Kőszegi and Rabin (2006) on Reference-Dependent Preferences

Botond Kőszegi and Matthew Rabin, “A Model of Reference-Dependent Preferences.” Quarterly Journal of Economics 121(4): 1133-1165, 2006. 

·  Utility depends on the consumption bundle and on a reference bundle (u(cǀr) for riskless bundles). Consumption utility, here but one component of overall utility, is the standard utility of microeconomic theory. The other component, “gain-loss utility,” reflects the reference bundle. 

·  Gain-loss utility is assumed to be related to the difference in consumption utility between the bundle chosen and the reference bundle. The model allows for uncertainty both in consumption bundles and in reference bundles. 

·  The reference bundle is not the status quo; rather, it reflects recent beliefs (probabilistic) about outcomes. Fixing the chosen bundle, a “lower” reference bundle leads to higher utility. The endowment effect follows from loss aversion, since the disutility from loss (for an owner) exceeds the utility from gain for a non-owner. People who acquire items expecting to trade will not suffer from the endowment effect, because they have a different reference bundle. 

·  A “preferred personal equilibrium” (PPE) is that (unique) consistent equilibrium with the highest expected utility. PPE reduces to the standard model when there is no uncertainty; reference dependence plays no role in a deterministic environment, as there are no surprises. 

·  Willingness-to-pay for a good depends on the probability that you expect to buy the good and the price you expect to pay. An increase in the likelihood of buying makes for a greater reference “loss” in the event you don’t buy – the “attachment effect.” The more you expect prices to be so low that they will induce you to buy, the more willing you are to buy when the price is higher than expected!

·  The “comparison effect” holds the expected likelihood of purchase constant. In this case, a decrease in the price you expect to pay means that a medium price feels like more of a loss, lowering the willingness to pay the medium price. (You would not buy at high prices in any case). This effect involves a contradiction of sorts with the law of demand: lower expected prices can lead to diminished interest in purchasing. 

·  Taxi drivers and target wages: a driver learns her afternoon wage after she completes her morning shift. If the driver had unexpectedly high morning earnings, she is less likely to drive in the afternoon. Higher expected wages increase the likelihood of working and of staying through the afternoon. As in many other dimensions of economics, whether an event is anticipated or unanticipated leads to large effects on behavior.