Basket of Goods and Services: Categories, Weights, and CPI Uses

The basket of goods and services in the CPI is a fixed, weighted list of everyday purchases that the Bureau of Labor Statistics prices every month to measure how the cost of living changes over time. It is organized into eight spending categories that together cover almost everything a typical urban household buys, from rent and groceries to haircuts and airline tickets. By holding the list relatively steady from month to month, the BLS can separate genuine price movement from shifts in what people choose to buy, and the resulting numbers feed directly into Social Security cost-of-living adjustments, federal tax brackets, and inflation-linked Treasury bonds.¹

The Eight Categories in the Basket

The BLS builds the basket from the Consumer Expenditure Survey, a national survey conducted by the Census Bureau that combines quarterly interviews about large purchases with weekly diaries capturing small everyday spending.² Those spending patterns are then grouped into eight categories. Their approximate share of the overall basket for 2026, based on December 2025 expenditure weights, looks like this:³

Housing, 44.5%. Rent, owners’ equivalent rent (an estimate of what homeowners would pay to rent their own home), and household energy costs like natural gas and electricity.
Transportation, 16.3%. New and used vehicle prices, airline fares, gasoline, auto insurance, and public transit.
Food and beverages, 14.5%. Groceries like milk, chicken, and cereal, plus meals eaten at restaurants.
Medical care, 8.4%. Prescription drugs, doctor and dentist visits, hospital services, and health insurance premiums.
Education and communication, 5.8%. College tuition, internet service, phone plans, and postage.
Recreation, 5.1%. Televisions, gym memberships, pets and pet supplies, and event tickets.
Other goods and services, 2.9%. Personal care products, tobacco, and funeral expenses.
Apparel, 2.4%. Clothing, footwear, and jewelry for all ages.

Those percentages shift slightly each year as spending habits change, but the ranking has been stable for a long time: housing dominates, and apparel sits at the bottom.

How the Basket Is Weighted

A 10% jump in the price of housing reshapes the overall index far more than a 10% jump in apparel, because housing eats up almost half of the average household budget while clothing accounts for about 2.4%.³ To reflect that reality, the BLS assigns each category a weight based on how much consumers actually spend there. If funeral costs rise 20% in a quarter, that barely registers in the overall number. If rents move even a few percentage points, most people feel it, and so does the headline figure.

Starting in January 2023, the BLS began updating those weights annually using a single calendar year of spending data, replacing the old practice of updating every two years based on two years of data.⁴ The annual refresh means the index reacts faster when spending habits shift, like the surge in grocery spending and drop in transportation costs many households experienced during the pandemic.

How the Basket Becomes the Monthly Inflation Number

Each month, BLS data collectors visit approximately 6,000 housing units and around 22,000 retail establishments across 75 urban areas to record actual prices paid by consumers. The total cost of the basket in the current month gets compared to its cost during a base period, which the BLS still sets at the 1982–1984 average.¹

The result is an index number. If the index reads 310, the basket costs roughly 210% more than it did during the base period. A move from 310 to 316 is about a 1.9% increase in the overall cost of living for whatever timeframe you’re comparing. Those monthly and annual percentage changes are what news reports mean when they say inflation came in at a given rate.

The approach has a built-in limitation. It measures price changes for the same basket, not changes in what people actually buy. If rents spike and you downgrade to a smaller apartment, the index still tracks the price of the apartment size you had before. That is deliberate, but it means the published number does not always match what a particular household feels.

Different Versions of the CPI Use the Same Basket

The BLS does not publish just one consumer price index. Several versions apply the same underlying basket to different populations or use different math, and the choice of version has real dollar consequences.

CPI-U: All Urban Consumers

The CPI-U is the broadest and most commonly cited version, covering roughly 87% of the U.S. population, including salaried professionals, part-time workers, the self-employed, the unemployed, and retirees.⁵ When news headlines quote the inflation rate, they are almost always referring to the CPI-U.

CPI-W: Urban Wage Earners and Clerical Workers

The CPI-W is a narrower index covering about 32% of the population. It tracks spending only in households where at least half of income comes from hourly wage or clerical jobs and at least one worker was employed for 37 or more weeks in those occupations.⁵ The Social Security Act specifies that COLAs are calculated from changes in the CPI-W, not the broader CPI-U.⁶ So the spending patterns of working-age wage earners determine the benefit increase for retirees, whose spending on medical care and housing often looks quite different.

Chained CPI-U

The Chained CPI-U uses the same population as the CPI-U but accounts for the way consumers switch to cheaper alternatives when a product’s price rises.⁵ Because it captures that substitution, it typically shows a lower inflation rate than the standard CPI-U. Since the Tax Cuts and Jobs Act of 2017, the Chained CPI-U has been used to adjust federal income tax brackets, standard deductions, and many other tax provisions for inflation. Tax thresholds now rise a bit more slowly each year than they would under the old formula.

R-CPI-E: Research Index for Americans 62 and Older

The BLS also publishes an experimental index called the R-CPI-E, based on spending patterns of Americans aged 62 and older.⁷ Because older Americans typically spend more on medical care and less on transportation, this index tends to run slightly higher than the CPI-U. No federal program currently uses it for benefit calculations. Proposals to switch Social Security COLAs from the CPI-W to the R-CPI-E have surfaced repeatedly in Congress but have not been enacted.

Headline Inflation vs. Core Inflation

Economists often distinguish between headline inflation (the full CPI including every category) and core inflation (the CPI with food and energy stripped out). That might sound like cherry-picking, but there is a practical reason. Food and energy prices are extremely volatile. A cold snap can spike natural gas costs for a month, and a supply disruption can swing gasoline prices by 20% and then reverse. Those swings create noise that makes it hard to see whether the underlying trend is speeding up or slowing down.⁸

Core inflation gives policymakers, particularly the Federal Reserve, a cleaner signal about where prices are actually headed. If headline inflation jumps 0.5% in a month but core inflation barely moves, energy or food probably drove the spike and it may not last. If both are climbing, the pressure is broader and more likely to persist. Both versions draw on the same basket.

What the Basket Ties to in Your Financial Life

The CPI is not just an economic statistic. Several concrete financial adjustments are pegged directly to it:

Social Security benefits. The annual COLA is based on changes in the CPI-W. For 2026, Social Security benefits increased 2.8%.⁹
Federal tax brackets. Income thresholds, the standard deduction, and many other tax provisions are adjusted each year using the Chained CPI-U.
Treasury Inflation-Protected Securities. The principal value of TIPS bonds adjusts based on CPI changes, so if inflation rises, the bond’s face value rises with it.¹⁰
Federal poverty guidelines. The thresholds used to determine eligibility for programs like Medicaid and SNAP are updated annually based on CPI data.
Private contracts. Many commercial leases, union wage agreements, and alimony orders include escalation clauses pegged to the CPI.

Because different versions of the CPI produce slightly different inflation numbers, the choice of index has real dollar consequences. The shift to Chained CPI-U for tax brackets, for example, produces a lower annual adjustment than the old formula would, giving taxpayers smaller inflation relief each year.

How the Basket Stays Current

The basket is not frozen in time. As consumer habits evolve, the BLS removes obsolete products and adds modern replacements. Streaming subscriptions replaced VHS tapes. Smartphones replaced pagers. The geographic sample of areas where prices are collected gets refreshed roughly every 10 years following the decennial census, and one-sixth of the rented housing unit sample is replaced every year using the latest Census data.¹¹

A subtler challenge is quality change. If a television costs the same as last year’s model but has double the resolution, the price has not really stayed flat. You are getting more for the same money. The BLS handles this through hedonic quality adjustments, a statistical method that estimates the dollar value of specific features and separates genuine price increases from quality improvements.¹² Hedonic adjustments apply to electronics like televisions and video equipment, major appliances like refrigerators and washing machines, and clothing categories. For personal computers and cars, the BLS uses cost-based adjustments instead, estimating the production cost of new features directly.

Some economists argue the BLS does not adjust enough for quality gains and therefore overstates inflation. Others say the adjustments are too aggressive and hide real price increases consumers feel at the register. The net effect on the overall index is small, but the debate is not.

Who and What the Basket Leaves Out

The CPI-U covers about 87% of the population, which sounds comprehensive until you look at who sits in the other 13%. The index explicitly excludes people living in rural areas outside metropolitan statistical areas, farm households, people on military installations, those in religious communities, and people in institutions like prisons and hospitals.¹³

Prices are collected from 75 urban sampling areas consolidated into 32 index areas for calculation purposes.¹¹ Regional cost differences can be substantial. State-level price data shows that living costs in the most expensive states run roughly 25% higher than in the cheapest ones. The CPI produces a national average, so it can understate inflation for someone in a high-cost city and overstate it for someone in a low-cost rural area. The BLS publishes indexes for specific metro areas, but those come out less frequently and with less detail than the national figures.

A retiree spending heavily on medical care and a young renter spending heavily on technology and transportation experience meaningfully different rates of inflation. The basket is a portrait of an average household, and no single average captures every household inside it.

1
U.S. Bureau of Labor Statistics. Consumer Price Index – Overview
2
U.S. Census Bureau. Consumer Expenditure Survey (CE)
3
U.S. Bureau of Labor Statistics. Consumer Price Index – January 2026
4
Federal Register. Updating Spending Weights Annually Based on a Single Calendar Year of Data
5
EveryCRSReport.com. The Chained Consumer Price Index: What Is It and Would It Be Appropriate for Cost-of-Living Adjustments?
6
Social Security Administration. Latest Cost-of-Living Adjustment
7
U.S. Bureau of Labor Statistics. R-CPI-E Homepage
8
Federal Reserve Board. What Should Core Inflation Exclude?
9
Social Security Administration. Social Security Announces 2.8 Percent Benefit Increase for 2026
10
TreasuryDirect. TIPS/CPI Data
11
U.S. Bureau of Labor Statistics. Handbook of Methods: Consumer Price Index Design
12
U.S. Bureau of Labor Statistics. Quality Adjustment in the CPI
13
U.S. Bureau of Labor Statistics. Consumer Price Index Frequently Asked Questions