The Yale Law Journal

October 2013

The Unbundled Union: Politics Without Collective Bargaining

abstract. Public policy in the United States is disproportionately responsive to the wealthy, and the traditional response to this problem, campaign finance regulation, has failed. As students of politics have long recognized, however, political influence flows not only from wealth but also from organization, a form of political power open to all income groups. Accordingly, as this Essay argues, a promising alternative to campaign finance regulations is legal interventions designed to facilitate political organizing by the poor and middle class. To date, the most important legal intervention of this kind has been labor law, and the labor union has been the central vehicle for this type of organizing. But the labor union as a political-organizational vehicle suffers a fundamental flaw: unions bundle political organization with collective bargaining, a highly contested form of economic organization. As a result, opposition to collective bargaining impedes unions’ ability to serve as a political-organizing vehicle for lower- and middle-income groups.

This Essay proposes that labor law unbundle the union, allowing employees to organize politically through the union form without also organizing economically for collective bargaining purposes. Doing so would have the immediate effect of liberating political-organizational efforts from the constraints of collective bargaining, an outcome that could mitigate representational inequality. The Essay identifies the legal reforms that would be necessary to enable such unbundled “political unions” to succeed. It concludes by looking beyond the union context and suggesting a broader regime of reforms aimed at facilitating political organizing by those income groups for whom representational inequality is now a problem.

author. Professor of Law, Harvard Law School. The author thanks David Barron, Cynthia Estlund, Catherine Fisk, Gerald Frug, Jack Goldsmith, Daryl Levinson, and Matthew Stephenson for helpful comments and discussion. The author is also grateful to the participants in the Harvard Law School Faculty Workshop. Nikolas Bowie, William Dreher, Anna Lvovsky, Joshua Segal, and Heather Whitney provided outstanding research assistance.


It is a good time to be wealthy in America and a tough time not to be. This is true not only because of the well-known economicproblems facing low- and middle-income Americans. It is true because the poor and middle class have a major political problem today. The problem is that the government is strikingly unresponsive to their views.1 As Martin Gilens concludes in his study of contemporary American politics, “the preferences of the vast majority of Americans appear to have essentially no impact on which policies the government does or doesn’t adopt.”2

No government, of course, is perfectly responsive to its citizenry, and perfect responsiveness is not even an aspiration of our democratic order.3 But it remains a fundamental democratic commitment that policies enacted by the government reflect the preferences of the polity. To borrow Dahl’s formulation, “a key characteristic of a democracy is the continuing responsiveness of the government to the preferences of its citizens, considered as political equals.”4 The degree of representational inequality that currently defines American political practice is thus a matter of substantial concern.

The wealthy have disproportionate influence over public policy because, to state the obvious, they have more money.5 Because it is the wealthy who make campaign contributions, fund independent electoral expenditures, and pay for lobbyists, policy is more responsive to those with money than to those without it.6 Given the political influence that wealth bestows, scholars and Congress have understandably focused political reform proposals on campaign finance.7 But these attempts to get money out of politics have devolved into a cat-and-mouse game in which political actors bent on avoiding regulation, and a Supreme Court bent on invalidating it, have rendered the reforms ineffectual.8 After all, the Court has now struck down all forms of independent expenditure limitations, and political actors have designed ways to frustrate even the most creative restraints on campaign spending.9

Fortunately, however, money is not the only source of influence in American politics. Political power also flows from political organization, and organization is a source of power available to all income groups.10 As this Essay will suggest, legal interventions designed to facilitate political organizing by the poor and middle class are thus a viable alternative to campaign finance reforms and a promising means of redressing representational inequality.

In the United States, the legal regime that has most successfully facilitated lower- and middle-class political organizing has been labor law, and the labor union has been a critical vehicle for lower- and middle-class political organization.11 At the peak of union strength, more than twenty million Americans—nearly all within the income classes for whom representational inequality is now a problem—exercised collective political voice through the union form.12 Unions have successfully mobilized their memberships to vote, and, by aggregating millions of small-dollar donations from these members, have built effective lobbying operations, led extensive independent electoral efforts, and positioned themselves as leading campaign contributors.13 At times and on certain issues, unions have been politically liberal; at other times and on other issues, they have taken conservative—even reactionary—positions.14 But, when they were active and strong, unions helped ensure that the government was responsive to the actual preferences of the poor and middle class.15

Today, however, labor unions face a major obstacle to their ability to organize workers politically. The obstacle derives from the fact that unions bundle political organization with a specific and highly contested form of economic organization. Under our labor law regime, that is, unionization requires workers to organize for the purpose of collective bargaining with their employers in order to organize for political action.

From the perspective of political organization, this is a problem, and for several reasons. First, in recent years, managerial opposition to collective bargaining has become widespread and highly effective, and this opposition has made traditional union organizing difficult and increasingly rare.16 Second, changes in the structure of markets and the way work is organized have made collective bargaining hard to sustain in the contemporary economy.17 And, third, at most, only about half of all workers now want to engage in collective bargaining, meaning that unions are not a viable political vehicle for approximately half of the labor force.18

These obstacles that collective bargaining poses to the viability of unions have contributed to a sharp decline in unionization rates. From a peak of thirty-five percent in the mid-1950s,19 unions now represent less than seven percent of private sector workers.20 And this decline in unionization rates has, in turn, contributed significantly to the declining responsiveness of American politics to the poor and middle class.21

For decades, scholars and policymakers have been proposing ways to reform labor law in order to better facilitate unionization.22 If our goal, however, is not to increase the prevalence of collective bargaining but, instead, to facilitate political organizing among politically underrepresented groups, then there is a new possibility for reform. Namely, we could unbundle the collective bargaining and political functions of unions and allow employees to organize politically through the union form without also organizing economically for collective bargaining.

In fact, there is nothing in the nature of unionization that requires the bundling of economic and political functions. Bundling is instead an artifact of history and, more to the point, of law. Workers who sought to improve workplace conditions through collective bargaining often turned to politics to achieve similar goals, and they found that their unions were well suited to act as collective political agents. Contemporary labor law reflects this historical practice of bundling and perpetuates it.23 But, while bundling has made sense in certain contexts, nothing in the history of the union movement suggests that collective bargaining and political action must go together. Moreover, the legal regime that has required a bundling of political and economic functions could just as well allow employees to organize unions for political purposes but not collective bargaining ones.

An unbundled labor law would allow workers engaged in new organizing efforts to form either a traditional union or what this Essay will name a “political union.”24 Political unions would be barred by statute from engaging in collective bargaining, but they would be able to serve as a vehicle for collective political voice for workers who decided to join the union. Unlike traditional unions, political unions would—for reasons this Essay will discuss—represent only workers who affirmatively desired to join and support the union: mandatory membership or mandatory dues payment arrangements of any kind would be out of place in this context.25

As this Essay will explain, the statutory work of unbundling would not be terribly complex. But an unbundled labor law would nonetheless have a critical role to play in facilitating the organization of political unions. In brief, traditional labor law has done four key things to enable workers to use the employment relationship as a locus for organizational activity and thereby to overcome what would otherwise be potentially insurmountable collective action problems.26 First, labor law allows workers to use the workplace as a geographic site for organizational activity, thereby significantly decreasing the coordination costs of organizing. Second, labor law allows unions to harness the employer’s administrative capacity—in particular, its payroll function—to fund union operations.27 Third, labor law allows unions to use the employer’s informational resources—in particular, data about employees—for organizational purposes, thus dramatically reducing the information costs of organizing. And, fourth, labor law prohibits employers from retaliating against employees engaged in organizational activity, thereby preventing the employer’s “rational predatory action” from impeding organizing efforts.28 An unbundled labor law that offered these same four legal advantages to political unions could enable workers to overcome the hurdles to collective political action and take advantage of the union form as a vehicle for collective political voice.

However achieved, unbundling would expand employees’ choice set in a critical way: it would allow employees to organize politically through the union form even if they oppose collective bargaining, and it would thereby expand the range of people for whom unionization is a viable form of political organization. Moreover, because workers would choose the political projects that each of these new unions would pursue—and because those projects would not have to be tied to any traditional union agenda—political unions might better capture the range of political preferences and views of their members than do traditional unions.29

Significantly, political unions would also likely generate less managerial opposition than collective-bargaining unions do, and for several reasons. One, some political unions would devote themselves to policies—on social issues, for example—that firm management would consider non-threatening, or irrelevant, or might even favor. Two, political unions could exercise influence only in the political arena and not at the bargaining table, and thus their power vis-à-vis the individual firm would be more diffuse than the power of traditional unions. Operating at the level of the polity would also mean that political unions would generally be less able to place their firms at a competitive disadvantage vis-à-visnon-union firms. And, because of the greater range of interests represented in the political arena than at the collective-bargaining table, politics-only unions would also be less likely to secure any anti-competitive demands that they might make. Last, and of equal importance, if and when management does oppose workers’ efforts to organize political unions, unbundling would change the social resonance of that opposition.

Indeed, as the Essay will show, there is some preliminary evidence that organizing for politics but not collective bargaining is feasible. There are, to be sure, no extant models of the kind of political unions this Essay proposes. But, in emerging sectors of the labor market where labor law does not apply a bundling requirement, unions have succeeded in organizing workers exclusively for political purposes.

More broadly, identifying the ways in which labor law can facilitate political organizing among workers points us towards a more comprehensive set of reforms designed to enable organizing by politically underrepresented groups.Such an approach to political reform, moreover, has a significant advantage over traditional modes of regulation. No matter how creative the design, campaign finance law does nothing to alter the underlying conditions that produce political inequality.30 To the contrary, traditional modes of political reform attempt to regulate the processes through which the power of wealth is exercised, but they leave in place the distribution of wealth that creates the problems for political equality in the first place. The result is the undoing of the reforms through repeated circumventions that have been aptly analogized to a hydraulic process.31 Legal interventions designed to facilitate organizing are fundamentally different because political organization, like wealth, is itself a source of political power.32 Thus, like wealth, the power that flows from organization can be exercised across processes of political participation—in elections, lobbying, media, and the rest. For this reason, reforms designed to facilitate political organizing are more likely to avoid the problems of circumvention that have undermined traditional modes of regulation.

Three brief words on the premises of the argument. First, the Essay assumes that, in light of the gross disparities in political influence currently enjoyed by different income groups, legal reforms designed to increase the political influence of lower- and middle-income groups are justified by a commitment to political equality. An argument of this sort can be met with the objection that, to succeed, it must articulate the optimal level of political influence that the groups in question ought to enjoy, along with a way of measuring when and whether that optimal level is achieved. Otherwise, the objection goes, the argument cannot allow us to evaluate whether the proposals go too far, or not far enough, in redistributing political influence.33 But, as Rick Pildes has written in a related context, “[i]n theory and in doctrine, we can often identify what is troublingly unfair, unequal, or wrong without a precise standard of what is optimally fair, equal, or right.”34 Thus, rather than attempting to articulate such an optimal distribution of political power or a way to measure it, the Essay instead proceeds on the more tractable assumption that, in Rawls’s terms, all citizens in a democracy—irrespective of income level—ought to have an “approximately equal” chance of influencing political decisions.35

Second, representational inequality among income groups is not the only type of representational inequality: policy may well be more responsive to different racial groups, or geographic groups, or age groups. Without adjudicating the priority of these different aspects of inequality, this Essay limits its attention to economic forms of political inequality and thus its proposals for intervention are similarly focused.

Third, this Essay suggests unbundling the union as a way to mitigate representational inequality: it is an Essay about political organizing. But the argument here does not imply that workers should not also have a collective voice in the workplace. To the contrary, such a voice is critical for a host of reasons, some of which I have pointed to elsewhere.36 The proposal in this Essay is meant as a complement to, rather than a substitute for, efforts to rethink collective economic representation at work.37

The Essay proceeds as follows. Part I briefly reviews the recent political science on representational inequality and describes the findings that point to substantial skew in policy responsiveness across income levels. The Part then discusses the failure of the traditional response to representational inequality—campaign finance regulation—and proposes an alternative: legal interventions designed to enhance the political-organizational capacity of low- and middle-income groups. Part II begins by identifying labor unions as an obvious source of organizational voice for such groups. It then identifies the advantages that the workplace offers to unions as a locus of organizational activity, and shows the ways in which labor law allows unions to harness these advantages. But the Part concludes by showing how the bundling of unions’ collective bargaining and political functions has impeded unions’ viability as a political-organizational vehicle. Part III argues that unbundling unions’ political and economic functions would increase their capacity to serve as a platform for political organization. It argues that political action would not replicate for political unions the impediments that collective bargaining has posed for traditional unions, and it shows why political organizing can succeed even when it is not grounded in the economic practice of collective bargaining. Part IV describes the statutory reforms necessary to unbundle unions’ political and collective bargaining functions, and Part V suggests other contexts in which the law might facilitate political organizing by lower- and middle-income citizens. The Essay then concludes.

I. representational inequality

A. Income and Responsiveness

Political equality is a core feature of democratic governance. While the definition and appropriate scope of such equality is contested, there is general agreement that citizens in a democracy ought to have an approximately equal opportunity to influence the political process.38 But theorists writing from a wide range of perspectives have long argued that economic inequalities threaten to subvert this democratic goal. Rawls, for example, was concerned that “those with greater property and wealth” would capture “the electoral process to their advantage.”39 Schattschneider believed that economic inequality builds an “upper-class bias” into democratic politics,40 and Walzer expressed concern that “the dominance of money in the sphere of politics” would render much of the populace politically powerless.41 In more contemporary work, Hall and Deardorff argue that the affluent can “distort[]” policymaking in their favor through lobbying.42

Contemporary empirical research in political science confirms the theorists’ concerns. Most prominently, in a book published last year, Martin Gilens reports the findings of an analysis of two decades of U.S. public policy.43 Gilens finds that as the gap between the preferences of the poor (those in the bottom income decile) and the preferences of the affluent (those in the top income decile) increases—that is, as the rich and poor disagree more—there is a major decline in the association between the poor’s preferences and policy outcomes.44 Put plainly, “when preferences between the well-off and the poor diverge, government policy bears absolutely no relationship to the degree of support or opposition among the poor.”45

These results might be consistent with democratic principles if the views of poor respondents are simply minority views. But Gilens finds that median-income earners fare no better than the poor when they part ways with the policy positions of the affluent. Where the preferences of these two income groups diverge by ten percentage points or more, policy responsiveness for the ninetieth income percentile holds steady and strong, but is statistically equivalent to zero for those at the fiftieth income percentile.46 Most importantly, Gilens finds that even when the poor and middle classes agree with one another and together disagree with the affluent, it is still the views of the affluent that get translated into policy.47 Gilens thus concludes that “for Americans below the top of the income distribution, any association between preferences and policy outcomes is likely to reflect the extent to which their preferences coincide with those of the affluent.”48

Although Gilens’s findings are striking, his research confirms earlier work by Larry Bartels and others.49 Bartels’s influential study found that U.S. senators are far more responsive to the policy preferences of their affluent constituents than they are to those of their lower- and middle-class constituents.50 As Bartels put it, senators “were vastly more responsive to affluent constituents than to constituents of modest means,”51 and “the views of constituents in the bottom third of the income distribution received no weight at all in the voting decisions of their senators.”52

The fact that policy is more responsive to the affluent than to other income groups does not tell us why this is so, and it is possible that responsiveness follows not wealth, but voting rates or other measures of political activity that the affluent also happen to exhibit. Here too, though, Gilens and Bartels reach similar conclusions. Both investigate the possibility that policy is particularly responsive to the affluent because the affluent vote more or are otherwise more active in politics. Both reject this possibility. Bartels, for example, shows that while there are significant voting gaps between high- and low-income groups, the differences are not large enough to account for the skew in policy responsiveness.53 In fact, Bartels concludes that accounting for differences in voting rates, political knowledge, and contact with public officials “reduces only modestly the substantial income-based disparities in responsiveness.”54

It is important to clarify two things that Gilens’s and Bartels’s work does not imply. First, their conclusions do not mean that the affluent always do—or always will—get their way. Government is highly responsive to the preferences of the affluent and more responsive to their preferences than to the preferences of the poor and middle class, but the government is not perfectly responsive to what the wealthy want. Indeed, some government policies are enacted despite the fact that majorities of all income groups—including the wealthy—oppose them, and many policies are not enacted despite the fact that broad cross-sections of the public—including the wealthy—support them.55 Second, and more important, the Gilens and Bartels analyses do not mean that low- and middle-income citizens never do—or never will—get their way. To the contrary, there are many issues on which the preferences of the poor and middle class find expression in enacted policy. The problem is that the poor and middle classes’ preferences tend to get enacted into policy only when their preferences align with the preferences of the wealthy.56

B. Addressing the Problem: From Finance to Organization

The traditional, and perhaps most obvious, response to representational inequality is campaign finance reform.If the wealthy have disproportionate influence over the political process because they are able to deploy their wealth to political ends, regulations that restrict the political uses of money make sense. And, for over a century, reformers in Congress have attempted to address the political influence of wealth by enacting restrictions on campaign spending.57 The efforts began at the opening of the twentieth century with the Tillman Act, a law that responded to the popular movement for “elections free from the power of money.”58 From 1907 on, legislative efforts at campaign finance regulation have followed a steady path forward, expanding to sweep in more spenders, more elections, and more forms of spending.59

Despite multiple interventions of increasing complexity and scope, however, campaign finance regulation has, to put it mildly, not succeeded in curbing the role and influence of money in American politics. This failure is well documented in the literature. In The Hydraulics of Campaign Finance Reform,60 for example, Issacharoff and Karlan showed that political actors adjust to campaign finance regulation by reorganizing and redirecting political spending in ways not reached by existing law.61 The election cycles since Issacharoff and Karlan’s writing lend support to their thesis.62

It is not only the ingenuity of political spenders that has stymied campaign finance reform, however. The Supreme Court, too, has limited the range of regulation permissible under the First Amendment. The Court’s restrictive interventions began with Buckley v. Valeo,63 which struck down limits on independent expenditures by individuals,and continue today, most notably with Citizens United v. FEC.64 As Michael Kang explains, Citizens United effectively makes any type of meaningful campaign finance regulation (other than restrictions on direct contributions to candidates and certain disclosure requirements) unconstitutional.65 The decision, in other words, “leaves virtually no constitutional space for new campaign finance regulation.”66

Given their view that campaign finance reform has reached a dead end, many election law scholars now argue that some other approach is needed to deal with the influence of wealth on politics.In Kang’s view, the appropriate move is away from what he calls the ex ante regulation of campaign spending and toward the ex postregulation of the legislative process, especially through lobbying reform.67 Although Kang’s account is helpful in moving the discussion beyond spending restrictions, lobbying regulation presents some of the same difficulties as campaign finance regulation. First, as Kang acknowledges, “lobbying reform faces its own constitutional challenges under the First Amendment.”68 Second, and perhaps more importantly, it is not clear why the same kind of hydraulics that plagued campaign finance restrictions would not reproduce themselves in the context of lobbying restrictions. The political actors who found ways around spending restrictions would likely find ways of avoiding lobbying regulation as well.

Heather Gerken is also attracted to the idea of lobbying reform,69 but rather than suggesting new restrictions on lobbying activity—an approach Gerken categorizes as leveling down—she suggests that we find ways to level up by expandingaccess to lobbyists. She thus characterizes her proposal as an analog to the public financing of elections: supplement privately funded lobbyists with publicly financed ones for those who otherwise would be locked out of the lobbying game.70

Conceptually, the leveling-up strategy for legal intervention into the political process embodies what Bruce Cain calls the “more voice, not less” approach to reform.71 As Cain puts it, “another way to neutralize political advantage aside from capping and prohibiting is to support countervailing voices.”72 Cain locates this approach in the Madisonian idea of fighting faction with faction, and in the pluralist idea of “expand[ing] the number of players in a political area to offset the advantages of the dominant players.”73 Like Gerken, Cain sees public financing of both elections and lobbying as the primary way to instantiate this approach to reform.

But the democratic tradition in which Cain grounds his idea supports other approaches as well. In particular, the democratic norm of equal representation can be advanced by legal interventions designed to address inequalities in organizational capacity—interventions aimed at facilitating political organizing by underrepresented groups.Rather than intervening later in the political process by moving from elections to lobbying, that is, we can intervene earlierin the process by facilitating the organizational development of underrepresented groups. Writing in a related context, Joshua Cohen and Joel Rogers made the point this way:

[I]nequalities in material advantage . . . translate directly to inequalities in political power. Groups can help remedy these inequalities by permitting individuals with low per capita resources to pool those resources through organization. In making the benefits of organization available to those whose influence on policy is negligible without it, groups help satisfy the norm of political equality.74

Moreover, legal interventions designed to facilitate political organizing have an important advantage over both the type of leveling-down restrictions that Kang advocates and the leveling-up proposals of Cain and Gerken. Neither campaign finance restrictions (or lobbying restrictions) nor public financing of elections (or public financing of lobbying) address the underlying inequalities in political power that flow from wealth. Rather, such regulations change something about one or another of the multiple processes through which such political power is expressed. The regulations might restrict campaign spending, or they might equalize lobbying spending, but they do not alter the background conditions that produce political inequalities in the first place. Because the political power that comes from wealth is portable across political processes—because “[t]he sheer versatility of material power is what makes it so significant politically”75—a circumvention problem plagues not only the leveling-down approach but these leveling-up approaches as well. If we succeed in equalizing election spending, either by leveling up or down, the political power that comes from wealth will be exercised through lobbying; if we succeed in equalizing lobbying spending, by leveling up or down, that power will be exercised through some other means.76

While elections and lobbying are processes through which political power can be expressed, organization, like wealth, is itself a source of political power. This distinction matters enormously. As we have seen, political reforms that restructure processes of participation but that leave background power asymmetries untouched can be undone by circumvention. Political reform aimed at organizational capacity is less apt to suffer this problem because it enables groups to build political power that, like power derived from wealth, is portable across processes of participation. A well-organized political group can mobilize voters and influence elections; it can lobby and influence legislation; it can buy media time and influence public opinion; and so on. Organization can therefore countervail wealth’s power irrespective of whether that power is expressed through elections, lobbying, media, or any other avenue.

An organizational approach to redressing representational skew is not only well grounded theoretically, but also recommended by the empirical realities of current democratic practice in the United States. Again, Gilens’s study is illuminating. As discussed, Gilens finds skew in the responsiveness of policy to different income groups. But Gilens also finds an exception to this general rule when the balance of organized interest group power aligns with the policy preferences of the poor and middle class. Thus, unlike the other policy domains he studied, when it came to certain social welfare policies—including Social Security, Medicare, school vouchers, and public works spending—Gilens found no evidence that the policy preferences of the poor and middle-class suffered when those preferences diverged from those of the affluent.77 The reason, Gilens suggests, is that on these questions “poor and middle-income Americans have powerful allies that tend to share their preferences.”78 Those allies are organized interest groups.79

Gilens thus observes empirically what the preceding discussion predicts theoretically: organization can compensate for the effects of wealth because it operates as an independent source of political power.80 When the balance of organized interest group power aligns with the preferences of the wealthy, as it generally does, the affluent’s preferences gain even more policy traction relative to the preferences of the poor and middle class. But where organized interests reflect the preferences of the poor and middle class, organization reduces representational inequality.

II. unions and representational inequality

A. Unions and Politics

Among the interest groups operating in the United States today, the “strongest positive associations between [the] groups’ [policy] positions and the preferences of the less well-off” are found in labor unions.81 Across a range of issues, unions’ policy positions are highly correlated with those expressed by individuals in the bottom nine income deciles.82 This should come as no surprise: the income profile of union membership resembles the population that Gilens’s work reveals to be lacking in political influence. In fact, more than eighty-five percent of union members fall into the “non-affluent” income categories against which policy responsiveness is currently skewed.83

Given that organization constitutes a non-wealth-dependent source of political influence, and given unions’ ability to organize and advocate for the policy preferences of the poor and middle class, unions are a clear source of political influence for these sectors of the polity. In fact, in the United States, unions have managed to organize lower- and middle-class Americans for political action in numbers unmatched by any other non-party actor. At their peak, unions represented more than one-third of wage-earners in the country,84 and counted twenty-one million workers as members.85

Historically, unions have mobilized their memberships for various forms of political action. They have, to start, done well registering and increasing voter turnout among lower- and middle-class workers. One early review of the literature concluded that union members were sixteen percent more likely to vote than unorganized workers with similar occupations, education, income, and status,86 and subsequent studies agree that union members turn out for elections at disproportionately high rates.87 Unions have also contributed substantial sums to candidates and parties,88 built effective lobbying operations,89 and engaged in extensive independent political advertising campaigns.90 Unions, moreover, have funded this political activity with small-dollar voluntary contributions from members. To take just one example, political contributions to the Service Employees International Union (SEIU)—perhaps the most politically active union in the nation and one that spent approximately $23 million on the last election cycle—average seven dollars per month.91

Although it is difficult to measure with precision, it is also clear that union organization has had a significant impact on American policymaking.92 Numerous historical case studies support the point. In the 1940s, the Congress of Industrial Organizations (CIO) had such a significant role in policymaking that President Roosevelt is rumored to have instructed his advisors to “[c]lear it with Sidney [Hillman]”—the CIO’s political director—before moving on major political or policy initiatives.93In the 1960s, according to historian Nelson Lichtenstein, labor played a central role in shaping the Civil Rights Act of 1964,94 in particular in ensuring that Title VII was included in the final bill.95 Indeed, the Act’s passage depended in large measure upon labor’s power to turn out legislative votes.96 As Representative Richard Bolling, one of the Act’s leading supporters, put it, “We never would have passed the Civil Rights Act without labor. They had the muscle; the other civil rights groups did not.”97 The enactment of Medicare the following year obeyed a similar logic: labor was “the most powerful single source of pressure” among supporters in the legislative struggle.98 And unions have been the primary force behind changes in federal minimum wage policy since the enactment of the Fair Labor Standards Act in 1938.99

Finally, quantitative data corroborate the case-study evidence of union influence. In one of the foundational studies in the field, Richard B. Freeman and James L. Medoff showed that federal legislators representing well-organized states in the 1970s were more likely to “vote union”—that is, vote in harmony with the expressed preferences of the AFL-CIO—than their peers from less well-organized jurisdictions.100 Subsequent scholarship confirms this correlation between union density and legislative voting in favor of the unions’ policy positions.101 Further, in the states, relative union density accounts for significant variations in social welfare, education, and tax policy,102 and greater union density correlates with more rigorous state statutory protections for workers.103

B. Organizing Through Work

There are numerous reasons why unions were successful in organizing low- and middle-income groups but one is particularly relevant to the analysis here: labor law enabled unions to take advantage of the workplace (in particular) and the employment relationship (more generally) as loci of organizational activity. As this Section will discuss, workplace organizing offers a number of advantages that allowed unions to overcome what might otherwise have been insurmountable hurdles to collective action. And, as this Section will also discuss, labor law allowed unions to harness these advantages through several relatively modest, but critical, interventions.

First, interpersonal relationships and social networks are important facilitators of organizational and political activity.104 Work often is a rich source of such relationships and networks,105 and can serve as an especially viable platform for organizing.106 In a similar vein, organizing depends on the development of a collective identity among participants.107 Again, work can—and often does—provide the basis for such an identity. As Robert MacKenzie puts it, “[t]he shared experience of work is a strong influence over the creation of social collectivity. Work as a collective experience . . . may act as a basis for group identity.”108

More tangibly, though no less importantly, the workplace is an important geographic sitefor organizing, a centralized location where employees gather as a group and where they can be reached as a group. Thus, the significant costs of identifying and contacting employees, who are otherwise dispersed across potentially large geographic areas, can be avoided when the workplace is available for organizational purposes.109 Here, labor law has made an important contribution: the National Labor Relations Act (NLRA) allows employees to use the workplace as a centralized location for organizing by granting employees the right to speak with one another about unionization in non-work areas of the workplace and during non-work time.110 These same rules also enable employees to speak to one another about, and encourage each other to participate in, a certain range of political activity in the workplace.111 For many years after the enactment of the NLRA, moreover, labor law also allowed non-employee union organizers to speak with employees at their worksites.112

Next, work has proved a fruitful locus for organizing because the employment relationship enabled unions to solve one of the more dogged problems in organizational development: the need for an administrable and sustainable financing mechanism. In short, the employer’s payroll system provides a channel through which dues payments can be made by employees to the union automatically and on a recurring basis. Here, again, a modest legal intervention has been important: the NLRA makes payroll deductions for union dues a mandatory subject of bargaining113 and treats an employer’s refusal to agree to payroll deduction as evidence of a failure to bargain in good faith.114 In some states, dues payments can be required as a condition of employment; in other states—so-called right-to-work states—dues can only be collected from employees who wish to join the union. But in either setting, the statute permits unions to harness the employer’s payroll system as a method of dues deduction.115

Employers also constitute a centralized source of information about employees, information that is important for organizing. Specifically, because it is often necessary for organizers to speak with employees outside of the workplace—particularly if there is no centralized geographic workplace—unionization requires that organizers be able to identify who the workforce is.116 Here, the NLRA grants union organizers the right to access the employer’s list of employee names, addresses, and—potentially—phone numbers and email addresses.117 Particularly in larger bargaining units, this grant of access to the employer’s informational resources reduces what could otherwise be prohibitive information costs.118

Finally, labor law compounds the advantages of the workplace as a site for organizing by making employer retaliation for such activity illegal. Section 7 of the NLRA gives employees the affirmative right to “form, join, or assist labor organizations,” and Section 8(a) makes it an unfair labor practice for an employer “by discrimination in regard to hire or tenure of employment or any term or condition of employment to encourage or discourage membership in any labor organization.”119 Accordingly, if an employee is penalized for union activity, the employer faces unfair labor practice liability and the employee is entitled to an appropriate remedy (including reinstatement and back pay). As Richard Posner explains: “The efforts of an employee to induce his fellows to [unionize] would often, in the absence of legal protection . . . be set at naught by the employer’s firing him.”120 By making such actions illegal, the law denies the employer “the natural advantage that he would have, as one facing many, in fending off organizing activities.”121

With respect to both workplace access and anti-retaliation protections, the robustness of these legal rights has declined rather dramatically in recent years. For example, the Court has curtailed access rights for nonemployee union organizers,122 and the National Labor Relations Board’s ability to remedy anti-union retaliation has declined as a result of both Court intervention and inadequate enforcement resources.123 The prevailing view is that the curtailment of these legal protections has contributed to the decline in unions’ organizational success.124 The point here, however, is simply that the right to use the workplace and the employment relationship for organizational activity, along with protections against retaliation, help employees overcome impediments to collective action and serve as important facilitators of organizational activity. The stronger those rights, and the more robust their enforcement, the more effective they are at facilitating such activity.

A word is also in order here about a role that labor law does not play in facilitating organizational success. Among the traditional impediments to collective action, the problem of free riding is central.125 Labor law does provide unions with a mechanism for dealing with part of the free-rider problem. Under the NLRA, collective bargaining agreements can require that all employees who are represented by the union pay dues to cover the union’s collective bargaining and contract administration expenses.126 But labor law’s resolution of the free-rider problem is limited in a way that is relevant here: the law does not entitle unions to overcome free riding with respect to political activity.127 This, of course, is the free-rider threat that political unions would face, and the success that workers historically have had in organizing for political action suggests that they have been able to overcome this free-rider threat even though labor law has not given them a mechanism for doing so.

C. Collective Bargaining and Union Decline

In part due to unions’ ability to take advantage of the workplace as a locus of organizational activity, unions historically have been an effective political voice for the poor and middle class. The problem, from the perspective of representational equality, is that unionization rates have been falling consistently across the last several decades and have now fallen to levels not seen since before the Wagner Act was passed in 1935.128 Scholars have devoted significant attention recently to the consequences for American politics of this decline in unionization rates.129 Nearly all—including Gilens, Bartels, Hacker and Pierson, and Schlozman et al.—point to the decline in union strength as a major explanation for the growth in representational inequality.130 Hacker and Pierson summarize the consensus this way:

No group better captures the mid-century influence of voluntary organizations representing middle- and working-class Americans than organized labor. . . . [But] as unions shifted from confident involvement in politics to embattled defense of their ever-smaller pocket of the workforce, they also ceased to be able, or always willing, to play the role as champions of the broad middle class they had carved out in their heyday.131

There is a long-running debate over the specific causes of this decline in unionization.132 The leading theories point to increasing competitiveness of product markets,133 shifts in production methods and systems of work organization,134 and managerial opposition to unionization.135 Fortunately, we need not attempt to resolve this debate here because all of the leading theories place collective bargaining at the center of the story of union decline. It might be that collective bargaining raises wages and benefits above competitive levels and thereby puts unionized employers at a competitive disadvantage vis-à-vis non-union firms in newly competitive markets.136 It might be that collective bargaining imposes a set of inflexible work rules that stand in the way of much-needed flexibility.137 Or it might be that, given both of the above, collective bargaining inspires deep managerial hostility to the union project.138 But, whatever the mechanism, the consensus view is that unions are in decline—largely, if not entirely—because of their collective bargaining function.139

To see this, take the concern about increasingly competitive markets. The argument is a plausible one and asserts that unions were able to thrive at a moment in U.S. history when markets were nearly oligopolistic.140 As markets globalized and became increasingly competitive, however, collective bargaining—which drove up labor costs at unionized firms—undermined the ability of union employers to compete with non-union ones. The result was significant loss of market share for unionized companies and dramatic declines in unionization rates.141

Or take the concern about the changing nature of work. As Katherine Stone documents, the era of mass industrial production was one in which “[e]mployers sought uniformity in products and processes in order to achieve economies of scale.”142 These production processes lent themselves to narrow job definitions and long-term employment relationships.143 In this setting, according to Stone’s account, collective bargaining agreements that enforced rigid job rules and strict seniority rights made sense. But because our economy is no longer defined by mass production industries organized in this fashion, Stone argues, unions’ insistence that collective bargaining agreements dictate work rules, seniority protections, and the like is out of place.144

Finally, take managerial opposition to unionization. Although there is some debate about the specifics, the prevalence, ferocity, and effectiveness of managerial opposition to unionization is well established. In one prominent study, for example, employers engaged in anti-union efforts in 96% of the union organizing campaigns they faced;145 in another study, the opposition rate was 98.4%.146 In about half of all union campaigns, moreover, employers threaten to close the business should employees choose to unionize.147 Employers also fire between 5% and 20% of active union supporters.148 Not surprisingly, these tactics work: when management threatens to close the business or fires union supporters, union win rates decline significantly.149 Even where there is a successful unionization campaign, moreover, management frequently engages in efforts to avoid concluding a collective bargaining agreement. These tactics, also often successful, have further contributed to the decline in unionization rates as initial union organizing victories fail to become institutionalized and lead instead to the decertification of the union as the employees’ bargaining representative.150

Managerial opposition to unionization has everything to do with collective bargaining. In fact, management opposes unions for two primary reasons: one, to secure competitive labor costs, and, two, to maintain control over the way work is organized and carried out.151 Collective bargaining agreements can impede both of these managerial objectives. First, collective bargaining often requires unionized firms to pay more in labor costs than their non-union competitors.152 Second, as Stone’s work emphasizes, collective bargaining also often restricts managerial control over the workplace. Thus, for example, collective bargaining agreements generally include just-cause dismissal clauses that restrict management’s ability to discharge employees and thereby to control how those employees behave.153 The work rules contained in most collective bargaining agreements have a similar effect: rather than allowing management to determine how work at the firm is performed, these clauses specify which workers are to perform which tasks and set out with some precision how those tasks are to be performed. Seniority agreements, job-bidding systems, bumping rights, and analogous collective bargaining clauses likewise diminish managerial control over how the workplace is organized and thereby fuel managerial opposition.154

Collective bargaining has significant merits. Over the last half century, the practice has been a key contributor to economicequality in the United States,155 and collective bargaining helps correct market failures that plague individual employment contracting.156 But, as the above discussion makes clear, collective bargaining is also a central factor in each of the primary explanations for the decline of unions.

III. the unbundled union: liberating politics from collective bargaining

The last Part argued that collective bargaining has, for a number of reasons, made it difficult for traditional unions to thrive under contemporary conditions. But because unions’ political and collective bargaining functions are bundled, all of the collective-bargaining-related reasons for unions’ decline—that is to say, essentially all of the reasons for union decline—are also impediments to workers’ ability to use unions as a vehicle for political organizing. Bundling, in short, holds unions’ political-organizational capabilities captive to the fortunes of collective bargaining.

Because unions’ political and collective bargaining functions are bundled, a resurgence of traditional unions would increase the political voice of lower- and middle-income groups.157 But rehabilitating unions as a vehicle for political organizing does not require a resurgence of traditional unions.158 To the contrary, political organizing can be advanced by unbundling the political and collective bargaining functions of the union.

This is true for a reason that will now be obvious. In an unbundled regime, the fact that collective bargaining can be incompatible with contemporary forms of work organization, that it can create problems for firms operating in modern markets, and that it inspires fierce managerial opposition would be irrelevant to workers’ political efforts because those efforts could now proceed independently of collective bargaining. Unbundling, that is to say, would immediately insulate political organizing efforts from the vulnerabilities of collective bargaining. Unbundling would also expand the range of employees for whom political organizing through the union form is a viable option. As we have seen, employees who wish to organize politically through a union can do so only if they also choose to organize for collective bargaining. But a substantial portion of the U.S. labor force does not desire to bargain collectively with their employers.159An unbundled regime would thus allow the half of the labor force that does not want collective bargaining to take advantage of the union as a political vehicle.

The third reason to predict that unbundling will facilitate political organizing is the focus of this Part: namely, political unions would likely generate less managerial opposition than traditional collective bargaining unions generate. Itis important to clarify at the outset, however, what this Part does and does not argue. The argument here is that unbundling the union would likely decrease—though not eliminate—managerial opposition to employee organizing efforts. There are groups other than firm management that oppose unions’ political efforts. The Chamber of Commerce and the National Right to Work Committee are prominent examples.160 More recently, individuals like the Koch brothers have played a similar role, sponsoring a 2012 ballot initiative in California that would have prohibited even voluntary dues deductions for political purposes.161Many elected officials also oppose unions’ political efforts. Scott Walker, the Republican Governor of Wisconsin, is a leading contemporary example.162

With respect to these groups and individuals, two points bear mention. First, whether groups that oppose traditional unions would also oppose political unions depends on what political unions end up doing. To the extent that political unions replicate the political agendas of traditional unions, groups like the Chamber of Commerce undoubtedly would continue their opposition. But in a world where politics is unbundled from collective bargaining, it is uncertain whether political unions would advance the same agendas as traditional unions. As this Essay has noted, some political unions might choose not to advance economic goals at all; in such cases, it is doubtful that the Chamber of Commerce and groups like it would focus much energy on opposing them.

The second point is that even if opposition by these groups remained undiminished, it is still relevant that managerial opposition is likely to decline. This is the case most generally because it would imply an overall decline in opposition. More particularly, this is true because non-managerial actors lack the set of resources that makes managerial opposition to union organizing so effective. Namely, these groups cannot fire employees who try to organize unions, and they cannot threaten to close firms in response to organizing drives. Only management can do that.

In addition to offering reasons to predict that unbundling will increase the prospects for political organizing, this Part also offers some very preliminary evidence that workers would organize political unions. The evidence consists of union organizing campaigns that have been restricted to political action. The Part also addresses more theoretically the question of whether union political organizing can succeed when it is not connected to collective bargaining.

A. Improving Prospects for Organizing: Diminishing and Altering Managerial Opposition

Management, to be sure, would sometimes oppose the formation of political unions. But political unions would likely generate less pervasive and less severe opposition than conventional unions do. The most basic reason for this prediction is that, while traditional unions have a legal obligation to bargain over wages, hours, and other terms and conditions of employment,163 political unions might not choose to pursue public policies that impact the firm at all. For example, while a traditional union must engage the firm over economics, a political union would be entitled to focus entirely on social issues, international affairs, public education, and the like. And while individual managers might have positions on these issues, the union’s advocacy of them would likely not threaten the economic interests of the firm. It is also the case that certain components of collective bargaining agreements that generate managerial opposition would be difficult, if not impossible, to replicate through political action: key examples are the work rules and job classifications contained in many collective bargaining agreements, which, as Section II.C explained, are understood to impede flexibility in work design.164

But even if political unions took on policy issues that impact the firm directly, they would still likely engender less managerial opposition. This is the case because political unions would have less power vis-à-visthe individual firm than a collective bargaining union does. Collective bargaining unions exercise collective power directly against the firm: they translate their ability to threaten production—or to otherwise intervene in the firm’s operations—into bargaining concessions that the firm itself must grant. By contrast, a political union would exercise its power through the political processes of some government. And, even if all the employees in a given firm join a political union, they would constitute only a small proportion of the voters in the relevant polity. The result is a diminution in the relative ability of a particular union to impact the individual firm where it is organized, a diminution that should lessen the intensity of managerial opposition.

Of course, if political union organizing is successful, and if unions formed at different firms were to affiliate, these unions could secure significant political power. But managerial opposition to political unionism on these grounds would be plagued by its own collective action problem. That is, even if it would be in the collective long-term interests of multiple firm managements to fight the development of individual political unions (because those unions might eventually unite and exercise significant political power), the incentives for individual managements to free-ride on the oppositional efforts of other firms would be significant.This collective action problem would thus reduce the likelihood that any particular firm’s management would fight the organization of any particular political union.

Political unions are also likely to generate less managerial opposition because they are less likely to impose competitive disadvantages on their firms. This is the case, in part, because collective bargaining generally takes place at the firm level and political action by definition takes place at the level of a polity that will encompass multiple firms and, at times, entire markets.165 To take a basic example: if a union, through collective bargaining, secures paid sick leave for the employees of a grocery store chain in a given state, that chain will face higher labor costs than, and be at a competitive disadvantage relative to, all the other chains in the state (and relative to any chains that move into the state to take advantage of the unionized firm’s higher labor costs). If, on the other hand, the union helps secure legislation requiring all employers in the state to provide paid sick leave, then the sick leave requirement produces no competitive disadvantage for any of the grocery stores in the state. In this scenario, employers are likely to vigorously oppose the formation of the collective bargaining union. They may also oppose the formation of a political union—a legislated sick-leave requirement imposes costs, even if those costs are borne by all the firms operating in the state. But because the political outcome takes sick-leave costs out of competition, managerial opposition is likely to be less intense.

Political action certainly can place firms at a competitive disadvantage relative to others. Indeed, any legislation that imposes costs on some firms in a market but not others—say, in-state manufacturers but not out-of-state manufacturers—will have this effect. Nonetheless, even where politics has the potential to impose competitive disadvantages on firms operating within the domestic political boundaries, there are many contexts in which politics is less likely to lead to these outcomes than is collective bargaining. This is so because a broader set of interests are represented in the political process than in the collective bargaining one. At the bargaining table, there are only two parties: the union and the employer. The union is accountable to its current membership, a dynamic that can lead the union to make bargaining demands that improve current conditions at the expense of long-term competitiveness.166 Notgiven a place at the collective-bargaining table are the many constituencies that may be negatively impacted by anti-competitive demands made by the union: future employees who lose job opportunities, other firms whose fortunes are linked to the unionized employer’s, community groups whose interests are connected to the success of the employer, and, indeed, the broader polity interested in the tax base to which the employer contributes.

In the political process, by contrast, there are multiple parties with multiple constituencies. When union political organizing is successful, unions can at most constitute one voice among these many others. Should a political union seek legislation that has anti-competitive effects, it will therefore be opposed not only by the firm but also by a host of other groups that would face the negative consequences of anti-competitive legislation. Where these interests can outvote the union, proposals that the union might have won at the bargaining table will fail to carry in the political environment.167

In sum, then, political unions may not implicate the firms’ economic interests at all. If they attempted to do so, their ability to negatively impact those interests would be tempered by the diffuseness of their organizational density (relative to the polity as a whole), by their need to build coalitions that would include groups less willing to compromise firm competitiveness, and by the fact that—should adequate coalitions be built—any political successes would likely impact multiple firms in the relevant market, thereby lessening anti-competitive effects.

It is worth addressing here the possibility that some firms would oppose political unions on the ground that they would constitute a first step towards the formation of a traditional union. For example, even though a political union could not itself engage in collective bargaining, it might contribute to the development of a collective identity among workers who previously lacked one and thereby make the arguments for collective bargaining more compelling.168 A political union might also conceivably serve to whet the appetite of workers for collective bargaining, making the possibility of bargaining directly with management over terms and conditions of employment seem more appealing.169 Or, more straightforwardly, the political union might make it easier for workers to deal with the coordination costs of organizing a collective bargaining union.

These arguments have some plausibility, but there are important rejoinders. First, employees form many types of collectives that do not elicit managerial opposition—and certainly not the type of opposition that management exhibits toward collective-bargaining unions. When employees form a book club or a softball team or a prayer group, we might predict that these groups would contribute to the development of a collective identity or make it easier to overcome the coordination costs of union organizing. Yet, management does not routinely fire workers who take a lead role in organizing these types of collectives. Second, while a political union might whet the appetites of employees for collective bargaining, it could also have precisely the opposite effect: it could satiate—or deflect—the workforce’s collective impulse, thereby making the formation of a collective bargaining union less likely.170 Indeed, management itself often establishes forms of collective organization within the workplace—self-managing production teams, employee involvement committees, and the like—in an attempt to “deflect workers’ group choice over workplace governance modes . . . away from . . . full collective bargaining.”171

A recent study by John Godard and Carola Frege sheds light on this issue.172 Godard and Frege conducted a survey of a thousand U.S. workers and investigated, inter alia, whether the presence of certain non-union forms of employee representation in the workplace impacted the propensity of workers to join a traditional union.173 The researchers asked workers whether “there [was] a nonunion, management-established system [in the workplace] where worker representatives meet with management,” and whether the worker was “a member of another type of association to assist with work related matters[, including an] association . . . based on your occupation, race, gender, or some other characteristic you identify with.”174 Although the authors found relatively high levels of both types of non-union representation systems in U.S. workplaces, they found no evidence that the presence of either type of organization impacted workers’ propensity to vote for a union. That is, having a workplace organization other than a union neither increased nor decreased the likelihood that a worker would decide to support unionization.175

It is not possible to predict with certainty how much managerial opposition would be generated by a concern about political unions as a precursor to traditional unionization. Given management’s own willingness to experiment with non-union forms of collective workplace organization, the lack of managerial opposition to other types of employee organizations, and the evidence that non-union forms of workplace representation do not increase workers’ propensity to support unions, however, it seems unlikely that opposition generated by this concern would be severe or pervasive.176

Finally, and of equal importance, when management does oppose political unions—as it will in certain circumstances—its opposition will likely have a different social resonance than does opposition to traditional collective-bargaining unions. Employees’ collective economic activity is sufficiently contested that management’s decision to oppose union organizing does not generate much social contestation.177 Perhaps the best piece of evidence for this is that although managerial opposition to unionization is intense and widespread, and quite often illegal, it produces essentially no public outcry. In 2010, for example, more than 1600 workers were offered reinstatement after being illegally fired for union activity, and more than 17,000 workers received back pay for illegal employer conduct related to union activity.178 Despite the remarkable scope of this anti-union activity, no major newspaper in the country made more than passing reference to any of it. Most of the largest papers did not report on it at all.179

While opposition to union organizing efforts fails to provoke public concern, managerial opposition to employee political activity would likely be viewed with greater skepticism. This is the case, in part, because political organizing is far more culturally mainstream in American public life than union organizing.180 A second factor is at work too. American social and legal culture largely accepts managerial control over employees’ worklife—over the way that work is organized, carried out, and compensated. But there is greater discomfort when management tries to use its economic power to control noneconomic aspects of an employee’s life, including the employee’s political activities.181 During the latest presidential campaign, for example, a number of corporations encouraged their employees to vote a particular way. Some firms, in pressing employees to vote for the Republican nominee, suggested that an Obama reelection would threaten the employees’ jobs. Analogous employer attempts to influence employee voting in union campaigns—including suggestions that voting for the union will threaten jobs—are routine and receive no attention. But the attempt by management to intervene in employees’ political voting was covered extensively, and with strong notes of disapproval, in the media. Among other examples, the New York Times ran a full-length, front-page article on the subject, titled Here’s a Memo from the Boss: Vote this Way.182 The Houston Chronicle published a condemnatory editorial arguing that “the boss has no business in your voting booth.”183And two prominent law professors excoriated the practice, calling it one of the “worst pathologies of this second gilded age.”184

The commitment to this kind of spheres separation is also reflected in the organizing principles of U.S. labor and employment law. Both legal regimes give management extensive control over how work is organized, performed, and compensated. Thus, the default rule in U.S. workplaces is individual employment contracting,185 which, in practice, means that management generally sets the terms of the employment contract and employees choose whether or not to accept employment under those terms.186 Even if employees unionize, thereby obligating management to bargain with the union over terms and conditions of employment, management nonetheless retains the ultimate right to refuse union proposals and implement its desired workplace policies.187

While the law gives management substantial control over an employee’s worklife, it is more resistant to managerial attempts to control employees’ political activities. In fact, as Eugene Volokh has recently shown, legal protection against employer interference with employee political activity has a rich historical tradition in the United States.188 Thus, “the very first American laws banning employment discrimination by private employers [were] voter protection laws, which barred employers from discriminating against employees based on how the employees voted.”189 Such voter protection statutes have been on the books since the mid-1800s, but more recent legislation protects a far broader range of employee political activity from employer interference. To take just a few examples: Connecticut prohibits employment discrimination based on the “exercise . . . of rights guaranteed by the First Amendment”; eight states prohibit employers from retaliating against employees for “engaging in political activities”; two states along with Puerto Rico and the District of Columbia prohibit discrimination based on employees’ political party membership; and three states prohibit discrimination against employees for “engaging in electoral activities.”190 Unlike in the case of union organizing, moreover, there is no evidence that management violates these statutes repeatedly or with impunity.

Outside the mainstream of labor and employment doctrine, further support can be found in cases involving the tort of discharge against public policy. In Novosel v. Nationwide Insurance Co.,191 the employer instructed its employees to engage in political-organizing activity—including canvassing and signature gathering—in support of an insurance reform bill. The plaintiff-employee in the case objected to the firm’s political stand, refused to engage in the political work, and was fired for doing so. The Third Circuit held that the discharge was tortious because it was in violation of public policy. As the court put it, where an employer “conditions employment upon political subordination,” the employer has exceeded the scope of its legitimate managerial authority.192

The point here is not that management discipline of employees for political-organizational activity would be illegal under existing law.193 Rather, the point is that these strands of labor and employment law reflect a view that the exercise of managerial power is more acceptable with respect to the economic terms and conditions of employment than with respect to employees’ political activity. To the extent that these legal principles capture a socio-cultural view of the appropriate reach of managerial power,194 they support the proposition that managerial opposition to political unions would be met with a less forgiving response than is managerial opposition to traditional unions.

B. Political Organizing Without Collective Bargaining?

The preceding Section provided reasons to predict that unbundling the union’s political and economic functions would increase the prospects for political organizing through the union form. Although there are no extant examples of the political unions this Essay envisions, this Section will offer some preliminary evidence that under an unbundled regime workers would in fact organize political unions. The evidence comes from recent organizing efforts in sectors of the labor market to which, for various reasons, traditional labor law does not apply. In these sectors, labor law does not require unions to bundle political and economic functions. Indeed, in some of these sectors, unions are legally precluded from collective bargaining but have nonetheless engaged in successful political organizing campaigns.

In the homecare sector, traditional collective bargaining has often been legally impossible because homecare workers are classified either as employees of the single clients for whom they work or as independent contractors.195 Despite the legal impossibility of collective bargaining, unions have nonetheless led campaigns to organize homecare workers for political action. In Illinois, for instance, at a time when homecare workers were barred from collective bargaining, the homecare workers union led a campaign to enact a “Homecare Workers Bill of Rights.”196 Throughout the campaign, union members hosted meetings with state representatives, testified at legislative hearings, and lobbied on behalf of the bill.197 The homecare union also took an active role in more traditional electoral politics, running the field operation for Mayor Harold Washington’s reelection campaign in two key wards on Chicago’s south and west sides.198 Members participated in the union’s door-to-door and onsite voter registration efforts, staffed phone banks that the union ran on Washington’s behalf, and took a central role in the union’s get-out-the-vote program on election day.199

The Service Employees International Union’s recent experience in the nursing home industry also offers some evidence of the potential of politics-only organizing. After years of unsuccessful attempts at traditional organizing campaigns in California nursing homes, the union sought to reorient its efforts away from collective bargaining and toward politics.200 This led the union to craft an agreement with several nursing home chains that granted the union access to employer property for the sole purpose of discussing political mobilization with the workforce.201 The union’s organizing efforts were thus strictly limited to political campaigns—ones designed to increase the state’s financial support for nursing homes—but the efforts were nonetheless successful in generating worker support and participation.202

It is worth noting that there is a strand of labor research suggesting employees are more likely to participate in a union’s political program if the union is successful at delivering economic goods through collective bargaining.203 This research indicates that a worker’s “commitment” to her union predicts much about if, and how, the worker will participate in union activities.204 And, as Herbert Asher and his colleagues report, commitment is highly correlated with the union’s economic performance: a worker who believes that her union has successfully improved wages and benefits is more likely to have a high level of union commitment than is a worker who has not benefitted from the union in these ways.205 Asher et al. also find that the relationship between union commitment and participation holds with respect to the union’s political activities.206

Despite these findings, this literature does not imply that the loss of the collective-bargaining function would preclude worker support for political unions. To start, even if it is a union’s economic performance that fuels worker support for the union’s political program, this does not entail a conclusion that collective bargaining is a necessary predicate for political participation. To the contrary, the union’s political program can itself generate economic returns for the membership.207 Next, Asher et al.’s work suggests that economic performance is important to union commitment because, traditionally, unions are formed for the purpose of securing economic goods. Thus, they write: “Unions came into being to deliver material benefits to their members: better wages and benefits, and better working conditions. Not surprisingly, union members judge their union based on how it delivers on these promises.”208 Accordingly, a political union formed initially for non-economic purposes—for example, to expand gun rights—would not need to deliver economic goods in order to secure member commitment.

Moreover, while it is true that union commitment is related to political participation, commitment is not the only explanatory variable. Asher et al. themselves report that several other factors are significant correlates of political participation: workers’ agreement with the political positions taken by the union, workers’ views of what the authors call the “appropriateness of union political activity,” and whether union membership is voluntary or mandatory.209 In fact, political agreement is as strong a predictor of participation as commitment is, and, when taken together, these three other factors have about three times stronger a correlation to participation than does commitment.210 This is relevant to our analysis because workers who join a political union will do so voluntarily.211 They will accordingly be quite likely to agree with the union’s politics and to view the union’s political activities as appropriate.

IV. law and organizing: designing an unbundled regime

The last Part argued that unbundling the union would improve the prospects for unionization as a political-organizational vehicle. But an important design question remains: what statutory work is required to unbundle the union and make political unions viable? This Part describes these necessary statutory reforms.

But before doing that, it is important to reiterate that even an unbundled labor law would need to offer several types of affirmative protection to employees engaged in the project of organizing political unions. These protections would not be extensive, but would be critical. In Section II.B, this Essay described the ways in which labor law enables traditional unions to take advantage of the workplace as a locus of organizational activity. The law functions this way by prohibiting employers from banning employee speech about organizational activity in non-work areas of the workplace and during non-work time. The law also allows unions to use the employer’s payroll system as a dues-deduction mechanism, gives unions access to the employer’s informational resources about employees for organizational purposes, and offers anti-retaliation protection to workers engaged in organizational activity. To facilitate the successful organization of political unions, an unbundled labor law should offer the same kinds of guarantees. Thus, employers should be prohibited from banning employee speech about political unions in non-working areas of the workplace and during non-work time. Allowing political unions to use payroll deduction as a mechanism for collecting dues from those employees who consent to such payments would enable political unions to overcome the collective action costs involved in dues collection.212 Political unions should be provided access to information about the composition of the workforce, at least subsequent to a showing of support by some significant portion of the workforce. And, finally, employers should be prohibited from retaliating against employees based on the employees’ support for or participation in political unions.213

To be sure, every provision in the National Labor Relations Act that deals with union organizing rights—and all of the administrative and judicial interpretations of those provisions—would impact political unions’ potential. Amending the statute to permit non-employee organizers to access company property, or to make punitive damages available against employers who discipline employees for union activity, or to add a private right of action, would all contribute to the success of organizational activity. But these avenues are beside the point. What matters here is that unbundling the union and preserving for political unions labor law’s basic and extant protections would itself improve the prospects for unionization as a political-organizational vehicle.

The actual statutory work involved in unbundling would not be extensive or difficult. The National Labor Relations Act gives employees the right to “self-organization,” which includes the right to “form, join, or assist labor organizations, to bargain collectively through representatives of their own choosing, and to engage in other concerted activities for the purpose of collective bargaining or other mutual aid or protection.”214 Amendments aimed at unbundling unions’ twin functions would establish that the right to “form, join, or assist labor organizations” includes the right to “form, join, or assist organizations not for the purpose of collective bargaining.” Concomitantly, amendments would clarify that the “concerted activities” protected by the law include concerted activities related to the range of political activities that a political union could undertake, whether or not such activities would qualify as “mutual aid or protection” under current law.215

Next, under current law, once a group of employees forms a labor organization, that organization has a statutory obligation to “bargain collectively with [the] employer,”216 and such collective bargaining must include bargaining over the economic terms of the employment relationship.217 An amended statute would make clear that, as to the new non-collective bargaining organization, there would exist no “mutual obligation” on the part of either an employer or the union to bargain over terms and conditions of employment. The unfair labor practices section of the law would reflect this change by clarifying that an employer does not violate the law by refusing to bargain with such a union, and that such a union does commit an unfair labor practice by attempting to engage the employer in bargaining over terms and conditions of employment.218 Finally, recall that the current regime requires the employer to negotiate—as part of its overall collective-bargaining obligation—dues checkoff provisions.219 Because political unions will not engage in collective bargaining, an unbundled labor law would need to establish that a political union designated by employees is entitled to collect dues directly from the employer through a payroll deduction system.220

Beyond these straightforward statutory moves, unbundling would require addressing one important conceptual issue. Under current law, an employer is obligated to recognize a union if and when a majority of employees in a relevant “bargaining unit”—essentially, a subgroup of the workforce that shares a set of common interests—votes for union representation.221 Moreover, under current rules, if a majority of workers votes to unionize, the union becomes the bargaining representative of all the workers in the unit, even those who opposed unionization.222 This is the so-called rule of exclusive representation, and it is designed to ensure the efficacy of collective bargaining: by establishing that the union speaks for all the workers in the unit, exclusive representation aims to prevent employers from playing different factions of workers off each other to the detriment of the whole.223 At least in non-right-to-work states,224 the rule also permits the union to collect mandatory dues from all the workers in the unit, even those workers who voted against unionization, in order to prevent any workers from free riding off the financial contributions of others.225

With regard to political unions, however, there is no call for exclusive representation: because the union will not bargain over a contract with the employer, the threat of the employer exploiting factions to undermine the union’s bargaining strength is inapposite. Moreover, given the Supreme Court’s jurisprudence in this area, mandatory dues collection would likely not be permissible in political unions, even in the private sector. As the Court has held, the NLRA permits unions to collect mandatory dues from all employees in a bargaining unit, but only for the purposes of collective bargaining and contract administration.226 With respect to the union’s political program, all dues payments must be fully voluntary, and employees accordingly must be provided a right to opt out of funding any of the union’s political expenditures. And although the Court’s holding on this score with respect to private sector unions is formally based on its interpretation of the statute, it is clear that the Court’s reasoning is either driven by the canon of constitutional avoidance or, at the least, informed by “constitutional values.”227 With respect to public employees and public-sector unions, the rule is explicitly a constitutional one.228

Accordingly, a legal regime that permitted unions to collect dues for political purposes from employees who objected to paying those dues would be flatly unconstitutional in the public sector and it would face significant constitutional difficulties in the private sector.229 Even if constitutionally permissible, however, requiring employees to pay dues to support a political union would be undesirable as a policy matter.Indeed, for the same reasons that employer opposition to employee political activity would be viewed skeptically, a rule requiring employees to fund union political activity with which they disagreed would be viewed with similar skepticism. This hostility would produce not only political opposition to the statutory changes necessary for unbundling, but would also impede the organizational potential of political unions once authorized by statute. Thus, rather than serving as exclusive bargaining agents, political unions should be designed to resemble the kind of “minority” or members-only unions that Charles Morris describes: employees could form a political union even if the union is supported only by a minority of the workforce, and the union, in turn, would represent and collect dues from only those workers who affirmatively desire to be members.230

V. beyond the workplace

Because unbundling the political and economic functions of unions will better enable political organizing by those in the lower and middle classes, unbundling is justified by a commitment to representational equality. But the goal of representational equality invites us to think beyond unions to a more comprehensive set of reforms designed to facilitate organizing by underrepresented groups. A full discussion of such a program is beyond the scope of this Essay, but a few preliminary observations are in order.

As this Essay has argued, labor law has facilitated unions’ organizational success by providing a geographic locus for organizational activity, offering a mechanism for funding organizational activity, making available the information necessary for organizing, and protecting employees against retaliation. These same basic legal interventions could be deployed to facilitate organizing in other contexts. In particular, where the government provides services to constituencies in the relevant income groups, it could supplement the services it provides with legal rules designed to facilitate political organizing around the service. Importantly, such interventions would be substantively neutral as a matter of design: they would enable organizing but would take no position on—and indeed exert no influence over—the types of policies the groups would pursue once organized.

This suggestion may sound quite foreign, but it has historical roots. During the mid-1960s, as part of President Johnson’s war on poverty, Congress established Community Action Programs (CAPs) as part of the new Office of Economic Opportunity (OEO). The CAPs were charged with involving low-income community members in the design and implementation of the anti-poverty programs that were to serve them. In fact, the Economic Opportunity Act of 1964 dictated that CAPs were to be “developed, conducted, and administered with the maximum feasible participation of the residents of the area and members of the groups served.”231

Although the statutory charge of maximum feasible participation was a highly contested one, to those running the OEO and the CAPs, “participation” implied more than simple involvement in the bureaucratic and administrative operation of welfare programs. Under the OEO, community participation came to imply governmental support—including financial support—for political organizing by the poor. Thus, in the leading study of the CAPs, J. David Greenstone and Paul Peterson explain that community action “became an attack on political poverty, oriented toward increasing the political participation of previously excluded citizens, particularly black Americans.”232 As Greenstone and Peterson describe:

In order to achieve the maximum in participation . . . CAPs were expected to redistribute political power through the mobilization of deprived groups. Techniques designed to achieve this goal included recruitment of issue-oriented community organizers, financial assistance to indigenous community organizations, formation of community corporations, voter registration drives, [and] sponsorship of protest demonstrations . . . .233

Indeed, the OEO itself suggested that local CAPs “assist the poor in developing autonomous and self-managed organizations which are competent to exert political influence on behalf of their own self-interest.”234

In the contemporary setting, public assistance programs provide a promising context for expanding on the labor law model developed in this Essay.235 Following this model, revised public assistance statutes could, first, make welfare centers available for political-organizational activity.236 Welfare centers, like workplaces, provide natural gathering places for recipients and would offer a centralized location for organizing. Next, just as labor law allows unions to collect dues through authorized payroll deductions, this regime could enable benefits recipients to authorize dues payments directly from their benefits.237 As labor law gives unions the names and addresses of workers in the bargaining unit, revised benefits laws could give welfare organizers the names and addresses of benefits recipients.238And, finally,the law would need to ensure that recipients who became active in political organizing efforts faced no negative repercussions for doing so.

Public education—one of the largest government-run programs in the United States239—is another context in which law might generate political organizing.240 By making public schools available for organizational purposes, providing information about parents to organizers, and allowing organizers to use the public schools’ administrative capacity to facilitate dues collection from parents, the state could help build a political organization of public school parents. Public hospitals, libraries, and even public recreation centers could similarly be used to support political organizing among the poor and middle class.


The risk that economic inequalities will produce political ones, long a concern of democratic theorists and political scientists, has led to several generations of campaign finance regulation designed to get money out of politics. But these efforts have not succeeded. Rather than struggling to find new ways to restrict political spending by the wealthy, this Essay has argued that representational equality can be advanced by legal mechanisms designed to facilitate the organizational capacity of those for whom the current system has become unresponsive.

Historically in the United States, labor law has been the most important legal mechanism of this type, and the labor union a primary vehicle for the political organization of the poor and middle class. Accordingly, as this Essay has shown, labor law offers important lessons about how to construct a program of legal interventions designed to facilitate political organizing among currently underrepresented groups. The unbundled union, in which political organization is liberated from collective bargaining, constitutes one promising component of such a broader attempt to improve representational equality.