Distinguishing between consensual and nonconsensual advantages of liability rules.

• Author: Ayres, Ian
• Position: Response to article by Louis Kaplow and Steven Shavell in this issue, p. 221

Louis Kaplow and Steven Shavell's thoughtful reply to our recent article contains powerful insights about the relative efficiency of liability and property rules.(1) While we are in agreement that liability rules can be more efficient than property rules when transaction costs are low, we disagree about the cause of this liability-rule advantage. Kaplow and Shavell believe that liability rules hold only a nonconsensual advantage over property rules (i.e., liability rules tend to induce efficient nonconsensual takings). While granting this oft-recognized nonconsensual advantage, we contend that liability rules may also have a consensual advantage in low-transaction-cost settings (i.e., liability rules facilitate trade). We use this Comment as a forum to articulate our side of the story.

Our answer consists of two parts. In Part 1, we locate the current debate within the broader context of entitlement form, transaction costs, and bargaining. In Part II, we provide an example that distinguishes between the consensual and nonconsensual advantages of liability rules.

1. LOCATING THE DEBATE

   Before plunging into the technical arguments, it may first be helpful to contextualize our initial contribution and its relation to the debate between ourselves and Kaplow and Shavell. In Solomonic Bargaining, we ventured into what we termed "the economic purgatory between the findings of Coase and of Calabresi and Melamed."(2) This phrase bears elaboration.

   Calabresi and Melamed argued persuasively that, when transaction costs make consensual transfer prohibitively expensive, liability rules are likely to ominate property rules because liability rules more closely replicate the outcome that transaction costs preclude.(3) When transacting is costless, Coase tells us that liability and property rules are equally efficient.'

   In the last twenty-some years, numerous law-and-economics scholars have attempted to connect the dots between these efficiency benchmarks characterized by Coase, on the one hand, and Calabresi and Melamed, on the other. In this so-called "economic purgatory," transaction costs are positive but not prohibitive.(5) Richard Posner has characteristically offered a lucid and powerful theory by arguing that, in settings with "low" transaction costs, property rules are more efficient than liability rules because they channel transactions into the market.(6)

   Solomonic Bargaining challenged this view in the context of a specific type of transaction cost: private information.(7) our core insight was that dividing an entitlement between two negotiators could cause more forthright and efficient bargaining.(8) Essentially, we argued that these so-called Solomonic divisions tend to obscure the titular boundary between "buyer" and "seller" during bargaining and, in so doing, dampen the parties' strategic incentives to "shade up" or "shade down" their privately known valuations. We argued that numerous entitlement divisions can cause this phenomenon, including divisions along use, space, and time dimensions; divided ownership through judicial randomness (i.e., "probabilistic" divisions); and divided ownership through untailored liability rules.(9)

   It is with our extended analysis of this last form of division that Kaplow and Shavell find their sole quarrel. They agree with us that liability rules can be more efficient than property rules.(10) They part company with us, however, in interpreting the cause of this postbargaining advantage. Kaplow and Shavell argue that our analysis of liability rules failed to distinguish between the consensual and nonconsensual attributes of liability rules, and they conjecture that any advantage liability rules possess stems not from bargaining, but rather from the persistence of the Calabresi-Melamed "head start" that liability rules enjoy in the absence of bargaining.(11) More explicitly, they argue that liability rules have a nonconsensual advantage over property rules when trade is not possible, and that this nonconsensual advantage tends to persist once bargaining is allowed. While Solomonic Bargaining focused on a consensual advantage of liability rules (i.e., that liability rules could facilitate trade), Kaplow and Shavell attribute the efficiency of liability rules in our example to the persistence of a nonconsensual advantage (i.e., that liability rules tend to induce taking by higher valuers). While we agree with Kaplow and Shavell that liability rules enjoy a "head start" in this horse race (before bargaining occurs), we disagree with their conjecture that this head start alone preordains the outcome of the race (after bargaining). Instead, we maintain that bargaining concerns can play an important (and possibly first-order) role in the normative choice between liability and property rules.

   A diagram representing the competing views may help flesh out the debate more fully. Figure I heuristically depicts the relative efficiency of property rules and liability rules as transaction costs vary between zero and infinity.(12) The origin of the diagram represents the Coase Theorem benchmark of zero transaction Costs.(13) Here, property and liability rules are indistinguishable from an efficiency standpoint--neither produces any inefficiency. The right side of the diagram corresponds to the other polar extreme, in which transaction costs are infinitely high. Here, there is no bargaining, and Calabresi and Melamed's nonconsensual "head start" of liability rules over property rules frequently prevails.(14)

   Between these two extremes, we can use Figure I to distinguish between the competing characterizations of relative efficiency outlined above, which, for convenience, we associate with the names of Posner, Kaplow and Shavell, and ourselves. For the purposes of exposition, we posit a fixed, linear relation between the size of transaction costs and the inefficiency under a property rule(15)--and show what the three theories say about the relative shape of the liability-rule curve.

   The Posnerian thesis is depicted by the "Liability Rule (Posner)" curve. The strictly concave shape of this curve is intended to depict the asserted dominance of property rules in low-friction settings and the eventual point where liability rules overtake property rules as transaction costs grow.(16) Kaplow and Shavell's argument is depicted by the linear "Liability Rule (Kaplow & Shavell)" schedule. The linearity of this curve illustrates their "persistence conjecture" in that the relative efficiency at the endpoints of the diagram will tend to dictate the position of the curves in the interior.(17) Finally, the possibility that we raised in Solomonic Bargaining--that liability rules might facilitate more efficient bargaining--appears as the "Liability Rule, (Ayres & Talley)" schedule. The convex shape of this curve is intended to depict our assertion that a postbargaining advantage of liability rules might reflect more than the simple preservation of the head start, because liability rules can also work to lubricate consensual transactions between the parties.(18)

   It is important to emphasize that these three conceptions of relative efficiency (depicted in Figure 1) also make different claims about the likelihood that a particular advantage will obtain. The Posnerian conception makes the strongest claim: When transaction costs are low, property rules will always be more efficient than liability rules.(19) Kaplow and Shavell make a more modest claim, that liability rules "tend to be" more efficient.(20) And, on this likelihood dimension, our thesis is the most modest. We only claim that liability rules might facilitate trade as indicated by "Liability Rule."(21) The goal of Solomonic Bargaining was to provide a "possibility theorem." By showing that liability rules could produce a more efficient equilibrium with more contracting, we refuted the strong-form Posnerian claim that property rules must be more efficient than liability rules when transaction costs are low.

   To be sure, we believe that property rules are often efficient, but a primary goal of Solomonic Bargaining was to show that the efficiency of property rules could not be justified by the traditional claims that they are "market encouraging"; liability rules can channel even more transactions into the market and produce a more efficient equilibrium. We developed an example that contradicted the Posnerian thesis in that an untailored liability rule produced larger gains from trade and more contracting than a property rule.22 This example is illustrated in Figure I by points P and L. The distance between P and the horizontal axis represents the amount of inefficiency that remains after bargaining under a property rule, and the distance between L and this axis represents the analogous amount under a liability rule. Even with the transaction costs in our model (i.e., asymmetric information), we found parties contracting in the shadow of either rule; and after contracting, liability rules were more efficient than property rules (i.e., Point L is below Point P).(23)

   Kaplow and Shavell's criticism of this core example can also be described using Figure 1. They agree that we have produced an example that refutes the Posnerian claim that property rules are more efficient when transaction costs are low, but they point out that the efficiency of the liability rule in our example may be due to the persistence of the rule's nonconsensual advantage and not to its ability to facilitate trade (a theoretical difference depicted by the fact that the "Liability Rule" schedule lies between point L and our "Liability Rule" schedule).

   We concede that Kaplow and Shavell are right that our original example did not adequately distinguish between the consensual and nonconsensual advantages of liability rules.(24) While the example does succeed in refuting the Posnerian hypothesis, our article's focus on what happens as one moves away from the point of Coasean...