The headline inflation rate

What inflation has really been for the past 55 years.

The Reality Index headline inflation rate: the cost of a fixed 1980 household basket — the same categories, held at their 1980 spending shares — priced year by year with real retail data, compared directly to the official Consumer Price Index.

Latest reading · May 2026

The Reality Index rose 4.95% year over year in May 2026, versus the official CPI’s 4.25% — 70 basis points hotter. Energy led the month (gasoline +40.7%, transportation +10.24%). Since 1980, a fixed household basket now costs $476 per the Reality Index versus $407 per CPI — a cumulative gap of 17.0%.

Annual series, 1980–2025. A fixed 1980 basket of goods and services that cost $100 in 1980 costs $463 per the Reality Index in 2025, versus $391 per the official CPI. The Reality Index ran 18.5% faster than CPI over 45 years — roughly 0.38 percentage points a year. The headline above is the latest monthly reading; this annual series is the long-run context.

Two inflation crises. Extending the data back to 1970 captures the previous inflation crisis (peak: 1980, Reality Index +16.8% year-over-year; CPI +13.6%) alongside the current one (peak so far: 2022, Reality Index +8.7%; CPI +8.0%). On a fixed 1980 basket the 1970s episode was the sharper of the two, and the single largest year-over-year gap between the two measures falls in 1978 — not in the post-2020 years. Over the full 55-year window, Reality Index ran 34.4% faster than CPI.

A fixed basket, real prices. The Reality Index weights each category by its share of household spending in 1980 and holds that basket fixed — a Laspeyres index anchored to the start of the series. This is the honest way to ask what the 1980 cost of living has actually done: fix the basket and let prices move. Two things then differ from CPI. First, we price the basket with independent retail data wherever it exists — Kaiser Family Foundation for family health premiums, NCES for tuition (published sticker price), FHFA for home prices, AAA for vehicles, EIA for gasoline, and BLS Average Price Data for retail food and electricity — falling back to BLS CPI subindexes only where no independent series exists. Second, we hold the 1980 basket fixed while CPI updates its weights over time. An earlier version of the index applied modern (2024) expenditure weights to every past year; that backcast over-weights categories like housing and health care that grew large only later, inflating the historical gap, so it has been retired to the weighting explorer. The gap persists under every defensible weighting scheme; only its size depends on the choice.

Annual series · 1980 → 2025

Reality Index 1980 → 2025

4.63×

$100 in 1980 → $463 today

Official CPI 1980 → 2025

3.91×

$100 in 1980 → $391 today

Reality Index ran

18.5%

faster than CPI over 45 years

Biggest single-year gap

1978

RI 11.5% vs CPI 7.6% (3.9 pp gap)

Cumulative inflation (annual series): Reality Index vs CPI, 1970 → 2025

Both indexes anchored to 1980 = 100 · fixed 1980 basket (Laspeyres), priced with real retail data

Reality Index (real retail prices, fixed 1980 weights) CPI All Items (BLS official)

Reading the chart. The 1970s show the previous American inflation crisis, and on a fixed 1980 basket it is the sharper episode: Reality Index and CPI both spike, with the Reality Index pulling ahead through the 1978–1980 peak. Through the 1980s and 1990s the two lines track closely, the Reality Index sometimes running below CPI. A modest, persistent divergence opens after about 2000 as housing and health care outpace what CPI captures, and the post-2020 period adds several years of Reality Index running roughly 1–2 percentage points above CPI year-over-year — real, but far smaller than the 1970s gap. The 1980–2025 cumulative effect: 18.5% more inflation per Reality Index than per official CPI. The 1970–2025 cumulative effect: 34.4% more.

Pre-1980 methodology disclosure. Independent retail dollar data for individual food items begins in 1980 with BLS's Average Price Data series. For 1970-1979, the Reality Index is constructed from BLS subindex chaining — the same methodology BLS uses to construct CPI itself. The methodological differentiation between Reality Index and CPI begins meaningfully in 1980 when APU retail dollar data starts. We include the 1970s for historical context, but the 1970s portion of the line is methodologically less differentiated from CPI than the 1980+ portion. The KFF substitution for health care premium (vs. BLS medical) begins in 1999. The FHFA home price substitution begins in 1975; for 1970-1974, FHFA is chained backward via the BLS rent subindex.

Where the gap came from, year by year

Below are the years with the largest single-year inflation gap between Reality Index and CPI. Each shows the year-over-year rate per Reality Index, the year-over-year rate per official CPI, and the gap in percentage points.

The bucket weights

These are the fixed 1980 expenditure weights — each category’s share of household spending in 1980 — that determine how each bucket contributes to the Reality Index headline rate. They sum to 100% and are held fixed across all years (a Laspeyres basket). CPI, by contrast, updates its weights over time; holding ours at 1980 is what makes this a clean fixed-basket measure. Alternative schemes (2024 CES, PCE, and others) appear side-by-side in the weighting explorer.

Methodology disclosures

This is the most ambitious composite in the Reality Index project. Several methodology choices are worth flagging honestly so they can be examined and critiqued. The weighting explorer shows what the headline rate looks like under several defensible weighting schemes — 2024 CES, PCE, and others. The gap persists under every one; only its size depends on the weighting choice.

A standing commitment. Reality Index is built to be as accurate as possible, not as accurate as it is today. If we identify a more rigorous data source, a methodology improvement, or a measurement error — whether through our own continued research, reviewer feedback, or critique from readers — we will publish the update, announce the change in /news, and revise the affected pages with a timestamped methodology note. The goal is to be the most honest reference number for what inflation has actually cost American families, which requires being willing to revise the methodology when revision is warranted.

Weights are fixed 1980 expenditure shares (a Laspeyres basket). We hold the 1980 household basket fixed and let prices move, rather than updating weights over time as CPI does. The “you cherry-picked the basket” critique is answered not by the weights but by the items: the Reality Index carries categories that ran slower than CPI (groceries, utilities, communications) at the same prominence as the ones that ran faster (housing, health care, education). An earlier version applied 2024 weights to every past year; it is retired to the weighting explorer.
Independent data wherever it exists. The Reality Index's core methodological principle: where independent retail or consumer data exists for a category, we use it; we fall back to BLS subindexes only where independent data does not exist. Specifically, health care uses Kaiser Family Foundation family premium data (1999+), education uses published sticker tuition (a 13.3× multiple), housing uses a tenure-weighted carrying cost — new-buyer and existing-owner mortgages plus rent, weighted by the homeownership rate, with FHFA prices and Freddie Mac rates feeding the owner side, transportation uses AAA Your Driving Costs, gasoline uses EIA retail data, electricity and food items use BLS Average Price Data (retail dollar series, not CPI subindexes). Dining out, communications, and recreation use BLS CPI subindexes because no comparable independent retail data exists for those categories. The principle: BLS's own data agencies that base their value proposition on getting numbers right are credible; BLS subindexes that are methodologically constructed are the fallback.
Direct taxes are NOT in the headline rate. CPI does not include taxes (they aren't consumer spending), so neither do we, for apples-to-apples comparison. The American Dream Index includes them separately, and the sensitivity analysis shows what the headline rate looks like with direct taxes added as an 11th bucket.
The communications bucket pre-1997 uses CPI All Items as a proxy. The BLS Communications subindex only begins in 1997. For 1980-1996, we assume communications inflated at the broad CPI All Items rate, then anchor to the BLS Communications series at 1997 and let it float forward. This avoids a methodological artifact that would have spiked the headline rate in the late 1990s when the category transformed from "landline phone" to mobile + internet + streaming. The pet and discretionary buckets use the same approach.
Some bucket data is partial pre-1999. The Reality Index extends back to 1980 with full data coverage for some buckets and proxy data for others. Most importantly, in early years we use BLS CPI subindex chaining where direct retail dollar data isn't available. We disclose this transparently for every individual chart on the site.
This is a price index, not a cost-of-living index. The Reality Index measures the cost of a fixed basket. It does not capture what families substitute toward when prices rise, nor whether wages keep pace. The American Dream Index looks at the household income side of the same picture.