Comportamiento financiero

The Endowment Effect

Praveen Kujal, Universidad Carlos III.
Vernon L. Smith, University of Arizona.

The Background
The emergence of empirical evidence suggesting divergence between the willingness-to-accept (WTA), for the sale of an object, and the willingness-to-pay (WTP), for the purchase of an object, has resulted in two explanations. One of the explanations was proposed by Kahneman,Knetsch and Thaler (1991) (hereafter KKT), that the discrepancy between the WTP and the WTA may be a manifestation of the endowment effect. Thus, “... endowment effects will almost certainly occur when owners are faced with an opportunity to sell an item purchased for use that is not easily replaceable” (p. 1344). Further they argue that the endowment effect will not apply when the goods arepurchased for resale and not for use; there is no endowment effect for the retail firm, only for the consumer purchasing the firm’s good. Similarly, they argue that the endowment effect does not apply to the exchange of tokens (or rights) to which private redemption values, or induced values have been assigned by the experimenter (Smith, 1976).
The second explanation came from Hanemann (1991)who showed that when close substitutes exist for the good in question (and given positive income elasticity) WTA and WTP can be shown to be very close, and vice-versa.
Given the explanations for the divergence between the average WTA and WTP the experimental literature focused on direct choice tests of these (theoretical) explanations, and there examination in market contexts.
The first choice,and exchange, experiments were run by KKT establishing the endowment effect for Cornell and other (emblem) coffee mugs but not for induced value tokens. These were followed by experiments from Shogren et al. (1994) and Franciosi et al. (1996). Shogren et al. establish that the Hanneman hypothesis under repeat play is robust and that the divergence between the WTA and WTP disappears with repeatinteraction for close substitutes but not for imperfect substitutes. Further, in their experimental setting (distinct from that of KKT) they show that under repeat interaction the endowment effect disappears. Franciosi et al. show that we can observe undertrading in markets even if the WTA-WTP discrepancy is negligible. This is the result of underrevelation of intramarginal units leading to veryflat reported inverse supply and demand so that very small deviations in reported WTA and WTP can lead to undertrading.
The Experiments
Kahneman-Knetsch-Thaler Choice Experiments:
In the typical experiment of KKT an undergraduate class is divided into equal parts. Half the subjects were randomly assigned to the role of buyers and the other half sellers. University emblem coffee mugs(Cornell, Simon Fraser, or University of British Columbia), costing around $6 in the local University bookstore, were then distributed to the sellers, and all the buyers were given the opportunity to examine the mug. The following forms were then executed.
I Will Sell I Will Keep
[Buy] [Not Buy] the Mug
If the price is $0 __________ __________
If the price is $0.50 ____________________
.....
If the price is $9.50 __________ __________

Next, a random price (KKT used the BDM procedure, in Becker et al., 1964) was drawn from the list between $0 and $9.50, and exchanges were conducted by the experimenter on the basis of this price. The typical result was a median selling price that was double the median buying price, an observation that is consistent with theendowment effect. KKT, however, recognized that this procedure did not control for any income effect. This problem was exacerbated by the fact that buyers in their experiments were required to use their own funds while the sellers were given the coffee mugs.
To address the need to control for income effects KKT (pp. 179-80) developed an ingenious variation on the above experiments. Instead of...