Is your raise actually a raise?

Every raise arrives as a single number on a letter, and that number is the least interesting thing about it. What decides whether your position improved is the gap between the increase and the change in prices over the same period. A 4% raise in a year when the things you buy got 5% more expensive is a pay cut with better wording. A 2% raise in a flat year is a genuine gain. The letter never mentions any of this, because the letter is written by someone whose job is to make one number sound generous.

The two numbers have names: nominal is what payroll says, real is what the money buys. Almost every vague complaint people have about pay — working harder each year and feeling no further ahead, a decade of respectable increases that never became breathing room — is the gap between those two, compounded quietly for long enough to become visible.

The arithmetic, exactly

A real raise is not the raise minus inflation, though that subtraction is fine as a mental check. The exact relationship divides rather than subtracts:

real raise = (1 + raise) ÷ (1 + price change) − 1

At a 3% raise against 3% price growth the two methods agree at zero. They part company as the numbers get larger: a 10% raise against 8% inflation is not a 2% gain, it is 1.85%. That is why subtraction survives as a rule of thumb, and why it quietly stops being safe in a high-inflation year or across a long stretch of them.

Here is a single year on a $70,000 salary, assuming prices rise 3% over the same twelve months:

The 3% row is the one worth staring at. That raise is neither stingy nor generous — it is nothing at all. It is the number that leaves you exactly where you were, and everything below it is a reduction in pay that nobody is ever required to describe as one.

A decade of small shortfalls

A single thin year is a rounding error you can absorb. The trouble is that raises are not independent events. The increase that came in a point light this year sets the base that next year's percentage is applied to, so the shortfall compounds — the same mechanism that makes saving work, aimed in the other direction.

Three careers, identical apart from the annual increase, with prices rising 3% a year throughout:

1 point belowMatches prices1 point above

What the same salary buys, ten years on

	1 point below	Matches prices	1 point above
Now	70000	70000	70000
2 yrs	68647	70000	71366
4 yrs	67321	70000	72758
6 yrs	66020	70000	74178
8 yrs	64744	70000	75625
10 yrs	63493	70000	77101

A $70,000 base pay measured in today's dollars, with prices rising 3% a year. All three salaries rise on paper; only one of them rises.

The person on 2% raises did not have a visibly bad decade. They were never passed over, never turned down, never told no. Ten years of near-miss increases cost them about $6,500 a year of purchasing power by the end — roughly a month's pay, given away a percentage point at a time, in a sequence of conversations that each felt like a small win.

The 4% line shows how narrow the winning margin is. One percentage point a year, sustained, is a 10% real raise over a decade without a single promotion. And the top row is the one that catches people who stay somewhere comfortable through a quiet stretch: three years with no increase at all, at 3% price growth, is an 8.5% pay cut nobody announced.

Working out your own number

Four things make the calculation honest:

• Compare like with like. Use gross base pay in both years. Folding overtime or a variable bonus into one side and not the other produces a number about your workload, not your pay.
• Match the windows. An increase effective in April should be measured against the twelve-month price change ending in April, not against a calendar-year figure that covers a different stretch.
• Check it again after payroll. Open-enrollment changes land in the same paychecks as the raise, and a health premium increase can absorb most of a modest one. Compare take-home to take-home for the number you will actually notice.
• Weight your own basket. Published indexes describe an average household. If rent is half your budget and it rose 8%, your personal price change is well above the headline and your real raise is worse than the table suggests.

The raise vs inflation calculator runs the first three from your own figures, and the inflation calculator shows what a longer stretch of price growth did to a salary you have held for several years without a serious review.

Run your numbersRaise vs inflation→

Flat pay is a cut, and that is the sentence to use

Pay conversations are conducted in nominal terms because that framing favors whoever is offering the number. "We can do 2%" sounds like a concession that cost someone an argument. Restated in real terms — "we are proposing to reduce your pay by about a percent" — it is the identical fact with the burden of justification moved. You do not need to say it adversarially; you need the other person to know you have done the arithmetic.

Three things are worth separating out loud. The cost-of-living adjustment, which holds you level and is not a reward for anything. The merit increase, which is the only part that reflects your year. And the band you sit in, which is a market question and moves independently of both. Plenty of employers quietly folded the first two into a single merit number years ago, which lets an entire performance conversation resolve at zero real change without anyone naming what happened.

Timing matters more than delivery. Merit pools are set weeks or months before review conversations, so the discussion that actually moves a number happens well before the letter — and the useful questions are logistical rather than emotional. When is the pool decided. What separates a mid-band increase from a top-of-band one. When was this role last benchmarked against what it costs to hire today. Asking those in a normal voice makes you someone who understands the machinery, which is a better position than someone appealing to it.

When raises structurally lag

Sometimes the arithmetic is not going to work no matter how the conversation goes, and it helps to recognize that case early rather than negotiating harder against it for three more cycles.

Internal increases are a percentage of a payroll budget, set annually and slow to respond. External offers are priced against what a role currently costs to fill, which moves continuously. When those two drift apart, the reliable way to get repriced is to be priced by the market — the unglamorous explanation for why moving jobs pays a premium over staying, and why the premium is largest in exactly the roles where internal budgets lag hardest.

The same mechanism produces pay compression: a new hire arrives at the market rate while everyone who stayed received budgeted increments. A newer colleague earning more than you is usually not a judgment about either of you. It is two pricing systems running side by side, and the one that pays better is the one you are not currently in.

What actually resets pay, short of leaving, is a change in what you are being paid for — a scope increase, a promotion into a different band, a credential or skill that makes the old band the wrong comparison. Increments do not reset anything; they modify a number that was set the day you were hired. Relocation is the other lever, and it is really the same calculation with a different price level attached: comparing job offers across cities works through the version where the salary and the cost of living both change at once. If you are weighing contract or hourly work against a salaried offer, the salary converter puts them on the same footing first.

What to do with the part that is real

Only the real portion of a raise can change anything. The rest is compensation for the fact that your existing life now costs more, and it will be absorbed by that life whether or not you make a decision about it. So decide about the real portion before it arrives, while it is still an abstraction rather than a slightly larger deposit.

The cleanest rule is to route the entire real increase into savings the same week the raise takes effect, leaving your take-home spending flat by construction. If your 401(k) contribution is set as a percentage rather than a dollar amount, part of that happens on its own; the rest is one form. Nothing about your month changes, which is exactly why it works — what counts as a good savings rate covers where that lands you relative to the goals, and how compound interest works is the reason a modest redirection made early outruns a large one made late.

The objective is not to win every year. It is to stop losing ground silently for ten of them, which requires nothing more than knowing, each time a number arrives, which of the two numbers it actually is.