Printer Friendly

Private production of public goods: liability for unrequested benefits.

This Article explores why the law treats negative externalities (harms) and positive externalities (benefits) differently. Ideally, from an economic perspective, both negative and positive externalities should be internalized by those who produce them, for with full internalization, injurers and benefactors alike would behave efficiently. In actuality, however, whereas the law requires that injurers bear the harms they create (or wrongfully create), benefactors are seldom entitled to recover for benefits they voluntarily confer on recipients without the latter's consent ("unrequested benefits").

One aim of this Article is to explore the puzzle of the law's differing treatment of negative and positive externalities and expose the reasons for this divergence. The Article's novel explanation for this phenomenon is found in the different types of obstacles possibly hindering agreement between injurers and victims, on the one side, and benefactors and recipients of benefits, on the other.

The Article also proposes a change to the law, with the adoption of an Expanded Duty of Restitution ("EDR"), under which, when certain conditions are met, recipients would compensate benefactors for unrequested benefits. This EDR would apply mainly to cases in which the benefits whose creation is mandated by efficiency are public goods. Public goods are characterized by their producer's inability to exclude others from consuming the good. Thus people tend to free ride on the producer's investment and refuse to share in the costs of producing the public good. As a result, without government intervention, many public goods whose production is efficient are currently not being created. An EDR would solve the free-riding problem in such cases and would facilitate efficient private production of public goods.
TABLE OF CONTENTS

INTRODUCTION

    I. UNREQUESTED BENEFITS UNDER PREVAILING LAW
   II. SOLVING THE HARMS-VERSUS-BENEFITS PUZZLE
       A. Entitlement Not to Be Injured Versus Entitlement
          Not to Pay for Unrequested Benefits
       B. Entitlement to Injure Versus Entitlement to Benefits
  III. INTRODUCING THE EXPANDED DUTY OF RESTITUTION
   IV. VALUATING BENEFITS
       A. The Overvaluation Problem
       B. Mitigating Overvaluation
       C. Applications
    V. OBJECTIONS
       A. Infringing on the Recipient's Autonomy
       B. The Role of the Government
          1. Information
          2. Negative Externalities
          3. Finance
       C. Enforcement Costs
       D. Other Market Mechanisms
CONCLUSION


INTRODUCTION

When people promote their own interests, they often create negative or positive effects for other people's interests, without the latter's consent. Economists refer to these effects as "negative externalities"--the harms injurers cause to victims--and "positive externalities"--the benefits benefactors confer on the recipients of those benefits (hereinafter "recipients"). Ideally, from an economic perspective, both the negative and positive effects should be internalized by those who produce them, for with full internalization, injurers and benefactors alike will behave efficiently. In actuality, however, whereas the law does require that injurers bear the harms they create (or wrongfully create), benefactors are seldom entitled to recover for benefits they voluntarily confer on recipients without the latter's consent (hereinafter "unrequested benefits").

One goal of this Article is to explore the puzzle of the law's diverging treatment of negative and positive externalities and expose the reasons for this. A second objective is to propose changing the law by adopting an Expanded Duty of Restitution ("EDR") under which, when certain conditions are met, recipients would compensate benefactors for unrequested benefits. The EDR suggested here would promote efficiency and provide incentives for private production of public goods. Whereas currently, public goods are produced principally by the government, under an EDR more public goods would be produced by private entities, to the benefit of all.

To demonstrate the puzzle and concretize the argument for adopting an EDR as proposed, consider the following illustrative example.

Example 1. The Construction. Owner contemplates constructing a park on his land. The park is expected to yield benefits to Owner and Neighbors by improving the appearance of the neighborhood and enhancing the value of Owner's and Neighbors' property. Owner's expected costs of construction exceed his expected benefits; thus without neighbors bearing at least some of the costs, Owner will not construct the park.

Let us ascribe numbers to this example. Suppose the expected costs of constructing the park are 15, while Owner's expected benefits from the park are 10 and neighbors' expected benefits are 10. Owner could try convincing Neighbors to pay him 5 or more for constructing the park; if he succeeds and the construction is carried out, a net social gain of 5 (20-15) will be produced. However, the transaction costs between Owner and Neighbors, which are typically the result of free riding in such cases, could be prohibitively high, making it implausible that the parties will reach an agreement. Specifically, each and every Neighbor could refuse to pay for the construction, knowing that he or she would be able to personally reap the benefits of the park without paying anything to Owner, thereby free riding on Owner's and other Neighbors' investments. This is a problem that characterizes the production of public goods: when excluding people from consuming goods is impossible or impractical, the production of those goods by private entities becomes impossible, or at least very hard. Thus, in this example, after failing to raise enough money from Neighbors, Owner would decide not to construct the park even though the project is cost-justified. The result would be different were Owner entitled to recover from Neighbors 5 or more, with or without their consent. The law, however, refrains (except in very limited categories of cases) from imposing such a duty of restitution for unrequested benefits. In so doing, it fails to recognize an internalization-of-benefits principle that could facilitate efficiency in numerous sets of circumstances and, in particular, enable private production of public goods.

In tort law, a problem very similar to the one illustrated by the Construction Example arises, but it is resolved in a different manner. As is well known, injurers are commonly not required by tort law to secure their victims' consent prior to the creation of risk, but once harm has occurred, they are often required to compensate them for their losses. Thus, if an injurer derives a benefit of 10 from his activity but exposes his victim to expected wrongful harm of 5, the injurer will often be entitled to continue his activity so long as he bears the resulting wrongful harm. (1) Tort law, like restitution law, allows injurers to unilaterally affect people's interests, but, unlike restitution law, combines this with an internalization-of-harms (or wrongfulharms) principle, thus facilitating efficiency.

Arguably, a uniform legal approach would allow unilateral creation of harms and benefits and mandate in both that their creators internalize them. The question that then emerges is why the law treats the two cases differently. Surprisingly, this basic and important query has garnered very little attention in legal scholarship. (2) A noneconomic approach might have no particular interest in responding to this question, with the a priori assumption that harm cases and benefit cases share very little in common. (3) An economic approach, however, should find the law's divergent treatment far more puzzling.

The Article begins in Part I with a short overview of restitution law's approach to unrequested benefits. It shows that a limited duty of restitution is applied in the law to well-defined categories of unrequested benefit cases, but fails to cover most of the situations where it has potentially efficient application.

Part II offers an innovative solution to the puzzle of the law's varying approach to harm and benefit cases. First, it reasons that given the creation of risks by injurers and the high transaction costs between injurers and victims, internalization of the resulting harms----or wrongful harms--by injurers is most crucial, as otherwise they would have no reason to restrict their injurious activities. There is no parallel concern in benefit cases. Second, it is argued that the real puzzle is then not why, when risks and benefits are created, the law treats the two cases differently. Rather, what is baffling is why, to begin with, the law allows injurers to force transactions on victims who are entitled not to be injured, by creating risks for them and then compensating them for resulting harms (or wrongful harms), yet at the same time, benefactors are not allowed to similarly force transactions on recipients by conferring benefits on them and recovering for those benefits. The explanation offered for this apparent inconsistency is that harm cases give rise to a unique problem that does not emerge in benefit cases. This problem manifests in cases where entitlements are allocated to victims. If, in such cases, injurers were not routinely allowed to create risks for victims without the latter's consent and compensate them if harm (or wrongful harm) materializes, then each and every potential victim would have veto power over the injurious activity. This would result in many beneficial activities being stymied. Benefit cases do not raise a similar risk, since no recipient has a similar power to veto beneficial activities. Of course, in many cases, that benefactors are prevented from recovering from recipients for the unrequested benefits they create for the latter leads to a free-riding problem, which results in some beneficial activities not taking place. But this free-riding concern is not of comparable seriousness to the veto-power problem inherent to harm cases.

The suggested solution to the harm-benefit puzzle implies that there is not as crucial a need for an EDR in benefit cases as a liability rule for harm cases. But this conclusion does not mean that an EDR is not crucial for promoting social welfare. On the contrary: the free-riding problem that emerges in benefit cases prevents many efficient transactions between benefactors and recipients from being made and hinders private production of public goods for the benefit of all. An EDR can solve this problem. The rest of the Article, therefore, develops an EDR and recommends its adoption by the law.

Part III introduces the proposed EDR and illustrates the typical situations in which it should apply. In all of those situations, the benefits whose creation is mandated by efficiency are public goods. In most of the cases, the benefits are either embodied in the enhancement of the value of real property or the protection of the interests of others in their bodily integrity and property. This Part demonstrates how only an EDR would facilitate the generation of those benefits. The valuation problems entailed in the implementation of the EDR are discussed in Part IV, which also contours the EDR so as to reduce to a minimum any risk of overvaluation of the unrequested benefits.

Part V then presents several possible arguments against recognizing the EDR. The first is that the duty of restitution would infringe on recipients' autonomy by obliging them to pay for benefits they never consented to purchase. A second objection could be that the production of public goods should be left to the government and not delegated to private entities through the EDR, for it is either unnecessary or, even worse, contradicts basic tenets of democracy. A third objection is that the EDR enforcement costs can be expected to be prohibitively high. A final argument would be that some of the benefits the EDR is aimed at creating are already produced, or could be produced, through alternative market mechanisms, rendering the duty of restitution superfluous. In this Part, I respond to these objections and show that careful construction of the EDR would decrease their otherwise persuasive force. As a result, none of the objections--the majority of which are applicable also in the tort context---constitutes a reason for not recognizing the proposed EDR.

In light of the possible objections to an EDR and based on the explanation provided for the law's differentiation between harm and benefit cases, the Article concludes by proposing the recognition of the EDR and sketching its outlines. Under this proposed rule, recipients would be obliged to compensate benefactors for unrequested benefits when the following six cumulative conditions are met:

1. High transaction costs preclude reaching an agreement between the benefactor and recipients for the latter to pay for the benefits conferred on them.

2. The risk of overvaluation of the benefits is not high enough to undermine the efficiency of the restitution duty or to substantially infringe on recipients' autonomy.

3. Enforcement costs do not exceed the benefit of the EDR.

4. The benefits cannot be created by market mechanisms.

5. The benefits will not be created by the government.

6. The beneficial activity does not produce significant welfare-reducing effects along with its welfare-enhancing effects.

To reduce the risk of overvaluation as well as the risk of infringing on recipients' autonomy, the measure of recovery should typically be the lower of two measures--either the indisputable benefit gained by the recipients or their relative share of the reasonable costs of producing the benefit. Several mechanisms for reducing those risks even further should also be employed in certain circumstances.

The adoption of an EDR as conceptualized in this Article would result in a substantial expansion of the categories of restitution for unrequested benefits currently recognized by the law. Whereas presently, such a duty is recognized only (or almost only) when a benefactor protected or preserved existing entitlements, the EDR proposed here applies also to instances in which new entitlements were created. Moreover, whereas presently, a duty of restitution is limited solely (or almost solely) to cases where there is a preexisting or other close relationship among the parties, the proposed EDR applies also in the absence of such a relationship. Consequently, the adoption of the EDR would render a dramatic change in the law and, more importantly, a substantial improvement in current incentives for the creation of public goods by private entities. (4)

I. UNREQUESTED BENEFITS UNDER PREVAILING LAW

When a benefactor voluntarily confers benefits at the recipient' s request, the contract between the benefactor and recipient typically regulates the rights and duties of the two parties. (5) However, when a recipient secures benefits by way of wrongful behavior on her part and those benefits are consequently nonvoluntarily conferred, the law of restitution often mandates the disgorgement of the ill-gotten benefits to the nonvoluntary benefactor. (6) In contrast, when benefits are voluntarily conferred but not at the recipient's request, the law does not impose any duty of restitution on the recipient, and she is allowed to keep the benefits at no cost to her. (7)

This rule, which this Article advocates changing, has certain exceptions. In the next paragraphs, a short overview of the main categories of those exceptional cases is presented, as well as their underlying rationales.

The first category of exceptions is rescue cases. (8) It includes all those instances where the benefactor has acted to protect the recipient's life, health, property, or other economic interest when the latter's consent could not be obtained due to the emergency nature of the circumstances. In cases of property or another economic interest, the law allows the benefactor to recover a reasonable charge for his beneficial actions. In cases of protecting life or health, the law allows such recovery only when the services granted were professional, as when a doctor provides first aid to an unconscious bystander. (9)

A second category includes cases of indemnity and contribution between co-obligors. In those cases one party has performed all or part of an obligation when he and a second party are jointly and severally liable toward a third party. (10) Based on a theory of restitution, the law allows the first party to recover from the second party in the amount of the latter's relative share of the obligation, even if he did not consent to the first party's performance on his behalf. This extends to cases in which there are no preexisting relations between the joint obligors, for example, two wrongdoers who separately cause inseparable harm for which they are jointly and severally liable toward the victim. (11)

In cases falling under a third category of exceptions, the benefactor, due to an innocent mistake (or other defect of will, such as fraud or duress), pays money to or creates a nonmonetary benefit for the recipient. (12) Under certain conditions, the benefactor is entitled to recover the benefits that are thus transferred to the recipient. (13)

A fourth category of cases encompasses those instances in which one party protects or preserves an interest he shares with another party, thereby benefiting the latter without her prior consent to pay for this benefit. (14) A common example is a co-owner of property who incurs expenses to maintain or protect it, thereby benefiting the other co-owners. (15) Generally, under a theory of restitution, the co-owner who bears the costs can recover from the others in the amount of their relative shares. (16)

A fifth and final category (17) of cases deals with common funds that are obtained through legal proceedings initiated by one party (or her attorney) but to which a group of people are entitled. (18) Under certain conditions, the initiator of the legal proceedings is entitled to collect from the other fund recipients their relative shares in the expenses he incurred in the process, even if they refused to back his efforts at the outset. (19) An illustration is the case of an heir who initiates legal proceedings and ends up increasing the estate's value, to the benefit of the other heirs as well. (20) The class-action mechanism works in a similar way and can be classified as failing under the fifth category. High rewards for a class-action initiator and his or her attorney are allowed under the law for conferring benefits on a large group of plaintiffs. (21)

From this brief review, three necessary--but not sufficient--conditions emerge for imposing a duty of restitution on a recipient for unrequested benefits: that reaching an agreement prior to the conferral of the benefit was unfeasible or impractical; that the benefactor was pursuing his own interests while the benefit to the other party was incidental (rescue cases being a major exception); and that the benefactor protected or preserved existing entitlements and did not create new ones.

There is a straightforward rationale to making the implausibility of reaching an agreement (or, in economic terms, high transaction costs) a requirement for imposing a restitution duty. Absent this condition, conferring unrequested benefits would replace consensual transactions. In practical terms, then, this condition mandates that, whenever a consensual transaction is plausible, the benefactor should not be able to take the restitution path. The rationales of the second and third conditions are less obvious. Clearly, they limit the range of cases in which a person can create a benefit for others and then charge them. The requirement that the benefactor be a recipient who is motivated by his own interests and the accrual of benefits to others incidental (the second condition) prevents the emergence of an extensive practice of sellers providing benefits through avenues other than market transactions. The requirement of protection and preservation of existing entitlements (third condition) reflects the law's preference for maintaining the status quo over a broader principle of maximizing Utility. (22) The EDR proposed in this Article incorporates the first condition but not the other two. That said, the maximum measure of recovery under the EDR--no more than the reasonable costs of producing the benefit (23)--would reduce the chances of the second condition not being met de facto.

Both the case law and commentary have raised as an additional condition for recovery of unrequested benefits the existence of a "proximity of interests," "closeness of interests," or "community of interests" between the parties (rescue cases as well as defect-of-the-will cases being the exceptions). (24) But the term "proximity of interests" and the like are ambiguous and cannot fully encompass the complexity of the issue of restitution for unrequested benefits. Indeed, were these terms to point to the existence of a contractual relationship between the parties, the rationale of this condition would have been to allow a duty of restitution as a default rule that the parties can opt out of if they wish. The EDR proposed in this Article, however, is certainly not limited to cases where a contractual relationship or other so-called proximity of interests exists. (25) On the contrary, the main appeal of the proposed duty of restitution derives from the parties' inability to regulate their relationship through contract due to high transaction costs.

II. SOLVING THE HARMS-VERSUS-BENEFITS PUZZLE

The preceding Part of the Article demonstrated that only in narrow categories of cases does the law recognize a duty of restitution for unrequested benefits. Except in such cases, benefactors are not entitled to recover for the benefits they conferred on recipients without the latter's consent. In contrast, in harm cases, injurers are often allowed to create risks for victims without the latter's consent but are then required to compensate them for the resulting harms (or wrongful harms). Thus, the law does not recognize an internalization-of-benefits principle despite the fact that it recognizes an internalization-of-harms (or wrongful-harms) principle. This differing approach to benefit and harm cases is puzzling. In both types of cases, allowing unilateral creation of benefits or harms, coupled with an internalization-of-benefits or -harms principle, would promote efficiency. Indeed, it seems unclear why the law has not adopted an EDR under which recipients should compensate benefactors for any unrequested benefits conferred on the former by the latter, thereby promoting efficiency. In this Part, I offer a solution to this harms-versus-benefits puzzle. There will be discussion of other solutions, which I believe to be unpersuasive, at a later stage in the Article. (26)

The comparison of the law's approach to harm cases with its treatment of benefit cases reveals the dual nature of the puzzle: First, why are injurers required to internalize harms or wrongful harms they created while benefactors are commonly not allowed to internalize benefits, even though full internalization in both cases is essential for efficiency? Second, why does the law, from the outset, allow injurers to force transactions on victims who are entitled not to be injured, by creating risks for them and compensating them for the resulting harms, yet, at the same time, commonly disallow benefactors to force transactions on recipients by conferring benefits on them and recovering for those benefits?

The first part of the puzzle is easy to explain: Given the creation of risks by injurers and the high transaction costs between injurers and victims, which preclude agreement between them, injurer internalization of the resulting harms--or wrongful harms--is most crucial, especially when regulating the risks is impractical. For without regulation or internalization, there would be no restriction whatsoever on injurers' harmful activities. (27) And without restriction, injurers would be completely indifferent to causing harms; even a small expected benefit would induce them to create huge losses for others. For this reason, the law only rarely allows injurers to negligently create harm without an accompanying duty of compensation. No parallel risk exists in the case of benefits. (28)

This in itself does not explain the second and more important part of the puzzle: why injurers are allowed to force transactions on victims, whereas benefactors are not. Injurers are in fact allowed to expose victims who are entitled not to be injured (or wrongfully injured) to risks and compensate them for the resulting harms (or wrongful harms). They are not required to get victims' consent prior to exposing them to risk. In Calabresi and Melamed's terms, victim entitlement not to be injured is commonly protected by a liability, rather than property, rule. (29) Protection under a property rule means that the entitlement holder has the exclusive right to choose to forego the entitlement; protection under a liability rule means that someone else can deprive the holder of his entitlement but with the accompanying obligation to compensate her for the value of the entitlement and for any other ensuing losses. (30) The law's choice of a liability rule to protect the entitlement of victims--which is so common in harm cases--amounts to authorization for injurers to force transactions on victims. (31)

But liability rules are almost never recognized for benefit cases. (32) Borrowing again from Calabresi and Melamed, a recipient's entitlement not to pay for unrequested benefits is protected by a property, rather than liability, rule. As a result, forced transactions are generally not allowed in benefit cases, and a duty of restitution is only rarely recognized. Why does the law's stance on this matter differ with regard to harm and benefit cases? The answer to this harm-versus-benefit puzzle is that even though in both types of cases, efficiency could be promoted by adopting a liability rule (or allowing forced transactions), such a rule is more essential in harm cases than in benefit cases. In particular, whereas in the former, liability rules typically solve a holdout problem, in benefit cases, it is a free-riding problem that is typically mitigated. (33) As will be argued below, efficiency is impaired far more by holdout than by free riding. (34)

To understand why, let us suppose that tort law were to not recognize liability rules. In such a world, the law would allocate to victims an entitlement not to be injured or, alternatively, an entitlement to injurers to create risks without an accompanying duty of compensation, and it would protect those entitlements with property rules. In some cases, the law would likely opt for the former option and, in others, for the latter. Let us consider now these two options and compare them with parallel options that are applicable to benefit cases.

A. Entitlement Not to Be Injured Versus Entitlement Not to Pay for Unrequested Benefits

Let us first contrast benefit cases in which beneficiaries have an entitlement not to pay for unrequested benefits with harm cases where victims have an entitlement not to be injured. In both cases, the entitlements are protected by a property rule. This means that injurers and benefactors are not allowed to force transactions on victims and recipients, respectively.

The prohibitively high transaction costs between injurers and victims, especially in cases with numerous victims, would prevent injurers from securing victims' consent to be exposed to risks of harm regardless of whether efficiency requires this. As a result, many of the modern activities of our society, like driving cars or manufacturing products, would be paralyzed, even if their benefits far exceeded their costs. (35)

Fortunately, in the real world, victims' entitlements are often protected by liability roles. Injurers are often entitled to create risks but remain liable for their victims' losses. Sometimes this liability extends to any harm that they cause and, other times, only to wrongful harms. Either way, allowing injurers to force transactions on victims by unilaterally creating risks for them and then beating all or a substantial proportion of the costs associated with those risks is essential for the occurrence of many important activities in modern society.

A liability rule is not as crucial in benefit cases as in harm cases. Indeed, society operates quite well without an EDR, and public goods are created on a regular basis. The divergence in the extent of the need for liability rules in harm and benefit cases is not just an empirical observation. The next paragraphs outline a crucial difference between two types of transaction costs: the one type existing between injurers and victims when victims are entitled not to be injured, and the other between benefactors and recipients when benefactors are not entitled to recover for unrequested benefits.

In harm cases, when victims are entitled not to be injured and this entitlement is protected by a property rule, each victim enjoys veto power over the potentially injurious activity. When a property rule is effectively enforced (36) (which, on many occasions, is a plausible scenario), this power means that every potential injurer should obtain the consent of all potential victims prior to subjecting them to risk of harm. In contrast, in benefit cases where recipients are not liable for unrequested benefits, the potential benefactor need not obtain the consent of all potential recipients prior to creating the benefit and, consequently, none of those recipients has veto power over the beneficial activity. There are two important ramifications to this difference between harm and benefit cases. The first is relevant regardless of whether victims and recipients behave strategically and is mostly applicable to cases of numerous victims and recipients; the second is relevant only when strategic behavior occurs and is applicable to all cases with more than one victim or recipient. These two ramifications fairly account for most cases in practice.

Assume an injurer with 1000 potential victims who are all entitled not to be injured and are protected by property rules. Assume also that the benefits of the injurious activity exceed its costs. Since, by definition, each potential victim has the power to veto the injurer's activity, the injurer should reach all of them and negotiate an agreement with each one. But even when victims express their true preferences and do not behave strategically, it will be impossible to reach all victims. Moreover, the injurer who starts reaching the victims can never know whether he will be able to reach all of them and knows that negotiating with "only" 999 of the 1000 would be both very costly and insufficient. Most injurers will thus abandon the negotiation attempt from the outset, and efficiency will not be achieved.

Now assume a benefactor with 1000 potential recipients who are not liable for unrequested benefits. Assume also that the benefits of the benefactor's activity exceed its costs. Here, by definition, no one enjoys any veto power. Consequently, for the beneficial activity to be worthwhile for the benefactor, it would be often more than enough for him to reach 900, 800, or perhaps even 500 of the recipients and convince them to share in the costs. Assuming recipients do not behave strategically, in many cases, the beneficial activity will take place regardless.

This is not to say that the high transaction costs of reaching the recipients and collecting payments from them could not obstruct efficient creation of benefits. Indeed, the central claim of this Article is that often such a risk does in fact exist. Nonetheless, there seems to be a more pressing need for a liability rule in harm cases than in benefit cases.

The second ramification of the difference between harm and benefit cases arises only when strategic behavior occurs. In such cases, however, the difference between harm and benefit cases is apparent even when there is only a small number of victims and recipients.

Assume an injurer with 5 potential victims who are entitled not to be injured and are protected by a property rule. Assume also that the benefits of the injurious activity exceed its costs. Since, by definition, each potential victim wields veto power over the activity, the injurer should reach all victims and negotiate an agreement with each one. Contacting all 5 victims could be relatively easy. Reaching an agreement with them, however, could be very hard due to holdouts: armed with veto power, each victim has incentive to hold out and extort the injurer. Since all victims share the same powers and incentives, the injurer would find it very difficult, sometimes impossible, to reach an agreement with them even when the benefits of his activity far exceed its costs. The greater the number of victims, the more acute this holdout problem becomes.

Now assume a benefactor with 5 potential recipients who are not liable for unrequested benefits. Assume also that the benefits of the benefactor's activity exceed its costs. Even though the recipients in this case have no veto power, each has the power and incentive to refuse to pay for the benefits and free ride on the benefactor's investment. Since all recipients have the same powers and incentives, the benefactor would not be able to reach an agreement with them even when the benefits produced by him are far greater than the costs of production. (37)

In both harm and benefit cases, victims and recipients are likely not to act according to their true preferences. In both cases, recognition of a liability rule is justified by the severe risk of strategic behavior, where the reasons for this behavior are not identical but the outcomes arguably are. Under closer scrutiny, however, this conclusion seems to be wrong. Assume, for the moment, that 4 victims and 4 recipients in each group, respectively, are motivated solely by their true preferences, while only 1 victim and 1 recipient behave strategically. Under this scenario, in the absence of liability rules, the veto power of the strategically acting victim will typically be far more detrimental than the free-riding power of the strategically acting recipient. This is because the strategic victim can veto the entire activity, whereas the recipient can refuse to pay only his share of the costs of creating the benefit.

It could be argued that this difference between the two cases is a fallacy: if only 1 victim uses his veto power, the injurer will negotiate with him and reach agreement, and the efficient outcome will materialize. But this is precisely the heart of the distinction between harm and benefit cases: the injurer will be willing to invest far greater resources in negotiating with the strategically acting victim than the benefactor will with the strategically acting recipient. The reason for this is that, in the injurer and victim context, the entire surplus of the injurious activity is at stake, whereas in the benefactor and recipient context, only the benefit that the latter can derive from the beneficial activity is at stake.

Indeed, strategic behavior is a genuine threat. If it is prevalent, the holdout and free-riding problems will lead to similar severe outcomes. In the real world, however, victims as well as recipients sometimes behave strategically and sometimes act only (or mainly) on their true preferences. (38) In this world, holdouts constitute a more detrimental problem than free riding. This leads to a corresponding variation in urgency in overcoming these two phenomena by way of a liability rule. (39)

B. Entitlement to Injure Versus Entitlement to Benefits

Harm cases do not, by nature, present a veto-power problem, nor do benefit cases inherently raise a free-riding problem. Rather these obstacles to reaching an agreement result from the current allocation of entitlements in society. The previous section discussed cases where victims have an entitlement protected by a property rule not to be injured and therefore have veto power over the injurious activity. But imagine a different allocation of entitlements, where injurers are allowed to create risks without bearing liability for resulting harms and the entitlement is protected by a property rule. If the costs of the injurious activity exceed its benefits so that stopping it would be efficient, the parties will find it difficult to reach an agreement with the injurer to that effect, not because of the victims' veto power, but due to free riding. Specifically, even if it is in the interest of all the victims that the injurious activity come to a halt, many will refuse to share in the burden of compensating the injurer for stopping the activity; they will be motivated by the desire to free ride on the investments of other victims. But, as explained above, this is "just" a free-riding problem, and exactly as in benefit cases, here, too, it is not as detrimental an impediment to reaching an agreement between the parties as posed by veto power in the hands of victims.

More interesting, however, is that the law fails to cure this free-riding problem exactly as it fails to cure the free-riding problem in benefit cases. (40) The law is consistent, therefore, in ignoring the free-riding problem in both harm and benefit cases. Consequently, in the above example it is likely that the injurer would create risks even if this were inefficient, and no potential victim would stop him. An EDR, however, as will be explained in the next Part of the Article, could remedy the problem, by allowing a victim who acted to stop the inefficient injurious activity to recover part of his costs from victims who benefited from his actions. (41)

Let us now consider the validity of the assumption that benefit cases present only a free-riding problem. A veto-power problem could arise in such cases were the law to grant potential recipients the entitlement to force benefactors to create benefits for them and were this entitlement protected by a property rule. Under such a legal regime, if the costs of the beneficial activity were to exceed its benefits so that its cessation would be efficient, the benefactor who would try to reach an agreement with the recipients to halt the activity would face the difficult hurdle of their veto power (each recipient could veto the agreement). But allocation of this type of entitlement to many recipients of one benefactor is a very rare, perhaps implausible, phenomenon in the private-law sphere.

The distinction between harm cases and benefit cases, then, can be summarized as follows. In harm cases, a veto-power problem arises when the entitlement is allocated to victims, and a free-riding problem arises when it is allocated to injurers. In benefit cases, it is only a free-riding problem that arises. The law provides a liability rule to remedy the veto-power problem but does not do so for the free-riding problem in either harm or benefit cases. This seems to make sense, since veto power is more severe an impediment to agreements than free riding.

Yet free riding can often block efficient outcomes as well. As demonstrated in the discussion above, in many situations the free-riding problem is an impediment to efficiency no less than the holdout problem. In such cases, as the next Part shows, an EDR is most essential for fostering private production of public goods currently not being produced. Indeed, the EDR proposed in this Article is not as broad as the liability rules applied in harm cases. Nevertheless, its adoption would represent a significant improvement in the law and promote social welfare in numerous cases.

III. INTRODUCING THE EXPANDED DUTY OF RESTITUTION

As shown in Part I, restitution liability for unrequested benefits is currently limited to a narrow set of well-defined categories of cases. This Article calls for recognizing a much broader duty of restitution. The basic argument made in this Part is that when high transaction costs preclude the creation of certain benefits and in the absence of difficulties of implementation, recipients of benefits, even if unrequested, should be subject to an EDR, under which they compensate benefactors for those benefits.

The cases in which the EDR would be most applicable are those that deal with the production of public goods. Public goods are characterized by the inability of their producers to exclude other people from their consumption. For this reason, people will tend to free ride on the producers' investment and refuse to share in the costs of production. (42) As a consequence, where there is no government intervention, many public goods whose production is efficient are not created. (43) An EDR could solve this free-riding problem and facilitate the efficient private production of public goods. (44) Absent implementation difficulties and based on efficiency considerations alone, the measure of recovery under the EDR should be equal to the amount of the benefits conferred on the recipients. Thus, full internalization of benefits will be secured and efficient creation of benefits promoted. But for substantive as well as practical reasons, to be discussed in the subsequent parts of the Article, limiting the EDR's scope as well as the remedy it offers to benefactors is justified. (45) In particular, recovery under the EDR would be limited to the lower of two measures: the recipient's indisputable benefit or the recipient's relative share of the reasonable costs of producing the benefit. (46)

The discussion below illustrates how free riding prevents the production of desirable benefits and how the EDR could lead to their creation. The Construction Example that opened this Article is illustrative: Some Neighbors may refuse to share Owner's costs, hoping to free ride on his and other Neighbors' investments. As a result, the construction work will not take place. (47) Given that the costs of construction are 15 and that the benefit of 20 is equally shared by Owner and Neighbors, imposing an EDR on Neighbors and making them liable for half of Owner's costs (7.5) would encourage Owner to construct the park for his and Neighbors' benefit.

Example 2 below represents another typical case in which free riding subverts the efficient creation of benefits.

Example 2. Stopping an Interference. X creates pollution to residents. A, one of the residents, pays X to cease the pollution or, alternatively, takes costly precautions to remove its cause, or, alternatively, sues X in court, where an injunction to stop the pollution is issued against X. A's successful efforts result in an increase in the market value of all the residents' houses and apartments. Should A be entitled to recover from the other residents part of her reasonable costs? (48)

Assuming A's costs exceed her benefits, she will not make the necessary effort at stopping the pollution even if this would be welfare enhancing, unless she is able to collect from the other residents at least part of her costs. Due to the difficulty in reaching all the residents and the accompanying free-riding problem, collection would be very difficult. A duty of restitution could solve the problem. (49)

A possible variation on Example 2 would be that X, the creator of the pollution, voluntarily and under no legal duty, stops or reduces it, bearing costs in so doing. Should the EDR be applied in these circumstances? If the answer is yes, risk creators will arguably be tempted to inefficiently increase risks to the maximum legally permissible level and then later reduce them and charge recipients accordingly. (50) But this concern is mitigated by the fact that benefactors' entitlements would be limited to reasonable costs rather than benefits. Given this limitation, they would gain nothing from such a maneuver. Thus, if X were to legally increase the risk by 10 and later reduce it at a cost of 1, he would recover no more than 1. As a result, he would have no incentive to increase the risk from the outset.

In contrast to Example 2, which demonstrates the protection of existing entitlements, the next Example is about creating new ones.

Example 3. Changing Zoning Plans. In a certain neighborhood, constructing more than four floors in apartment buildings is prohibited. A, the owner of a fourth-floor apartment, spends a substantial sum of money on convincing the local government to change the zoning plans and permit adding a fifth floor. As a result--in addition to the increase in market value of A's apartment--the market value of another twenty fourth-floor apartments increases by 10 percent, with no one in a worse-off position. Should A be entitled to any reimbursement for any of his costs from the owners of the other fourth-floor apartments?

As in Example 2, free riding can, in Example 3, hinder reaching an agreement between A and the other apartment owners prior to his incurrence of costs. Note the resemblance between this Example and common-fund cases, in which a fund creator is entitled to collect from the recipients their relative share in the expenses he incurred. (51) In both cases, the benefactor acts in conjunction with a public authority (the court and the local government, respectively), and his or her successful efforts yield a considerable and measurable benefit to a group of people. The difference between the two cases is that, in common-fund cases, unlike in Example 3, the benefit is monetary. But this divergence in itself (52) should not change the outcome: in both cases, applying an EDR could be a practical way of resolving the free-riding problem. (53)

Below, Example 4 illustrates cases in which a benefactor is able to act to protect her own and other people's interests in bodily integrity and property but, absent an EDR she would refrain from doing so.

Example 4. The Security Firm. A lives in a small neighborhood with thirty residential homes that is substantially threatened by crime. Posting a guard at each house is too costly. The most desirable option for both A and residents is to hire a security firm to patrol the streets at night. After failing to reach agreement among the residents on the matter, A hires a security firm. The threat to the neighborhood decreases significantly, as do the residents' insurance premiums. Should A be reimbursed by the residents for part of her costs?

While prevailing restitution law often recognizes rescuers' entitlement to recover from those they rescued in emergency circumstances, (54) it fails to solve the free riding problem that may prevent a group of people from acting in concert to protect themselves against risk of injury, as in Example 4. An EDR in the above Example would allow A to recover from the other residents their relative shares of the reasonable costs of hiring the security firm. This could solve the problem.

One caveat to this discussion is in order, however. In all four of the examples, free riding would not be a problem, and an EDR not necessary, were the recipients of the unrequested benefits significantly motivated by social or other extralegal norms to cooperate with the benefactor or among themselves. Such motivation could reduce the risk of free riding to the point where the EDR becomes superfluous. This would be likely either in close-knit communities or in the case of repeat-player recipients with a long-term relationship with the benefactor or each other. A classic example is the landowners in Shasta County in California, who shared the costs of fencing their lands according to informal norms, which are not influenced by the formal law. (55) Those landowners do not need restitution law to encourage them to build fences, even though a free-riding problem could have arisen absent the informal norms that govern their relationship. It is beyond the scope of this Article to identify those situations in which extralegal incentives would suffice for producing public goods. Indeed, in the presence of informal norms, law in general becomes less essential, as does the EDR. However, when benefits are not created, even though their value to recipients is clear and indisputable, it is evident that informal norms are not strong enough to overcome the free-riding problem, and the law should supplement them.

To sum up, whenever a free-riding problem is an impediment to private production of public goods, the EDR has the potential to overcome this.

However, the next two Parts of the Article raise some practical as well as substantive difficulties entailed in applying the EDR. These difficulties justify setting some limits on the EDR, the most important of which is restricting recovery to no more than the recipient's relative share of the reasonable costs of producing the benefit.

IV. VALUATING BENEFITS

This Part turns to the possible problem of overvaluation of the unrequested benefits that could be encountered in implementing the Expanded Duty of Restitution described in Part III. It proposes mechanisms for minimizing this risk and illustrates their application.

A. The Overvaluation Problem

The overvaluation of the benefits conferred on recipients would result in the creation of non-cost-justified benefits. The risk of this occurring would thus be a major concern if the EDR were to be adopted. (56) To illustrate this problem, let us return to the Construction Example, where Owner expects to garner a benefit of 10 from constructing a park at a cost of 15 and Neighbors expect a benefit of 10 as well. Under these circumstances, an EDR would allow Owner to recover from Neighbors more than 5 and induce him to efficiently construct the park. But assume now that the actual benefit to Neighbors is less than 5, but the court evaluates it as 10. Here, the application of an EDR would incentivize Owner to construct the park even though it is not cost justified.

This over-valuation concern intensifies if two additional factors are taken into account: the first is lost opportunities and the second liquidity. To understand the lost opportunities issue, assume that in Example 1, one of the Neighbors could have created a benefit that is a substitute for the benefit created by Owner and that that benefit could have been greater than the benefit created by Owner and even produced at lower cost. Under those conditions, the benefit created by Owner would be illusory, for without his activity, an even greater benefit would have been created. Any valuation of benefits that ignores such lost opportunities will result in overvaluation. It will also lead to inefficiencies: if Owner is allowed to recover in the amount of the benefit he creates, he will create the benefit even if someone else could do it more efficiently.

The liquidity issue also affects the risk of overvaluation. When a recipient receives a nonmonetary benefit and is required to pay for its market value, he may find himself--because of a lack of liquidity--in a position that is worse than the position he would have occupied had he not been conferred the unrequested benefit. (57) Thus, if the recipient is a property owner and the benefit she receives is an increase in her property's market value, obliging her to compensate the benefactor for this benefit might force her to take a loan, sell the land, or use other resources differently from how she would prefer. This would create a burden for her that, if ignored, would lead (58) to overvaluation of her benefits and to the inefficiencies that ensue.

But the risk of overvaluation does not pose a fatal challenge to adopting an EDR. (59) It is, however, a compelling concern that should be taken seriously when the parameters of the EDR are shaped. The next Section discusses various possible methods of mitigating the risk of overvaluation.

B. Mitigating Overvaluation

Several methods can be used to reduce the risk of overvaluation to a minimum as well as to overcome the lost opportunities and liquidity problems raised in the previous section. In this Section below, I discuss these methods and offer three special mechanisms--delaying payment and placing a lien on the enhanced property, licensing, and voting--that can reduce the overvaluation problem even further.

As in benefit cases, misvaluation is a concern that arises in harm cases. Tort law obliges injurers to compensate victims for harms, as measured by objective criteria. Consequently, damages paid by injurers are often higher or lower than the true harms incurred by victims, due to discrepancies between the value victims ascribe to their life, health, and property and the value the law assigns them. (60) Over- and undervaluation of victims' harms result in over- and undercompensation of victims, which makes injurers inefficiently create risks that are too low and too high respectively. Courts and legislatures have developed various techniques for handling the misvaluation problem. (61) Some examples of these mechanisms are placing caps on damages, limiting liability for nonpecuniary losses, and excluding liability for speculative damages.

In contrast to harm cases where the risk of over- and undervaluation is equally troubling for tort law, in benefit cases, overvaluation would be the main concern if an EDR is applied, for several reasons. First, undervaluation would result in the risk of valuation error being allocated to benefactors rather than to recipients, which is justified by distributive justice: the benefactor is the active party who has promoted his own interests and affected those of others, whereas the recipient is the passive party being affected without her consent by the benefactor. (62) Second, undervaluation of benefits would mitigate the concern of the EDR infringing on the recipient's autonomy.(63) Third, undervaluation would create an incentive for benefactors to engage in a market transaction when plausible. Indeed, the inability to create the benefits through the market is the rationale, and a precondition, for adopting a duty of restitution; (64) the EDR is a second-best solution, advocated only in cases of benefits that market mechanisms failed to produce. However, setting high damages under the EDR would create the danger of benefactors intentionally avoiding the market and then later convincing the court that a consensual transaction was not a viable option. It would also provide benefactors with disincentives to develop markets for creating the benefits in question, when such markets do not exist.

A practical way of reducing the risk of overvaluation would be to impose a duty on recipients to compensate benefactors by the lower of the following two measures: either the indisputable benefits the recipients gained from the beneficial activity, or the reasonable costs of producing those benefits. (65) Thus, in the Construction Example, the court could award recovery of 7.5 even if it were to measure the benefit at 10. In some cases, those costs should include a quantum meruit fee, which is a fee for the time and effort invested by the benefactor in creating the benefits, and even a premium for the risk the benefactor took on himself in producing the benefits. That would be essential particularly in cases where there was an ex ante risk that the benefactor would fail to produce the eventual benefits and would have to bear the costs alone. Of course, limiting recovery to reasonable costs has its own costs. (66) But as has been explained, (67) the risk of overvaluation is more troubling and therefore should be minimized.

Furthermore, in measuring the indisputable benefits and the reasonable costs of their production, lost opportunities should also be taken into account. In particular, the possibility that somebody else, or even the benefactor himself, could have produced an alternative benefit that would have been greater than that actually produced or created at lower cost should be considered. For this reason as well, the benefactor should recover no more than his reasonable--rather than actual--costs. Thus, if more than one person can produce the benefit, each of the potential producers should be certain enough before acting that he can do so in the most efficient way. (68)

Finally, when the benefit is an increase in the market value of the recipients' property, it could be appropriate to allow them to delay compensation until the benefits materialize in the form of monetary profits from the sale of the property (or otherwise). This will overcome the liquidity problem. To secure the benefactor's interest, a lien could be placed on the enhanced property; (69) a market would then likely develop in which firms purchase from benefactors their interests in the property for immediate payment.

Two mechanisms could further reduce the risk of overvaluation, although they should be resorted to only in exceptional cases. The first mechanism is licensing. This mechanism would be suitable for cases in which a substantial risk of overvaluation were to exist and the recipients were likely to be required to make a large payment for the benefits they received. Under these conditions, the law should allow an EDR only with prior government authorization. In addition to reducing the overvaluation risk, licensing could also be a plausible response to the assertion that public goods should be produced--or their production should be controlled exclusively by the government. (70) The second mechanism is voting. Conditioning the application of an EDR on recipients' advance vote in favor of production of the benefits could reduce the risk of overvaluation and mitigate the autonomy concern, discussed in Section V.A, (71) as well. One version of this mechanism would be that recipients vote unanimously in favor of paying for the unrequested benefits as a condition for applying the EDR. This version of the voting mechanism would eliminate the risk of overvaluation completely. It could be practical, however, only in cases of a relatively small number of recipients. Alternatively, a supermajority vote requirement (like three-quarters or two-thirds of recipients) or even a simple majority could be required. Under this version, since a greater proportion of the recipients would be required to obligate all recipients to pay for the unrequested benefits, the total risk of overvaluation would be reduced. The problem with the voting mechanism, however, is that it could trigger a free-riding problem, in that there would be the risk of recipients' making their vote conditional on their paying less than others to the benefactor. A secret vote could presumably mitigate this problem, but it could be objectionable on other grounds. (72)

C. Applications

To illustrate how the risk of overvaluation can be mitigated, let us return to Examples 1 through 3 (the Construction Example, Stopping an Interference, and Changing Zoning Plans). In all three, a defined group of recipients gained benefits through the benefactor's activity by an increase in the market value of their property. Such an increase can easily be verified and could set a limit to the benefactor's recovery. Another limit could be the recipients' share of the reasonable costs of producing the benefit. Both limitations would reduce the risk of overvaluation. To avoid the liquidity problem, the recipients should have the choice between paying the benefactor immediately and placing a lien on their property in his favor. Furthermore, in Example 3 (Changing Zoning Plans), a quantum meruit fee for the benefactor should be a component in the costs, as perhaps should a modest premium for the risk of failure he took upon himself. Finally, in all three examples, the possibility that someone else could have provided the same benefit or its substitute at lower cost should also be taken into account in determining the recovery amount.

Example 4 (the Security Firm) should also not raise any insurmountable valuation hurdles. If a decline in insurance premiums subsequent to the hiring of a security firm could be reasonably attributed to the firm's activity and that decline is in an equal or higher amount than the recipients' relative share of the reasonable costs of hiring the firm, there should be no difficulty recovering those costs under an EDR.

Applying the EDR in Examples 1-3 could be much trickier were there no reliable objective criteria to valuate the benefits or were the benefits more ambiguous. Thus, in Example 1 (the Construction Example), had Owner constructed the park and had the construction not impacted the market value of Neighbors' property, there would be a very high risk of their benefits being overvalued, thus justifying the nonapplication of an EDR. Indeed, intangible benefits, which are affected by idiosyncratic preferences, are by nature hard to measure. As a result, it is difficult, sometimes impossible, to define the indisputable benefits accrued to the recipients, which is a precondition for the application of the EDR. This is not to say that the EDR would never apply to intangible benefits, but only that when the benefits are intangible, the rule should be applied with great caution and awareness of the high risk of overvaluation.

In contrast, the EDR would easily apply in all three cases were all recipients wealth maximizers, such as commercial firms or other businesses. In such a circumstance, the risk of overvaluation would be minimal, since typically this type of recipient lacks any idiosyncratic preferences or values. To illustrate, were all recipients property owners who lease apartments and offices to customers and the beneficial activity were to enable them to raise the rent and earn more, evaluating their benefits would be simple.

V. OBJECTIONS

I now proceed to several potential objections to the EDR. Even though none is compelling enough to rule out recognizing the proposed EDR or to explain the puzzle of the law's diverging approaches to internalization of harms and benefits, these objections should certainly be taken into consideration when shaping the EDR and implementing it.

A. Infringing on the Recipient's Autonomy

One possible reason for not recognizing an EDR is that it would infringe on recipients' autonomy, for given such a duty, they would be required to pay for benefits they had never sought to buy. Arguably, it would violate their right "to be left alone" as well as sharply clashing with the basic principle of freedom of contract, in that it would allow benefactors to impose an exchange on recipients. (73)

A number of responses can be offered to this autonomy concern. First, the EDR would apply only in situations where paying for the benefit received is likely to be consistent with the parties' will. Thus, the EDR would be limited solely to those cases in which a market transaction is not a practical option and where there is only a very small risk that recipients would pay more than the value they ascribe to the benefit. The risk would be small because of the two caps that would be set on recovery under the EDR: the recipient's indisputable benefit and her relative share in the reasonable costs incurred in creating the benefit. (74) With these caps, and given the infeasibility of consensual transaction, applying the EDR would typically be consistent with the parties' will. (75)

This would be first and foremost evident in cases where the benefactor created a wholly monetary benefit. In Example 4, for instance, most of the benefits are monetary: the recipients' insurance premiums were reduced due to the benefactor's hiring of the security firm. It is difficult to see any infringement on the recipients' autonomy if they were to be required to pay less than the reduction in their premiums. Thus, imposing an EDR on recipients in such cases would hardly clash with their will, since after transferring part of the benefit to the benefactor, they would clearly be in a better position than they would have occupied had the benefit not been created.

Furthermore, when the recipients are firms, businesses, or other wealth maximizers, it can typically be assumed that the objective value of the benefit--even if not wholly monetary--is identical to, or at least approximates, the value the recipients actually ascribe to it. Here, too, all recipients would be better off were an EDR applied, which would be consistent with their will.

In other cases, the risk that recipients would pay more than what they would be willing to pay could be attenuated if payments were put off to a later date when the benefits are realized, as explained above. (76) Thus, in Examples 1 through 3, if recipients were to have the option to pay for the benefits after they have sold their property, they would clearly be better off given an EDR and their autonomy would hardly be infringed.

A second line of response to the fear of infringing on recipient autonomy is that in similar contexts to those in which the proposed EDR would be applied, infringements on people's autonomy are tolerated by the law. Thus, as is well known, governments create public goods on a daily basis and charge their recipients through taxes. An EDR can be construed as the privatization of the government's power to produce public goods and collect payment for them. (77) The counterargument to this would be that the authority to levy taxes is the exclusive prerogative of the government, and delegating it to private entities would infringe on basic principles of democracy. I address this argument in the next section below. (78)

Another field where severe infringements on people's autonomy are taken for granted is torts. In this field, by (de facto) allowing injurers to impose transactions on victims who are entitled not to be injured infringes on the latter's autonomy far more severely than an EDR could. Tort law enables injurers to inflict bodily and property injury on victims and obliges the injurers to compensate the victims for their wrongful losses. These losses are measured by objective criteria, and victims are often not fully compensated. The infringement of the autonomy of a pedestrian who lost his ann in a road accident due to a driver's negligence and who is compensated according to objective criteria is incomparable to an infringement of the autonomy of a recipient who is required to pay for a benefit he gained from a benefactor's activity. (79)

Third and finally, the EDR could be used in conjunction with the voting mechanism, discussed above, (80) to further mitigate the autonomy concern. As has been explained, (81) one version of the voting mechanism is to condition the application of the EDR on a unanimous vote. This version, if adopted, would eliminate any autonomy concern. More practical is a version requiring a super majority for applying the EDR. This would mitigate, albeit not eliminate, the autonomy concern for two reasons. First, the autonomy of those recipients constituting the majority that supported the application of the EDR will not be infringed. Second, if a large supermajority is required, say, of three-quarters or more of the recipients, it is probable (at least when the recipients are relatively homogenous) that the discrepancy between what the minority opposing the EDR pay and what they get in return will not be that great. Thus, even though the minority's autonomy would be infringed, the magnitude of the harm done will typically not be large.

In sum, the goal of an EDR would be to enhance people's welfare by freeing them from situations they would prefer to avoid but, due to bargaining failures, are unable to do so. Indeed, when the law allows coerced exchanges only when there is a barrier to a consensual transaction (high transaction costs) and there is a strong presumption of the recipients' will in favor of the exchange (the risk of overvaluation of benefits is very small), the autonomy concern is minimized. This is not to say that this concern disappears. Indeed, there is always the risk that some recipients will be required to pay for benefits they had never sought to buy. But as long as the EDR allows recovery for no more than the reasonable costs of producing the benefits and no more than the indisputable benefits gained by recipients, the advantages of recognizing the EDR would far exceed the risk of its adverse effect on some people's autonomy.

B. The Role of the Government

A second possible objection to adopting the EDR could be that it would lead to the delegation to individuals of powers that are, and should be, vested in the government. As already explained, the benefits the EDR is mainly aimed at generating can be characterized as public goods. (82) One of the traditional roles of government is to either produce such goods or else enable their production when the market fails to do so. (83) Therefore, it can be claimed, an EDR would be redundant. (84) Moreover, it could be argued that allowing private entities to recover from recipients for the benefits the former produced would contradict a basic tenet of democracy, that only the elected body or its authorized representatives are empowered to levy taxes on citizens in order to finance the production of public goods.

A short digression to tort law will help to understand why this argument is unpersuasive. It can be argued in the context of tort law as well that, instead of imposing liability on injurers and letting them decide if and how to create risks, the state should always regulate their behavior in a centralized way, leaving no room for free choice with respect to risk creation. (85) However, regulations are not free of flaws. On the one hand, when, under a liability regime, injurers are expected to externalize costs or benefits, regulation could work better than liability. Yet on the other hand, political constraints, in some cases, and prohibitive costs, in others, undermine the feasibility of efficient regulation and make a liability regime more attractive. (86) But more relevant for our purposes is that, very often, the individual possesses better information than the state about the costs and benefits of her behavior. If, in such situations, she also internalizes all or most of those costs and benefits, she may be better suited to bring her behavior to its optimal level. Liability, and not regulation, could be a better solution in such cases.

The same argument holds with respect to public goods. On many occasions, the government is better suited than individuals to produce the public good or finance its production, either due to superior information or because individuals may externalize costs if production is left in their hands. But this is not to say that, under certain conditions and for many types of public goods, individuals are not the better producers. In these latter situations an EDR is essential.

Lastly, the argument that an EDR, by allowing, in practice, private entities to levy taxes, is counter to basic tenets of democracy is misplaced. First, this objection falters if the "tax" is not greater than the benefit to the recipients and the EDR therefore has no redistribution effects. The major reason that in democracies the power to levy taxes is typically vested in the government is that collecting taxes is directed at redistributing wealth from one group of people to another, and only the elected body or its representatives has the legitimacy to implement such a redistribution. (87) But since the EDR does not redistribute wealth--i.e., it does not transfer wealth from one group of people to another it should not raise any democracy concerns. Second, two safeguarding mechanisms could further ameliorate this democracy concern. The first is licensing by the government and the second voting by the potential recipients. (88)

The next Sections discuss three factors that, combined, yield the "best producer of the public good." When it is individuals, rather than the government, who constitute such a best producer, an EDR is the desirable solution.

1. Information

The question of whether it is necessary to produce a certain public good can be very complex. Sometimes the government possesses more information on the matter than individuals, while at other times the reverse is true. (89) Indeed, even when individuals have greater information, they could attempt to convey this information to the government and try to convince it either to produce the public good or finance it. But the government would not necessarily be convinced. Among other things, it might suspect that individuals would prefer overproduction of public goods, especially when they do not internalize all the costs of production. Alternatively, the government could take a different stance from individuals regarding the need for a particular public good. (90) Either way, in such cases, an EDR could encourage individuals to both produce and finance the public good in question.

Thus, in Example 1 (the Construction Example), Owner could know much more than the government about whether the construction work is beneficial to him and to Neighbors; the same is true with respect to Example 2 (Stopping an Interference). In the latter example, if the interference is illegal, the government has both the information and obligation to enforce the law. But this notwithstanding, if the government fails to act, the individual who takes it on herself to step in should be able to recover her costs from the recipients of the benefit, since otherwise no one will act in the circumstances. The failure of the government to act could serve as a proof that the government doesn't have the information necessary for acting (or just lacks the willingness to do so), while the individual who stepped in has the information and therefore is qualified as the best producer of the public good. Notice, however, that an alternative solution, which also has the potential of resolving the free-riding problem in Example 2, is to let the individual who accrued the expenses recover from the government itself. Indeed, some courts have granted fee awards to plaintiffs who enforced the law through their legal actions to the benefit of others. (91) But this latter solution would not be viable if the interference is legal; in that case, the EDR alone would provide efficient incentives for stopping the interference.

2. Negative Externalities

One troubling issue related to public goods is the negative externalities that can result from their production. (92) When creating public goods could result in externalization of costs to third parties, the government, not individuals, is more suitable a producer. Thus, if constructing the park in Example 1 produces negative effects for third parties, and those effects are not internalized by the benefactor, an EDR should not be applied: it is essential to leave it to the government to weigh costs and benefits and act accordingly. (93) In general, when creating benefits externalizes significant costs to third parties for which the benefactor is not liable, applying the EDR would be unjustified because it would inefficiently induce overproduction of benefits for one group of people at the expense of another group. If, however, the benefits clearly exceed the costs and the costs are small, the application of the EDR would be justified.

A slightly modified version of Example 3 (Changing Zoning Plans) illustrates this latter point. Recall that, in this Example, permitting the addition of a fifth floor to apartment buildings creates substantial benefits for owners of fourth-floor apartments. Suppose, now, that some of the residents living in the neighborhood will bear costs due to the change in the zoning plans. Indeed, if the local government is persuaded to allow this change, it can be reasonably assumed that the benefits of the change will likely exceed the costs. Still, the true social benefit in this case will not equal the sum of the benefits garnered by all winners, since the losers' costs should be factored into the calculations. Under these circumstances, an individual who can recover from the winners but is not liable to the losers will have inefficiently excessive incentive to induce the local government to change the zoning plans and allow the construction of a fifth floor in apartment buildings.

This problem would be mitigated, however, as long as the measure of recovery was limited to the benefactor's reasonable costs in persuading the local government to change the plans. With this limitation, and given that the local government is supposed to consider both potential costs and benefits before changing the plans, the risk of excessively motivating the benefactor to act would be rather small, and recognizing an EDR would be warranted.

Externalization is not the only matter of concern with allowing individuals to recover for benefits they created when there are also negative effects. As Lee Anne Fennel has persuasively argued, allowing actors to recover for benefits alongside tort liability for harms "encourage[s] actors to gratuitously expand the reach of the risky activity capable of producing harmful as well as beneficial effects." (94) This would occur when people affected by the activity are heterogeneous: some will suffer losses and others will benefit from the same activity. Indeed, even if valuation is not a concern, benefactors should not be allowed to carry out substantial transfers of wealth from one group of people to the other. This should be left to the government to do. Note, however, that this argument has force only when the negative effects are significant. If the negative effects are minor or nonexistent, allowing recovery to benefactors for the benefits they conferred on recipients would be justified.

In sum, only when the beneficial activity does not produce significant welfare-reducing effects along with its welfare-enhancing effects should an EDR be applied. This condition for applying the EDR is of utmost importance when the welfare-reducing effects are not internalized by the benefactor; but it is recommended that this condition be kept, even if internalization takes place.

3. Finance

The definition of the best producer of the public goods also raises the question of who should finance their production. This question should be distinguished from the question of who could technically produce them better, for example, due to his or her superior information. Thus, if some individuals are in possession of better information than the government with respect to the production of a certain public good, it is still possible to argue that an EDR should not apply because the government should finance the production. In other cases, the reverse can be true: even if the government has better information and abilities than individuals to produce a certain public good, it should not necessarily finance it. Instead, either individuals should produce and finance the production of the public good or else the production should be undertaken by the government, while the recipients of the benefits bear the costs of production, all or in part.

Example 4 (the Security Firm) illustrates this latter point. Suppose that Example 4 transpires in a wealthy neighborhood in New York. The Residents ask the municipality to send police patrols to their neighborhood at night. Even though such an activity would appear to be welfare enhancing, the City of New York refuses to do so due to limited resources. In fact, due to the City's budgetary constraints, improving security in the wealthy neighborhood would come at the expense of more valuable activities elsewhere and therefore would be welfare reducing. Thus, even though the City could technically be a better producer of the public good (extra security), it is not the best financer: there is no justification to burden the City with the costs of extra security for the wealthy neighborhood. (95) One possible solution is to allow the City to provide the extra security but also to allow it to levy taxes on those who are expected to directly benefit from it. (96) But this is quite a rare practice and usually not a viable option. An alternative solution could be to let the residents produce the extra security themselves. The EDR could be a practical way of achieving this goal.

C. Enforcement Costs

Another objection that could be raised against recognizing an EDR is that if the costs of enforcing the duty will typically exceed the benefits, then the EDR will not be welfare-enhancing. (97)

Admittedly, the enforcement costs in benefit cases would pose a greater barrier to recovery than in harm cases, mainly in situations in which there are numerous benefactors and victims. This stems from the fact that the potential plaintiffs in harm cases are the victims, who possess more information than injurers about their losses, (98) whereas in benefit cases, the potential plaintiffs are the benefactors, who possess less information than the recipients about the latter's benefits. Thus, in harm cases, when there is one injurer and many victims, it is relatively easy to locate the injurer, bring a suit against him, prove the harm suffered by each victim, and subsequently collect damages. In contrast, enforcement will be a much harder task for the benefactor in benefit cases. The recipients will have good reason to hide as well as to understate their benefits, and at times, the benefactor will not be able to collect anything from them.

Another argument in support of a distinction between harm and benefit cases on enforcement-costs grounds is that the volume of litigation in the latter, were an EDR recognized, would by far exceed the litigation in harm cases under a negligence rule, even if assuming, counterfactually, that proving harms and benefits and collecting damages for them are equally difficult in both types of cases. In fact, with full enforcement of the law and absent court or injurer error, in equilibrium, under a rule of negligence, no harms are caused negligently, when harms caused nonnegligently are not recoverable. (99) Consequently, no claims are brought against anybody. In contrast, if a duty of restitution were recognized in benefit cases, the benefactors would create more and more benefits and would be repeat plaintiffs for recovery from their recipients. (100)

Let us respond to the second argument first. Implicitly assumed in its reasoning is that most claims are brought when liability exists, for in the absence thereof, there are no grounds for suing. In fact, in an ideal world, with full enforcement of the law and no court, injurer, or benefactor error, claims would never be brought in either harm or benefit cases. In such a world, there would be no disputes between parties regarding liability, but there also would be no disputes over the amount of damages to be paid by defendants to plaintiffs. In our nonideal world, however, both liability and damages are regularly controverted; it is an empirical matter and thus hard to predict whether an EDR would trigger more or less litigation than triggered by a negligence rule (not to mention a strict liability rule).

The first argument is the more compelling one. A possible response would be that so long as enforcement costs are fully internalized by benefactors, there is no reason not to allow them to decide for themselves whether or not to enforce recipients' duty of restitution, rather than rejecting such a duty from the outset. This is a weak response if a substantial proportion of the costs are borne either by the state, in providing a judicial system and other enforcement mechanisms, or by defendants when they win on the merits but are not fully compensated for their litigation costs. This concern would be serious were a major part of the benefit cases to involve minor benefits. In such cases, an EDR would be either superfluous or detrimental. It would be superfluous if the benefactor were to internalize most of the enforcement costs; then he or she would not enforce the EDR. It would be detrimental if the benefactor were to externalize most of the enforcement costs; then he or she would create the minor benefits even when it is not cost-justified to do so.

For these latter reasons, it is essential, before applying the EDR to a certain category of cases, to evaluate the expected enforcement costs of the rule and compare them to its expected benefits. In general terms, enforcement costs increase and the practical significance of a duty of restitution decreases when the benefits spread across many recipients, are nonuniformly allocated, and are of a low average value. Thus, in Example 1 (the Construction Example), an EDR would not be cost justified were there thousands of Neighbors and the benefit enjoyed by each small and varying across individuals. In contrast, when the group of recipients is defined and the benefits relatively high, with each individual's share easily verifiable, recognizing an EDR will be warranted. Examples 1 through 4 fall into the latter category.

D. Other Market Mechanisms

A fourth, final objection to an EDR could be that it would suppress the market mechanism for producing benefits of the same type at which the EDR is aimed. Indeed, at times, even when the parties cannot reach an agreement due to high transaction costs, there are still market mechanisms that could facilitate the creation of the benefit in question. In such cases-admittedly--the EDR should not be applied. (101)

Consider a variation of Example 1 (the Construction Example) in which Owner constructed a park on his land and thereby benefited Neighbors who can visit and enjoy the park. A free-riding problem arguably does not arise in this case or, alternatively, if one does, it is resolvable. For example, Owner can charge a fee for using the park and be thus reimbursed for all or part of his costs. If this is a realistic option, an EDR should not be applied. However, this can be a very costly--and inefficient--process due to the many users and many more instances of use on a daily basis. (102) Imposing a duty of restitution on Neighbors and charging them the lower amount of the indisputable increase in the market value of their property or their relative share of the reasonable costs of constructing the park could be far more practical.

Indeed, the argument of the market mechanism as an alternative should be applied with caution in any particular case. Thus, with respect to some of the examples presented in the Article, it is tempting to argue that were the recipients willing to allow the creation of the given benefit, they could organize in such a way so as to enable its production. For example, they could all live in a condominium or in a Business Improvement District, (103) which would enable the hiring of a security firm as in Example 4. Thus--the argument goes--the recipients should not be forced to pay for benefits they have chosen not to have. But this argument is flawed. Many recipients, although preferring not to live in condominiums or be involved in similar associations, will, at the same time, be willing to receive the benefits that an EDR would facilitate. An EDR would provide recipients with an ad hoc solution, which is often more suited to their needs and preferences than an institutional solution. Therefore, the choice not to live in a condominium or similar association should by no means be interpreted as reluctance to receive the benefits, and bear the costs, associated with an EDR.

In other cases, the market allows benefactors to recoup all their costs in producing the benefits, even though many recipients do not pay for the benefits they receive. Should an EDR be applied in such cases even though it is evident that the market would have incentivized the benefactor to produce the benefit regardless of whether he or she could recover under the EDR? In an ideal world absent any enforcement costs or overvaluation risk, a general principle of internalization of benefits would be warranted. This principle would provide benefactors with efficient incentives not only to take steps to produce a certain benefit but also to choose efficient levels of benefit-producing activities. (104) In the nonideal world, however, if certain benefits are expected to be created in any event, an EDR is less crucial.

To illustrate cases where benefits are expected to be created for recipients even in the absence of an EDR, consider a scenario in which Owner builds a shopping mall from which Neighbors benefit. There are several reasons why an EDR should not apply in this case. The valuation problem is one reason. (105) That Owner expects to recoup his costs from Neighbors by charging them, directly or indirectly, for shopping at the mall is another reason. Moreover, constructing the mall could also produce negative externalities (noise, air pollution, traffic congestion, and a shortage of parking), which, if not internalized, could offset some of the benefits. (106) But more importantly, even if it is reasonable to assume that some benefits are externalized to Neighbors, Owner's expected profits would often be high enough to provide him with sufficient incentive to build the mall. Indeed, if the mall will not be constructed without payments from Neighbors to Owner, a government subsidy would be a better solution than an EDR, as is often the case in the production of public goods. (107)

CONCLUSION

This Article has considered the issue of how the law should deal with positive externalities. The central claim has been that the existing categories of restitution law dealing with voluntary conferral of unrequested benefits should be replaced with a consolidated, principled approach to restitution. Indeed, the EDR proposed in this Article is much broader than the currently recognized duty of restitution. In particular, the EDR would enable recovery when benefactors create new entitlements and do not merely preserve existing ones and even when no proximity of interests exists among the parties involved. (108) Even though the EDR has the potential of fostering production of public goods that are currently not produced, the EDR would not take the place of other mechanisms for producing public goods, mainly those employed by the government. Instead, an EDR would be just one more mechanism supplementing existing ones.

Admittedly, an EDR in benefit cases is less essential from a social-welfare perspective than a liability rule in harm cases. The reason for this is that whereas in harm cases liability rules typically solve a holdout problem, in benefit cases the EDR resolves a free-riding problem. As explained, efficiency is impaired far more by the former problem than the latter. The lesser essentiality of the EDR relative to liability rules in tort law mandates caution in shaping the duty's scope and parameters. Drawing on the possible objections to the EDR presented above, this Article recommends applying the EDR to a defined set of cases when the market, the government, or the parties through consensual transactions are incapable of creating the benefit in question themselves; the risk of overvaluation of the benefits is low; the costs of enforcement are not so high as to make the application of the EDR inefficient; and the beneficial activity does not produce significant welfare-reducing effects along with its welfare-enhancing effects.

Even though theoretically possible, it would be unrealistic to apply the EDR on a case-by-case basis. Instead, it should be developed in specific categories of cases where the above conditions are typically met. The cases most suitable to the EDR would be those in which the value of the recipients' property is enhanced and where the recipients receive monetary benefits, although the duty could be applied in other contexts. When the recipients are wealth maximizers, the argument for an EDR has even greater appeal.

In order to reduce the risk of overvaluation and to alleviate the concern of infringing on recipients' autonomy, recovery under the EDR should be limited to either the recipient's indisputable benefit or the recipient's relative share in the reasonable costs of producing the benefit, whichever is lower. Delaying payment would also sometimes be necessary to reduce the burden on the recipients. Finally, under certain circumstances, licensing and voting mechanisms could be used as a precondition for recognizing an EDR. Since the EDR would allow benefactors to recover from recipients no more than their relative share of the reasonable costs of producing the benefits--which are determined by taking into account alternative ways of production--only benefactors with a self-interest in creating the particular benefits, as well as the ability to produce them efficiently, would have incentive to create them in the first place. (109)

In his seminal book The Death of Contract, published in 1974, Professor Grant Gilmore predicted that "we are fast approaching the point where, to prevent unjust enrichment, many benefits received by a defendant must be paid for unless it was clearly meant as a gift." (110) I do not think that we are close or should be close to that point. It is my claim, rather, that we are too far away from this point.

(1.) This is true under both strict-liability and negligence rules. Note that under a negligence rule, the injurer is expected not to cause harm negligently, but the choice of whether to do so or not is ultimately his.

(2.) But see Saul Levmore, Explaining Restitution, 71 VA. L. REV. 65 (1985) (justifying the law's different approaches to harm and benefit cases); Ariel Porat, Offsetting Risks, 106 MICH. L. REV. 243 (2007) (discussing tort liability when the wrongful act reduces risks ex ante either to the victim or to third paries, thereby creating positive externalities).

(3.) See Daniel Friedmann, Unjust Enrichment, Pursuance of Self-Interest, and the Limits of Free Riding, 36 LOY. L.A.L. REV. 831 (2003) (arguing that the negative aspect of freedom of contract entails that the recipient of unsolicited benefits be under no duty to pay for them); Scott Hershovitz, Two Models of Tort (and Takings), 92 VA. L. REV. 1147, 1160 (2006) (arguing that "there is an underlying moral asymmetry between harms and benefits" and that "it is perfectly intelligible that the institution which redresses the harms that we inflict on one another is more robust than the institution which allows recapture of the benefits that we confer on one another").

(4.) Extending the duty of restitution as proposed here can be expected to trigger the development of at least six rules for handling benefit cases, although only the second and third rules are comprehensively discussed in this Article. These rules would be as follows: (1) A benefactor is free to create a given benefit but is not entitled to any recovery from the recipients. (2) A benefactor is free to create a given benefit and is entitled to recovery from the recipients in the amount of their relative share of his reasonable costs. (3) A benefactor is free to create a given benefit and is entitled to recovery from the recipients in the amount of their benefit. (4) Recipients are entitled to the creation of a given benefit and are not liable toward the benefactor. (5) Recipients are entitled to the creation of a given benefit, but are liable in the amount of their relative share of his reasonable costs. (6) Recipients are entitled to the creation of a given benefit, but are liable in the amount of the benefit.

(5.) But sometimes the contract is unenforceable and restitution law prevails. See, e.g., RESTATEMENT (THIRD) OF RESTITUTION & UNJUST ENRICHMENT [section][section] 31-37 (Tentative Draft No. 3, 2004) (describing when restitution law will be applied to unenforceable contracts); 2 GEORGE E. PALMER, THE LAW OE RESTITUTION [section] 6.1 (1978) (same).

(6.) See RESTATEMENT (THIRD) OF RESTITUTION & UNJUST ENRICHMENT [section][section] 2, 13--14 (Tentative Draft No. 1, 2000) (discussing when a duty of disgorgement will apply); 1 PALMER, supra note 5, ch. 2 (same).

(7.) See RESTATEMENT (THIRD) OF RESTITUTION & UNJUST ENRICHMENT [section] 23 (Tentative Draft No. 2, 2002) (referring to unrequested-benefit cases as cases of "[s]elf-[i]nterested [i]ntervention").

(8.) See id. [section][section] 20-21; 2 PALMER, supra note 5, [section] 10.4; see also Hanoch Dagan, In Defense of the Good Samaritan, 97 MICH. L. REV. 1152 (1999) (analyzing rescue cases and supporting a broad duty of restitution).

(9.) E.g., Cotnam v. Wisdom, 104 S.W. 164, 166 (Ark. 1907) (awarding damages to doctor who performed emergency surgery on unconscious, injured passerby).

(10.) See RESTATEMENT (THIRD) OF RESTITUTION & UNJUST ENRICHMENT [section] 25 (Tentative Draft No. 2, 2002); HANOCH DAGAN, THE LAW AND ETHICS OF RESTITUTION 126--27 (2004); 2 PALMER, supra note 5, [section] 10.6; Friedmann, supra note 3, at 852-54; Levmore, supra note 2, at 100. Similar to this second category of cases are instances of equitable subrogation, where one party performs an obligation toward a third party, thereby discharging a second party from performing his or her separate obligation toward the same third party. See Ford v. United States, 88 F. Supp. 263, 263 (Ct. C1. 1950) (holding that U.S. military authorities that compensated a victim of a crime committed by a U.S. soldier were entitled to recover the amount from the soldier's confiscated money); RESTATEMENT (THIRD) OF RESTITUTION & UNJUST ENRICHMENT [section] 26 (Tentative Draft No. 2, 2002); 1 DAN B. DOBBS, LAW OF REMEDIES" DAMAGES, EQUITY, RESTITUTION [section]4.3 (2d ed. 1993).

(11.) See DAN B. DOBBS, THE LAW OF TORTS [section] 386--87 (2000).

(12.) See RESTATEMENT (THIRD) OF RESTITUTION & UNJUST ENRICHMENT [section][section] 9--10 (Tentative Draft No. 1, 2000); 2 DOBBS, supra note 10, [section][section] 11.1--.10; 2 PALMER, supra note 5, [section][section] 11.1-.6. It could be argued that, strictly speaking, the benefits in this category of cases cannot be classified as having been voluntarily conferred, since there was a defect in the benefactor's volition. RESTATEMENT (THIRD) RESTITUTION & UNJUST ENRICHMENT [section] 23 cmt. a (Tentative Draft No. 2, 2002) ("The limitation [of the rule denying liability for unrequested benefits] to intentional acts excludes transfers that are subject to avoidance on the grounds of fraud, mistake duress, or similar invalidating cause.... ").

(13.) See, e.g., Challenge Air Transp., Inc. v. Transportes Aereos Nacionales, S.A., 520 So.2d 323, 324-25 (Fla. Dist. Ct. App. 1988) (suggesting that an airline carrying passengers with tickets issued by another airline, when mistakenly believing a reimbursement agreement to exist between the two, may recover from the issuing airline under certain conditions).

(14.) RESTATEMENT (THIRD) OF RESTITUTION & UNJUST ENRICHMENT [section] 24 (Tentative Draft No. 2, 2002); see also DAGAN, supra note 10, at 125-26; Levmore, supra note 2, at 100-01.

(15.) 2 PALMER, supra note 5, [section] 10.7(c); Friedmann, supra note 3, at 855-58.

(16.) See, e.g., United Carolina Bank v. Caroprop, Ltd., 446 S.E.2d 415, 416-17 (S.C. 1994) (holding that when one cotenant stops paying his share of taxes and mortgage payments, other cotenants may pay his share and recover from him).

(17.) Another category of cases, which I will mention only briefly, is instances in which a benefactor, without any duty on his or her part to do so, performs the recipient's duty without the latter's consent, to the benefit of a third party or to promote a social interest. RESTATEMENT (THIRD) OF RESTITUTION & UNJUST ENRICHMENT [section] 22 (Tentative Draft No. 2, 2002).

(18.) See John P. Dawson, Lawyers and Involuntary Clients in Public Interest Litigation, 88 HARV. L. REV. 849 (1975); John P. Dawson, Lawyers and Involuntary Clients: Attorney Fees From Funds, 87 HARV. L. REV. 1597 (1974); Levmore, supra note 2, at 95-99.

(19.) Section 23(b) of the Third Restatement of Restitution and Unjust Enrichment (Tentative Draft No. 2, 2002), allows recovery in cases where "the benefit is a money payment" thereby substantially broadening the common-funds category of cases.

(20.) For examples of suits brought by an heir against his or her coheirs, see the Third Restatement of Restitution and Unjust Enrichment [section] 30, illustrations 23-25 (Tentative Draft No. 3, 2004), and 2 PALMER, supra note 5, [section] 10.7. See also Feick v. Fleener, 653 F.2d 69, 75-79 (2d Cir. 1981) (holding that heirs who hired a lawyer whose representation was successful and led to an increase in the amount they received could not obtain restitution from other beneficiaries of the lawyer's actions); Friedmann, supra note 3, at 858-61 (discussing cases of co-heirs when an indemnity claim was allowed and cases in which it was denied). Moreover, some courts have granted fee awards to plaintiffs who enforced the law through their legal actions to the benefit of others, thereby helping parties to overcome a free-riding problem. See infra note 91 and accompanying text. Lastly, allowing derivative actions by shareholders and obliging the firm to cover the derivative plaintiff's litigation costs also can mitigate a free-riding problem. Jonathan R. Macey & Geoffrey P. Miller, The Plaintiffs' Attorney's Role in Class Action and Derivath, e Litigation: Economic Analysis and Recommendations for Reform, 58 U. CHI. L. REV. 1, 19-20 (1991) ("Members of the plaintiff class in a large class action or shareholder's derivative suit often have claims so small that the litigation is a matter of relative unimportance to them. Even though the claims in the aggregate may be very large, the small size of the individual claims creates enormous free-rider effects...."); see also Mills v. Elec. Auto-Lite Co., 396 U.S. 375, 395 (1970) (stating that because shareholder-derivative suits are brought on behalf of the corporation, courts reason that the corporation should pay for any benefit it receives as a result of the suit).

(21.) Phillips Petroleum Co. v. Shutts, 472 U.S. 797, 809 (1985) ("Class actions ... may permit the plaintiffs to pool claims which would be uneconomical to litigate individually.").

(22.) It is not always clear whether the third condition focuses on protection and preservation of the benefactor's entitlement or that of the recipient. Interestingly, in most of the categories of cases, the benefactor's acts clearly protected and preserved both. See 2 PALMER, supra note 5, [section] 10.1 (discussing two categories of cases: those in which the person seeking restitution acted primarily to protect the defendant's interests and those in which he sought primarily to protect his own interests).

(23.) See infra Part III.

(24.) Friedmann, supra note 3, at 856-58; see also RESTATEMENT (THIRD) RESTITUTION & UNJUST ENRICHMENT [section] 24 cmt. a (Tentative Draft No. 2, 2002) ("Situations in which one person's payment to protect an interest has the consequences of conferring a benefit on another are those in which the parties share interests.... In either settings the interests of the parties have already been linked.... The claimant must have acted to protect an interest in property, and there must be a nexus of interests.... ").

(25.) See infra Part II.

(26.) See infra Part V.

(27.) ROBERT COOTER & THOMAS ULEN, LAW AND ECONOMICS 340 (5th ed. 2007) ("[N]o liability gives the injurer no incentive to take precaution.").

(28.) See supra Part I.

(29.) A third possibility, which will not be elaborated on here, is protection by an inalienability rule, under which the entitlement is nontransferable. See Guido Calabresi & A. Douglas Melamed, Property Rules, Liability Rules, and Inalienability: One View of the Cathedral, 85 HARV. L. REV. 1089, 1111-15 (1972) (discussing the various ways in which the law protects entitlements).

(30.) See id. at 1106-10 (distinguishing property rules from liability roles).

(31.) An alternative classification of the rule of negligence, and perhaps of other rules governing accident cases, is as a compensation, rather than liability, rule. Under this approach, the limitations of the protection offered to potential victims of an accident stem from practical enforcement difficulties and do not reflect permission to injurers to create risks and bear the costs of resulting harms. Specifically, since, in most such cases, it is virtually impossible to enjoin an injurer ex ante from negligently creating risks, the best the law can do is impose an ex post duty of compensation when negligence is the cause of the harm. For arguments in this vein, see Jules Coleman & Jody Kraus, Rethinking the Theory of Legal Rights, 95 YALE L.J. 1335 (1986) (arguing that a negligence rule is not a liability rule); Benjamin C. Zipursky, Rights, Wrongs, and Recourse in the Law of Torts, 51 VAND. L. REV. 1, 55-70 (1998) (same). See also Mark A. Geistfeld, Punitive Damages, Retribution, and Due Process, 81 So. CAL. L. REV. 263 (2008) (arguing that risk creation will respect the victim's rights only if the injurer expects to compensate the victim in the amount prescribed by their hypothetical agreement at the time of the risk creation). But for our purposes, it is immaterial: de facto, even if not de jure, injurers are allowed to impose negligent risks on victims, thereby forcing transactions on them.

(32.) See supra Part I.

(33.) Notice, however, that sometimes the opposite is true. See discussion infra Section II.B.

(34.) Lloyd Cohen, Holdouts and Free Riders, 20 J. LEGAL STUD. 351 (1991) (comparing holdout and free riding and indicating that sometimes free riding is a more severe obstacle to efficiency than holdout).

(35.) Indeed, one could argue that even under a property role, when that rule is not effectively enforced, injurers would sometimes proceed with their injurious activities and compensate victims for the resulting harms (or wrongful harms). But this argument does not undermine the assertion implied in the text, that it is hard to imagine the modern world without a rule that in practice allows the creation of risks accompanied by a duty of compensation (whether conditioned on the injurer's wrongdoing or not).

(36.) See supra note 35.

(37.) Example 1 (The Construction), supra text accompanying note 1, illustrates just such a scenario: Owner would fail to collect payments from Neighbors when each one of the latter expects to free ride on Owner's and other Neighbors' investments toward constructing the park.

(38.) See ROBERT NOZICK, THE NATURE OF RATIONALITY 50-55 (1993) (arguing that rational actors would cooperate and not free fide); Ernst Fehr & Urs Fischbacher, Why Social Preferences Matter--The Impact of Non-selfish Motives on Competition, Cooperation and Incentives, 112 ECON. J. 478 (2002) (arguing that identification with the group can also prevent free riding); David M. Kreps et al., Rational Cooperation in the Finitely Repeated Prisoners' Dilemma, 27 J. ECON. THEORY 245 (1982) (arguing that repeat players will tend not to free ride); Daphna Lewinsohn-Zamir, Consumer Preferences, Citizen Preferences, and the Provision of Public Goods, 108 YALE L.J. 377 (1998) (arguing that the main reason for people's reluctance to take part in financing public goods is not their desire to free fide but, rather, their belief that others will do so); Thomas J. Miceli & Kathleen Segerson, A Bargaining Model of Holdouts and Takings, 9 AM. L. & ECON. REV. 160 (2007) (arguing that allowing developers to exercise eminent-domain powers only after negotiations with owners resolves the holdout problem); Gary J. Miller, The Impact of Economics on Contemporary Political Science, 35 J. ECON. LIT. 1173, 1179-83 (1997) (describing experiments that show that the free-riding problem does not always preclude cooperation).

(39.) It seems that a liability rule does not solve the holdout problem but, rather, merely changes the identity of the person creating the problem. Thus, in instances of one injurer and one victim and where the victim has an entitlement, protected by a property rule, not to be injured, the victim has the power to hold out (by refusing to permit the injurious activity unless he or she is well paid). If, instead, the victim's entitlement is protected by a liability rule, the injurer has the power to hold out (by refusing to stop the injurious activity unless well paid, which could happen when damages are not fully compensatory). But the situations discussed in this Section are different. I have assumed all along that, in a typical harm case, there is one injurer and many victims. In such cases, when the victim's entitlement not to be injured is protected by a property rule, the holdout problem is severe because there are many actors, not just one, who can hold out. Therefore, the move from property-rule protection for the victim's entitlement to liability-rule protection for that entitlement also represents a move from a situation in which many victims can hold out to one in which many victims can free ride. The latter move is a desirable one since the holdout problem is more severe than the free-riding one. A similar analysis applies to benefit cases.

(40.) In allocating the entitlement to injurers, the law could offer a liability rule under which victims could stop injurers from harming them but would have to compensate them for the resulting costs. Such a liability rule was proposed by Calabresi & Melamed, supra note 29, at 1115-24, which they entitled "Rule four." But in our Example, such a liability rule would not be of much assistance, since even if injurers were to be required to halt their activities upon victim demand and be compensated, free riding would still be an obstacle to agreement among potential victims. Specifically, each victim would have an incentive to free ride on other victims' investments, refusing to share with them the costs of compensating the injurer for halting his activity.

(41.) Example 2 (Stopping an Interference), infra text accompanying note 48, illustrates such cases.

(42.) See RESTATEMENT (THIRD) OF RESTITUTION & UNJUST ENRICHMENT [section] 23 cmt. b (Tentative Draft No. 2, 2002) (explaining that the protection commonly afforded to property fights and contractual liberty (by denying restitution) comes at an important cost: an invitation to behave strategically and free fide).

(43.) See Friedmann, supra note 3, at 846 (noting free tiding as an obstacle to efficient creation of benefits).

(44.) DAGAN, supra note 10, at 130-39 (arguing that the rationale for allowing recovery for unrequested benefits under prevailing restitution law is to avoid free riding); see also Consol. Edison Co. of N.Y.v. Bodman, 445 F.3d 438, 442 (D.C. Cir. 2006) C[C]ourts have found it sensible to apply the unjust enrichment principle ... because doing so answers a potential free-rider problem.").

(45.) Infra Parts IV-V.

(46.) Infra Section IV.B.

(47.) Cf. Green Tree Estates, Inc. v. Furstenberg, 124 N.W.2d 90 (Wis. 1963) (denying restitution to a developer who made improvements for its own benefit and for the benefit of local residents).

(48) Cf. Ulmer v. Farnsworth, 15 A. 65 (Me. 1888) (denying restitution to owner of a flooded quarry who drained it to his own benefit and that of a neighboring quarry owner).

(49.) If the interference in this Example is illegal, the question will be whether A should recover from the recipients or from the enforcement agency that failed to stop the illegal interference, if at all. See discussion infra Section V.B. 1.

(50.) In the original version of Example 2, as well, one could imagine collusion between A and X, where X would increase risks and, subsequently, in line with A's demands and payment, agree to stop creating or to reduce those risks.

(51.) See supra text accompanying notes 17-21 (fifth category of cases).

(52.) But the overvaluation problem can make a difference here. See infra Section IV.C.

(53.) Cf. Sprague v. Ticonic Nat'l Bank, 307 U.S. 161 (1939) (awarding plaintiff reasonable attorney's fees and litigation expenses, when, by virtue of stare decisis, fourteen other claimants, who were not parties to the litigation, could recover from the defendant).

(54.) See supra text accompanying notes 8-9 (first category of cases).

(55.) ROBERT C. ELLICKSON, ORDER WITHOUT LAW: HOW NEIGHBORS SETTLE DISPUTES 65-76 (1991) (arguing that landowners in Shasta County are influenced by informal norms regulating the fencing of their lands).

(56.) Levmore, supra note 2, at 69-72 (claiming that the law may be seen as normally disallowing restitution claims because of valuation difficulties); see also DAGAN, supra note 10, at 139-48 (discussing the recipient's subjective devaluation of the conferred benefit).

(57.) See DAGAN, supra note 10, at 141 (describing the duty of restitution of unrequested benefits as an obligation to exchange money for nonmonetary values without an opportunity to refuse the exchange). For another articulation, see Levmore, supra note 2, at 74-79, who detaches the misvaluation argument from another argument, according to which even if the recipient is required to pay no more than the value of the benefit, because people's decisions to spend money depend on their wealth, a duty of restitution might force some recipients to spend their money in a way that deviates from their preferences. Levmore calls this latter argument the "Wealth Dependency" argument. Id.

(58.) The liquidity issue could also pose an autonomy concern. See infra text accompanying note 73. On the possible ways to solve this concern, see infra text accompanying note 74-81.

(59.) Valuation difficulties could be a barrier to a restitution rule when consensual transactions are a plausible option. Thus, in a recent (unpublished) paper, Bar-Gill & Bebchuk show why a market operating under a restitution rule, where sellers provide goods and services without buyers' consent but are entitled to recover the value of the benefits, will probably not survive. Oren Bar-Gill & Lucian Arye Bebchuk, Consent and Exchange (Harvard Univ. Law & Econ. Research Paper Series, Discussion Paper No. 590, 2007), available at http://ssrn.com/abstract=998486. According to their thesis, when courts are imperfectly informed about the value of goods and services, a restitution rule will induce excessive entry of low-quality sellers and excessive exit of low-valuation buyers. Court adjustment of the value estimate upward to reflect the exit of low-valuation buyers will induce the exit of more buyers, and the market could thus completely unravel. But Bar-Gill & Bebchuk's argument is limited by its own terms to situations in which transaction costs are low and consensual exchange is possible, Id. at 2-4. Therefore, their argument does not apply to the cases for which the EDR presented here is designed, where transaction costs are high and consensual exchange is not an option.

(60.) Calabresi & Melamed, supra note 29, at 1108 (discussing inaccuracies in awarding damages).

(61.) But then there is the risk of undervaluation arising from efforts to prevent overvaluation, and vice versa.

(62.) A reverse argument can be made for harm cases, according to which preference of overvaluation over undervaluation is justified in tort law. See Stephen R. Perry, The Moral Foundations of Tort Law, 77 IOWA L. REV. 449, 467 (1992) (arguing that distributive considerations favor the victim over the wrongdoer). A possible response to this argument would be that in torts, the victim often takes an active role in creating the situation that resulted in his or her injury, whereas the typical recipient in the cases discussed in this Article is almost always a passive party.

(63.) See infra Section V.A.

(64.) See infra Section V.D.

(65.) See RESTATEMENT (THIRD) OF RESTITUTION & UNJUST ENRICHMENT [section] 24 cmt. e (Tentative Draft No. 2, 2002) (explaining when the measure of recovery should be reasonable costs); 2 PALMER, supra note 5, [section] 10.7 ("In some cases the value of the benefit has been measured by... the cost of the improvement").

(66.) Thus if Owner in Example l can choose between two uses of his land, and one is more beneficial to him than the other but less beneficial to Neighbors, under a rule that allows him to recover reasonable costs only, he would choose the first use even if the total benefit to him and Neighbors from the second use would be greater. A rule allowing Owner to recover at least some of Neighbors' benefits could prevent this inefficient outcome. Nevertheless, I am not advocating allowing recovery for more than the reasonable costs of producing the benefits, since the overvaluation problem is severe enough to rule out any recovery beyond that level.

(67.) See supra text accompanying notes 62-64.

(68.) Alternatively, when there are many potential benefactors, a licensing mechanism can be used. See infra text accompanying note 70. In class actions, a parallel question arises when there are many candidates for the class-action plaintiff and only one can be chosen. See JOHN J. COUND ET AL., CIVIL PROCEDURE CASES AND MATERIALS 707--08 (6th ed. 1993) (describing a situation in which several members of a class want to be affirmed as class representatives).

(69.) 1 PALMER, supra note 5, [section] 1.5(a) (explaining how equitable lien is used to protect plaintiff rights). The Arkansas Supreme Court reasoned similarly:
   It is well settled that a tenant in common has the right to make
   improvements on the land without the consent of his cotenants; and,
   although he has no lien on the land for the value of his
   improvements, he will be indemnified for them, in a proceeding in
   equity to partition the land between himself and cotenants, either
   by having the part upon which the improvements are located allotted
   to him or by having compensation for them, if thrown into the
   common mass.


Graham v. Inlow, 790 S.W.2d 428, 430 (Ark. 1990). See also In re Mach, 25 N.W.2d 881 (S.D. 1947) (discussing an equitable lien in circumstances of performance of another's duty); Bennett v. Bennett, 36 So. 452, 453 (Miss. 1904) (holding that a tenant in common is entitled to a lien on the shares of his cotenants in the land for taxes paid by him beyond his proportionate share and for any sum due him for improvements or rent from his cotenants); RESTATEMENT (THIRD) OF RESTITUTION & UNJUST ENmCHMENT [section] 24 cmt. e (Tentative Draft No. 2, 2002) ("Where a claimant has made expenditures to protect an interest in common property, the basic requirement that a liability in restitution not prejudice an innocent defendant ([section] 49) is frequently observed by limiting the remedy in restitution to subrogation or equitable lien.").

(70.) See infra note 88 and accompanying text.

(71.) See infra Section V.A.

(72.) One possible objection is that a secret vote would allow cheating. For example, a recipient could ask for a side payment from other recipients as a condition for voting in favor of the application of the EDR, but then in fact vote against it. Indeed, aware of this possibility in advance, recipients might condition the side payment on the outcome of the voting rather than on the particular vote cast by said recipient. Then again, the latter might get paid his side payment even if he votes against the EDR's application but the majority decides in favor of its application. In sum, the main problem with a secret vote is the lack of transparency--i.e., despite its advantages, it could also have costs. Furthermore, if even with a secret vote side payments become effective, the free-riding problem could reappear and the difference between open and secret vote will narrow.

(73.) See RESTATEMENT (THIRD) OF RESTITUTION & UNJUST ENRICHMENT [section] 23 cmt. b (Tentative Draft No. 2, 2002); Friedmann, supra note 3, at 846-47.

(74.) See supra Section IV.B.

(75.) As more people are expected to behave irrationally, this argument weakens.

(76.) See supra Section IV.B.

(77.) See infra Section V.B.

(78.) See infra Section V.B.

(79.) A counterargument could be that since the need for a liability rule for harm cases is greater than for an EDR in benefit cases, the justification for infringing on people's autonomy is stronger in the former cases than in the latter. See supra Part II.

(80.) See supra text accompanying notes 71-72.

(81.) See supra text accompanying notes 71-72.

82. Pure public goods are characterized by the inability to exclude people from consuming them ("non-excludability") and by the inability of one person's consumption to detract from or prevent another person's consumption ("nonrivalry"). See JOSEPH E. STIGLITZ, ECONOMICS OF THE PUBLIC SECTOR 128-29 (3d ed. 2000).

83. See id. at 129 (explaining how non-excludability and nonrivalry result in market failure); Paul A. Samuelson, Aspects of Public Expenditure Theories, 40 REV. ECON. & STAT. 332 (1958) (discussing market provision of public goods); Paul A. Samuelson, Diagrammatic Exposition of a Theory of Public Expenditure, 37 REv. ECON. & STAT. 350 (1955) (discussing the market's ability to provide public goods); Paul A. Samuelson, The Pure Theory of Public Expenditure, 36 REV. ECON. & STAT. 387 (1954) (discussing the need to focus on optimal public expenditure).

84. It seems that this is a generally accepted argument among the reporters of the Third Restatement. See RESTATEMENT (THIRD) OF RESTITUTION & UNJUST ENRICHMENT [section] 23 cmt. b. (Tentative Draft No. 2, 2002).

(85). See, e.g., GUIDO CALABRESI, THE COSTS OF ACCIDENTS: A LEGAL AND ECONOMIC ANALYSIS 95-129 (1970) (discussing specific deterrence, where the state, through its agents, regulates safety).

(86.) See id. at 68-129 (discussing the pros and cons of market deterrence and specific deterrence as a mean to reduce accident costs); RICHARD A. POSNER, ECONOMIC ANALYSIS OF LAW [section] 6.1 (7th ed. 2007) (discussing the advantages of private enforcement in torts and contracts over public enforcement in criminal law).

(87.) J.A. Mirrlees, An Exploration M the Theory of Optimum Income Taxation, 38 REV. ECON. STUD. 175 (1971) (discussing principles and methods of optimum income taxation for redistributing wealth).

(88.) See supra Section IV.B.

(89.) See GARETH D. MYLES, PUBLIC ECONOMICS 311 (1995) ("One aspect of public goods that prevents the government making efficient decisions is the government's lack of knowledge of households' preferences and willingness to pay for public goods.").

(90.) There could be different views on the necessary conditions for justifying the production of public goods by the government. For a thoughtful discussion, see Barak Medina, "Economic Constitution," Privatization and Public Finance: A Framework for Judicial Review of Economic Policy, in ZAMIR BOOK ON LAW, SOCIETY AND POLITICS 583 (Yoav Dotan & Ariel Bendor eds., 2005) (arguing that the public-good theory is not value free, but rather depends on the normative considerations underlying the government's goals) (translation by author).

(91.) See Scott J. Jordan, Awarding Attorney's Fee to Environmental Plaintiffs Under a Private Attorney General Theory, 14 B.C. ENVTL. AFF. L. REV. 287 (1987) (discussing those cases and supporting them). Class actions are another mechanism by which the problem can be solved.

(92.) See JOHN G. HEAD, PUBLIC GOODS AND PUBLIC WELFARE 184-210 (1974) (discussing externalities and their relevance to the theory of public goods).

(93.) Cf. Abraham Bell & Gideon Parchomovsky, Givings, 111 YALE L.J. 547 (2001) (arguing that in takings, the public authority will act efficiently only if it internalizes both the negative and positive externalities of the takings).

(94.) Lee Anne Fennel, Property and Half-Torts, 116 YALE L.J. 1400, 1450-52(2007)(arguing against recovery of positive externalities conferred on landowners, when either they or the uses to which they may put their property are heterogeneous).

(95.) It has been argued that there should be less willingness on the part of the government to produce public goods that are consumed by high-income earners. This argument was raised with respect to activities for the preservation of the environment. See Richard J. Lazarus, Pursuing "Environmental Justice": The Distributional Effects of Environmental Protection, 87 Nw. U. L. REV. 787 (1992).

(96.) A similar practice is the creation of a "business improvement district" ("BID"). A BID is a public-private partnership in which property and business owners of a defined area elect to make a collective contribution to the maintenance, development, and promotion of their commercial district. BIDs require legislative authorization from the local government. They typically provide services such as street and sidewalk maintenance, public-safety officers, park and open-space maintenance, marketing, capital improvements, and various development projects. The services provided by BIDs supplement the services already provided by the local government. BIDs are funded through special assessments collected from the property owners in the defined boundaries of the district. Like a property tax, the assessment is levied on the property owners who can, if the property lease allows, pass it on to their tenants. For further details, see the website of The Los Angeles Downtown Center Business Improvement District, http://www.downtownla.com (last visited May 14, 2009), and Downtown DC BID, http://www.downtowndc.org (last visited May 14, 2009).

(97.) Cf. Donald Wittman, Liability for Harm or Restitution for Benefit?, 13 J. LEGAL STUD. 57 (1984) (arguing that the choice between encouraging actors to create benefits by sanctions and by subsidies should depend to a great extent on the litigation costs entailed in each method).

(98.) STEVEN SHAVELL, FOUNDATIONS OF ECONOMIC ANALYSIS OF LAW 578-81 (2004) (arguing that the reason the entitlement to collect damages for harm done is allocated to victims and not the state is that the former have better information about their harms).

(99.) Injurers will not behave negligently, because if they do, their expected liability will be higher than the precautions they could have taken to avoid liability. This is a simple application of the Hand Rule. See POSNER, supra note 86, [section] 7.1 (describing and explaining the Rule).

(100.) This argument is inspired by a different argument made by Giuseppe Dari-Mattiacci, Negative Liability (George Mason Univ. Law & Econ. Research Paper No. 03-29, 2008), available at http://ssrn.com/abstract=422961, as part of an endeavor to explain why a duty of restitution, which he calls "negative liability" is so rare. Dari-Mattiacci argues that under a rule of negligence, it is sufficient for the law to have one sanction at its disposal, since in equilibrium, there is no negligence and the sanction is never implemented. In contrast, in benefit cases, the "sanction," or, more accurately, the subsidy, should be implemented again and again, whenever a benefit is created by one person for another person.

(101.) Occasionally, legislatures create institutions that facilitate market solutions to free riding. A good example is the patent registration afforded inventors under intellectual property law. The ability to collect a fee from users of their inventions through the market without competition for a limited period of time protects patent holders from free riders, resulting in sufficient incentives for inventors to invent. See POSNER, supra note 86, [section] 3.3.

(102.) Note also that after the park has been built, the investment in its construction is a sunk and fixed cost. In a perfectly competitive market, the price-competing parks charge would not reflect the fixed costs of construction; in fact they would be zero, assuming the marginal cost of giving each additional person access to the park is negligible. Competition, however, is seldom expected to be perfect. In our Example specifically, since there are no identical substitutes for A's park, he would probably have the ability to charge Neighbors a positive price even if all of his costs are fixed and sunk. See, e.g., JEAN TIROLE, THE THEORY OF INDUSTRIAL ORGANIZATION 280 (1988) (showing formally how, in the case of competition between a small number of firms that do not provide perfect substitutes, prices remain above the cost of supplying the marginal unit).

(103.) See supra note 96.

(104.) Cf STIGLITZ, supra note 82, at 131 (explaining how a benefactor who expects to gain a large benefit from a public good will produce it, but not in the efficient quantity).

(105.) See supra Part IV.

(106.) See supra Section V.B.2.

(107.) A subsidy could work better than an EDR, because the valuation problem is especially severe here: recipients are very heterogeneous (everyone enjoys the mall in a different way); it is hard to define the class of recipients (should it include only people living next to the mall or encompass a broader vicinity?); and in measuring the net benefits received by recipients, the negative externalities, which are difficult to measure, should also be factored in.

(108.) See supra note 24 and accompanying text.

(109.) And therefore, the currently prevailing condition under restitution law, that the benefactor was pursuing his own interests while the benefit to the other party was incidental, will typically be met.

(110.) GRANT GILMORE, THE DEATH OF CONTRACT 88 (1974). I thank Ofer Grosskopf for this reference.

Ariel Porat, Alain Poher Professor of Law, Tel Aviv University, Fischel-Neil Distinguished Visiting Professor of Law, University of Chicago. For very helpful comments on a previous draft of the Article, I thank Oren Bar-Gill, Hanoch Dagan, Arye Edrei, Melvin Eisenberg, David Enoch, Jill Fisch, David Gilo, Suzanne Goldberg, Zohar Goshen, Ofer Grosskopf, Ehud Guttel, Michael Heller, Adam Hofri-Winogradow, Avery Katz, Gregory Keating, Roy Kreitner, Yoram Margalioth, Barak Medina, Gideon Parchomovsky, Ronen Perry, Timna Porat, Eric Posner, Joseph Raz, Robert Scott, Uzi Segal, James Spindler, Lior Strahilevitz, Avraham Tabbach, Doron Teichman, Shay Wozner, Omri Yadlin, and Benjamin Zipursky. My thanks also to the participants at the 2008 American Law and Economics Association Annual Meeting, the faculty workshops at the University of Chicago, Columbia University, the Hebrew University, the University of Southern California, San Diego University, and Tel Aviv University, as well as the Tort Group at NYU Law School. I am particularly grateful to Saul Levmore and Stephen Sugarman for their very insightful comments. I also thank Arik Brenneisen for his very able research assistance and Dana Rothman-Meshulam for her invaluable language editing. Lastly, I thank the Michigan Law Review Editorial Board, and especially Katie Ladewski and Sarah Tremont, for their excellent editorial assistance.
COPYRIGHT 2009 Michigan Law Review Association
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2009 Gale, Cengage Learning. All rights reserved.

 
Article Details
Printer friendly Cite/link Email Feedback
Author:Porat, Ariel
Publication:Michigan Law Review
Date:Nov 1, 2009
Words:21046
Previous Article:Free speech federalism.
Next Article:A sea of confusion: the Shipowner's Limitation of Liability Act as an independent basis for admiralty jurisdiction.
Topics:

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters