Consumer preferences, citizen preferences, and the provision of public goods.
A standard justification for state regulation and state provision of certain services is the market failure known as public goods. Public goods are characterized by two features: jointness of supply and nonexcludability. Jointness of supply concerns the cost function of the good and the degree to which its consumption is nonrivalrous. The marginal production costs of public goods are very low or even zero; adding more users would not detract from the benefits enjoyed by others. Nonexcludability means that it is impossible or impractical to prevent those who do not pay for the good from enjoying its benefits. These features make the supply of public goods through the market mechanism infeasible or suboptimal, thus calling for state intervention. Consequently, the government plays a major role in providing public goods such as national defense, roads, national parks, health and education systems and in guarding against air pollution, extinction of endangered species, and the demolition of historic buildings.
Assuming the private market cannot react appropriately to the true demand for public goods,(1) the regulator must decide which public goods to supply and in what quantities. A policymaker wishing to base these decisions on people's preferences faces not only the problem of conflicts among the preferences of different people but arguably also the difficulty of conflicts between the revealed preferences of the same persons. As against the model of human beings who are rational actors, having a single, complete and transitive set of preferences, it has been claimed that in fact people have dual or even multiple preference orderings, reflecting the different roles they play in different situations.(2) In particular, some scholars argue that people hold and express different preferences in their "consumer" role and in their "citizen" role.(3) While the former role, evoked in market settings, reflects people's self-regarding wants and interests, the latter, aroused in political settings, reflects their opinions, values, and beliefs regarding the good of society as a whole. In its most extreme manifestation, the consumer/citizen distinction brings to mind the famous story of Dr. Jekyll and Mr. Hyde. Citizen Jekyll is a relatively altruistic person who attributes great value to--and is willing to make sacrifices for--the common good, whereas Consumer Hyde is an egoistic maximizer of personal wants and desires.
This dismal description of daily life has drawn criticism, and alternative explanations have been suggested for the apparent inconsistency in people's behavior. These explanations attempt to vindicate consumer behavior and often portray citizen preferences as misinformed, capricious, or insincere.(4)
The conceivable conflict between consumer and citizen preferences requires a crucial decision from policymakers: Should regulators base their decisions on preferences that reflect consumer wants, citizen opinions, or some compromise between the two? The possible divergence between consumer and citizen preferences demands a preliminary resolution with respect to not only the type of "informational input" that should be used by the regulator but also the type of process by which this information should be gathered and handled. A consumer approach will use market techniques, such as cost-benefit analysis and questionnaires concerning individuals' willingness to pay for public goods (e.g., contingent valuation studies). The regulator will attempt to identify actual preferences, aggregate them, and maximize their satisfaction. In contrast, a citizen approach will focus on preferences that are revealed in nonmarket contexts, such as votes in elections, political debates, and public opinion surveys. Decisions regarding the appropriate quantity of public goods will not be based on an aggregation of identified citizen preferences but rather on a more flexible process of deliberation and reasoning.(5)
This Essay critically analyzes various explanations for the discrepancy between consumer and citizen revealed preferences, as well as their normative implications. It claims that neither the explanations based on the supremacy of citizen preferences and political processes nor those resting on the authenticity and realism of consumer preferences can persuasively and exclusively account for the discrepancy. Rejecting the contention that consumers and citizens hold radically distinct preference orderings, the Essay offers a different explanation for the greater manifestation of other-regarding views in the public arena. The suggested explanation is twofold. First, drawing on insights from game theory, I argue that differences in behavior can be understood within the framework of an Assurance Game. The game-theoretic examination explains why in market settings people may fail to achieve their most favored collective goals, even if their preferences are markedly other-regarding rather than egoistic. In particular, the perceived sense of "hopelessness" may play a crucial role in the behavior of consumers. This "hopelessness" is absent in political settings, which therefore may authentically reflect the other-regarding preferences that exist--but are not expressed--in the private arena as well. The second part of the explanation focuses on why preferences concerning public goods tend to be other-regarding. I argue that the nature of the good, rather than the nature of the process by which it is produced, determines the preference orderings and, in particular, the tendency for socially-oriented orderings. People's preferences for public goods are more other-regarding than their preferences for private goods.
This argument has important normative implications. It supports the claim that citizen preferences should be given substantial weight in the formulation of public policy, particularly with respect to public goods. Moreover, the proposed explanation offers the soundest, least controversial grounds for interventionist regulation, more so than any other explanation for the differences between consumer and citizen behavior. Government intervention assists individuals in realizing their highest-ranking preferences, those they hold but cannot effectuate in the marketplace. At the same time, because the espoused reliance on citizen preferences rests on neither the inferiority of consumer preferences nor their characterization as self-centered, a sweeping rejection of economic tools, such as cost-benefit analysis, is unwarranted.
The Essay opens in Section II.A with an account of the consumer/citizen distinction and the explanation for it suggested by its proponents. Section II.B then argues that the common replies to this explanation do not dispense with the issues raised by the consumer/citizen distinction. Section II.C critically examines additional explanations offered for the differences between consumer and citizen behavior. The first explanation focuses on the Prisoner's Dilemma. Other explanations emphasize various derogatory aspects of citizen behavior, highlighting such factors as misconception, insincerity, and caprice. The main Part of the Essay proposes an alternative explanation. First, Section III.A introduces the motivation of "hopelessness" and applies the Assurance Game model to the market and political arenas. Section III.B then examines how the nature of the good determines the type of preferences that people hold. Part IV then presents several normative implications of this favored explanation, and Part V briefly summarizes my conclusions.
II. THE CONSUMER/CITIZEN DISTINCTION AND ITS CRITIQUE
A. The Basic Argument
A widely accepted model of the economically rational actor assumes, among other things, that the preferences of an individual can be compared and ranked so as to constitute a complete and transitive hierarchy.(6) The model posits the existence of a single preference ordering, capable of representing every interest and value. This ordering determines the actor's choices among all possible alternatives and explains her behavior in each and every case.
The single-ordering assumption has come under severe attack over the years. One powerful critique claims that people have dual or even multiple preference orderings, reflecting the different roles they play in different situations. These distinct rankings cannot be reduced to one all-purpose ordering. In particular, critics of the single-ordering hypothesis distinguish between "personal" or "self-interested" preferences on the one hand and "social," "ethical," or "moral" ones on the other.(7) The variant of this distinction most directly bearing on provision and regulation of public goods focuses on the difference between the preferences people express in their "consumer" role and those they express in their "citizen" role.
The mere identification of different sets of preferences does not imply that the preferences declared in a person's "social" role are more virtuous or noble than the ones revealed in her "private" role, or that the former alone should form the basis for government regulation.(8) Proponents of the more specific consumer/citizen distinction, however, usually make the further claim that citizen preferences represent finer, more benevolent views concerning the common good. They therefore argue that this type of preference is the proper input for social decisionmaking.(9) The superiority of citizen preferences is commonly deduced from quasi-empirical observations of human behavior in various situations. For example, in their capacity as consumers, people may show no interest in the natural environment and refrain from visiting beautiful sites. Indeed, in day-to-day life, individuals frequently behave in a way that actively damages the environment--for example, by littering and polluting wilderness areas and parks. Yet, when answering public opinion polls, the same people (or at least some of them) demand that additional public funds be used to ensure the protection of the environment.(10) Many individuals do not donate to the poor or the sick and deliberately avoid homeless persons on the sidewalk. The same individuals may vote for candidates advocating increased welfare payments and universal health insurance. Similarly, people send their own children to expensive private schools, yet enthusiastically endorse integration and the improvement of public education.(11)
Explanations of these conflicting choices emphasize the virtues of the political process. Private or consumerist behavior reflects people's egoistic regard for their own interests whereas citizen behavior reflects their values and beliefs about society's interests. The differences in behavior result from the nature of the process through which the preferences are revealed: Market processes evoke self-interested choices whereas political processes encourage other-regarding ones.(12) From this assumption it follows that issues pertaining to the community as a whole, and to public goods in particular, should be resolved according to citizen preferences.
Consumer preferences are evoked not only through the interaction of individuals in the private market but also when the government uses economic techniques--like cost-benefit analysis or contingent valuation studies--to assess public policies and to determine the level of production of public goods. Such economic techniques employ the Kaldor-Hicks criterion of efficiency maximization. Thus, they recommend policies that produce enough potential gains to outweigh their costs. These gains are measured by the aggregate willingness of individuals to pay for the adoption of a given policy.(13) In other words, the value of any good is determined by estimating the amount of money with which people might be willing to part in order to satisfy their preference for it.
Some writers claim that this mode of inquiry is not appropriate for social policy and that it fails to capture important ways in which people value public goods. In the framework of cost-benefit analysis, every good is valued according to the extent that people wish to appropriate it and only as a means for satisfying their wants. The focus on the instrumental value of the good in promoting human welfare elicits self-interested and egoistic preferences. It does not capture the high intrinsic value that people attribute to certain goods, like beautiful landscapes, wild animals, or historic buildings. Goods of this kind are often valued for their own sake, independent of their impact on one's own welfare. People may want to protect endangered animals even if they have no wish or intention of ever "using" them to their benefit. Therefore, the fate of such goods should not depend on consumer preferences and on the results of a technical cost-benefit analysis.(14) In the political arena, in contrast, decisions are reached through argument and deliberation. The strength of a person's reasons for supporting a policy and the effectiveness of one's persuasion, rather than the quantity of money one is willing and able to pay, determine the outcome. Consequently, the political process encourages a mode of thinking that concentrates on the good of the community as a whole and its appropriate collective goals. Other-regarding preferences, concerning the welfare of other people and the intrinsic value of certain goods, are naturally generated by such a process. This social frame of mind is present when people answer a public opinion survey or cast a vote. Arguably, these citizen preferences should play the major role in formulating social policy and regulation.(15)
B. Critical Responses
Both the alleged sharp distinction between consumer and citizen preferences and the claim that economic analysis typically evokes and employs self-centered preferences have attracted criticism.
First, critics claim that the consumer/citizen distinction is based on an unrealistic, romanticized view of political life and an overly derogatory perception of private life. In reality, the argument goes, there is no qualitative difference between preferences manifested in these two spheres; other-regarding and self-interested preferences play a role in both market and political settings.(16)
Second, and more importantly, there is nothing in economic theory that denies the existence, or precludes the consideration, of other-regarding preferences in economic calculus. The economist is concerned with maximizing people's preferences, whatever their content.(17) These preferences certainly may be altruistic and benevolent,(18) as well as the products of informed reasoning and deliberation.(19) Furthermore, it is inaccurate to claim that economic calculus limits the benefits it attributes to any good to those accruing from people's use of the good. In particular, economic analysis recognizes the concept of "existence value," which represents the value derived from mere knowledge that certain goods exist. Existence value is the importance that people attribute to things, such as wilderness areas, which they have never seen and have no intention of ever visiting. This concept of "nonuse" value tries to capture altruistic and philanthropic values that are independent of any direct or personal benefit that a person may derive from a good.(20) The recognition of existence values diminishes, in a sense, the instrumentality of economic analysis.(21)
It should be noted that economists themselves are partly responsible for the impression that economic analysis is focused on egoistic maximizers. Although today most economists agree, in theory, that other-regarding sentiments may influence people's utility function, they usually prefer to explain behavior using narrowly defined self-interest. Indeed, they often point out that even behavior that is conventionally seen as ideological or altruistic is, in fact, motivated by self-interest.(22) The tendency to avoid explanations based on altruism is due, in part, to a belief that altruism is a relatively rare phenomenon outside the family sphere and therefore has little practical significance. In addition, if other-regarding preferences are admitted into economic models, then the ability of these models to predict and explain human behavior would be diminished.(23)
Be that as it may, the critical responses, though correct in principle, do not dispense with the problem raised by the consumer/citizen distinction. Even assuming that the preferences expressed in both private and public settings are a mix of self-centered and other-regarding preferences, one still needs to explain why they are not the same mix. In other words, people do behave differently in different settings--often expressing less self-interested views in political ones--and accounting for this phenomenon is necessary.(24) Only then will it be possible to decide which of the conflicting expressed preferences should be adopted by policymakers.
C. Other Explanations for the Distinction
In this Section, I examine other prevalent explanations for the differences between consumer and citizen behavior. Generally, these explanations focus on derogatory aspects of either consumer or citizen behavior. The Prisoner's Dilemma or Tragedy of the Commons explanation highlights consumers' motivation to free ride on the efforts of others. Other explanations focus on the insincerity or misjudgment underlying citizens' choices. I claim that none of these explanations can adequately account for the discrepancy in behavior.
1. Prisoner's Dilemma or Tragedy of the Commons
A sharp consumer/citizen distinction is rejected by standard economic theory. According to conventional economic wisdom, the differences between behavior in the market sphere and behavior in the political one are simply attributable to the free-riding behavior predicted by the Prisoner's Dilemma (PD) or Tragedy of the Commons (TOC) model.(25)
A PD or TOC problem may exist with regard to private provision of goods having the "public" characteristics of nonrivalry and nonexcludability.(26) The impracticability of excluding noncontributors from the benefits of the good (once it exists) and the inability of any one person to provide it independently create two motives that prevent market provision: "greed and fear."(27) The first motive results from the knowledge that the absence of one's small contribution will bear no noticeable effect on the outcome. Therefore, profits can be derived from free-riding on the efforts of others. The second motive is created by the possibility that one's own efforts will not be accompanied by enough contributions from others. Consequently, the good will not be provided and efforts spent on the collective enterprise will be wasted. Therefore, the rational course of action for individuals in the private sphere is noncontribution to the collective goal, regardless of the action taken by others. The PD or TOC scenario assumes the following ranking of preferences: (I) I do not contribute, but enough others do; (II) we all contribute; (III) no one contributes; (IV) I contribute, but not enough others do.(28) If individuals can neither communicate nor enter into enforceable agreements to cooperate, they may reason as follows: If sufficient people will contribute, it is preferable not to contribute and freely enjoy the good provided by others--receiving outcome (I) instead of (II). If not enough people will contribute, it is again better not to contribute, thereby insuring against the worst result, (IV). Thus, noncontribution is the dominant strategy in a PD game. The ensuing collective outcome--universal non-cooperation (III)--is, however, inferior to universal cooperation (II).(29)
The PD or TOC explanation rejects the view that people display different rankings of preferences (as distinguished from different choices) in different social roles, such as consumer and citizen. Take, for example, the discrepancy between the private behavior of polluting the environment and the public behavior of voting for stringent environmental protection laws. A PD or TOC explanation for this discrepancy does not assume that people assign environmental protection a higher ranking in the political sphere than in the private sphere, or that in the latter they prefer nonpreservation over preservation. The ranking is identical in both spheres. People's support for the environment is reflected in the fact that its protection forms their first- and second-best outcomes. The first outcome means that individuals wish for the environment to be preserved but prefer to be exempted from the costs. The second-best outcome is that everyone contributes to this goal. Nonpreservation (either as a result of universal non-contribution, or of too few contributions) is ranked lower (Outcomes III and IV). In daily life, individuals end up with the less-preferred third outcome due to the aforementioned motives of greed and fear. These motives are absent or ineffectual in the political sphere.(30) Within the political sphere, people usually cannot record or realize their preference for the highest ranked outcome of a PD--free-riding. At most, they can vote or voice an opinion for their second-best outcome--personal contribution if others do their share as well.
In contrast, proponents of the consumer/citizen distinction believe that people behave differently in public life because they give more weight and consideration to the welfare of others and to the good of the community as a whole. For example, I may be personally indifferent to the low quality of public schools in my city, since I can afford to send my children to elite, private institutions. Yet, when casting a vote On this issue, I may feel that all children deserve quality education and support policies likely to further this goal. Similarly, I may abhor certain kinds of music, preferring to purchase and listen to classical CDs only. Nevertheless, since I believe diversity is beneficial, I may endorse public funding for types of music that I neither appreciate nor enjoy.
The explanation offered by the consumer/citizen distinction theory for the failure of market provision of social goods seems to rest on the assumption that, in their consumer roles, people truly do not value the goods they decline to purchase or support. If this is so, in market settings individuals display the following preference ranking: (I) I do not contribute, but enough others do; (II) nobody contributes; (III) everybody contributes; (IV) I contribute, but not enough others do. In contrast to PD or TOC, mutual cooperation is not ranked as the second-best option. Consequently, there is no real dilemma in choosing the course of behavior in the private sphere: Because universal noncontribution is preferred to universal contribution, there is no reason to regret a decision not to cooperate. Noncooperation both guarantees the highest minimum payoff (Outcome II) and represents the only chance of realizing the most favored result (Outcome I).(31) However, this explanation for the consumer/citizen distinction must assume further that these consumer preferences will undergo a radical change in political settings: The alternative ranked third in the market scenario will climb to the top of the hierarchy in the political scenario.
Neither the PD explanation nor the suggestion of radical changes in preference orderings can exclusively or adequately account for the discrepancies between consumer and citizen behavior. As a sole (or primary) explanation, the PD or TOC theory represents an impoverished perception of human preferences and motivation. It is a simplistic and unrealistic portrayal, incompatible with both daily experience and behavioral studies.(32) The PD or TOC explanation implicitly acknowledges the accusation that choices in private life are largely motivated by egoism; greed has a significant impact on the final outcome, and people's most preferred option is that others toil while they enjoy a free ride. Individuals' behavior, however, displays a much wider range of motivations and interests. Greed, admittedly, may be one of them, but solidarity, fairness, and altruism are also powerful motivations. Thus, although PD or TOC may explain some cases in which private and public behavior diverge, this explanation is clearly insufficient.
Similar reservations apply to the contention that in the political sphere people usually hold different, more altruistic preference orderings. Individuals may indeed forgo choices that best promote their own interests and opt for those giving greater consideration to the interests of others. Yet it is doubtful that the preference rankings of people who are basically self-interested in their private lives will undergo radical transformation in their public lives.(33) It seems unwise to attribute too much weight to motivations based on self-sacrifice. Therefore, one must seek additional explanations for the more prevalent manifestation of other-regarding choices in the political arena.
2. Insincerity and Misjudgment Underlying Citizen Preferences
Some explanations for the greater manifestation of other-regarding preferences in political settings are based on individuals' perceptions regarding the effect of their choices on their lives. These explanations come in two somewhat different variants. The first concerns the impact of people's politically revealed preferences on their welfare. Because people know that it is improbable that their own vote will be decisive for any particular outcome, they feel free to support views that are inconsistent with their self-interest. They can rest assured that satisfying their conscience or whim for idealistic expression will not affect their welfare in any other way.(34) In other words, voting or answering public opinion surveys according to one's views regarding the common good is perceived as a "cheap" way of displaying altruism. In contrast, other-regarding decisions in market settings are much more expensive because they result in immediate and direct consequences for the individual altruist; thus, they are rarer. People do not put their money where their mouths are.
A second related explanation is based on people's shortsightedness and misjudgment. Adoption of any new public policy is a complex process. It involves a long chain of events, decisions by numerous individuals and much time. People may not realize that their vote or publicly expressed opinion--along with numerous other opinions and factors--will eventually result in a decision that will directly affect their lives and welfare. They fail to see that they will have to bear the costs of the policies they support.(35) For example, in public opinion surveys it is not rare to find answers to the effect that certain goals--such as environmental protection--should be pursued regardless of costs.(36) It is unlikely that these people really wish their preferences to be taken literally. Similarly, when expressing an opinion on a specific issue, a voter may not keep in mind that there are additional worthy causes, so that not all can be satisfied, and that her generosity, on this particular issue necessarily comes at the expense of other causes.
Both explanations offer unflattering reasons for the greater manifestation of other-regarding preferences in the political sphere. The first portrays people as insincere: They express concern for others only if they think that it will have no effect. The second explanation presents people as fools, incapable of understanding the long-term effects of their decisions. In the context of the consumer/citizen debate, these explanations suggest that citizen preferences may be based on caprice, misconceptions, or cognitive errors. People are unable to consider adequately the effects of social policies on their private lives and on things they hold dear. Therefore, consumer preferences, although they express more self-centered wants and interests, are as true and authentic a reflection of people's wants and values as citizen preferences.(37)
Undoubtedly, there is some truth in the above explanations, and their vindication of consumer preferences is especially important. Consumer preferences cannot be disregarded in social regulation, as there is a real danger that citizen preferences may be too idealistic for anyone's good. Like the above PD theory, however, these explanations are too simplistic and one-dimensional. Insincerity and misjudgment may form part of the explanation for the attitudes people express in the public sphere, but true commitment, responsibility, and solidarity seem just as important. The following Part of this Essay offers a different explanation for the discrepancy between consumer and citizen behavior. It rests on neither the superiority nor the inferiority of any type of preferences but rather on an alternative game theory model and on the nature of the pertinent good.
III. AN ALTERNATIVE EXPLANATION
In this Part, I offer a twofold explanation for the differences between consumer and citizen behavior. Citizen preferences may indeed manifest more socially-oriented or other-regarding views than do consumer preferences. This phenomenon, however, need not be due to the egoism of consumers or the inferiority of the market process. The first Section utilizes a game-theoretic approach to explain this phenomenon. The second Section argues that the nature of the good rather than the nature of the process determines the type of revealed preferences.
A. The Political Process as a Cooperative Assurance Game
As indicated earlier, the conventional economic explanation for the failure of private markets to provide public goods is the Prisoner's Dilemma (PD) or Tragedy of the Commons (TOC) model. Of the two characteristic motivations in a PD--greed and fear--the former is usually emphasized and seen as controlling. Thus, the PD or public goods problem is often referred to as a "free rider problem."(38) Greed is indeed a possible reason for this market failure. It is, however, not the only plausible reason. This phenomenon may be no less convincingly explained by a different kind of motivation, closer to that of fear, which I label "hopelessness." Both the knowledge that one's own contribution will have a minuscule effect on the desired outcome and the fear that not enough others will contribute may create a feeling of hopelessness: People reason that, regardless of what they choose to do, the collective goal will not be achieved. Therefore, they despair and do nothing. This motivation helps explain why people fail to achieve desired collective goals, even if they prefer to contribute and not to free ride. A more cooperative model of behavior--the Assurance Game (AG)--may clarify this idea and elucidate both the failures and successes of collective action.(39)
In an Assurance Game, the ranking of preferences is as follows: (I) everyone contributes; (II) no one contributes; (III) I do not contribute, but others do; (IV) I contribute, but others do not.(40) The two noncoordinated outcomes (III and IV) are less preferred, both because the desired goal is not achieved and because the contribution of some individuals is wasted. It seems reasonable to assume that the worst outcome is that one's own efforts, rather than those of others, are wasted.(41) In contrast to the PD game, mutual contribution, rather than free-riding, is ranked highest in the Assurance Game. While a PD has a dominant strategy, not to cooperate regardless of what other players do, there is no dominant strategy in an AG. Alternatives I and II are both equilibrium points, since no player has an incentive to depart from them once the other players' choices are revealed. Universal noncooperation (Outcome II) is, however, pareto inferior to universal cooperation (Outcome I). The outcome of the AG depends on the expectations of each person regarding the behavior of others. The optimal outcome can be achieved only if everyone is assured that everyone else will also "do the right thing."(42) Such assurance depends to a great extent on the existence and quality of information regarding the action that is likely to be taken by others (or likely to be expected by everyone to be taken by others).(43) If individuals cannot communicate or coordinate their behavior, and if they have insufficient trust in each other, they may opt for the maximin choice, which is non-cooperation. This choice guarantees that the worst possible outcome will be (III). Cooperation creates the danger of ending up with the least preferred outcome (IV).(44)
The AG model offers a realistic and sympathetic explanation for the failure to achieve collective goals in private settings. It does not assume radical changes in preference orderings between consumer and citizen roles and is not focused on one-dimensional, derogatory motivations.(45) This model shows that noncooperation is not necessarily the result of greed or a desire to free ride. Even those who rank mutual cooperation highest may behave noncooperatively, because they doubt the feasibility of attaining what they desire most.(46) In other words, even if the preferences people hold in private settings--indeed, in their roles as consumers--are extremely other-regarding and concerned with the public good, one might still outwardly observe what seems to be greedy and egoistic behavior. In these circumstances, identifying preferences with behavior is both mistaken and unfair.(47) It is not argued that hopelessness as to the expected behavior of others and acknowledgment of the futility of one's lone efforts will always obstruct other-regarding and cooperative behavior. The existence of private charities, nonprofit environmental groups, voluntary blood donations, and numerous other undertakings attest to the plausibility of successful collective endeavors by consumers. I do suggest, however, that the above motivations convincingly explain why the behavior observed in private settings is noncooperative or self-interested to a greater extent than in public settings. Frequently, revealed choices do not reflect people's most favored preferences, which may be markedly other-regarding.
The causes of failure in the private sphere are either absent or ineffectual in the political sphere. Political processes--such as elections and public opinion surveys--resemble a cooperative AG, through which people can test, with little risk to themselves, the possibility of realizing their most preferred options.(48) They do not stand to lose from expressing their preference for mutual cooperation in achieving collective goals and supplying public goods (Outcome I in the AG). Only if enough people voice similar views will these opinions translate into political power and eventually be adopted by regulators. Such adoption will result in mutual funding of the good (through government taxation, for example), which is what the AG player desires most.
Although a person's views may not receive sufficient support from others, the failure to realize one's highest-ranking preference is less personally damaging in the political sphere than in the private one. In the latter, the individual is faced with the danger of incurring substantial costs of time and money to no avail. In the former, the potential loss is mainly the relatively low cost of recording one's preferences in a ballot or a poll.(49) The political process successfully overcomes the grave informational problems that often prevent AG players from cooperating in private settings. This is because, in contrast with the private sphere, people do not have to know or estimate in advance what others are likely to do. It suffices that the results be known after everyone has expressed his or her preference. The operative action, implementing these results, will be carried out by regulators at a later time.(50)
The proposed explanation for the discrepancy between consumer and citizen behavior is contrary to the explanations based on people's assumption that the preferences they express as citizens will have no practical effect.(51) According to these explanations, other-regarding citizen preferences may reflect whims or misconceptions, rather than what people truly desire. Therefore, they should be taken with a grain of salt. The hopelessness explanation asserts that this discrepancy is caused by people's perception that there is a greater chance of realizing collective goals through political processes. Therefore, people are induced to express their most desired preferences in their civic role. It is not claimed that individuals falsely believe that their single vote (or view expressed in a public opinion survey) will be decisive for the adoption of any public policy. The AG model focuses neither on the phenomenon of voting nor on the question of why people bother to vote but rather on the content of their votes. It explains why preferences revealed in political settings have a greater tendency to be other-regarding. Since the political process guarantees mutual funding, people may view it as an effective and safe vehicle for expressing and achieving their favorable, authentic, collective goals.
Furthermore, the suggested reasoning implies that there is no sharp difference between the preference orderings of consumers and citizens. In this respect, the AG explanation coincides with the PD explanation.(52) The divergence between the two games is, however, more significant. A PD explanation assumes that individuals' highest-ranking preference is to free ride whenever possible. This motivation is the main reason for the market's failure to provide sufficient public goods. In contrast, the AG explanation rejects the supremacy of the free-riding motivation. It claims that in both consumer and citizen roles, cooperation may be ranked highest. The differences in observed choices occur because hopelessness causes lower-ranking preferences to be adopted in market settings.(53)
Empirical research confirms that people's estimation of the prospect of successful cooperation weighs heavily on their decision to contribute. Individuals are more willing to commit resources if they perceive that the probability of achieving the desired goal is high. In such circumstances, contrary to the assumptions of the PD, they often do not choose to free ride. When the probability of successful collective action is sufficiently low, however, even "hard core" cooperators may feel compelled to opt for the noncooperative choice. A few examples from the psychological research will demonstrate these claims.(54)
Various empirical studies have examined the correlation between the perceived efficacy of collective action (which often declines with group size) and the extent of individual cooperation. A highly relevant study by Norbert Kerr addressed not only the issue of perceived self-efficacy (the effect of one's own contribution on the production of a public good) but also the issue of perceived collective efficacy (the prospects of successfully joining efforts to achieve the desired goal).(55) It explored the connection between collective efficacy and group size. The most illuminating findings concerned the effect of individuals' errors regarding the prospects of successful cooperation. The participants thought that the chances of providing the public good were higher for smaller groups, even when the opposite was true. In other words, even when collective efficacy increased with group size, they still presumed that it decreased, and their willingness to cooperate declined accordingly.(56) This decline in willingness to cooperate is incompatible with the PD hypothesis, since in a PD noncontribution is the dominant strategy regardless of other people's behavior.(57) In contrast, people's perceptions of collective efficacy are highly relevant for the AG player, because her most preferred alternative, mutual contribution, is conditioned on the contribution of others. Errors in estimation and illusions of inefficacy enhance people's sense of hopelessness and the natural consequence is a decline in their willingness to cooperate.
The advantages of the AG model are also manifest in studies that examine the effect of communication on cooperation. Numerous experiments have found that communication between strangers increased cooperation significantly, even when subsequent choices were made anonymously and no binding and enforceable agreements were feasible.(58) For instance, one experiment investigated the effects of four levels of communication: (1) no communication; (2) communication about topics that are irrelevant to the social dilemma; (3) relevant communication; and (4) relevant communication with nonbinding announcements by all participants of their intended choice.(59) The cooperation rate in the first two levels of communication was around thirty percent, whereas the last two levels yielded cooperation rates of over seventy percent.(60) Another interesting result related to participants' predictions about other people's behavior. More defection was predicted when people could not communicate, and noncooperators predicted approximately four times as much defection as did cooperators. More accurate estimations were made in the two cases in which relevant communication was allowed and by people who actually participated in the game (as opposed to mere observers).(61)
Two explanations offered in the literature committed to the PD model for the observed increase in cooperation are enhancement of group identity and formation of commitments(62) The first explanation claims that communication enhances group solidarity, thereby causing people to place more weight on collective welfare and cooperation. This argument, however, is not strongly supported by the results of the experiment, since communication about issues unrelated to the social dilemma did not affect cooperation.(63) The second explanation is not very persuasive either. It asserts that communication affords opportunities for commitments to cooperation and that such nonbinding agreements encourage individuals to keep their promises. The difficulty lies in the fact that each individual can benefit from choosing noncooperation despite any nonbinding promises made by others. Were free riding ranked highest, then in circumstances in which agreements to cooperate are unenforceable and uncooperative behavior remains anonymous (and thus cannot be socially sanctioned), communication should not increase cooperation. In contrast, the AG model provides a simple solution to the puzzle. Since an AG player prefers mutual contribution over free riding, she will choose the former whenever she estimates that enough others will do the same. Communication generates information about the likely behavior of others, promotes trust between people, and fosters confidence in the possibility of collective action. Thus, it reduces the sense of hopelessness that obstructs successful cooperation.(64) Additional experiments support similar conclusions.(65)
The above experimental results are more compatible with an AG model than with a PD model. Two caveats, however, are in order. First, I do not claim that cooperation always fails in PD situations. For example, studies show that under certain, specified conditions, indefinitely-iterated plays between the same persons (the "supergame") enhance the chances of cooperation.(66) Yet, the required conditions did not exist in the studies presented above, in which strangers interacted only once and their decisions remained anonymous. Second, the hopelessness explanation is not founded on a naive belief that greed does not exist. I do, however, argue that the significance and importance of the free-riding motivation have been highly exaggerated. Consumer preferences for public goods are less self-centered and more other-regarding than is commonly believed. The AG model and the notion of hopelessness provide an explanation for the failures of voluntary cooperation that is as valid as the PD explanation, and the AG model is a superior explanation for the cases in which cooperation succeeds.
B. The Nature of the Good, Not the Nature of the Process
The AG model and the notion of hopelessness explain why, in market settings, people may fail to achieve their favored cooperative goals regarding public goods--even if they prefer to contribute and not to free ride. These causes of failure are absent in political settings, which therefore may authentically reflect the other-regarding preferences that exist unexpressed in the private sphere as well. This Section completes the argument by explaining why preferences concerning public goods tend to be other-regarding. I will argue that the conclusive factor is the nature of the good, rather than the nature of the process by which the good is produced. Thus, the relatively low manifestation of other-regarding preferences in the marketplace is largely due to the type of goods the market process can produce rather than to certain qualities of the process itself.
At least part of the reason for the prevalence of self-interested behavior in market settings is that the goods ordinarily produced in markets (and allocated through individuals' willingness to pay) are those that cannot be meaningfully enjoyed or used simultaneously with others: for example, apples, books, cars, cinema tickets, and residences. Such goods are intended and suited mainly for individual consumption (which may include some participation of family and close friends), and their consumption generally does not affect other people.(67) For this reason, the community ordinarily has (and should have) little interest in the purchase and use of these goods. We usually would not like (or allow) individuals to don their "citizen" cap and intervene in any person's decisions regarding these goods by claiming that benevolent consideration of that person's own well-being requires that she make different choices. In other words, the individualistic, self-centered character of market activities need not be viewed pejoratively. There is much virtue--at least in many private scenarios--in people minding their own welfare rather than the welfare of others.
In contrast, the goods typically produced in the political sphere are collaborative and nonrivalrous: for example, parks, roads, environmental protection, public education, and preservation of culturally significant buildings. To a great extent, they are indivisible and necessarily shared with others. Since common use and enjoyment are unavoidable, it is only natural that other people's interests will be considered in the decisionmaking process and that individuals will think in terms of the common good.(68) Moreover, because in many cases this type of good can be produced adequately only in the public sphere, we observe a greater manifestation of other-regarding preferences in political settings. In other words, the nature of the good is a dominant factor in determining both the plausible process of production and the prevalence of other-regarding preferences. The market process and the consumer role people play in it does not conjure, in itself, egoistic preference orderings. Self-centered behavior results from two interrelated phenomena. On the one hand, the market mechanism is best suited to supply rivalrous and excludable goods. On the other hand, goods that require common use and therefore invoke other-regarding sentiments are often incapable of market production.
The inherently common enjoyment of the goods characteristically produced through political processes also provides a justification for people's involvement with the well-being of others. If goods like environmental protection or public television broadcasting are provided at all, they are provided for everyone, supporters and opponents alike. Therefore, a distinction between personal preferences, which concern a person's own life and enjoyment of goods, and external preferences, which concern the welfare of others,(69) is inapplicable. Any choice a person makes based on her own preferences for these goods necessarily affects other people and may conflict with their preferences. Yet if we believe that it is legitimate for a person to decide what is best for herself, then we must also accept the inherent, unavoidable effects that her decision will have on others.(70)
Empirical research sustains the claim that the nature of the good--rather than the nature of the process--often determines the type of preferences that people form. In a study of the price-elasticity of private and public goods, Professor Green found that preferences regarding private goods are considerably more price-elastic than preferences for public goods. Small increases in the price of private goods (such as refrigerators, housing, or hardcover books) sharply reduce people's willingness to pay for them. In contrast, the willingness to fund or pay more taxes for public goods (like environmental protection or shelters for the homeless) is much less responsive to changes in prices.(71) The study demonstrates that preferences for public goods behave differently than those for private goods. They are less sensitive to considerations of personal cost and are more other-regarding. For our purposes, it is most revealing that the vehicle used to elicit preferences--people's willingness to pay for private and public goods--did not affect the results. In other words, although the experiments were framed in a consumerist way, social goods were treated differently from private ones.(72)
Another study examined the impact of surveys' questioning methods on the willingness of people to fund public goods.(73) Variations among the questionnaires included the payment vehicle (tax or voluntary contribution) and the inclusion or omission of a reminder about the number of additional people that will be asked to contribute. The most striking, statistically significant effects were related to the reminder. Although individuals' mean willingness to pay dropped by half (and, in one experiment, by two-thirds) when they were reminded that others would be asked to contribute as well, the reminder reduced the proportion of people that refused to contribute at all.(74) In contrast, the effect of the payment-vehicle variable was much less significant.(75)
The reminder effect cannot be adequately explained by the common assertion that people are strongly motivated to free ride when possible. Although mean contribution declined sharply, this was not due to an increase in refusals to pay. The percentage of free riders--those unwilling to contribute--actually decreased. This implies that people feel more obliged to participate in a collective endeavor when they know that others will participate as well. The expectation of reciprocity both creates a social norm and indicates the viability of the collective effort. Because additional people will share in the burden, however, individuals perceive that their own contribution can be less substantial.(76)
Both studies support the argument that the nature of the good--rather than the nature of the process by which it is produced--determines the type of preferences and, in particular, the tendency for other-regarding ones. The lesser manifestation of other-regarding preferences in market settings is due largely to the lower chances of realizing collective goals in the private arena. The higher prospects of successful cooperation through the political process encourage people to express their most favored social preferences in these settings.
IV. NORMATIVE IMPLICATIONS
I have suggested a twofold explanation for the greater manifestation of other-regarding preferences in the public arena: People's preferences regarding public goods are less self-centered than their preferences for private goods; and individuals fail to express these other-regarding preferences in their daily lives, primarily because they are aware of the implausibility of attaining them in the private sphere. This explanation has important normative implications.
First and foremost, it supports the claim that citizen preferences should be accorded substantial weight in social regulation. Moreover, the proposed explanation offers the soundest, least controversial grounds for interventionist regulation--more so than any other explanation discussed in this Essay, including the one suggested by the proponents of the consumer/citizen distinction. The consumer/citizen explanation rests on a sharp dichotomy between two distinct preference orderings. It assumes that citizen preferences accord substantially more weight to the welfare of others and to the good of the community. It therefore presumes that people support in their public lives things that they truly do not value--or value only to a lesser degree--in their private, self-centered lives. Consequently, further justification is required to favor one set of preferences over the other. This opens the door to much debate and to legitimate criticism of the excessive debasement of daily life. It is far from clear that the alleged idealistic set of preferences is a superior representation of people's genuine desires or that it better advances their well-being. The emergence of rival explanations, which ridicule citizen preferences and present them as misconceived, misinformed, whimsical, and hypocritical, is an expected, natural rejoinder to the attack on consumer preferences. In contrast, the explanation offered above suggests that there is no significant gap between consumer and citizen preferences. By adopting the latter set of preferences in social regulation, we neither abandon nor disregard the former. On the contrary, we assist consumers in realizing their most preferred alternatives, those they could not hope to realize in the marketplace.
The proposed explanation affords a more solid ground for government intervention than the common Prisoner's Dilemma (PD) or Tragedy of the Commons (TOC) explanation. A PD or TOC scenario assumes that people's highest preference is to free ride. They rank mutual contribution as the second-best option. By pursuing their most favored option, however, they end up with an even worse one--universal noncooperation. Government intervention can only help individuals realize their second-best option. In contrast, if it is true that a major driving force in private scenarios is hopelessness (not greed) and that the appropriate game-theory model is the Assurance Game (AG), then government intervention assists individuals in realizing their highest-ranking alternative.
The AG explanation does not imply that governmental provision of public goods or other forms of regulation are always necessary or justified. Whenever the sense of hopelessness does not exist or can be successfully eliminated, the other-regarding considerations may be reflected not only in people's preferences but also in their actual behavior. Both proponents of the consumer/citizen distinction and advocates of the PD or TOC explanation are necessarily skeptical of voluntary cooperation. If, however, contrary to these theories, individuals are generally other-regarding in their views on public goods, efforts to reduce feelings of hopelessness and facilitate cooperation in the private sphere may be fruitful.
Historic preservation may provide a useful illustration of this claim. Elsewhere I have argued that, under certain circumstances, deregulation may be successful with regard to historic preservation.(77) Although the benefits from preserving buildings of cultural importance are often a pure public good, self-preservation by their owners is sometimes feasible. This will generally be the case when a whole district or neighborhood is worthy of conservation and successful preservation will enhance the buildings' market value. Where enjoyment of the benefits of preservation depends on mutual cooperation of the district's homeowners, each owner's most-favored preference may be to preserve her building if (and only if) there is assurance that others do so as well. The prospects of realizing this favored outcome are high, due to certain characteristics of building preservation. Inter alia, cooperation in this context is extremely visible and easy to scrutinize. Noncooperation, that is, demolishing one's building, will immediately be noticed. In addition, a building owner's contribution to preservation is largely retrievable. A decision to preserve may be reversed by tearing down the building. Consequently, property owners need not fear that their efforts would be wasted and do not stand to lose much from postponing demolition decisions until the possibility of cooperation is tested and assured. There is no need to estimate in advance the likely behavior of others; one need only wait and see.(78) The government can assist and encourage this cooperative process by creating economic incentives for voluntary self-preservation, such as granting transferable development rights, tax benefits, or convenient loans.(79) More generally, it may be worthwhile to develop institutions that can coordinate people's expectations and thus facilitate the voluntary provision of public goods.(80)
Another implication of the above analysis concerns the role of economic tools in defining the scope of state intervention and the quantity of public goods to be produced. My justification for adopting citizen preferences neither rests on the inferiority of consumer preferences nor accepts the claim that such preferences are markedly egoistic. Empirical studies demonstrate that social goods elicit other-regarding responses even in consumerist settings. Therefore, the contention that economic tools, such as cost-benefit analysis, should not be used because they evoke self-interested preferences is unwarranted. The results of cost-benefit analysis should be used cautiously, however, for other reasons. First, since every individual's private preferences for public goods necessarily affect other individuals, one cannot rely on a process of aggregation and maximization alone. Usually people can avoid neither "enjoying" public goods they do not value nor participating in their funding. Although the inherent coercion of minority opposition cannot be avoided, its negative impacts can be decreased through evaluation of the reasons underlying the majority's preferences. One needs to be convinced that there are good arguments for supporting the adoption of any given policy. It does not suffice that the expected gains for the majority outweigh the expected losses for the minority.(81) Thus, cost-benefit analysis alone should not determine the fate of any social proposal. Second, one should examine whether cost-benefit analysis can accurately capture people's valuation of public goods. Severe problems in valuing these goods may prompt us to be cautious about--although not to forgo--reliance on economic tools for practical, rather than theoretical, reasons.(82)
This Essay analyzes the discrepancy between consumer and citizen behavior and offers a twofold explanation for this phenomenon. First, people's preferences for public goods are more other-regarding than their preferences for private goods. Second, individuals may fail to express these socially-oriented preferences in daily life because of the perceived hopelessness of realizing their highest-ranking collective goals in private settings. Thus, the phenomenon that citizens' revealed preferences have a greater tendency to be other-regarding is due neither to the inferiority of consumer preferences and market processes nor to the superiority of the political process. The differences between the preference orderings that people hold in their roles as consumers and citizens are considerably less significant than proponents of the consumer/citizen distinction assume. This, however, does not amount to a rejection of multiple utility frameworks. I only deny the generalization that consumers and citizens hold distinct preference orderings that are markedly self-interested for the former and markedly other-regarding for the latter.
The proposed explanation has broad normative implications, both for regulation and for deregulation. This explanation supports the claim that citizen preferences should receive substantial weight in social regulation, and justifies the major role of the government in the provision of public goods. It further legitimizes a careful use of economic tools in the regulatory process. At the same time, the suggested explanation leaves room for optimism regarding the feasibility of voluntary provision of public goods, when feelings of hopelessness do not prevail or may be successfully eliminated.
To the extent that this Essay's proposition is correct, people are less schizophrenic than the Jekyll-Hyde metaphor suggests. Neither heroic nor villainous, individuals wish to achieve collectively beneficial goals, but they are conscious of their own limitations and weakness. Yet the very existence of these aspirations facilitates success in the public arena.
(1.) This assumption will not always be true. Sometimes public goods may be turned into private ones and supplied by the market. For instance, an entrepreneur may construct a road or a park. He may exclude free riders and recoup his expenses by fencing, installing toll booths, and charging entrance fees. It is commonly accepted, however, that the market cannot supply the entire demand for all public goods. See. e.g., Thomas W. Merrill, The Economics of Public Use, 72 CORNELL L. REV. 61 (1986). Therefore, nonmarket decisions regarding the choice and appropriate quantity of public goods are unavoidable.
(2.) A leading article making this claim is Amartya K. Sen, Rational Fools: A Critique of the Behavioral Foundations of Economic Theory, 6 PHIL. & PUB. AFF. 317 (1977). Subsequent contributions include Amitai Etzioni, The Case for a Multiple-Utility Conception, 2 ECON. & PHIL. 159 (1986); and Howard Margolis, A New Model of Rational Choice, 91 ETHICS 265 (1981).
(3.) See ELIZABETH ANDERSON, VALUE IN ETHICS AND ECONOMICS 144-47, 158-59, 203-10 (1993); MARK SAGOFF, THE ECONOMY OF THE EARTH: PHILOSOPHY, LAW, AND THE ENVIRONMENT 714 (1988); Cass R. Sunstein, Endogenous Preferences, Environmental Law, 22 J. LEGAL STUD. 217, 242-43 (1993) [hereinafter Sunstein, Endogenous Preferences]; Cass R. Sunstein, Social Norms and Social Roles, 96 COLUM. L. REV. 903, 923-25 (1996) [hereinafter Sunstein, Social Norms].
(4.) See, e.g., Geoffrey Brennan & James Buchanan, Voter Choice: Evaluating Political Alternatives, 28 AM. BEHAV. SCIENTIST 185, 194-99 (1984); Sunstein, Endogenous Preferences, supra note 3, at 242-43.
(5.) The consumer/citizen distinction is only one problem faced by policymakers wishing to consider or satisfy preferences. For example, preferences can be a function of the initial allocation of legal entitlements (the "endowment effect") and may adapt to what is perceived by people as available (the "sour grapes" phenomenon). In both cases, the result may be a strong bias in favor of the status quo. See JON ELSTER, SOUR GRAPES: STUDIES IN THE SUBVERSION OF RATIONALITY (1983); Tyler Cowen, The Scope and Limits of Preference Sovereignty, 9 ECON. & PHIL. 253 (1993); Sunstein, Endogenous Preferences, supra note 3, at 221-42; Richard Thater, Toward a Positive Theory of Consumer Choice, 1 J. ECON. BEHAV. & ORG. 39, 44 (1980). Such problems have important implications for social regulation. Nevertheless, consideration of people's preferences is essential in a liberal democracy. Thus, if people indeed manifest conflicting preferences in different social roles, the choice between consumer and citizen preferences merits close examination.
(6.) A notable example is GARY S. BECKER, ACCOUNTING FOR TASTES 3-23 (1996).
(7.) See, e.g., Sen, supra note 2, at 326; John C. Harsanyi, Cardinal Welfare, Individualistic Ethics, and Interpersonal Comparisons of Utility, 63 J. POL. ECON. 309, 315 (1955); Etzioni, supra note 2; Robert E. Goodin, Laundering Preferences, in FOUNDATIONS OF SOCIAL, CHOICE THEORY 75, 87-91 (Jon Elster & Aanund Hylland eds., 1986) James G. March, Bounded Rationality, Ambiguity, and the Engineering of Choice, in RATIONAL CHOICE 142, 161 (Jon Elster ed., 1986); Margolis, supra note 2. For a general appraisal and critique of the multiple utility theories, see Timothy J. Brennan, A Methodological Assessment of Multiple Utility Frameworks, 5 ECON. & PHIL. 189 (1989).
(8.) Two examples are Amartya Sen's and John Harsanyi's positions. Sen distinguishes between a choice that maximizes an individual's personal welfare, which may include consideration of other people's well-being as long as their welfare affects her own, and a choice that gives a preference to an alternative that results in a lower level of personal welfare for the actor. The former is labeled "sympathy" and the latter "commitment." See Sen, supra note 2, at 326. Commitment involves counter-preferential choice and may be caused by factors like altruism, education, or social responsibility. Consequently, an individual may choose to forgo her subjective preference and behave in accordance with other people's preferences or the best interests of the community as a whole. Sen does not imply, however, that preferences based on commitment are in any sense "better" than those based on personal-welfare maximization. See id. at 326-29. He suggests a technique of meta-ranking (or rankings of preference rankings) to capture the process of choice between the different orderings and admits that such meta-ranking may result in an incomplete ordering of all possible alternatives. See id. at 337-38; cf. Amartya Sen, Behaviour and the Concept of Preference, in RATIONAL CHOICE, supra note 7, at 60, 74-77 (discussing an application of his theory to collective action problems).
Harsanyi distinguishes between "personal" or "subjective" preferences, which convey one's own personal point of view, and "ethical" or "moral" preferences, which reflect impersonal and impartial considerations of what is good from a social point of view. Like Sen, Harsanyi does not think that personal/subjective preferences are necessarily egoistic or should rank lower than ethical/moral ones. See Harsanyi, supra note 7, at 315; John C. Harsanyi, Morality and the Theory of Rational Behaviour, in UTILITARIANISM AND BEYOND 39, 47-48 (Amartya Sen & Bernard Williams eds., 1982). On the subject of "second order preferences," or preferences over preferences, see Harry G. Frankfurt, Freedom of the Will and the Concept of a Person, 68 J. PHIL. 5 (1971), which argues that second order preferences are a distinguishing characteristic of human beings.
(9.) Throughout the Essay, I shall assume that regulators must decide which form of actual preferences--consumer or citizen--should be chosen as an input for policy-making. It should be noted, however, that although commonly used in theory and practice, actual preferences are not the only plausible criteria for measuring people's well-being. Two other candidates are "hypothetical preferences" and "objective list" theories. A hypothetical preferences theory (also called "ideal preferences," "informed desire," or "true preferences") does not focus on the preferences a person actually has but rather on those she would have had if she possessed all the relevant information and had the emotional and intellectual capabilities to conduct an accurate examination of all alternatives and their possible consequences. See JAMES GRIFFIN, WELL-BEING: ITS MEANING, MEASUREMENT, AND MORAL IMPORTANCE 11-17 (1986); SHELLY KAGAN, NORMATIVE ETHICS 38 (1998); HENRY SIDGWICK, THE METHODS OF ETHICS 109-13 (Hackett 1981) (1907); Harsanyi, supra note 8, at 55-56. Justice Marshall's concurring opinion in Furman v. Georgia, 408 U.S. 238, 361-62, 369 (1972) (Marshall, J., concurring), in which he argues that a fully informed citizenry would reject the death penalty, uses similar hypothetical reasoning.
An objective list theory focuses on how to maximize people's access to and use of certain good things. This theory presents some list of things that people should have, regardless of whether these same people actually desire them. See GRIFFIN, supra, at 40-55; KAGAN, supra, at 39-40; DEREK PARFIT, REASONS AND PERSONS 493 app. I at 499-501 (1984); Thomas Scanlon, Value, Desire, and Quality of Life, in THE QUALITY OF LIFE 185, 186-91 (Martha Nussbaum & Amartya Sen eds., 1993). The consumer/citizen debate is hardly relevant in the context of an objective list theory, as this theory is not tied to preferences, actual or hypothetical.
Inasmuch as hypothetical preferences depart from the subjective desires and capabilities of individuals, they are significantly distinct from the preferences of real human beings. In the extreme, it may be argued that the hypothetical preferences theory is not a theory of preferences at all. However, to the extent that hypothetical preferences are taken to be similar to actual preferences and can thus assume either a consumer or citizen character, the debate is relevant to these preferences as well.
(10.) Hardin points out the enormous divergence between the number of Americans who profess their concern for environmental issues (over 100 million) and the number of Americans who actually contribute to environmental organizations (about one million). Even the one percent contributing spend very low sums to further their desired goals--around $10 per person, per year on average. The divergence with respect to other honored causes--such as civil rights, women's movements, and gun control--is even more dramatic. See RUSSELL HARDIN, COLLECTIVE ACTION 11, 105-07 (1982).
(11.) See SAGOFF, supra note 3, at 7, 50-53; Mark Sagoff, Economic Theory and Environmental Law, 79 MICH. L. REV. 1393, 1403 (1981); Sunstein, Endogenous Preferences, supra note 3, at 242-43. Another example given in support of the claim that the political process evokes behavior that is not aimed at maximizing self interest is the phenomenon of voting. People must realize that in large elections no single vote is likely to alter the outcome. Since the act of voting involves some costs (such as time and transportation costs), it is seemingly irrational to vote. Nevertheless, the turnout in elections is usually high. Therefore, political participation in elections is not guided by the desire to maximize private gains but by a desire "to do the right thing," to fulfill a civic duty, etc. See DONALD P. GREEN & IAN SHAPIRO, PATHOLOGIES OF RATIONAL CHOICE THEORY 47-71 (1994); IAIN MCLEAN, PUBLIC CHOICE: AN INTRODUCTION 45-49 (1987); Goodin, supra note 7, at 89-90; Sen, supra note 2, at 332-33.
(12.) See ANDERSON, supra note 3, at 141-47, 158-63, 190-216; SAGOFF, supra note 3, at 7-14; Etzioni, supra note 2, at 164-65, 176; Goodin, supra note 7, at 86-91; Steven Kelman, "Public Choice" and Public Spirit, PUB. INTEREST, Spring 1987, at 80, 85-86, 88-90, 93-94; Margolis, supra note 2, at 266, 277-79; Sagoff, supra note 11, at 1394-95, 1402-03, 1410-12; Sunstein, Endogenous Preferences, supra note 3, at 243-47, 254; Cass R. Sunstein, Legal Interference with Private Preferences, 53 U. CHI. L. REV. 1129, 1140-45 (1986); Sunstein, Social Norms, supra note 3, at 924, 960. It should be noted that Sunstein's view of the consumer/citizen distinction is less extreme. He is careful not to label either type of preferences as "true" and not to draw a sharp line between the egoism of the private market and the benevolence of the political process. Sunstein tends to agree, however, that the disparities between consumer and citizen behavior support the use of citizen preferences in matters of social regulation. See Sunstein, Endogenous Preferences, supra note 3, at 247; Sunstein, Social Norms, supra note 3, at 960.
(13.) See, e.g., ROBERT SUGDEN & ALAN WILLIAMS, THE PRINCIPLES OF PRACTICAL COST-BENEFIT ANALYSIS 89-91 (1978); DAVID L. WEIMER & AIDAN R. VINING, POLICY ANALYSIS: CONCEPTS AND PRACTICE 259-75 (2d ed., 1992). "Willingness To Pay" (WTP) is most commonly used to estimate the gains of any project. Significantly different estimations result from the use of another criterion--"Willingness To Accept" (WTA). This represents the sum of money potential losers will require as compensation for putting up with the project. People usually demand much higher sums for agreeing to a project they oppose than they would be willing to pay for one they wish to endorse. See ROBERT CAMERON MITCHELL & RICHARD T. CARSON, USING SURVEYS TO VALUE PUBLIC GOODS: THE CONTINGENT VALUATION METHOD 30-38 (1989); Daniel S. Levy & David Friedman, The Revenge of the Redwoods? Reconsidering Property Rights and the Economic Allocation of Natural Resources, 61 U. CHI. L. REV. 493, 506-15 (1994). For the purposes of this study, however, there is no need to distinguish between the two criteria. Like WTP, WTA evaluates goods in monetary terms and focuses on their instrumental value in advancing people's welfare. Therefore, the two criteria are subject to many of the same criticisms.
(14.) According to this critique, cost-benefit analysis is a theoretically and inherently deficient tool with regard to social policy. This claim is more radical than the more common one, which focuses on the severe difficulties encountered in making an accurate estimate of costs and benefits or on the manipulability of the results of such an analysis.
(15.) The most forceful and elaborate presentations of this argument are found in the writings of Elizabeth Anderson and Mark Sagoff. See ANDERSON, supra note 3, at 190-95, 203-16; SAGOFF, supra note 3, at 26-29, 40-42, 50-57, 68-72, 92-97; Sagoff, supra note 11, at 1410-18. For additional arguments against "scientific" reasoning and commensurability, and in defense of decision-making that utilizes emotions and imagination, see MARTHA C. NUSSBAUM, POETIC JUSTICE: THE LITERARY IMAGINATION AND PUBLIC LIFE (1995); and Martha C. Nussbaum, The Discernment of Perception: An Aristotelian Conception of Private and Public Rationality, in LOVE'S KNOWLEDGE: ESSAYS ON PHILOSOPHY AND LITERATURE 54 (1990). On various models of the political process, see generally Jon Elster, The Market and the Forum: Three Varieties of Political Theory, in FOUNDATIONS OF SOCIAL CHOICE THEORY, supra note 7, at 103.
(16.) The most forceful exposition of this argument is Carol M. Rose, Environmental Faust Succumbs to Temptations of Economic Mephistopheles, or, Value by any Other Name Is Preference, 87 MICH. L. REV. 1631, 1635-39 (1989). For others, see Bruce Chapman, Rational Environmental Choice: Lessons for Economics from Law and Ethics, 6 CAN. J.L. & JURIS. 63, 65-66 (1993); Daniel A. Farber, Environmentalism, Economics, and the Public Interest, 41 STAN. L. REV. 1021, 1028-31, 1039-40 (1989); and Daniel A. Farber, From Plastic Trees to Arrow's Theorem, 1986 U. ILL. L. REV. 337, 342-44 [hereinafter Farber, Arrow's Theorem].
(17.) See Rose, supra note 16, at 1633-35.
(18.) See BECKER, supra note 6, at 4, 139; see also WILLIAM F. BAXTER, PEOPLE OR PENGUINS: THE CASE FOR OPTIMAL POLLUTION 5-7, 16 (1974) (setting forth criteria for the analysis of environmental problems that do not exclude private altruism).
(19.) See Amartya Sen, Foundations of Social Choice Theory: An Epilogue, in FOUNDATIONS OF SOCIAL CHOICE THEORY, supra note 7, at 213, 233-35. For a similar argument, see Chapman, supra note 16, at 64.
(20.) Existence value represents the value a person attributes to a good's enjoyment by other people and future generations and to the intrinsic worth of certain goods. This concept was developed and is mainly used in relation to environmental protection and wildlife preservation. See MITCHELL & CARSON, supra note 13, at 63-67; John V. Krutilla, Conservation Reconsidered, 57 AM. ECON. REV. 777, 781 (1967); Alan Randall, Human Preferences, Economics, and the Preservation of Species, in THE PRESERVATION OF SPECIES 79, 84-85 (Bryan G. Norton ed., 1986). This relatively new concept supplements other "use-oriented" methods of estimating the value of unmarketed goods; examples of these latter methods are comparing the prices of houses that enjoy beautiful views or clean air to those of houses that do not, or examining the costs people incur in traveling to national parks. See THOMAS M. POWER, ENVIRONMENTAL PROTECTION AND ECONOMIC WELL-BEING: THE ECONOMIC PURSUIT OF QUALITY 71-74, 96-104 (2d ed. 1996); SUGDEN & WILLIAMS, supra note 13, at 148-167. The criticism of economic tools, discussed in supra notes 13-14 and accompanying text, is more applicable to these "use-oriented" methods of valuation.
(21.) It is true that existence value is estimated through the criterion of willingness to pay and is still a factor in the welfare functions of individuals. However, those who emphasize the intrinsic value of certain goods, such as the environment, do not usually claim that the effects of the goods on people are irrelevant or that the goods are important regardless of the existence of human beings or their consciousness of these goods. Indeed, they focus on the appropriate ways that people should value the goods. See Donald H. Regan, Duties of Preservation, in THE PRESERVATION OF SPECIES, supra note 20, at 195, 196-203. Thus, the theoretical gap between the two views is narrower than it might seem at first glance. Whether existence value can be accurately measured by economic tools is a different question. See infra notes 81-82 and accompanying text.
(22.) See, e.g., JON ELSTER, ULYSSES AND THE SIRENS: STUDIES IN RATIONALITY AND IRRATIONALITY 141-44, 153-56 (1979); DANIEL A. FARBER & PHILIP P. FRICKEY, LAW AND PUBLIC CHOICE: A CRITICAL INTRODUCTION 12-37 (1991); Brennan, supra note 7, at 201; Russell Hardin, Autonomy, Identity, and Welfare, in THE INNER CITADEL: ESSAYS ON INDIVIDUAL AUTONOMY 189, 190-91 (John Christman ed., 1989) ("[M]uch of actual economic reasoning is based on the assumption that actors are rational in the sense of being serf-interested."); see also BECKER, supra note 6, at 151 (discussing the possibility that potential beneficiaries act as if they are altruistic out of self-interested motives); cf. Amartya Sen, Rationality and Social Choice, 85 AM. ECON. REV. 1, 15-16 (1995) (criticizing the conventional economic view).
(23.) See Brennan, supra note 7, at 201; Etzioni, supra note 2, at 161-63,171; see also Joseph P. Kalt & Mark A. Zupan, Capture and Ideology in the Economic Theory of Politics, 74 AM. ECON. REV. 279, 280 (1984) (noting that economic models of regulation generally assume that the publicly interested goals of individuals are "empirically uninteresting and dispensable").24. See supra notes 9-11 and accompanying text.
(25.) See Farber, Arrow's Theorem, supra note 16, at 342; Kalt & Zupan, supra note 23, at 281-82. In general, the TOC may be regarded as a specific manifestation of the PD model: Common resources are over-exploited by individuals due to their PD-type preference orderings. For the purposes of this Essay, however, the terms PD and TOC are interchangeable, since they employ the same game-theoretic model and involve the same ranking of preferences. For conventional analyses of public goods as an instance of a PD or TOC, see, for example, ALLAN M. FELDMAN, WELFARE ECONOMICS AND SOCIAL CHOICE THEORY 112-14 (1980); GREEN & SHAPIRO, supra note 11, at 72-77; HARDIN, supra note 10, at 16-28; DENNIS C. MUELLER, PUBLIC CHOICE II, 9-15 (1989); EDNA ULLMAN-MARGALIT, THE EMERGENCE OF NORMS 49-53 (1977); and Garrett Hardin, The Tragedy of the Commons, 162 Sci. 1243, 1244-46 (1968).
(26.) Classical examples of pure public goods are national defense and lighthouses. See FELDMAN, supra note 25, at 106-07; MUELLER, supra note 25, at 10-11. But see FRED FOLDVARY, PUBLIC GOODS AND PRIVATE COMMUNITIES 7-9 (1994) (noting that some private provision of public goods, like lighthouses, does exist).
(27.) Anatol Rapoport, Prisoner's Dilemma--Recollections and Observations, in RATIONAL MAN AND IRRATIONAL SOCIETY?--AN INTRODUCTION AND SOURCEBOOK 71, 77 (Brian Barry & Russell Hardin eds., 1982).
(28.) The payoff matrix of a Prisoner's Dilemma Game is as follows (with the payoffs of Player One listed first and the parenthetical roman numerals indicating Player One's preference ordering):
Player Two Contribute Do Not Contribute Player One Contribute (II) 4,4 (IV) 0.8 Do Not (I) 8,0 (III) 2,2 Contribute
(29.) See R. DUNCAN LUCE & HOWARD RAIFFA, GAMES AND DECISIONS 94-97 (1957); see also ANDREW M. COLMAN, GAME THEORY AND ITS APPLICATIONS IN THE SOCIAL AND BIOLOGICAL SCIENCES 115-18, 201-09 (2d ed., Butterworth-Heinemann 1995) (explaining and providing examples of PD games and analogous multiperson social dilemmas); HARDIN, supra note 10, at 22-25 (explaining the genesis, function, and appeal of the PD game).
(30.) For further explanation and elaboration, see infra Section III.A.
(31.) See GREEN & SHAPIRO, supra note 11, at 75-77; Daphna Lewinsohn-Zamir, The "Conservation Game": The Possibility of Voluntary Cooperation in Preserving Buildings of Cultural Importance, 20 HARV. J.L. & PUB. POL'Y 733, 759-60 (1997).
(32.) For various examples of behavior incompatible with the assumption of narrow self-interest maximization, see AMITAI ETZIONI, THE MORAL DIMENSION: TOWARD A NEW ECONOMICS, 51-66 (1988); HOWARD MARGOLIS, SELFISHNESS, ALTRUISM, AND RATIONALITY: A THEORY OF SOCIAL CHOICE 17-25 (1982); Robert C. Ellickson, Bringing Culture and Human Frailty to Rational Actors: A Critique of Classical Law and Economics, 65 CHI.-KENT L. REV. 23, 50-54 (1989); and Jon Elster, Selfishness and Altruism, in BEYOND SELF-INTEREST 44 (Jane J. Mansbridge ed., 1990). See also Norman Frohlich & Joe Oppenheimer, Beyond Economic Man: Altruism, Egalitarianism, and Difference Maximizing, 28 J. CONFLICT RESOL. 3 (1984) (showing that altruism and egalitarianism play a significant role in individuals' choices); Daniel Kahneman et al., Fairness and the Assumptions of Economics, in RATIONAL CHOICE: THE CONTRAST BETWEEN ECONOMICS AND PSYCHOLOGY 101 (Robin M. Hogarth & Melvin W. Reder eds., 1986) (demonstrating the willingness of people to enforce fairness at some cost to themselves). For a survey of experiments questioning the validity of the free-riding assumption, see infra notes 54-66, 73-76 and accompanying text.
(33.) On the rarity of this phenomenon, see ROBERT SUGDEN, THE POLITICAL ECONOMY OF PUBLIC CHOICE: AN INTRODUCTION TO WELFARE ECONOMICS 32 (1981); and Harsanyi, supra note 7, at 315-16. Empirical evidence supports the assumption that self-interested, personal preferences do not undergo radical changes in the political sphere. See, e.g., Donald Philip Green & Ann Elizabeth Gerken, Self-Interest and Public Opinion Toward Smoking Restrictions and Cigarette Taxes, 53 PUB. OPINION Q. 1 (1989) (showing that nonsmokers are far more supportive of tightening smoking restrictions and increasing cigarette taxes than smokers). In addition, psychological experiments have shown a direct correlation between individuals' values and their propensities to cooperate with others. People having cooperative social value orientations behave more cooperatively than individualists and competitively motivated persons. See Roderick M. Kramer et al., Social Values and Cooperative Response to a Simulated Resource Conservation Crisis, 54 J. PERSONALITY 576 (1986); Wim B.G. Liebrand & Godfried J. van Run, The Effects of Social Motives on Behavior in Social Dilemmas in Two Cultures, 21 J. EXPERIMENTAL SOC. PSYCHOL. 86 (1985); see also COLMAN, supra note 29, at 221-23 (reporting and summarizing the findings of the psychological experiments). In the same vein, a Dutch survey that measured people's willingness to cooperate in helping the government achieve environmental goals--e.g., recycling of chemical household waste and energy saving in the household--has found a high rate of consistency between reported motives, rankings of preferences, and choices. In particular, respondents' motives (values, norms, desires, and personal tastes) toward environmental issues constrain their preference rankings and their decisions about whether to cooperate. See Robert J. van der Veen, How Motives Speak to Preferences 13-32 (June 1996) (unpublished manuscript, on file with The Yale Law Journal). Thus, one may reasonably assume that a self-centered person is less likely to alter his behavior in political settings, and that a socially-oriented person is likely to cooperate in private settings as well.
(34.) See Brennan & Buchanan, supra note 4, at 194-99; Kalt & Zupan, supra note 23, at 281-82; Kelman, supra note 12, at 90; see also MUELLER, supra note 25. at 367 (summarizing Brennan and Buchanan's argument).
(35.) See Sunstein, Endogenous Preferences, supra note 3, at 242-43. On problems in the estimation of risks, see generally Amos Tversky & Daniel Kahneman, Judgment Under Uncertainty: Heuristics and Biases, 185 SCIENCE 1124 (1974).
(36.) See, e.g., COUNCIL ON ENVTL. QUALITY ET AL, PUBLIC OPINION ON ENVIRONMENTAL ISSUES: RESULTS OF A NATIONAL PUBLIC OPINION SURVEY 3, 12 (1980); John M. Gillroy & Robert Y. Shapiro, The Polls: Environmental Protection, 50 PUB. OPINION Q. 270, 279 (1986); see also infra note 82 (describing a comparable phenomenon regarding people's stated willingness to pay in contingent valuation studies).
(37.) Indeed, these explanations may support a stronger claim that consumer preferences represent a more accurate and realistic expression of people's preferences. See, e.g., Brennan & Buchanan, supra note 4, at 199. For forceful elaborations of the claim that preferences are context-dependent and that one cannot speak about "true" preferences, see Sunstein, Endogenous Preferences, supra note 3; and Sunstein, Social Norms, supra note 3, at 960-61.
(38.) See supra notes 25-29 and accompanying text.
(39.) The "Assurance Problem" was introduced by Amartya Sen. See Amartya K. Sen, Isolation, Assurance and the Social Rate of Discount, 81 Q.J. ECON. 112, 114 (1967). For a discussion of public goods that focuses on the Assurance Game model, see Carlisle Ford Runge, Institutions and the Free Rider: The Assurance Problem in Collective Action, 46 J. POL. 154 (1984). Professor Runge's analysis differs from mine in several important respects. While I argue that a sense of hopelessness obstructs mutual cooperation in private settings, Runge argues that the problem lies in people's sense of fairness. Even if their own contribution may further the desired goal, they are unwilling to be the lone contributor because bearing such a burden alone seems unfair. See id. at 160-61. While I claim that in the context of public goods people naturally consider the interests of others, see infra Section III.B, Runge denies the presence of any altruistic motives. See Runge, supra, at 160-61. Furthermore, whereas my analysis focuses on a critique of the consumer/citizen distinction and is primarily directed at the evaluation and justification of governmental provision of public goods, Runge centers on the conditions under which public goods may be provided voluntarily. See id. at 155, 158, 171-72. I will discuss a "fairness" explanation in the context of the consumer/citizen distinction, see infra note 53, as well as the possibility of voluntary provision of public goods, see infra text accompanying notes 77-80.
(40.) The payoff matrix of an Assurance Game is as follows (with the payoffs of Player One listed first and the parenthetical roman numerals indicating Player One's preference ordering):
Player Two Contribute Do Not Contribute Player One Contribute (I) 8,8 (IV) 0,2 Do Not (III) 2,0 (Il) 4,4 Contribute
The term "Assurance Game" is sometimes used to describe a slightly different preference ordering: (I) everyone contributes; (II) I do not contribute but others do; (III) no one contributes; (IV) I contribute, but others do not. See, e.g., ELSTER, supra note 22, at 20 (referring to the version of the Assurance Game not used in this Essay); MICHAEL TAYLOR, THE POSSIBILITY OF COOPERATION 38 (1987) (describing both versions). The difference between the two versions lies in the reversal of the second and third priorities. That variation reflects differences in the models' underlying assumptions: While the model 1 use assumes that mutual contribution is a precondition to achieving the desired goal, the other version assumes that a person can successfully free ride on the efforts of others (her second-ranking preference). Although both versions assume that people prefer mutual cooperation to free riding, and thus their normative implications largely converge, the version I employ better captures the hopelessness explanation that I describe in this Essay.
(41.) The ranking of Outcome III before Outcome IV should not be regarded as a manifestation of greed. In both situations, not enough people contribute, so some people's efforts are wasted. The magnitude of this waste may be the same. The AG player is not assumed to be a "saint." One may realistically presume that she favors a situation in which she does not belong to the group that toils in vain.
(42.) Sen, supra note 39, at 114, 122 (using the phrase in his analysis of people's decisions regarding optimum saving rates); see also TAYLOR, supra note 40, at 38-39 (applying AG analysis to the maintenance of dikes and ditches for irrigation or flood control by two cultivators, and to the weeding of a shared vegetable patch); Carlisle Ford Runge, Common Property Externalities: Isolation, Assurance, and Resource Depletion in a Traditional Grazing Context, 63 AM. J. AGRIC. ECON. 595. 600-05 (1981) (examining overgrazing of a common range as an application of the AG model); Runge. supra note 39, at 155, 160-61 (discussing the development of institutions that coordinate expectations and facilitate voluntary provision of public goods). For further applications, see ELSTER, supra note 22, at 21-22; THOMAS C. SCHELLING, THE STRATEGY OF CONFLICT 83-93 (1960); and Edna Ullman-Margalit, Coordination Norms and Social Choice, 11 ERKENNTINIS 143, 145-50 (1977).
(43.) See SCHELLING, supra note 42, at 92-93, 96, 134-35; Runge, supra note 39, at 164: Ullman-Margalit, supra note 42, at 149-50.
(44.) See ELSTER, supra note 22, at 20-21. In some cases, the uncooperative outcome can be avoided if individual contributions are retrievable. In such circumstances, one does not stand to lose much from testing the possibility of cooperation, since there is no fear of being "stuck" with wasteful contributions. See Jean Hampton, Free-Rider Problems in the Production of Collective Goods, 3 ECON. & PHIL. 245,259-60, 263 (1987). In many other cases, however, contributors to the provision of public goods cannot retrieve their costs--e.g., a person who restrains her whaling operations without succeeding in avoiding whale extinction or who contributes money to the campaign of an unsuccessful candidate.
(45.) See supra notes 32-33, 36-37 and accompanying text.
(46.) Compare this preference ordering with that of a PD, in which the most preferred alternative is to free ride on the efforts of others. See supra text accompanying note 28.
(47.) Feelings of "hopelessness" can explain the choice of lower-ranked preferences in other game scenarios as well. In the Chicken Game (CG), the ordering of preferences is as follows: (I) I do not contribute, but enough others do; (II) we all contribute; (III) I contribute, although others do not; (IV) nobody contributes. Like the AG, the CG is more cooperative than the PD. In comparison to the PD, the third and fourth priorities are reversed: While a PD player prefers universal noncooperation to the "sucker's" lone contribution, a CG player prefers nonreciprocated contribution to nonprovision of the good. Yet, as with the PD, the literature focuses on the impact of greed on a CG. Much of the discussion is dedicated to the potential role greed plays in bringing about the least preferred result (Outcome IV); each CG player has an incentive to precommit herself (irrevocably, if possible) to noncooperation. By doing so she may force others into cooperation that will avoid the worst outcome and enable herself to realize the most preferred option (Outcome I). Mistakes regarding the probable choices of others, or mutual maintenance of previous commitments, may bring about the disastrous, least preferred option. See COLMAN, supra note 29, at 111-15; TAYLOR, supra note 40, at 39; Anatol Rapoport & Albert M. Chammah, The Game of Chicken, in GAME THEORY IN THE BEHAVIORAL SCIENCES 151 (Ira R. Buchler & Hugo G. Nutini eds., 1969). I argue that a desire to free ride is only one possible explanation for an observed noncooperative outcome. Although a CG player prefers lone contribution (Outcome III) to universal noncooperation (Outcome IV), this contribution will depend on her estimation of the probability of independently providing the good. She will contribute only if her individual efforts (or the combined efforts of enough cooperative others) will suffice. If she estimates that, no matter what she does, cooperative efforts will not reach the minimum level needed to provide the good, then she will not waste her efforts on a hopeless project. See HARDIN, supra note 10, at 58-59. Thus, feelings of hopelessness may play an important role in CGs as well and cause people to behave in a way that realizes their least preferred outcome. Therefore, one should be cautious before deducing preferences from observed choices.
(48.) For a discussion of circumstances under which people's other-regarding preferences may be effectuated in the private sphere, see infra notes 77-80 and accompanying text.
(49.) Furthermore, people derive utility from the very act of voting. See GREEN & SHAPIRO, supra note 11, at 51. The relatively low cost of expressing preferences in political settings has important distributive implications: Poor people may lack sufficient resources even to attempt to realize their socially-oriented preferences in market settings. In contrast, their vote carries equal weight in the political arena. See Sunstein, Endogenous Preferences, supra note 3, at 246.
(50.) To be sure, this is a simplistic and stylized account of the political process. In reality, it may not be easy to realize desired goals in the political arena. For example, problems associated with majority voting, logrolling, and rent-seeking, or the fact that a person's preferred options may not appear on the agenda or only as a bundle with additional and less desired proposals cast doubt on the feasibility of accurately transforming citizen preferences into political outcomes. Nevertheless, it is still true that the perplexing feeling of hopelessness is significantly diminished in a political setting. Therefore, such a setting can genuinely reflect the preferences that people hold but do not always reveal in private settings.
(51.) See supra Subsection II.C.2.
(52.) For a discussion of the PD explanation, see supra text accompanying notes 29-30.
(53.) Besides hopelessness, people's notions of fairness can also explain why cooperation fails in AGs played in market settings. A person may rank other-regarding, collective goals highest (Outcome I in an AG) but condition her contribution on that of others. In the absence of similar contributions by other people, she may prefer not to contribute at all (Outcome II in an AG). Even if a person believes that her lone contribution can have a positive impact on a desired goal (e.g., improving the condition of one homeless individual or saving one person from starvation), she may still believe that it is unfair for her to carry the burden alone. Consequently, if people do not trust others to do their fair share in the private sphere, only the second-best outcome will be realized. In contrast, the highest-ranking preferences can be attained in the public sphere, since there is a guarantee that if the favored policy is adopted by the government or other regulatory agencies, individuals will be required to share the financial burden among them. Cf. Sunstein, Endogenous Pre, fences, supra note 3, at 245-46 (mentioning motives of fairness, shame, and victimization in the context of a "solution of prisoners' dilemmas").
(54.) The experiments described in the text below were structured as PD situations, so they do not directly test the AG hypothesis. They show, however, that people often choose to cooperate, even when they are able to free ride. These results support the plausibility of the AG model.
(55.) See Norbert L. Kerr, Illusions of Efficacy: The Effects of Group Size on Perceived Efficacy in Social Dilemmas, 25 J. EXPERIMENTAL SOC. PSYCHOL, 287 (1989) [hereinafter Kerr, Illusions]. Studies focusing on self-efficacy show that cooperation increases when people believe their own contribution will significantly affect the outcome. See COLMAN, supra note 29, at 217; Norbert L. Kerr, Efficacy as a Causal and Moderating Variable in Social Dilemmas, in SOCIAL DILEMMAS: THEORETICAL ISSUES AND RESEARCH FINDINGS 59 (Wim B.G. Liebrand et al. eds., 1992); Norbert L. Kerr & Cynthia M. Kaufman-Gilliland, Communication, Commitment. and Cooperation in Social Dilemmas, 66 J. PERSONALITY & SOC. PSYCHOL. 513 (1994). Such studies can also support the claim that individuals are encouraged to cooperate rather than to free ride when they feel they can personally influence the prospect of successful cooperation. Since individual contributions are inconsequential with respect to many public goods, however, one should concentrate on the effect of perceived collective efficacy.
(56.) See Kerr, Illusions, supra note 55, at 303-09.
(57.) See supra notes 26-29 and accompanying text. One might even argue that the free tiding assumption requires an opposite correlation between collective efficacy and cooperation. A free rider can mason that the higher the prospects of successful cooperation, the less his own contribution is needed to attain the desired goal. Since there is a greater chance of freely benefiting from the efforts of others, the rate of cooperation should decline when collective efficacy rises.
(58.) See, e.g., COLMAN, supra note 29, at 218-21; Linnda R. Caporael et al., Selfishness Examined: Cooperation in the Absence of Egoistic Incentives, 12 BEHAV. & BRAIN SCI. 683,692-93, 696 (1989); Kerr & Kaufman-Gilliland, supra note 55, at 513-15; A.J.C. van de Kragt et al., Doing Well and Doing Good as Ways of Resolving Social Dilemmas, in EXPERIMENTAL SOCIAL DILEMMAS 177, 197-99 (Henk A.M. Wilke et al. eds., 1986).
(59.) See Robyn M. Dawes et al., Behavior, Communication, and Assumptions About Other People's Behavior in a Commons Dilemma Situation, 35 J. PERSONALITY & SOC. PSYCHOL. 1, 3-5 (1977).
(60.) See id. at 5, 9.
(61.) See id at 5-7, 10.
(62.) See COLMAN, supra note 29, at 219-21; Kerr & Kaufman-Gilliland, supra note 55, at 514, 525-27; John M. Orbell et al., Explaining Discussion-Induced Cooperation, 54 J. PERSONALITY & SOC. PSYCHOL. 811,812, 818 (1988).
(63.) See also Kerr & Kaufman-Gilliland, supra note 55, at 525 (providing additional reasons for concluding that the group identity hypothesis cannot persuasively explain the communication effect).
(64.) The absence of universal cooperation can also be explained within the AG framework: It is reasonable to assume that not everyone will be convinced by nonbinding promises of cooperation. Those having insufficient trust in others may opt for the maximin choice, which is noncooperation. For an explanation, see supra notes 41-44 and accompanying text. A different rationalization is that the uncooperative minority (or at least some subset of them) indeed prefers to free ride. The AG model, however, still applies to the large majority of participants.
(65.) See, e.g., Peter Bohm, Estimating Demand for Public Goods: An Experiment, 3 EUR. ECON. REV. 111 (1972) (demonstrating that people do not always cheat or free-ride whenever they can benefit from it, even when the opportunity for strategic behavior is understood). For additional experiments questioning the generality and validity of the strong free-riding hypothesis, see Gerald Marwell & Ruth E. Ames, Experiments on the Provision of Public Goods I: Resources, Interest, Group Size, and the Free-Rider Problem, 84 AM. J. SOC. 1335, 1349-50, 1356-59 (1979); Gerald Marwell & Ruth E. Ames, Experiments on the Provision of Public Goods II: Provision Points. Stakes, Experience, and the Free-Rider Problem, 85 AM. J. Soc. 926, 936-37 (1980); and Gerald Marwell & Ruth E. Ames, Economists Free Ride, Does Anyone Else?: Experiments on the Provision of Public Goods IV, 15 J. PUB. ECON. 295, 307-08 (1981).
(66.) In the supergame, the long-run gains from cooperation and the fear of retaliation may outweigh the short-run gains of defection (getting the most preferred outcome once, but receiving the third-best outcome thereafter). Even in the supergame, however, there is no guarantee that the desired outcome will be realized. The parties might find themselves locked in the noncooperative outcome. For discussions of the possibility of cooperation in repetitive plays of PD, see COLMAN, supra note 29, at 135-60; LUCE & RAIFFA, supra note 29, at 97-102; MCLEAN, supra note 11, at 133-9; ROBERT SUGDEN, THE ECONOMICS OF RIGHTS, CO-OPERATION AND WELFARE 107-11, 138-40 (1986); and TAYLOR, supra note 40, at 60-108.
(67.) See Sen, supra note 2, at 326, 330 (maintaining that commitment, as described supra note 8, is mostly irrelevant to choices of consumer goods).
(68.) See Kelman, supra note 12, at 93 (explaining that people "display their concerns for others" when "political decisions involve the community as a whole"); Sen, supra note 2, at 330-32 (questioning the assumption that with regard to public goods, as distinguished from private ones, the preferences people reveal aim at maximizing their personal gains); Sunstein, Social Norms, supra note 3, at 924-25,960 (arguing that citizen choices are other-regarding because, unlike consumer choices, they directly affect the collectivity and therefore are governed by social norms that discourage and sanction selfishness).
(69.) See Ronald Dworkin, DeFunis v. Sweatt, in EQUALITY AND PREFERENTIAL TREATMENT 63, 77-81 (Marshall Cohen et al. eds., 1977). On the distinction between personal and external preferences and its policy implications, see BRIAN BARRY, POLITICAL ARGUMENT 62-65, 71-72 (1965); Kurt Baier, Welfare and Preference, in RATIONAL MAN AND IRRATIONAL SOCIETY?, supra note 27, at 284, 289-91; Brian Barry, Lady Chatterley's Lover and Doctor Fischer's Bomb Party: Liberalism, Pareto Optimality, and the Problem of Objectionable Preferences, in FOUNDATIONS OF SOCIAL CHOICE THEORY, supra note 7, at 11, 35-36, 41; Brennan, supra note 7, at 196, 202, 204; and Margolis, supra note 2, at 277-78. Cf. ELSTER, supra note 22, at 81-85 (discussing the legitimacy of manipulating people's preferences by coercion, seduction, and persuasion).
(70.) Dworkin forcefully argues against consideration of people's external preferences, i.e., those concerning the "assignment of goods and opportunities to others." Dworkin, supra note 69, at 77. He claims that considering altruistic external preferences will result in "double counting" and thus will infringe upon the norm that people be treated equally. Id. at 78. This is because some people will benefit not only from the weight accorded to their own preferences but also from the weight accorded to altruistic preferences that others have concerning their well-being. Individuals holding opposing personal preferences may be unable or unwilling to develop countering external preferences to offset this influence. See id. at 78-79, 81. For criticisms of this claim, see Joseph Raz, Professor Dworkin's Theory of Rights, 26 POL. STUD. 123, 131-32 (1978). See also GRIFFIN, supra note 9, at 24-25 (rejecting the claim that other-regarding desires should not be considered). It should be noted that Dworkin's argument addressed problems in the utility calculus--i.e., the process of quantifying and aggregating individual preferences in order to maximize overall social utility. Inasmuch as the problem of "double counting" exists, it is particularly relevant to such a process and considerably less relevant to one that is based on more flexible, less technical deliberation and reasoning. In other words, Dworkin's criticism of external preferences supports a political process that accords much weight to the reasons for supporting various policies and the merits of different views. See infra Part IV. Furthermore, as explained in the text above, Dworkin's recommendation that external preferences be disregarded cannot be applied to public goods. The inherent collaborative character of these goods causes even strictly personal preferences--those regarding one's own enjoyment--to have significant external effects.
(71.) See Donald Philip Green, The Price Elasticity of Mass Preferences, 86 AM. POL. SCI. REV. 128, 129, 132, 136-39 (1992) (employing data from both verbal expressions of willingness to pay and actual payments for goods); cf. Etzioni, supra note 2, at 173, 176, 178 (offering a similar, though nonempirical, argument).
(72.) There is a tension between this finding and the claims of proponents of the consumer/citizen distinction. See supra notes 11-15 and accompanying text.
(73.) See Donald Philip Green et al., How the Scope and Method of Public Funding Affect Willingness To Pay for Public Goods, 58 PUB. OPINION Q. 49, 64-65 (1994) (questioning subjects on their readiness to fund teaching English to immigrants and saving seabirds from oil spills).74.
(74.) See id. at 54-56, 59, 61-62.
(75.) If anything, these results suggest that the taxation mode inspired people to contribute more. There appeared to be no correlation, however, between the reminder and the type of payment vehicle. See id. at 58-60, 62.
(76.) See id at 60-61, 64.
(77.) See Lewinsohn-Zamir, supra note 31.
(78.) See id at 755-57, 763-66.
(79.) See id. at 783-87.
(80.) See Runge, supra note 39, at 155-58.
(81.) For arguments supporting the necessity of examining and evaluating the reasons for public policies, see Baler, supra note 69, at 291-94; Brennan, supra note 7, at 203-05; Elster, supra note 15, at 111; and Sagoff, supra note 11, at 1412.
(82.) The most severe estimation problems are posed by goods that have a large component of "existence" value, such as environmental and wildlife protection. See supra notes 20-21 and accompanying text. Such values cannot be captured by market-revealed evaluations, and so economists have turned to personal surveys. The questionnaire technique known as the contingent valuation method (CVM) tries to elicit people's responses regarding the additional costs they are willing to incur (in donations, higher taxes or bills, and the like) in order to assure some level of the public good. These estimates are then employed in a cost-benefit analysis. CVM's ability to estimate existence values accurately is debated among economists and psychologists. For example, some studies have shown that people's stated willingness to pay is implausibly large, given their support for other public goods and their budgetary constraints. See, e.g., Natural Resource Damage Assessments Under the Oil Pollution Act of 1990, 58 Fed. Reg. 4602, 4603-04 app. 1 (1993); MITCHELL & CARSON, supra note 13, at 231-59, 295-306; William H. Desvousges et al., Measuring Natural Resource Damages with Contingent Valuation.' Tests of Validity and Reliability, in CONTINGENT VALUATION 91 (Jerry A. Hausman ed., 1993); Peter A. Diamond & Jerry A. Hausman, On Contingent Valuation Measurement of Nonuse Values, in CONTINGENT VALUATION, supra, at 3; Peter A. Diamond et al., Does Contingent Valuation Measure Preferences? Experimental Evidence, in CONTINGENT VALUATION, supra, at 41; Daniel Kahneman & Ilana Ritov, Determinants of Stated Willingness To Pay for Public Goods: A Study in the Headline Method, 9 J. RISK & UNCERTAINTY 5 (1994); Charles R. Plott, Contingent Valuation: A View of the Conference and Associated Research, in CONTINGENT VALUATION, supra, at 467.
Although it is difficult to elicit "consumer" preferences accurately through CVM surveys when existence values are involved, it is far from clear that abandoning any attempt at estimation will achieve better results. For example, in order to overcome the tendency to overestimate values in CVM surveys, it was suggested that respondents be explicitly reminded of substitute goods and budgetary constraints. See Natural Resource Damage Assessments Under the Oil Pollution Act of 1990, 58 Fed. Reg. at 4611 app. 4612-13. It may be argued that this fear is magnified in political settings that do not mention the issue of cost at all, such as elections and public opinion polls. Thus, although economic information should be gathered and utilized with caution and restraint, it seems unwise to disregard it. Consumerist estimations of people's valuation of public goods may be used ,,s a kind of "check" on citizen preferences. It is worth noting that, in the final analysis, even scholars who severely criticize the use of economic quantification tools do not recommend that they be abandoned. Rather, they advocate recognizing the limits of quantification and abandoning the search for a single, "magic" figure or an accurate numerical point estimate. See, e.g., Lisa Heinzerling, Regulatory Costs of Mythic Proportions, 107 YALE L.J. 1981, 2042-70 (1998) (criticizing the use of cost-benefit analysis in the field of risk regulation). In particular, see id. at 2042-43, 2054-56, 2068-69.
Daphna Lewinsohn-Zamir, Faculty of Law, Hebrew University of Jerusalem. LL.B., LL.D., Hebrew University of Jerusalem. The Essay was written during my stay at the Yale Law School in 1996-97. I am grateful to Robert Ellickson and Eyal Zamir for invaluable comments and suggestions and to the Yale Law School for its hospitality. The research was supported by a grant from the Israel Science Foundation.
|Printer friendly Cite/link Email Feedback|
|Publication:||Yale Law Journal|
|Date:||Nov 1, 1998|
|Previous Article:||Religion in Politics: Constitutional and Moral Perspectives.|
|Next Article:||Waiving conflicts of interest.|