Diminishing Sensitivity

What it is

Prospect theory replaced the straight line of expected utility with a value function that has three properties: it is defined over changes from a reference point rather than over final states, it is steeper for losses than for gains, and it is concave in both directions away from that point. The third of those is diminishing sensitivity.

The consequence is that the psychological size of an amount depends on what it is being subtracted from. The first ten pounds of a saving does a great deal of work. The same ten pounds attached to a large base does very little, because the perceptual distance it covers on a flattening curve is small.

The classic demonstration, from Tversky and Kahneman in 1981, is the calculator and jacket problem, in which people are told they can save a fixed sum by driving to another branch of a shop, and the sum is held constant while the price of the item varies. The willingness to make the trip falls sharply as the base price rises. The version that circulates in marketing books, with a clock radio and a television, descends from that problem by way of Richard Thaler's work on mental accounting.

In effect

The practical output is the rule that a discount should be expressed in whatever form makes it look larger, as a percentage on cheap items and as an absolute amount on expensive ones, with the switch happening somewhere around a hundred units of currency. That rule is a straightforward consequence of the curve and it is used constantly in retail.

The same curve explains why an upsell attached to a large purchase meets so little resistance. Leather seats costing a thousand pounds are evaluated against a thirty thousand pound car rather than against a thousand pounds, and on the flattened part of the curve the addition barely registers.

The evidence should be handled in two pieces. The core patterns of prospect theory, including the shape of the value function, were reproduced across nineteen countries by Ruggeri and colleagues in 2020, which is about as good as this kind of confirmation gets. The specific same-saving-different-base demonstration is another matter: a registered replication of the Thaler problems with around a thousand participants, published by Li and Feldman in 2025, found no support for the calculator and jacket effect. Keep the value-function claim and stop using the shopping trip as proof of it.

What it does not say

It does not say that people are indifferent to money at high prices. It says the same increment covers less psychological distance the further it sits from the reference point, which is a claim about the curve and not about indifference.

It does not rest on the survey figures that circulate with it. The eighty-three per cent and thirteen per cent numbers often quoted for the clock radio and the television come from an author's informal hundred-person survey, not from a published experiment, and they are an illustration rather than evidence.

It does not survive as a demonstration in the form everyone uses. A registered replication of the classic shopping-trip problem did not reproduce the effect, and that result has to travel with the example.

And it does not belong to Kahneman alone. The 1979 paper is joint with Amos Tversky, and secondary coverage regularly drops him.

---