To check whether a salary has kept pace, multiply the old salary by the later price index and divide by the earlier one. A $60,000 salary in 2015 needs to be $83,837.02 in 2026 to buy the same amount, using the US consumer price index — a rise of 39.73 per cent, or 3.09 per cent a year compounded.
Anything below that figure is a real-terms pay cut, however it looked at the time. This is the single most useful thing a price index does for an individual, and it is a two-number calculation.
How do you actually run the test?
Four steps, and the only judgement call is which starting year to use.
- Take the salary at the earlier date and the year it applied to.
- Look up the price index for that year and for now.
- Multiply the salary by the later index and divide by the earlier one.
- Compare the answer with what you are paid today.
Worked through: $60,000 in 2020 against an index of 258.81, compared with 331.18 in 2026, gives $60,000 × 331.18 ÷ 258.81 = $76,777.26. A salary that has risen to $72,000 over the same period has lost $4,777.26 of purchasing power despite a 20 per cent nominal increase.
What has the index actually done?
The recent period is unusual, which is why the answer feels wrong to people who remember the 2010s.
| From | $100 becomes | Total change | Annual rate |
|---|---|---|---|
| 1990 | $253.39 | +153.4% | 2.62% |
| 2000 | $192.32 | +92.3% | 2.55% |
| 2010 | $151.88 | +51.9% | 2.65% |
| 2015 | $139.73 | +39.7% | 3.09% |
| 2019 | $129.54 | +29.5% | 3.77% |
| 2020 | $127.96 | +28.0% | 4.20% |
| 2021 | $122.22 | +22.2% | 4.09% |
Read the last column rather than the middle one. The long-run annual rate sits around 2.6 per cent; the windows starting in 2019, 2020 and 2021 run at 3.8 to 4.2 per cent, because they contain the price surge that followed. A salary benchmarked against a long-run average will have fallen behind over exactly the period most people are asking about.
The same method works on any recurring cost. A rent of $1,400 in 2019 that has risen to $1,800 has gone up 28.6 per cent against a general price level that rose 29.5 per cent over the same window — so in real terms it is very slightly cheaper, which is not how it feels and is what the index is for.
Why is the average a compound rate rather than a simple one?
Because each year’s rise applies to the price level the previous year left behind. Prices rose 153.4 per cent between 1990 and 2026 across 36 years. Dividing gives 4.26 per cent a year, which is wrong; the compound rate that turns 130.7 into 331.18 over 36 years is 2.62 per cent.
The simple average always overstates, and the gap widens with the length of the period. It is the same arithmetic that makes compound interest work, running in the direction nobody enjoys.
Why does the index not match your own experience?
Because it measures an average basket and you do not buy the average basket. The index weights housing, transport, food, medical care and everything else by national spending patterns, and your own weights are different — most obviously if your rent moved sharply while the national average did not.
Three specific mismatches account for most of the disagreement.
- Housing. Whether you own with a fixed-rate mortgage, own outright or rent changes your exposure more than any other single item.
- Life stage. Childcare, tuition and medical costs have their own trajectories and appear in some households and not others.
- Substitution. The index assumes people switch to cheaper alternatives as prices move. If you do not, your personal inflation runs higher.
What does this change about a pay negotiation?
It gives the conversation a defensible number instead of a feeling. "I would like more" is a preference; "the same package in 2020 money would be $76,777 today" is an arithmetic claim from a published federal series, and it can be checked.
The figure is a floor rather than a target. It describes what standing still costs, and any argument about progression, added responsibility or market rate sits on top of it.
Questions people ask
Which index is this? The US city average, all items, all urban consumers — the CPI-U published by the Bureau of Labor Statistics, and the series most commonly used for this comparison.
Is the current year complete? No. A year in progress uses the data published so far, so the most recent figure moves as later months arrive.
Does it work for other countries? The method does; the numbers do not. Every statistics agency publishes its own index, and applying a US index to a European salary answers a question nobody asked.
What about house prices? They are not in the index as prices. Owner-occupied housing enters through an estimate of rental equivalence, which is why the index and the property market can diverge for years at a stretch.
Should I use CPI or a wage index? CPI answers what your money buys; a wage index answers what comparable people are paid. They are different questions, and a negotiation usually needs both.
Why does the annual rate differ depending on the start year? Because it is the compound rate across that specific window. A window containing a price surge shows a higher average than one that does not, even when it ends on the same day.
One division answers it, and the answer is the floor rather than the ceiling. The inflation calculator carries the full CPI series back to 1913, the percentage calculator handles the raise arithmetic on top, and the paycheck calculator shows how much of any increase actually reaches your account.