Economic Rationales for Mediation. 80 Virginia Law Review 323 (1994) (with Jennifer Brown)
Return to Professor Ayres' Homepage
Return to Professor Ayres' Resume

INTRODUCTION

Much of the economic literature on Alternative Dispute Resolution ("ADR") displays a surprising failure to differentiate between types of dispute resolution devices. Often when economists purport to examine the efficiency of ADR, they focus exclusively on arbitration or other forms of private adjudication. Mediation-negotiation facilitated by a neutral third party- has received far less attention than arbitration in the economic literature.

The neglect of mediation is particularly surprising because mediation is more "alternative" than arbitration. Arbitration (in both its binding and nonbinding forms) asks the arbitrator to replicate the decision of a court.

A mediator, by contrast, stops short of recommending how the dispute should be resolved. By focusing on arbitration when they examine ADR, economists have failed to provide a coherent rationale for mediation.

The ADR literature has succeeded in fleshing out noneconomic explanations for mediation. Robert Mnookin and Lee Ross, for example, have suggested several ways in which a mediator might overcome psychological barriers to conflict resolution, but they have not explained how a mediator might overcome the barriers created by the strategic interaction of two rational, self- interested negotiators. We agree that mediators are valuable in helping parties overcome a broad variety of psychological barriers. The goal of this Article, however, is to identify how mediation also could increase the efficiency of bargaining from an economic or strategic perspective.

This Article explores how mediators (acting specifically as mediators, rather than as adjudicators or advisors) can create value in negotiations between rational actors. Mediators in practice provide a number of services. Many of these services, however, could be performed by advocates, advisors or neutral third-party adjudicators. For example, commentators often note that an important function of the mediator is to serve as a "reality check" or "agent of reality" who impresses upon the parties the costs of failing to reach a negotiated agreement. But such a function could be performed equally well by a nonbinding arbitrator working with public information (and performing a reality check in the presence of both parties) or by an expert advisor working with private information (and delivering the bad news privately). We believe that mediators create value by providing reality checks and other services, but that these services are not uniquely mediative and thus do not adequately explain mediation.

We focus instead on how mediators resolve disputes by "caucusing" privately with the individual disputants. Sequential caucusing is uniquely mediative: arbitrators generally do not conduct ex parte meetings with the parties and negotiation necessarily requires the parties to meet together. By shuttling back and forth between meetings with individual disputants, mediators can collect and distribute private information. An economic rationale for mediation centers on caucusing because it is here that the mediator most clearly controls the flow of information between the disputants. Although our title refers to "economic rationales for mediation," probably a more illuminating description would be "economic rationales for mediative caucusing."

ADR theory, however, has been particularly unhelpful in explaining how caucusing creates value. Although psychological explanations are offered for caucusing, no one has ventured an economic or strategic justification for the practice. Why, for instance, would rational disputants disclose private information to a mediator that they would not disclose directly to the other side? And what should mediators do with this information? Mediators and analysts often claim that the mediator should maintain the confidentiality of all private communications learned in the caucus. If confidentiality is scrupulously maintained, however, the mediator is doing nothing more than what an advisor individually retained by each party could do. This extreme form of "Chinese wall" caucusing thus seems not only a poor example of mediation, but also an activity that could be more cheaply accomplished by independent advisors concerned only with one side of the dispute.

In practice, however, mediators often do make disclosures of private information, if only indirectly. By caucusing, for example, the mediator might "determine whether a zone of agreement exists" or "figure out a set of trades that will bring the participants as close to agreement as they can possibly get." Revealing that there are gains from trade or that a particular set of trades might be acceptable to the other side has the effect of indirectly disclosing to each party some of the mediator's private discussions with the other side. The ADR literature has often failed to acknowledge the tension between the mediator's duty of confidentiality and the need for indirect disclosure. Indeed, generally when ADR scholars have written about confidentiality, the focus has been on whether the mediation will be privileged from discovery by outsiders, rather than on the extent to which one party's caucus information can be disclosed to the other side.

Knowing, however, how a mediator is allowed to transmit caucus information is not only important in assessing the ethics of a particular mediator's behavior, but also determines whether private parties will disclose information and whether that disclosure will be truthful.

This Article's thesis is that mediators can create value by controlling the flow of private information (variously eliminating, translating, or even creating it) to mitigate adverse selection and moral hazard. Adverse selection is caused by hidden information that distorts the terms of a contract; because of adverse selection, for example, unhealthy people are more likely than healthy people to opt for life insurance. Moral hazard is caused by hidden conduct; because of moral hazard, insured people are more likely than uninsured people to take risks. Thus, adverse selection problems involve hidden precontractual information; moral hazard problems involve hidden postcontractual conduct. Both adverse selection and moral hazard are caused by the disputants' ability to hide information about themselves or their conduct. Adverse selection can create inefficiency in negotiation when parties hide information about their valuation or other characteristics prior to agreement. Moral hazard can create inefficiency when one or both parties take hidden actions after an agreement that reduce the joint gains from trade. Moral hazard in the ADR setting is a problem of coordination, because the parties' inability to directly observe each other's postcontractual conduct makes it difficult for the parties to coordinate their performance.

Sequential caucusing is particularly adept at responding to informational problems because it is a uniquely mediative way to elicit and channel private information. This Article shows how mediators can reduce adverse selection in three ways:
(1) by committing parties to break off negotiations when private representations to a mediator indicate that there are no gains from trade;
(2) by committing parties to equally divide the gains from trade; and
(3) by committing to send noisy translations of information disclosed during private caucuses.
Each of these commitments allows a mediator to increase the amount and accuracy of disclosure and thus decrease the impact of the parties' private information. Our third conclusion-that mediators can enhance communication by adding imprecision (or "noise") to privately disclosed information-is especially illuminating. Although adding noise would seem to degrade the quality of information communicated, this result is only true if one mistakenly assumes that the amount of information disclosed is unaffected by the mediator's commitments. When a mediator commits to translate imprecisely the private disclosures of one party, the mediator may induce that party to make more precise disclosures to the mediator, because the mediator's imprecise translation reduces the ability of the other side to use the disclosure to the detriment of the disclosing party. We also show that noisy (but correlated) signals can help mediators mitigate the inefficiencies of moral hazard.

We reach these conclusions with a series of five game-theoretic models of mediation that we refer to for convenience as:
1. "solicit offer" mediation;
2. "imprecise translation" mediation;
3. "job selection" mediation;
4. "battle of the sexes" mediation; and
5. "joint venture" mediation.
These models are admittedly reductive and simplistic. Mediators will never observe situations adhering precisely to the assumptions we adopt. Most importantly, parties to a mediation rarely behave in a purely rational way; overcoming psychological barriers to negotiation will always be an important part of the mediator's function. Nonetheless, our hope is that these models will demonstrate that the mediator can overcome both psychological barriers and economic inefficiency in negotiation. Although some law and economics scholars have been skeptical about how mediation could create value, these stylized models show how mediation commitments can reduce the strategic inefficiencies of adverse selection and moral hazard.

The Article also has implications for practitioners and lawmakers. The models suggest innovative mediation strategies that practitioners must now weigh against alternative mechanisms. For example, we show that mediators can increase the expected gains from trade not only by indirectly or imprecisely disclosing the parties' private information, but even by committing to misrepresent private information a certain proportion of the time. We provide examples showing how mediators in different contexts can productively control the flow of information between the parties by filtering or inserting noise into their private disclosure. Contrary to the standard claims that mediators need to keep all privately disclosed information confidential, we show that some form of controlled disclosure is central to any economic rationale for mediation.

Because credible commitments to particular forms of disclosure are crucial to the effectiveness of mediators, our analysis has implications for how mediation agreements should be drafted and negotiated. In particular, we propose that courts enforce mediation agreements that allow mediators to disclose private information in indirect ways or even break off negotiations.

We also tentatively suggest that courts should enforce, in the absence of contrary language, an implicit provision requiring strict mediator confidentiality-to encourage more explicit contracting about how mediators will use caucus information.

Our economic analysis also suggests the proper contours for legally mandated mediation. We argue that mediation should not be legally mandated when the process aims at mitigating moral hazard, but that mandatory mediation may be necessary to overcome barriers to resolution caused by adverse selection.

Legally mandated mediation may be necessary to mitigate adverse selection because the private information that causes the adverse selection may also impede parties from voluntarily adopting the efficient form of mediation. We show, however, that there are no such strategic barriers to resolving moral hazard through voluntary mediation, which suggests that private mediation contracts should be sufficient to respond to problems of moral hazard. The Article has three Parts. Part I shows how mediation can mitigate the inefficiencies of adverse selection. Part II shows how mediation can mitigate the inefficiencies of moral hazard. And Part III discusses the implications of our analysis for lawmakers and practitioners.


Return to top of the Page