CalcMyFin

What inflation actually does to money that is sitting still

Inflation does not take money away from you; it quietly changes what the same number of dollars can buy, and the official index that measures this is more specific — and more arguable — than the headline suggests.

Updated September 2026

What the CPI-U actually measures

The Consumer Price Index for All Urban Consumers (CPI-U) tracks the change in prices paid by urban consumers for a representative basket of goods and services. The Bureau of Labor Statistics sorts that basket into more than 200 expenditure categories grouped under eight major headings, including food and beverages, housing, transportation and medical care. CPI-U covers urban consumers making up over 90 percent of the U.S. population, excluding rural non-metropolitan, farm, military-installation and institutional residents. A narrower companion index, the CPI-W, represents roughly 30 percent of the population.

What the index leaves out matters as much as what it includes. Sales and excise taxes tied to purchases are in, and so are government user fees such as water charges and tolls. Income and Social Security taxes are out. So are investment items — stocks, real estate and similar assets are treated as capital, not consumption. BLS is explicit that the CPI is only a conditional cost-of-living index: it does not attempt to price public safety, education quality or the environment.

How the basket is built

BLS collects the prices of about 80,000 items a month. The sample design draws from 75 urban primary sampling units, consolidated into 32 publication areas; crossed with 243 item strata, that yields the 7,776 basic indexes which aggregate into the headline number. Prices are collected each month from about 23,000 retail establishments and about 6,000 housing units.

The weights — how much each category counts — come from the Consumer Expenditure Surveys, and they lag. Those surveys measure consumer units rather than households — a distinction with real consequences for reading spending averages, unpacked in the 50/30/20 budget rule guide. BLS now updates the weights annually using a single year of spending data rather than two, which shortened the average lag between a purchase and its use in the index from 36 months to 24. The 2026 weights are built on 2024 expenditure data, so the basket priced in 2026 reflects how consumer units were spending two years earlier.

What $100 is actually worth now

The figures below use the CPI-U, U.S. city average, all items, not seasonally adjusted — BLS series CUUR0000SA0, published on FRED as CPIAUCNS. Because that series is not seasonally adjusted, each comparison runs from August of the earlier year to August 2026: the same calendar month in both years, so no seasonal pattern is mistaken for inflation. August 2026 is the most recent month published as of early October 2026; it was released on September 11, 2026, and September 2026 figures are not due until October 14. The index stood at 334.980 (1982-84 = 100).

MonthCPI-U index$100 then, in Aug 2026 dollars$100 held as cash, buying power nowCompound annual rate
August 2000172.800$193.85$51.592.58%
August 2010218.312$153.44$65.172.71%
August 2020259.918$128.88$77.594.32%

Read the third column as “prices then versus now” and the fourth as “what idle money kept.” A $100 bill put in a drawer in August 2000 now buys what $51.59 bought back then — a 48.4 percent reduction in purchasing power. The same bill from August 2020 has lost 22.4 percent in six years, because that six-year window contained the June 2022 peak and the fastest price increases in four decades. You can run any pair of months yourself with the inflation calculator.

Why one household’s inflation is not the headline inflation

The headline rate does not describe a particular household. BLS says plainly that the CPI does not necessarily measure any individual’s experience of price change, because the weights describe an average across urban consumers. The spread inside a single month shows why. In the 12 months to August 2026, with the all-items index up 3.4 percent, the components diverged sharply: energy up 16.3 percent, gasoline up 27.4 percent, food up 2.7 percent, shelter up 3.0 percent, and everything excluding food and energy up 2.4 percent.

A household that drives a long commute every day experienced something far above 3.4 percent that year. A household that walks to work experienced something close to the 2.4 percent core figure. Neither is wrong, and neither is the headline.

Owners’ equivalent rent

The largest single line in the index is not food or fuel. As of December 2025, owners’ equivalent rent of residences carried a relative importance of 26.204 percent, with rent of primary residence at 7.840 percent — about a third of the basket between them. Because BLS treats an owned home as a capital good, house purchase prices, mortgage interest, property taxes and improvements all sit outside the index. Instead, the Consumer Expenditure Survey asks owners to estimate what their home would rent for monthly, unfurnished and without utilities; that answer sets the weight, while the actual price change comes from the rental sample. Each sampled rental unit is repriced only every six months, in one of six rotating panels. That is the mechanical reason shelter inflation looks slow to turn when housing markets move fast, and a legitimate criticism of the index.

Substitution

The second standard criticism is that a fixed basket overstates inflation, because people switch toward whatever got cheaper. BLS has partly addressed this: since 1999 most item indexes use a geometric mean formula that allows modest substitution within a category, and a separate Chained CPI-U allows substitution across categories too. From August 2000 to August 2026, CPI-U rose 93.9 percent while the Chained CPI-U rose 81.6 percent — 2.58 percent versus 2.32 percent a year, a gap of about 0.26 percentage points annually that shows up as 12.3 percentage points of cumulative difference over 26 years. One caveat: BLS publishes the Chained CPI-U in preliminary form and revises it three times, with final values posted 10 to 12 months after first release, so the August 2026 chained figure used here will move slightly. Revisions to its 12-month changes have generally run 0.2 percentage points or less.

The weights themselves

BLS publishes what its weight choices are worth in index points. For January 2026, the official index using 2024 expenditure weights came out at 325.252, against 325.364 under the previous weights — a one-month change of 0.37 percent rather than 0.40 percent. Small, but it demonstrates that the index is a construction with defensible alternatives.

The 2021-2023 episode, in real numbers

The 12-month CPI-U increase peaked at 9.1 percent in June 2022, which BLS described as the largest 12-month rise since the period ending November 1981. That month the index rose 1.3 percent in a single month on a seasonally adjusted basis; energy was up 41.6 percent over the year, gasoline 59.9 percent, food 10.4 percent and shelter 5.6 percent. The core measure peaked later, at 6.6 percent in September 2022, the largest since August 1982, while the headline rate had already eased to 8.2 percent.

Cumulatively, the index rose 18.1 percent in the three years from August 2020 to August 2023. The benchmark for that has to be the pace of the two decades before the episode, not an average that contains it: from August 2000 to August 2020 the CPI-U rose 2.06 percent a year, which over three years compounds to about 6.3 percent. The episode ran close to three times that. Put in dollars, $10,000 of cash held across those three years retained about $8,466 of purchasing power.

Nominal versus real, worked out

A nominal figure is the number of dollars. A real figure is that number restated in the purchasing power of a chosen month. The relationship is a division, not a subtraction: real growth equals (1 + nominal) ÷ (1 + inflation) − 1. At an illustrative 6 percent nominal growth and 3.4 percent inflation, the real rate is 2.515 percent, not the 2.6 percent that subtraction suggests. The compound growth calculator shows what that small gap does over decades.

Years$10,000 at 6% nominalSame sum in today’s purchasing powerError from subtracting instead
10$17,908$12,819+$107
20$32,071$16,433+$276
30$57,435$21,065+$533

BLS publishes the cleanest illustration of the distinction. In August 2026, average hourly earnings for all employees were $37.75, up 3.1 percent over the year — a nominal gain. Deflated by the CPI-U, real average hourly earnings came to $11.30 in constant 1982-84 dollars and were down 0.3 percent over the same 12 months. Pay rose; purchasing power did not. The investment growth calculator projects a balance and restates it in today’s dollars, so the two are never confused.

Money that sits still

Three things decide what sitting still costs: how long the money sits, what the price level does during that window, and how soon the money is needed. The trade-off is between accessibility and purchasing-power erosion, and it cuts both ways. Money needed in four months has almost nothing to lose to inflation and a great deal to lose to being inaccessible; the emergency fund calculator and the savings goal calculator frame the same question as how much needs to stay reachable, and by when. Money that sits for a decade with no nominal growth would, at the 3.4 percent rate of the year to August 2026, retain roughly $7,158 of each $10,000; at the 2.58 percent average of the past 26 years, about $7,751. The cost of waiting calculator puts a delay of a given length into dollars.

Whether any particular account or asset produces a nominal return, and whether it suits a given purpose, is a question this page takes no position on; the arithmetic above is indifferent to where nominal growth comes from. Any past price can be tested the same way: deflate it by the CPI-U over the same months and compare. The gold and silver calculator restates past metal prices in 2026 dollars for exactly that measurement; it is not a suggestion to hold anything.

Limitations worth knowing

Used for what it is — a carefully built, publicly documented, arguable measure of average urban price change — the CPI-U is the right deflator whenever the question is what a sum of money will actually buy.

Common questions

What is the difference between CPI-U and the Chained CPI-U?

Both measure consumer price change, but the Chained CPI-U allows for consumers shifting spending between categories as relative prices change, while CPI-U allows substitution only within categories, via a geometric mean formula used for most item indexes since 1999. The difference is small per year and large over time: from August 2000 to August 2026, CPI-U rose 93.9 percent and the Chained CPI-U rose 81.6 percent, or 2.58 percent versus 2.32 percent a year. One caveat on the chained series: BLS publishes it in preliminary form and revises it three times, with final values posted 10 to 12 months later, so the recent chained figure will move slightly.

Why does the CPI exclude house prices when housing is the biggest part of the basket?

BLS treats an owned home as a capital good rather than something consumed, so purchase prices, mortgage interest, property taxes and improvements all sit outside the index. What the CPI prices is the shelter service a home provides, measured as owners' equivalent rent. As of December 2025 that line alone carried a relative importance of 26.204 percent, with rent of primary residence at 7.840 percent.

Why doesn't the headline inflation rate match a household's own spending?

Because the weights describe an average across urban consumers rather than any one household, and because category rates diverge sharply. In the 12 months to August 2026 the all-items index rose 3.4 percent while gasoline rose 27.4 percent, energy 16.3 percent, shelter 3.0 percent, food 2.7 percent, and everything excluding food and energy 2.4 percent. A heavy commuter and a non-driver experienced very different years inside the same headline.

How high did inflation actually get in the 2021-2023 episode?

The 12-month CPI-U increase peaked at 9.1 percent in June 2022, the largest since the period ending November 1981. Core inflation peaked later at 6.6 percent in September 2022, the largest since August 1982. Across the three years from August 2020 to August 2023 the index rose 18.1 percent, against about 6.3 percent if the 2.06 percent annual pace of the two preceding decades (August 2000 to August 2020) had continued.

How is a nominal return converted into a real one?

Divide rather than subtract: real growth equals (1 + nominal) divided by (1 + inflation), minus 1. At an illustrative 6 percent nominal and 3.4 percent inflation the real rate is 2.515 percent, not 2.6 percent. Over 30 years on $10,000, subtracting instead of dividing overstates the result by about $533.

Which CPI series and months suit a long comparison?

The CPI-U, U.S. city average, all items, not seasonally adjusted (BLS series CUUR0000SA0, shown on FRED as CPIAUCNS) is the usual choice, because unadjusted index levels are not revised the way seasonally adjusted ones are. Compare the same calendar month in both years to avoid seasonal distortion. Note that no October 2025 index was published, so any figure that needs that specific month — including the October and November 2025 one-month changes — cannot be computed; comparisons between two other months are unaffected.

Sources

Calculators for this

Inflation calculatorWhat money will be worth, and what things will cost, later.Investment Growth calculatorHow your investments could grow over time.Emergency Fund calculatorYour cash-cushion target and progress.Savings Goal calculatorHow much to save each month to hit a goal.Cost of Waiting calculatorWhat putting off investing could cost you.Gold & Silver Value calculatorWhat your gold, silver or platinum is worth by weight and purity.

Read the other guides, or browse every calculator.