Key takeaways
- Inflation is an index change. A single price, even a visible one like gasoline, is not inflation.
- Divide, do not subtract, when converting amounts. The shortcut is close only at low rates.
- Compounding cuts both ways. 3% for ten years is 34% higher prices and 26% less purchasing power.
- Indexes differ. CPI-U, CPI-W, chained CPI, and PCE track different baskets and different populations.
- Your inflation rate is personal. Housing, health care, and transportation weights decide it.
- Real values need one index at both ends. Mixing base years produces nonsense.
What a price index measures
A price index tracks the cost of a fixed market basket over time, with one period set to 100. The Consumer Price Index is built from prices collected for goods and services that urban households actually buy, weighted by how much of the budget goes to each category. When the index moves from 300.0 to 315.0, the basket costs 5% more, and that is the inflation rate for the period.
An index is not a cost-of-living measure, and it is not a forecast. It is a backward-looking average that is revised when new price data arrive. The Fed watches several indexes — the CPI, the producer price index, and the personal consumption expenditures index — because they cover different things and can point in different directions for a few months.
The formula for the rate
The one-period rate is π = (CPI1 ÷ CPI0 − 1) × 100. From 300.0 to 315.0 that is (315 ÷ 300 − 1) × 100 = 5.0%. Over several years, the average annual rate is (CPIn ÷ CPI0)1/n − 1, which is the compound growth rate of the index, not the arithmetic mean of the yearly rates.
The distinction matters. An index that rises 5% one year and falls 5% the next is not back where it started: 1.05 × 0.95 = 0.9975, a 0.25% loss. Prices and rates compound, so averages of percentages are a poor description of a path.
Purchasing power: the reciprocal
If prices rise by π, one dollar buys 1 ÷ (1 + π) of what it bought before. At 3%, a dollar buys 0.9709 of last year’s basket. After n years the factor is 1 ÷ (1 + π)n. At 3% for ten years, 1 ÷ 1.0310 = 0.7441, so the same dollar buys about 74 cents.
The reciprocal is where most errors happen. Two years of 3% inflation is not a 6% loss of purchasing power; it is 1 − 1 ÷ 1.032 = 5.74%. The gap grows with the rate and the horizon, and at high inflation the shortcut is badly wrong.
Worked example: one year and one decade
Suppose the index goes from 300.0 to 315.0 in a year. The rate is 5.0%. A $50,000 salary would need to reach $52,500 to keep pace. A worker who receives a 2% raise in that year loses ground: 1.02 ÷ 1.05 − 1 = −2.86%, which on $50,000 is about $1,429 of purchasing power.
Now take 3% a year for ten years. The index rises by a factor of 1.3439, so prices are 34.4% higher, and $1 buys 0.7441 of the original basket. A $60,000 salary in the first year needs to reach $80,635 to have the same purchasing power ten years later. A car priced at $30,000 becomes about $40,317 for the same model, before any change in features.
The long-run US record is in the same range. CPI-U averaged 172.2 in 2000 and 313.7 in 2024, a cumulative rise of about 82%, or roughly 2.5% a year. In that data, $100 of 2000 spending costs about $182 in 2024, and $100 of 2024 spending buys what about $55 bought in 2000.
Converting nominal prices to real prices
To express an old price in today’s dollars, multiply by today’s index and divide by the old index: real = nominal × (CPInow ÷ CPIthen). A salary of $60,000 earned when the index was 250 is worth $60,000 × (313 ÷ 250) = $75,120 in a period when the index is 313. The same operation in reverse deflates a current figure to an older base.
Two rules keep this honest. Both numbers must come from the same index and the same base period; CPI-U and CPI-W are not interchangeable, and an index rebased to a new year is not comparable to the old levels without conversion. And the result is a real quantity, so label it — a “real” salary is not the number on the offer letter.
Real wages and real returns
Real wage growth is (1 + nominal wage growth) ÷ (1 + inflation) − 1. A 4% raise with 3% inflation is 0.97% real, not 1%. A 2% raise with 3.5% inflation is −1.45% real. The same arithmetic applies to investment returns: 8% nominal with 3% inflation is 4.85% real, while the subtraction shortcut gives 5%.
Taxes make the picture worse because they are usually levied on the nominal gain. If the 8% return is taxed at 15%, the after-tax nominal return is 6.8%, and the after-tax real return is 1.068 ÷ 1.03 − 1 = 3.69%. A taxable account can post a nominal gain and a much smaller real one.
Which index to use
CPI-U covers urban consumers, about 93% of the US population, and is the headline number quoted in the news. CPI-W covers urban wage earners and clerical workers and is used to index Social Security benefits. The chained CPI-U allows the basket to change as relative prices move, which usually produces a slightly lower rate. The PCE price index, produced by the Commerce Department, is the Fed’s preferred gauge because it covers a wider range of household spending.
The choice changes the answer. Use CPI-U for general consumer purchasing power, CPI-W when you are checking a Social Security COLA, chained CPI when a contract or tax provision specifies it, and PCE when you are following Fed policy. Do not switch indexes mid-calculation to get a friendlier number.
Why your inflation rate is different
CPI weights reflect the average urban household, not yours. Shelter is roughly a third of the index, and medical care, transportation, and food carry large weights too. A renter in a fast-growing city, a household with high prescription costs, and a family paying for child care can each face a rate well above or below the national figure.
BLS publishes regional and area indexes, and some metro areas have their own series, which is the closest official approximation to local inflation. For planning, the practical approach is to use the national index as a baseline and test a higher rate for the categories you spend the most on.
Common mistakes
- Subtracting inflation from a nominal rate when converting amounts. Divide: 1.08 ÷ 1.03 − 1 = 4.85%, not 5%.
- Treating a one-month or one-item price move as the inflation rate.
- Comparing index levels from different base years, or mixing CPI-U with CPI-W.
- Assuming a raise keeps pace without checking real wage growth.
- Using CPI as a cost-of-living index for an individual household.
- Forgetting that taxes are charged on nominal gains while inflation erodes the real value.
Practical use
Inflation assumptions show up in retirement projections, lease escalators, child-support and contract adjustments, and any long-term savings goal. A single number is a planning convenience, not a measurement of the future. Run the plan at the baseline and at a rate one or two points higher, and see whether the conclusion changes.
For a goal several years away, inflate the target rather than discounting the return, or do both consistently. Mixing a nominal return with a target in today’s dollars is the most common way a careful plan ends up short.
Frequently asked questions
What is the difference between inflation and a price increase?
Inflation is a sustained rise in the general price level, measured by an index. A price increase is one item changing. A single item can rise far more than inflation, and the index can rise while that item falls.
How do I convert an old price to today’s dollars?
Multiply by the current index and divide by the index for the old period: real = nominal × (CPInow ÷ CPIthen). Both indexes must be the same series and base.
Why is my personal inflation rate higher than CPI?
Because your basket is not the average basket. If housing, health care, or child care takes a larger share of your budget than it does for the average urban household, your rate can run above CPI.
Does inflation hurt savers?
It reduces the purchasing power of cash and of any return below the inflation rate. A nominal gain with a lower real return still loses ground, which is why real returns are the honest comparison.
Which index does Social Security use?
The CPI-W, which covers urban wage earners and clerical workers. That is one reason a COLA can differ from the CPI-U figure reported in the news.
Further reading