The Efficiency of Bargaining under Divided Entitlements
Ever since Professor Guido Calabresi and A. Douglas Melamed published their seminal article Property Rules, Liability Rules, and Inalienability: One View of the Cathedral, the question of how to best define property rights has been a central question in the legal literature. Calabresi and Melamed highlighted the fact that, if bargaining is impossible, a liability rule (in which an individual can take another’s entitlement by paying damages) can make the allocation of the entitlement depend on the taker’s privately-known valuation. Thus, they argued, while simple property rights are preferable when bargaining is possible, liability rules can serve a market-mimicking role in cases in which it is not. Indeed, by setting damages equal to the plaintiff’s expected value for the entitlement, the defendant would be induced to take precisely when his valuation was greater than this expectation, thus enhancing efficiency over what could be achieved with a simple property right held by the plaintiff.
Subsequently, scholars beginning with Professors Ian Ayres and Eric Talley and Professors Louis Kaplow and Steven Shavell began to question the presumption that simple, undivided property rights are always best when bargaining is possible. Noting that when bargaining is perfect, the Coase Theorem indicates that all forms of entitlements achieve a first-best outcome, they asked whether certain forms of entitlements such as liability rules might perform better than simple property rights when bargaining is possible but inefficient due to asymmetric information. As Kaplow and Shavell suggested, liability rules might be expected to perform better because of their “head start,” the fact that they outperform property rules when bargaining is impossible. Ayres and Talley argued that, moreover, liability rules provide an “information-forcing” benefit by encouraging victims to reveal their harm truthfully. To date, however, the literature has considered this question using very particular examples and bargaining procedures.
In this Article, we address the “horse race” between property rules and liability rules by taking a mechanism-design approach. In particular, we ask what the best form of entitlements is given that bargaining always takes a form that is “optimal” given the informational constraints that are present.
Through most of our analysis, we do so by focusing on a benchmark measure of bargaining efficiency, the subsidy that would be required for the parties to achieve an efficient outcome given the informational constraints they face. Under most conditions of asymmetric information, bargaining will fall short of achieving efficiency, regardless of the bargaining procedure parties follow. The dual pressures of what economists call “incentive-compatibility” (bargainers’ inclination to understate or overstate their true benefit or harm when bargaining) and “individual rationality” (the fact that an injurer or victim can always refuse any proposed bargain and revert to his or her entitlement under the law) prevent always reaching an efficient outcome. Our benchmark measure is the subsidy that is needed to relax the individual-rationality constraints to a sufficient degree to permit efficiency if the best possible bargaining procedure were to be followed.
As a general matter, liability rules with unrestricted choices of damages must be weakly better than simple, undivided property rules—after all, a simple property rule can be viewed as a special case of a liability rule in which damages are infinitely large, or at least large enough to deter any taking. Thus, for there to be a horse race between property rules and liability rules, we must either be comparing general, unrestricted forms of property rules (in other words, allowing for what Ayres and Talley called “fractional property entitlements”) or be considering a restricted class of liability rules (for instance, liability rules with damages equal to the expected harm). As noted by both Ayres and Talley and Kaplow and Shavell, fractional property entitlements are sometimes physically possible—an asset may be owned in partnership, or a firm may have the right to pollute up to some specified limit—and can also be created by legal uncertainty over who will be determined to be the proper holder of the entitlement.