Salience in Experimental Tests of the Endowment Effect

Starting with Knetsch (1989), experiments on the “endowment effect” (Thaler 1980) typically rely on a two-stage procedure. In the first stage, subjects are endowed with a good, such as a mug. In the second stage, the same subjects are given the opportunity to trade this good for another good of similar value, such as a pen. The endowment effect holds that very few subjects chose to trade, sometimes as few as ten percent. In related experiments, subjects state selling prices for their endowment that are much higher than their buying prices for the same good. These patterns are hard to reconcile with standard choice theory, which predicts that about half the subjects would trade and that selling prices and buying prices are similar. The common explanation of this evidence relies on prospect theory’s loss aversion (Tversky and Kahneman 1979). Because the pain of parting with the endowment looms larger in the decision maker’s mind than the pleasure of acquiring a good of similar value (Kahneman, Knetsch, and Thaler 1990), a decision maker endowed with a mug is unwilling to trade it for a pen (or states a high selling price). Recent experimental evidence, however, suggests that loss aversion relative to expectations may not be the whole story. Perhaps the most revealing fact is that the endowment effect is sensitive to the type of goods involved and to the