Anticipated Regret or Endowment Effect? A Reconsideration of Exchange Asymmetry in Laboratory Experiments

Following the literature (Plott and Zeiler 2007; Engelmann and Hollard 2010), the phenomenon is referred to as “exchange asymmetry” to distinguish it from “endowment effect”, which is a theoretical explanation in this article.

Thus, this study’s hard-closure treatment was similar to experiments in which exchange asymmetry is often reported.

To support such interpretations, Knetsch and Wong (2009) and Ericson, Marzilli, and Fuster (2010) showed that exchange asymmetry was sensitive to various treatments that were intended to manipulate expectations regarding final outcomes. Abeler et al. (2011) showed differences in participant behaviour when their expectations over final outcomes were exogenously varied in a different context.

In his words: “Plott and Zeiler’s favored version of Knetsch’s exchange experiment is similarly biased: It does not allow the owner of the good to have physical possession of it, which is crucial to the effect.”(Chapter 27, footnote 297). Other studies have shown that lack of physical proximity could affect the valuation of endowments (Bushong et al. 2010).

These students responded to a request for participation from a pool of around 1,400 students registered for laboratory experiments at FEEL.

One US dollar roughly equalled 6.4 Yuans on the days the experiment was conducted. The informal minimum wage rate in China varies with province and city. As per a business consulting report, “Understand China,” the average minimum wages in Xiamen and Beijing are roughly similar to skilled manufacturing (1,947 and 2,000 Yuans/month, respectively). Wage estimates are listed at http://understand-china.com/?page_id=678. A rough estimate of the minimum wage for unskilled work (e.g. a store attendant) is about 5.5 Yuans/hour in Xiamen (1,100 Yuans/month, 25 working days, and 8 hours/day).

The laboratory manager had raised the fee (prior to phase 2) to adjust for inflation and improve participation rates.

Plott and Zeiler (2007) discussed important articles in the exchange experiment literature. These studies have retained the key features of Knetsch’s (1989) design.

Note that ownership may not be established unambiguously despite its emphasis in language.

Please see examples of these differences in the relevant journal articles in the appendix.

This experiment’s earnings task was similar in nature to participants being asked to complete a questionnaire after or before endowments were given to them.

Our approach was similar to Knetsch and Wong’s (2009), in that they described the endowment as an earned object for showing up for the experiment.

Even if such tasks alter the nature of endowments, as alleged by Plott and Zeiler (2007), they do not bias participant behaviour against exchange asymmetry.

Kahneman (2011) also suggested that physical proximity to endowments (possession) could affect exchange decisions.

Note that endowments were introduced before the alternative object. Thus, asymmetric physical access to objects could create such misconceptions.

Most experiments do not reveal prices to eliminate the impact of prices on exchange and valuation decisions. However, in experiment 7 reported by Kahneman, Knetsch, and Thaler (1990), price tags were not removed from mugs. Despite revealing prices, “undertrading” and “valuation disparity” were reported in their experiment, which suggests that the impact of price revelation on exchange decisions is minimal.

First, note that, at the time of writing this article there is no definitive theory that prescribes the process of reference formulation. Therefore, unique identification of reference points is not possible. In the following, an unknown reference lottery (over final outcomes) is denoted for participant in treatment H as such that is the probability of finally selecting the drink bottle and is the probability of selecting the bookmark. Despite differences in closure conditions, the domain of expectations (over final outcomes) remains the same across treatments. It is not obvious from Koszegi and Rabin (2006) how the guaranteed return/exchange would impact the reference lottery (expectations). Assuming that it does, the following property of the guarantee option holds. Since (a) the guaranteed return/exchange option in treatment S applies symmetrically to both objects and (b) the costs of exercising the guarantee to reverse decisions are equal (for both objects) for a participant, the guarantee does not bias participants’ exchange decisions and expected final outcomes. Thus, even if the guarantee option in treatment S impacts the reference lottery, the net effect is such that it leaves the reference lottery unaltered for a participant. Thus, . If a participant underestimates the costs of reversing his or her decisions, then the underestimation applies symmetrically to both objects, that is, there is no bias in choosing one object over the other. Thus, even in this scenario, the reference lottery over objects remains unchanged in treatment S. In Section 4, other possible effects of these procedures on expectations that may qualify claims made in this study are discussed.

Thus, the test examines final allocations over objects (conditional on endowment) rather than rates of exchange of objects.

A preference bias would tend to move and in the same direction, and, therefore, the difference of proportions would remain unaffected due to this bias alone.

The information in Table 2a is derived from Table 2 but presented differently to allow easy comparison of differences in the final selection of objects conditional on objects earned.

The relevant participant in phase 1 sent an email to staff but did not show up at the laboratory to complete the exchange. In phase 2, the relevant participant expressed a desire to exchange immediately after the session (while the staff members were processing payments for other participants) but did not wait to complete the transaction.

An anonymous referee is credited for this observation.