A funny thing happened on the way to the (alternative) forum: Reexamining Alexander v. Gardner-Denver in the wake of Gilmer v. Interstate/Johnson Lane Corp.

Sarah Rudolph Cole*

I. INTRODUCTION

In its landmark 1991 decision in Gilmer v. Interstate/Johnson Lane Corporation,1 the Supreme Court held that a predispute agreement to arbitrate statutory claims contained in a securities representative's registration application was enforceable.2 This decision triggered an exponential increase in the use of arbitration agreements3 as employers interpreted Gilmer to authorize the insertion of such agreements into contracts with existing employees and prospective hires.4 Employers have embraced these agreements, hoping that arbitration will deliver what it promises: inexpensive and speedy decisionmaking, finality, and confidentiality.5 Moreover, many employers are convinced that together with providing a more expeditious, less expensive system of justice, arbitration will improve their bottom line by lowering potential damage awards.6 Employers' enthusiasm for the perceived benefits of inserting arbitration clauses into employment agreements, together with the judicial approval of these clauses,7 ensures the continued use of such agreements in the nonunionized workplace.8

A certain inevitability surrounded the emergence of arbitration as a preferred method for resolving employment disputes in the nonunionized sector. After all, arbitration had long been the preferred means for resolving employment disputes in the unionized workplace.9 Perhaps it was the extraordinary success of arbitration in resolving disputes in the unionized sector that precipitated nonunionized employers' adoption of arbitration to resolve their own ever-increasing number of employment disputes. Whatever the reason for its increased use, arbitration of employment disputes in the nonunionized sector is here to stay.

Although arbitration originated in organized labor and commercial settings, in at least one respect, nonunionized arbitration has developed more rapidly than arbitration in the unionized sector.lo While Gilmer approved the use of predispute agreements to arbitrate discrimination claims in the nonunionized workplace, the use of such agreements is not permitted in the unionized world because of the 1974 Supreme Court decision in Alexander v. Gardner-Denver. 11 In GardnerDenver, the issue was whether an unionized employee, who, as required by his union's collective bargaining agreement, had submitted his claim under the agreement's nondiscrimination clause to final arbitration, retained the right to bring a Title VII claim in federal court following the arbitration. The Court determined that an unionized employee's right to a trial de novo on a Title VII claim is not precluded by prior submission of a claim to arbitration under a collective bargaining agreement's nondiscrimination clause.12

The continued viability of Gardner-Denver following Gilmer remains an open question, at least in cases where the parties expressly agree to abide by antidiscrimination laws. Gilmer superficially addressed Gardner-Denver's continuing validity in response to Gilmer's argument that statutory claims could not be the subject of a predispute arbitration agreement.13 While the Court concluded that Gilmer's reliance on Gardner-Denver was misplaced,l4 its willing acceptance of arbitration as a method for resolving statutory claims in the employment context seemed to cry for a re-evaluation of Gardner-Denver that the Court failed to provide. Thus, the question remained: what effect does the Gilmer decision have, if any, on the enforceability of predispute arbitration agreements between unions and employers?

Some commentators have suggested that Gilmer does not provide an opportunity for revisiting Gardner-Denver.l5 Yet at least one circuit court and several district courts have used Gilmer as the basis for enforcing an agreement to arbitrate statutory claims in a collective bargaining context.ls Rejecting the continued application of Gardner-Denver, the Fourth Circuit in Austin v. Owens-Brockway Glass Container, Inc.,17 held that a represented employee's statutory claim, based on alleged violations of Title VII and the Americans with Disabilities Act, could be arbitrated.ls After emphasizing that the parties had agreed that external law should govern the dispute, the Austin Court rejected Gardner-Denver, holding that the arbitral setting is an adequate forum for the resolution of statutory disputes where the parties have voluntarily agreed to arbitrate their claims. 19

The Austin decision has sparked heated debate among both bench and bar over the continuing viability of Gardner-Denver.20 While the reasoning of Austin is not entirely sound, its conclusion is correct-Gardner-Denver's absolute prohibition on arbitration of statutory claims is inconsistent with the Supreme Court's changed attitude toward the suitability of arbitration announced in Gilmer and numerous other judicial decisions.21 Moreover, the rationale underlying Gardner-Denver, while somewhat understandable given Gardner-Denver's facts, is no longer compelling in cases where parties have agreed to arbitrate their statutory claims in accordance with federal antidiscrimination provisions. In those cases, Gilmer requires a re-evaluation of the foundation for the decision in GardnerDenver.

Part II of this article conducts just such a re-evaluation, and concludes that the bases for the Gardner-Denver decision are unsound. Part III contends that since the reasoning of GardnerDenver is unsound the only remaining question is whether any reason exists for continuing to reject collectively bargained agreements to arbitrate statutory claims. Part III demonstrates that if the "Gilmer agreement," which requires an employee, as a condition of employment, to sign a predispute arbitration agreement foregoing all access to jury trials, is enforceable, then the "Gardner-Denver agreement," an agreement between the union and employer to arbitrate employees' claims, must be similarly enforceable. This article uses a game theoretic analysis to argue that there is actually greater reason to enforce "Gardner-Denver agreements" than "Gilmer agreements." Game theory demonstrates that the structural protections inherent in the collective bargaining context cannot be duplicated in cases involving agreements to arbitrate individual statutory claims. Thus, this article contends that even if Gilmer is ultimately overturned,22 "Gardner-Denver agreements" should remain enforceable.

II. ALEXANDER V. GARDNER-DENVER NEED NOT APPLY

In Gilmer, the Court compelled arbitration of Gilmer's age discrimination claim because he had signed a predispute agreement to arbitrate claims arising out of his employment. In an attempt to avoid arbitration, Gilmer raised a number of issues, including his theory that statutory claims could not be the subject of a predispute agreement. The Court rejected this argument, as well as Gilmer's argument that arbitration provided an inadequate forum in which to vindicate statutory rights because of its lack of procedural safeguards.23 Emphasizing how the judicial attitude toward arbitration had changed since the days of Wilko v. Swan24 and Gardner-Denver, the Court enforced the arbitration agreement. In addition, the Court rejected Gilmer's claims that the agreement should not be enforced because it was the result of unequal bargaining power.25 Gilmer has prompted commentators and courts alike to reconsider the Gardner-Denver ruling.26 In the aftermath of Gilmer, some courts have taken the decision at face value, concluding that Gardner-Denver is still good law.27 Other courts have concluded that Gardner-Denver is inconsistent with the principles Gilmer advocated and have, therefore, used Gilmer to reject Gardner-Denver.23 Analyzing Gardner-Denver in light of Gilmer demonstrates that many of the reasons offered in support of Gardner-Denver were wrong when proffered; other reasons, such as the unsuitability of the arbitral forum to resolve statutory claims, no longer correctly state the law. Given the changed perception of arbitration and the lack of remaining justifications for the Gardner-Denver decision, where the parties have agreed to resolve their statutory claims using external antidiscrimination law, Gardner-Denver should be reversed.

The Gardner-Denver Court confronted and resolved four separate issues in reaching its conclusion. The Court's first concern was whether an unionized employee's right to a trial de novo on his Title VII claim should be precluded because of his prior submission of the dispute to an arbitrator.29 Second, the Court expressed reservations about the adequacy of the arbitral forum as a substitute for litigation.30 Third, in a footnote, the Court raised the concern that the interests of the individual might be subordinated to those of the group if the union were permitted to waive an employee's right to select a forum.31 Finally, the Court suggested that an employee's right to be free from racial discrimination is an individual statutory right that the union is not authorized to waive.32

The Gilmer Court rejected Gardner-Denver to the extent that it rested on the ground that arbitration was inferior to litigation for the resolution of statutory claims. According to the Court, [I]n our recent arbitration cases we have already rejected most of these arguments as insufficient to preclude arbitration of statutory claims. Such generalized attacks on arbitration "res(t] on suspicion of arbitration as a method of weakening the protections afforded in the substantive law to would-be complainants," and as such, they are "far out of step with our current strong endorsement of the federal statutes favoring this method of resolving disputes."33

The Court also stated, "so long as the prospective litigant effectively may vindicate [his or her] statutory cause of action in the arbitral forum, the statute will continue to serve its remedial and deterrent function."34

Yet the decision in Gilmer agreed with Gardner-Denver's theory that a potential disparity in interests between the union and an employee means that subsequent litigation of a statutory claim is permissible. 35Moreover, the Gilmer Court reiterated its belief that collective bargaining arbitration involved contractual, not statutory rights. 36 The Court also espoused the GardnerDenver theory that the union has the power to waive collective rights, but not individual rights." Thus, the Gilmer Court concluded that: Gardner-Denver remains good law because Gilmer involved an express agreement to arbitrate statutory claims while Gardner-Denver's agreement extended only to contract-based claims.38

Following Gilmer, then, the remaining justifications for denying enforcement of collectively bargained predispute agreements to arbitrate statutory claims are that (1) the potential disparity in interests between a union and an employee prohibits the enforcement of the agreement; (2) the union's ability to waive the collective rights of its constituents does not include the ability to waive an employee's statutory right to select a forum for the adjudication of his or her statutory discrimination claim; (3) the labor arbitrator has no authority to resolve statutory disputes; and (4) labor arbitrators are experts in resolving contractual disputes, not statutory ones. Careful examination of each of these concerns establishes that they are inadequate as justifications for rejecting collectively bargained agreements to arbitrate statutory claims, especially in light of the Court's strong preference for enforcement of arbitration agreements.

A. The Disparity Between the Interests of the Union and the Represented Employee is Insignificant The Court, both in Gardner-Denver and later in Gilmer, raised the important concern that the union, as labor's exclusive representative, might use its power to bargain to the detriment of the interests of a certain employee or group of employees.as The theory is that unions might sacrifice individual or protected groups' preferences in order to obtain benefits for the majority. By definition, exclusive representation involves individual employee sacrifice.40 It is theoretically possible that, as a large entity, the union might have prejudices or, at the least, be more interested in responding to, and satisfying, the needs of the majority. Moreover, as an elected entity, the union may recognize that if it is able to increase the number and type of claims it handles, it will become more powerful.41 The union's desire to increase its importance to the employees and thereby become indispensable may contribute to its motivation to give away the rights of individuals too easily. The question then is whether the union's ability to disregard or bargain away protected groups' interests should invalidate a collectively bargained agreement to arbitrate statutory discrimination claims.

1. Public choice theory suggests that protected group interests are not compromised

It is entirely possible that protected groups-those whose rights are protected by antidiscrimination statutes-actually receive greater attention and representation from the union than does the majority. Unions may be more responsive to the needs of the protected classes because they articulate those concerns to the union while the majority remains silent.

This is an application of the theory of public choice. Public choice theory involves the application of microeconomics and game theory to legislative decisionmaking.42 While public choice theory has typically focused on the production of law by legislators, regulatory agencies and courts, the theory applies equally well to any large, elected group that must respond to its constituency. Public choice theorists explain that the compromises reached by legislators, as codified in statutes, depend, in great part, upon the influence of special interest groups. Theoretically, the groups with greatest influence are those who can organize easily to promote their ends. To develop an effective political coalition an organizing group must be able to reduce informational and organizational costs. Otherwise, the costs of creating the organization will be prohibitive. As Professor Frickey emphasizes, the groups most likely to organize and influence the legislature are those who can minimize organization costs by identifying similarly situated individuals 43 Protected groups are more likely to succeed in organizing due to an ease of identification and because they are more "likely to receive discrete benefits or suffer disproportionate burdens . . . than the diffuse public."4 As a result, Frickey predicts that "lots" of statutes provide "concentrated, unjustified benefits to small groups at the expense of the general public."45

The theory of public choice arose to explain legislative results that seemed to prefer the interests of the political minority to those of the political majority. Public choice explained such results using the theory of "interest group capture" articulated above. To apply public choice theory in the union setting requires proof that unions act like legislatures when making decisions. In the context of public choice analysis, this requires a showing that the union's decisions reflect a preference for the political minority at the expense of the political majority. This proof, if it exists, can be obtained by examination of the analogue to the "legislation" produced by the union-the "collective bargaining agreement"-as well as other quantifiable union activities, such as lobbying.

The universal inclusion of nondiscrimination clauses in collective bargaining agreements would seem to suggest protected group capture. So too would the continuing union efforts to eliminate sexual discrimination, 46and fetal protection policies.47 More recently, union efforts on behalf of the disabled culminated in the passage of the Americans with Disabilities Act, demonstrating the powerful influence of protected groups. 48

Judicial decisions also reflect union efforts on behalf of protected groups. In one famous case, United Steelworkers u. Weber,49 the Supreme Court considered the validity of an affirmative action plan which created a training program and reserved fifty percent of the program's positions for black employees. The program was intended to eliminate the racial imbalances in the employer's skilled craft positions. The Court upheld the plan against a challenge by a white employee. For purposes of this article, it is important to note only that the program was developed at the behest of the union. The union's behavior demonstrates some evidence of a protected group's ability to influence union decisionmaking.

Similarly, in another well known case, Wygant v. Jackson Board of Education, 50the Court considered a provision of a collective bargaining agreement that retained probationary minority employees during layoffs over nonminority employees with greater seniority. While this provision was ultimately stricken, this case highlights unions' efforts on behalf of protected groups.

Risks of marginalization and erasure of protected classes by unions exist-historically, unions were hardly thought of as protectors of minority rights. Yet even commentators who believe that unions tend to favor majority interests, concede that many unions have "attempted to foster gender consciousness as well as class consciousness through their organizing and pressure strategies," and that "the labor movement has done more than any other social institution to improve women's economic status."sl Thus unions, who at first blush would seem more likely to marginalize protected groups, may in fact do much to advance minority interests.

Applying public choice theory to the operation of unions, one would expect to see groups who may suffer disproportionately if predispute agreements to arbitrate statutory claims are included in collective bargaining agreements organizing and using their influence to ensure that the union does not concede their right to bring statutory claims in federal court. After all, organization is not difficult because protected groups can identify potential members and define interests easily. The protected groups share the well-defined interest of wishing to avoid the detrimental effects discrimination potentially causes. Moreover, the protected groups tend to be small. This helps such groups overcome potential free-rider problems52 and concentrates the benefits of the union's actions to create strong incentives for the group to lobby for beneficial results.

A concern that may arise when public choice theory is applied to nonlegislative decisionmaking is whether the protected classes will be able to convince their elected representatives that they should act in the protected group's interest. In the legislative process, this goal is accomplished with pledges of money for reelection campaigns and public support, among other promises. In the union setting, contributions of money are not likely to occur. Yet members of unions wield significant power. Under the National Labor Relations Act (NLRA), a petition for a National Labor Relations Board (NLRB) representation can be filed by unions, employers, or the employees themselves.53 In order to file a representation petition, a union must first demonstrate that a substantial number of employees support the union.54 If the NLRB believes that at least thirty percent of the employees in an appropriate bargaining unit support the union, it will call for an election.55 In that election process, a rival union can obtain a position on the ballot upon the showing of the support of just a single employee.56 If the rival union wants full intervention in the representation hearing, it need only show that ten percent of the employees support it.57

Because the NLRA's requirements for appearing on a ballot in an election are minimal, the protected class that is able to organize has a great deal of power to influence the kinds of provisions adopted in the negotiation process. The influence of a protected class is enhanced by its ability to threaten the incumbent union every twelve months to three years.58 With such frequent challenges possible, a union that wishes to remain in power must pay attention to organized groups. Failure to attend to their needs might very well result in a change of leadership to a rival union more interested in satisfying the needs of the protected group. This result is made more likely because, as in the legislative process, the majority is likely to remain unorganized and, thus, without a voice.

In another way, the case for application of public choice theory to union decisionmaking is even more compelling than it is in the legislative process. Commentators have repeatedly expressed concern with the ability of the public choice theory to explain legislative outcomes. The possibility always exists that a statute will be intentionally, or unintentionally, misconstrued during the judicial review process or, before that, substantially changed through the Congressional committee system.59 The legislation might also receive a Presidential veto or may be delegated to an administrative agency for further consideration.so The possibility that proposed legislation will be substantially altered or misinterpreted decreases the incentive of the interest group to influence the legislation. Yet public choice theory remains a popular method for explaining legislation.

Although similar problems exist in the union setting, the opportunities for misinterpretation are greatly reduced. A protected group interested in particular "legislation" in the collective bargaining agreement need only concern itself with the possibilities that management will change the proposed provision or that an arbitrator might subsequently misinterpret it. The reviewing court might also commit error, but this possibility is not significant because only limited judicial review is available. While these are real possibilities, the universe of potential effects in the union setting is smaller than the possibility of alteration of legislation in the legislative process. Thus, the protected groups' incentives to influence union "legislation" is at least as strong as an interest group's incentive to influence legislation.

While empirical evidence supporting application of public choice theory to unions has not been gathered, in at least one well-known NLRB decision, the Board recognized the analogies. In Jubilee Manufacturing Co.,sl the Board stated that the effect of employer discrimination might "cause minorities to coalesce, and it is possible that this could lead to collective action with nonminority group union members."62 The Board emphasized that given their organizational ability, the minority groups, acting alone, should have the strength to eliminate discriminatory practices by the employer.63

Applying public choice theory in the union setting, one would predict that agreements to arbitrate statutory claims would not be included in collective bargaining agreements if the wellorganized protected groups believed that such agreements were not in their best interest. Because these agreements are becoming more frequent rather than less, the protected groups may well believe that such agreements are not disadvantageous.

2. Title VII of the Civil Rights Act of 1964 prohibits discrimination against protected groups In the absence of empirical evidence supporting the application of public choice theory to union action, concerns that the union will prefer majority interests at the expense of protected groups may still be overcome by two alternative legislative protections against union abuse of the power. Both the duty of fair representation and Title VII of the Civil Rights Act of 1964 ensure that the union's increased power will not be accompanied by an increase in discrimination against protected groups.

a. The duty of fair representation. The duty of fair representation (DFR) obligates the union to represent fairly all members of the bargaining unit and process grievances in good faith, without hostility or discriminatory intent. 64While courts give unions fairly wide latitude in negotiating agreements and resolving grievances, in order to avoid liability for breaching the DFR, the union must provide a legitimate and rational explanation for its conduct. 65 In determining whether the union's decisions are reasonable, courts consider the basis for the union's decision. If the union's decision is based on "impermissible" or "invidious" factors, the union is held to be in breach of its duty. "Impermissible factors" include the members' race, sex, national origin, political positions or status as union members. To the extent that most nondiscrimination clauses in collective bargaining agreements have been expanded to include other protected statuses, union decisionmaking that relied upon such information is likely to be considered a breach as well. 66

Courts hold that the fair representation duty imposes on labor unions both the duty not to discriminate and an "affirmative duty to take corrective steps to ensure compliance with Title VII.'47 Thus, the fair representation duty, at least in the context of members' discrimination claims in contract negotiation and administration, imposes a significant burden on the union to avoid even the appearance of discriminatory decisionmaking.

Some commentators criticize judicial analysis of the fair representation duty, suggesting that the courts' limited judicial review of DFR challenges renders the DFR ineffective.

According to the critics, the "DFR" is meaningless because it is based on a principle of "fairness" that is extremely difficult to judge.69 Thus, in their view, the DFR rarely results in the second-guessing of union decisions.70

While judicial review of DFR violations is a potential problem, this criticism is less compelling when the union's actions result in discrimination against a discrete group. Where a union relies on "invidious factors" such as those articulated in Title VII, courts are quick to find a DFR violation."1 Moreover, judicial understanding of the nature and scope of the "invidious" categories makes it easy for courts to find a DFR breach.72 Thus, concerns that union decisions are rarely struck down on the principle of distributive fairness should not affect the vitality of the DFR claim as a means to limit discrimination against protected classes, at least when the union's decision is based on an invidious factor. Instead, the good faith duty stands as a bar to the union's ability to prefer majority interests.

b. Title VII protection. If the duty of fair representation were insufficient to ensure that the union did not discriminate against any of its members, Title VII provides overlapping protection to employees against union discrimination on the basis of race, color, religion, sex, or national origin.73 While the union is still occasionally a defendant in a Title VII action instituted by an employee,74 more often the union's role has been as an active player in the effort to eliminate unlawful employment discrimination in the workplace. Many labor unions have advocated vociferously for the elimination of sexual discrimination,75 disability discrimination and fetal protection policies.76

When the union has discriminated, courts do not hesitate to impose liability under Title VII.77 Courts also emphasize that Title VII not only imposes a duty on unions to avoid active discrimination, but also to eliminate existing discriminatory practices.78

This is not to suggest that unions have resolved the dilemma of responding to majority needs while still protecting minorities or that unions are never guilty of racial discrimination. Yet it would seem that in light of the severe penalties that can be imposed for discriminatory behavior, unions would have little incentive to negotiate an agreement to arbitrate statutory claims if such an agreement could be considered discriminatory. As Samuel Estreicher noted, under current law, an employee claiming inadequate union representation may disregard the collective bargaining agreement's finality provisions and go directly to court.9 Consequently, the union will be forced to defend its decision to negotiate a clause or process a grievance in front of a jury at its own expense. If a breach is ultimately found, the union will have to pay damages. Because unions are organizations with limited resources, it would not be surprising for them to avoid the risk of trial on a DFR or Title VII claim even if the consequence is overprotecting protected classes.so

B. The Union May Waive the Right of Its Members to Select a Forum for the Airing of Discrimination Claims

Another objection to the enforcement of a union's agreement to arbitrate its members' statutory claims is that the union only has the power to waive collective rights, such as the right to strike, and not the power to waive individual rights, such as the right to select a forum for the airing of a Title VII discrimination claim. In Austin, the Fourth Circuit stated that the union's power to waive the right to strike and other rights protected by the NLRA,al included the power to exchange the right to a forum for statutory claims for some other benefit.82 The Fourth Circuit believed that the employees' designation of the union as their representative empowered the union to bargain all terms and conditions of employment, including agreements to arbitrate employee's statutory claims. Since the employees voluntarily elect the union, the court reasoned, the agreement to arbitrate statutory claims is also voluntary and, therefore, enforceable.

According to the dissent, a labor union cannot waive a member's right to a judicial forum for a statutory claim because that right belongs to the individual.83 According to the dissent, the power to waive collective rights does not include the power to waive individual ones.84 The supposition underlying the dissent's position in Austin and the Court's position in Gardner-Denver is that a union's agreement with the employer to arbitrate individual employees' statutory claims is invalid because the individual is not offered the opportunity to waive his right to a forum. In other words, only the individual can waive forum selection and the selection of the union as bargaining agent does not constitute such a waiver. 85

By contrast, the Gilmer Court clearly stated that an arbitration agreement signed by an employee as a condition of employment is a legitimate waiver of the employee's right to a forum. In the securities industry, where Gilmer worked, arbitration agreements in individual employees' contracts are presented on a take-it-or-leave-it basis. The only alternative to signing such an agreement is turning down the job.ss This concern is magnified for securities industry employees because all jobs in that industry require the signing of such an agreement. The decision to reject the arbitration agreement is the decision to work outside the securities industry. Yet in Gilmer and subsequent decisions, courts have emphasized that the fact that the agreement is a condition of employment does not render the agreement unenforceable. Such agreements are routinely enforced in the securities industry and have been rejected only once outside the securities industry.s? Supporters of Gardner-Denver argue that it is the ability of the employee to refuse the agreement that makes his waiver meaningful. Interestingly, the choices presented to the represented employee are remarkably similar to those presented to the "Gilmer employee." Once an employee becomes aware that the union and the employer have agreed to arbitrate employees' statutory claims, the employee has the following options: abide by the union's agreement or look for another job. This is the identical dilemma unrepresented employees face. Yet the basis for the Gilmer Court's decision to enforce such agreements as voluntary is the understanding that every employee has the option to reject the arbitration agreement when it is offered. Represented employees are offered the same choices: to continue working under the agreement to arbitrate statutory claims or look for another job. The distinction Gilmer attempts to draw between the choices of represented employees and unrepresented employees is empty formalism and should be rejected. If the "Gilmer agreement" waiver has legal significance, so too should the union's waiver of a right to a forum for adjudication of its members' statutory claims.

At the time Gardner-Denver was decided, the concept of using the grievance arbitration machinery to resolve a statutory discrimination claim was unthinkable. Yet in the twenty years since the Gardner-Denver ruling, the acceptance of arbitration as a means of resolving statutory disputes and the recognition of arbitrators as potential experts in employment law issues has increased dramatically. Absent evidence that a union is derogating its minority members' interests in favor of majority desires, there is no justification for ignoring the parties' wish to resolve statutory disputes in arbitration. That the represented employee does not have the option to reject an arbitration provision should not prove troublesome since all employees are offered the same choice: take the agreement or leave the job. Since the bases for Gardner-Denver have eroded over time, it should be overruled. As has been shown above, none of the reasons Gilmer cited in opposition to the use of predispute agreements to arbitrate statutory rights are valid concerns in the union setting.

C. Arbitrators Are Authorized to Utilize External Law When Interpreting the Collective Bargaining Agreement if the Parties Have Agreed to It

In Gardner-Denver, the Court emphasized that when an employee submits his grievance to arbitration, he seeks to vindicate his contractual rights, not his statutory rights.? According to the Court, statutory rights are independent of the contract, and must be adjudicated separately even if they arise from the same factual occurrence.89 It is crucial to note that while the Gardner-Denver collective bargaining agreement contained a standard nondiscrimination clause, 90it did not expressly require statutory claims to be arbitrated. As a result, for purposes of Gardner-Denver, the Court was right-the parties agreed only to arbitrate contractual disputes. What the Gilmer Court subsequently misunderstood when it confronted the identical issue, is that grievance arbitration is not always limited to resolving contractual disputes; parties to collective bargaining agreements can agree to arbitrate noncontractual diputes. While they rarely did so in 1974, it has become increasingly common to see parties agreeing to arbitrate statutory disputes.91 In these agreements, the parties agree that the arbitrator will apply "external law"-the same law a court would apply were it resolving the dispute. As a result, arbitrators have become more comfortable interpreting and applying Title VII and other antidiscrimination statutes.92 Thus, it is no longer true that grievance arbitration is a forum solely for the resolution of contractual claims.

It is essential for parties who wish to resolve their disputes using external law to make that intention clear in the language of the collective bargaining agreement. Where the parties choose to incorporate external law into their agreement, the arbitrator is required to interpret and apply that law. 93This was true even in the days of Gardner-Denver. In the absence of a stated intention, an arbitrator should reject external law because his principal task is to interpret and apply the terms of the contract. If it is not clear that the parties wish to use external law to resolve the dispute, the arbitrator will be prohibited from using it.

The debate questioning the propriety of using external law to assist arbitrators in interpretation of an agreement arose because most agreements contained broadly drawn contractual provisions rather than specific commands regarding the use of external law to resolve disputes.94 In such cases, the courts decided to leave to the arbitrators the decision whether external law should be used to resolve disputes. 95 The Court in United Steelworkers v. Enterprise Wheel & Car Corp., 96suggested that an arbitrator could look beyond the agreement to the external law for assistance in "determining the sense of the agreement."97 At the same time, it emphasized that an arbitrator must still look to the collective bargaining agreement and draw the "essence" of any award from the agreement, not external law.98 But Enterprise Wheel involved a broadly written clause, not a clause that specifically articulated which external laws the arbitrator was to consider. Where the contract states that the parties will comply with external laws prohibiting discrimination and that disputes regarding discrimination claims will be arbitrated, the ambiguity surrounding the arbitrator's interpretive role is reduced. 99 In such a case, the arbitrator is not simply authorized to utilize external law; the arbitrator must use external law when resolving the dispute.100

The parties may go even further. In Austin v. OwensBrockway Glass Container, Inc.,lol the collective bargaining agreement stated that the arbitrator was to resolve disputes that arise under federal antidiscrimination laws.loo If the parties authorize it and are satisfied that they can find an arbitrator capable of interpreting those laws, there is no legitimate basis for prohibiting arbitrators from resolving these disputes.103 Thus, the statement in Gardner-Denver, that arbitration is designed to resolve disputes involving contractual rights and not statutory rights, is simply inaccurate.l04 Concerns regarding the arbitrator's power to decide cases using external law provide no basis for disturbing the parties' agreement.105

D. Labor Arbitrators are Experts in Statutory Disputes as Well as Contractual Disputes

According to the Supreme Court in Gardner-Denver, a major impediment to the enforcement of arbitration agreements between unions and employers is that the arbitrators chosen to resolve these disputes are educated in the "law of the shop" rather than the "law of the land."loo If arbitrators are not competent to analyze and decide statutory claims, then the propriety of expanding the arbitrator's role as an interpreter of law would be questionable.

Although the Gilmer Court did not address arbitrators' ability to interpret statutes in resolving disputes between represented employees and employers, it quite clearly authorized arbitrators to resolve exactly these questions when unrepresented employees and employers are involved. Arbitrators for labor arbitration and private employment disputes are drawn from the same pool, typically the American Arbitration Association. Thus, there is no reason to think that the arbitrator who resolves disputes between the unrepresented employee and his employer is any less qualified to resolve the identical dispute when a represented employee raises it.

That the Gilmer Court empowered arbitrators to resolve statutory disputes does not answer the question of whether arbitrators should resolve these disputes. Yet most concerns regarding arbitrators' ability to interpret antidiscrimination statutes are largely misplaced. As Elkouri and Elkouri emphasize, labor arbitrators have long utilized Title VII and other discrimination statutes to resolve discrimination claims.107 For this reason, "qualified observers in the field of labor law and arbitration" believe that arbitrators do have the requisite ability to interpret and apply external law. 108

Where external legal issues are raised, it would be beneficial to both parties if the selected arbitrator was a lawyer or former judge with experience in employment law. In a recent revision of the American Arbitration Association's (AAA) rules, the AAA stated that arbitrators who resolve employment disputes "shall have familiarity with the employment field."los Even without these precautions the selected arbitrator will likely have the expertise the parties wish her to have. Because both parties have an incentive to choose an expert in the employment law field when a discrimination claim has been raised, their joint decision to select a particular arbitrator should be respected.

Gilmer's approval of the use of arbitrators to resolve discrimination claims, together with the arbitrators' proven experience in deciding such claims, leaves little basis for invalidating an arbitration agreement on the basis that the arbitrator is unqualified to decide these issues. The argument that arbitrators are qualified to decide statutory disputes is especially compelling in the employment discrimination context, where cases most often turn on factual not legal issues.llo Thus, in employment cases, the effect on the underlying dispute of an arbitrator's misunderstanding of the statute is minimized. In addition, judicial review of the legal issues is always possible.ll

III. ARE "GARDNER-DENVER AGREEMENTS" INVALID BECAUSE "GILMER AGREEMENTS" REMAIN SUSPECT?

In evaluating the Gilmer decision, commentators have found disturbing the Court's emphatic rejection of Gilmer's argument that the existence of unequal bargaining power between an employee and an employer should render that agreement invalid."2 The Court stated that while validity of consent can be examined on a case-by-case basis, only agreements that are the result of fraud or overwhelming economic advantage would be unenforceable.'l3 Despite the Gilmer Court's forceful statement on the matter, the question of whether unequal bargaining power should invalidate an agreement to arbitrate statutory claims has remained unsettled.

The question is of continuing importance because of unceasing calls for Gilmer's reversal on the basis that predispute agreements to arbitrate are unfair to employees. The primary basis for arguing that "Gilmer agreements" should be rejected is that they are not entered into voluntarily. Voluntariness in this context seems to be a euphemism for unequal bargaining power. Legislators argue that the agreements are not voluntary because the employee confronted with the agreement has no ability to reject it or negotiate its terms.TM The employee's inability to make a legitimate choice stems from the fact that the agreement is presented on a take-it-or-leave-it basis and the employee has no bargaining power to challenge the employer's provision. As a result, the argument goes, the agreement is mandatory and should not be enforceable. 115

Thus, a question arises: if Gilmer is ultimately reversed by legislation, should that reversal have any effect on a proposed overruling of Gardner-Denver? Regardless of Gilmer's fate, Gardner-Denver should still be overruled. "Gilmer agreements" are objectionable primarily because the parties who agree to them have disparate negotiating incentives. Because the parties to a collective bargaining agreement have similar negotiating incentives, there is no basis for invalidating their agreement on the basis of unequal bargaining power.

In fact, it is this difference that actually provides greater reason to enforce "Gardner-Denver agreements" than "Gilmer agreements." Application of game theory to both the Gilmer and Gardner-Denver agreements highlights the important differences between a "Gilmer agreement" and a "Gardner-Denver agreement" and serves to explain why the perceived unfairness of the "Gilmer agreement" is not present in cases involving "Gardner-Denver agreements."

Arbitration of public law issues in a nonunion setting is troubling because the structural protections inherent in collective bargaining are not present.lls Unlike interactions in the collective bargaining context, in which both the employer and the union are regular participants in negotiation and arbitration, only the employer is a "repeat player"1" in individual employment arbitration. The employee, by contrast, is a one-shot player.lls An analysis of the interactions between one-shotters and repeat players demonstrates that repeat players have a distinct and systematic advantage in interactions with one-shot players.119

By contrast, the possibility of overreaching will rarely play a part in negotiation and other interactions between repeat players. In such interactions, external nonlegal interests, such as each party's interest in maintaining a smooth working relationship, together with equivalent negotiating power, provides the incentive not to overreach.

A. The Repeat Player Interacting With the One-Shot Player: Employer-Employee Negotiations

The repeat player's greater experience, expertise and sophistication in contract negotiation will provide it significant advantages in interactions with one-shot players. For instance, in the dispute resolution context, the repeat player is likely to have a much better understanding of the risks and benefits of various dispute resolution mechanisms. Through this understanding, the repeat player may be able to choose the dispute resolution mechanism that best favors both parties, or one that is more favorable to it.120 A one-shot player, by contrast, will be unable to evaluate intelligently the proposed clause, because of a lack of experience in dispute resolution and inadequate resources to investigate the benefits and drawbacks of the clause.

The repeat player may also enjoy significant benefits during the dispute resolution process. A greater understanding of the process and an ability to influence that process through repeated informal relations with the decisionmaker provides the repeat player notable advantages. Moreover, the repeat player's institutional memory will lead to more informed choices in selecting an arbitrator.121 The one-shot player will not have a similar ability to influence the arbitrator and cannot afford to keep track of different arbitrators' decisions.

Finally, the repeat player may benefit from the fact that oneshot players, such as employees, tend to value improperly the inclusion of an arbitration agreement in an employment contract. Employees suffer from judgmental bias as a result of their personal experiences;l22 that is, they systematically ignore or de-emphasize the likelihood that a low probability event will affect them because the event has not occurred in the past. In the employment context, this judgmental bias may cause an employee to misapprehend the risk that he will engage in litigation with his employer. This informational problem may lead the employee to demand lower wages and fewer benefits than if he was fully cognizant of the risks present in the proposed arbitral agreement.

In the context of a relationship between an employer and an employee, the employer's systematic advantage as a repeat player over the one-shot-player employee manifests as follows: Repeat players maintain their advantage during the negotiation of the arbitration agreement because of their greater experience and superior knowledge, as well as in the selection of the arbitrator and in the arbitration itself.

1. Arbitration agreement negotiation

If repeat players use contracts in their negotiations with oneshot players, they will attempt to maximize profits and benefits from economies of scale by using standardized forms presenting limited opportunity for negotiation of terms. When presented with a standardized agreement, a one-shot player can only attempt to gain concessions on the negotiable terms if he fully appreciates the disadvantages or costs arising from the nonnegotiable portions of the agreement. To appreciate the value of the nonnegotiable terms, the employee would need to read and understand the proposed agreement.l23 Yet the rational employee will not invest substantial resources in reading or analyzing a proposed agreement. Such behavior is rational because the expected benefits from undertaking such an investigation would be significantly outweighed by the costs associated with such investigation. The expected benefits of reading and understanding the agreement may be reduced even further if the nonnegotiable terms concern the consequences of unlikely occurrences and appear in small print and/or are defined using obscure language 124

Empirical evidence supports the theory that the rational employee does not expend his limited resources reading and analyzing terms other than wages or benefits. According to David Charny, employees typically learn about crucial issues such as dispute resolution mechanisms, job safety or compensation a substantial amount of time after beginning employment.l25

By contrast, a rational repeat player will have included the dispute resolution system of its choice in the employment agreement with a one-shot player because it will have developed an understanding of the different methods of dispute resolution available and identified which method affords the greatest benefit. The employer's greater understanding of the value of such clauses, together with its ability to maximize its surplus by determining which provisions should be included in the agreement, will enable the employer to structure the employment agreement in a way that furnishes it the most advantage.l26 Moreover, the rational employer's position is further enhanced by its ability to present the arbitration agreement on a take-it-or-leave-it basis. If an employee actually understands the arbitration provision and attempts to negotiate the elimination of the provision, the employer will simply refuse and make a job offer to someone else.

An analysis of arbitration demonstrates that the process of negotiation is not the only area where repeat players have the potential to obtain significant advantage.127 The employer's repeat player status also creates a systematic bias in its favor in the arbitration proceedings. The bias results from the employer's incentive to foster relations with the arbitrator and create a precedent system for tracking arbitration decisions.

2. Interaction with the arbitrator

An individual using arbitration to resolve disputes has the incentive to compile information about potential arbitrators and their past decisions and develop a relationship with those arbitrators. The former will allow better predictability of arbitral outcomes. The latter will potentially allow the individual to influence the outcome of the arbitration. The employer's position as a repeat player enables it to accomplish both of these goals. It makes economic sense for the repeat player to monitor arbitrators' decisions and acquire advance intelligence about each arbitrator because it is likely that it will use that information repeatedly in the future. Not surprisingly, it is common for large organizations and law firms that represent those organizations to keep databases containing extensive background information on each potential arbitrator, including how the arbitrator ruled in a number of cases, as well as the quality of his decision."

For the same reason, the repeat player will take the opportunity to develop facilitative informal relations with the arbitrator, investing resources in attending events or conferences where the arbitrator will be present in order to establish a friendly relationship with the arbitrator that may result in bias in favor of the employer in the future.l29 A one-shot player who devoted any substantial time and resources to obtaining information about arbitrators or developing relationships with them would, by contrast, be acting irrationally because he would never have use for the information again.l30

The structure of the current arbitrator selection system does not eliminate the one-shot employee's disadvantages. Quite to the contrary, the current system provides significant benefits to the employer, at the expense of the employee. The arbitrator is likely to feel pressure to find in favor of the permanent party, the employer, in most cases because industry members will more frequently appear before the arbitrator. In addition, in many employment arbitrations the employer pays the arbitrator's entire fee.l31 The sense that the employer "owns" the process as a result may influence the arbitrator's ultimate resolution of the case. An arbitrator who regularly finds in favor of complaining employees may be sure that the employer will be reluctant to rehire her in the future.

Thus, the employer maintains significant advantages over the employee in structuring and executing a dispute resolution clause.l32 On that basis, a persuasive argument can be made that predispute arbitration agreements between employers and employees should not be enforced.l33 Yet the Supreme Court firmly rejected this argument in Gilmer, holding that, as a general rule, statutory claims are subject to binding arbitration, at least outside the collective bargaining context. Despite the perceived unfairness such agreements generate, the arbitral agreements of the unrepresented will be enforced.

B. The Repeat Player Interacting With the Repeat Player: Union-Management Relations

As the D.C. Circuit recently acknowledged, arbitration in the collective bargaining context is not as troubling as arbitration between employers and individual employees because the structure of the negotiation and arbitral process does not confer benefits on one party at the expense of the other.l34 In unionmanagement relations, both parties are frequent participants, or "repeat players" in negotiation and arbitration.

1. Drafting the arbitration agreement

The dynamics of the relationship between two repeat players temper many of the defects present in the relationship between a repeat player and a one-shot player. In repeat player relationships, both actors should have similar experience and expertise in negotiation and dispute resolution. Economies of scale do not favor either party; nor does one party have a greater understanding of the dispute resolution process than the other. More importantly, in a transaction involving two repeat players, both parties will have an economic incentive to avoid self-serving behavior. In most instances nonlegal sanctions, such as the desire to maintain a profitable business relationship with the other party, induces the repeat player to keep its commitments. These nonlegal sanctions, together with the awareness that both parties have similar knowledge and access to information about negotiation and dispute resolution, motivate the drafting party to apportion fairly the agreement's surplus.

The drafting party creates the agreement with the knowledge that an experienced negotiator will review it. As a result, the party presenting the first draft is aware that drafting an unfair or oppressive agreement may result in the kind of ill will that might ultimately trigger the relationship's demise. Further, even if the self-serving behavior went undetected initially, the self-serving party would have difficulty dealing with the other party throughout the life of the agreement and would certainly face tough opposition in subsequent negotiations. Thus, in drafting an agreement with another repeat player, the drafter has the proper incentives both to draft an efficient contract, and to distribute equitably the economic benefits.

In game theory terms, the strategy that motivates a repeat player engaged in continued interactions with other repeat players to avoid overreaching is the game of "tit for tat."135 Using the "tit for tat" strategy, a party's optimal strategy is to begin by cooperating and continue to cooperate as long as one's opponent does. If one's opponent engages in an act of betrayal, the affected party should retaliate. This strategy discourages noncooperative behavior while permitting a pattern of mutual cooperation to develop. Thus, "tit for tat" is the best strategy in a repeat-move game involving repeat players.l36

Applying this theory to the union-management relationship yields predictable results. First, the union is a repository of information about both the benefits particular contractual terms provide and the history of the negotiating relationship. The employer has no incentive to propose a one-sided provision, because it is quite probable that the union will fully appreciate the disadvantages and costs associated with such a provision and will reject it. Unlike the one-shot player who would be irrational to invest substantial resources in reading and analyzing a proposed agreement, it is perfectly rational for the union to read and analyze each and every one of the provisions of a proposed agreement. The equal bargaining power present on both sides reduces the likelihood that either side will attempt to garner a disproportionate benefit from the agreement. Moreover, if an employer, for example, attempts to capture the surplus, the union, as a repeat player, is capable of retaliating against the breaching employer by playing "tit for tat" either during the life of the agreement or during the next round of negotiations.

Thus, both sides are conscious of the need to avoid overreaching. Yet on those occasions when overreaching does occur, repeat players prefer not to engage in formal litigation with each other. Because formal litigation might jeopardize their ongoing relationship, the parties prefer an informal dispute resolution method that is less disruptive and which elevates the interest in maintaining the relationship over the need for a "correct" resolution of the dispute. In most repeat player relationships the preferred method of dispute resolution is arbitration.

2. Dispute resolution between repeat players

A study of repeat player behavior establishes that arbitration is the dispute resolution mechanism of choice for a repeat player engaged in a dispute with another repeat player.137 According to Galanter, repeat players dealing with other repeat players have the expectation that they are dealing and will continue to deal with each other frequently.l38 Because both parties are interested in "continued mutually beneficial interaction," they will prefer to use informal controls to govern their relations. The use of informal controls is preferable because the potential loss of their continuing relationship outweighs any official remedy available.l39 Thus, as is quite well known in labor-management circles, unions and employers prefer a dispute resolution form "detached from official sanctions." 140

Arbitration provides just such detachment. Arbitration is typically conducted in private. Moreover, the parties have considerable freedom in designing the procedures governing their arbitration because no formal rules govern the proceedings. Because parties can customize the proceedings to suit their interests, arbitration also has the potential for providing an acceptable result at a low cost. As a result, arbitration limits wealth transfers between parties, enabling the parties to retain resources rather than expending them on lengthy litigation that does not produce income for either side. Moreover, for repeat players, it is irrelevant that errors may occur in determining the outcome of a particular dispute, as long as no systematic bias presents itself. Repeat players are aware that wealth effects should balance out over the long term.

As repeat players, both union and management also have an interest in avoiding the unsettling impact of constant litigation. Because the two players raise their interest in a continuous working relationship above the need for obtaining a correct answer to a dispute, they prefer arbitration to litigation. As Julius Getman noted: "[T]he feeling that awards are likely to be equalized over the long run and that erroneous awards can be dealt with through negotiation, all have contributed to the common labor-relations practice of routinely obeying awards, even those that the losing side considers erroneous.""4' The method for dealing with awards that are perceived to be unfair to the union or the employer is not reprisal or disobedience, both of which might cause negotiation difficulties in the future, but rather simply the decision not to hire the offending arbitrator again in the future.

Neither should the issue of judgmental bias present a problem in union-management interactions.142 Because the union is a professional negotiator, it will have engaged in frequent dispute resolution and will understand fully the needs of its constituents, the employees. Thus, the union is in a much better position, in fact the same position as the employer, to resist the judgmental biases that plague unrepresented employees. Thus, there is no reason to suspect that unions are at a disadvantage if they select arbitration as a means for resolving disputes. Rather, the contrary is true; courts should presume that arbitration protects the represented employee's rights, whether those rights are contractual or statutory.

IV. CONCLUSION

Predispute arbitration agreements between employers and employees are unquestionably enforceable following Gilmer v. Interstate/Johnson Lane Corporation. By contrast, the Supreme Court has been reluctant to enforce similar agreements in the unionized workplace, citing Alexander v. Gardner-Denver, for the proposition that such agreements are not enforceable.

Over the past twenty years, however, the Supreme Court and lower federal courts have articulated an increasingly favorable attitude toward the use of arbitration to resolve statutory disputes. The Court has also repeatedly announced that the arbitral forum is a perfectly adequate venue for the resolution of statutory rights. This new vision of arbitration, when considered in conjunction with the erosion of Gardner-Denver's foundation, mandates a re-evaluation and reversal of Gardner-Denver. A number of lower courts have already begun the movement, placing collectively-bargained agreements to resolve statutory disputes using external antidiscrimination laws on equal footing with "Gilmer agreements." While some differences between the two types of agreements remain, these are distinctions without a difference that should be acknowledged as such and rejected. Furthermore, even if Gilmer were ultimately overruled or reversed by legislation because the employee's agreement to arbitrate is viewed as involuntary and therefore unfair, GardnerDenver should still be overturned because "Gardner-Denver agreements" are unquestionably voluntary and the result of exactly the kind of vigorous, two-sided negotiation that eliminates questions of compulsion.

[Author Affiliation]

34 Visiting Associate Professor of Law, University of Oklahoma Law School. J.D. University of Chicago, 1990. Thanks to Douglas R. Cole and Stephen Ware for their helpful comments and Sean Gillen for his able research assistance.