The Endowment Effect's Disappearing Act

The "endowment effect," a tendency for an individual to state a minimum amount for which he might be willing to sell an item that is greater than the amount the same individual is willing to pay for the same item as a buyer, is quite popular in the legal literature and elsewhere these days (it even has a Wikipedia entry, as does the rather lengthy "list of cognitive biases").  An unrestricted Westlaw search (JLR) turns up 541 articles mentioning the endowment effect.  189 of these articles appeared over the past three years, and 46 in 2005 alone.  I will be the first to admit that I have not read all of these articles.  However, I have read enough to get a sense that the template goes something like this: (1)the "endowment effect" exists; (2) Coase was wrong; (3) there is a market failure; and (4) insert proposal for regulatory intervention.

A recent American Economic Review article by Charles Plott and Kathryn Zeiler, " The WTP-WTA Gap, the 'Endowment Effect,' Subject Misconceptions, and Experimental Procedures," is a must read for scholars contemplating entry into this literature.  Some highlights: 

The issue explored here is not whether a WTP-WTA gap can be observed.  Clearly, the  experiments of KKT and others show not only that gaps can be observed, but also that they are replicable.  Instead, our interest lies in the interpretation of observed gaps.  The primary conclusion derived from the data reported here is that observed WTP-WTA gaps do not reflect a fundamental feature of human preferences.  That is, endowment effect theory does not seem to explain observed gaps.

And:

"by implementing different procedures, the phenomenon can be turned on and off.  When procedures used in studies that report the gap are employed, the gap is readily observed.  When a full set of controls is implemented, the gap is not observed.  The fact that the gap can be turned on and off demonstrates that interpreting gaps as support for endowment effect theory is problematic."

Problematic? I'd say.  That experimenters can eliminate or create the effect by modifying experimental procedures is at least problematic.  So why is the effect so easily replicated if it doesn't really exist?  Plott and Zeiler's answer: "[t]he thesis of this paper is that observed gaps are symptomatic of subjects' misconceptions about the nature of the experimental task."  That is, experimental procedures and not biases in individual preferences explain the gap.  Anyway, read the whole thing. 

A final thought.  To my mind, proponents of endowment effect-based regulation must confront a burden that looks something like this characterization offered by Steve Bainbridge a few years ago here:

The regulator ought to be confident that (a) the endowment effect is really present; (b) as a result of the endowment effect, an inefficient outcome is proving sticky; (c) that market experience cannot remove the anomaly; and (d) there is a superior regulatory solution.

Prior to reading Plott & Zeiler, I was fairly confident that the number of "endowment effect" regulatory proposals meeting this standard was small, and possibly zero.  Pending convincing contradictory evidence that the effect actually exists in markets, the set of qualifying proposals now appears to be null.  My only question is: What would the law reviews do with all that extra space?

Comments

[Kate Litvak]

Arlen, Spitzer, and Talley have a lovely experimental paper showing that the alleged endowment effect disappears in the agency context.
See here: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=276110 (I get no commission for their downloads, I swear).

I was always suspicious about the existence of the endowment effect. As I recall, most European cultures have an equivalent to the old English saying “the grass is always greener on the other side,” meaning “whatever the other guy has is inevitably better than what I’ve got, no matter what we are talking about.” Sounds exactly the opposite to the endowment effect – and widespread enough to become a major part of folk cultures across the continent.

[Paul Gowder]

I'm skeptical.

Check out footnote 10, page 12-13 of the Plott-Zeiler piece. That language suggests that the experimenters very carefully trained a whole host of errors and misconceptions out of their subjects. (not considering opportunity cost, gambler's fallacy, no notion whatsoever of expected value, etc.)

The question then arises: do ordinary people participating in market behavior have those misconceptions trained out of them? And if not, is this study important at all? Or is it the case that the endowment effect *for all intents and purposes* functions in a marketplace full of non-specialists who aren't being gleefully trained by a team of economists on every transaction?

Also, this training might modify preferences, i.e. people might subjectively experience a lower harm to an opportunity cost -- which is another way of stating the endowment effect -- until the economist kindly trains that out of them by repeatedly explaining to them that there's an opportunity that they're losing.

[Joshua Wright]

Paul, thanks for your comments, but I'm not sure if your characterization of the study is accurate.

What is different about Plott & Zeiler is not that they provided explanations of the benefits of particular strategies to subjects or had subjects engage in practice rounds. Many experimenters (including those finding gaps) do this (see Table 1 in the AER version of the paper, I'm not sure if it is in the SSRN link). And for good reason. One would imagine that experimenters want subjects to have a good understanding of the procedures being used, available alternatives, and mapping of choices to consequences in order to ensure that the results reflect individual preferences and not simply subject misconceptions about the experimental procedures.

You also seem to suggest that if ordinary people do not have misconceptions trained out of them in actual markets (whether "gleefully by economists" or otherwise), then the study is not valuable. While I agree that evidence from actual markets is what matters most, this seems like an odd conclusion to draw. The question that comes to my mind is: "do ordinary people act in markets in a way that is systematically related to how they perform in experimental laboratories at their undergraduate universities?"

[Paul Gowder]

Joshua -- you're right of course, but my only point (perhaps a bit overstated) was that the answer to your last question moves closer to "no" as the experimenters carefully explain and train people to avoid the very sorts of errors that they purport to test.

The line between "a good understanding of the procedures being used, available alternatives, and mapping of choices to consequences" and "train people to avoid the very sorts of errors that they purport to test" seems very thin indeed when the experiment is supposed to be measuring inability to accurately map choices to consequences.

The opportunity cost thing in particular strikes me as odd, that they'd just casually note that people would disregard opportunity costs until they were carefully trained out of it. Well, isn't that closely related to the endowment effect? (Both are an irrational preference for the status quo, etc.)

It seems like Plott and Zeiler were testing a strawman version of the endowment effect theory, viz. "the endowment effect exists as an immutable psychological/neurobiological reality no matter how much work we do to help people avoid it." But their study does nothing to cast any doubt on the weaker version of the endowment effect and most other cognitive biases, which freely acknowledges that trained actors aren't as susceptible. (For example, how many statisticians are going to fall for the various probability errors that have been shown time and time again to be true?)

Also, as I understand it, part of the underlying theory behind the whole notion of cognitive biases is that they exist as heuristics to minimize decision cost for low-importance choices. If this is the case, as the perceived importance of a choice increases, people are more likely to take care to avoid their normal mental processes. This too is fully consistent with "weak endowment effect."

[Michael Guttentag]

Joshua. As in so many things, I think we have to consider here both Scylla and Carbides. I agree with you about the Scylla – the wild and unjustified extrapolation from generic experimental findings to policy recommendations that one frequently finds in legal scholarship.

But, aren’t there potentially some policy implications to be drawn from what economic experimentalist have been finding in the lab over the past twenty-five years. I agree with Paul’s comments above that Charles Plott and Kathryn Zeiler are far too aggressively conservative in interpreting their own results. That the endowment effect can be turned on and off is not surprising at all – Plott and Zeiler are simply starting the process of deconstructing (can I use that word on this blog?) the factors that may lead to an endowment effect.

As for your question, "do ordinary people act in markets in a way that is systematically related to how they perform in experimental laboratories at their undergraduate universities?” I think the answer is probably yes. A nice example of the way that hypotheses about heuristics can be used to generate testable hypotheses showed up in last month’s Journal of Financial Economics: Prospect theory, mental accounting, and momentum, Mark Grinblatt and Bing Han. Volume 78, Issue 2 , November 2005, Pages 311-339. To disregard what’s being learned by experimental economists is to fall too far to the side of Carbides.

The issue, as your cite of Bainbridge alludes to, is this: can we relate experimental findings to regulatory intervention? If we are clever, and think it is sufficiently important, the number of such proposals might exceed zero.

[Joshua Wright]

Michael: I agree with you entirely. Well, almost entirely. I certainly agree that the optimal number of such proposals may well be positive (namely, all those that satisfy the burden in the post). I also agree experimental evidence might be valuable. Perhaps I should have been more clear on both of those points.

Here is where I disagree: you write that if "we are clever, and this it sufficiently important, the number of such proposals might exceed zero." I guess my concern is that we might be a bit too clever. I read Plott & Zeiler as doing a bit more than initiating the "deconstruction" of the gap. Rather, it seems to me that their finding is sufficient to shift the burden of proof re: existence of the gap at all.

I'm not saying this is an insurmountable gap, just that I am not going to read one of these papers unless it purports to satisfy the burden above. Then again, I guess I am probably infra-marginal ...

[Robert Schwartz]

"What would the law reviews do with all that extra space?"

Run articles about law that ordinary lawyers can read and understand. I read the dialogue above, and I couldn't make heads or tails out of it. But, I am a lawyer not a social scientist.