CPI versus PCE: two measures of inflation

CPI and PCE differ in coverage, weights, formulas and revision practices. Choose the measure that fits your question.

Different baskets answer different questions

The Consumer Price Index (CPI) follows prices for goods and services bought by consumers in its covered population. The personal consumption expenditures (PCE) price index follows prices for a broader set of consumption expenditures, including purchases made on behalf of households. Health care paid by an employer or government program is one reason their scopes differ.

The measures also use different spending weights and index formulas. CPI weights draw on consumer expenditure information; PCE weights use national-accounts consumption spending and can change as spending patterns change. A difference between their inflation readings is not automatically a mistake in either series.

Headline, core and timing

Headline indexes include food and energy. Core versions exclude those categories under each program’s definition. Exclusions can help analysts examine less volatile movements, but food and energy still affect household costs. A core reading is not the inflation rate of a household that buys no food or fuel.

A monthly change and a twelve-month change describe different intervals. Keep adjustment consistent within each ratio. On this site, the default CPI twelve-month view uses unadjusted observations; the separate monthly CPI view can use the adjusted series. PCE views clearly retain the BEA adjustment basis.

Revisions matter

PCE is part of national accounts and its history can change when accounts are updated. CPI-U unadjusted indexes generally are not revised routinely, while seasonally adjusted history can be updated with revised seasonal factors. Do not apply one program’s revision policy to every price index.

For a report, record the exact series code, adjustment, reference month, transformation and snapshot date. Use CPI and PCE side by side to understand their different scopes, rather than silently swapping one for the other in a historical calculation. The comparison workspace retains separate panels and labels.

Where each index comes from

The CPI is produced by the Bureau of Labor Statistics, which publishes CPI data reaching back to 1913. The CPI for All Urban Consumers (CPI-U), the index most often quoted in news coverage, was introduced in 1978 alongside the older index for urban wage earners and clerical workers (CPI-W). The CPI is widely used to adjust Social Security benefits, tax brackets, wages and contracts, so its history is rarely rewritten after publication.

The PCE price index is produced by the Bureau of Economic Analysis as part of the national income and product accounts. It is derived from the same consumption spending data that feed GDP, which ties it to a consistent accounting framework. The Federal Reserve’s longer-run inflation objective, first stated formally in 2012, is 2% as measured by the annual change in the PCE price index. That choice is a policy convention, not a ruling that the CPI is wrong.

Why the two readings usually differ

BEA and BLS have published reconciliation work that splits the gap between the two measures into several effects. A formula effect arises because the PCE index is a chain-type Fisher index that updates weights continuously, while the aggregate CPI is a Laspeyres-type index built on a fixed expenditure basket that is updated periodically. A weight effect arises because the same item, such as shelter, carries a different share of spending in each index.

A scope effect arises from items in one index but not the other, such as employer-paid health insurance or the services of nonprofit institutions in PCE. Because shelter carries a larger weight in the CPI, periods of rapid rent growth tend to widen the gap. A reader who sees CPI inflation above PCE inflation in a given month should look to these structural effects before suspecting a data error.

Common reporting errors

A frequent error is to describe a CPI reading as the measure the Federal Reserve targets, or to compare a CPI value with the 2% objective as if the two were defined identically. Another is to compare a twelve-month CPI change with a monthly PCE change, or to place a core measure beside a headline measure without labeling either. Each mismatch can create an apparent gap that reflects the comparison rather than the economy.

A clear citation names the producer, the index, headline or core, the adjustment basis, the interval and the reference month. For example: “CPI-U, all items, not seasonally adjusted, twelve-month percent change, BLS.” A comparable PCE citation would read: “PCE price index, excluding food and energy, seasonally adjusted, twelve-month percent change, BEA.” Writing both in full makes any difference in basis visible to the reader.

Worked example: how weights alone can open a gap

Suppose two simplified, hypothetical price indexes track the same three categories with identical price changes but different spending weights. These weights are illustrative round numbers, not the actual CPI or PCE weights.

  1. Assume shelter prices rise 5%, medical care prices rise 2% and all other items rise 1% over the same twelve months in both indexes.
  2. Index A weights shelter at 40%, medical care at 10% and other items at 50%. Its inflation rate is (0.40 × 5) + (0.10 × 2) + (0.50 × 1) = 2.0 + 0.2 + 0.5 = 2.7%.
  3. Index B weights shelter at 20%, medical care at 20% and other items at 60%. Its inflation rate is (0.20 × 5) + (0.20 × 2) + (0.60 × 1) = 1.0 + 0.4 + 0.6 = 2.0%.
  4. The gap between the two readings is 2.7 − 2.0 = 0.7 percentage point, even though every individual price change is the same in both indexes.
  5. Raising shelter inflation to 7% while holding everything else fixed lifts Index A by 0.40 × 2 = 0.8 point to 3.5% and Index B by 0.20 × 2 = 0.4 point to 2.4%, widening the gap to 1.1 percentage points.

Takeaway: Different weights can produce different inflation readings from identical price changes. Real CPI and PCE gaps also reflect formula and scope differences, so neither reading is the single correct rate.

Frequently asked questions

Which inflation measure does the Federal Reserve target?

The Federal Reserve’s longer-run objective is inflation of 2% as measured by the annual change in the PCE price index produced by BEA. Policymakers also watch core PCE, the CPI and many other indicators. The CPI is not the target measure, so comparing a CPI reading directly with the 2% objective mixes two differently defined indexes.

Why is CPI inflation usually higher than PCE inflation?

Over long periods the CPI has tended to run above the PCE price index, mainly because of formula and weight differences. The PCE index uses a chain-type formula that reflects substitution between items as relative prices change, and it gives shelter a smaller weight than the CPI. The gap varies from month to month and can occasionally reverse.

Which measure should I use to adjust a salary or contract for inflation?

Many contracts and government programs specify the CPI because it is published monthly on a fixed schedule and the unadjusted CPI-U is generally not revised. If a contract names an index, use exactly that series. For research about consumer prices across the whole economy, the PCE price index may fit better, but its history can change with national-accounts revisions.

Is core inflation more accurate than headline inflation?

Neither is more accurate. Core measures exclude food and energy to reduce volatility and help analysts identify underlying trends. Headline measures include everything in scope and better describe the full change in consumer prices. Choose the version that matches the question and label it clearly.

What can we help you find?

Type at least two letters to start searching.

Browse economic guides Esc to close