In George Watts & Son v. Tiffany & Co., 248 F.3d 577 (7^th Cir. 2001), then Circuit Judge (now Chief Judge) Frank H. Easterbrook of the United States Court of Appeals for the Seventh Circuit said: “What the parties may do, the arbitrator as their mutual agent may do.” 248 F.3d at 581. Chief Judge Easterbrook made this statement in the course of defining the “manifest disregard” standard of review. Applying his “agency model,” he concluded that “the ‘manifest disregard’ principle is limited to two possibilities: an arbitral order requiring the parties to violate the law. . . , and an arbitral order that does not adhere to the legal principles specified by contract, and hence unenforceable under § 10(a)(4).” Id.

Chief Judge Easterbrook’s “agency” model of arbitral authority is instructive. Just as agents derive their authority by the consent of the principal (subject to the rules of apparent and implied authority), arbitrators derive their authority from the parties via the arbitration agreement and the submission. Subject to any restrictions in the arbitration agreement, the arbitrators’ powers to resolve a dispute under a broad arbitration agreement are arguably co-extensive with those of the parties that appointed them.

But the model is not perfect. First, unlike agents, arbitrators are not subject to the control of their principals and owe them no fiduciary duties. Second, analogizing arbitrators as agents of the parties in the way Chief Judge Easterbrook does effectively empowers arbitrators not only to decide cases, but to negotiate settlements that the parties could have entered into. It therefore does not require arbitrators to even arguably interpret the contract or apply the law: As long as the arbitrators do not require the parties to violate the law, and as long as the arbitrators are at least arguably faithful to the parties’ expressed choice-of-law, if any, they can reach whatever decision they wish, whether by application of facts to legal norms or by a compromise settlement that may or may not be rooted in the parties’ agreement. That arguably does not comport with the parties’ presumed, legitimate expectations. For the arbitrator’s job is to decide cases; settlement is a matter for the parties, and should be subject to the parties’ control.

University of Chicago Law School Professor Tom Ginsburg has written an excellent white paper that argues that the deferential standard of review espoused by Watts and other courts does not necessarily make arbitration an attractive substitute for litigation. See Tom Ginsburg, John M. Olin Law & Economics Working Paper No. 502 (2d Series), The Arbitrator as Agent: Why Deferential Review Is Not Always Pro-Arbitration (Dec. 2009) (copy available here). He argues that a more searching standard of review would make the market for arbitrators more transparent, and thus more effective. He advocates using Chief Judge Easterbrook’s agency model as an analytical framework for allowing parties to choose whether they prefer a very deferential standard of review, like that prescribed in Watts; something akin to de novo review, like that available in litigation; or something in between the two. Professor Ginsburg is in the process of publishing in the University of Chicago Law Review an article based on his white paper.

As respects the tie-in between the standard of review and the effectiveness of arbitration, Professor Ginsburg explains that arbitrators are frequently experts in the subject matter of dispute, but, unlike “generalist judges,” they are not necessarily experts on the law or on the application of law to facts, and therefore are “prone to make mistakes, or at least presumptively more prone to do so than are judges.” Since arbitration is a surrogate for adjudication, the U.S. legal system has had to decide what level of scrutiny best advances the federal policy in favor of arbitration. Most arbitration statutes and the Convention on the Recognition and Enforcement of Foreign Arbitral A... have eschewed mere legal error as a ground for judicial review of arbitration awards. Only a particularly egregious error will warrant vacatur, and even then, courts differ as to what – if anything – constitutes an egregious error amounting to manifest disregard of the law (or perhaps manifest disregard of the agreement). Courts consider this deferential approach as the one best suited to implementing a policy in favor of arbitration.

But, as Professor Ginsburg observes, there is “a spectrum of possible alternative regimes, ranging from de novo review to complete non-reviewability,” and choosing which one best promotes arbitration involves a “tradeoff:”

If the standard of review is too rigorous, the benefits of arbitration in terms of speed, cost, and finality may be lost because parties will frequently appeal arbitral awards to the courts. On the other hand, if review is too limited, arbitrators might deliver very poor quality decisions that undermine the attractiveness of arbitration as a whole.

Professor Ginsburg argues that Watts, Hall Street Assoc., L.L.C. v. Mattel , Inc., 552 U.S. ___, slip op. (March 25, 2008), and other court decisions trade off too much in favor of speed, cost and finality by adopting a standard of review that amounts to practically no review at all. He assumes, for the sake of argument, that there are “good arbitrators” that “always interpret the law accurately,” and “bad” ones who “do so only with a probability p < 1.” He assumes that, in selecting arbitrators, a sufficiently large number of purchasers will seek “good” arbitrators (or “bad” arbitrators whose probability of interpreting the law correctly is close to 1.) Parties will attempt to mitigate the risk of choosing a “bad” arbitrator through screening, but market imperfections tend to make the screening process unreliable:

Will screening serve to adequately reduce the agency problem of arbitrators? If the market for arbitrators is sufficiently robust, it might. But the market for arbitrators has certain imperfections. For example, there are no public records of arbitrator performance. Arbitrators need not produce publicly available opinions, and parties generally have no incentive to allow them to reveal the basis for the award. Only when an arbitrator makes an egregious error leading to a vacatur petition (such as manifestly disregarding the law outside the Seventh Circuit) will there be a public record of performance. Hence it is difficult for the parties to a contractual dispute to evaluate potential arbitrators in terms of their ability to interpret law. Reputational considerations are, of course, a factor, but in the absence of reasoned decisions, even past users of a particular arbitrator cannot be sure their favorable outcome resulted from skilled arbitration or a combination of lazy arbitration and luck. There will thus be certain informational asymmetries in the market for arbitrators, allowing bad arbitrators to remain in the market.

A heightened level of judicial monitoring might compensate for some of these imperfections, says Ginsburg. But decisions like Hall Street and Watts allow for only perfunctory review. And while those decisions were intended to make arbitration more attractive, and to promote judicial economy as a result, Professor Ginsburg says their practical effect may be quite the opposite:

One perverse result of the Hall Street decision might be greater pressure on courts to resolve full contract disputes. One rationale for not allowing parties to contract into higher levels of judicial scrutiny (though not fully articulated in the Hall Street decision which relied on a textual analysis of the FAA) would be to enhance judicial economy — the public should not have to subsidize private dispute resolution. But after Hall Street, parties who want a legally proper decision cannot submit to arbitration, or at least will be less likely to do so, because there can be only minimal ex post monitoring of arbitral awards. Like Watts, Hall Street limits the scope of review. But unlike Watts, it does not follow an agency perspective. By preventing courts from policing arbitrator interpretations of law, Hall Street may end up reducing the number of cases sent to arbitration and, perversely, shifting contract disputes to the courts, precisely because there is no alternative way for parties to ensure that arbitrators do follow the law. The Hall Street logic may end up sacrificing judicial economy in an attempt to preserve it, and hurt arbitration in the name of helping it. (footnotes omitted)

But if the parties could “designate the standard of review,” that “would improve the functioning of the market for arbitrators by allowing good arbitrators to signal their status[:]”

The point is that the standard of review will affect the mix of agents in the labor pool. A policy of no scrutiny will draw bad types. A policy of minimal scrutiny, such as under the FAA, will keep some bad types out: it will prevent arbitrators, for example, from applying New York law when they are instructed to apply Wisconsin law. But it will do nothing to hinder an arbitrator who applies Wisconsin law so poorly as to produce an obvious error. Given the existence of agency problems, the hands-off approach of the FAA after Hall Street may end up undermining the arbitration regime by drawing bad arbitrators.

So, working within the framework of the Federal Arbitration Act, and decisions other than Hall Street interpreting it, how do we justify a more heightened standard of review of arbitration awards? Professor Ginsburg says the agency model itself provides the answer:

An agency perspective would allow the parties more freedom in stipulating legal grounds for review. If parties want a decision that is accurate, they could require that arbitrators not make clear errors of law. High quality arbitrator-agents could trade on their ability to interpret law by promising not to make clear errors. Low quality arbitrator-agents would not want to make such enforceable promises and so might be driven from the market. The Hall Street approach limits contractual freedom; the Watts approach might expand it, and in doing so, may in fact enhance arbitration. It might thus allay concerns about a possible ‘flight from arbitration.’ (footnotes omitted)

. . . .

Arbitration is contractual dispute resolution, and this means that arbitrators are agents. The standard of review of arbitral awards sets the level of monitoring of the agents’ interpretation of law. Some agents will prefer not to be subject to monitoring of their performance, and these are the agents who are happy with Hall Street. Other agents have no fear of monitoring. While specifying a universal standard of review applicable for all cases may be an inherently unstable venture, it seems clear that allowing the parties to set the standard, and to choose higher levels of monitoring by contract, will reduce agency slack and allow parties to determine what type of arbitrator they are hiring. Deferential review, in short, is not always pro-arbitration.

We have quoted extensively from Professor Ginsburg’s work, but do not let this post be a substitute for reading and studying his white paper and the upcoming University of Chicago Law Review article. Professor Ginsburg is a second-generation law and economics theorist, and like the first generation – whose ranks include Associate Justice Antonin G. Scalia, Circuit Judge Richard A. Posner and Chief Judge Easterbrook – he writes clearly and well, and more importantly — like those of the first generation – his theories make good sense, which is all too uncommon these days.

NOTE: The preceding article was originally posted on April 7, 2010 at The Loree Reinsurance and Arbitration Law Forum.