Money Illusion: Why a Raise Feels Bigger Than It Is

Every amount of money has two values at once, and only one of them is printed on the check. The nominal value is the number of dollars. The real value is what those dollars buy, which depends on prices. Ann's raise of 2 percent in a year with no inflation left her 2 percent better off in real terms. Barbara's raise of 5 percent in a year with 4 percent inflation left her about 1 percent better off. The larger number belongs to the smaller raise.

The nominal figure wins attention because it is the one that exists without any work. It appears in the letter, it is the number a person repeats to a friend, and it needs no second calculation. The real figure has to be computed from a second quantity that arrives separately, late and in the news rather than in the mail. Shafir, Diamond and Tversky argued that people hold both representations and that judgments draw on whichever one the question makes salient, which is why the same two people can be ranked one way on economics and the other way on satisfaction. The split did not require anybody to be confused about what inflation is.

The consequences are easiest to see in wages. A pay cut of 2 percent in a year with no inflation and a raise of 2 percent in a year with 4 percent inflation leave an employee in a similar real position, but the first reads as an insult and the second as a modest win. That asymmetry is one reason nominal wages are hard to cut, which economists call downward nominal rigidity. Fehr and Tyran showed in a laboratory market that even a small amount of individual money illusion produced considerable aggregate inertia after a negative nominal shock, and much less after a positive one.

There is also evidence that the illusion reaches the reward system rather than only the explanation a person offers. Weber, Rangel, Wibral and Falk gave participants prizes in two conditions that were identical in real terms but differed in nominal terms, and found that activity in the ventromedial prefrontal cortex, an area associated with the valuation of rewards, tracked the nominal difference. On its own that study shows a neural correlate, not a cause; it does rule out the idea that money illusion is only a way of talking.

Eldar Shafir, Peter Diamond and Amos Tversky published "Money illusion" in the Quarterly Journal of Economics in 1997. The paper is a series of questionnaire problems given to respondents who were not economists, designed to find out whether people reason about money in nominal or in real terms when both are available.

The first problem is the one reproduced in the demo above. Ann and Barbara start at the same salary; one receives a 2 percent raise with no inflation, the other a 5 percent raise with 4 percent inflation. Respondents were asked one of three questions about them. Asked who was doing better in economic terms, 71 percent said Ann, which is the correct answer in real terms. Asked who was happier, 64 percent said Barbara, whose raise was nominally larger and really smaller. Asked who was more likely to leave her job for another position, 65 percent said Ann.

The pattern across the three questions is the paper's argument in miniature. The three questions went to three separate groups drawn from the same pool: 150 answered the economic question, 69 the happiness question and 139 the job question, so no one person did the arithmetic and then rated satisfaction. The authors took that as evidence that nominal and real representations coexist, with the nominal one easier to evoke, rather than evidence that people cannot compute real values.

Later problems in the paper extended the result to transactions, contracts and the perceived fairness of wage changes, and the authors argued that money illusion helps explain price stickiness, the rarity of inflation-indexed contracts and the way economic discussion is conducted in nominal terms. The questions have been repeated since: Ziano and colleagues ran a direct replication of the first four problems with 604 participants online in 2021 and reproduced them.

The rows below are illustrations of where the nominal figure and the real one come apart, not findings from the studies cited on this page.

The illusion lives in the gap between two numbers, so the counter is to carry the second one.

• Convert before comparing. Any comparison of money across years needs both amounts in the same year's dollars, and an inflation calculator does it in a few seconds.
• Learn the current inflation rate the way you know the weather. A raise, a rate or a rent increase can then be judged on arrival rather than in hindsight.
• Subtract. Write real change is roughly nominal change minus inflation on the same page as the offer. The subtraction is an approximation, close enough at the rates most raises and savings accounts involve; once either figure runs into double digits, divide 1 plus the nominal rate by 1 plus inflation instead, and then subtract 1. The arithmetic is not hard; remembering to do it is the hard part.
• Treat a freeze as a cut when prices are rising. The absence of a visible loss is a feature of the wording, not of the outcome.
• Check the indexing. Pensions, contracts, alimony, child support and long leases either adjust with prices or quietly erode; which one is a question with an answer.
• Expect your own satisfaction to track the nominal number. In the original survey the economic judgment and the happiness judgment came apart, so the way a raise feels is not a reading of what it did.

Money illusion is a confusion between face value and buying power. Several neighbors are about money and judgment without being this.

• It is not the framing effect, where the same outcome described as a gain or as a loss changes which option people pick. Framing changes the description of one quantity; money illusion substitutes one quantity for another.
• It is not mental accounting, the habit of keeping money in separate mental pots so that a dollar from one pot is spent differently from a dollar in another. That is about which account money belongs to, not about what it buys.
• It is not the denomination effect, in which the same sum is spent more readily as small bills than as one large one. That one depends on the physical form of the money rather than on prices.
• It is not simple innumeracy. The group asked in economic terms mostly answered correctly; the group asked about happiness mostly picked the nominally larger raise.