Explaining the major paradoxes and anomalies of single choices under risk and uncertainty

 

Study: Plott, C., R. & Zeiler, K., (2003). The Willingness to Pay / Willingness to Accept Gap, the "Endowment Effect", Subject Misconceptions and Experimental Procedures for Eliciting Valuations.

 

 

 

 

 

 

 

Introduction:

 

    Plott and Zeiler explored the gap between the Willingness to Pay and the Willingness to Accept, and wanted to question its link with the endowment effect.

 

The Willingness to Pay / The Willingness to Accept:

 

    The Willingness to pay (WTP) / Willingness to accept (WTA) gap refers to the different amount of money individuals give to lotteries they can either buy (WTP) or sell (WTA), considering in this latter case that they are owners of the lottery. Fundamentally, this amount should be the same or close as far as rationality is at stake.

 

The Endowment effect:

 

    The endowment effect is a psychological effect that seems to account, in this precise case, for the gap between WTP and WTA since it is based on people’s preferences and decision-making exhibited by Kahneman and Tversky in their prospect theory (1979).

More generally, the endowment effect appears to be a robust anomaly to what economic theories logically predict. Basically, the endowment effect can be summarized as the fact that people value more an object they possess than another one they do not own.

 

Aim of present paper:

 

    Plott and Zeiler aimed to perform several tests to seek whether there was a consensus on the accuracy of the endowment effect or if it does not exist, but only in some cases; thus proving that there has been some misinterpretation of results in past experiments. In the wake of a non-always existing phenomenon, Plott and Zeiler research could then question the WTP/WTA gap and the linked endowment effect as being or not a true “feature” of human decision process and preferences. Authors thought that when results are not convincing, there might has been a misconception for subjects to understand the task they were asked for. In order to get rid from this problem, they tried to use a “revealed theory methodology” to control for subjects confusion, keeping in mind that controlling for subject misconceptions cannot be supposed, it will be an ex-post backward assumption.

 

    Then the main question was: if Plott and Zeiler use every tools existing in the literature to control for subject misconceptions, will there still be a WTP / WTA gap?

 

    They conducted their experiment under two procedures. First by replicating procedures where the gap is observed. Second, by controlling for subjects misconception (“in the sense of the union of procedures for control found in the literature”) and seeking if the gap still exists. If not, then there has been some incorrect interpretation of the gap. Conversely, if the gap is related to subject misconceptions, authors will use the union on controls that exist in the literature to eliminate the gap. The literature provided several experiments that control, more or less, for this misconception. For example, one step (experiment #1 in the current study: non-incentive compatible mechanisms) was to reveal WTP and WTA by asking participants if they want to buy/sell the corresponding lottery at a pre-fixed price. In that case, they do not have to provide the price. Other techniques (sealed bid one price) consist in gathering bids from buyers and sellers and deriving then the supply and demand curve where the intersection represent the market price. Alternative mechanisms were based on double auctions, Vickery auction or the Becker, Degroot, Marschak (BDM) procedure for example. For a convenience purpose, we will not provide more details on these techniques in order to focus more on Plott and Zeiler findings and results.

 

The model:

 

    It is stated that if the WTP/WTA gap is the reflection of an endowment effect, then it should be observed in the current experiment. Plott and Zeiler made their 74 participants perform at each session: 2 training rounds (unpaid), 14 paid rounds (lotteries) and 1 last paid round (with mugs as reward). 

Here is the model they used to evaluate the WTP/WTA gap:

 

    1: Subjects are explained in details the mechanism and how they will have to assess lotteries, in order to get rid of so-called strategic behaviors[1]. To ensure that participants will give their actual valuation, a demand-revealing elicitation device is used, such as the BDM procedure.

 

    2: They have a training and practice sessions (with tutorial[2] and asking questions) using the implemented mechanism.

 

    3: Participants took part in binding outcome experiences that consists in exposing them the consequences of their decisions before experimenters record their choices. Subjects received incentives related to the consistency in the decisions they made before the gap was measured. This mechanism enabled to keep an internal consistency in participants’ answers and to ensure that there are no more strategic or random (due to misconception) responses during the task.

 

    4: Anonymity, in order to emancipate from the concern of how others see you when you behave, as fair as money is concerned, in the tasks.

 

Results:

 

    Based on Horowitz et al. research (2000), the WTA is frequently higher than the WTP and the ratio WTA/WTP is higher than expected in basic economic theories. This gap between both wills has been attributed to the responsibility of the “endowment effect”, according to Knetch, Tang and Thaler (1998).

 

    On the other hand, Plott and Zeiler ended up with the conclusion that, under certain circumstances, “there is no support for the hypothesis that WTA is significantly greater than WTP”. These four circumstances that were implemented in their experimental were described above in their model (with the BDM mechanism chosen to elicit both wills and binding outcome experiences). Authors explain that their finding does not seem to exist in other experiments (the gap does not exist, according to them), because no experiment design to study this gap has fully implemented the controls provided in the literature. Indeed, some statistical tests have been performed and revealed that in each experiment (experiments #1 and #3), where the null hypothesis that WTP ≥ WTA is tested against WTA > WTP, the first one cannot be rejected. Another striking result is that the gap does not exists – because the difference between WTP and WTA is not statistically significant – even in the experiment (experiment #2) where there was no binding outcome experiences.

The main conclusion is so that teaching well to participants, thanks to training and practicing (binding outcome experiences are even not necessary), is enough to get rid from subjects’ misconceptions.

 

    Eventually, the income effect[3] was tested. This effect is derived from the fact that participants earn money during the experiment and they can play with this money, which does not come from their pockets. It turned out that the results do not support this effect.

 

 

Conclusion:

 

    Of course, the gap between the WTP and the WTA exists stricto sensu. In the current and previous experiments, WTP is never strictly equal to WTA. The question is then the significance of this difference and its interpretation. Plott and Zeiler concluded that there has been a misinterpretation of the gap between both wills because it does not reflect a fundamental feature of human preferences[4]. Indeed, experimental designs are responsible, to them, of the gap’s existence. In that sense, authors break the link between the observed gap and the endowment effect, as part of the process of preference disclosure stated in the prospect theory.

 

    Five interpretations can be drawn to understand why the WTP/WTA exists in some cases, experiments and literature and why it disappears in other cases:

 

    1: When the measurements includes unexpected aspects of behaviors: every incentive that biases the valuation participants give, the gap exists. This interpretation is consistent with the fact that subjects do not necessary understand everything in the task they have to perform. In this precise case, the endowment effect has nothing to do with the gap.

 

 

    2: The design of experiments and procedures do not promote the understanding of the difference between owning a lottery (and aiming to sell it) and giving an entrance fee to play the lottery. In particular, as far as mugs are at stake as tokens, participants lose the feeling of ownership because they play on both side: the selling one and the buying one, before playing with lotteries. Therefore, procedures transform the preferences influenced by loss aversion into preferences not influenced by the latter.

 

 

 

    3: Subjects can have the feeling that experimenters want them to lose the feeling of ownership when they have to give a price that fits with their true value or preferences. Indeed, they may have some wrong expectations about a “correct” value, corresponding to a “thoughtful” or “altruistic” response. Obviously, that kind of misinterpretation is not the goal when performing the task.

 

 

    4: In the wake of the previous interpretation, subjects do not always manage to differ some suggested values given by the procedures, that actually do not exist, and the true and personal values there are ask to give, consistent with their own preferences.

 

 

 

    5: Eventually, the WTP/WTA gap may embody some features of a decision-making process and not a pure preference elicitation. At the beginning of the task, when participants are not familiar with the lotteries and the elicitation procedure, they can have a cognitive bias to reveal their true preferences (because of a framing effect due to the display of the first lottery choice with ambiguous outcomes, as opposed to a sure probability associated to the outcomes). But as the stages go by, they learn from previous stages and the value they provide might derives from their process of decision-making in accordance to what they did before, more than a reveal of their true preferences.

 

 

    Ultimately, authors provide a psychological fact that triggers an increase in the WTA: the ownership is closely related to the feeling of familiarity and reduce uncertainty. Therefore, the WTA rises. They argue that people’s value is sought and elicited during several phases of learning. By practicing on the lotteries, people tend to reach their true value by increasing their WTA.

 

 

 

Personal comment:

 

I was very interested in this subject for two main reasons:

 

    - First, my personal interest and research (master thesis) focuses on, among others, this gap between the WTP and the WTA.

 

    - On top of that, my personal motivations to understand people behaviors drove me toward this topic since I have been confronted to one particular idea pointed out by Plott and Zeiler. When I ran my last year experiment, I was present in the lab to ensure that the experiment was going well and as a matter of fact, many participants asked me more explanations about the task they were actually asked regarding a part of the experimental design, involving WTP and WTA questions. I understood that subjects could have a misconception of the problem since it was hard to explain them what WTP is, not refereeing to the WTA. For instance, if subjects could not understand “how much could you give to participate in this lottery?”, it was very hard to explain them alternatively than “how much would you sell this lottery if you owned it?”. Plott and Zeiler refers to “confusion” and “misconceptions” and argue that these concepts are not clearly and quantifiably defined.

 

At the end of their conclusions, authors acknowledge a part of weakness of their study, arguing that: even if they have demonstrated some misinterpretations that led to false conclusions, they do not propose any substitution solution. But it seems to me that Plott and Zeiler play on words. Because no one could control for a perfect and total understanding from participants. Of course, it depends on their current mental state when they participate in the experiment (are people tired or not? Are they inclined to understand well enough the instructions? Do they hear loud enough? And so on). As far as I am concerned, Plott and Zeiler gave a sufficient model to control for people understanding: the four principles and steps of their model.

 

Suggested extensions:

 

    In order to get rid of the fifth interpretation misconception, the task can be asked backwards, meaning that subjects will automatically understand what they prefer between a sure amount (probability 10/10) over an purely not-sure amount (probability 0/10). This manner of completing the task still holds a framing effect, but a framing that enables understanding, to me, much faster, since everyone is going to choose, once again the highest outcome associated to the 100% probability. I am currently applying this backwards elicitation task in my current experiment.

 

 

    As another extension path, it could be interesting to check if these results can apply both in the lab and on the field. According to Plott and Zeiler, this effect exists both in the lab (for economic experiences) and on the field (on the market), but John List (2004) prove that if novice dealers are subjects to this effect, experts or experienced dealers are not. Indeed, Plott and Zeiler acknowledge that asymmetries between the WTP and the WTA exist on the field, with theories not linked to the “reference point” exhibited by prospect theory. In his experiment, John List observed that by giving as a token collection objects for baseball fans, 45% of experienced dealers trade the good when only 20% of inexperienced dealers did not trade. In this sense, experts are not subject to the endowment effect. We could then imagine that when explaining the task and the token (through practice, tutorials and other) as Plott and Zeiler suggest to avoid misconceptions, non-experts could become experts and their preferences would not be biased by the endowment effect anymore. Precisely, what authors prescribed in their model look like a true training, such as how you become an expert, or an experienced dealer. After having replicated his experiment in the lab, List came to the same observations, concluding that market experience makes anomalies disappear and rationality arise, anomalies such as the gap between the WTP and the WTA. As far as inexperienced dealers are concerned, prospect theory seems to account well for their behaviors.

 

    As a matter of fact, Plott and Zeiler’s current paper came in response to John List works, arguing, once again, that the initial endowment effect is derived from the experimental procedures because the first token is perceived as a gift. But, to go further, it could be interesting to apply Plott and Zeiler’s model (4 requirements) to List experiment design, to see if the gap disappears, questioning the expectation of an endowment effect. 

 

 

    Ultimately, I would find enriching to study the effectiveness of a learning effect when performing the task. If such an effect exists, it could be an argument in favor of the bias when eliciting true, or wrongly-expected preferences. This learning effect could be evaluated by a reaction time measurement. Some reduction in the reaction time can support the existence of a learning effect.

 

 

 

 

 

[1] Strategic behaviors consist in giving a selling price higher than the “true” personal valuation of the token for example. This is a common psychological reflex when you do not understand clearly the task. You try to “protect” yourself from giving a “too easy” answer (here a too low/high amount for your WTP and WTA).

 

 

[2] The tutorial also displays an illustration where it is explained why giving other prices in the WTP/WTA elicitation, than the true personal valuation, is a dominated strategies in the BDM procedure.

 

 

[3] The income effect is often confounded with the “house money effect”. It consists in playing with the previous gains accumulated in the game/experiment.

 

 

[4] Thus, the gap exists only because the experimental procedure that was implemented created it, by essence. With enough controls (as in Plott and Zeiler experiment), the phenomenon is “turned off”. 

 

 

 

 

 

 

 

Bibliography:

 

 

Horowitz J.K. & McConnell, K.E. (2000). Working Paper. A Review of WTA/WTP Studies.

 

 

Isoni, A., Loomes, G., Sugden, R. (2011). The Willingness to Pay–Willingness to Accept Gap, the “Endowment Effect”, Subject Misconceptions, and Experimental Procedures for Eliciting Valuations: A Reassessment American Economic Review, 101(2): 991-1011.

 

 

Kahneman, D., Tversky, A. (1979). Prospect Theory: An Analysis of Decision Under Risk. Econometrica, 47(2), pp. 263-291.

 

 

Knetsch, J.L. & Sinden, J.A. (1984). Willingness to Pay and Compensation Demanded:Experimental evidence of an unexpected disparity in measures of value. Quarterly Journal of Economics.

 

 

List. J. (2004) The Nature and Extend of Discrimination in the Marketplace: Evidence from the Field. Quarterly Journal of Economics 119(1), pp. 49-89.

 

 

Plott, R., Zeiler, K. (2003). The Willingness to Pay/Willingness to Accept Gap, the “Endowment Effect,” Subject Misconceptions and Experimental Procedures for Eliciting Valuations. Social Science Working Paper 1132R.

 

 

Plott, R., Zeiler, K. (2011). The Willingness to Pay—Willingness to Accept Gap, the “Endowment Effect,” Subject Misconceptions, and Experimental Procedures for Eliciting Valuations: Reply. American Economic Review 101, pp. 1012–1028.