Bureau of Labor Statistics CPI-U, 1913 to 2024.
Inflation is the average rise in prices over time, measured in the US by the Bureau of Labor Statistics' Consumer Price Index for All Urban Consumers (CPI-U) — a basket of goods and services tracked monthly since 1913. To convert any past dollar amount to today's purchasing power: multiply by the ratio of current CPI to past CPI. The math compounds — small annual rates produce large multi-decade gaps. Average US inflation since 1913 has been ~3.2% per year, but the journey has included 1979's 13% peak, 2009's brief deflation, 2022's ~9% surge, and stretches of 1-2% near-stability.
| Era | Average annual inflation | Notes |
|---|---|---|
| 1913-1945 | ~2.0% | WWI/WWII shocks, mid-decade deflation, no formal Fed target |
| 1946-1965 | ~2.0% | Bretton Woods era — generally stable |
| 1966-1982 | ~7.5% | Stagflation. 1980 peak: 13.5% |
| 1983-1999 | ~3.4% | Volcker disinflation, gradual normalization |
| 2000-2019 | ~2.2% | Globalization era, low inflation |
| 2020-2024 | ~4.7% | Pandemic supply chain + fiscal stimulus surge to 9% in 2022 |
Long-run average since 1913: ~3.2%. The Fed's formal 2% PCE target dates to January 2012; before that, "low and stable" was the implicit goal without a numeric target.
A "$2 million retirement" projected in 2055 nominal dollars sounds substantial — but at 3% inflation that's only ~$1M of today's purchasing power. Run all retirement, college-savings, and FIRE math in real (inflation-adjusted) dollars, or apply a discount rate at the end. Mixing nominal and real produces optimistic illusions.
HYSA at 4.3% APY at 3% inflation = real return of 1.3% — barely positive. T-bills yields below CPI produce negative real return (common during 2010s ZIRP era). Holding multi-decade emergency reserves in cash mathematically guarantees loss of purchasing power. The trade-off is liquidity vs growth — for emergency funds the trade is correct; for long-horizon savings, equities and TIPS protect better.
The Fed's 2% target is on PCE inflation, which runs ~0.3-0.5 points lower than CPI inflation due to methodology differences (chain-weighting, broader basket, lower housing share). When news quotes "3.2% CPI" and "Fed has not yet hit its 2% target," the corresponding PCE figure is closer to 2.7% — much closer to target. Always check which measure is being cited.
Sources: US Bureau of Labor Statistics CPI-U series (CUUR0000SA0, all items, all urban consumers, US city average); BLS Handbook of Methods, Chapter 17 (CPI methodology); Federal Reserve FOMC Statement on Longer-Run Goals (2012, updated 2020); Bureau of Economic Analysis PCE methodology; Cleveland Fed median CPI / 16% trimmed-mean CPI; Robert Shiller online historical CPI data; Damodaran historical inflation series.
BLS measure of average price changes for a representative urban-household basket. Includes housing (~33%), food (~14%), transportation (~17%), medical (~8%), other. CPI-U is the standard reference; CPI-W drives Social Security COLA. Released monthly with ~2-week lag. Series since 1913 with revisions including 1983's shift to Owner's Equivalent Rent for housing.
Compounding inflation. Average annual US inflation 1980-2024 was ~3.0%, compounding to ~3.8× cumulative price increase over 44 years. $100 in 1980 = ~$382 in 2024 dollars; conversely, $100 in 2024 buys what $26 bought in 1980. The 1979-80 stagflation peak (~13%) and 2021-2022 spike (~7-9%) drove an outsized share of the increase.
Adopted formally January 2012 — 2% PCE inflation as part of the Fed's dual mandate (price stability + max employment). PCE runs ~0.3 points lower than CPI. Above 2% PCE → Fed tightens. The 2021-2022 surge to ~9% CPI / 7% PCE drove the most aggressive tightening since the 1980s. Through 2025-26 both measures gradually returning toward target.
Both measure prices but differ. CPI (BLS): fixed basket updated every 2 years, ~33% housing share. PCE (BEA, Fed-preferred): chain-weighted basket updated monthly, ~15-20% housing share, broader "consumed by households" definition including employer-provided healthcare. PCE typically prints 0.3-0.5 points lower. Fed's 2% target is on PCE. News usually quotes CPI because released earlier.
Headline includes everything in the basket. Core excludes food and energy — both highly volatile due to weather, geopolitics, oil shocks. Core is a better measure of underlying trends and what monetary policy primarily targets. 2022 example: headline CPI peaked at 9.1%, core peaked at 6.6%. Trim-mean and median CPI (Cleveland Fed) exclude only the extreme components in any given month.
Erodes real (purchasing-power) value of nominal amounts. 4% HYSA at 3% inflation = 0.97% real — barely keeping up. S&P 500's 10% nominal long-run = 7% real. Cash and short bonds typically lose to inflation over decades; equities, real estate, and TIPS are the main classes that beat inflation reliably. Real return = (1 + nominal) ÷ (1 + inflation) − 1. Plan multi-decade goals in real dollars.
TIPS (Treasury Inflation-Protected Securities): US Treasuries whose principal adjusts with CPI-U. Pay fixed real coupon on inflation-adjusted principal. Effectively zero-real-yield insurance. I Bonds (Series I): TreasuryDirect-only savings bonds with composite rate (fixed real + CPI variable). Capped $10k/yr per SSN. Simpler for retail (no broker, no markup) but 12-mo minimum hold and lose 3 mo interest if redeemed before 5 years.
No — modern targets are recent. 1913-1945: WWI/WWII shocks. 1946-1965: Bretton Woods ~2%. 1966-1982: stagflation, 13.5% peak 1980. 1983-2007: Volcker disinflation ~3%. 2008-2020: ~1.8% (some deflation fears). 2021-2022: pandemic surge to ~9%. 2023-2025: gradual return to target. The "normal 3%" expectation is a 1980-2020 averaging artifact.
BLS methodology adjustments. Hedonic: when product quality improves (faster computers), BLS adjusts price for the quality increase. Substitution bias: when consumers switch to cheaper alternatives (chicken instead of beef), fixed-basket CPI overstates true cost-of-living impact. PCE addresses substitution via chain-weighting; CPI partially via biennial basket updates. Critics argue these understate experienced inflation; defenders argue they correct genuine measurement errors.
Several reasons. (1) CPI is an average; specific demographics experience differently. (2) Salience bias: rising prices more visible. (3) Cumulative effect: 25% rise over 5 years feels jarring even at 4.5% annual. (4) Sticky baseline: people remember pre-inflation prices for years. (5) Asset price inflation (housing, education) partly excluded — OER captures rental but not home-buying prices.