PCE in an economic calendar can mean consumption spending or its price index. Those are different quantities: spending can rise because people buy more, pay higher prices, or both. For inflation, choose the PCE price index, its headline or core variant and the comparison period. Then compare the release with the recorded forecast and revisions. The acronym alone does not tell you whether prices fell or where the dollar will move.

Slower inflation, a higher price level

A fictional annual price index is 100, 104 and 106.08 in three successive years. The first annual change is 4%, the second 2%. The pace slows, but the level still rises: 106.08 exceeds 104. Disinflation means slower positive price growth, not a falling price level.

Scroll sideways to see the full diagram

Slower growth with a rising levelThe index rises from 100 to 104, then 106.08. Annual changes fall from 4 to 2 percent. Prices in the model do not return to their earlier level.Fictional index: level, not rate100104106.08Year 1Year 2Year 3Year 2: 104/100 − 1 = +4%Year 3: 106.08/104 − 1 = +2%Growth slows; the price level still rises.
The gentler second segment means slower growth in this model, not a reversal.

When describing inflation, state both its rate and the index level. A return to 2% does not undo the earlier 4% rise; the compounded increase from the first date is 6.08%. A negative change in the same measure over a specified period would indicate a falling level. This arithmetic predicts neither USD nor the Fed’s next decision.

Assumptions: One invented index series, three equally spaced annual dates, no revisions or definition changes. Price growth means a change in the index level; we reproduce neither a particular household’s costs nor actual PCE. Official reference: BEA — PCE Price Index.

What the PCE price index is, and who publishes it

BEA publishes the monthly PCE price index in Personal Income and Outlays. It covers consumer purchases in the United States and spending on their behalf, including health care financed by others. It is different from current-dollar PCE spending and inflation-adjusted real PCE. The statistical spending mix also need not match one household’s bills.

The FOMC’s longer-run 2% objective uses the annual change in the headline PCE price index. It is neither a monthly target nor a promise to hit exactly 2% at each release. Core excludes specified food and energy categories to help examine underlying pressure; it is not a second target or an infallible forecast. BEA’s excluded food category is food and beverages for off-premises consumption. Restaurant services remain outside that excluded category. I lay out the mandate and the logic of how the central bank reacts in more detail in the piece on how the Fed decision moves the dollar.

How PCE differs from CPI

BEA groups CPI/PCE differences into formula, weight, scope and other effects, including seasonal adjustment. PCE uses a Fisher formula and includes purchases on behalf of the personal sector; CPI-U focuses on out-of-pocket urban-household spending. Their coverage and weights can change the contribution of a category’s price. There is no fixed number to subtract from CPI to obtain PCE.

The PCE price index uses a Fisher chain formula with weights from adjacent periods. CPI also allows some substitution within item categories; BLS publishes the separate C-CPI-U to account for substitution between categories. Record the exact series, period and seasonal treatment. A formula that reflects changing expenditure patterns is not a real-time household observation.

Why real components may not add: our two-good model

At the reference date, A and B each cost USD 1 and 100 units of each are bought. Later, A costs USD 2 and its quantity falls to 50; B remains USD 1 and 100 units. Nominal spending stays USD 200.

Fictional two-period quantities and real aggregation
MeasureReference periodLater period
A: price / quantityUSD 1 / 100USD 2 / 50
B: price / quantityUSD 1 / 100USD 1 / 100
Nominal spendingUSD 200USD 200
Sum of separate real valuesUSD 200USD 150
Fisher real aggregateUSD 200About USD 141.42

The quantity ratio at reference prices is (50 + 100) / 200 = 0.75; at later prices it is (2 × 50 + 100) / (2 × 100 + 100) = 2/3. Fisher’s geometric mean is √(0.75 × 2/3) = √0.5 ≈ 0.707107. Multiplying the reference USD 200 gives about USD 141.42. Separate real values are 50 and 100, summing to 150.

The roughly USD 8.58 difference comes from aggregation weights, not an unrecorded purchase. This illustrates one Fisher link; it does not reconstruct PCE. BEA warns that chained-dollar components generally are not additive outside the reference period. Use its published aggregate and growth contributions rather than adding real rows. Nominal values remain additive in this model.

Differences in formula, weights and coverage mean CPI and PCE can show different inflation rates. There is no fixed gap that should be subtracted from one to predict the other. Compare the same reference period and type of change, such as year-over-year core inflation, while keeping their methodological differences in view. The CPI article discusses the other measure.

Monthly, annual and annualised changes

A month-on-month change compares the current index with the previous month; year-over-year compares the same month a year earlier. Annualising a monthly rate asks what twelve identical changes would produce. It is not observed annual inflation or a BEA forecast. Use the same series and seasonal treatment when comparing figures.

Four calculations for one fictional core price index

Assume an index of 120.000 a year ago, 123.600 last month and 123.9708 now, and a recorded 0.20% monthly forecast. These are invented inputs, not a PCE release.

Percentages with different denominators and horizons
MeasureCalculationResult
Month-on-month(123.9708 / 123.600 − 1) × 100%0.30%
Year-over-year(123.9708 / 120.000 − 1) × 100%3.309%, about 3.31%
Annualised monthly pace(1.003¹² − 1) × 100%About 3.66%
Monthly surprise0.30% − 0.20%+0.10 percentage points

Multiplying 0.30 by twelve gives 3.60%, only an approximation to compounding. Do not compare 0.30% monthly directly with the Fed’s annual 2% objective. Use unrounded index values where available; identical displayed rounded rates can hide different changes.

Keep the previous value known before the release, its revised value and the current result separate. If last month’s 0.20% becomes 0.30%, the revision is +0.10 percentage points. A current 0.30% then means no acceleration against the revised month, even though the previously known history was lower. Preserve the data vintage and the forecast provider/time; a later revised table cannot stand in for information available to an earlier decision.

What a PCE print means for the dollar

An inflation surprise can affect policy expectations, alongside employment, income, spending, revisions and other simultaneous information. This is an interpretation to examine, not an equation fixing the currency response. A higher core rate does not guarantee USD appreciation, and a lower rate does not guarantee depreciation. A CPI/PCE volatility comparison requires a defined pair, sample, event window and execution costs; this article performs no such study.

Compare CPI and PCE for matching periods while keeping their different coverage and formulas in view. Agreement can clarify the inflation picture, but it does not guarantee a stronger or more durable dollar signal. If they differ, inspect the components rather than forcing them into one story. Use the economic calendar to check release dates and consult the fundamental analysis section for wider context.

Your next step

  1. Check dated release schedules. Take the CPI and Personal Income and Outlays dates from the publishers’ calendars, preserve the source timezone and convert it for the observation date. Do not infer a fixed middle/end-of-month date or one year-round local hour. Record headline/core and monthly/annual fields before comparing the results.
  2. Compare matching series. Record core CPI and core PCE for the same reference months, horizon and seasonal treatment. Keep each exact series and its data vintage. Inspect their components if they disagree; neither a fixed gap nor agreement alone establishes a currency signal.
  3. Explain agreement or divergence. Compare the CPI and PCE components and methods for the same period. Record what each measure contributes to the inflation outlook; agreement alone is not confirmation of a trading direction.