A $50 discount from $150 and a $50 surcharge on $50 both produce a $100 price, yet they do not necessarily feel like the same offer. The difference points toward an idea in Daniel Kahneman’s account of risky choices: people often judge outcomes as changes from a reference point, rather than as final amounts alone. In Tversky’s and Kahneman’s research, the way a choice was framed could shift preferences even when the underlying outcomes were equivalent.
Prospect theory is a descriptive account of choices under risk. It examines patterns in how people evaluate possible gains and losses and the probabilities attached to them. It is not a rule that determines what any individual must choose, or a replacement for a broader decision-making guide. Its value is more specific: it exposes features of a choice that can alter its appeal before the arithmetic changes.
Tversky and Kahneman: What the Theory Tried to Explain
Before prospect theory, expected utility theory offered a leading account of choice under risk. It evaluates possible final outcomes by their utility to the decision-maker and weights them by their probabilities. Utility need not equal money: an extra dollar may matter less to someone already wealthy than to someone struggling to pay rent.
In their 1979 paper, the researchers presented choices in which many respondents showed patterns difficult to reconcile with standard expected utility theory. Many of the questions involved hypothetical monetary prospects, so their answers are evidence about responses to those questions, not proof that every person behaves identically with real money.
The paper proposed an alternative in two stages. First, people may edit or simplify the prospects before evaluating them. Second, they evaluate outcomes as gains or losses relative to a reference point and apply decision weights to probabilities. For Kahneman and Tversky, prospect theory was a model of these evaluations, not a general claim that emotions always defeat calculation.
The Reference Point: Where a Gain or Loss Begins
Imagine two employees who end the year earning $70,000. One expected $65,000; the other expected $75,000. The same salary can be experienced as a gain in the first case and a disappointment in the second. Expectations provide one possible reference point; an existing salary, a promised raise, or an available alternative might provide another.
This does not mean the two people are objectively equally well off in every respect, or that disappointment can be inferred from salary alone. Their obligations and opportunities may differ. The example isolates a narrower point: the meaning of a final amount depends partly on the comparison used to evaluate it.
Reference points are useful precisely because they direct attention, but they can also shift. A customer might initially see a $90 ticket as a $10 saving from a $100 list price. After becoming accustomed to paying $80, the same ticket may feel like a $10 loss. The ticket itself has not changed; the comparison has.
Why Losses Often Carry More Weight
The theory’s value function represents outcomes relative to the reference point. It is typically drawn as an S-shaped curve: gains rise more slowly as they grow, while losses become less sensitive to each additional dollar as they deepen. The curve is also generally steeper for losses than for comparable gains, a feature called loss aversion.
A simple illustration is a coin toss offering a $100 gain on heads and a $100 loss on tails. Its expected monetary value is zero, but someone may reject it because the possible loss matters more to them than the equal gain. Another person might accept it, and a person facing a small budget may have sound practical reasons to avoid the loss. The model describes a common tendency; it does not diagnose the reason behind a particular refusal.
Diminishing sensitivity adds another detail. The difference between losing $10 and losing $20 may feel more consequential than the difference between losing $510 and losing $520, even though both increments are $10. The same principle can apply to gains. These are qualitative patterns in the model, not a promise that every dollar amount will be experienced in a fixed proportion.
How Risk Preferences Can Change
Consider a choice between receiving $100 for certain and a 50% chance of receiving $200, with nothing otherwise. Both have an expected monetary value of $100. A preference for the certain $100 illustrates risk aversion in the domain of gains. It is a plausible choice, especially if the money is needed immediately, but it is not a prediction for every reader.
Now consider a sure loss of $100 versus a 50% chance of losing $200, with no loss otherwise. Both have an expected monetary loss of $100. Someone might prefer the gamble, hoping to avoid the loss entirely; that illustrates risk seeking in the domain of losses. The gain and loss examples show the model’s typical curvature, not a recommendation to accept either gamble.
Probability matters too. In the original model, decision weights are not simply identical to stated probabilities. Small probabilities can receive disproportionately large weight, while the move from a high probability to certainty can have a distinctive effect. This helps explain why a small chance may attract attention, but it does not establish that everyone overestimates every rare event. Later research built on Tversky and Kahneman’s work to refine the treatment of probability; the initial account should not be mistaken for the last word.
Framing the Same Outcomes Differently
A frame is the presentation or mental organization of a choice. In their 1981 study, Tversky and Kahneman presented versions of a public-health problem involving a hypothetical disease expected to kill 600 people. In one version, programs were described in terms of lives saved; in another, equivalent outcomes were described in terms of deaths. The reported group preferences differed between the versions.
The comparison is striking because saving 200 of 600 people and 400 people dying describe the same outcome under the stated assumptions. Yet the study compared responses to different presentations; it does not establish that each individual would reverse their own choice, or that the result applies unchanged to every medical decision. Real public-health choices also involve uncertainties and ethical considerations absent from a short vignette.
A frame need not be deliberately manipulative. A store may naturally advertise a discount from a listed price, while a bank may describe a fee for a different payment method. Either description can direct attention toward a different baseline. To check whether two offers truly match, calculate the final price under each option and check for conditions before comparing the wording.
Using the Insight Without Turning It Into a Rule
The most useful question is often “relative to what?” When a new offer feels like a gain or a setback, identify the baseline and write down the possible final outcomes. Then separate the stated probabilities from the emotional weight attached to them. This does not eliminate uncertainty, but it makes the choice easier to inspect.
The prospect theory Kahneman developed with his collaborator is often illustrated with curves that can look universal and exact. They are better read as a simplified account of observed tendencies. Kahneman and Tversky’s model of choice drew on stylized examples, and real decisions can involve learning, social obligations, changing goals, and probabilities that nobody knows. Loss aversion itself should not be treated as a precise constant for every person and situation.
For a practical comparison, consider a $100 purchase with two payment plans. One offers a guaranteed $10 discount, producing a $90 price. The other offers a 50% chance of a $20 discount and otherwise no discount: the final price is either $80 or $100, with an expected price of $90. A person who favors certainty might choose the first.
The same choice can be described from an $80 baseline: pay a certain $10 surcharge, or face a 50% chance of a $20 surcharge and otherwise pay no surcharge. The actual prices and probabilities are unchanged, but the second description emphasizes losses. Prospect theory suggests why that change in frame may affect the choice; it does not say which payment plan the person ought to select.
