|
Article Excerpt
INTRODUCTION
I. THE LAW OF ARBITRATION II. POLICY ARGUMENTS FOR AND AGAINST ARBITRATION: A FIRST LOOK AT THE EMPIRICAL DATA A. Empirical Research Limitations: Systemic Differences of Arbitration and Adjudication Systems B. Prior Empirical Studies 1. Win/loss rates 2. Repeat players and the due process protocol 3. Disposition time 4. Damages 5. Conclusions from prior empirical research III. MANDATORY ARBITRATION SYSTEMS: POSSIBLE REASONS FOR THEIR ADOPTION IV. A NEW PATH FOR EMPIRICAL RESEARCH A. Background Information and Data B. Results and Discussion CONCLUSION
INTRODUCTION
In Gilmer v. Interstate/Johnson Lane Corp., (1) the Supreme Court held that employers could require employees to agree, as a condition of employment, to arbitrate federal statutory age discrimination cases. (2) Its reasoning strongly suggested that predispute arbitration agreements covering virtually any federal and state statutory claims were also enforceable. Gilmer spawned a debate that resulted in a small Alexandrian library of law review articles (3) a series of Supreme Court decisions, (4) hundreds of federal and state court opinions, (5) and various state and federal legislative proposals. (6) Scholars, judges, legislators, management attorneys, and employee advocates debated whether employers should be able to require employees to sign agreements under which all employment disputes, including statutory claims, would be arbitrated and not litigated. Gilmer ultimately turned on an interpretation of the Federal Arbitration Act of 1925 (FAA) as establishing a federal presumption of arbitrability covering any and all disputes, both statutory and common law, arising between parties to arbitration agreements. (7) Both advocates and critics of predispute mandatory arbitration agreements (what we will henceforth call "mandatory arbitration") advanced legal and policy arguments supporting their opposing positions. Legal arguments from both sides drew on statutes and decisional law for support. (8) Critics also complained that widespread resort to arbitration would compromise the rights of employees and make it more difficult for aggrieved plaintiffs to recover. (9) Policy arguments from arbitration advocates, in contrast, dwelled on the inefficiencies of the Equal Employment Opportunity Commission (EEOC)/litigation system with its attendant costs and inequities for employers and employees. (10)
The passage of time has brought greater clarity to the legal positions advanced by arbitration critics and proponents. Fourteen years after Gilmer, the applicable law is relatively stable and clear: employers outside of the transportation industry (11) may require employees to agree to arbitrate all employment disputes as a condition of employment so long as certain due process requirements are met. For example, the arbitrator must be permitted to award statutory remedies in the event of a violation, and a fair procedure for selection of the arbitrator must be afforded.
Unfortunately, we have not achieved comparable stability and clarity in the underlying policy debate. Legal and policy questions incident to mandatory arbitration, many of which are intrinsically empirical, continue to attract sustained scholarly attention. At first, critics and advocates alike leveled assertions that were generally devoid of empirical support. One problem with this approach is obvious: it makes little sense to answer empirical questions without empirical evidence. The emergence of empirical research on these legal and policy questions has changed the pitch and tenor of the arbitration discourse. (12)
The first wave of empirical arbitration research focused on those claims that made it to the formal stages of arbitration. This initial work remains vital because one cannot evaluate the viability and fairness of a dispute resolution system without analyzing the outcomes of actual adjudications. Of course, while a greater understanding of arbitration outcomes is essential, arbitration does not operate in a vacuum. Indeed, arbitration systems are implemented partly to replace the EEOC/litigation system. Analyses of arbitration systems are necessarily comparative. Consequently, a second wave of empirical work compared arbitration and litigation outcomes.
This Article proposes a new path for empirical research in this area. We seek to expand the empirical inquiry to take into account cases that are not ultimately tried or arbitrated. Not all employers implement an arbitration system solely to avoid courts or even juries. Such concerns may be a principal motivator for some employers. For others, we suspect that the attraction of alternative dispute resolution (ADR) systems is that they provide a relatively low-cost alternative for resolving a high volume of relatively low-value cases, and do so in a manner that does not necessarily entail the dissolution of the employment relationship. Replacing litigation with an arbitration system allows such employers and their employees to address issues in a relatively nonadversarial, low-cost forum. (13) An important element of fairness, we submit, is promoted if adjudicative costs do not overwhelm the claim resolution process.
If our contentions are well founded, the cases that are arbitrated or tried do not capture the full value of the arbitration policy, nor are adjudicated cases the best indicator of whether either system is "fair" to employees. Instead, the cases that never make it to formal adjudication may provide a superior vantage for evaluating dispute resolution systems. Until now, it has been difficult to measure the effects of arbitration on these types of cases because arbitration systems were too new to provide any meaningful data. Now, because a significant number of large employers have administered their policies for several years, meaningful, albeit preliminary, data are available to permit tentative, though empirically grounded, conclusions.
One major employer provided us with longitudinal data from its dispute resolution systems (DRSs). We analyzed these data and compared them to comparable EEOC and federal court data. This Article presents results from this case study. Before examining the company's experience, however, we provide an overview of the law and briefly summarize the arguments for and against arbitration. We then assess these arguments in light of previously published empirical work as well as our new data.
I. THE LAW OF ARBITRATION
Gilmer launched the modern arbitration revolution. Before Gilmer, judges, practitioners, and academics widely accepted the view that predispute mandatory arbitration agreements were unenforceable to the extent arbitration was sought for statutory and other public policy--rather than strictly contractual--claims. (14) Gilmer changed the law, but left open a number of issues. (15) Specifically, Gilmer did not settle the question of whether the FAA applied to the majority of employment contracts. (16) In addition, while it provided some guidance, Gilmer did not explicitly establish procedural standards that might condition enforceability. (17) Finally, a number of courts held that Gilmer did not apply to claims arising under Title VII of the Civil Rights Act of 1964. (18)
Most of the legal issues left open by Gilmer and its progeny have been settled. In Circuit City Stores v. Adams, (19) the Supreme Court held that the FAA applies to all employment contracts, except those in the transportation industry. In EEOC v. Luce, Forward, Hamilton, & Scripps, (20) the Ninth Circuit abandoned its view--alone among the circuits--that Gilmer did not apply to Title VII claims. Finally, with the assistance of the American Bar Association, the American Arbitration Association, and leading plaintiff and management lawyers, a "due process protocol" was developed that has won broad judicial and academic acceptance and informs the drafting of arbitration policies satisfying due process concerns in almost every, if not actually every, jurisdiction. Thus, the law is relatively settled: courts will enforce mandatory arbitration agreements so long as the employee is not in me transportation industry and the agreement satisfies certain due process criteria. (21)
II. POLICY ARGUMENTS FOR AND AGAINST ARBITRATION: A FIRST LOOK AT THE EMPIRICAL DATA
While legal challenges to arbitration have largely subsided, policy debates continue with undiminished force in the academy. As is often the case with controversial issues, those opposing arbitration are more vocal than those favoring arbitration. Critics insist that mandatory arbitration should be prohibited because it: (1) does not allow for the development of the law; (22) (2) is private and does not provide for public accountability; (23) (3) is unfair to employees because it can be expensive, limit damages, reduce the statute of limitations, alter the burden of proof, allow for untrained arbitrators to decide cases, limit discovery, and is biased in favor of employers; (24) and (4) is the product of contracts of adhesion and unequal bargaining power. (25) In addition to addressing these arguments, arbitration supporters maintain that arbitration is a better system for most employers and employees because it is less expensive and faster. (26)
The contract-of-adhesion argument is difficult to resolve empirically as it is principally an issue of perception. The question is whether a take-it-or-leave-it arbitration policy should be prohibited in the employment context. We have argued elsewhere that employers set numerous policies over which individual employees rarely negotiate. (27) For example, health insurance, life insurance, pension plans, as well as noncompetition agreements, vacation pay, sick time, holiday pay, and severance pay are conditions of employment that employers almost always offer to employees on a take-it-or-leave-it basis. Even in the union-represented sectors, individual employees typically give up any right to negotiate terms and conditions of employment. (28)
Unlike the contract-of-adhesion argument, the other criticisms of mandatory arbitration are not based on perception. Instead, these are empirical questions that presumably can be answered by data drawn from experience. Put simply, we can compare mandatory arbitration and the EEOC/litigation system and assess whether mandatory arbitration (1) significantly retards the development of the law, (2) precludes public accountability because of its private nature, or (3) creates an unfair dispute resolution system compromising the rights of employees. Of the critics' arguments, the first wave of empirical work primarily examined the unfairness claim, focusing on factors such as the win/loss ratio, "repeat player" effect, and damages. We summarize and evaluate this literature below. We conclude that plaintiffs do not fare significantly better in litigation, that arbitration provides a quicker resolution than litigation, and that available data do not indicate whether damages are fairer under either system.
A. Empirical Research Limitations: Systemic Differences of Arbitration and Adjudication Systems
The arbitration literature is fast growing, though often shedding more heat than light. We evaluate a number of representative articles that seek to marshal available data for answers to three questions: (1) Is arbitration unfair to employees?; (2) Is arbitration faster than litigation?; and (3) Is arbitration less expensive than litigation?
To answer the question of whether arbitration is fair to employees, researchers typically compare the results of cases adjudicated in arbitration against those adjudicated in litigation. Specifically, win/loss rates and monetary awards are compared across the two systems. There is, of course, a problem with this type of analysis: the stream of litigated and arbitrated cases differs. Consequently, one can never be sure if the reason for a disparity in outcomes, if any, involves the adjudication system or some other factor, such as the strength of the case, or perhaps a selection factor determining which cases go to court and which cases end up in arbitration.
To illustrate, assume that there were only two discrimination cases available to study. One case was adjudicated in court while the other was resolved in arbitration. The employee was successful in arbitration and the employer successful in litigation. No reasonable person would conclude from such a limited study that arbitration was more favorable for employees and litigation for employers. If, however, the same case was both arbitrated and litigated and the results were different, a tentative hypothesis about the effect of the two different systems might be possible. Still, we would be uncomfortable reaching any definitive conclusion based on a sample of two occurrences. If, however, we had a universe of one hundred cases that were both arbitrated and litigated and there was a systematic discrepancy between the employee win rates in one system over the other, we could begin to formulate tentative interpretations based on the correlations. Unfortunately, in the real world it is doubtful we could find a significant number of discrimination cases that were both arbitrated and litigated. (29)
The absence of such data poses consequential research design problems. As a second-best option, researchers are limited to comparing two distinct streams of cases: those that were arbitrated and those that were litigated. Perhaps over time the sheer volume of cases will blunt the research design problem. One might argue that if researchers selected, at random, one thousand discrimination cases filed in federal court and randomly assigned these cases to either arbitration or litigation, we could arrive at defensible conclusions about the different systems.
Critical research design concerns persist, however, as the stream of adjudicated and litigated cases is likely to differ systematically. Discrimination cases resolved through arbitration invariably flow from employers that have arbitration policies. Most of these policies have a number of internal steps that employees go through before formal arbitration begins. These internal steps perform critical filtering functions. Thus, before the cases are arbitrated they have likely undergone some form of internal review--whether by an ombudsman, peer review, or a human resources department--often coupled with mediation by an outside neutral. These ADR processes allow the parties to analyze the case and provide several opportunities for settlement. Alternatively, while cases in litigation may have gone through the EEOC investigation, conciliation, and/or mediation, many cases are filed in court with only perfunctory EEOC processing. If the company's internal review and mediation steps have any value at all, there may well be a systematic difference in the "quality" of the cases that make it to arbitration as opposed to those cases that make it to the final stages of litigation.
Differences in internal filtering between arbitration and litigation systems likely skew win/loss arbitration results in favor of employers: Employers might use the internal filters to better determine if a case has merit and then to settle the meritorious cases. (30) Therefore, such cases would not be arbitrated. Because of the likely lower costs of pursuing a claim to arbitration--a point we develop below--cases that are arbitrated are likely to have less merit, on average, than cases that are settled, and an employer with a longstanding program featuring such filtering mechanisms will have a higher win rate than a comparable employer lacking such internal filters.
Exacerbating this disparity in the types of cases that end up in arbitration is the rarity of summary judgment motions in arbitration. Thus, while cases lacking merit, as a matter of law, are excluded from litigation outcome statistics, they may remain part of the arbitration outcome statistics. Because the vast majority of successful motions are made by employers, summary judgment motions, when granted, skew litigation trial numbers in favor of employees. Put simply, a large number of employee losses are excluded from...
|
