Legal entitlements as auctions: property rules, liability rules, and beyond.

• Author: Ayres, Ian

This Essay concerns the different ways that policymakers can protect legal entitlements. The notion of a "legal entitlement" is an expansive one, encompassing such diverse rights as the right to bodily security, the right to a pollution-free atmosphere, the right to build a house that blocks another's view, or the right to damage another's reputation by false accusation.(1) Twenty-five years ago Guido Calabresi and Douglas Melamed distinguished between property rules and liability rules as techniques for protecting entitlements.(2) Property rules discouraged nonconsensual takings. Liability rules permitted nonconsensual takings in return for payment of damages.(3)

Recent articles have reconceptualized the distinction between property rules and liability rules in terms of options.(4) A liability rule gives at least one party an option to take an entitlement nonconsensually and pay the entitlement owner some exercise price. Thus, if the right against pollution is protected by a liability rule, a polluter may pollute if she is willing to pay damages.(5)

From this perspective, the only difference between liability and property rules is the price of exercising the option - the damages to be paid for the nonconsensual taking.(6) Property rules set the exercise price so high that no one is likely to exercise the option to take nonconsensually, while the lower exercise prices of liability rules presuppose that some people will take nonconsensually.(7)

The option analysis deconstructs the original distinction between property rules and liability rules. Whereas Calabresi and Melamed assumed that property rules involved consensual agreements and liability rules involved nonconsensual takings, the options analysis shows that both property and liability rules involve options for nonconsensual taking. In other words, property rules are actually a special case of liability rules: Property rules are liability rules with an exercise price so high that the option is almost never taken.

In this Essay, we will argue that the analysis of property and liability rules in terms of options does not go far enough. Both property and liability rules are special cases of an even larger family of possible legal regimes: auctions. A deeper and more valuable way to think of legal entitlements is as species of auctions of differing lengths and with differing rules for the distribution of proceeds. From this perspective, both property rules and liability rules are truncated auctions of legal entitlements.

Why change our focus from property and liability rules to auctions? Calabresi and Melamed were not merely interested in offering a novel nomenclature, and neither are we. They argued that when transaction costs were high - because parties lacked information or bargaining was impractical - liability rules in the form of damage awards were a more efficient way to protect entitlements.(8) Conversely, they concluded (incorrectly, as we shall argue) that property rules were a more efficient way to protect entitlements when transaction costs were low.(9) Thus, Calabresi and Melamed's distinction offered an important gloss on the Coasean point that imperfect information and other transaction costs might prevent the efficient allocation of resources.(10) Calabresi and Melamed showed that under these conditions not only the ownership but also the form of protection of the entitlement could affect allocational efficiency.

Although this basic insight remains correct, the actual arguments for the relative efficiency of property and liability rules now seem more problematic. It is by no means clear that property rules are always more efficient when bargaining is possible. The problem is that when parties have private information about their valuations of an entitlement, they face a classic case of asymmetric information. Ian Ayres and Eric Talley, as well as Louis Kaplow and Steven Shavell, have shown that where information is asymmetric, property rules do not necessarily produce the most efficient result, even when transaction costs are otherwise low.(11)

We shall show in this Essay that efficiency sometimes might be further enhanced by allowing additional rounds of "bidding" - that is, by allowing both sides successive options to take the entitlement back at successively higher prices. Just as ordinary liability rules sometimes dominate property rules, "higher-order" liability rules that feature successive and reciprocal options to take can sometimes dominate ordinary liability and property rules.

Our argument can be understood as an extension of a basic caveat to the Coase Theorem. Coase argued that regardless of the initial allocation of entitlements, efficient deals would be struck under ideal bargaining conditions, which include full information.(12) But many transactions in the real world occur under conditions of asymmetric information, where each party knows only her own private valuation of the bargained-for entitlement. Asymmetric

information will often prevent efficient negotiation - even when there are only two people bargaining. We have found that higher-order liability rules might be able to produce nonconsensual transfers that are more efficient than those produced by consensual trade under a property or a first-order liability rule.

In this sense, our result is also an extension of the basic insight of Calabresi and Melamed that the form of entitlement protection matters as well as its ownership. They saw that the choice between property rules and liability rules could make a difference where transaction costs were high. We shall now argue that one must sometimes go beyond property and liability rules to enhance efficiency, when bargaining in the shadow of these more traditional regimes fails to capture all the gains from trade.

Part I of this Essay explains how one can view liability rules with reciprocal taking options as forming a class of "internal" auctions - auctions where the proceeds are distributed among the bidders rather than to a third party. Ordinary or first-order liability rules feature only one round of possible takings; allowing successive and reciprocal taking options creates second- and higher-order liability rules. Part II discusses the relative efficiency of secondand higher-order liability rules under a relatively strict set of assumptions. Using a standard model featuring asymmetric information between two parties, we show that second- and higher-order liability rules are more efficient than first-order liability rules when the taking regime is costlessly administered, and when bargaining is not possible. Part III relaxes the first assumption by considering the relative efficiency of different rules when nonconsensual taking is not costless. Part IV relaxes the second assumption by considering how the possibility of bargaining affects the relative efficiency of different legal regimes. Part V applies the theory to some real world examples, with particular emphasis on contract negotiations.

1. Entitlements as Internal Auctions

   Viewing entitlements as auctions implies that after one party exercises its option to take nonconsensually, the other has an option to "take back," and so on, for some number of rounds. However, almost all analyses of liability rules have implicitly assumed that the law deters the initial entitlement holder from taking back after an initial nonconsensual taking. For example, if a liability rule regime gives Calabresi an option to take some entitlement of Melamed for $100, most analysts assume that after this taking, Melamed (and others) would not have a viable option to take the entitlement back from Calabresi. In other words, most people have assumed that liability rules are protected by property rules.

   Kaplow and Shavell are among the few scholars who have recognized and attempted to defend this assumption.(13) They offer two arguments against retaking entitlements - one of impossibility, the other of inefficiency. First, they point out that when an entitlement concerns a harmful externality like pollution, it is effectively impossible to retake the entitlement once it has been taken the first time. A polluted stream, for example, cannot easily be returned to its pristine state. Second, they argue that giving the original owner a takeback option might lead to an infinite sequence of takings and retakings if the exercise price for the take-back option (i.e., the damages assessed at each round) is set too low. If an object is worth at least $100 to both Calabresi and Melamed, but the price of taking and subsequently taking back is only $75, then giving take-back options could lead to an infinite number of takings: "Such reciprocal takings are problematic because they will lead inevitably to destructive contests to retain or to take control of things, and teus to the use of force."(14) Hence, "[tlhe only apparent solution to the problem of reciprocal takings lies in a mixed system that would employ a liability rule for the initial taking combined with property rule protection of the taker's possessory right afterwards."(15)

   As to the first objection, it is by no means clear that most retakings of legal entitlements are impossible. A chattel like the Maltese Falcon or a parcel of real estate like Blackacre could, in theory, be taken and retaken many times.(16) Kaplow and Shavell are correct that some takings, like the taking of an entitlement to a pollution-free river, are not easily undone after the fact." But this objection only applies to harm that has already occurred, and not to additional harm that might happen in the future. In nuisance cases, for example, it is hardly unusual for courts to award damages for existing pollution combined with an injunction against future harm. As we shall see presently, there is no reason why a court could not offer to dissolve the injunction upon payment of an option price to the plaintiffs, followed by an offer to reinstate it upon payment...