But seriously, folks, what do people want?
INTRODUCTION I. ARLEN AND TONTRUP: THE ENDOWMENT EFFECT AND THE ROLE OF REGRET A. Experimental Design 1. Agents as responsibility diffusers 2. Responsibility shifting through conformity with a group B. Interpreting the Results II. HOFFMAN AND WILKINSON-RYAN: TAKING PRECAUTIONS BEFORE AND AFTER ENTERING INTO A CONTRACT A. Experimental Design B. Interpreting the Results III. So WHAT DO PEOPLE WANT AND WHAT DIFFERENCE DOES IT MAKE? A. Staying Within the RC Model B. Leaving RCT Behind
In a lecture he gave at Stanford many years ago, Richard Rorty recounted the following story: A friend, an eminent decision theorist, had to decide between competing job offers from two universities. Unable to choose, he called up Rorty for advice. "Why don't you make one of those fancy decision trees you're always writing about?" suggested Rorty. His friend's response: "Oh, come on, Dick, this is serious."
The story came to mind in reading through the papers in behavioral economics presented at the Seventh Annual Conference on Empirical Legal Studies (CELS). One of the most exciting bodies of work to come out of the social sciences over the last fifty years, the heuristics and biases (H&B) wing of behavioral economics has identified robust patterns in human decisionmaking that undermine many of the core assumptions of rational choice theory (RCT): that we have stable preferences, that we act "rationally" to optimize those preferences, and that utility depends on end states (e.g., total wealth), rather than gains and losses off of a reference point (e.g., of prior wealth).
But RCT and much of the H&B research on consumer behavior share one presupposition, which is arguably more important than all of the ones they disagree about: that the ultimate carriers of utility in consumer transactions are commodity bundles, that consumers' "true" preference is to optimize on those bundles, and therefore that prospect theory and other violations of expected utility theory documented by H&B scholars "must lead to normatively unacceptable consequences." (1) The question I want to raise here is: what if that supposition is wrong? Suppose, for example, that in the typical purchase decision, the consumer's "true" preference is not to maximize some function of wealth (absolute, relative, or changes in wealth measured against a reference point), but instead to minimize the time and mental energy spent on trades, because once she has reached an acceptable level of material consumption at a fairly low level of demandingness, she gets little if any additional utility from optimizing, compared to the other things she could be doing with that time. What difference would that make in how we interpret experimental findings in behavioral economics and psychology?
For RCT, it makes no analytical difference. "True preference," like maximand, has no operational meaning. Optimizing on a choice always means optimizing on the outcome minus the costs of getting there. Once costs are accounted for, the optimal decisionmaking strategy is, in Herbert Simon's words, more appropriately described as "satisficing," meaning settling for a good enough (satisfactory) outcome. (2) Thus, the difference between a consumer who seeks to "maximize" on a toaster, subject to keeping search costs tolerable, and one who seeks to "minimize" the time and psychic energy spent on search, subject to the need to satisfice on a toaster, is just a verbal difference. Mathematically, they are identical decision problems; as long as we hold constant the measure of costs and benefits, we end up at the same place. When academics refer to consumers' "true" preferences, I take it to mean the preference to which consumers give the greatest weight in a given decision; that is the sense in which I use the term here.
Furthermore, RCT is officially indifferent as to the content of consumer preferences and which preference (wealth; consumption; fairness; ideology; religious convictions; or time available to think, garden, or surf the web for porn) dominates any particular consumer choice. In Sam Peltzman's words, all of these possibilities are just "'taste' variables about which economists have nothing much useful to say." (3)
At the same time, RC theorists perceive themselves to have a large professional stake in what those preferences actually turn out to be, and they may well be right. As Peltzman acknowledged thirty years ago, if noneconomic preferences (ideology, etc.) are the dominant influence on consumer (or legislator) behavior, "pessimism about the future of economic versus sociological analysis of [such behavior] would be warranted." (4) Hence, Peltzman's self-described ambition was to "see how much the role of categories like 'political ideology,' which are unfamiliar to economists, can be reduced by simple and fairly crude manipulation of economic categories." (5) In its own way, this is an astonishing admission and gives every reason to regard with skepticism the standard RCT account of human motivation. Yes, division of labor and simplifying assumptions can be enormously valuable research tools, but there's a limit. If your doctor tells you that the pain in your abdomen must be caused by a problem in your ear, because ears are what he knows about, it's time to find a new doctor.
Behavioral economists, in contrast, have no professional stake in the content of our "true" preferences. Their project is to describe consumer behavior accurately and identify the psychological motivations that drive it. They will not be out of a job, whatever those motivations turn out to be. In addition, individual behavioral economists and psychologists have pressed the argument that maximizing on commodity bundles does not necessarily maximize utility, because consumers care about nonmaterial objectives as well, (6) or because the endowment effect and other framing effects, if they persist and affect the experienced utility of a choice, are not cognitive biases but simply part of the experienced utility of that choice; (7) or because treating "sacred" nonmarket goods (e.g., intimate relationships, friendships, work relationships, family, babies, body parts, etc.) as commensurable with money destroys much of their value. (8)
Finally, many of the same people who brought us behavioral economics are a major force in the hedonics literature, which has insisted that any form of utilitarianism worth its name must look beyond RCT's formal axioms of rationality (invariance, dominance, transitivity, etc.) to substantive criteria of rationality, in Daniel Kahneman's words, we must ask "whether choices maximize the (expected) utility of their consequences, as these consequences will actually be experienced." (9) As discussed below, the large and growing body of hedonics research suggests that, once basic needs are met, increased income or consumption levels have at best an uncertain relationship to experienced utility, and that more choice can be bad--not just because we lack the cognitive capacities to process it correctly (which RCT can account for as part of search costs) but because it can lead to greater unhappiness, both during the choice process and in living with the consequences of our choices. (10)
Notwithstanding the foregoing, much of the experimental literature on consumer choice coming out of behavioral economics still defaults to the assumption that the "true" preference of individuals in evaluating potential trades is to maximize consumption, and that that preference accurately predicts experienced utility. In a sense, the discussion that follows boils down to a suggestion that the H&B wing of behavioral economics take more seriously the emerging lessons of its brethren in the hedonics wing. (11)
I use two of the papers on contracts presented at the CELS conference to illustrate the implications of relaxing that assumption: Jennifer Arlen and Stephan Tontrup's The Endowment Effect: Voluntary Debiasing Through Agents and Markets, (12) and David Hoffman and Tess Wilkinson-Ryan's The Psychology of Contract Precautions. (13) In interpreting their findings, both papers presuppose that consumers' "true" preference in evaluating trades is to maximize wealth/consumption. In both cases, they may well be right, at least under experimental conditions, and their results are interesting and suggestive in their own right. In the spirit of a friendly amendment, I suggest other interpretations of experimental findings that are opened up if we take seriously the possibility that consumers' "true" preferences lie elsewhere.
I. ARLEN AND TONTRUP: THE ENDOWMENT EFFECT AND THE ROLE OF REGRET
Arlen and Tontrup's experiment is designed to test the psychological motivations behind the widely observed disparity between the price we are willing to pay and the price we are willing to accept for the same object. The standard explanation in the H&B literature for this observed disparity is the "endowment effect": we form an emotional attachment to whatever we possess (i.e., whatever we are "endowed with"), which leads us to value it more than we would if we did not possess it. Assuming that the emotional attachment is transitory, H&B scholars have generally concluded that the endowment effect is irrational, because "people's valuation of an entitlement can depend on their relationship to the object in ways that do not affect its actual utility to the person." (14) How can the same mug be worth four dollars to you if you don't own it and eight dollars if you do? (15)
The experiment focuses on one of the alternative explanations for the willing-to-pay/willing-to-accept disparity advanced in the H&B literature: our reluctance to trade is motivated by our fear of future regret should the choice turn out to be the "wrong" one. (16) (The explanation in this case relies on the plausible assumptions that we code a trade as an act and a failure to trade as an omission, and are more prone to regret actions that turn out badly relative to "doing nothing at all," than we are to regret failures to act that turn out badly, relative to the opportunities we forewent.) Consistent with that explanation, Arlen and Tontrup hypothesized that they could "mute the behavioral factors that cause endowed people to resist trading" by interposing mechanisms that deflect onto others the responsibility for a potentially bad outcome. (17)
A. Experimental Design
Each subject was assigned heads or tails for a prospective coin flip and given the opportunity to swap sides before the coin was flipped. If the coin came up on the subject's chosen side, the subject would receive 8 euros (roughly 10 dollars). In addition, if but only if the subject opted to trade her initially assigned side for the other, she would receive an additional 25 euro-cents (roughly 30 cents). Assuming that the subjects' objective is to maximize expected wealth, the strictly dominant choice is to trade, provided that (as the subjects were instructed) it is a fair coin flip. Either way, they have a 50% chance of getting 8 euros. But if they trade, they will be guaranteed an extra 25 euro-cents. "Thus, a rational subject, who is not biased by the endowment effect, should exchange his ticket for the alternative (equally valuable) ticket in order to obtain the guaranteed monetary bonus. A subject who retains the ticket is exhibiting the endowment effect." (18)
Arlen and Tontrup tested several variations of this basic experiment. In the Base treatment, where subjects made the decision whether to trade for themselves, only 30% of the subjects (19 out of 64) opted to switch--a result that the authors attribute to the endowment effect. (19) The authors then introduce two intermediating mechanisms that they hypothesized would diffuse the subjects' sense of responsibility for the choice. The first was to interpose an agent who would make an initial recommendation whether the subject should switch, with the subject retaining veto power. (20) The second was to expose subjects to information about the choices made by a majority of others facing the identical decision. (21)
1. Agents as responsibility diffusers
In the agency treatments, the authors told subjects that an agent would (or could) make the initial decision for them whether to trade or not, but that they retained an absolute veto over the recommendation of the agent. Subjects were told that the agent would be paid more if he recommended a trade, information correctly interpreted by the subjects to mean that agents would pretty much always recommend a trade. This piece of information was intended to remove any tendency for the subjects to think that the agent's recommendation added any new information about the optimal choice. If the subject didn't veto the agent's recommendation to trade, the trade went through. If the subject vetoed the agent's recommendation to trade, then the subject kept the originally assigned side.
In the first version of the experiment (Mandatory), subjects were instructed that they had to employ an agent to make the initial recommendation, although they retained the right to veto the recommendation. In the second version (Optional), subjects were allowed to choose whether to employ an agent at all. Thus, in the Optional version, the subjects had to make two decisions rather than just one: (i) whether to employ an agent to make the initial recommendation; and (ii) whether to veto the agent's recommendation once made.
In the Mandatory condition, 69% (31 out of 45) of the subjects opted to trade. In the Optional condition, about half (29 out of 59) opted to delegate the initial decision to the agent. Of those who delegated, 90% (26 out of 29) opted to trade. (22)
2. Responsibility shifting through conformity with a group
In what the authors call the social conformity treatment, subjects were informed that a majority of subjects in a prior, identical experiment had opted to trade. Current subjects were told that the prior group had only the information that the current subjects had been given--a statement meant to blunt any tendency to defer to the past group on the assumption that they had better information to assess the value of trading. This treatment was designed to test whether conformity with others' decisions would be enough to mute the subject's fear of future regret, even without shifting the actual affirmative decisionmaking to a third party. Once again, a much higher percentage of the treatment group opted to trade than in the Base condition--55.1% (27 out of 49) versus 29.7% (19 out of 64). (23)
B. Interpreting the Results
The authors interpret the results to support the regret avoidance hypothesis. In particular:
(a) The experiments were carefully designed to rule out (or at least make very unlikely) any rational explanation for the observed behavioral changes in moving from the Base condition to the treatment conditions. Subjects were instructed ahead of time that the agent and the prior group had been given exactly the same information as the subject had, and (implicitly) that the agents were certain to recommend trading because they had a self-interested motive to do so. As a result, the only rational explanation for subjects' increased willingness to trade when an agent or a prior group's decision was introduced--that they believed the agent or prior groups knew more than they did about the optimal choice--is ruled out. (24)
(b) The endowment effect cannot account for the increased willingness to trade, since the distribution of endowments did not change from the Base condition to the treatment conditions. The regret avoidance hypothesis, on the other hand, seems to fit the results rather well. If nothing changes from the Base condition to the treatment conditions except the introduction of an intermediary (past or present) to opine on the "right" choice, the mere presence of the intermediary seems to be what makes the difference, presumably because it provides subjects with someone else to deflect the responsibility onto, and with that, reduces their ex ante anxiety about ex post regret. (25)
(c) The regret avoidance hypothesis also seems to account for the fact that roughly 50% of subjects in the Optional condition chose to employ an agent. Since the subjects knew ahead of time what the agent would recommend and that the agent wasn't privy to any information that they didn't already have, they could not rationally have expected to make a better decision by employing an agent. And in any event, they would still have to decide whether to veto the decision or not, without having garnered additional information from the recommendation. Why not just make the decision themselves?
Thus, the authors conclude, "[I]n this design, the decision to delegate is properly viewed as a decision to shift responsibility because principals do not obtain any other advantage from delegating to the agent." (26)
This strikes me as a plausible and interesting interpretation of the data. But, like the conventional interpretation that the willing-to-pay/willing-to-accept disparity is the result of an endowment effect, it presupposes that the subjects' "true" preference (that is, primary motivation) with respect to the proposed trade is to maximize wealth. Suppose we were to take seriously the possibility that consumers' "true" preference in evaluating most potential trades is something else--say, to minimize the time and mental energy we spend on decisionmaking in order to maximize the time available for work, play, thought, family, or friends. This alternative account of consumers' motivations fits many of the authors' findings pretty well, and some arguably better than the 'regret avoidance' explanation.
First, deflecting responsibility for choice onto others may be a completely rational rule of thumb (conscious or unconscious) for happiness if your "true" preference going through life is to minimize time and effort spent on consumption decisions. It will misfire in some cases (possibly including the Optional condition here, since it increases the subjects' necessary decisions to two).
Second, it could help explain the (at least to me) surprising results in the Base condition, before intermediaries are introduced. Recall that 70% of the subjects refused to trade in the Base condition. (27) The authors attribute that result to the endowment effect, (28) although they understandably don't dwell on it as they are interested in the difference between the Base and treatment conditions. But given the lengths to which the authors went to make the optimal choice transparent (choosing a simple and familiar bet which most people would process as a 50/50 shot, making the superiority of trading obvious by monetizing its advantage and making the advantage a certainty, and finally instructing subjects on its superiority), I find it surprising that the endowment effect would be powerful enough to lead 70% of the subjects in the Base condition to refuse to trade.
If, however, our "true" preference when offered a trade is not to maximize wealth but to minimize the time spent thinking about the trade, presumptively sticking with the status quo, whatever it is, is a perfectly rational strategy. It amounts to saying, "I won't waste time thinking about whether to sell or whether to buy unless the disparity in values is great enough to make the trade not a close case." That motivation will, rationally, produce a disparity in the price a consumer is willing to pay and the price the consumer is willing to accept, because it puts the "buy" and "sell" decisions at different ends of a band of indifference (that is, the range of values in which trading is a close case). It isn't worth my while to consider an offer to buy my mug until the price gets high enough (say ten dollars), at which point it is no longer a close case; and it also isn't worth my while to consider an offer to sell me the identical mug until the price gets low enough (say five dollars), at which point it is no longer a close case.
II. HOFFMAN AND WILKINSON-RYAN: TAKING PRECAUTIONS BEFORE AND AFTER ENTERING INTO A CONTRACT
When parties cross over from negotiations to a binding contract, they typically decrease precautionary behavior sharply (for example, a party may stop comparative shopping even if they have the option to get out of the contract, or stop monitoring the other side or taking precautions against possible opportunistic behavior). Hoffman and Wilkinson-Ryan's experiment is designed to test the motivations behind that observed behavioral kink.
Over the years, scholars have offered a number of explanations for the decrease in precautionary behavior post-contract.
Economic rationality. Beginning with Stewart Macaulay's famous 1963 study entitled Non-Contractual Relations in Business, (29) contracts scholars coming from a wide range of perspectives have argued that trusting your contract partner is often rational, given the mutual benefits of cooperation and the fact that monitoring the other side can be both difficult and, if visible to them, counterproductive. (30)
Early scholarship in the relational contracting vein focused mostly on the advantages of informality and trust once in a contractual relationship. More recent scholarship has expanded the focus of inquiry to include precautionary behavior before a contract is finalized and after contractual relations have ruptured, concluding that trust and precautionary behavior are both useful responses over the life of a contract, and that the moment of finalizing a contract may well be a rational point at which to shift from precaution-taking to trust. (31)
The moral bindingness of promises. The Kantian-inflected literature on the sanctity of promising argues that the moment at which we commit to do X we assume a moral obligation to do X, either as a vindication of our own autonomy, or out of respect for the autonomy of our promisee, or both. (32) Within that literature, however, there is serious disagreement about what acts constitute a commitment to do X.
In the realm of moral (as opposed to legal) duties, many reject the promissory basis of commitment (we commit ourselves to do X only when we expressly or impliedly promise to do X) in favor of a detrimental reliance-based view of commitment (we commit ourselves to do X whenever we lead others reasonably to expect we will do X and reasonably to rely on it). (33) The same disagreement about whether "promissory" duties are rooted in promise or in reliance is replicated in the legal domain, in the longstanding arguments over whether promise or reliance is the moral root of contractual obligation, and whether an express or implied promise to do X is necessary or sufficient to require that X (or some substitute) be done. However, pace Grant Gilmore, (34) even those who believe it is not necessary still regard promising as the root of contractual obligation, and detrimental reliance as an exception in cases in which a formal exchange of promises is premature (e.g., precontract negotiations) or unlikely to occur (e.g., family relationships). The fact that all promises are conditional, and that some (including the one at issue here) explicitly give one or both parties the legal option to terminate the contract with no penalty, strains the moral justification for fetishizing the contract/no-contract line. Be that as it may, once the association is normalized, it would naturally produce a behavioral kink at the moment of signing a contract: "Sure, it was fine to shop the field as long as I didn't sign on the dotted line, but now that I did, I'm morally committed to this contract."
Humean rationality. David Hume's conventionalist account of the bindingness of promises in effect harnesses moral reasons in service of economic rationality. In Hume's account, the primary problem with bilateral commitments is that whichever party performs first opens itself up to opportunistic behavior by the other side. ("Thanks so much for helping me harvest my corn, but now that you've completed your side of the promise, I see no personal advantage to completing mine by helping you harvest yours.") Thinking about solutions to the problem in an extralegal context, Hume concludes that the only dependable constraint on defection is reputation. (35) But in order to punish transgressors, we have to agree on what constitutes a transgression--that is, when it is that we cease to have the right to behave opportunistically vis-a-vis the other side. Hume seizes on the moment of promising as an obvious Schelling-like focal point, but presumably he would have been perfectly happy to substitute another if it served the purpose. Foreseeable pre-contract detrimental reliance is one obvious alternative.
Economic irrationality. H&B scholars, in contrast, have focused on the hypothesis that when contracting parties decrease precautions at the moment of finalizing the contract, they are responding to cognitive biases. Two explanations in particular have been the object of extensive study in the H&B literature.
The first is cognitive dissonance. As Hoffman and Wilkinson-Ryan put it: "We trust people we are in contracts with more than those we aren't, believing we chose our contracts counterparties precisely because we found them trustworthy." (36) Rather than entertaining the possibility that we were wrong about that, we persuade ourselves that the contract we have entered into is the best of all choices in the best of all possible worlds and call a halt to shopping around, further negotiations over the terms of the contract, monitoring our contracting partners closely, or insuring against their default. (37)
The second grows out of prospect theory and, in particular, loss aversion and reference point theory. Loss aversion posits that we attach more weight to losses than we do to equal gains. Since losses and gains are measured in relation to some prior state, in order to know whether an outcome is coded as a gain or loss, we have to know what prior state people are using as a reference point. (38) H&B scholars have posited that the contract/no-contract line serves as a compelling reference point for evaluating outcomes. (39) Precautions (shopping around, buying insurance, etc.) that are taken before the contract is finalized are coded as part of a transaction that produces an overall gain. Equivalent precautions taken after the contract is finalized are coded as a separate loss and therefore assigned a much greater disutility. In short, "the moment of contracting resets the status quo and primes a cognitive script that favors trust in contract." (40) As with cognitive dissonance, the result is an observable kink in precautionary behavior before and after the moment of contracting.
In practice, it can be very difficult to tease apart these various explanations for the observed kink. Hoffman and Wilkinson-Ryan's experiment was designed to eliminate economic rationality by stripping the contract/no-contract line of all economic significance. Any remaining drop-off in precautionary behavior at the point of contract, they reasoned, "cannot be explained by straightforward analysis of economic costs and benefits, even when we take into account transactions costs," and hence looks presumptively irrational. (41)
A. Experimental Design
The authors ran three different experiments. I focus on the last of the three, because it provides the strongest support for the authors' conclusions.
Subjects (incoming law students) were assigned randomly to a "No Contract" or a "Contract" group. Both groups were instructed that they were in the market to lease a new Chevy Blazer and that they had found one for $300 per month at Tim's Auto World. Both groups were told that they had just signed a lease agreement with Tim's Auto World and had taken the car home. Subjects in the Contract group were told that the lease agreement went into legal effect immediately, but that they could cancel it, with no penalty, within three days of signing. Subjects in the No Contract group were told that the lease agreement would not go into effect for three days, and they could opt out any time before then without penalty. (42)
As the authors state, "This description was intended to make the notion of contract as [economically] meaningless as possible. Thus, in both cases, doing nothing leads to contract, and returning the car within three days of purchase cancels the contract and has no other legal repercussions." (43) The only distinction between them was the legal formality that the Contract group was bound to a contract immediately and the No Contract group was not. The authors underscored the presence of that formal distinction and the absence of any material consequences to it by ending instructions to the Contract group with the statement, "Remember: You are under a contract, but you can walk away without consequence," and ending instructions to the No Contract group with the statement, "Remember: You are not under contract." (44)
Both groups were told that on the second day of their three-day option, they learned that they could lease the same car more cheaply from another dealer. They were then asked if, having learned that, they would return the car to Tim's Auto World and go after the lower-priced one. Those in the Contract condition were less likely to return it and required greater savings ($32.31 versus $15.55) to do so. (45) In other words, they were less willing to protect their pecuniary interests after the contract with Tim's Auto World was signed than before, although "the formalization of the contract is essentially arbitrary and without legal consequence." (46)
B. Interpreting the Results
The authors conclude that the contract/no-contract line "produces a behavioral kink, divorced from concerns about cost, or projection of future legal remedies." (47) Although the experiment was designed to rule out any rational explanations for that kink, the authors acknowledge that even if fetishizing the contract/no-contract line is economically irrational in this particular case, it could still be an economically rational rule of thumb to follow. (48) Nonetheless, the authors incline more towards the view that cognitive biases like prospect theory and cognitive dissonance explain the observed behavior, and that such biases "may discourage optimal precaution-taking even when economic analysis suggests that they ought to protect themselves." (49)
The authors' interpretation of the results seems plausible, but, like the economically rational "rule of thumb" interpretation, it presupposes that consumers' "true" preference when presented with this choice is to optimize on the car lease. Once again, if what you really want is to spend as little time as possible on the choice, subject to satisficing at a fairly undemanding level, adopting arbitrary lines at which you cut off search may be a perfectly rational strategy.
One obvious response is: "Okay, maybe satisficing on consumer choices leads to greater life happiness than trying to maximize on them, (50) but surely it can't be rational to peg satisficing to the empty legal formality of passing from the No Contract condition to the Contract condition." Yes and no. The contract/no-contract line may be arbitrary with respect to the optimal tradeoff between consumption and freedom from shopping in a given case, and in some cases it may miss by a lot. (Note, however, that a priori we don't know which direction it will miss in. By the time the subject shows up at Tim's Auto World in real life, she may well already have spent far more time on search than is optimal for her happiness, and, going forward, the rational cutoff point is whatever point comes next.) On the other hand, it may be a quite serviceable utilitarian rule of thumb for pulling the plug, along with a host of other seemingly arbitrary cutoff rules ("I'll look in two stores, no more," or "I'll spend an hour on this, and then I'm out of here"). In light of experimental results suggesting that giving people the opportunity to revise their decision leaves them less satisfied with the decision, it may well be less arbitrary than many others. (51)
III. So WHAT DO PEOPLE WANT AND WHAT DIFFERENCE DOES IT MAKE?
A. Staying Within the RC Model
In both the Arlen and Tontrup and Hoffman and Wilkinson-Ryan studies, it is impossible to answer the question of what people want based only on observed behavior, since the subjects' responses are logically consistent with all of the motivations explored above, and no doubt many more. In the end, one is thrown back on one's own intuitions. For what it's worth, here are mine.
The "I have better things to do with my time" explanation for the willingness-to-pay/willingness-to-accept disparity in the Arlen and Tontrup experiment and for the abrupt end to precautionary behavior in the Hoffman and Wilkinson-Ryan experiment faces an obvious problem. In both cases, the subjects volunteered to participate, in one case in return for money (albeit a nominal amount). (52) Whatever the subjects' "true" preferences would be in the real world, one might reasonably expect that--having agreed to participate in the study--they would take the assigned task seriously for the few minutes required to complete it without reopening the question of the opportunity costs of having agreed to do so. For the same reason, the findings in other studies that agents acting on behalf of others do not exhibit the endowment effect (53) and that agents do not exhibit an endowment effect when making exchanges with their principal (54) are not surprising. In both cases, the agent's job is to make the trade in question, and it is reasonable to expect that fact to concentrate her mind wonderfully on making the "rational" (meaning wealth-maximizing) choice for herself or her principal.
On the other hand, opportunity costs could be a factor even in Arlen and Tontrup's experimental design. The booths in which subjects were placed were "very hot and humid," subjects were permitted to leave whenever they concluded the questionnaire, and 25 euro-cents may well have been a small price to pay to get out of the experiment more quickly. (55) Physical exit is not the only competing opportunity here; there is also mental exit. After giving the questionnaire one quick pass-through, subjects might well decide that even if they are going to be trapped in that hot and humid space for another X minutes, they would rather spend those X minutes thinking about other things. In follow-up studies, one could try to control for both possibilities by requiring subjects to stay put for a set period, whether they finished the questionnaire or not, and giving them similarly cognitively taxing tasks if they finish early.
Once one turns to consumer behavior in the wild, however, the "I have better things to do with my time than optimize on wealth/consumption" explanation for any observed disparity in willingness to accept and willingness to pay or the behavioral kink at the contract/no-contract line seems much more plausible. In a sense, that explanation has to be a factor almost all of the time. In these two experiments, as in experimental studies of consumer behavior more generally, subjects have one choice to make: whether to trade heads for tails or tails for heads, or whether to pursue the cheaper auto lease option or not. By contrast, in real life we are presented with infinite opportunities at every moment to improve our commodity bundles through trade, and the notion of rationality built into RCT assumes, in Kahneman's words, that "no significant opportunity will remain unexploited." (56) If spending your life assessing those opportunities against the status quo doesn't strike you as a recipe for happiness, then the rational rule of thumb probably is to just say no to almost all of them without giving them a moment's consideration, and to spend as little time as possible on most of the rest while still satisficing on consumption in an undemanding sense.
Of course, in the real world, each of us cares a lot about some consumer choices, and, with respect to those, rationally chooses to invest considerable time and energy in trying to optimize them. I do not, however, share Hoffman and Wilkinson-Ryan's intuition that as a general matter, "in real world contracts with more at stake, individuals may tend to behave in a more economically maximizing manner" than in experimental conditions. (57) While maximizing on a real car lease is more important to most people than maximizing on a hypothetical one, the opportunity costs of doing so are also much, much higher. At the very least, we should recognize that when we study consumer decisionmaking under experimental conditions, in which subjects have nothing to do with their time but focus on the decision at hand, we automatically render consumption the most plausible maximand from the subjects' perspective. The only way to avoid that is to move out of the laboratory and into the field, to assess consumer behavior in its real-world context.
If the default assumptions of RCT and H&B have gotten wrong the relative values consumers attach to maximizing consumption and minimizing time spent on consumption decisions, that fact has important implications, both in assessing what behavior is irrational and (if one is so inclined) in designing policy interventions to mute or eliminate it. To take the Hoffman and Wilkinson-Ryan study, if the authors are correct that some combination of reference point theory and cognitive dissonance explains consumers' mentally closing the transaction once the contract is signed, (58) then sensible policy interventions would look to debiasing consumers from that automatic response or manipulating the choice architecture to mute it. One could, for example, require that all consumer contracts that are revocable within a specified time period at no cost to the consumer be labeled "Option to Purchase" rather than "Contract." If, on the other hand, consumers' mentally closing the transaction at the contract line is a rational rule of thumb for limiting time and mental energy spent on consumption decisions, a very different set of policy interventions might be in order, including curtailing options to back out so as to force consumers to move on with their lives, (59) reducing choices, or encouraging consumers to end their searches even earlier or (in many cases) not even get started.
Similarly, in the Arlen and Tontrup study, if subjects' refusal to make obviously superior trades results from the endowment effect or excessive regret avoidance, then (as the authors suggest) sensible policy interventions would look to debiasing consumers' irrational attachment to what they happen to have or encouraging them to make greater use of responsibility-dispersing mechanisms. (60) If, on the other hand, it reflects a preference not even to think about possible trades unless they present a clear opportunity for significant gain, we might just as well leave the rule of thumb alone. It will misfire in individual cases (although a priori it is not clear whether it will screen out too many or too few potential trades), but may still be a pretty good recipe for happiness over the long run.
B. Leaving RCT Behind
Up until now, the discussion has stayed within the conventional RCT framework, substituting one presumed maximand (time and mental energy) for another (the value of the commodity bundle). But suppose people's "true" preference most of the time with respect to trades is not to think about them at all, not (as argued above) because of the time and energy required to evaluate them, or because of people's cognitive limitations in evaluating them, or because commodifying the value of "sacred goods" destroys that value, but because people do not want to go through life relating to their environment as a potential source of gains from trade. They want to live a different kind of life, one immersed in the lived experience, oblivious to the fact that it could be characterized as an opportunity cost with respect to the infinite other possible lives one could be living. The metapreference for such a life would pose a much more serious challenge to conventional RCT, because it leaves the tools of RCT behind completely. Yes, one could expand the lists of "tastes" consumers potentially bring to the bargaining table to include a metapreference not even to come to the table, but it is hard to see what we might hope to learn by doing so.
How plausible is this alternative construction of people's "true" preferences? One need not fall back on intuition alone here. There is a large and growing literature in hedonistic psychology relying on self-reports, observations, and experimental data to support the conclusion that once individuals have enough income to meet basic needs, increased wealth does relatively little, if anything, to enhance their a sense of well-being throughout their lives, compared to a host of other nonmaterial "goods" (e.g., intimate relationships, friends, social networks, social skills, mental absorption in work or hobbies, volunteer work, and inherited temperament and personality). (61) Let us suppose for the moment that those findings are correct. For a welfarist, it would suggest a profound shift in public policy in a host of substantive areas, including (just for starters) transportation policy, housing design, the provision of local social services, regulation of advertising, educational methods, and tax policy. For those inclined to intervene at the point of preference formation, it might also suggest trying to debias individuals from the mistaken belief that increasing consumption will increase their future happiness. Finally, as many have urged, it suggests that we need to reevaluate the social indicators we use to measure societal well-being. The traditional indicators, including GDP and income, make sense if material consumption is strongly and positively correlated with happiness. If it is not, the answer to the question, "how happy are we?" lies elsewhere.
The stakes are even higher when it comes to solving the collective problems we face as a nation and as part of the world community. As long as the solutions to long-term crises (global warming, the collapse of our educational system and physical infrastructure, etc.) are cast as tradeoffs between the well-being of the current generation who will have to pay for the solution and future generations who will reap the benefits of it, any meaningful solution faces an uphill battle politically. But at least in the developed world, if the current generation could be convinced that the financial sacrifice required of them is unlikely to affect their well-being very much, if at all, the battle starts looking a little more winnable.
These grand observations may seem far afield from the immediate question at hand: what motivates consumer choice, and are those motivations and resulting choices rational, in a means-end sense, when measured against the ex post experienced utility of those choices? But I hope I have at least made the case that the two are connected, and that understanding consumer behavior may require us to step outside of it, and think about what makes people happy more generally in their lives.
(1.) Daniel Kahneman & Amos Tversky, Prospect Theory: An Analysis of Decision Under Risk, in CHOICES, VALUES, AND FRAMES 17, 31 (Daniel Kahneman & Amos Tversky eds., 2000). Prospect theory models decisions between alternatives with uncertain outcomes. Under these conditions, the theory states that individuals make decisions based on the expected gains and losses from the status quo, rather than the expected value of the final outcome (as expected utility theory assumes), giving rise to inconsistencies in expected utility theory that present "significant normative challenges" for economists. Jennifer H. Arlen & Eric L. Talley, Introduction to EXPERIMENTAL LAW AND ECONOMICS, at xxv, xxv-xxvi (Jennifer H. Arlen & Eric L. Talley eds., 2008). See generally Kahneman & Tversky, supra. For an early criticism of behavioral economists' willingness to sign on to expected utility theory without inquiring whether it accurately predicts experienced utility, see Graham Loomes & Robert Sugden, Regret Theory: An Alternative Theory of Rational Choice Under Uncertainty, 92 ECON. J. 805 (1982).
(2.) Herbert A. Simon, Rational Choice and the Structure of the Environment, 63 PSYCHOL. REV. 129, 129, 136 (1956).
(3.) Sam Peltzman, Constituent Interest and Congressional Voting, 27 J.L. & ECON. 181, 183 n.5 (1984). For a less pessimistic view about the relevance of economics in studying rational behavior given noneconomistic preferences, see Botond Koszegi & Matthew Rabin, Choices, Situations, and Happiness, 92 J. PUB. ECON. 1821, 1830-32 (2008).
(4.) Peltzman, supra note 3, at 192.
(5.) Id. at 192-93.
(6.) For just a small sample of the literature taking seriously a broadened menu of objectives, see Elizabeth Hoffman & Matthew L. Spitzer, Entitlements, Rights, and Fairness: An Experimental Examination of Subjects' Concepts" of Distributive Justice, 14 J. LEGAL STUD. 259, 259-60 (1985); Daniel Kahneman et al., Fairness as a Constraint on Profit Seeking: Entitlements in the Market, 76 AM. ECON. REV. 728, 728-29 (1986); Russell Korobkin, The Endowment Effect and Legal Analysis, 97 NW. U. L. REV. 1227, 1227-29 (2003) [hereinafter Korobkin, Endowment Effect]; Russell Korobkin, The Status Quo Bias and Contract Default Rules, 83 CORNELL L. REV. 608, 611-12, 617-24 (1997) [hereinafter Korobkin, Status Quo]; Koszegi & Rabin, supra note 3, at 1821-27; Cass R. Sunstein, Legal Interference with Private Preferences, 53 U. CHI. L. REV. 1129, 1150-52, 1168-69 (1986).
(7.) Daniel J. Keys & Barry Schwartz, "Leaky" Rationality: How Research on Behavioral Decision Making Challenges Normative Standards of Rationality, 2 PERSP. ON PSYCHOL. SCI. 162 (2007); Graham Loomes & Robert Sugden, Disappointment and Dynamic Consistency in Choice Under Uncertainty, 53 REV. ECON. STUD. 271, 279-81 (1986); Loomes & Sugden, supra note 1, at 818-19.
(8.) Barry Schwartz, Some Disutilities of Utility, 23 J. THOUGHT 132 (1988); Philip E. Tetlock, Coping with Trade-Offs: Psychological Constraints and Political Implications, in ELEMENTS OF REASON: COGNITION, CHOICE, AND THE BOUNDS OF RATIONALITY 239, 242-45, 249-54 (Arthur Lupia et al. eds., 2000).
(9.) Daniel Kahneman, New Challenges to the Rationality Assumption, in CHOICES, VALUES, AND FRAMES, Supra note 1, at 758, 760.
(10.) See infra Part III.B.
(11.) Others have voiced similar concerns that behavioral economics' right hand doesn't always know (or take seriously enough) what its left hand is doing. See Keys & Schwartz, supra note 7, at 162-67 (arguing that Kahneman and other prominent behavioral economists, in their hedonics research, have defended experienced utility as the appropriate standard for assessing the rationality of decisionmaking, but that the same theorists have assumed in prior decision theory work that framing effects, etc., had to be irrational because they conflicted with the formal principles of rationality).
(12.) Jennifer Arlen & Stephan Tontrup, The Endowment Effect: Voluntary Debiasing Through Agents and Markets (CELS Version, Nov. 2012). A revised version of this paper is available at http://ssrn.com/abstract=2263447.
(13.) David A. Hoffman & Tess Wilkinson-Ryan, The Psychology of Contract Precautions (CELS Version, Nov. 2012) [hereinafter Hoffman & Wilkinson-Ryan, CELS Version]. Hoffman and Wilkinson-Ryan's CELS paper has since been published. David A. Hoffman & Tess Wilkinson-Ryan, The Psychology of Contract Precautions, 80 U. CHI. L. REV. 395 (2013).
(14.) Arlen & Talley, supra note 1, at xxv.
(15.) For articulations of the standard position, see Jack L. Knetsch & J.A. Sinden, Willingness to Pay and Compensation Demanded. Experimental Evidence of an Unexpected Disparity in Measures of Value, 99 Q.J. ECON. 507 (1984); Korobkin, Endowment Effect, supra note 6, at 1250-55; Korobkin, Status Quo, supra note 6, at 625-30; Gary H. McClelland & William D. Schulze, The Disparity Between Willingness-to-Pay Versus Willingness-to-Accept as a Framing Effect, in FRONTIERS OF MATHEMATICAL PSYCHOLOGY: ESSAYS IN HONOR OF CLYDE COOMBS 166, 166 (Donald R. Brown & J.E. Keith Smith eds., 1991); and Richard H. Thaler, Toward a Positive Theory of Consumer Choice, in CHOICES, VALUES, AND FRAMES, supra note 1, at 269, 273-76. For more recent findings that cast doubt on the existence of an endowment effect, see Arlen & Tontrup, supra note 12, at 4 & n. 12.
(16.) For an excellent overview of literature on regret avoidance, see Korobkin, Status Quo, supra note 6, at 657-60.
(17.) Arlen & Tontrup, supra note 12, at 3.
(18.) Id. at 12 (citations omitted).
(19.) Id. at 21.
(20.) Id. at 13-15.
(21.) Id. at 15-16. For a complete description of the experimental design, see id. at 9-17 (base and agent); and id. at 26 (information exposure).
(22.) Id. at 21.
(23.) Id. at 27.
(24.) Id. at 10-11.
(25.) Id. at 27-28.
(26.) Id. at 15; see also id. at 27-28.
(27.) See supra note 19 and accompanying text.
(28.) See Arlen & Tontrup, supra note 12, at 30.
(29.) Stewart Macaulay, Non-Contractual Relations in Business. A Preliminary Study, 28 AM. SOC. REV. 55 (1963).
(30.) Lisa Bernstein, Merchant Law in a Merchant Court: Rethinking the Code's Search for Immanent Business Norms, 144 U. PA. L. REV. 1765, 1787-95 (1996); David Charny, Nonlegal Sanctions in Commercial Relationships, 104 HARV. L. REV. 373 (1990); Ronald J. Gilson et al., Braiding: The Interaction of Formal and Informal Contracting in Theory, Practice, and Doctrine, 110 COLUM. L. REV. 1377 (2010); Hoffman & Wilkinson-Ryan, CELS Version, supra note 13, at 9-10, 24.
(31.) Bernstein, supra note 30, at 1791-94; Gilson et al., supra note 30, at 1379-87.
(32.) In the legal academy, Charles Fried and Seana Shiffrin are the most prominent exponents of this view. See CHARLES FRIED, CONTRACT AS PROMISE: A THEORY OF CONTRACTUAL OBLIGATION (1981); Seana Valentine Shiffrin, The Divergence of Contract and Promise, 120 HARV. L. REV. 708 (2007). Within philosophy itself, there is an enormous literature on the moral obligations we assume by promising to do X. Notable works include J. Raz, Promises and Obligations, in LAW, MORALITY, AND SOCIETY: ESSAYS IN HONOUR OF H.L.A. HART 210 (P.M.S. Hacker & J. Raz eds., 1977); and Thomas Scanlon, Promises and Practices, 19 PHIL. & PUB. AFF. 199 (1990).
(33.) PATRICK ATIYAH, THE RISE AND FALL OF FREEDOM OF CONTRACT (1979); Raz, supra note 32, at 215-19; Scanlon, supra note 32, at 200-01.
(34.) See GRANT GILMORE, THE DEATH OF CONTRACT (1974).
(35.) DAVID HUME, A TREATISE OF HUMAN NATURE, bk. III, pt. II, sec. V ("Of the Obligation of Promises").
(36.) Hoffman & Wilkinson-Ryan, CELS Version, supra note 13, at 18 (emphasis omitted).
(37.) For some of the legal literature on the subject, see id. at 25 n.89.
(38.) The literature here is enormous, much of it tracing back to Kahneman & Tversky, supra note 1.
(39.) See Ernst Fehr et al., Contracts as Reference Points--Experimental Evidence, 101 AM. ECON. REV. 493 (2011); Oliver Hart & John Moore, Contracts as Reference Points, 123 Q.J. ECON. 1 (2008); Yuval Feldman et al., Reference Points and Contractual Choices: An Experimental Examination (Jan. 21, 2012) (unpublished manuscript), available at http://ssrn.com/abstract=1989556.
(40.) Hoffman & Wilkinson-Ryan, CELS Version, supra note 13, at 42.
(41.) Id. at 5.
(42.) Id. at 18-19.
(43.) Id. at 18.
(44.) Id. at 19. Having recently exercised a "free" option to return a mattress within sixty days of purchase and then having to endure the salesman's protracted haranguing and guilt-tripping for my decision to "cancel" the contract, I am less confident than the authors that the line is economically meaningless. At least where negotiations are face to face or over the telephone with a live, psychologically astute and motivated seller, exercising one's right to terminate in the Contract condition is likely to be more unpleasant than exercising it in the No Contract condition, thereby raising the real costs of terminating. After all, that's why the mattress store puts us in the Contract condition to begin with.
(45.) Id. at 20.
(47.) Id. at 28.
(48.) See id. at 24-25.
(49.) Id. at 25-26, 29.
(50.) In support of that hypothesis, see BARRY SCHWARTZ, THE PARADOX OF CHOICE: WHY MORE IS LESS (2004) (arguing that a multiplicity of options decreases happiness); Sheena S. Iyengar, Rachael E. Wells & Barry Schwartz, Doing Better but Feeling Worse: Looking for the "Best" Job Undermines Satisfaction, 17 PSYCHOL. SCI. 143, 147 (2006) (finding that job seekers who maximized their number of options were less happy with the search, despite achieving higher pay); Barry Schwartz et al., Maximizing Versus Satisficing: Happiness Is a Matter of Choice, 83 J. PERSONALITY & SOC. PSYCHOL. 1178, 1193-94 (2002) (finding that "maximizers" achieved objectively better outcomes, but subjectively experienced their conditions with less satisfaction than "satisficers"); and Erin A. Sparks et al., Failing to Commit: Maximizers Avoid Commitment in a Way That Contributes to Reduced Satisfaction, 52 PERSONALITY & INDIVIDUAL DIFFERENCES 72, 74-75 (2012) (finding that maximizers were less committed to their choices and arguing that the lingering uncertainty about the choice reduced their happiness).
(51.) See Lottie Bullens et al., Keeping One's Options Open." The Detrimental Consequences of Decision Reversibility, 47 J. EXPERIMENTAL SOC. PSYCHOL. 800, 803-04 (2011) (finding that decision reversibility can lead to feelings of regret because rejected options continue to be weighed in working memory).
(52.) See Arlen & Tontrup, supra note 12, at 6-7 (volunteers with modest compensation based on performance); Hoffman & Wilkinson-Ryan, CELS Version, supra note 13, at 18 (volunteers without compensation).
(53.) Knetsch & Sinden, supra note 15, at 518; James D. Marshall et al., Agents' Evaluations and the Disparity in Measures of Economic Loss, 7 J. ECON. BEHAV. & ORG. 115, 120 (1986); see also Russell Korobkin & Chris Guthrie, Psychology, Economies, and Settlement: A New Look at the Role of the Lawyer, 76 TEX. L. REV. 77, 99-101 (1997) (finding that lawyers acting on behalf of a client in a hypothetical settlement exercise did not display the endowment effect, while the clients did).
(54.) Jennifer Arlen et al., Endowment Effects Within Corporate Agency Relationships, 31 J. LEGAL STUD. 1, 18-19 (2002).
(55.) See Arlen & Tontrup, supra note 12, at 7, 15 n.40, 22 (describing the experimental conditions). In her oral commentary on the Arlen and Tontrup paper at CELS, Alison Morantz raised a different concern about subjects' motivations: that the 25 euro-cent reward might well have been a disincentive to reach the right answer, because it crowded out intrinsic motivations for doing so. The thought here is that the subjects, who might otherwise have faithfully executed the task at hand out of a sense of duty or curiosity, regarded it as demeaning to be thought to do so in order to get a mere 25 euro-cents. Alison Morantz, Professor of Law, Stanford Law Sch., Commentary at the Seventh Annual Conference on Empirical Legal Studies (Nov. 10, 2012) (discussing Arlen & Tontrup, supra note 12).
(56.) Kahneman, supra note 9, at 758. For a general discussion of these issues, see Mark Kelman, Consumption Theory. Production Theory, and Ideology in the Coase Theorem, 52 S. CAL. L. REV. 669 (1979).
(57.) Hoffman & Wilkinson-Ryan, CELS Version, supra note 13, at 39.
(58.) Id. at 24-26.
(59.) Of course, this could have the opposite effect from the one intended: it could induce consumers to spend more time on search before entering into the contract, knowing that once the contract is signed, it is binding. Which effect would predominate in any given case and with what consequences for consumer happiness are empirical questions, the answers to which I suspect will depend upon the nature of the choice involved. I thank the editors of the Stanford Law Review for pushing me on this point.
(60.) See Arlen & Tontrup, supra note 12, at 31-32.
(61.) The literature here is voluminous. Many of the leading articles are collected in INTERNATIONAL DIFFERENCES IN WELL-BEING (Ed Diener, John F. Helliwell & Daniel Kahneman eds., 2010); and WEEL-BEING: THE FOUNDATIONS OF HEDONIC PSYCHOLOGY (Daniel Kahneman et al. eds., 1999). See also MIHALY CSIKSZENTMIHALYI FLOW: THE PSYCHOLOGY OF OPTIMAL EXPERIENCE 36, 39-40 (1990); SCHWARTZ, supra note 50, at 106; Ed Diener & Robert Biswas-Diener, Will Money Increase Subjective Well-Being? A Literature Review and Guide to Needed Research, 57 Soc. INDICATORS RES. 119 (2002) (concluding that economic growth in developed countries over the past few decades has been accompanied by little rise in subjective well-being and that people who prize material goods more than other values tend to be less happy unless they are rich).
Barbara H. Fried, William W. and Gertrude H. Saunders Professor of Law, Stanford Law School. I thank Jennifer Arlen, Joe Bankman, Paul Brest, Dick Craswell, Dan Ho, and the editors of the Stanford Law Review for their helpful comments.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||whether presupposition that consumers' "true" preference is wealth/consumption maximization, or whether their preference is something else; Seventh Annual Conference on Empirical Legal Studies: The Empirical Revolution in Law|
|Author:||Fried, Barbara H.|
|Publication:||Stanford Law Review|
|Date:||Jun 1, 2013|
|Previous Article:||The Twiqbal puzzle and empirical study of civil procedure.|
|Next Article:||Answering questions, questioning answers, and the roles of empiricism in the law of democracy.|