1. Traditional vs. Behavioral Law and Economics

1This paper contrasts scholarship in traditional law and economics with more recent scholarship in behavioral law and economics. A number of prominent advocates of applying behavioral economics to the law have made the claim that behavioral law and economics is simply a refinement of traditional law and economics. The key difference is that behavioral law and economics uses more realistic descriptions of human behavior as its foundation and thus provides a more accurate and useful understanding of the relationships between economics and the law.

2This paper takes issue with that claim. While behavioral law and economics may have important insights that legal scholars and policy makers may wish to embrace, behavioral economics is not simply a refinement of traditional law and economics. We first demonstrate through a series of examples that the strongest adherents of behavioral law and economics take positions that can be seen as subversive of the fundamentals of the core rationality principles underlying traditional law and economics. Second, and less appreciated, the assessment of welfare within behavioral economics differs sharply from the traditional economic paradigm used in law and economics.

3To understand these claims, we first need to provide some background on behavioral economics, traditional law and economics, and behavioral law and economics.

9This paper focuses on the differences between traditional and behavioral law and economics. The conventional view of this relationship—typically offered by proponents of behavioral law and economics—is that it is simply a refinement of traditional law and economics by incorporating more realistic behavioral assumptions. This is my interpretation of the motivation of Jolls, Sunstein, and Thaler (1998, 13) who strongly advocate for incorporating behavioral principles into legal analysis to increase our understanding of law: “Our goal in this chapter is to advance an approach to the economic analysis of law that is informed by a more accurate conception of choice, on that reflects a better understanding of human behavior and its wellspring.” In this view, behavioral law and economics refines the analysis in traditional law and economics by simply incorporating more realistic economic modelling assumptions into the law and economics paradigm. Richard Thaler (2006, 1597) expressed precisely this sentiment for behavioral economics: “If economics does develop along these lines the term ‘behavioral economics’ will eventually disappear from our lexicon. All economics will be as behavioral as the topic requires.”

12Second, the paper explores the differences in welfare analysis between traditional and behavioral law and economics. Drawing on recent scholarship, I emphasize the fragile nature of the conclusions about welfare that can be drawn from behavioral economics. Unlike traditional law and economics, behavioral models often feature unstable preference orderings which creates difficulties for systematic welfare analysis.

13To make these arguments, I present canonical treatments of traditional and behavioral law and economics. I argue that the alternative perspectives on rationality and welfare in behavioral law and economics often will undercut the typically strong conclusions drawn from traditional law and economics. I should emphasize that the focus of this paper is on the logical consequences of accepting a set of behavioral principles as true facts about the world, but I take no definitive position in this paper on the veracity of those principles.

2. What is Traditional Law and Economics?

15To illustrate the operation of traditional law and economics, I will start with an example from Louis Kaplow and Steven Shavell in Fairness and Welfar (2002). Their book analyses simple models of torts, contracts, legal procedures and law enforcement through an economic lens. While their book attracted notoriety because it claimed that principles of “fairness” were not consistent with welfare, our focus here is simply on the logic of their economic models, which we highlight because they make the role of economic analysis so clear.

17Now the analysis of the legal rule depends on the rationality and sophistication of two parties. The buyer of the table should not want to have a system in place which requires the producer to manufacture the table when the costs exceed $200. The reason is that the deal that would be struck between the parties would be better off ex ante for the buyer if production did not occur in this case. So the buyer should not want a system of specific performance in the incomplete contract case, even though it would guarantee that he would get the table regardless of the outcome. The example also requires that the buyer’s valuation cannot be highly fragile or contingent. For example, we must assume that the value of $200 that the buyer places on the table is not whimsical and does not suddenly rise to $300 if he learns that the seller decides not to produce the table. This example also assumes basic rationality on the part of the producer. When costs rise above $200, the producer breaches the contract and pays $200—the producer cannot feel guilty about disappointing the customer by not fulfilling an obligation.

18Moving away from the contract example, another classic set of law and economics cases involve externalities. Consider the classic Coase example where a rancher and a farmer face a situation where their actions adversely affect each other. A rancher’s cattle can cause damage to crops and a farmer’s fencing will impede profitable cattle production. With costless bargaining, the solution is to reach the outcome that jointly maximizes benefits. For example, if the benefits to the rancher from additional grazing exceed the damage to the farmer’s crops and the law requires the rancher to pay the farmer for damages, then grazing will occur with the rancher compensating the farmer. If the law does not require compensation to the farmer, then the rancher will simply graze on the farmland and farmer will pay the rancher to cut back only when the damages to the crops exceed the additional benefits from grazing. The legal rule will affect the distribution of gains but not the outcome in terms of the scope of grazing—this is Coase’s famous result in the costless bargaining environment.

19This example presumes that the legal rule does not frame the parties’ interactions by introducing new preferences or valuations. The fact that under one legal rule the farmer is entitled to compensation whereas under an alternative he is not, should not affect the rational calculations that the farmer and the rancher make in reaching their bargain. Again, there can be no contingent psychological changes in valuation based on the legal rule.

21We must, however, make behavioral assumptions in several layers of this process to ensure this equivalence. First, under both regimes, the parties are assumed to make decisions based on either the actual costs they incur by paying a tax or purchasing a permit or the opportunity costs dictated by the market from not selling a permit. Opportunity costs must be treated symmetrically to out of pocket costs. As we discuss below, behavioral economics denies this equivalence. Second, the actual market for tradeable permits is assumed to operate in an efficient manner in which prices accurately convey the relative cost information of the buyers and sellers and is fully reflected in the equilibrium price for the tradeable permit. Individuals respond to the market in a fundamentally similar way they would respond to a direct tax.

3. The Innovations from Behavioral Law and Economics

22There are many different strands in behavioral economics that are now in common usage. In order to focus on the contrast between traditional and behavioral law and economics, I will highlight just two of several themes in behavioral economics. These two themes are context-dependent preferences and the endowment effect, with its related phenomena.

24The endowment effect is based on the idea that individuals place extra value on an object once it is in their possession. In other words, they may value an object at $x if someone else has it, but once they have it in their possession, the object becomes worth more, say $y. The gaps between $y and $x can be extremely large, as extensive empirical work has demonstrated (Ariely, 2008, chapter 8). Once someone has possession of the object, they will require $y to part with it whereas they would only require $x to purchase it.

26Loss aversion is probably one of the most robust results in behavioral economics; indeed, scholars have suggested in many different contexts—including legal contexts—to design incentives around loss aversion (Jolls, 2002, 44). Fryer et al. (2012), for example, have suggested that incentive systems for teachers involve removing prepaid bonuses for teachers rather than rewarding them with varying bonuses after student performance is determined. They provide evidence from a field experiment that the loss-aversion frame leads to improved student performance.

28Let’s apply these two simple concepts to the methods and approaches underlying traditional law and economics. First, consider context-dependence preferences in the case of contract negotiation. As we noted in our discussion of a contract for producing a table valued by a buyer at $200, the conclusions we reached concerning why it is optimal to breach a contract in the high cost case required that the value placed by the buyer of the table did not depend on the realized cost of production. With that assumption, we were able to conclude in the incomplete contract case that only the expectations damage principle achieved efficiency.

29But suppose that preferences are context-dependent so a buyer might simply feel that a table is worth more to him or her personally if the costs of production are high. Don’t we all change our valuations if we think something is more costly to produce? Now, if both the buyer and seller understood these preferences ex ante, they would reach a very different outcome than predicted by traditional law and economics. The buyer would insist on specific performance which would reduce the surplus to be shared by both parties, relative to the efficient contract as dictated by the theory. On the other hand, if the buyer did not realize that his or her preferences would change, they would be dissatisfied with a payment of $200 under the expectations damage regime. How the individual would respond in this case is not clear, but one could envision court proceedings or subsequent litigation.

31Consider a few more examples that take into account loss aversion and the lack of parity between out of pocket costs and opportunity costs. First, a farmer in the Coase setting should consider the opportunity cost of not letting a rancher pay for damaging his crops. If that opportunity cost is not valued correctly, the two parties will reach an outcome that is not economically efficient. Second, consider the market for tradeable pollution permits. It is essential for an effectively functioning market that the holders of permits who can reduce pollution at a cost lower than the current market price for the permit be willing to sell their permits into the market. But again, the funds gained from selling a permit are opportunity costs. If an endowment effect mentality is operative, then a seller may not wish to depart with a tradeable permit, even if this would be the economically efficient outcome. In this case, the market itself would not work efficiently and there no longer would be a presumed equivalence between pollution taxes and a tradeable permit regime. This would be inconsistent with traditional law and economics.

32The environmental arena provides a number of other examples where the endowment effect challenges the predictions from traditional law and economics. Economists have long recognized that, as an empirical phenomenon, the willingness of individuals to pay for pollution abatement is much less than the willingness of individuals to accept additional pollution. For example, consider a lake that is 5 percent polluted and survey individuals to determine how much they would pay to restore the lake to its pristine character with zero pollution. That amount would typically be much less than they claim would be required to pay them to reduce the quality of a lake from its pristine character to one in which there is 5 percent pollution. While this finding may seem quite intuitive—after all, who wants to be responsible for polluting a lake?—it runs counter to the basic tradeoffs involved in comparing the marginal benefits and costs of pollution.

33According to basic economic principles, individuals may value pollution abatement, but should have stable preferences in trading off dollars against pollution. These should use these preferences to calculate the optimal level of pollution—whether it is 1 percent, 5 percent, or 10 percent economic theory is silent. But this calculation should not depend on whether the lake is originally polluted or not. This is just another example of the endowment effect in action—possessing a pristine lake will change one’s preferences.

34As this example intimates, preferences generated through the endowment effect may lead to inconsistent rankings of outcomes (Knetch, 1989). With inconsistent rankings, not only does it make it extremely difficult to ascertain the optimal level of pollution—is it 1, 5 or 10 percent?—but it also makes it especially difficult to analyze the consequences of alternative legal rules. In order to make definitive assessments, traditional law and economics assumes underlying economic rationality in evaluating the consequences of alternative legal rules. Without the underlying rationality, it no longer becomes possible to draw out the legal consequences of alternative legal regimes.

35These examples illustrate how both context-dependence preferences and the endowment effect undercut the most basic analyses of traditional law and economics. Of course, whether these behavioral phenomena are pervasive is contested intellectual terrain.

37Second, there is less agreement on how ubiquitous the endowment effect is beyond simple laboratory settings in which the endowment effect is well established. John List (2003), for example, has shown that sophisticated professional baseball card traders grow less susceptible to these biases over time. The presumption would be that in sophisticated commercial markets one would not expect to see the endowment effect operating. On the other hand, Fryer and his co-authors (2012) (which include List) also believe that loss aversion can create powerful incentives for behavior.

39Another prominent example where markets cannot guide individual choices is the phenomenon of internalities, or inefficient decisions made by individuals because of either defective decision making or some type of inefficient preferences. The classic example here is time-inconsistent preferences in which an individual at time t will plan to take action x at time t+1 but, in fact, they end up choosing action y at time t+1 instead. You may smoke or overeat now, but believe you will stop smoking and go on a diet in the future. When the future rolls around, however, you continue smoking and overeating. Markets will also not save you in this case—the fault lies directly with the individual (Gruber and Koszegi, 2008).

41Perhaps this argument against traditional law and economics is overdrawn. Here is one defense. It could be the case that behavioral principles become so well established and empirically grounded that they can be used to replace or complement traditional rationally assumptions. As the title of Dan Ariely’s best seller implies, perhaps humans are Predictably Irrational (2008). By this phrase we imply that non-rational behavior is not simply any wild alternative to classic rational principles—instead there is a well-defined set of biases that individuals possess and through careful study we can just use these in our modelling of behavior and legal rules. Loss aversion may be a good example of one. Moreover, we do not have to establish perfection in this quest—after all, practitioners of traditional law and economics have long recognized deviations from rationality in their analysis, they made no pretense that their models were universally accurate. However, they simply argue that economic rationality provides the most reliable approach to the analysis of the law. In a similar vein, one could simply argue that behavioral law and economics does not achieve perfection but simply deepens the analysis.

42Kaplow and Shavell recognized this possibility in their treatise Fairness versus Welfare (2002, 461-463), in a section entitled, “The Difficulty of Predicting the Behavior of Individuals Who Are Not Always Rational Maximizers of Their Own Well Being”. Their underlying position is quite balanced:

We believe that the standard assumption of rational maximization is apt for many legal applications—notably the study of the behavior of private enterprise—and has proven to be useful in a wide range of other settings. Nevertheless, when behavioral economics, cognitive psychology, evolutionary biology, sociology, or anthropology yields valid insights, they should be incorporated in legal policy analysis. (Kaplow and Shavell, 2002, 463, citations omitted.)

43They emphasize that although these alternative approaches may yield better predictions of behavior, improvements in positive economics, we still should rely on traditional principles of normative assessment based on maximization of stable individual preferences.

44Kaplow and Shavell present two examples where behavioral insights may be important in developing better positive economic models of behavior. Cognitive psychology may establish that physicians may overreact to highly publicized jury verdicts on medical malpractice. This finding could change the relative balance of alternative legal rules in determining the best approach to induce the proper level of precautionary behavior on the part of doctors. In another example, suppose that widespread non-compliance in the tax system spills over to individuals and causes them also to become more non-compliant. In this case, the optimal penalties for non-compliance and the level of resources devoted to detection of non-compliance by the tax authorities would change (Kaplow and Shavell, 2002, 463). Despite the nod that two prominent practitioners of traditional law and economics give to behavioral approaches, there are two potential problems with this strategy of accommodation and assimilation of behavioral factors into law and economics. The first is the lack of quantitative precision of behavioral economics as it stands today. The second, which is perhaps even more fundamental and we defer to next section, is the difficulties that behavioral economics poses for the assessment of welfare.

46While there are experimental studies of the endowment effect, most are laboratory studies with the recognized problems of external validity in generalizing to phenomenon outside the narrow confines of the laboratory. Dan Ariely conducted a fascinating field experiment with Duke basketball tickets by comparing the willingness to pay and accept for losers and winners of a lottery among passionate basketball fans. His finding that winners of the lottery would only sell their tickets for $2400 whereas losers would only buy tickets for $170 was striking but well out of the range of typical estimates of the endowment effect. Typical estimates of willingness to accept to willingness to pay are usually less than two to one, not over the twelve to one that Ariely found (Kahneman et al., 1990). Thus, while most studies do find evidence for an endowment effect, its quantitative measure is uncertain. In that case, how can behavioral law and economics analyze the alternative consequence of legal rules, except by making them very context-dependent?

48While the same critique could be leveled at traditional economics, there is typically more theoretical guidance from basic price theory as to where to measure and look for the effects of policies. Perhaps someday behavioral economics could reach that point, but the open ended experimentation of the governmental entities and their frank admissions that their findings are context-dependent, suggest otherwise.

4. The Welfare Conundrum in Behavioral Economics

51All of this is based on some deep notion of revealed preferences. From what individuals choose, we can infer what individuals prefer. That is the essence of the revalued preference approach—actual choices reflect true and stable preferences.

52Behavioral economics effectively denies that preferences are necessarily stable. Context-dependent preferences are rooted in the idea that other factors in the environment—such as arbitrary frames—can shift preferences. One example of an arbitrary frame would be to cast outcomes as losses from some status quo point as opposed to gains as viewed from another vantage point. This is the genesis of the endowment effect.

54At this abstract level, it is clear that behavioral factors pose problems for welfare orderings. Suppose an individual can rank a set of outcomes i=1 to N, but the ranking depends upon the specific frame. If the frame is F, there is one ranking, whereas if the frame is F’ there is another ranking. Under frame F, outcome 1 is preferred to 2, whereas under frame F’, outcome 2 is preferred to outcome 1. What can we say about the rankings in this case?

55Douglas Bernheim and Antonio Rangel (2009) have developed an extensive theoretical treatment of these key issues. They first suggest that one approach would be to simply rule out a priori that some frames are valid. For example, if frame F’ had not even a remote relationship to the relevant choice set—say it was whether the outcome of a roulette wheel in Macao was red or black—this would be ruled out as a valid frame. Note that even in this case, the social judgment of states we would make would differ from the individual judgment made by the person with frame F’ but in this case we are resigned to that because we are willing to rule out F’ as a relevant frame.

56The more common case, however, is when we are not able to rule out a particular frame. Bernheim and Rangel then suggest that we have to rely only on orderings that dominate others under all relevant frames. In our previous example, we could not say whether state 1 was more desirable than state 2 because their ordering depended upon the frame. But suppose under either frame, both state 1 and 2 were preferred to state 3; then we could say that 3 was inferior to either 1 or 2. Notice that this approach may still allow for some orderings, but there will always be a set of outcomes that cannot be ranked because their relative preference depends on the specific frame.

58Jonathan Gruber and Botond Koszegi (2008) used this model to analyze the welfare effects of taxes on cigarettes. In their welfare analysis, they assumed that β<1 represented a deficiency of decision making capacity, and analyzed the effects of taxes on welfare evaluated when β was set equal to 1.

59Using this assumption, they concluded that high cigarette taxes could actually make low-income smokers better off, contrary to the normal conclusions about the welfare effects of cigarette taxes. The reason for this result was that smokers were making their decisions inefficiently (they were subject to an internality and should not have been smoking or smoking as much) and higher taxes were particularly effective in dissuading low income individuals to quit. Since quitting improved their welfare, even though they did not have the willpower to quit without the taxes, it made them better off.

61Now you may side with the foresighted self in the smoking case. But if you side with that self in the smoking case, who would you side with in the case of a woman in labor and suffering excruciating pain who asks for pain relief which, prior to labor, she told her doctor in absolute terms not to administer to her (Schelling, 1984)?

62There is not a simple and general answer to this conundrum. For smoking, our intuitions today suggest that we should favor the foresighted self, particularly because so many people wish to cease smoking. However, attitudes towards smoking have changed radically in a generation and in 1950 the vast majority of individuals would have sided with the smoker. And in the case of pain relief under labor, almost everyone would side the woman in labor, even if they know she would regret her decision later.

63To reach a policy conclusion on the optimal level of cigarette taxes, behavioral law and economics must take a position on this issue. But the decision about what aspects of preferences to respect does not have a strong scientific basis. In some cases, we may want to take a paternalistic stand and in others we may not. We may want individuals to save more for retirement and use a series of clever nudges to push them in that direction. But in other cases, we may not have the requisite social agreement. Our degree of moral consensus may be high for some social phenomena but not for others. But even when we do have a moral consensus, it is just that—a moral and not scientific consensus, and one not rooted in any systematic treatment of individual welfare that we can simply write down or use to distinguish one case for another. The seeming precision of welfare economics is replaced by the thoughtful but equivocating field of applied ethics.

5. The Limits to Behavioral Law and Economics

65This would be a perfectly satisfactory state of affairs if behavioral law and economics provided us with robust alternative paradigms. But for the reasons we discussed, it does not achieve this level of generality. There is “predictable irrationality” but it is not as predictive or reliable in terms of predicting behavior as we require. Instead, we must rely on field studies to test our theories of nudges and examine their efficacy and power in individual settings. But it is unlikely that a nudge in one area works precisely the same as a nudge in another area.

67Behavioral law and economics provides a valuable but piecemeal set of tools and insights to assist us in social policy. It may help us to think through social policies towards obesity, smoking, alcoholism, compliance with the law, and individual savings behavior. Many areas of law already reflect different standards for transactions directly affecting individual and those that are primarily business-to-business. For example, truth-in-lending laws apply to individual borrowers. The contributors to Sunstein’s edited volume (2000) and the Oxford Handbook (Zamir and Teichman, 2014) also explored, with a behavioral perspective, traditional legal topics such as punitive damages, contract formation, jury awards for pain and suffering, and the regulation of risk. These are important areas of law and policy and intellectual and policy contributions in this area will be welcome. But the contributions are not likely to be all the same—they will be dependent on context, social attitudes and norms, and other general cultural phenomenon. As long as we recognize these limitations, behavioral law and economics can inform our legal and social deliberations and provide new insights for social policy.

Thanks are due to the editors of this special issue and to the anonymous referees.