EXPECTED VALUE ARBITRATION JOSHUA DAVIS* Table of Contents I. Introduction ............................................. 47 II. Defining EVA and How It Would Work....................... 52 A. Expected Value — A Familiar Concept in the Law ............ 52 B. EVA is a Distinct Form of Dispute Resolution ............... 54 C. EVA is Unlike Other Proposals for Partial Recovery........... 58 III. The Benefits of EVA ..................................... 70 A. Vindicating Legal Rights Without the Risks of Trial ........... 71 B. EVA Minimizes Errors .................................. 85 C. Encouraging Desirable Expenditures on Litigation ............ 94 IV. Assessing EVA from Various Theoretical Perspectives ......... 106 A. Law and Economics ................................... 106 B. Rights Theorists ...................................... 112 C. The Public-Life Conception of Trial ...................... 114 V. Practical Concerns about EVA ............................. 116 A. Predictability and Reliability in EVA...................... 116 B. Biases that May Affect Arbitrators ........................ 117 C. Factors in Assigning the Expected Value of Trial ............ 119 VI. Conclusion ............................................ 121 Appendix I ............................................... 122 Appendix II .............................................. 123 Appendix III .............................................. 124 Appendix IV.............................................. 125 I. Introduction This Article proposes a new form of dispute resolution: Expected Value Arbitration (EVA). In some ways, EVA resembles traditional forms of arbitration: disputants choose to enter into EVA, they present their legal arguments and evidence to a neutral party, and the neutral party imposes a binding resolution. What makes EVA unique is its standard for decision * Professor, University of San Francisco School of Law. I thank Rhonda Andrews, Morton Davis, David Franklyn, Susan Freiwald, Alice Kaswan, Doug Lichtman, Josh Rosenberg, Stephen Ware and participants in the USF School of Law Faculty Workshop for helpful comments. Allison Campos, Langston Edwards, and Jon Eldredge provided excellent research assistance. All errors remain my own. 47 48 OKLAHOMA LAW REVIEW [Vol. 57:47 making. The neutral decision maker is to award her estimate of the expected value of the outcome at trial — that is, the average of the possible outcomes with each weighted by its likelihood of occurring.1 The expected value of trial is already often used as a point of departure for settlement negotiations.2 By relying on expected value, EVA essentially offers the imposition by a neutral party of an objectively reasonable settlement. EVA lies at the intersection of two trends in the law. One trend is in the practice of law — a trend toward informal dispute resolution. Recent decades have seen a much-remarked explosion of alternatives to trial.3 They include mediation,4 nonbinding arbitration,5 binding arbitration,6 and early neutral evaluation,7 to name a few. The second trend is more academic in nature. Scholars over recent decades have discussed the potential of partial recoveries. In particular, they have challenged the prevailing assumption that a court must adopt a winner-take-all approach to dispute resolution.8 They have suggested instead that a court 1. See John J. Donohue III, Opting for the British Rule, or if Posner and Shavell Can’t Remember the Coase Theorem, Who Will?, 104 HARV. L. REV. 1093, 1096 (1991); Note, Settling for Less: Applying Law and Economics to Poor People, 107 HARV. L. REV. 442, 444 (1993). 2. See, e.g., Paul J. Mode, Jr., & Deanne C. Siemer, The Litigation Partner and the Settlement Partner, 12 LITIG. 33 (1986). Various scholars have long relied on models of litigation that recognize this use of expected value. See, e.g., RICHARD A. POSNER, ECONOMIC ANALYSIS OF LAW § 21.5, at 554-56 (4th ed. 1992) [hereinafter POSNER, ECONOMIC ANALYSIS]; Richard A. Posner, An Economic Approach to Legal Procedure and Judicial Administration, 2 J. LEGAL STUD. 399 (1973) [hereinafter Posner, Economic Approach]; George L. Priest & Benjamin Klein, The Selection of Disputes for Litigation, 13 J. LEGAL STUD. 1 (1984); Steven Shavell, Suit, Settlement, and Trial: A Theoretical Analysis Under Alternative Methods for Allocation of the Legal Costs, 11 J. LEGAL STUD. 55 (1982) [hereinafter Shavell, Suit, Settlement]. 3. See, e.g., Lisa Bernstein, Understanding the Limits of Court-Connected ADR: A Critique of Federal Court-Annexed Arbitration Programs, 141 U. PA. L. REV. 2169, 2172 n.2 (1993); Janeen Kerper, Creative Problem-Solving vs. the Case Method: A Marvelous Adventure in Which Winnie-the-Pooh Meets Mrs. Palsgraf, 34 CAL. W. L. REV. 351, 354 (1998); Roderick W. Macneil, Contract in China: Law, Practice, and Dispute Resolution, 38 STAN. L. REV. 303, 396 (1986); Richard C. Reuben, Constitutional Gravity: A Unitary Theory of Alternative Dispute Resolution and Public Civil Justice, 47 UCLA L. REV. 949, 963 (2000). 4. For a practical description of mediation, see JAY E. GRENIG, ALTERNATIVE DISPUTE RESOLUTION WITH FORMS §§ 2.16, 7.1-7.3 (2d ed. 1997). 5. See, e.g., MYRA WARREN ISENHART & MICHAEL SPANGLE, COLLABORATIVE APPROACHES TO RESOLVING CONFLICT 130-33 (2000). 6. Id. 7. THOMAS CROWLEY, SETTLE IT OUT OF COURT: HOW TO RESOLVE BUSINESS AND PERSONAL DISPUTES USING MEDIATION, ARBITRATION, AND NEGOTIATION 192 (1994); GRENIG, supra note 4, §§ 2.59, 18.10-.11. 8. See, e.g., STEVEN SHAVELL, ECONOMIC ANALYSIS OF ACCIDENT LAW § 5.3 [hereinafter 2004] EXPECTED VALUE ARBITRATION 49 could award a portion of the compensation to which a wronged plaintiff would be entitled, discounted to reflect the court’s uncertainty about whether the plaintiff should prevail.9 EVA improves on the existing proposals for partial recovery in at least two ways. First, although scholars have directed their proposals to courts, voluntary forms of dispute resolution provide a better setting for this kind of creativity. Parties engage in arbitration only if they select it.10 Using a novel standard for resolving disputes that all parties have chosen is far less troubling than having a court impose the new standard, regardless of the parties’ preferences.11 SHAVELL, ECONOMIC ANALYSIS]; Michael Abramowicz, A Compromise Approach to Compromise Verdicts, 89 CAL. L. REV. 233, 234-36 (2001); John E. Coons, Approaches to Court Imposed Compromise — The Uses of Doubt and Reason, 58 NW. U. L. REV. 750, 751-52, 787-93 (1964) [hereinafter Coons, Approaches]; John E. Coons, Compromise as Precise Justice, in COMPROMISE IN ETHICS, LAW, AND POLITICS 190 (J. Roland Pennock & John W. Chapman eds., 1979), reprinted in 68 CAL. L. REV. 250 (1980) [hereinafter Coons, Compromise]; David Kaye, The Limits of the Preponderance of the Evidence Standard: Justifiably Naked Statistical Evidence and Multiple Causation, 1982 Am. B. Found. Res. J. 487, 491; Joseph H. King Jr., Causation, Valuation, and Chance in Personal Injury Torts Involving Preexisting Conditions and Future Consequences, 90 YALE L.J. 1353, 1354 (1981); Saul Levmore, Probabilistic Recoveries, Restitution, and Recurring Wrongs, 19 J. LEGAL STUD. 691, 692 (1990); Charles Nesson, The Evidence or the Event? On Judicial Proof and the Acceptability of Verdicts, 98 HARV. L. REV. 1357, 1382-85 (1985); Neil Orloff & Jery Stedinger, A Framework for Evaluating the Preponderance-of-the-Evidence Standard, 131 U. PA. L. REV. 1159, 1160 (1983); Mario J. Rizzo & Frank S. Arnold, Causal Apportionment in the Law of Torts: An Economic Theory, 80 COLUM. L. REV. 1399, 1399-1405 (1980); Glen O. Robinson, Multiple Causation in Tort Law: Reflections on the DES Cases, 68 VA. L. REV. 713, 715 (1982); David Rosenberg, The Causal Connection in Mass Exposure Cases: A “Public Law” Vision of the Tort System, 97 HARV. L. REV. 849, 866-70, 897-900 (1984); Steven Shavell, Uncertainty over Causation and the Determination of Civil Liability, 28 J.L. & ECON. 587, 589 (1985) [hereinafter Shavell, Uncertainty]. 9. Coons began with a modest proposal to split an award “fifty-fifty” when a judge was in equipoise between the competing positions of the parties. Coons, Approaches, supra note 8, at 757. Rosenberg and Shavell explored awarding a plaintiff recovery proportionate to the likelihood that the defendant caused the injury to the plaintiff. Rosenberg, supra note 8, at 881- 86; Shavell, Uncertainty, supra note 8, at 589. Levmore suggested a hybrid approach that would entail using the standard for recovery recommended by Rosenberg and Shavell in certain cases involving recurring wrongs, depending on the confidence of the fact finder in the proper outcome of litigation. Levmore, supra note 8, at 721-25. Abramowicz extended the sources of potential uncertainty beyond causation to other issues of fact (and, to some extent, law) and suggested his own hybrid approach, in which the finder of fact would use a winner-take-all approach or award a partial recovery, depending on the likelihood that plaintiff should win. Abramowicz, supra note 8. 10. United Steelworkers of Am. v. Warrior & Gulf Navigation Co., 363 U.S. 574, 582 (1960). 11. A creative possibility would be to have the court impose a compromise at the parties’ 50 OKLAHOMA LAW REVIEW [Vol. 57:47 EVA offers a second advantage over existing scholarly proposals using the expected value of trial. Scholars have not focused on this measure of partial recovery, usually recommending instead an award proportionate to the odds that a plaintiff should win.12 This remedy has been called “proportionate damages”13 or “proportional liability.”14 Drawing on analytic tools from the academic literature on rights theory, law and economics, and game theory, this Article argues that an outcome based on the expected value of trial has virtues absent from other proposals for partial recovery. The central claims of this Article pertain to trial: EVA is likely to be both more attractive to many disputants and may better approximate justice than trial. However, EVA compares favorably in various ways to proportionate damages as well.15 Part II explains how EVA would work and what makes it distinctive. Part III then describes three likely virtues of EVA. First, EVA would allow parties to insist on their legal rights without the risks of winner-take-all litigation. In doing so, EVA, unlike other forms of partial recovery, does not vary from the average result at trial. This characteristic increases the likelihood of adoption of EVA by all parties to a dispute and means that EVA is true to the conception of justice embodied in current law. request. This would not be easy to achieve, however. Parties are somewhat constrained in their ability to choose the standard in court. They decide whether to go to trial but, if they do, they must accept imposition of the law as the court interprets it. See, e.g., Stephen J. Ware, Consumer Arbitration as Exceptional Consumer Law (with a Contractualist Reply to Carrington & Haagen), 29 MCGEORGE L. REV. 195, 207-09 (1998) (noting mandatory rules in court, including strict products liability, the warranty of habitability, usury laws, and certain restrictions on insurance and employment contracts). Courts may be similarly unwilling to adjust the standard of appellate review to accommodate the desires of disputants. See Kyocera Corp. v. Prudential-Bache Trade Servs. Inc., 341 F.3d 987, 1003 (9th Cir. 2003) (refusing to allow parties to set standard for review in court); see also Lee Goldman, Contractually Expanded Review of Arbitration Awards, 8 HARV. NEGOTIATION L. REV. 171, 178-79 (2003); Margaret M. Maggio & Richard A. Bales, Contracting Around the FAA: The Enforceability of Private Agreements to Expand Judicial Review of Arbitration Awards, 18 OHIO ST. J. DISP. RES. 151, 154-55, 187 (2002). 12. Abramowicz relies primarily on this measure, although he does so with a proposed hybrid rule that would at times take a winner-take-all approach. Abramowicz, supra note 8, at 236-37. 13. Nesson, supra note 8, at 1382-85. 14. Rosenberg, supra note 8, at 859; Shavell, Uncertainty, supra note 8, at 591. 15. The juxtaposition to proportionate damages serves various purposes, including: (1) to clarify how EVA would work, (2) to emphasize its novel features, and (3) to highlight some of its distinctive virtues. Still, it should be noted at the outset that the two ideas are not mutually exclusive. If courts were to award proportionate damages, an arbitrator in EVA could take that into account in assessing the expected value of a case. 2004] EXPECTED VALUE ARBITRATION 51 Second, EVA may be better than trial, binding arbitration, or proportionate damages at minimizing errors in adjudication. In particular, EVA should produce the same average error across a set of cases as trial (often called “expected error costs”16) and should avoid the largest errors that occur at trial. EVA, like trial, should also produce a lower average error across a set of cases (or lower expected error costs) than proportionate damages. In a sense, then, when it comes to minimizing errors in dispute resolution, EVA offers the best of both worlds. Third, EVA may be more likely than trial to encourage desirable expenditures on litigation and, in particular, may provide incentive for a risk- averse party to make those investments in litigation (and only those investments) that will produce a net gain on average in dollars for that party. For this point, I rely on a line of analysis that has not yet been explored in the legal academic literature, one that could have significant implications. Specifically, I use utility functions to assess the interaction between risk aversion, the continuity or discontinuity of results from a standard for dispute resolution, and expenditures on litigation. A similar analysis would be relevant to other fields of the law, including, for example, the choice between contributory and comparative negligence. Part IV assesses EVA from various theoretical perspectives, including law and economics, rights theory, and the “public-life conception”17 of trial.18 Part V responds to some likely concerns about EVA, including whether the results it produces can be reliable and predictable, whether biases may limit its benefits, and how expected value arbitrators are to identify the factors that should and should not inform their awards. Part VI concludes by recommending that providers of dispute resolution services include EVA as an option for clients. 16. See, e.g., ROBERT G. BONE, CIVIL PROCEDURE: THE ECONOMICS OF CIVIL PROCEDURE 131 (2003). I am assuming that the social cost of an error at trial is the difference between the correct result and the actual result. Posner seems to make this assumption at times. See, e.g., POSNER, ECONOMIC ANALYSIS, supra note 2, at 554. Note, however, at other times he focuses more directly on the social cost caused by inefficient incentives from anticipated errors in adjudication. See, e.g., id. at 549. The latter approach measures social costs more directly, but does not lend itself to a general analysis of the harm from adjudicative errors. 17. The phrase comes from David Luban. David Luban, Settlements and the Erosion of the Public Realm, 83 GEO. L.J. 2619, 2634-60 (1995). 18. Unfortunately, no ready label is available for this last group, as it includes various perspectives. One possibility, although it is underinclusive, is civil republicanism. See Joshua P. Davis, Toward a Jurisprudence of Trial and Settlement: Allocating Attorney’s Fees by Amending Federal Rule of Civil Procedure 68, 48 ALA. L. REV. 65, 124-26 (1996) [hereinafter Davis, Toward a Jurisprudence]. 52 OKLAHOMA LAW REVIEW [Vol. 57:47 II. Defining EVA and How It Would Work At the heart of EVA is the concept of expected value, which is of great use in making decisions that involve risk. The concept has found its way into various aspects of legal practice and should be familiar to litigators and potential arbitrators.19 A. Expected Value — A Familiar Concept in the Law Expected value is the mean of the possible outcomes in a situation with each outcome weighted by its likelihood of occurring.20 Expected value finds a natural application in the resolution of legal disputes. The expected value of a trial, for example, is the sum of each possible outcome in a case multiplied by its odds of being adopted by a court.21 The concept of expected value may once have been foreign to lawyers. If so, it no longer is. It currently has many practical uses in litigation. It can clarify important decisions, including whether to sue and, once litigation has begun, whether to settle and on what terms.22 An example will illustrate how expected value works. Imagine a car accident between Penelope and Dwayne. Penelope claims that Dwayne ran a red light; Dwayne claims that Penelope ran a red light. In either case, the parties agree that they arrived in their cars simultaneously at an intersection and Penelope swerved to avoid a collision. She struck a telephone pole and suffered significant injuries. Dwayne was unharmed. The testimony of each party is the only evidence available regarding fault. Assume Penelope stands a 50% chance of persuading a court to award her $100,000 and a 50% chance of losing. The sum of each possible result multiplied by its likelihood of occurring yields the expected value of trial: (.5 x $100,000) + (.5 x $0) = $50,000. The expected value of trial is $50,000. 19. A literature has developed commenting on the use by practitioners of risk analysis based on expected value and discussing the topic for the benefit of practitioners. See, e.g., Marjorie Corman Aaron & David P. Hoffer, Using Decision Trees as Tools for Settlement, 14 ALTERNATIVES TO HIGH COST LITIG. 71, 72-73 (1996) (suggesting an analysis for decisions in litigation based on expected value); David R. Johnson, Screening the Future for Virtual ADR, 51 DIS. RESOL. J., Apr.-Sep. 1996, at 116 (discussing the use of software to assist litigators in undertaking risk analysis based on expected value); James E. McGuire, Practical Tips for Using Risk Analysis in Mediation, 53 DIS. RESOL. J., May 1998, at 15, 21 (noting lawyers’ familiarity with risk analysis). 20. See Donohue, supra note 1, at 1096; Note, supra note 1, at 444. 21. Donohue, supra note 1, at 1096; Note, supra note 1, at 444. 22. See, e.g., Aaron & Hoffer, supra note 19, at 72-73 (discussing the use of expected value in settlement negotiations, mediation, and the decision whether to sue). 2004] EXPECTED VALUE ARBITRATION 53 This expected value can be helpful in various ways. Penelope would be wise, for example, to consider it in deciding whether to sue.23 Once she calculates an expected value of $50,000, she should compare that to her cost of litigating. If she expects to expend more than $50,000 by the end of trial, it will be, on average, a losing proposition. Indeed, even if she expects her costs to be only $40,000, she will make a net gain of $60,000 half of the time, but she will suffer a net loss of $40,000 half of the time.24 In this case, another factor, her appetite for risk, will be essential in determining if trial is worthwhile. If she is averse to risk, given the modest net expected value of $10,000, she may prefer not to sue at all.25 Her decision will not be easy, and it should account for other factors, including the likelihood that Dwayne will settle before trial.26 The key point, however, is that a useful analysis of whether to sue begins with an estimate of the expected value of trial. Expected value can play a similar role in assessing whether to make or accept a settlement offer. Penelope and Dwayne, for instance, would do well to take this into account in negotiating.27 If Penelope is risk-neutral, she should settle for no less than the expected value of trial less her costs of 23. Id. at 72. 24. Of course, this analysis will be altered if a lawyer accepts her case on the basis of a contingency fee. Even under those circumstances, the lawyer will then have the incentives ascribed in the text to the plaintiff and will have significant influence over the plaintiff’s actions. For a discussion of a contingency lawyer’s decisions in light of risk, see, for example, Peter H. Huang, A New Options Theory for Risk Multipliers of Attorney’s Fees in Federal Civil Rights Litigation, 73 N.Y.U. L. REV. 1943 (1998). Huang’s analysis is quite insightful, although it suffers significantly from a failure to incorporate the limited ability of lawyers to cease litigating a case when they have decided it is no longer a good investment. 25. Scholars generally assume that litigants are averse to risk. See, e.g., Abramowicz, supra note 8, at 240 & n.37. As Abramowicz points out, even litigants who have a taste for risk are unlikely to indulge that taste in protracted litigation and would tend to explore other high-risk ventures, like hang gliding or poker. Id. (citing Richard Caswell, Deterrence and Damages: The Multiplier Principle and Its Alternatives, 97 MICH. L. REV. 2185, 2230 (1999)). Anyone who has seen litigation up close knows that it rarely is the kind of process that one would expect to excite people with an appetite for risk. See also POSNER, ECONOMIC ANALYSIS, supra note 2, §1.1, at 12. 26. Anticipation of the possibility of settlement provides an economic explanation of so- called “strike suits,” in which a plaintiff brings a claim without merit to extract a settlement on favorable terms from a defendant. For a discussion of this possibility, see David Rosenberg & Steven Shavell, A Model in Which Suits Are Brought for Their Nuisance Value, 5 INT’L REV. L. & ECON. 3 (1985). For an overview of the problem, see BONE, supra note 16, at 45-50. 27. In this sense, they will “bargain in the shadow of the law.” See Robert Cooter et al., Bargaining in the Shadow of the Law: A Testable Model of Strategic Behavior, 11 J. LEGAL STUD. 225 (1982); Robert H. Mnookin & Lewis Kornhauser, Bargaining in the Shadow of the Law: The Case of Divorce, 88 YALE L. J. 950 (1979). 54 OKLAHOMA LAW REVIEW [Vol. 57:47 litigating.28 If Dwayne is similarly risk-neutral, he should settle for no more than the expected value of trial plus his costs of litigating.29 Of course, aversion to risk may decrease the minimum amount Penelope will accept or increase the maximum amount Dwayne will pay.30 Further, these calculations offer only a range within which both parties will do better to settle than they would on average at trial. Various other factors will determine where in the range the parties resolve their dispute, if they do at all, including their skill at negotiations, their willingness (or apparent willingness) to endure the costs and risks of litigation, and their psychological needs and desires.31 Nevertheless, what matters for present purposes is that the expected value of trial is an important point of departure in settlement negotiations. None of this is new. Academics have long known that expected value is fundamental in taking a systematic approach to decisions involving risk.32 Litigation always involves risks of one sort or another. It is therefore unsurprising that many lawyers have come to recognize the use of expected value in counseling clients, in making their own decisions whether to accept or continue to prosecute cases, and in crafting settlement offers.33 Because of these uses, lawyers should be familiar enough with the concept behind EVA to feel comfortable recommending the process to their clients. Further, there should be no shortage of arbitrators who are experienced in both practicing in a given area of the law and assigning an expected value to a case. B. EVA Is a Distinct Form of Dispute Resolution EVA, then, could be a practical form of dispute resolution that draws on lawyers’ experience with expected value in various litigation contexts. Nevertheless, it is a novel proposal. This is in part because, unlike the most common forms of dispute resolution, EVA imposes compromise. To understand this claim, it is important to explore the two distinctions on which it relies: (1) distinctions between imposed and voluntary outcomes and (2) distinctions between determined and compromised outcomes. 28. For a standard analysis along these lines, see POSNER, ECONOMIC ANALYSIS, supra note 2, § 21.5, at 554-56. 29. Id. 30. Id. at 557. 31. For a discussion of the role of strategic behavior in the amount for which parties settle, see, for example, DOUGLAS G. BAIRD ET AL., GAME THEORY AND THE LAW 79-158 (1994). 32. See, e.g., POSNER, ECONOMIC ANALYSIS, supra note 2, § 21.5, at 554-56; George L. Priest & Benjamin Klein, The Selection of Disputes for Settlement, 13 J. LEGAL STUD. 1 (1984); Shavell, Suit, Settlement, supra note 2, at 55; Posner, Economic Approach, supra note 2, at 399. 33. See, e.g., supra note 19 and accompanying text. 2004] EXPECTED VALUE ARBITRATION 55 1. Imposed Versus Voluntary Outcomes A first useful distinction is between forms of dispute resolution in which parties have an outcome imposed on them and those in which the parties must voluntarily accept an outcome for it to bind them. In EVA, an arbitrator imposes an outcome on the parties and the parties have no choice but to accept it. In this regard, EVA resembles trial or traditional binding arbitration.34 In settlement or mediation, on the other hand, the parties must choose to accept an outcome before it can bind them.35 Note that, as used here, the terms "voluntary" and "imposed" describe how the parties reach a particular outcome for resolving their dispute, not how they choose the method of dispute resolution. Binding arbitration, for example, is a voluntary process for resolving a legal claim. Both parties must agree to arbitrate for the result to be binding.36 Trial, by contrast, is often imposed on a party against its will. Nevertheless, much like a court in trial, at the end of the day, the arbitrator imposes an outcome on the parties.37 Thus, binding arbitration, like trial, involves imposed outcomes. 2. Compromised Versus Determined Outcomes A second distinction is between compromised and determined outcomes. EVA involves a form of compromise. The decision maker does not choose one party’s version of the facts and the law or even a third, independent view. Rather the decision maker compromises among the various plausible interpretations of the facts and the law, creating an average by weighting each 34. See generally THOMAS E. CROWLEY, SETTLE IT OUT OF COURT: HOW TO RESOLVE BUSINESS AND PERSONAL DISPUTES USING MEDIATION, ARBITRATION, AND NEGOTIATION 171- 73 (1994) (discussing binding arbitration); ISENHART & SPANGLE, supra note 5, at 130-33 (same). 35. See STEPHEN GOLDBERG ET AL., DISPUTE RESOLUTION: NEGOTIATION, MEDIATION, AND OTHER PROCESSES 4 (4th ed. 2003). 36. United Steelworkers v. Warrior & Gulf Navigation Co., 363 U.S. 574, 582 (1960). 37. Imposed outcomes should not be confused with situations where a method of dispute resolution is imposed on the parties, although an outcome is not. The parties, for example, may be required to participate in a mediation process before trial, although the process may prove unsuccessful and any result would have to be accepted by the parties voluntarily. See, e.g., Bernstein, supra note 3, at 2248-51. XXSimilarly, a voluntary resolution of a dispute may require enforcement by a third party. One party may have to pursue litigation to enforce the terms of a settlement agreement. The settlement agreement itself, however, would constitute a voluntary result. XXAdditionally, some outcomes may blur the line between voluntary and imposed. If the parties settle after resolution of a potentially dispositive motion — or, as often occurs, after the judge has threatened each side with an adverse result on, for example, a motion for summary judgment the outcome might be described as partially voluntary and partially imposed. Little turns on the existence of this grey area. 56 OKLAHOMA LAW REVIEW [Vol. 57:47 possible outcome at trial according to its likelihood of occurring.38 In trial and traditional binding arbitration, however, the decision maker (at least officially) adopts a particular view of the law, finds the facts, and applies the law to the facts.39 In other words, the decision maker is supposed to determine all legal and factual issues. Settlement and mediation, much like EVA, ordinarily involve a compromise that takes into account uncertainty about how a decision maker would resolve ambiguous issues.40 The following chart suggests the possible combinations of these characteristics in the most prevalent forms of dispute resolution: 38. My definition of the term “compromise” is somewhat broader than Abramowicz’s focus. See Abramowicz, supra note 8, at 236 & n.24. My definition includes any decision that does not select one among the various possible understandings of the facts or the law, whether the decision relies on a single determination of the odds that the plaintiff should win or splits the difference between different views of those odds. Abramowicz focuses on the former while recognizing the latter. Id. 39. This view is consistent with the U.S. Supreme Court’s position that arbitration merely changes the process for deciding a case, not the substantive standard. See, e.g., Mitsubishi Motors Corp. v. Soler Chrysler-Plymouth, Inc., 473 U.S. 614, 628 (1985). On the other hand, as Stephen Ware points out, many arbitrators and the parties who appear before them believe arbitrators do not act as if they are bound by the law. See, e.g., Stephen J. Ware, Default Rules from Mandatory Rules, 83 MINN. L. REV. 703, 720 nn.78-81 (citing Soia Mentshikoff, Commercial Arbitration, 61 COLUM. L. REV. 846, 861 (1961); 1 GABRIEL M. WILNER, DOMKE ON COMMERCIAL ARBITRATION § 25.01, at 391 (rev. ed. 1995); Dean B. Thomson, Arbitration Theory and Practice: A Survey of AAA Construction Arbitrators, 23 HOFSTRA L. REV. 137, 154-55 (1994); JOHN S. MURRAY ET AL., PROCESSES OF DISPUTE RESOLUTION 514, 636 (2d ed. 1996); 4 IAN R. MACNEIL ET AL., FEDERAL ARBITRATION LAW § 40.5.2.4, at 40:47 (1994); Harry T. Edwards, Arbitration of Employment Discrimination Cases: An Emprical Study, in PROCEEDINGS OF THE TWENTY-EIGHTH ANNUAL MEETING OF THE NATIONAL ACADEMY OF ARBITRATORS 59 (1975); Patricia A. Greenfield, How Do Arbitrators Treat External Law?, 45 INDUS. & LAB. REL. REV. 683, 688 (1992); Edward Brunet, Arbitration and Constitutional Rights, 71 N.C. L. REV. 81, 85 (1992)). 40. Of course, these definitions of voluntary, imposed, determined, and compromised do not follow necessarily. They are merely useful. Moreover, a chosen approach to an outcome is neither irrevocable nor absolute. Parties may receive some imposed determinations on issues that decrease the scope of disagreement and lead to settlement. XXFurther, whether determined or compromised, an imposed outcome will often not be the final stage in resolving a dispute. It may well leave the parties with opportunities for negotiating for their mutual benefit. Indeed, the negotiations may result in a second-stage, voluntary outcome that reflects the cost of enforcing the imposed outcome from the first stage. For purposes of clarity and simplicity, I do not attempt to capture these complexities in the chart below.