The Supreme Court in Janus v. American Federation of State, County, and Municipal Employees, Council 31 upended public sector labor law by finding a novel First Amendment right of public employees to refuse to pay union fees and declaring unconstitutional scores of laws and thousands of labor contracts. This Article assesses the constraints on public sector labor law post-Janus, examines the variety of legislative responses, and proposes a path forward.
Janus makes it difficult to address the collective action problem facing all large groups. Although it is in the interest of every member of a group to engage in collective action to provide common goods, it is also in each individual’s interest to let others incur the costs of doing so. The Janus Court misstated the nature of the collective action problem when it said the problem was free-riding on union-negotiated benefits. The problem is that, without some way to require all who benefit to share the costs, unions will not negotiate effectively for the benefits in the first place, so there will be no common goods to free ride on.
Many proposals for ameliorating the collective action problem exacerbated by Janus continue unions’ financial solvency in the short-term but sacrifice unions’ fundamental nature as membership organizations governed by and for workers. Some adopt a form of members-only representation, thus abandoning the principles of majority and exclusive representation. Others have government employers subsidize the cost of union representation. And yet others treat union fees like health insurance: subject to an annual open-enrollment period. But four major public sector unions have condemned many of these approaches, even as legislatures have considered or enacted them. Close analysis of the unintended effects of these approaches to the collective action problem shows why they are problematic.
Returning to the economic theory of groups and public goods, the Article assesses legislation that seeks to give public employee unions some of the attributes of small groups, in which a mix of social norms and individual benefits provide the incentives for individuals to incur the costs of providing public goods. The Article concludes by explaining why the options we propose could survive the inevitable post-Janus legal challenges and enable unions to be majoritarian democratic institutions that are accountable to those whom they represent.
Public sector unions and the workers they represent face dramatic challenges and opportunities. On the one hand, teachers across the country, including in states and cities that have little or no public sector bargaining, successfully went on strike in 2018 and 2019 to protest steady declines in education spending and endemic teacher and staff shortages caused by years of tax cuts. On the other, the Supreme Court in Janus v. American Federation of State, County, and Municipal Employees, Council 31, upended public sector labor law by creating a novel First Amendment right to refuse to pay union fees. In so doing, the Court declared unconstitutional scores of laws and thousands of contracts in 22 states and the District of Columbia. These cataclysmic events suggest both that collective action remains essential to an adequately trained and compensated public sector workforce and that the Supreme Court made it much harder for states to create collective institutions. This Article assesses the constraints on public sector labor law, examines the post-Janus legislative responses that have emerged, and proposes a path forward.
The constraints are considerable because Janus makes it difficult to address the collective action problem facing all large groups. Economists since Mancur Olson have known that, although it is in the interest of every member of a group to engage in collective action to provide common goods (such as roads, schools, firefighters, or labor contracts guaranteeing fair wages and safe working conditions), it is also in the each individual’s interest to let others incur the costs of doing so. Olson explained: “If the members of a large group rationally seek to maximize their personal welfare, they will not act to advance their common or group objectives” absent either compulsion or incentives that will benefit the members apart from the group benefits. Rational individuals realize that their individual contributions will not likely have any significant impact on advancing the group effort to secure common goods and, therefore, decide not to incur the costs. As a result, large groups will not form effective organizations and all will be worse off. The Janus Court misstated the nature of the collective action problem when it said the problem was free-riding on union benefits. The problem is not that some employees will free ride on the benefits the union secures for all workers. The problem is that unions will not form and, if they do, they will not negotiate effectively for the benefits in the first place. There will be no common goods to free ride on.
This Article explores solutions to the collective action problem that are consistent with Janus and that enable unions to be majoritarian democratic institutions that are accountable to those whom they represent. For over a century, union relationships with employers and with unionized workers have operated on the model of electoral democracy. A union elected by a majority represents all workers in the unit, just as a legislative or executive official represents everyone. But a union, unlike a political leader, owes a duty of fair representation to every employee in the unit and cannot act arbitrarily or discriminatorily in deciding whose interests to prioritize. The duty applies whether the union is negotiating a contract or enforcing it. Democracy is foundational to everything unions do, from the way they govern their internal affairs to their efforts on behalf of workers to create workplace democracy to their role in civil society. Their responsibility to respect the interests and rights of minorities is what makes unions different from political leaders and what has made the contemporary fight over how unions fund their work so significant. Janus requires unions to continue to act like governments—indeed, to provide better representation to the minority than governments do by preventing elected leadership from advancing the interests of supporters over those of non-supporters—but denies them the tool that governments and all viable large organizations have, as the Framers put it, to “promote the general Welfare.”
Unions therefore face a choice among three legislative paths forward. First, they can abandon majoritarianism. Second, they can risk their independence by accepting funding from the government. Or, third, they can enhance solidarity through a mixture of incentives and organizing. Several pieces of proposed legislation or proposals made in the academic and popular press advocate either the first or the second. The majority opinion in Janus suggested the first: unions could abandon majoritarian representation. Along those same lines, two scholars have suggested that unions abandon exclusive representation and their focus on negotiating conditions of employment; instead, unions should embrace their political identities and draw funding from employees and others who support their agendas. One scholar has proposed the second path: employers would pay union dues directly to the union without the money ever being credited to the employees’ paycheck, thus risking union independence. However, four of the national organizations of public employee unions joined in a statement in July 2018 rejecting both of these proposals. AFSCME, AFT, NEA, and SEIU oppose members-only bargaining or any other incursion on the principles of majority-rule representation in both contract negotiation and contract enforcement. They unite in opposing the representation of employees by attorneys or other representatives not appointed by the union. They also unanimously oppose creating fee-for-service arrangements for nonmembers, including the system of per-capita payment by the government to the union.
Public sector unions have chosen a third way. Their multi-pronged strategy to ameliorate the collective action problem refuses to pursue short-term union financial solvency at the expense of sacrificing the fundamental nature of unions as membership organizations governed by and for workers. This Article analyzes the approaches taken by legislatures to address union security, offers empirical evidence of the effects of those approaches to the collective action problem, and explains why some proposals might be harmful to the interests of public sector employees.
Part I explains the union collective action problem that Janus misunderstood and shows the difficulty of solving it in a way that is both consistent with Janus and enables unions to remain accountable majoritarian organizations. Part II assesses three approaches that have been proposed by scholars and legislatures post-Janus: members-only representation, cost-shifting to the employer, and an annual open enrollment comparable to those found in benefits. Part III considers various proposals that would more effectively address the collective action problem. First, it considers possibly transplanting the way that the National Labor Relations Act (NLRA) and some states have treated religious objectors in fair share fee states. It finds some merit in that approach but explains why the limited analogy to the treatment of religious objectors should be coupled with measures to enhance solidarity. Part III then returns to the economic theory of groups and public goods. It assesses approaches that seek to give public employee unions some of the attributes of small groups in which a mix of social norms and individual benefits provide the incentives for individuals to incur the costs of providing public goods.
The Article provides tools for legislators and policymakers in states and municipalities across the United States to revise their public sector labor laws in ways that address the Janus Court’s concern about compulsion and protect the right of the majority to form organizations capable of delivering common goods.
Like other large groups, union employees face a collective action problem. Indeed, theirs is the paradigmatic collective action problem, the one used by economists to explain the theory of collective action. The problem itself is as follows: it is in the interest of every member of the group to engage in collective action to improve wages and working conditions. It is also in the interest of every member of the group to let others incur the costs of engaging in the collective action. But an economically rational worker would choose not to join the union and pay dues because the worker’s own dues, or lack thereof, are unlikely to have an appreciable effect on the union. So if every person acts rationally, as an individual, by free-riding on the efforts of others, everyone would be worse off because no one will get the benefit of collective action. Union security provisions address this collective action problem by requiring everyone to support the collective representative. This prevents individuals’ rational decisions to refrain from joining the collective and promotes the economically optimal collective action.
Mancur Olson explained the collective action problem at length in his seminal 1965 book, The Logic of Collective Action: Public Goods and the Theory of Groups. Large organizations form to further the interests of their members: labor unions to improve working conditions, corporations to obtain a favorable return on stockholders’ investments, and farm organizations to improve the situation of farmers through grain or dairy co-ops, government subsidies, or trade policies. These organizations exist “primarily for the common interests of their members.”  In contrast, Olson explained, “personal or individual interests can be advanced, and usually advanced most efficiently, by individual, unorganized action.” One essential characteristic of a common good is that no one in the group is excluded from the benefit of it. Fire protection is a common good because no one can be protected from fire if individuals can opt out. So, too, job protections for workers: all workers benefit from a system that curbs arbitrary supervisory authority, that provides health insurance to the group, and that ensures adequate safety protection. But the rational individual knows that “his own efforts will not have a noticeable effect on the situation of his organization, and he can enjoy any improvements brought about by others whether or not he has worked in support of his organization.”
Olson’s crucial theoretical insight is that organizations will never form, and the individuals who comprise them will never get the benefit of the common good, if there is no mechanism mandating the sharing of the costs. This is why, for as long as there have been governments, there have been systems of taxation, rules regulating fire safety, road safety, sanitation, and other such systems that share the cost among all those who benefit.
The compulsory aspects of group governance, as the framers of the United States Constitution put it in the preamble, are designed to “insure domestic Tranquility, provide for the common defence, [and] promote the general Welfare.” Long ago, courts prevented unions from solving their collective action problem through mandatory membership. The most unions could do was require represented workers to share in the cost through payment of what was known as an “agency” or “fair share” fee. The Court eliminated fair share fees in Janus; thus, public employee unions are now prohibited from charging nonmember employees anything for the services the union provides.
The Janus Court misstated the nature of the collective action problem. The problem is not simply that some employees will free ride on the benefits the union secures for all workers. The problem is that it is an economically rational decision for every employee to refrain from joining the union unless assured that everyone else will also join. Consequently, unless the collective action problem is dealt with, the union will be unable to negotiate for the benefits in the first place. As Justice Kagan explained in her dissent, citing empirical studies proving the Olson thesis,
Without a fair-share agreement, the class of union non-members spirals upward. Employees (including those who love the union) realize that they can get the same benefits even if they let their memberships expire. And as more and more stop paying dues, those left must take up the financial slack (and anyway, begin to feel like suckers)—so they too quit the union. And when the vicious cycle finally ends, chances are that the union will lack the resources to effectively perform the responsibilities of an exclusive representative—or, in the worst case, to perform them at all. The result is to frustrate the interests of every government entity that thinks a strong exclusive-representation scheme will promote stable labor relations.
The Court’s responses to the collective action argument misunderstand economic theory. The majority assumed that unions will exist, and will effectively negotiate collective benefits, even if only some workers incur the cost of maintaining the union and the contract. They also assumed that fair share fees serve just one interest: preventing nonmembers from free-riding on the existing benefits. But if unions are unable to compel support, there will be no common benefits in the first place.
The Court in Janus addressed several arguments about why fair share fees are necessary. First, the majority considered the argument that fees are necessary to secure labor peace by avoiding “dissension within the work force,” “conflicting demands from different unions,” and the “[c]onfusion [that] would ensue if the employer entered into and attempted to enforce two or more agreements specifying different terms and conditions of employment.” To this, the Court responded with evidence of unionization in the federal service and the Postal Service, both of which prohibit agency fees. But this response does not logically follow. Both the Federal Service Labor Management Relations Statute (FSLMRS) and the Postal Reorganization Act, as the Court recognized, make a union chosen by the majority the exclusive representative of all. Neither statute allows members-only unions nor members-only contracts. Moreover, under both statutes, as discussed below, costs that the union would otherwise incur are shifted to the employer.
We can assume the majority did not mean that members-only bargaining would address the collective action problem; instead, the majority probably meant that the experiences of federal government employment and the Postal Service show that agency fees are unnecessary. However, the empirical evidence does not support such a generalization. As noted, both the FSLMRS and the labor law governing the Postal Service allow unions to negotiate contracts under which the union conducts most of its representational activities on paid time. This essentially shifts the cost of securing and administering the common goods to the employer. As we explain more fully below, this makes unions less accountable to their membership in ways that should concern both the Janus majority and union supporters. It also fails to address the collective action problem.
There is considerable empirical evidence that the collective action problem is real. Olson drew support for his theory from evidence of NLRB-supervised elections conducted when the Taft-Hartley Act required any compulsory union membership provision of a collective bargaining agreement to win a majority vote of the represented employees. This provision of the statute, which was in effect only from 1947 to 1951, required a majority vote of all the represented employees, not just a majority of votes cast. In the four years the statute was in effect, unions won 97 percent of the elections and nearly 45,000 union shops were authorized. Yet, Olson pointed out, union members tended not to attend union meetings; the attendance figures he cited showed that fewer than ten percent of members attended meetings even though nearly one hundred percent supported compulsory union membership.
A large study (n=11,668) of the effects of union security provisions on state and local government workers published in 1993 demonstrated that union security provisions had a significant positive effect on union status and wages. The study suggested that compulsory union fee contracts produced higher wages and greater union density, and that “group norms are important in determining the propensity of covered workers to join unions.”
A more recent piece of evidence comes from Indiana. There, public school teachers have statutory collective bargaining rights. Under a 2017 amendment to Indiana’s teacher collective bargaining statute, exclusive bargaining representatives are required to annually certify to employers the percentage of the bargaining unit who are members of the union. If less than a majority of bargaining unit members are members of the union, the employees are notified of their right to change representatives and their right to decertify their union. The 2018 report on teacher exclusive bargaining representatives shows membership density ranging from 100% to 12% with many below 50%. However, there is no evidence that teachers have ever decertified their union, at least not without replacing the union with a different one. This suggests that union-represented employees find it economically rational to refrain from joining the union, but that they nevertheless have no desire to eliminate union representation. In other words, Indiana’s experience is the latest evidence supporting the Olson thesis.
One last point should be noted about whether fair share fees are necessary for unions to be effective representatives. The Janus majority asserted that “millions of public employees in the 28 States that have laws generally prohibiting agency fees are represented by unions that serve as the exclusive representatives of all the employees.”  According to the Court, this makes it “undeniable” that agency fees are unnecessary to serve the legitimate interest in stable labor-management relations.  However, the majority failed to note that union locals in those twenty-eight states have long been subsidized by the national federation, which collected dues from members in the states and territories that permitted fair share fees. Once those subsidies disappear—because every state is now a right-to-work state for public sector employment—economic theory suggests that the quality of representation will decline everywhere.
The remainder of the majority’s arguments concerned free-riding on union benefits. As noted above, the majority assumed that unions will survive to negotiate benefits and that the problem is free-riding on existing benefits rather than the inability of the organization to form and secure the benefits in the first place. But assuming arguendo that the concern is free-riding, the Court’s account of this issue and its responses to the arguments against the necessity of fair share fees are problematic.
The Court responded to arguments about free-riding on benefits in six ways. First, the Court observed that “[m]any private groups speak out with the objective of obtaining government action that will have the effect of benefiting nonmembers.”  The Court’s examples included unnamed groups representing senior citizens, veterans, and physicians. Mancur Olson dealt with this argument at length in his analysis of collective action and public goods; indeed, he devoted an entire chapter to it. As he showed, “the large economic groups that are organized do have one common characteristic which distinguishes them from those large economic groups that are not. . . . [These organized groups] obtain their strength and support because they perform some function in addition to lobbying for collective goods.” They either have the legal or de facto ability to compel membership, or they offer inducements to members to join, as is the case of the American Medical Association, AARP, American Bar Association, and a host of others that Olson studied. For example, Olson demonstrated that the Chamber of Commerce provides considerable individualized benefits to those large business organizations that effectively control it. These benefits perhaps come at the expense of other business groups on whose behalf they claim to speak but whose interests they may not serve. As Olson explained, there are several groups that exist on an entirely voluntary basis but that advocate for common goods. Perhaps the Janus majority’s example of veterans’ groups fits in that category. Others obviously exist: the National Rifle Association and the American Civil Liberties Union, for instance. But none of these organizations, unlike labor unions, have a statutory obligation to represent the interests of all people on whose behalf they claim to speak.
Second, the Court said that unions gain benefits from being designated the exclusive representative; these benefits include the right to bargain with the employer, the right to obtain information about employees, and the right to have dues deducted from payroll. The Court asserted that “[t]hese benefits greatly outweigh any extra burden imposed by the duty of providing fair representation for nonmembers.” But it cited nothing in support of that proposition—nor could it. As Justice Kagan pointed out in her dissent, the “key question” is “whether unions without agency fees will be able to (not whether they will want to) carry on as an effective exclusive representative.” While employees benefit from being able to present a united front to the employer in contract negotiations, the greater bargaining leverage does not itself pay for the lawyers, accountants, economists, researchers, organizers, and other union staff and consultants who are necessary for unions to effectively negotiate and administer contracts.
The higher wages that unions gain by presenting a united front does not go to the union staff; it goes to the employees. And when those who choose to join the union are required to pay higher dues to secure the same contractual wage premium that they would pay if all their coworkers paid, they are effectively subsidizing their nonunion co-workers. That is not a benefit to the dues payers; it is a loss as compared to the situation that would exist if all were required to share the cost of getting the wage premium. Moreover, the nonpayers enjoy the same union wage premium as the payers, but they actually get a higher premium because they do not pay dues. This unfair advantage creates the incentive to resign from the union. And once everyone responds to the incentive, the union collapses and the wage premium disappears. This may explain why corporate funders are behind the Freedom Foundation’s campaign to persuade workers to resign their union membership. The campaign appeals to individuals’ self-interest—resign your union membership and give yourself a raise. However, companies only seek to eliminate unions to lower labor costs.
Third, the Court suggested that it is unnecessary to provide any financial incentive for the union to represent nonmembers because the duty of fair representation prohibits the union from discriminating against nonmembers in contract negotiation or contract administration. As with the Court’s other arguments, this assumes the union will exist and will succeed in negotiating the union wage premium. The only question is whether the nonpayers will enjoy the benefits without incurring the costs.
Inconsistent with this third argument, the Court acknowledged in its fourth argument that it is unfair for nonpayers to enjoy the same benefits as those who pay union dues. To address the unfairness, the Court proposed that unions could refuse to represent nonmembers in grievance proceedings or could require nonmembers to pay for grievance handling. However, elsewhere in the opinion, the majority cast doubt on the viability of denying representation to nonmembers. Specifically, it noted that protection of nonmembers’ “interests is placed in the hands of the union, and if the union were free to disregard or even work against those interests, these employees would be wholly unprotected.”
Fifth, the Court asserted that union members benefit from union grievance handling on behalf of nonmembers because “the resolution of one employee’s grievance can affect others.” This is a different way of saying the Court’s second argument: that dues-paying members gain benefits from the union being the exclusive representative, except here, the Court is considering who should pay the costs of contract administration as opposed to contract negotiation. The answer is still the same. The benefits gained by union-represented employees are available only because the union bears the significant costs of obtaining them. These costs include lawyers’ fees and the union’s share of the arbitrator’s fee, the arbitrator-appointing agency’s fee, the court reporter’s fee, and the rental of a hearing room. It is unfair to require some employees to pay these costs for others.
Finally, the Court in a footnote referenced states that provide employees with religious objections to unionization the option to contribute an amount equal to unions fees to approved nonprofits. Later we address the possible expansion of the treatment of religious objectors to all who object to union membership. We show why it partly ameliorates the collective action problem but would be difficult to administer without raising the same constitutional problem as fair share fees.
Thus, Janus leaves unions in a quandary. The Court has offered no solution to the actual collective action problem that unions face (as opposed to the more limited problem of free-riding that the Court addressed). To the extent that unions respond to the collective action problem by abandoning their effort to seek common goods and instead pursue only the narrow interests of their actual members, they abandon their long heritage as majoritarian democratic institutions. Alternatively, to the extent they address the collective action problem by shifting the costs of representation onto the employer (and thereby onto taxpayers or onto all workers in the form of lower wages), they risk compromising their independence. That leaves only two ways to address the collective action problem. One is to assure union supporters that most of their co-workers will also join. A second is for unions to act more like small groups than large ones by undertaking actions that benefit more workers as individuals while also creating social norms that make shirking unacceptable. We explore these post-Janus options in Parts II and III below.
Three alternative solutions to the collective action problem have been considered. All three have been implemented in the United States to one degree or another, and there is by now evidence of their effects. Below, we describe each solution and what is known about their effects. Although each has some promise as a way to enable unions to continue to provide services to workers, to public agency employers, and to the public, none solves the collective action problem without fundamentally altering the nature of unions as majoritarian bargaining representatives.
The first approach is to abandon the majoritarian nature of unions by giving a union the option to represent only its members, at least to some extent. Members-only representation has existed or does exist in two forms. In one form, the union bargains only on behalf of its members and the contract it negotiates covers only its members. In the second, a union bargains on behalf of all employees in the unit, and the contract covers all, but employees who refuse to financially support the union cannot obtain individual representation services in enforcing the contract. We discuss each of these options below.
Members-only representation allows employees to choose another union or no union at all. Such representation addresses the Janus concern that majority unionism compels those who oppose any or all union speech or association to be free of union representation without restricting the freedom of those employees and employers who favor collective bargaining. Such a regime requires statutory changes to the union selection process, to the employer’s duty to bargain, to the union’s duty of fair representation, and to the law defining what is prohibited discrimination by a union and an employer. In particular, it would be necessary to amend current law to clarify whether it is unlawful discrimination on the basis of union membership if the employer negotiates different terms with the union than with non-union employees, or different terms with different unions.
Experience with members-only representation shows it is fraught with problems for both employers and employees. Prior to 1976, members-only representation existed in California primary and secondary schools under the Winton Act. Neither school districts nor teachers preferred the members-only system. It created administrative difficulties for districts, dissension among employees, and perceptions that terms of employment were unfairly different among teachers in the same district. It was replaced by majority rule exclusive representation.
In Tennessee, members-only representation, similar to that under the Winton Act, has been the only form of union negotiation for school teachers since 2011. Each entity that receives more than 15 percent of the votes gets proportional representation on a “collaborative conferencing” committee that meets with the school district. This system allows the district to run out the clock on the time for contract agreement so that it can set terms unilaterally. This means that Tennessee does not have true collective bargaining even on a members-only basis. Nevertheless, the Tennessee experience has not improved stability of labor relations, teacher working conditions, or the quality of education.
Wisconsin retains exclusive representation but with strict limits on the topics of negotiation. These limits are so strict that they have led to an increase in individual bargaining. Some school districts began to offer signing bonuses and higher salaries to attract successful teachers from other districts. Some districts have even paid the “resignation fees” that districts charge teachers who leave mid-year. A leader of the Wisconsin Association of School Boards said, “It’s like the world of baseball economics hitting public schools,” as teachers with valuable skills or excellent performance evaluations seek out districts with more money. This exacerbates inequalities among schools, making it much harder for states to improve education in the schools that most need to recruit and retain excellent staff. In short, the increased freedom of successful employees in a members-only regime, or any other that allows individual contracting, comes at a considerable cost for employers, employees, and consumers of public services.
An alternative form of members-only representation is majority bargaining but members-only grievance handling. This is the approach that Janus suggested when the Court said “[i]ndividual nonmembers could be required to pay for that service or could be denied union representation altogether.” The Court also noted that, in states that accommodated religious objections to paying an agency fee, religious objectors who had paid no fees could be charged for the reasonable cost of union representation in grievance handling. The Court explained, apparently with favor, that “[t]his more tailored alternative, if applied to other objectors, would prevent free ridership while imposing a lesser burden on First Amendment rights.”
Some unions have proposed a post-Janus legal order in which unions would be the exclusive representative of all employees in the bargaining unit for purposes of negotiating the contract but would represent only their members for purposes of administering the contract. International Union of Operating Engineers Local 150 filed a lawsuit alleging that its duty to represent nonmembers, who pay nothing toward its representation, violates the union’s rights under the First Amendment. The suit alleges that “[i]f . . . it violates the First Amendment right of a non-member to be compelled to pay fees to the union that is required by law to provide representation and services, it equally violates the rights of the union and its members to require them to use their money to speak on behalf of the non-member.” According to Local 150,”[f]orcing unions to advocate on behalf of non-members who object to the very reasons they exist is a severe violation of unions’ First Amendment rights to association.” Therefore, compelling unions to expend funds to represent nonmembers drains money that the unions would otherwise spend on First Amendment protected activity.
There is a certain logic to the distinction Local 150 draws between contract negotiation and contract administration. Negotiation of the contract is something the union would do regardless of whether it is negotiating on behalf of the entire bargaining unit or only on behalf of its members. The union resources that are diverted from members to nonmembers are less obvious in the case of contract negotiation than in the case of grievance handling. On the other hand, the time and resources required to negotiate on behalf of a diverse group of members and nonmembers are greater than those necessary for a smaller group of members only. It is more time-consuming to survey, involve, and address the perspectives and needs of a large bargaining unit. And the contract that applies to everyone may look different from one that would apply only to members.
Although the Janus majority focused on grievance processing, representation of individual employees during the term of a collective bargaining agreement can encompass a good deal more. For example, public employees protected by civil service statutes or collective bargaining agreements that require cause for discharge have a constitutional due process right to a hearing before they may be discharged. It is common for unions to represent employees in these pre-disciplinary hearings. Similarly, in many states, employees questioned in investigatory interviews have a right to union representation if they reasonably believe that the interview could lead to disciplinary action. In New York, this right is expressly provided for by statute. In other states, the right has been found to be part of the general right to engage in concerted activity for mutual aid and protection, with courts and labor boards following the analogous law under the National Labor Relations Act upheld by the Supreme Court in NLRB v. J. Weingarten, Inc.
In the private sector under the NLRA, unions may not refuse to represent nonmembers in contract administration or require nonmembers to pay the costs of their representation. In the public sector, some states allow unions to refuse to represent non-payers or to charge them for representation services. These jurisdictions prohibited agency fees by statute before Janus made fees unconstitutional and conditioned the union’s refusal to represent nonmembers on the nonmembers’ having the right to process their own grievances without union representation. For example, the Florida public sector collective bargaining statute expressly provides: “All public employees shall have the right to a fair and equitable grievance procedure administered without regard to membership or nonmembership in any organization, except that certified employee organizations shall not be required to process grievances for employees who are not members of the organization.”
Similarly, Nebraska’s statute provides that “[a]ny employee may choose his or her own representative in any grievance or legal action regardless of whether or not an exclusive collective-bargaining agent has been certified.” The statute goes on to state, however: “If an employee who is not a member of the labor organization chooses to have legal representation from the labor organization in any grievance or legal action, such employee shall reimburse the labor organization for his or her pro rata share of the actual legal fees and court costs incurred by the labor organization in representing the employee in such grievance or legal action.” 
In Nevada, court and administrative agency decisions allow unions to charge nonmembers for contract enforcement services. In Cone v. Nevada Service Employees Union, the Nevada Supreme Court considered SEIU’s policy that allowed nonmembers to hire their own attorneys to handle their grievances. The policy required any nonmember who wanted union representation in a grievance to pay for it according to a stated fee schedule. The court held that the union did not violate Nevada’s right-to-work statute because paying the service fee was not a condition of employment; it was merely a condition of having the union represent the nonmember in a grievance. The court also rejected the non-payers’ argument that charging for individual services violated the union’s statutory obligation as an exclusive representative, opining that the exclusive representative’s duty of fair representation does not require a union “to provide all services for free.” In addition, the court observed that Nevada law “explicitly authorizes a nonunion member to act on his own behalf ‘with respect to any condition of his employment.’” The court reasoned that the individual right to refuse to pay fees, and to forego union representation in a grievance, allows the union to charge if the employee requests union representation in the grievance proceeding. In line with Cone, Nevada’s public sector labor board has held that where a member opts to proceed through the grievance procedure with her own attorney, the union may decline to represent her.
In anticipation of the Court’s ruling in Janus, New York decided to follow the example of Florida, Nebraska, and Nevada and now allows a limited form of members-only representation. New York amended its public sector collective bargaining statute, commonly known as the Taylor Law, to allow unions to decline to represent nonmembers in discipline grievance and administrative proceedings where nonmembers have the right to represent themselves. The amendment also allows unions to decline to represent nonmembers in disciplinary interviews. Similarly, Rhode Island amended its statutes to allow unions representing police, firefighters and teachers to decline grievance representation of employees who have elected not to maintain membership in the union for at least ninety days before the event giving rise to the grievance. Nonmembers have the right to pursue their grievances at their own expense. However, the union has a right to be present at any grievance or arbitration hearing, and any grievance resolution must be consistent with the collective bargaining agreement. The common theme in Florida, Nebraska, Nevada, New York and Rhode Island is that where employees can process their own grievances with their own counsel or other representative, the union may refuse to represent nonmembers or condition such representation on paying a fee.
Allowing unions to decline to represent nonmembers in contract administration or to condition their representation on the nonmember paying for it eliminates the free rider problem with respect to contract administration but does not solve the collective action problem. Employees who do not pay dues cannot rely on the union to provide representation during the term of a collective bargaining agreement for free. At first blush, this would appear to go a long way to dealing with the collective action problem as well. The heart of the collective action problem is that most services that the union provides are collective goods. It is economically rational for a union supporter to decide not to join the union so long as enough others do. Allowing unions to decline to represent nonmembers in grievance processing and other matters of contract administration converts a collective good into an exclusive good. However, even with unions’ ability to restrict representation in grievance and contract administration to members only, it remains an economically rational decision for a union supporter to decline to join. There is evidence of this from Florida and Nebraska.
In Florida, there are many teacher union locals where fewer than 50 percent of the bargaining unit are union members. Many nonmembers support representation by their union. An amendment to Florida’s public sector labor relations statute that took effect July 1, 2018 requires teacher unions to certify the percentage of bargaining unit employees who are union members, and to submit to recertification elections if fewer than half of the bargaining unit are dues-paying members. A group of Florida teachers who are represented by an exclusive bargaining representative but are not members of the union have sued alleging that the new requirement violates the right-to-work provision of the Florida Constitution by coercing those who support union representation into becoming members of the union.
There is even more compelling evidence of the serious collective action problem from other Florida public sector employees. Patrick Wright, Vice President for Legal Affairs of the Mackinac Center for Public Policy, a conservative think tank that supports right-to-work legislation and litigation, gathered data showing that 74,266 Florida state government employees are covered by collective bargaining agreements but only 7,689 of them, or 10.4%, have authorized payment of membership dues by payroll deduction. There may be additional state employees who are union members but pay their dues other than by payroll deduction. Even assuming that the number of members paying dues directly rather than through payroll deduction equals the number paying by payroll deduction, a very generous assumption, membership density among state employees is very low. This is consistent with the observation of a prominent Florida public sector labor lawyer, who said that membership density across the board, except for law enforcement and firefighters, is very low in Florida and that a union that has 30 percent membership is doing well.
In Nebraska, there is evidence that teacher unions enjoy very high membership density because teachers want to protect against having to pay for union grievance representation if they do not join. An attorney who represents the Nebraska State Education Association (NSEA) explained that Nebraska teachers have come to regard legal assistance as “a privilege of membership” in the union, and not a right of all teachers in the state. The lawyer explained that NSEA members are entitled to non-lawyer representation in grievances and lawyer representation in teacher dismissal proceedings, professional practices disciplinary proceedings, or judicial proceedings that involve enforcing statutory or constitutional rights. The lawyer concluded:
I do believe there is a perception that membership in the NSEA offers value in the area of job security, whether in the form of staff representation in grievances and performance issues, or legal representation in dismissals. The cost of legal fees to go it alone in a dismissal is intimidating and is likely why the statute has never been used to access legal assistance by nonmembers.
Thus, among Nebraska teachers, the exclusive membership benefit is access to legal representation, which is used primarily outside of the collective bargaining agreement’s grievance procedure.
Moving beyond teachers, the evidence from Nebraska is consistent with the evidence from Florida: members-only unionism does not solve the collective action problem even if it addresses the free-rider problem. Wright’s study found that only 1,573, or 15.4 percent, of the 10,247 Nebraska state employees covered by collective bargaining agreements pay union dues by payroll deduction.
The evidence of low membership density among state employees in Florida and Nebraska is not surprising. As we discuss later, collective action problems are greatest in very large groups that are made up of a wide diversity of constituents. State government bargaining units tend to be state-wide in scope, cut across state agency lines, and cover a large diversity of occupations and job titles. State employee unions are the epitome of the types of organizations beset by collective action problems.
Members-only grievance representation does not solve the collective-action problem; even when unions exclude nonmembers from union grievance representation, it likely remains economically rational for union supporters not to become union members. Essentially, union membership provides insurance in case one needs grievance representation. The employer decision that poses the greatest threat to an employee’s economic security is the discipline or discharge of an employee. But most employees believe the likelihood that they will need grievance representation is quite low because they do not believe there is a significant risk that they will be disciplined or fired. Presumably, this is particularly true for those employees who receive satisfactory performance evaluations. As a result, employees believe it is economically rational to forgo paying union fees and, in effect, self-insure against the small risk of needing grievance representation.
Beyond offering at best limited help in dealing with the collective action problem, members-only grievance representation has another potential drawback: the line between contract negotiation and contract administration is very blurry. The contract language in these agreements is often too general, leaving much to be refined in subsequent contract administration. Consequently, the grievance and arbitration procedures are regarded generally as a continuation of the collective bargaining process.
Having a grievance and arbitration procedure enables parties to conclude contract negotiations with general language that will apply to a myriad of situations which defy specification in the collective bargaining agreement and to agree on language even though they disagree on what that language means. Parties defer the refinement of the general language into specific contractual rights to case-by-case negotiation through the grievance procedure with the understanding that if they cannot agree on what their contract means in a given situation they agree to be bound by the interpretation of their mutually selected arbitrator.
For example, it is common for parties to include a “relative ability clause,” which provides that in filling vacancies, the employer will select the most qualified applicant, but where qualifications are relatively equal seniority will govern. The parties recognize that in the abstract they disagree over what “relatively equal” means. The Union interprets the term broadly while the employer interprets it narrowly. But they defer to the grievance procedure if a specific issue with this interpretation arises.
By leaving nonmembers to process their own grievances, a union may also limit its ability to protect workers’ collective interests. For example, in a grievance over a promotion denial, the individual grievant’s interest is focused on receiving the promotion. However, the union’s collective interest is focused on finding the right mix between seniority and management-assessed qualifications. The Court recognized this conflict in Janus: “Representation of nonmembers furthers the union’s interest in keeping control of the administration of the collective-bargaining agreement, since the resolution of one employee’s grievance can affect others. And when a union controls the grievance process, it may, as a practical matter, effectively subordinate ‘the interests of [an] individual employee . . . to the collective interests of all employees in the bargaining unit.’” Even in Florida, unions will represent nonmembers in grievance processing without charge when a grievance affects more employees than just the grievant.
For all these reasons, members-only grievance representation is inconsistent with the underlying values of exclusive representation. The basic concept of exclusive representation is expressed in the old union adage that “an injury to one is an injury to all.” A grievance asserting one employee’s right to fair treatment asserts the right to fair treatment for all employees.
Exclusive representation, which evolved in the U.S. and Canada in reaction to an anti-union business climate in which workers lacked national political power, recognizes that the bargaining power of the collective is greater than the bargaining power of the individual. and comes from the elimination of competition among workers, at least with respect to particular employers, that drives wages and working conditions down. In other words, the legal concept of exclusive representation implements the basic concept of worker solidarity that is at the heart of the union movement.
As a result, members-only grievance representation presents a tradeoff between solving part of the free-rider problem, marginally solving the collective action problem, and maintaining worker solidarity. Evaluating this tradeoff is best conducted at the local level. Where worker solidarity is high, the tradeoff may be worthwhile. The party in the best position to make this decision is the union itself, either through its by-laws or in negotiations with the employer. The best approach for state legislatures is to allow, but not mandate, members-only contract administration.
A second approach to the collective action problem involves shifting some or all of the cost of union negotiation and contract administration to the employer. This does not attempt to solve the collective action problem. Rather, it ignores it and makes government the primary funding source for the exclusive representative. Shifting costs to the employer already occurs in the federal sector and other open-shop environments. A much more radical version of employer subsidies for unions has been proposed by Professor Aaron Tang. We discuss the existing cost-shifting systems first and the more radical proposal thereafter.
In many jurisdictions that even before Janus did not allow agency fees, unions made up for the lost revenue by shifting to the employer some of the costs of representing workers that unions would traditionally bear. The most prominent example was the federal sector.
In maintaining that agency fees were not required to achieve the state’s interest in labor peace, the Janus Court pointed to the experience in the federal sector. The Court posited that despite a prohibition on agency fees, approximately 27 percent of the federal workforce are union members. Professor Samuel Estreicher estimated that 81 percent of federal government employees covered by collective bargaining agreements pay union dues, a remarkably high rate of membership density for an environment in which employees cannot be compelled to pay fair share fees or join the union. While the Court’s estimate and Professor Estreicher’s estimate are based on data from the Current Population Survey (CPS), there is good reason to believe that in non-agency fee environments, CPS data greatly overstates union membership density. The Court also cited the number of union members in the Postal Service, where agency fees are prohibited, and alluded to union representation of employees in states which prohibited agency fees. The Court’s reliance on these examples was flawed because the Court failed to account for the cost shifting that occurs in each of them.
Conspicuously absent from the Court’s discussion is any recognition that in environments where agency fees are prohibited, many expenses otherwise borne by unions are instead absorbed by employers. In federal agencies, the government subsidizes the union by allowing union representatives to perform representation functions on paid working time, known as “official time.” The Federal Service Labor Management Relations Statute (FSLMRS) requires that union representatives who are also employees of the agency be granted official time for negotiating a collective bargaining agreement, including participation in impasse proceedings. “The grant of official time allows the employee negotiators to be paid as if they were at work, whenever they bargain during hours they would otherwise be on duty.” The FSLMRS also authorizes the union and agency to agree on reasonable amounts of additional official time for representation functions. A federal employee union may bargain for union representatives to spend 100 percent of their time on representational functions, and in large bargaining units, it is common for employees serving as union representatives to spend 100 percent of their time representing workers. In environments with agency fees, such representatives would instead be full-time employees of the union, perhaps on unpaid leave from the employer.
The substantial financial subsidy provided by the employer to the union in the form of official time is not confined to federal agencies. In the Postal Service, where the governing statute prohibits agency fees, unions have negotiated for significant amounts of official time for stewards and other representatives to perform representational functions. It is also common for public employers in states that prohibit agency fees to agree to paid release time for employees serving as union representatives.
In the federal sector, it is also common for the employer to provide the union with rent-free office space, office furniture and equipment, and telephone and internet service. Additionally, in some cases, unions are able to compel the employer to reimburse the union’s attorney fees. In most other sectors, collective bargaining agreements typically provide that each party bears its own costs of representation in grievance arbitration and awards of attorney fees are quite rare. However, under the federal Back Pay Act, an employee-grievant that “ha[s] been affected by an unjustified or unwarranted personnel action which has resulted in the withdrawal or reduction of all or part of the pay, allowances, or differentials of the employee is entitled . . . to receive . . . reasonable attorney fees related to the personnel action.”
Shifting costs of representation from unions to employers, as detailed above, does not deal with the collective action problem or the free rider problem exacerbated by Janus. Instead, it ignores these problems and provides a level of financial support to the exclusive representative. This approach has many drawbacks beyond ignoring those key post-Janus problems. We focus on two.
First, shifting the cost of representation from unions to employers makes unions financially dependent on the employers with whom they negotiate and against whom they advocate. It also leaves their financial health vulnerable to changes in the political climate and to elected officials who are opposed to public employee collective bargaining. Recent actions by the Trump Administration illustrate this concern.
The Trump administration has made recent changes to personnel that will likely impact a union’s ability to bargain effectively. Under the FSLMRS, when an agency and its union are at impasse in collective bargaining, either party or the parties jointly may petition the Federal Service Impasses Panel (FSIP) for assistance. FSIP consists of at least seven members appointed by the President. The President may remove FSIP members at any time without cause. FSIP assists the parties in resolving the impasse “through whatever methods and procedures . . . it may consider appropriate,” and has the ultimate authority to “take whatever action is necessary . . . to resolve the impasse.” As is common when there is a change in the party occupying the White House, President Trump removed all of the Obama-appointed FSIP members and replaced them with his own appointees. The Trump appointees made clear their hostility to official time.
In United States Department of Agriculture, USDA Rural Development and AFSCME Local 3870, the parties had been at impasse over, inter alia, official time. The agency proposed two days per week of official time and the union proposed three days per week. The Trump FSIP, however, awarded only one day per week, half of what the agency was willing to provide. It is startling, and may well be unprecedented, that any third-party neutral would make an award outside the parameters of the parties’ final offers. It would have been just as startling had FSIP awarded four days per week. The clear message that FSIP sent to unions with this case is that, at least with respect to official time, unions should accept whatever the agency has offered because resorting to FISP could result in even less.
FSIP’s apparent hostility to official time is shared and fostered by the current President. In Executive Order 13837, President Trump directed that agencies should negotiate official time such that the “union time rate,” defined as the total number of hours of official time in a fiscal year, including the official time mandated by the FSLMRS, divided by the number of employees in the bargaining unit, not exceed one. If an agency proposes or agrees to official time that will exceed a union time rate of one, the agency head must report to the President “explain[ing] why such expenditures are reasonable, necessary, and in the public interest, describe the benefit (if any) the public will receive from the activities conducted by employees on such taxpayer-funded union time, and identify the total cost of such time to the agency.” The Executive Order prohibits any employee from being on more than 25 percent official time. It actually hinders most of the advantages explained in the context of federal agencies and the Postal Service. It prohibits union representatives from using official time to prepare or pursue grievances, including arbitration. It also prohibits free or discounted use of office space and other government property by the exclusive representative unless it is available on the same terms to other non-federal organizations for non-agency business, and prohibits reimbursement of employee expenses (such as travel and per diem) incurred in union representation duties.
Following the executive order, some agencies evicted unions from offices that they had used for many years. The United States District Court for the District of Columbia enjoined most of the executive order, including the restrictions on official time and office space, as violating the FSLMRS but the Court of Appeals for the D.C. Circuit reversed on jurisdictional grounds.
Many agencies, taking their cue from the executive order, may bargain moving forward for significant reductions in official time and other agency subsidies of union representation. If they do so, they will bargain with confidence knowing that the FSIP will likely back them up, or go even further than they had proposed, in limiting official time. As the recent federal sector experience demonstrates, when we rely on shifting costs to employers, unions’ financial health is too vulnerable to changes in the political climate. In other words, the benefits of agency subsidies of union representation can easily fall away.
Second, shifting the costs of representation to the employer can substantially diminish the incentives for union officials to engage the workforce and involve them in the union. One of the authors of this Article, Martin Malin, saw evidence of this when he served as a member of FSIP by appointment of President Obama. Although most unions with whom he dealt with appeared to have dedicated, committed leadership, and some federal employee bargaining units had high union density, a few local unions had very low membership density and it appeared that the local leadership had incentives to keep it that way. Low membership density disenfranchised the bulk of the bargaining unit, leaving the local leaders accountable only to a small number of employees, often their friends, who they could count on to continuously reelect them to their positions. In these positions, leaders enjoyed large amounts (sometimes 100 percent) of official time and were able to use an agency-provided union office with furniture, telephone, computer, and other office equipment. Shifting the costs of representation to employers can impede workplace democracy and worker solidarity and disincentivize local union leadership from engaging with the workforce.
Agency fees theoretically could incentivize workers to become union members, although there is no rigorous empirical evidence that they choose to become members for any reason other than solidarity and to get a voice and a vote. It is true that membership density is higher in workplaces where agency fees are charged to nonmembers than it is in workplaces that do not charge agency fees. In agency shops, the marginal cost of becoming a member, which is the difference between full dues and agency fees, tends to be small. Thus, there is little economic incentive to free ride; employees have good reason to become members to gain a voice and a vote in the union’s governance and operations for little extra cost. In bargaining units in which agency fees are required, there is often concern over whether a particular contract proposal would pass a union ratification vote because the negotiating committee is accountable to the membership. However, experience in federal sector mediation, where agency fees are prohibited, suggests that such negotiators often do not worry about whether contract proposals will be ratified by the membership, presumably because so few workers are union members with the right to vote on contract ratification.
Professor Tang proposes another form of employer cost-shifting: unions should be permitted to negotiate with public employers as to whether they will pay the union directly for the costs incurred in administering a collective bargaining agreement. This alternative, according to Professor Tang, eliminates Janus’ First Amendment concerns even under strict scrutiny, which is substantially identical to the form of “exacting scrutiny” that the Court applied in Janus. Tang addresses concerns that direct reimbursement would lead unions to be less effective in representing employees because of their financial dependence on the employer. He notes that most government employers already provide significant benefits such as access to workspace, employee time, contact information, and, most significantly, agency fee clauses. These are all financially significant forms of cooperation. Thus, Tang believes there is no reason to worry that additional support in the form of direct reimbursement would jeopardize the union’s role as an independent advocate for employee interests. Most importantly, Tang insists that even in a direct reimbursement scheme, the union’s members still direct the union’s activity; as a result, a union would not risk decertification to soft-pedal negotiations.
While it is not clear how much funding direct reimbursements would provide unions, the sum will not be set by the union; instead, it will be set by the union and employer jointly or by some third party. Government employers could be expected to insist on reimbursing only certain activities by adopting a narrow definition of “bargaining expenses.” Tang proposed to address this problem by creating an independent board that reviews union expenses to determine whether they qualify for reimbursement. The board can also delegate the question to an arbitrator. That system would only be as good as the independent board, and experience in the federal sector addressing disputes over how much official time union officials may have suggests that it would be difficult to depoliticize the amount of money the union will receive. Boards would be under significant pressure to reduce payments to unions, especially during budget austerity times.
Implementing a direct reimbursement model, as Professor Tang suggests, would effectively turn unions into government contractors. Ethics rules would need to be created to address the possibility or the perception of compromised loyalties. Professor Tang analogizes this to indigent criminal defense lawyers who are paid by the government to handle criminal defense and who must oppose the government that is funding them. But lawyers are governed by professional ethical rules that mandate loyalty to their clients, even when the government, against whom they litigate, pays their fees.
Currently, there are no similar rules governing public sector union representatives; they would need to be created to address these issues. Lawyers are inculcated from early in law school in a culture that prizes loyalty to client, and the adversary system reinforces that sense of undivided loyalty. Of course, commitment to the cause and to the people they represent is a crucial—probably the most crucial—motivator for both lawyers and union leaders. Nevertheless, however dedicated many public employee union representatives are to the union and to its members, it is unclear whether unions will remain as independent if they depend on government funding for their survival.
A direct reimbursement model suffers from the same drawbacks as the system of official time and related subsidies. It leaves exclusive bargaining representatives’ financial health vulnerable to changes in the political climate. It can diminish the incentives for local union officers to engage the workforce and involve them in the union, thereby decreasing worker solidarity and workplace democracy.
Direct reimbursement has additional drawbacks not present in the official time and related subsidy systems that currently exist. The major drawback is that the Tang proposal will require significant changes in law regarding union financial independence and employer domination of unions. As for financial independence, the Tang proposal would require repeal of any state law comparable to section 302 of the Labor Management Relations Act, which prohibits employers from paying, lending, or giving money or anything of value to any labor organization. Many states have such statutes. While it is clear, as Professor Tang points out, that federal labor law allows an employer to cooperate with a labor organization to some degree, it does not appear to allow for direct payment to a union. Section 302 specifically exempts dues check-off systems from the ban on employer payments, but it does not likewise exempt direct and complete financial support of a union. Professor Sachs concludes that section 302 forbids direct employer payments.
Even states that do not have provisions like section 302 have labor law provisions that prohibit employer domination of or assistance to labor organizations. The NLRA has such a provision in section 8(a)(2), which prohibits employer financial support of any labor organization. Professor Tang cites to a few employer-domination decisions under section 8(a)(2) that differentiate between unlawful support of the union on the one hand and permissible cooperation on the other. But these are cases under section 8(a)(2) where the employer allows union members to use company space and resources. Section 302, however, explicitly prohibits employers from paying, lending, or delivering money or anything of value to any labor organization.
Still, in states that forbid employer domination, interference, and assistance, as section 8(a)(2) does, but do not explicitly outlaw direct payments to unions, as in section 302, parties could bargain for a direct reimbursement model. States could do so effectively without statutory change if they avoid impermissible employer intrusion in the governance or priority-setting of the labor organization. Tang and Sachs note that member participation can be a cure for the possibility of employer domination under the direct reimbursement model. Labor boards and courts do look to a totality of the circumstances to determine whether the employer’s assistance–the lowest level of intrusion of the three categories of influence–threatens the independence of the union. These authorities do allow some level of cooperation. But, even the inclusion of a direct reimbursement clause in a collective bargaining agreement negotiated at arms-length will not be enough to protect against a claim of impermissible support.
One solution to the section 302 problem is for the state to create a fund from which unions may draw; this would not constitute a direct payment to the union. This is the approach that Hawai’i proposed to take under a bill, H.B. 923, that was introduced in 2017 but did not become law. The Hawai’i bill would have established a collective bargaining fund that would have had allocated to it a minimum amount of money based on a percentage of the total compensation of all public employees. The fund would be disbursed to a union according to the terms of a contract between the union and the state department of budget and finance. The amount would be based on the share of total number of public employees represented by that union. Any money left over in the fund would go to the state’s general fund. The Hawai’i bill would have allowed unions to provide other services that are not paid by the collective bargaining fund to employees. It would also have allowed unions to charge nonmembers for such services. The bill explicitly stated that unions selected by a majority of employees would retain exclusive representation, but would allow unions to adopt members-only representation at the union’s election.
However, even this approach has its drawbacks. The fund might be perceived as a political payback to unions that supported the political leaders who ultimately decide how much money the unions will receive. In states in which public support of unions is precarious, a perception that the public is now directly paying for unions through the government fund could fuel legal and political attacks on the system.
As demonstrated above, direct payment systems can take many forms, and not all of them necessarily result in diminished independence and accountability to members. For example, unions have long used dues to provide services to workers and the public, which could be provided through a system of direct payments from the government instead. Home care unions have operated call centers that provide information of the sort that an HR office would ordinarily provide (“Why is my pay check smaller than I expected?” “Where can I get training to improve my skills?”). Many unions offer trainings that were financed by dues and could instead be financed by government payments. Some unions operate job training and labor market intermediary institutions such as apprenticeship programs or hiring halls, which could be funded by employers rather than by dues. And a program could be administered by a trust, rather than directly by the union, in order to increase its autonomy and accountability to the government and to the public.
If the government subsidized such service programs and union dues funded core contract negotiation and administration functions, unions would not be subject to financial vulnerability and dependence issues, as dues would still be the largest source of union income. Nor would such programs reduce member engagement with the union, as could be the case with full direct payment systems explained above. And because the union would continue to depend on member dues, the union would be working to solve the collective action problem by relying on solidarity-building engagement with members and dues collection, while ensuring that the government-funded service programs served the interests of members and the public.
The creation of a collective bargaining fund, or perhaps even a more modest service program like those describe above, would likely draw a constitutional challenge. Anti-union organizations might argue that the money paid into the fund would otherwise have been paid to the employees as wages. They would argue that this is effectively a compelled subsidy by employees to the union. The issue here is whether the employees ever had a legal entitlement to the fund payments–if they did, the constitutional challenge would be the same as the one in Janus. Professor Tang argues that employees would never have had “a legitimate claim of entitlement” to the fund payments, unlike the fees deducted from payroll, and therefore there is no compelled subsidy. It would be important, therefore, for the government employer not to offset the cost of reimbursement with reductions in either current employee wages or future increases specified in a collective bargaining agreement. This would demonstrate a legitimate claim of entitlement. But this system would allow the employer to reduce other future pay raises and cut benefits like “turning down the thermostat or putting limits on employee phone usage,” without constitutional problems.
Apart from the legal obstacles discussed above, there are a number of drawbacks to the direct payment model. For one, in difficult negotiations, the employer could have greater leverage with direct payments rather than with fair share fees and dues checkoff. After all, the cognitive distance between a worker and her support for the union is greater when the employer pays, even if the funds originate from future raises or other discretionary benefits. Tang acknowledges that this arrangement could lead to members being less vigilant about excessive union spending, which could then require negotiating annual caps on budget increases based on growth. But it could also be that workers who are not paying dues have less personal economic and psychological investment in maintaining the union. As a result, they may not be inclined to start paying dues when the employer plays hardball. 
Moreover, the concern about taxpayer funds supporting union activities could gain force if union-represented employees perceive a lack of transparency in how unions govern themselves and fund their operations. However, existing law requires unions to provide a high degree of transparency to workers and the public, and additional protections could be added in states where it might be needed. Most unions that represent public sector employees also represent private sector employees and are therefore covered by the extensive reporting and disclosure requirements of the federal Labor Management Reporting and Disclosure Act (LMRDA).  Even for those that are not, unions under a government funding system might require the same kind of transparency and financial accountability rules that are applied to any other government contractor.
About ten states have enacted laws like the LMRDA. Others have adopted similar, but lesser, protections. California is an example of the latter group. California’s statewide public labor relations laws do not, for the most part, contain all the transparency and internal governance provisions of the LMRDA. For instance, public employee unions are neither required to adopt particular rules governing the election of union officers or dues increases, nor must they adopt rules regarding use of union funds for elections for internal union leadership positions; these are left to union constitutions and bylaws. However, despite any state requirement, many of the major public sector unions have provisions in their union constitution that are applicable to all local affiliates requiring some of the same transparency and financial probity measures as under the LMRDA. California does, however, require its public employee unions to make financial disclosures, although the particular requirements vary union to union. Organizations affiliated with exclusive representatives are not required to disclose their finances.
In sum, any system of direct financial support of unions, be it from the government or the employer, makes unions dependent on employers, or politicians, rather than on members. This leaves unions vulnerable to cost-cutting and could lead to a loss of member-driven union priority-setting and autonomy. Direct financial support can reduce incentives for union leadership to engage bargaining unit members in their representation and thereby reduce internal union democracy. In some cases, it may undermine the ability of union-represented public employees to take a strong stance against government policies that have been harmful to the public, as illuminated in the 2018-2019 teachers’ strikes. For all of these reasons, most of the major public employee unions have not embraced any system of direct financial support as the simple solution to the collective action problems engendered by the decision in Janus.
Post-Janus concerns involving whether employees can resign union membership or revoke an agreement to pay a representation fee at any time have led to the idea of treating resignation in a manner comparable to open enrollment periods in health insurance. The possibility that employees will resign or revoke an agreement to pay a representation fee at any time makes budgeting extremely difficult. This is especially important because a union requires a predictable budget to gauge staffing, and staffing dictates the kinds of services the union can offer to its members and represented nonmembers. Moreover, if unions respond to the collective action problem by providing financial incentives to members that are not available to nonmembers—such as additional insurance or, as suggested by Janus, free representation in grievances—and employees can join and quit the union freely, there will be a huge adverse selection problem. People will join the union only when they need a service it provides and will resign the union whenever the need for services ends. This kind of adverse selection problem has been known to insurance companies for decades and explains why they, and employers that contract with insurers, restrict the times when employees may enroll in or drop life, health, accident, and other kinds of insurance.
Various jurisdictions have interpreted Janus differently in terms of employee union resignation. The proponents of the Janus rule insist that Janus means employees can resign their union membership at any time. That approach appears to have been adopted in some jurisdictions. But other jurisdictions have borrowed from the model of insurance and are treating resignation from union membership like open enrollment: people get a window once a year to elect coverage. Once they make their election, they cannot opt out of coverage until the next annual open enrollment period.
New Jersey amended its public sector labor relations law in May 2018 in anticipation of Janus. The New Jersey statute takes two approaches to the collective action problem. One is discussed in Part III below: it enhances union access to employees to discuss the benefits of unionization. New Jersey’s other approach is intended to promote unions’ financial stability.
The New Jersey statute treats membership similar to insurance with an “open enrollment” period that regulates when and how employees can revoke their authorization for deducting dues. The New Jersey statute repeals a provision allowing employees to revoke their authorization of payroll deduction at any time and instead provides a ten-day window each year during which revocations may be made. Under the old law, a revocation could be made in writing at any time, though it would be effective only on the next succeeding January 1 or July 1 to give the union fair notice of the loss of revenue. The new law makes the revocation effective within thirty days. Similarly, Hawai’i now limits employee revocation of dues deduction authorization to a window period within thirty days of the anniversary date of the authorization.
Laws that either create a window for dues revocation or allow a labor contract to create a window are justified by organizations’ legitimate needs for financial stability and predictability. As noted above, these laws apply to unions the same rules of open enrollment that are applied to health insurance and other employee benefit plans: workers get a once a year opportunity to make an election. (Except in the insurance context, the restriction is on signing up for and dropping coverage; in the union context the restriction would be on resigning.) The once a year open enrollment rule is a mandate of the Affordable Care Act, Medicare, and other benefits programs.
Laws that restrict the revocation of dues authorizations to a window period (particularly if they are analogized to restrictions on resignation from the union) are potentially vulnerable to the same argument about compelled speech and association that brought down fair share fees. As critics of the Supreme Court’s fair share fee jurisprudence have observed, the Court has perceived more First Amendment content in union membership and union fees than it has in similar compulsory activity, such as compulsory bar membership and dues for lawyers. And it has not yet determined whether compulsory payments of other sorts—such as homeowner’s association dues and employer sponsored health insurance—raise any First Amendment issues at all. Courts have not treated compelled payments to insurance programs as speech, even though such payments subsidize speech of insurance company representatives just as union dues and fair share fees subsidize speech of union representatives.
A slight variation on a statutory window period for revoking dues deduction authorizations that will avoid First Amendment problems is to allow unions themselves to set the terms for revocation. When employees join or authorize payroll deductions for dues, they do so subject to the terms offered by the union. The result is a private contract between the employee and the union. Thus, the government has not mandated the terms and there would appear to be an absence of state action necessary to trigger First Amendment scrutiny. As one district court observed, “the First Amendment does not confer . . . a constitutional right to disregard promises that would otherwise be enforced under state law.”
The open enrollment approach to union dues authorization has advantages of predictability, but it does not solve the collective action problem. It does not give new employees any reason to join the union. And, by making exiting difficult, it may deter employees from joining in the first instance.
In sum, members-only representation, cost shifting, and restricting resignation of membership to designated window periods do not solve the collective action problem that Janus ignored. Four major public employee unions have thus far declined to endorse them and have, instead, promoted strategies to strengthen solidarity. We turn to those now.
As discussed previously, the free-rider problem arises when an employee declines to join the union but receives the valuable benefits of union representation for nothing. The collective action problem arises because employees who support the union makes a rational decision not to join because the impact of their membership will be minimal. Membership becomes economically rational when the employee is assured that most coworkers will also join the union. Here we explore several systems that might address the collective action problem by removing the incentive not to join and creating incentives to join. First, we discuss the possibility of treating workers who object to paying fees similarly to workers who oppose supporting union causes because of religious objections. Next, we analyze different economic arrangements proposed, such as the imposition of arbitration process fees and members-only benefits. Last, we explore the possibility of using education to allay the concerns that lead to the collective action problem.
Some have analogized workers who refuse to join the union for ideological reasons to those who refuse to join or financially support the union for religious reasons. Prior to Janus, the only employees allowed to refuse to pay fees to the union were those who refused as a matter of conscience, not as a cost saving. Janus found the agency fee to be compelled speech because of the political nature of public employee bargaining. But what if employees were free to make their payments to another organization that engaged in speech they favored? That would eliminate the financial incentive not to join and yet not compel employees to subsidize the speech of an organization they oppose. The NLRA and many states already have such arrangements for people who have religious objections to unions.
Recall that agency fees deal with the free-rider problem by preventing employees from getting something for nothing and they deal with the collective action problem by making the marginal cost of union membership small. Consequently, the benefits of union membership outweigh the low marginal cost, and it becomes an economically rational decision for an employee to join the union. Employees always have the choice of whether to join, but economically rational employees are assured that if they join the union their economically rational coworkers will likely also join. After Janus, however, the marginal cost of joining is the full amount of union dues, making it an economically rational decision for even union supporters to refrain from joining.
The treatment of religious objectors to fair share fees offers one approach to addressing the collective action problem that Janus ignored. Under Section 19 of the NLRA, added in 1974, a collective bargaining agreement may not require employees who have bona fide religious objections to join or financially support a labor organization to pay a union fee. The exemption extends to “[a]ny employee who is a member of and adheres to established and traditional tenets or teachings of a bona fide religion, body, or sect which has historically held conscientious objections to joining or financially supporting labor organizations.” Employers and unions may agree to require such religious objectors “to pay sums equal to such dues and initiation fees to a nonreligious, nonlabor organization” that is a tax exempt organization under section 501(c)(3) of the Internal Revenue Code. The employer and union may create a list of at least three organizations from which fee objectors may choose, or, if there is no such list, fee objectors may pay the sum to a nonreligious, nonlabor organization of their choice. As the majority in Janus observed, apparently favorably, religious objectors can be required to pay the union’s cost of handling their grievances.
Because religious objectors still have to pay an amount equal to dues, the system removes the financial incentive to shirk while not forcing employees to subsidize an organization of which they disapprove. In a sense, it is like obtaining conscientious objector status from military service; the objector is not forced to engage in conduct that violates his conscience but still must do something comparable. Because the employer does not dictate what speech the nonmember must subsidize, it avoids the constitutional problem of Janus.
The religious objector system could be expanded to require any employee who opts, for any reason, not to join the union to pay an amount equal to union dues to a tax-exempt charitable organization. This approach would address the Janus compelled speech concern unless a court were persuaded that compelling employees to support any organization is unconstitutional. However, in Board of Regents v. Southworth, the Court rejected the contention that all compelled payments are unconstitutional. The Court held that a mandatory student activity fee imposed by the University of Wisconsin was constitutional even though it subsidized activities that some students objected to. As with agency fees, the university required the complaining students to subsidize the expressive activities of third parties with whom they disagreed. The Court found the fees constitutional, provided that funds were allocated to student groups in a viewpoint neutral manner. If the organizations eligible to receive the nonmembers’ payments are selected in a viewpoint neutral manner, the required payments might be considered constitutional by analogy to Southworth.
The student activity fees upheld in Southworth are, however, different from a union fee objector system in two respects. First, in Southworth, the Court emphasized that the university allocated funding to groups without regard to political or ideological viewpoint. In the approach discussed here, the union and employer would jointly select the organizations eligible for fee objector payments. There may not be a practical way to assure that the selections are done in a viewpoint-neutral fashion. And even if the selections are made in a viewpoint neutral fashion, a nonmember might still be philosophically opposed to supporting any group on the list. This concern could be resolved by allowing nonmembers to designate any 501(c)(3) organization to receive their payments.
The second reason Southworth could be distinguishable, even if nonmembers are allowed to choose their donee organization, is that the student fee program provided all students the benefit of “dynamic discussions of philosophical, religious, scientific, social, and political subjects in their extracurricular campus life outside the lecture hall.” The unionized public sector workplace funded by fair share fees is not designed to provide a rich extracurricular life. On the other hand, union fees benefit all workers by enabling the negotiation and enforcement of good working conditions and a fair disciplinary process. Thus, as in Southworth, all benefit. By failing to treat these like cases alike, the Janus majority was transparent in its hostility to public sector collective bargaining.
Even if a requirement that nonmembers pay an amount equal to union dues to whatever 501(c)(3) organization they wish survives constitutional attack, it may not solve the collective action problem. Nonmembers could simply pay the money to charitable organizations to which they would have donated anyway. Under these circumstances, it may remain a rational economic decision for even a union supporter to decide not to join the union and pay dues. To address this problem, all bargaining unit members should be required to contribute to a nonideological fund to which they would not otherwise have contributed, and the union could reimburse its members as a benefit of union membership. This approach is a variation on the fund created by the Portsmouth, New Hampshire school district and teachers’ union when agency fees were lawful. Their collective bargaining agreement required nonmembers to make a contribution equal to union dues to a fund that provided scholarships for Portsmouth high school graduates in the name of the union.
The Portsmouth system might be subject to constitutional challenge under Janus because it arguably discriminates against nonmembers by requiring only them to contribute to the scholarship fund. Because the scholarships are given in the name of the union, this contribution arguably compels nonmember association with the union. But this problem could be addressed. Rather than requiring only nonmembers to support the scholarship fund, the parties could agree to create an independently administered scholarship fund. The fund could, for example, provide scholarships to any eligible graduate of the district’s high school and require all employees, not just nonmembers, to support the scholarship fund. The union, as a benefit of membership, could reimburse its members for the required contributions.
Compelled subsidization of a scholarship fund is not compelled subsidization of an expressive activity. Thus, this variation on the Portsmouth approach may have a better chance of surviving a post-Janus constitutional attack. A school district certainly has a legitimate interest in providing scholarships for its graduates to further their education. Requiring teachers to fund those scholarships sends a positive message to students and parents that the teachers are personally invested in their students’ success. Although such a system may encourage union membership, the encouragement stems not from the requirement that all teachers contribute to the fund but from the union’s decision to reimburse its members as a benefit of membership. In workplaces outside of public education, the union and employer might agree to use the fund for other purposes related to the employer’s public mission, including safety campaigns, programs for low-income children, and so forth.
Creating nonideological funds to which all employees are required to contribute, and reimbursing members for their contributions as a benefit, could solve the collective action problem similarly to religious objector provisions. Both ensure that workers will decline to join the union out of conscience rather than out of economic expediency.
Another alternative would not completely solve the collective action problem, but would ameliorate it. Collective bargaining agreements provide employees with numerous benefits. One of those benefits is access to a grievance procedure that culminates in arbitration.
Arbitration is not free. The parties pay the arbitrator’s fee and, where applicable, travel expenses. Other costs depend on how the parties proceed. If they select their arbitrator by obtaining a list from an outside organization, such as the American Arbitration Association or the Federal Mediation and Conciliation Service, they may have to pay a fee for the organization’s services. They may also incur costs for renting the hearing room or hiring a court reporter. Typically, the collective bargaining agreement will provide that the employer and the union will be equally responsible for the costs of the proceeding.
We suggest that the parties treat grievance arbitration as they do other employment benefits by having the employer and the employees share in the costs. For example, health insurance is an employee benefit that typically has cost sharing, with the employer paying a certain percentage and each employee paying the rest. Employers who self-insure, as most do, project costs each year based the expenses incurred in prior years, and derive “premiums” from these projections. Public employers who self-insure maintain separate funds into which the employer’s share and the employees’ share of the “premiums” are deposited. The employer takes money from these funds to pay claims.
Similarly, the employer could project anticipated arbitrator fees and related costs for the coming year and deposit 50 percent of that amount into a separate fund. Employees in the bargaining unit would be assessed their pro rata share of the other 50 percent with money withheld from their pay—like health insurance premiums. The fund would pay arbitrators, court reporters, and the other expenses of arbitration. Each year the employer will project costs and make new assessments based on those projections. The union, as a benefit of membership, could reimburse its members’ payments to the arbitration fund. Alternatively, employers could charge all employees who resort to arbitration an arbitration process fee. The union would pay the fee for its members, and those employees who choose not to join would have to pay the fee themselves.
We believe this approach is likely to survive constitutional challenge under Janus. Arbitrators, court reporters, hearing room providers, and arbitrator-appointing organizations are not engaged in First Amendment protected expressive activity, regardless of whether the advocates representing the union and the employer may be. We discuss unions providing members-only benefits below, but we believe there is no basis for attacking the union’s reimbursement of member payments as a benefit of union membership. Similarly, there would be no basis for nonmembers to attack a union should it decide as a benefit of membership to reimburse some of each member’s health insurance premiums.
The arbitration process fee is a model that could be expanded to a whole range of other services and benefits. For example, a similar framework could be used to meet other required or encouraged obligations like professional development (for teachers or health care professionals), wellness programs, and so forth. The union could position itself to provide the required training for the best value on a members-only basis, or it could contribute the members’ cost of participation and leave nonmembers to satisfy their own requirements.
The degree to which this approach will ameliorate the collective action problem will vary depending on the size of the bargaining unit and the number of grievances generated in the unit. In small bargaining units with few grievances, the employee’s assessed costs will be minimal, perhaps even zero. However, the collective action problem is least likely to arise in small groups. In larger state-wide bargaining units that can have tens of thousands of employees and many grievances, the arbitration costs are likely to be substantial. Assessing employees their pro rata share of half of those costs with the union reimbursing its members for those assessments will contribute to making union membership a rational economic decision.
An arbitration process fee addresses the collective action problem by assessing each individual employee for the pro rata share of the costs of a collective good for which the union negotiated and providing a members-only benefit of reimbursement of that assessment. On a more general level, unions may provide members-only benefits that will further incentivize workers to join and assure them that their coworkers will also join because membership is an economically rational decision. We address this in the next Section.
Another way to deal with the collective action problem is for unions to provide benefits that increase the value of union membership and thereby make it a rational decision for employees to join and pay dues. The suggestion above that employers assess employees their pro rata share of one-half the costs of the arbitration forum and unions reimburse their members for the assessment is one way of increasing the value of union membership. But it needn’t be limited to arbitration.
Unions provide a range of benefits to their members funded through union dues and, presumably where costs warrant, by additional individual contributions to defray the cost of the benefits. These include supplemental benefits such as disability insurance, free legal representation, liability insurance, maternity/paternity benefits, professional development, and many types of trainings. Home health care provider unions have run call centers in multiple languages so that home health aides (who work in the isolated setting of a home and do not have an HR office where they can ask questions) could call in. Some questions might focus primarily on the worker’s concern (“Why is my pay check delayed or lower than I was expecting?”) but others would benefit the employer and the public too (“How can I get training to deal with a particular situation?”).
Beyond the benefits that unions currently provide to members only, many benefits that are collectively bargained for in unionized sectors (including health and disability benefits and life insurance) were once provided unilaterally by unions only to members. Unions in a post-Janus world could use their resources to improve benefits for their members rather than use resources to negotiate and administer a contract that provides those benefits to all bargaining unit members.
Members-only benefits have drawn Janus-style legal challenges but, thus far, the challenges have failed. In Bain v. California Teachers Association, for example, nonmembers of a teachers union alleged that the union violated their First Amendment rights because it only provided benefits to members, rather than through collective bargaining. Those benefits included “disability insurance, free legal representation, life insurance, death and dismemberment benefits, and disaster relief,” as well as maternity benefits. They also challenged the union rules that limited voting in union elections to members. The court dismissed the case for want of state action, and other similar suits have likewise failed. Even if there were state action, the First Amendment argument seems weak because many membership organizations—automobile clubs, health clubs, and so forth—provide benefits only to those who join and pay dues or other fees to defray the cost of service.
A more substantial argument might be that a union breaches the duty of fair representation (DFR) by choosing to provide benefits only to dues-paying members rather than to negotiate with the employer to provide benefits to all represented workers. The DFR, which governs all actions the union takes in its role as exclusive representative, contemplates that the union must necessarily prioritize the interests of some employees over others: negotiating for better health insurance will benefit those who are ill at the expense of those who are healthy and might prefer a wage increase. Recognizing this, the Supreme Court said that only arbitrary or invidious discrimination violates the DFR, and “[t]he complete satisfaction of all who are represented is hardly to be expected.”
Among the impermissible bases of discrimination are distinctions based on union-membership, but, not all discrimination against nonmembers violates the DFR. Using the example of Bain, the California Unemployment Insurance Code section 710.4 allows a school district to elect whether to become an employer covered by the paid family leave law. The plaintiffs in Bain alleged that the California Teachers Association and its affiliates decided against seeking coverage under the state system so that the unions could provide the benefits directly to their members only. The statutory regime, they alleged, “give[s] the unions a veto over teachers’ participation” in state disability insurance and paid family leave programs. Although the plaintiffs in Bain did not bring a DFR claim, their First Amended Complaint allegations regarding the union’s decision to provide benefits itself rather than negotiate for the school district to provide them had the same theory—the union was discriminating against nonmembers. Framed as a DFR claim, their theory might fail on the grounds that a union’s decision to negotiate for benefits that protect some more than others is generally subject to the deferential “wide range of reasonableness” standard, even when the line the union draws benefits union supporters more than opponents.
The Bain plaintiffs’ argument may prove too much as any union could negotiate for an employer to provide a benefit (such as legal services, life insurance, or death and dismemberment insurance) that the union could also provide itself. The history of unions reveals that unions once provided almost all benefits directly to their members and only began to negotiate for employer-provided benefits in the mid-twentieth century. On the other hand, a court might find a union’s actions unreasonable if it focuses on the ability of the union to prevent all teachers from accessing certain state programs and to provide those benefits only to union members in order to induce teachers to join. The CTA could certainly argue, though, that teachers can still access disability and paid family compensation through other means.
To sum up, in most cases it appears that unions can provide significant benefits without fear of violating the DFR. But when the union alone is the gatekeeper to a particular service, courts may feel more pressure to forgo deference and scrutinize the union’s decisions. Of course, this concern could be met by having an independently-managed and transparent trust administer the system for the entire bargaining unit.
In addition, unions negotiate benefits that individual employees are unable to enjoy on their own. An example is firefighter duty trades. Firefighters usually work 24-hour shifts followed by 48 hours off. In a given firehouse, there will be three shifts that rotate the days that they work. A benefit that firefighter unions frequently negotiate for is the right to trade shifts with a fellow firefighter of equal rank. Of course, that benefit is worthless unless a firefighter desiring a particular day off is able to find another firefighter to trade with. Firefighter union leaders with whom we have spoken have said that it is likely that union member firefighters will refuse to trade shifts with nonmembers. Although the union does not coordinate such duty trades, the operation of duty trades in the fire service suggests possible analogous benefits where unions may provide a coordination role for their members only.
One such benefit is sick leave banks to which employees may donate their unused sick leave. Contributors to the sick leave bank may draw sick leave from the bank if they face a lengthy illness and have insufficient personal sick leave to cover it. Many employers are reluctant to agree to sick leave banks because of the costs of coordinating them. But an employer might agree to the creation of a sick leave bank provided that the employer did not have to administer it. The employer and the union could provide for sick leave banks administered by groups of employees and set whatever minimum number of employees necessary for the sick leave bank to be viable. Unions could then offer coordinated sick leave banks as a benefit of membership.
Alternatively, unions could offer participation in their sick leave banks to all employees but charge an administration fee to nonmembers who wish to participate. Such administrative fees would be consistent with the Janus Court’s apparent approval of unions conditioning grievance representation of nonmembers on the nonmember paying for the costs of such representation. They would also be consistent with the service fees that unions charge nonmembers for use of their hiring hall referral services, which have been uniformly upheld.
Depending on the type of employee, unions may find other benefits to offer their members that will contribute significantly to making membership an economically rational decision. Teachers’ unions provide many professional development services. Health care workers’ unions provide safety and other training. Several public sector unions are pursuing initiatives to fund independently managed, transparent trust funds to administer services for entire bargaining units that might otherwise fall to unions, employers, or other organizations to provide.
Free legal services can provide an incentive for employees to join the union. For example, it appears that free legal services for members plays a role in Nebraska teacher unions’ membership density. Free legal services also play a role in law enforcement union membership density. Unlike typical employees who likely discount the value of free representation in grievance proceedings, law enforcement officers face the potential need for legal representation on a more frequent basis than other government workers. For example, they cannot predict when they may become involved in an officer-involved shooting which leads to an automatic investigation. Officers facing such investigations are exposed not only to potential discipline or discharge but also civil and criminal liability. They need legal representation in such investigations and law enforcement unions generally provide it as a benefit of membership. Law enforcement union leaders with whom we have spoken cite this benefit as contributing to their high levels of membership density.
As the Court recognized in Janus, employees can save hundreds of dollars per year by deciding not to join their union or pay the agency fee. Unless educated about the benefits of membership and the level of membership solidarity in the bargaining unit, many will elect to not join, believing they will get the benefits of a union contract without paying the costs. Recognizing this, several states have reacted to Janus by amending their public employee collective bargaining laws to require employers to promptly provide exclusive bargaining representatives the names and contact information for new members of the bargaining unit (either new hires or transfers). The amended laws also require employers to afford union representatives time to meet with new bargaining unit members during the members’ regular working time, i.e. on the clock.
These statutes deal with the collective action problem by enabling exclusive bargaining representatives to emulate, in some respect, smaller organizations where these problems are less likely to arise. Individuals are more likely to perceive their contributions to small organizations as having a noticeable effect. Additionally, small organizations are likely to support group solidarity and develop social norms against shirking. As Olson recognized, social pressures and social incentives work best in small groups where members have face-to-face contact.
As discussed previously, firefighter unions tend not to have collective action problems because firefighters literally live with each other in the same firehouse. Their lives depend on each other when they are performing their job duties, and they enjoy benefits that require group solidarity, such as duty trades and meal pools. The combination of social norms and incentives felt at the level of the individual firehouse accounts for the high level of solidarity even in large urban fire departments. Large fire departments with social incentives felt at the level of the individual firehouse exemplify what Olson referred to as a “‘federal’ group—a group divided into a number of small groups, each of which has a reason to join with the others to form a federation representing the large group as a whole.” Olson urged that federal groups were the one type of large organization where special incentives could overcome the collective action problem because “small constituent organizations . . . may be induced to use their social incentives to get the individuals belonging to each small group to contribute toward the achievement of the collective goals of the whole group.”
Statutes that require that employers notify exclusive bargaining representatives of new members of the bargaining unit and provide the opportunity to meet with new unit members on the clock can afford unions the opportunity to develop some of the attributes of small organizations, even in statewide bargaining units with thousands of members. They replicate features of a firehouse in a large urban fire department and can help turn large bargaining units into “federal groups.”
Contact with the new employee should come from a local union representative who is also a coworker. The representative/coworker can educate the new employee about the importance of paying annual dues, the benefits of union membership, and why it is an economically rational decision to join. (And, of course, why the new employee might wish to join the union for reasons of solidarity and community engagement wholly apart from economic self-interest.) Such meetings at the start of employment and regularly thereafter will socialize employees into the local group. This will establish norms of solidarity so that the new employee does not perceive that dues are going to a large faceless organization where they have virtually no impact.
There is a successful private sector model for building solidarity through new worker orientations and ongoing meetings. The United Automobile Workers union and the Big Three automakers have a robust program in which the union participates in new hire orientation and in annual meetings at which workers are informed about company goals, policies, and programs. Union involvement in these meetings helps workers see the union as part of the fabric of the workplace rather than a company adversary or a troublemaker. Everyone benefits in a unionized workplace if workers see that the union and management can work together. And this is especially true in the public sector where the union-management contract negotiations are not about dividing up the profits of the enterprise, but instead are about solving problems in the delivery of public services and the administration of government.
It is likely that these new statutory provisions will be challenged by opponents of public sector collective bargaining. Those challenges should be summarily rejected.
Granting a union a statutory right to participate in new employee orientation presents no constitutional problem. The Janus Court recognized that the status of exclusive bargaining representative confers benefits including “obtaining information about employees.” The Court cited this favorably in explaining why the duty to represent nonmembers did not justify charging nonmembers agency fees. Moreover, the Court has already rejected a constitutional challenge to a similar program that allowed a teacher’s union but no other organization to communicate with teachers through interoffice mail. In Perry Education Association v. Perry Local Educators’ Association, the Court rejected a First Amendment challenge brought by a rival union, holding that granting access only to the certified union was “reasonable in light of the purpose” served by the office mail system. The Court held that the access granted to the exclusive bargaining representative was reasonable because it enabled the union to perform its obligation to represent all of the district’s teachers.
Perry Education Association should control challenges to recent state enactments requiring employers to provide exclusive representatives names and contact information for new bargaining unit members and the opportunity to meet with new unit members during working time. As with access to the school’s mail system, this privilege is accorded based on the union’s status as exclusive representative and not based on the viewpoints it espouses. Meeting with employees during working time is not a public forum, and the grant of access facilitates the exclusive representative’s ability to represent all employees in the bargaining unit.
In Janus, the Supreme Court stripped public sector labor unions and employers of a key tool in dealing with the collective action problem facing all large organizations providing collective goods. Among the strategies that states are considering to fill the void, we believe only those that preserve unions as majoritarian democratic organizations serve the public interest in good labor relations.
Several proposed strategies erode the concept of exclusive representation by allowing unions to limit their representation to dues-paying members, either completely or in administering the collective bargaining agreement. We find these approaches wanting. Members-only representation, either for contract administration or for bargaining and administration, would require a significant overhaul of public employee collective bargaining laws, allowing parties to discriminate on the basis of union membership. It would likely increase workplace dissension and, in most cases, would complicate an already complex public sector personnel system. That was California’s experience with members-only representation before 1976. Reviving it will not solve the collective action problem.
A variation on members-only representation is members-only representation in contract administration. This approach will solve the free-rider problem but is highly unlikely to solve the collective action problem. It also runs a significant risk of impeding worker solidarity. We recommend that evaluation of the risks and benefits of members-only contract administration be made at the local level. States should amend their public sector labor relations acts to allow, but not to mandate, members-only contract administration, conditioned on nonmembers being able to process their own grievances with their own representatives. Where the collective bargaining agreement does not give individual employees a right to pursue their own grievances, unions should remain under a duty of fair representation to all members of the bargaining unit.
Other strategies ignore the collective action problem entirely and seek to ensure the solvency of exclusive bargaining representatives with heavy subsidies from public employers or the state in general. We recommend against such strategies because they erode worker solidarity and workplace democracy. They create union financial dependence on political bodies, making unions vulnerable to government cost-cutting, and they reduce the accountability of union leadership to the workers whom they represent.
One potentially promising strategy is to treat all fair share fee objectors like religious objectors and allow them to donate an amount equal to union dues to a charitable organization. A significant drawback to this approach is that specifying a group of organizations from which a nonmember may choose to donate to is vulnerable to attack under Janus. In addition, allowing nonmembers to donate to any charitable organization they select impedes the usefulness of the scheme in solving the collective action problem. Some employers, such as school districts, might mandate contributions from all employees to funds that are not expressive and that further the employer’s mission (such as scholarship funds or job training), with the union reimbursing its members’ contributions as a benefit of membership.
More generally, we advocate a multi-prong strategy for ameliorating the collective action problem. First, we urge that employers and unions treat the availability of grievance arbitration as they treat other employee benefits, with employers and their employees sharing the costs. Employees could be assessed, through payroll deduction, their pro rata share of the forum costs, such as arbitrator and court reporter fees, arbitrator appointing agency fees, and hearing room rentals. Unions might then reimburse their members’ payments as a benefit of membership. Second, unions should provide, and employers should facilitate, members-only benefits that will incentivize workers to join or continue their membership. These benefits can include, in appropriate instances, legal defense services, insurance benefits, and coordination of sick leave banks, among others.
Most importantly, unions should strive to emulate smaller organizations where member understanding of the impact of their dues and norms of solidarity and mutual support guard against collective action problems. States should facilitate this by mandating that employers provide exclusive representatives with contact information for new members of the bargaining unit and the opportunity to meet during working time with new unit members shortly after their hire or transfer into the unit. Such statutory mandates have been enacted in California, Maine, Maryland, New Jersey, New York, Oregon and Washington and are very likely to survive constitutional attack. Although our recommended multi-prong strategy may not solve the collective action problem as well as agency fees did, it will go a long way toward ameliorating the problem and preserving the system of majority rule and exclusive representation that is the foundation for many states’ public sector labor relations statutes.
Catherine L. Fisk: Barbara Nachtrieb Armstrong Professor of Law, University of California, Berkeley.
Martin H. Malin: Professor and Co-director, Institute for Law and the Workplace, Chicago-Kent College of Law, Illinois Institute of Technology. The authors gratefully acknowledge comments from Marion Crain, William Herbert, Michael Oswalt, John Rumel, Joseph Slater, and Aaron Tang, as well as from a large number of lawyers representing governments, unions, and management, including Thomas Brooks, Judith Scott, Sarah Cudahy, Scott Nobly, Phillip Roberts, David Rosenfeld, Don Slesnick, Alaine Williams, and John Wirenius. Responsibility for errors remains with us.
. Air Line Pilots Ass’n Int’l v. O’Neill, 499 U.S. 65 (1991) (holding that the duty of fair representation applies to contract negotiation); Vaca v. Sipes, 386 U.S. 171 (1967) (holding that the duty of fair representation applies to contract enforcement).
. See 29 U.S.C. §§ 411–15 (2012); Sheet Metal Workers’ Int’l Ass’n v. Lynn, 488 U.S. 347, 358 (1989) (explaining union democracy provisions of federal labor law and holding that elected union officials cannot be removed from office in retaliation for expressing views critical of proposed dues increase).
. The Seventh Circuit observed, in upholding a Wisconsin law that stripped collective bargaining rights from employees whose unions opposed Scott Walker’s gubernatorial candidacy while preserving the rights of employees whose unions supported Walker, “political favoritism is a frequent aspect of legislative action. . . . [T]here is no rule whereby legislation . . . becomes constitutionally defective because one of the reasons the legislators voted for it was to punish those who opposed them during an election campaign. . . . Indeed one might think that this is what election campaigns are all about: . . . the winners get to write the laws.” Wisconsin Educ. Ass’n Council v. Walker, 705 F.3d 640, 654 (7th Cir. 2013). Compare that with Barton Brands Ltd. v. N.L.R.B., 529 F.2d 793 (7th Cir. 1976), and Truck Drivers & Helpers, Local Union 568 v. N.L.R.B., 379 F.2d 137 (D.C. Cir. 1967), both of which suggest or hold that elected union leaders violate the duty of fair representation by adopting policies to favor those who supported their election and punish those who opposed them.
. Janus v. Am. Fed’n of State, Cty, & Mun. Emps, Council 31, 138 S. Ct. 2448, 2467 (2018) (observing that no union is compelled to seek the designation of exclusive representation); id. at 2468–69 (suggesting that unions could refuse to represent nonmembers in grievance processing).
. Aaron Tang, Public Sector Unions, the First Amendment, and the Costs of Collective Bargaining, 91 N.Y.U. L. Rev. 144, 183–90 (2016) [hereinafter Tang, Public Sector Unions]. Tang has since modified his proposal, proposing instead a rebate system in which unions collect dues but then rebate a portion to members based on the amount of money received from the government. See Aaron Tang, Life After Janus, 119 Colum. L. Rev. 677 (2019) [hereinafter Tang, Life After Janus]. Both of Tang’s proposals differ somewhat from the approach theorized by Benjamin I. Sachs, who argues that it has been a mistake to consider fair share or agency fees deducted from paychecks as being the employees’ money in the first place. Rather, agency fees are state money that are paid to the union for services the state requests, such as collective bargaining and contract enforcement. Benjamin I. Sachs, Agency Fees and the First Amendment, 131 Harv. L. Rev. 1046, 1075 (2018).
. These positions are spelled out in two documents issued jointly in July 2018 by the national offices of the NEA, AFT, AFSCME, and SEIU. AFSCME et al., Public Policy Priorities for Partner Unions (2018) [hereinafter Public Policy Priorities], http://nashtu.us/wp-content/uploads/2018/05/Maryann-Parker-Partner-Unions.pdf [https://perma.cc/QSC5-FXUH]; AFSCME et al., Together We Rise: Bargaining Proposals to Promote Membership Sign-Up, Unit Stability, Bargaining Unit Communications, and Alternative Dues Payment Systems (2018).
. Even under section 8(a)(3) of the NLRA, which expressly authorizes unions and employers to require employees to become members after 30 days of employment, the most that can be required is so-called “financial core” membership, which is functionally the same as paying a “fair share” or “agency fee.” NLRB v. Gen. Motors, 373 U.S. 734 (1963). See generally Catherine L. Fisk & Benjamin I. Sachs, Restoring Equity in Right-to-Work Law, 4 UC Irvine L. Rev. 857 (2014).
. Private sector unions governed by the NLRA remain able to charge fair share fees, because there is no constitutional issue in a contract between a private employer and a union. Commc’ns Workers of Am. v. Beck, 487 U.S. 735 (1988). Under the Railway Labor Act (RLA), which governs railway and airline employees, fair share fees remain constitutionally permissible, although that rule may be vulnerable if the Supreme Court adheres to an old and dubious ruling that RLA preemption of state right-to-work laws makes union security provisions in railway collective bargaining agreements subject to constitutional scrutiny. See Int’l Ass’n of Machinists v. Street, 367 U.S. 740 (1961); Railway Emps.’ Dept. v. Hanson, 351 U.S. 225 (1956).
. Janus v. Am. Fed’n of State, Cty, & Mun. Emps, Council 31, 138 S. Ct. 2448, 2491 (2018) (Kagan, J., dissenting) (citing Ichniowski & Zax, Right–to–Work Laws, Free Riders, and Unionization in the Local Public Sector, 9 J. Labor Economics 255, 257 (1991)).
. Id. § 20-29-5-8(b); 560 Ind. Adm. Code § 2-2.1-20 (2018). Where less than a majority of the bargaining unit are union members, the employer provides the Indiana Educational Employment Relations Board (IEERB) with a list of all bargaining unit members and their work email addresses. The IEERB sends all bargaining unit members an email advising them of their right to decertify their union or to change exclusive representatives by filing a petition supported by 20% of the bargaining unit. IEERB, Guide to Exclusive Representative Affidavit and Teacher Letter (2018), https://www.in.gov/ieerb/files/2018%20ERO%20Guide.pdf [https://perma.cc/A7WJ-8ZVR].
. Report Builder: ERO Upload Report Report, Ind. Gateway (2018), https://gateway.ifionline.org/report_builder/Default3a.aspx?rpttype=collBargain&rpt=ero_uploads_public&rptName=ERO%20Upload%20Report [https://perma.cc/VAF5-2CW7].
. The wage premium is the basis for Professor Sachs’ argument that courts erred in imagining that compulsory fees were the union property in the first place. Sachs, supra note 13, at 1063–69. In his analysis, absent majority representation, the workers will be paid less. The small part of the wage premium that workers pay in the form of compulsory fees is not a loss to the employees compared to what they would be paid if there were no union.
. See Bloomberg, Group Funded by Conservative Billionaires Launches Anti-Union Campaign Following Supreme Court Ruling, L.A. Times (June 28, 2018) [hereinafter Bloomberg, Anti-Union Campaign], https://www.latimes.com/business/la-fi-freedom-foundation-20180628-story.html [https://perma.cc/J79B-NHVG].
. See Josh Eidelson, Besieged American Unions Face New Conservative Legal War, Bloomberg (Sept. 24, 2018), https://www.bloomberg.com/news/articles/2018-09-24/with-u-s-labor-under-siege-union-opponents-launch-new-attack [https://perma.cc/PX3N-5HAD].
. Bloomberg, Anti-Union Campaign, supra note 49. Shortly after the Court granted certiorari in Janus, the Freedom Foundation, anticipating a victory, sent out a fundraising letter seeking donations to fund a campaign to lobby workers to resign their union memberships. The letter urged that overturning Abood “should take government unions out of the game for good – yet we know the unions won’t go away without a fight . . . They won’t go away until we drive the proverbial stake through their hearts and finish them off for good.” Fundraising Letter from Freedom Foundation (Oct. 2017) (emphasis in original), https://www.scribd.com/document/379234042/Freedom-Foundation-Letter https://perma.cc/66AW-QFAY.
. Members-only representation could take two other forms not discussed here. One is that a union represents its members only until it gains majority support and then becomes the exclusive representative of all. This form of members-only representation has been proposed as a solution to the difficulties of organizing. Charles J. Morris, The Blue Eagle at Work: Reclaiming Democratic Rights in the American Workplace (2005); Catherine L. Fisk & Xenia Tashlitsky, Imagine a World Where Employers Are Required to Bargain with Minority Unions, 27 A.B.A. J. Lab. & Emp. L. 1 (2011). This form of minority representation is currently lawful in the private sector, although the employer has no duty to bargain with a minority union. Retail Clerks Int’l Ass’n v. Lion Dry Goods, 369 U.S. 17, 29 (1962) (holding that a members-only contract is enforceable under section 301 of the NLRA); Consol. Edison v. N.L.R.B., 305 U.S. 197, 237 (1938) (holding that “in the absence of . . . an exclusive agency the employees represented by the Brotherhood, even if they were a minority,” had a statutory right to join a union and have it contract on their behalf); Dana Corp., 356 N.L.R.B. No. 49, at 256 (2010) (suggesting in dictum that members-only bargaining is permissible); Memorandum from Office of the Gen. Counsel, N.L.R.B., Advice Mem. GC 07-02, Dick’s Sporting Goods (June 22, 2006). However, this is not a solution to the Janus problem because the union that gains majority support becomes an exclusive representative subject to the usual rules governing union security. Another form of members-only representation, which is being sought in litigation by union opponents, would prohibit a union from negotiating on behalf of any employee who rejects union representation. See infra note 59.
. The same interest groups that succeeded in having the Court declare the entire public sector an open shop as a matter of constitutional law, are now seeking to extend Janus to relegate the entire public sector to members-only representation. See, e.g., Bierman v. Dayton, 900 F.3d 570 (8th Cir. 2018), cert. denied, 139 S. Ct. 2043 (2019); Uradnik v. Inter Faculty Org., No.18-1895, 2018 WL 4654751 (D. Minn. Sept. 27, 2018), aff’d, No. 18-3086 (8th Cir. Dec. 3, 2018), cert. denied, 139 S. Ct. 1618 (2019).
. Cf. Cal. Fed’n of Teachers v. Oxnard Elementary Sch., 272 Cal. App. 2d 514, 543 (1969) (describing operation of the Winton Act and rejecting various legal challenges to it notwithstanding differences in the rights given to employee groups and noting the absence of evidence that different treatment was based on invidious discrimination).
. Stats. 1965 ch. 2041, repealed Stats. 1975, ch. 961, formerly codified at Cal. Educ. Code § 13080-88. See Pacific Leg. Found. V. Brown, 29 Cal. 3d 168, 175-78 (1981) (explaining the evolution of California public sector labor law including the repeal and replacement of the Winton Act to replace members-only representation with a system of exclusive representation and mandatory bargaining).
. Catherine Fisk, Challenge to ‘Fair Share’ Union Fees Unfair and Unworkable, Edweek.org (Oct. 6, 2015), https://blogs.edweek.org/edweek/on_california/2015/10/challenge_to_fair_share_union_fees_unfair_and_unworkable.html [https://perma.cc/XJ6K-KUAV].
. California abandoned members-only representation because neither teachers, students, nor school administrators benefited from the difficulty in negotiating fair and consistent terms of employment across schools or districts. See Ophelia H. Zeff, California’s Alternative to Collective Bargaining for Teachers: The Winton Act, 1965-1974, and Proposals for Change, 5 Pac. L. J. 698 (1974) (collecting cases interpreting Winton Act, assessing effectiveness of Act in peacefully resolving conflict between teachers and school districts and in improving teacher salaries, and recommending abandoning the proportional representation Winton Act system in favor of exclusive representation and collective bargaining); Fisk, supra note 63 (explaining that California school districts found the proportional representation system of the Winton Act to be expensive and cumbersome, to be an obstacle to creating uniform policies across schools, and to generate rather than resolve conflict among different groups of teachers).
. Chris Brooks, The Cure Worse than the Disease: Expelling Freeloaders in an Open-Shop State, New Labor Forum (Aug. 2017), https://newlaborforum.cuny.edu/2017/08/24/the-cure-worse-than-the-disease [https://perma.cc/RK9E-KWV8].
. Edgar Mendez, In Wake of Act 10, School Districts Changing Teacher Pay Formulas, Milwaukee J. Sentinel (Aug. 18, 2014), http://archive.jsonline.com/news/education/in-wake-of-act-10-school-districts-changing-teacher-pay-formulas-b99321049z1-271617971.html [https://perma.cc/7L5D-J2LN].
. Id. Unsurprisingly, there has been a fair amount of partisan controversy over the effects of Act 10, with some studies funded by conservative organizations finding a decline in teacher salaries but no decline in the number or quality of teachers as compared to other states. Other studies have found that Act 10 seemingly exacerbated a teacher shortage and inequalities among school districts. See Annysa Johnson, Act 10 Impact on Public Education Muted, Study Says, Post Crescent (June 21, 2016), https://www.postcrescent.com/story/news/education/2016/06/21/act-10-impact-public-education-muted-study-says/86212178 [https://perma.cc/X4X9-NHAB] (reporting the findings and criticisms of one such study).
. Id. at ¶ 25. In linking the lawsuit to Janus¸ Local 150’s President wrote, “It isn’t fair for dues-paying members to subsidize the representation of those who want to cheat the system and save a buck, and our belief in fair treatment of members will be applied. If you have a right not to associate with us, we also have a right not to associate with you, and we will exercise that right. If a bargaining unit has multiple members exercise this choice, we will be forced to consider whether or not we can continue to represent the group at all.” James M. Sweeney, Fight Back (July 2018), http://local150.org/presidents-corner [https://perma.cc/7P5D-J884]. The US District Court for the Northern District of Illinois denied the state defendants’ motion to dismiss, but ruled only with respect to standing and ripeness, not with respect to the First Amendment issues that the lawsuit raises. Sweeney v. Madigan, 359 F. Supp. 585 (N.D. Ill. 2019).
. 420 U.S. 251 (1975); see, e.g., Redwoods Cmty. Coll. Dist. v. Public Emp. Rel. Bd., 205 Cal. Rptr. 523 (Ct. App. 1984); Wayne-Westland Educ. Ass’n v. Wayne-Westland Cmty. Schs., 439 N.W.2d 372 (Mich. Ct. App. 1989); In re Univ. of Medicine and Dentistry of N.J., 144 N.J. 511 (1996); Commonwealth v. Pa. Lab. Rel. Bd., 677 A.2d 721 (N.J. 2007); Warwick Sch. Comm. v. State Lab. Rel. Bd., No. M.P. NO. 13058, 1979 WL 196143 (R.I. Sup. Ct. Dec. 11, 1979) (unpublished opinion). But see City of Round Rock v. Rodriguez, 399 S.W.3d 130 (Tex. 2013).
. Fla. Stat. Ann. Tit XXXI, § 447.401 (West 2018). In Sherry v. United Teachers of Dade, 368 So. 2d 445 (Fla. Dist. Ct. App. 1979), a nonmember filed a grievance which the union refused to process unless she paid a fee according to a fee schedule the union had for nonmembers. Sherry sued claiming that the union’s refusal to process her grievance violated the right to work provision of the Florida Constitution. The court appears to have held that Sherry lacked standing. It wrote, “It is clear that the gravamen of Sherry’s complaint is that the U.T.D. would not process her grievance free of charge, although she is not a dues paying member and could process the grievance by herself. We, therefore, conclude that in the posture of this case the issue of the constitutionality of Section 447.401, Florida Statutes (1977) is not properly presented and Sherry lacks the requisite interest to bring the instant action under Section 86.011 et seq. Florida Statutes (1977).” Id. at 447.
. See Leslie Postal, Florida Teachers Union Sues State Over ‘Busting Unions’ Law, Orlando Sentinel (July 3, 2018), https://www.orlandosentinel.com/news/education/os-teachers-union-florida-lawsuit-20180702-story.html [https://perma.cc/4Y6X-BJN6].
. Complaint at IV, Florida Educ. Ass’n et al. v. Poole, No. 74352318 (Fla. Cir. Ct. 2018). As recent evidence from Indiana shows, nonmembership in the union does not equate to opposition to union representation. See supra notes 34–37 and accompanying text.
. Patrick Wright, Finding Quality Evidence of Union Survivability in the Absence of Agency Fees: Is the Current Population Center’s Public Sector Unionism Data Sufficiently Reliable, 2017 U. Chi. Legal F. 563, 584 (2017).
. See Martin H. Malin, Does Public Employee Collective Bargaining Distort Democracy: A Perspective from the United States, 34 Comparative Lab. L. & Pol’y J. 277, 293 (2013) (citing observations of University of Nebraska Professor Steven Willborn).
. The lawyer said: Provision of legal assistance in any proceeding is a privilege of membership in the NSEA and subject to the Legal Assistance Policies of the NSEA. Generally, those policies do not contemplate the provision of legal assistance in grievances. Any assistance provided a member in the presentation or administration of a grievance is typically handled by local association officers or NSEA non-legal staff. Legal assistance is generally limited to due process dismissals, professional practices disciplinary proceedings and judicial enforcement of statutory and constitutional rights. In Nebraska, educator disciplinary proceedings are not subject to grievance procedures. Grievances are typically limited in educational employment to disputes arising over economic terms of employment. Reprimands, suspensions and dismissals are subject to statutory criteria and procedures. Email from Scott Norby, Norby & Wade, LLP, to Martin Malin (Dec. 15, 2017).
. This phenomenon, of course, is not necessarily universal. For example, in Nebraska, it appears that teachers generally value access to union legal representation highly, resulting in high levels of membership density for teacher unions. See supra note 100 and accompanying text.
. See Martin H. Malin & Robert R. Ladenson, Privatizing Justice: A Jurisprudential Perspective on Labor and Employment Arbitration from the Steelworkers Trilogy to Gilmer, 44 Hastings L. J. 1187, 1192–94 (1993); see also Martin H. Malin, The Supreme Court and the Duty of Fair Representation, 27 Harv. C.R.-C.L. L. Rev. 127, 171–77 (1992).
. Perhaps because of the significant inroad that members-only representation in contract administration may have on exclusive representation and worker solidarity, AFSCME, the American Federation of Teachers, the National Education Association, and SEIU, which together represent the lion’s share of public employees, have eschewed it. See supra note 15 and accompanying text.
. For example, solidarity is generally very high among firefighters because they live together at the fire house and because, in performing their job duties, they are responsible for each other’s lives. It is also often very high among police officers. This may explain why Rhode Island limited members-only grievance representation to police and firefighters.
. Samuel Estreicher, The Paradox of Federal-Sector Labor Relations: Voluntary Unionism Without Collective Bargaining over Wages and Employee Benefits, 19 Emp. Rts. & Emp. Pol’y J. 283, 299 tbl.3 (2015).
. Bureau of Alcohol, Tobacco & Firearms v. F.L.R.A., 464 U.S. 89, 91 (1983). However, it does not require the employer to pay employee travel and per diem expenses, id. at 107, although unions may negotiate for that. See, e.g., Decision and Order, Dept. of Interior, Nat’l Park Serv. & Nat’l Treas. Emps. Union, 10 F.S.I.P. 119 (2011). The FSLMRS further empowers the Federal Labor Relations Authority to require official time for any employee appearing for or on behalf of a union in a proceeding before the Authority. 5 U.S.C. § 7131(c) (2012).
. See, e.g., Decision and Order, Dept. of Labor & Local 12, Am. Fed’n of Gov’t Emps., 12 F.S.I.P. 104 (2013); Decision and Order, Dep’t of Homeland Sec., Bureau of Customs & Border Protection and Nat’l Treas. Emps. Union, 10 F.S.I.P. 010 (2011) (one of the authors – Malin – was the FSIP Panel member who decided this case).
. See Collective Bargaining Agreement between the Postal Service and American Postal Workers Union, art. 17.4 (effective May 21, 2015 – Sept. 20, 2018), Collective Bargaining Agreement between the Postal Service and the National Postal Mailhandlers Union § 17.4 (effective May 21, 2016 – Sept. 20, 2019); Collective Bargaining Agreement between the Postal Service and the National Association of Letter Carriers art. 17-4 (2016–19).
. See Cheatham v. DiCiccio, 379 P.3d 211, 213 (Ariz. 2016) (holding that provision in Phoenix police collective bargaining agreement granting official time to employees to perform representational functions, including 100 percent official time to several, did not violate the Gifts Clause of the Arizona Constitution); Idaho Freedom Found. v. Indep. Sch. Dist. of Boise City, No. CV-OC-2015-15153 (Idaho Dist. Ct. Oct. 25, 2016) (upholding against state constitutional attack provision of Boise School District collective bargaining agreement giving union president release time with employer paying salary and benefits equal to a first year teacher and union paying the rest); Martin H. Malin, Life After Act 10?: Is There a Future for Collective Representation of Wisconsin Public Employees?, 96 Marquette L. Rev. 623, 657 (2012) (discussing practices in Tennessee).
. See, e.g., Opinion and Decision, Dept. of Homeland Sec., U.S. Coast Guard & Local 3313, Am. Fed’n Gov’t Emps., 12 F.S.I.P 157 (2013); Opinion and Decision, Dep’t of Veterans Affairs, 60 F.L.R.A. 479 (2004) (enforcing arbitration award which had found that agency breached its agreement with union concerning union office space); Opinion and Decision, United States Geological Survey, Caribbean District Office, 53 F.L.R.A. 1006 (1997) (holding union office space to be a substantively negotiable condition of employment); Dept. of Health and Human Servs., Region IX and Chapter 212, Nat’l Treas. Emps. Union, 89 F.S.I.P. 157 (1990); Opinion and Decision, Dept. of the Army Lexington – Blue Grass Army Depot, 34 F.L.R.A. 247 (1990) (holding that agency breached it duty to bargain by unilaterally terminating the union’s office space).
. 5 U.S.C. § 5596(b)(1)(A)(ii) (2012). The act requires that the award of attorney fees be made in accordance with the standards set forth in 5 U.S.C. § 7701(g), which governs attorney fee awards by the Merit Systems Protection Board (MSPB). That section requires that the employee be the prevailing party and that the adjudicator “determine that payment by the agency is warranted in the interest of justice, including any case in which a prohibited personnel practice was engaged in by the agency or any case in which the agency’s action was clearly without merit.” Allen v. U.S. Postal Serv., 2 M.S.P.B. 582, 586 (1980). In Allen, the MSPB detailed a list of situations where an award of attorney fees would be in the interest of justice. The MSPB made clear that the list was illustrative and that awards of attorney fees could meet the interests of justice requirement in other circumstances. The MSPB opined that attorney fee awards would be in the interest of justice where the employer had committed a prohibited personnel practice, the employer’s actions were clearly without merit or wholly unfounded or the employee was substantially innocent of the charges, the employer’s action was brought to harass the employee or to pressure the employee to act in a particular manner, the employer committed a gross procedural violation which prejudiced the employee or prolonged the proceedings and the employer knew or should have known that it would not prevail on the merits. Id. at 593.
. See Joe Davidson, Social Security, HUD Act on Trump’s Orders in Move to Emasculate Unions, Wash. Post (June 15, 2018), https://www.washingtonpost.com/news/powerpost/wp/2018/06/15/social-security-hud-act-on-trumps-orders-in-move-to-emasculate-unions/?utm_term=.6be7a7ce1945 [https://perma.cc/5VBF-Y3HF].
. See Noam Scheiber, Federal Workers Brace for New Push on Trump Anti‑Labor Goals, N.Y. Times (Sept. 2, 2018), https://www.nytimes.com/2018/09/02/business/economy/trump-labor.html [https://perma.cc/W4EL-LUEL].
. See supra note 133 and accompanying text; See also Soc. Sec. Admin. & Am. Fed’n Gov’t Emps., 2019 FSIP 19 (May 29, 2019) (eliminating most of the union’s free office space and greatly reducing the union’s official time).
. See George Brooks, The Strengths and Weaknesses of Compulsory Unionism, 11 N.Y.U. Rev. L. & Soc. Change 29 (1982); see also Joint Legislative Committee on the Taylor Law, 1971–72 Report at 36, New York Legislative (1972), http://www.perb.ny.gov/wp-content/uploads/2018/04/1971JLC-1.pdf [https://perma.cc/D3GB-EV2X] (reporting that most unions favored amending the Taylor Law to allow agency shop but some “insist that agency shop fees mandated by an agreement would remove the pressure on an organization to press for goals with universal appeal to the employees in the unit”).
. 138 S. Ct. at 2465 (adopting an “exacting scrutiny” test to just the permissibility of compelled subsidies, and under this level of scrutiny “a compelled subsidy must serve a compelling state interest that cannot be achieved through means significantly less restrictive of associational freedoms”). Although the Court said this is a “more permissive standard” than strict scrutiny, it is unclear how, as it is a “compelling interest” test. Id.
. See also Daniel Hemel & David Louk, Is Abood Irrelevant, 82 U. Chi. L. Rev. Dialogue 227, 236 & n.51 (2015). In Executive Order 13,837, President Trump attempted to restrict federal sector unions use of official time to exclude grievance processing. See supra note 134 and accompanying text.
. See supra note 133 and accompanying text (discussing USDA Rural Development and AFSCME Local 3870, 17 F.S.I.P. 060 (2018)). Although some statutes provide that one labor board member represents labor, one represents management, and another represents the public, all are appointed by the governor. Although it is common for appointees to serve fixed terms and for the governor to be prohibited from firing them (but not always, the President has statutory authority to remove FSIP members at will), but there are ways around that. For example, in Illinois it is common for a new governor to get the legislature to amend the statute by abolishing the existing board and then creating a new board which the governor gets to fill completely with his appointees. See 5 Ill. Comp. Stat. §§ 15/5(a-5)(b), 315/5.1(a), (b) (2018).
. Professor Tang might argue that his proposal encourages workers’ engagement with their union much like agency fees, as both reduce the marginal cost of union membership. Under Tang’s proposal, unions are reimbursed for their costs of collective bargaining and contract administration, i.e., they are reimbursed in approximately the same amount as they would have been able to charge non-members under agency fee regimes. Unions may then rebate to their members each member’s pro rata share of the reimbursement, thereby reducing the member’s dues to the difference between dues and agency fees under the agency-fee regime. Arguably, this reduced marginal cost of membership will motivate most workers to pay dues similar to agency fees. But there is a difference. In an agency-fee regime, every employee was required to choose to be a member or a fee payer. Confronted with that choice, most employees opted for full membership. They did so because the marginal cost of full membership was low and because that low marginal cost assured them that most of their coworkers would do the same. Under Janus, employees are not required to make any choices. The default is non-membership and employees must affirmatively opt to become members. For that to happen, local union leaders will have to engage workers and educate them about the benefits of union membership and the importance of solidarity. But, experience with official time and other employer subsidies in the federal sector suggests that some, perhaps many, local union leaders, assured of their union’s financial viability due to employer subsidies, will not engage the workers they represent and may see it in their personal interest to keep membership levels low.
. Massachusetts, California, Illinois, Michigan, New York, Connecticut, and Washington, among others. See Sachs, supra note 13, at 1056–57. Minnesota, Pennsylvania, Vermont also ban employer payments to unions. See Hemel & Louk, supra note 151, at 238 n.63.
. Id. at 216 (Tang arguing that in both the agency fee and reimbursement systems the employer has the power to threaten the union’s financial security during negotiations, so there is no greater threat for domination in a reimbursement system).
. See Lee v. N.L.R.B., 393 F.3d 491, 497 (4th Cir. 2005) (finding that requiring workers to wear uniforms showing the union and employer logo side-by-side violated the NLRA even though the CBA required both logos).
. H.B. 923 was introduced in January 2017 and was carried over to the 2018 legislative session, but it did not pass and appears not to have been re-introduced. See Legislative Research: HI HB923 2017 Regular Session, Legiscan (2017), https://legiscan.com/HI/research/HB923/2017 [https://perma.cc/ZK7G-EUWN]. The text of the bill is available at https://www.capitol.hawaii.gov/Archives/measure_indiv_Archives.aspx?billtype=HB&billnumber=923&year=2018 [https://perma.cc/Z7PZ-DN3Q].
. See Hemel & Louk, supra note 151, at 243 (framing the political challenge in terms of “political salience” where the perception of who pays a tax has a strong effect on the acceptability of policy proposals despite two regimes being functionally identical).
. Professor Benjamin Sachs has addressed these concerns in part by suggesting that the alignment of union goals with member values generates union commitment, and that it is possible mandatory fees generate as much resentment as commitment. A direct payment, on the other hand, eliminates the free rider problem along with disgruntled non-supporters, and the union is, of course, still free to charge some level of voluntary dues. Sachs, supra note 13, at 1075.
. Corey Fine & Paul Baktari, Public Sector Union Democracy: A Comparative Analysis, 22 J. L. Research 391, 392–95 (2001) (stating that ten states have union democracy statutes like LMRDA, but only Iowa, Montana, Ohio and Wisconsin statutorily guarantee equal rights to nominate officers, vote in elections and on union business, attend meetings, and participate in decision making. However, every state except South Dakota prohibits both unions and employers from using force or coercion against union members).
. See Cal. Lawyers Ass’n, California Public Sector Labor Relations Chs. 30–31 (Kirsten L. Zerger et al., eds., 2017) (chapters on organizational rights and obligations under California public sector labor relations laws reveal absence of any laws regulating union financial disclosure, elections, secret ballot, trusteeship, or fiduciary duties analogous to LMRDA). Compare with section 401(g), 29 U.S.C. § 481(g) (2012), of the LMRDA.
. Under EERA (public school employees), HEERA (higher education employees), and TEERA (transit employees) “employee organizations” that are certified or otherwise recognized must maintain and disclose annually to PERB and employees a report of their financial transactions. Cal. Gov. Code § 3546.5 (EERA) (2018); id. § 3587 (HEERA); Cal. Pub. Util. Code § 99566.3 (TEERA) (2018). Under the Dills Act (civil service employees) and remaining statutes only unions that have in place an agency fee agreement are required to report their spending. Cal. Gov. Code § 3515.7(e) (Dills Act); id. § 3502.5(f) (MMBA); id. § 71632.5(f) (Trial Court Employees, TCEPGA); id. § 71814(f) (Trial Court Interpreters, TCIELRA).
. Pub. Emp’t Relations Bd. Order Ad-123, Cal.Teachers Ass’n 5 (1981) (No. SF-CO-134) (finding that petitioner could not access the financial information of the California Teachers Association and the National Education Association even though his local chapter is affiliated with those organizations).
. See, e.g., Liz Perlman, ‘Company Unions’ Deepen Post-Janus Threat to Labor, Capitol Weekly (Sept. 5, 2018), http://capitolweekly.net/company-unions-make-post-janus-threat-labor-even-worse [https://perma.cc/4EJH-26XU].
. See Louis Freedberg, Battle Ramps Up to Convince California School Employees to Withhold Union Fees, EDSOURCE (July 11, 2018), https://edsource.org/2018/battle-ramps-up-to-convince-california-school-employees-to-withhold-union-fees/599841 [https://perma.cc/R5LY-S6T6] (describing a campaign by the Mackinac Center for Public Policy and other organizations to get employees to resign their union membership and looking at the sophisticated information campaign, including staff resources and websites, (mypaymysay.com; optouttoday.com) to facilitate union resignations and revocation of dues authorizations in California, Oregon, Washington, and other states with many unionized teachers).
. For example, in Memorando Especial Conjunto Num. 2018-02, dated July 18, 2018, the Puerto Rican government determined that Janus requires public employee union members to be able to resign their membership at any time by completing and filing a “Solicitud de Desafiliacion del Representante Sindica,” which will terminate the payroll deduction of union dues or agency fees.
. The law provides: Employees who have authorized the payroll deduction of fees to employee organizations may revoke such authorization by providing written notice to their public employer during the 10 days following each anniversary date of their employment. Within five days of receipt of notice from an employee of revocation of authorization for the payroll deduction of fees, the public employer shall provide notice to the employee organization of an employee’s revocation of such authorization. An employee’s notice of revocation of authorization for the payroll deduction of employee organization fees shall be effective on the 30th day after the anniversary date of employment.” N.J. Stat. Ann. § 52:14-15.9e (West 2018).
. Hawai’i adopted a similar approach, requiring notice 30 days before the anniversary of an employee’s dues deduction authorization to revoke the authorization. Haw. Rev. Stat. § 89-4 (2018). California, Delaware, and New York allow the union to specify in the dues deduction authorization the window during which it may be revoked. See Cal. Educ. Code § 45060 (2018); Del. Code Ann. Tit. 19, § 1304 (2018)); N.Y. Civ. Serv. L. § 208(b)(i).
. See McCahon v. Penn. Turnpike Comm’n, 491 F. Supp. 2d 522, 527 (M.D. Pa. 2007) (granting preliminary injunction against enforcement of a maintenance of membership provision in a CBA that allowed members to resign from union only during a fifteen-day window prior to expiration of a three-year CBA, reasoning that the term compelled membership in violation of the First Amendment).
. In an op-ed. anticipating the outcome in Janus, Professor Samuel Estreicher suggested unions consider such an approach. Samuel Estreicher, How Unions Can Survive a Supreme Court Defeat, Bloomberg Opinion (May 2, 2018), https://www.bloomberg.com/view/articles/2018-03-02/how-unions-can-survive-a-supreme-court-defeat [https://perma.cc/HHW4-MFHS]. We analyze this approach in detail below and reject it.
. In Janus, the agency fee was 78.06 percent of union dues, which amounted to $44.58 per month. Janus v. Am. Fed’n of State, Cty, & Mun. Emps, Council 31, 138 S. Ct. 2448, 2461 (2018). Thus, the marginal cost of union membership was $12.43 per month.
. Id. Congress added Section 19 in 1974 when it amended the NLRA to cover health care employers, many of which were affiliated with the Catholic Church or other religions. Wilson v. N.L.R.B., 920 F.2d 1282, 1286 (6th Cir. 1990). As the amendment was originally written, only health care employees could be religious objectors. When Congress extended the religious exemption in 1980 to apply to all employees, it added the language about charging for grievance handling. H.R. Rep. 96-496 (1979). An earlier version of what became Section 19 did not include the sentence about paying the union for the cost of using the grievance-arbitration process. See H.R. Rep. 95-768 (1977). A court has suggested that Congress extended the religious objector exemption to all employees in order to reconcile it with the employer’s duty to accommodate religious practices of employees under Title VII. Int’l Ass’n of Machinists v. Boeing Co., 833 F.2d 165, 170 (9th Cir. 1987). There are a number of cases under Title VII as to whether a union reasonably accommodates, and therefore avoids discrimination, when it requires an objector to pay full dues to the charity as opposed to the lower amount paid by employees who exercise their rights under Communication Workers of America v. Beck, 487 U.S. 735 (1988) to pay a reduced fee that excludes that portion of dues which covers expenditures not related to collective bargaining. Compare Madsen v. Associated Chino Teachers, 317 F. Supp. 2d 1175 (C.D. Cal. 2004) and O’Brien v. City of Springfield, 319 F. Supp. 2d 90 (D. Mass. 2003); see Christopher J. Conant, Toward a More Reasonable Accommodation for Union Religious Objectors, 37 McGeorge L. Rev. 105 (2006) (arguing that objectors should only have to pay an amount equal to agency fees).
. NLRA section 19 says, “If such employee . . . requests the labor organization to use the grievance-arbitration procedure on the employee’s behalf, the labor organization is authorized to charge the employee for the reasonable cost of using such procedure.” 29 U.S.C. § 169. The Janus majority quoted similar language from California public sector labor law and also cited a similar provision in Illinois’ public sector law. Janus, 138 S. Ct. at 2469 n.6 (citing Cal. Govt. Code Ann. § 3546.3 (West 2010);Ill. Comp. Stat., ch. 5, § 315/6(g) (2016)).
. Note that the requirement that religious objectors pay an amount equal to dues to a charitable organization does not solve the free-rider problem. Religious objectors remain entitled to union representation even though they are not contributing to the costs of that representation. The authorization of unions charging religious objectors for grievance representation alleviates the free-rider problem.
. See United States v. Seeger, 380 U.S. 163 (1965) (explaining the religious objector exemption to the Selective Service Act governing the military draft). See generally Selective Service Mission, Selective Serv. Syst., www.sss.gov/about/alternative-service [https://perma.cc/KF9W-EW3U].
. In Southworth, the Court remanded for consideration of whether an alternative approach to the application procedure whereby a student organization could obtain finding through a student referendum operated without in a viewpoint neutral manner. Id. at 235.
. For example, the majority opinion described Illinois’ budget problems, poor credit rating, and allegedly unsustainable pension and healthcare costs at considerable length, attributing them to public employee collective bargaining agreements and suggesting that “employment-related debt is squeezing core programs in education, public safety, and human services, in addition to limiting the State’s ability to pay its bills.” Janus v. Am. Fed’n of State, Cty, & Mun. Emps, Council 31, 138 S. Ct. 2448, 2475 (2018) (internal punctuation omitted). During oral argument, Justice Kennedy made clear that his antipathy to public employee collective bargaining drove his thinking about the case: 1. FRANKLIN [Solicitor General of Illinois]: You know, the state’s interest here, if I can spend just a few moments talking about that, is, first, we have an interest in dealing with a single spokesman for –e – for the employees. Second, we have an interest in imposing on that spokesman a legal duty to represent everyone. But–s — but as regards agency fees, they are complementary to those first two interests. They serve our managerial interests in two ways. First, they allow us to avoid a situation where some employees bear the cost of representing others who contribute nothing. That kind of two-tiered workplace would be corrosive to our ability to cultivate collaboration, cohesion, good working relationships among our personnel. Second, independent of that, we have an interest at the end of the day in being able to work with a stable, responsible, independent counterparty that’s well-resourced enough that it can be a partner with us in the process of not only contract negotiation — JUSTICE KENNEDY: It can be a partner with you in advocating for a greater size workforce, against privatization, against merit promotion, agai–t — for teacher tenure, for higher wages, for massive government, for increasing bonded indebtedness, for increasing taxes? That’s — that’s the interest the state has? . . . [D]oesn’t it blink reality to deny that that is what’s happening here? Transcript of Oral Argument at 46–47, Janus, 138 S. Ct. at 2475 (No. 16-1466), https://www.supremecourt.gov/oral_arguments/argument_transcripts/2017/16-1466_bocf.pdf [https://perma.cc/QR3P-NS49].
. The Agreement between the Portsmouth School Board and the Association of Portsmouth Teachers (effective July 1, 2014 – June 30, 2018) provides in Article 3: It is recognized that the negotiations for, and administration, of the AGREEMENT entails expenses which appropriately should be shared by all employees who are beneficiaries of this AGREEMENT. To this end, if an employee in the bargaining unit does not join the ASSOCIATION, such employee will, as a condition of employment by the BOARD, execute an authorization for the deduction of a “representation fee” which shall be a sum equivalent to membership dues and assessments required to be paid by members of the ASSOCIATION, which sum shall be retained for a scholarship fund. The committee to award the scholarship shall be made up of two administrators, two members of the ASSOCIATION, and one member of the “representation fee” group. The scholarship shall be given in the name of the ASSOCIATION OF PORTSMOUTH TEACHERS. The ASSOCIATION agrees to indemnify and defend the BOARD, the Portsmouth School District and SAU, the City of Portsmouth and any employee, official, agent, representative or attorney of any such entity from any claim arising out of or in any way connected with the “representation fee.” Agreement between the Portsmouth School Board and the Association of Portsmouth Teachers at 5, City of Portsmouth, http://files.cityofportsmouth.com/hr/contracts/School/APTAGREEMENToneyearextension2019.pdf [https://perma.cc/NG3R-7NGM].
. See Elkouri & Elkouri, supra note 126, at 1-14 to 1-15. Occasionally, an agreement will provide that the losing party pays the entire fee of the arbitrator. Id. There are other variations as well. See, e.g., Internal Revenue Service and National Treasury Employees Union, 2016 National Agreement, art. 43, § 4(A)(1) (providing for each party to pay 50 percent of the arbitrator’s fees and expenses unless the grievant substantially prevails in which case the employer pays 75 percent).
. We placed premiums in quotation marks because these are not payments made to an insurance company; instead they represent each employee’s pro rata cost of the projected costs of the benefit for the coming year.
. Note that this approach is significantly different from refusing to process grievances for nonmembers unless they pay the costs of their representation. This approach shares the costs of a process that is valuable to the employer and all of the employees while providing a valuable benefit to union members that helps defray some of the costs of membership. We discuss other forms of members-only benefits below in Part C.
. There has been considerable controversy in the non-union sector where some employers have, as a condition of employment, required their employees to agree to arbitrate any claims arising out of the employment relationship and to pay half of the arbitrator’s fee. The Supreme Court has recognized that such fee-splitting requirements could act as a barrier to an employee’s ability to effectively vindicate statutory rights in the arbitral forum and, when it does, such fee-splitting requirements will not be enforced. Green Tree Fin. Corp.-Ala. v. Randolph, 531 U.S. 79 (2000). At one time, the D.C. Circuit held that to be enforceable, employer-imposed arbitration mandates must provide that the employer would be responsible for the entire arbitrator fee. Cole v. Burns Int’l Sec, Servs., 105 F.3d 1465 (D.C. Cir. 1997). Although the holding of Cole has been superseded by the decision in Randolph, the American Arbitration Association requires that employers using its services agree to pay the entire arbitrator fee. See Employment/Workplace Fee Schedule, Am. Arbitration Ass’n (Oct. 1, 2017), https://adr.org/sites/default/files/Employment_Arbitration_Fee_Schedule.pdf [https://perma.cc/4X3L-63Z7]. Our proposal differs significantly from non-union employer-imposed mandates that employees share the costs of arbitration of their individual claims. We do not propose that individual grievants be assessed half of the arbitrator’s fee and related costs. Rather, we propose that all employees in the bargaining unit be assessed their pro rata share of one-half the projected costs of the negotiated benefit of access to grievance arbitration. The union would cover the assessments made on its members by direct payment or reimbursement. For example, assume in a bargaining unit of 500 employees, the parties project that there are likely to be ten arbitrations in the coming year with arbitrator fees and related costs estimated to be $6,000 each, for a total projected cost of $60,000. Each employee in the bargaining unit would be assessed a pro rata share of half the cost. In other words $30,000 ÷ 500 employees = $60.00 per employee, with the union covering the $60.00 assessment for each of its members. Our proposal does not provide for the ten grievants whose cases are arbitrated to be responsible for $3,000 each—half of the fees incurred in each grievance.
. Treating the arbitration process the same way as other employee benefits with cost sharing between the employer and the employees may facilitate a system of members-only grievance representation. In such a system, nonmembers would have access to the grievance and arbitration procedures but would have to pay the union to represent them or provide, and pay for, their own representation and related costs, including their share of the arbitrator’s fee. Arbitrators may be reluctant to accept appointments in such cases because of insecurity concerning eventual payment by the nonmember. Having the arbitrator paid from a fund to which the employers and employees contribute solves that problem and makes the arbitration procedure more accessible for self-funded nonmembers.
. Among the many kinds of union-provided training, one could include apprenticeship programs, safety trainings, and professional development trainings for teachers and other skilled employees. (The Writers Guild of America, West, for example, runs programs on everything from how to write a teleplay to how to become and remain an effective showrunner (head writer-producer) of a television series). See Television and Digital Media, WGAW https://www.wga.org/members/programs/tv-digital [https://perma.cc/GK3E-MU4A].
. Id. at 1153; Branch v. Commonwealth Emp. Relations Bd., 120 N.E.3d 1163 (Mass. 2019) (rejecting, for lack of state action, constitutional challenge to union rules limiting voting on bargaining representatives, contract proposals, and bargaining strategy to members only).
. Hines v. Anchor Motor Freight, 424 U.S. 554, 564 (1976) (observing that only arbitrary or discriminatory decisions violate DFR); Ford Motor Co. v. Huffman, 345 U.S. 330 (1953) (holding union did not breach DFR by agreeing to grant seniority credit for pre-employment military service); Am. Fed’n of Gov’t Emps. v. FLRA, 812 F.2d 1326 (10th Cir. 1987).
. Am. Fed’n of Gov’t Emps, 812 F.2d at 1328; see also Nat’l Treas. Emps. Union v. F.L.R.A., 800 F.2d 1165 (D.C. Cir. 1986) (holding the union did not violate the duty of fair representation in declining to represent an employee seeking statutory relief under the Civil Service Reform Act for an adverse employer action); cf. Del Casal v. Eastern Airlines, 634 F.2d 295 (5th Cir. 1981) (holding the union violated the duty of fair representation by not representing a non-union member in his dismissal grievance). With respect to public teachers, the structure of California statutes allows unions to provide members-only benefits while avoiding DFR claims in some cases. For example, it may not violate the duty of fair representation if a union declines to represent a nonmember in a proceeding under statutory just case protection because the protection does not arise under the contract and the union does not control access to the proceeding. Cal. Educ. Code §§ 44934, 44938, 44944 (2018).
. For example, a coalition of the Chicago District Laborers Council, IUOE Locals 150 and 399, the Chicago Regional Council of Carpenters, and International Brotherhood of Electrical Workers Local 134 formed the Midwest Coalition of Labor providing an array of members-only benefits. See Member Only Benefits, Midwest Coalition of Lab., https://coalitionoflabor.org [https://perma.cc/U4G5-Z85J]; New Exclusive Member Benefits – Midwest Coalition of Labor, Chi. Laborers’ District Council, (Aug. 21, 2018), http://www.liunachicago.org/about/news/details/2074/new-exclusive-member-benefits-midwest-coalition-of-labor [https://perma.cc/82YK-T7DP]. Another example is drawn from Europe, in which unions play a significant role in administering benefits that are in the US provided (if at all) by government bureaucracies. See Kate Andrias, An American Approach to Social Democracy: The Forgotten Promise of the Fair Labor Standards Act, 128 Yale L. J. 616, 702–04 (2019); Matthew Dimick, Labor Law, New Governance, and the Ghent System, 90 N.C. L. Rev. 319 (2012).
. Examples, besides those in the text, include the work that SEIU Local 721 does with Los Angeles Department of Health Services to provide services to homeless people in Los Angeles. The idea is that unions can seek grants for public-facing work that advances labor-management goals and public services.
. See, e.g., Cal. Gov’t Code § 3558 (2018); Md. H.B. 811, 2018 Gen. Assemb., Reg. Sess. (Md. 2018) (codified as Md. Const. art. II, § 17(b)); Md. H.B. 1017, 2018 Gen. Assemb., Reg. Sess. (Md. 2018) (codified as Md. Const. art. II, § 17(b)); N.J. Stat. Ann. § 34:13A–5.13(c) (West 2018); N.Y. Civ. Serv. L. § 208(4)(a) (McKinney 2018); An Act Providing Labor Unions with Reasonable Access to Current and Newly Hired Public Sector Workers, 2019 Maine H.P. 1063 (adopted June 19, 2019), to be codified as 26 Me. Rev. Stat. Ann. §§ 975, 979-T, 1037; Or. H.B. 2016, § 5, 2019 Leg. Assemb., Reg. Sess. (Or. 2019) (signed into law June 20, 2019).
. Id. Reliance on small groups within the larger organization to maintain membership and solidarity is not restricted to unions. Megachurches have used the technique very effectively to solve collective action problems. See Malcolm Gladwell, The Cellular Church, The Atlantic (Sept. 4, 2005), https://www.newyorker.com/magazine/2005/09/12/the-cellular-church [https://perma.cc/22BS-AE9Z].
. See Michael Wasser, Making the Case for Union Membership: The Strategic Value of New Hire Orientations, Jobs with Justice (2016), http://www.jwj.org/wp-content/uploads/2016/09/JWJEDU_BestPractice_Report_2016_FINAL.pdf [https://perma.cc/7BWX-JS4L].