How the Inflation Calculator works

The formula

Step by step

1. Both dates are resolved to a month in the committed CPI series — the same monthly series the investment backtest deflates by, so the two tools can never disagree.
2. The ratio of the two CPI readings does all the work. Multiplying by it gives what a sum would cost in the later month; dividing by it gives what a sum still buys.
3. Both directions are always reported. They are reciprocals, they answer different questions, and showing only one is how readers end up with figures that appear to contradict each other.
4. The annual figure compounds the ratio over the exact number of months between the two dates, not the number of whole years.
5. Sub-dollar answers keep four decimal places instead of two: $1 of 1871 money has about 3.7 cents of purchasing power today, and rounding that to 4 cents would overstate it by 8%.

Assumptions & limitations

• CPI measures a national basket of goods and services. It is not your household's inflation rate, which depends on where you live and what you buy — housing, health care and tuition have run well above the headline rate for decades.
• Figures before 1913 come from scholarly price indices spliced onto the official CPI-U series, because CPI-U does not exist before then. They are reliable for the shape of an era and not for arguing over a dollar; the calculator flags any span that touches them.
• No seasonal adjustment is applied, and no attempt is made to correct for changes in CPI methodology over 150 years — the basket of 1890 is not the basket of today, and no long price series escapes that limitation.
• The forward projection compounds a constant rate the reader chooses. It defaults to the measured long-run figure rather than a number picked to look reasonable, but it remains an assumption about the future, not a measurement.