8
U.S. Bureau of Labor Statistics. Using Seasonally Adjusted and Unadjusted Data
9
TreasuryDirect. TIPS/CPI Data
10
Social Security Administration. Social Security Cost-of-Living Adjustments and the Consumer Price Index
11
Social Security Administration. Cost-of-Living Adjustments
12
Bureau of Labor Statistics. Frequently Asked Questions About the Chained Consumer Price Index for All Urban Consumers (C-CPI-U)
13
Office of the Law Revision Counsel. 26 USC 1 – Tax Imposed
14
Internal Revenue Service. Inflation-Adjusted Tax Items by Tax Year
15
U.S. Bureau of Labor Statistics. Common Misconceptions About the Consumer Price Index – Questions and Answers
16
Federal Reserve Bank of Cleveland. Infographic on Inflation – CPI Versus PCE Price Index
17
Federal Reserve. Minutes of the Federal Open Market Committee, March 17-18, 2026
18
TreasuryDirect. Treasury Inflation-Protected Securities (TIPS)
19
Federal Reserve Bank of St. Louis. Critiquing the Consumer Price Index
20
U.S. Bureau of Labor Statistics. Schedule of Releases for the Consumer Price Index