Reference-dependent preferences
Also known as: Koszegi–rabin model
A formal model where utility depends on outcomes relative to rationally expected reference points.
What it means
Reference-dependent preferences is the broad class of models, formalized influentially by Koszegi and Rabin, in which a person's utility has two components: a standard consumption utility and a gain-loss utility measured against a reference point. Their key innovation was to derive the reference point endogenously from the agent's recent rational expectations about outcomes, rather than treating it as the status quo or assuming it exogenously. This makes the reference point a forward-looking equilibrium object: what you expect shapes what counts as a gain or loss, and your expectations must be consistent with your behavior. The framework rationalizes the endowment effect, attachment to expected outcomes, and how planned purchases feel different from surprises, and it yields sharp, testable predictions absent from earlier ad hoc reference-point assumptions. It is the leading attempt to put prospect theory's reference dependence on rigorous, predictive footing.
Examples
Someone who expected to buy a car feels a loss if the deal falls through, while someone who never expected to buy feels nothing — their differing expectations set different reference points.
Experienced traders show little endowment effect: they arrive expecting to part with their stock, so handing it over is no loss — unlike the collector who never planned to sell.
A commuter who has already pictured her morning flat white feels a real loss when the café is shut; a passer-by who never expected one walks past unbothered.
First described in Koszegi & Rabin (2006, 2007).