Transaction Utility

What it is

Standard consumer theory treats a purchase as a comparison between the value of the goods and the price paid. Thaler's 1985 paper builds a different model by combining cognitive psychology with microeconomics. It begins with how people code combinations of gains and losses, using the value function of prospect theory. Transaction utility is the next piece: it explains how a buyer judges whether a purchase feels worthwhile, which the value of the goods alone does not settle.

The third piece is household budgeting, the sorting of money into accounts that the paper's title calls mental accounting. The three fit together. Gains and losses are coded, the deal is judged as a deal, and the purchase is charged to an account. The paper ends by drawing implications for marketing, especially for pricing.

Transaction utility is therefore a component of a model and not a freestanding law. Its job in the model is to let two purchases of the same goods at the same final cost differ in how good they feel, because one was a bargain and the other was not.

In effect

A calculator question from the experimental literature shows the kind of evidence involved. Respondents were asked whether they would travel to another store to save five dollars, and the price of the item varied between a cheap calculator and an expensive one. As reported in the 2025 registered replication by Mengfei Li and Gilad Feldman, the original result was that 68 per cent would make the trip to save five dollars on a calculator costing 15 dollars, and 29 per cent when the calculator cost 125 dollars.

One reading that fits transaction utility is that five dollars off a small price is a larger bargain than five dollars off a large one, so the deal feels better. That reading is ours and not a finding of the replication.

The replication did not reproduce the pattern. In its data, 11 per cent were willing to travel when the calculator cost 15 dollars and 7 per cent when it cost 125 dollars, with no meaningful difference between the two conditions, and the authors classed the problem as an unsuccessful replication. It was one of 17 problems from Thaler's 1999 review tested on 1,007 participants, of which 11 were clearly supported, three received mixed support and three received none.

The 1985 paper draws implications for marketing, especially for pricing. This page does not claim a measured size for the effect.

What it does not say

It does not say that every discount creates a sense of gain. The calculator problem did not replicate in 2025, and the replication sets a limit on how far one hypothetical problem can carry the idea.

It does not say that people are mistaken to enjoy a bargain. Whether the pleasure of a deal is an error or a legitimate part of what a purchase gives is a separate question, and Daniel Kahneman's own caution about similar problems is that feelings such as regret and frustration are real consequences.

It is not a measure of how much people spend, and it does not predict what any particular shopper will do.

---