Freedom of Expression in the Workplace (2024)

The workplace is a distinctive expressive domain because the ‘censor’ and the speaker are typically bound together by an employment contract that affords the former a large measure of hierarchical control over the latter. The employer, having hired the employee to do a job, has legitimate interests in regulating some employee speech. The employee is typically dependent on the employer for their livelihood, and vulnerable to the employer’s overreaching beyond those legitimate interests. Those features of the employment relationship give rise to a distinct set of questions about the value and limits of free speech in the workplace setting, public or private. This chapter will focus on how US law—primarily constitutional law but also non-constitutional law—has dealt with those questions.

The United States—an outlier among Western democracies in many respects—obviously has its own distinctive history and constitutional make-up. Two features of US constitutionalism are especially relevant here. The first is the famously robust commitment to freedom of expression relative to other political values with which it can sometimes conflict. A second feature cuts in the opposite direction: the ‘state action’ doctrine blocks the entry of constitutional free speech principles into ‘horizontal’ relations among citizens, including the employment relations through which most individuals derive their livelihood.

Beyond those distinctive features of US constitutionalism, the US law of work is distinguished by the persistence of ‘employment-at-will’—the presumptive right of both parties to terminate employment at any time and for any reason or no reason at all. That principle potentially empowers employers to censor, suppress, or punish employees’ speech and associations—on-duty or off-duty—on pain of dismissal. In the private sector that power is unconstrained by the Constitution, though it has come to be constrained by a variety of narrower non-constitutional speech protections. In the public sector, employer action is state action that is at least potentially constrained by the First Amendment—though not, as it turns out, by the famously robust First Amendment principles that govern in the public square, but by a weaker workplace-specific version of those principles. The employment relationship between the speaker and the censor sharply inflects the former’s free speech rights and the latter’s power to control speech.

While the US law governing freedom of expression in the workplace is unique in some ways, the problems it addresses will arise in any society that both recognizes the value of freedom of expression and channels labour into the production of goods and services largely through the institution of employment. In particular, clashes between employees’ freedom of speech and association and employers’ claimed prerogatives may arise as to expression that is either at work or off-duty and outside the workplace; expression whose content is either related or wholly unrelated to the employment; or expression that is actually part of the employee’s job. Several of the many possible permutations have arisen in US Supreme Court decisions on the First Amendment rights of public employees. The answers the Court has given in those decisions might be instructive, whether or not they are persuasive, both in the US private sector and in other societies grappling with these issues.

As a preface to those doctrinal matters, however, we should begin by asking whether and why we should care about freedom of expression in the workplace. From that preface, the chapter will turn to the free speech rights of public employees under the First Amendment, and then more briefly to the non-constitutional free speech rights of employees in the private sector, before concluding.1

22.1 Why Freedom of Expression in the Workplace Matters

As a way into the question of why we should care about free speech in the workplace, imagine a society with none of it. Specifically, imagine an electoral democracy with free speech protections against the sovereign, and with pure employment-at-will: both employers and employees are free to end the employment relationship at any time and for any reason or no reason. In that system, employers could freely dictate their employees’ associations and expression, on and off the job, and could enforce those dictates by the threat of discharge. Employees would be free to quit, but the employer’s commands would effectively be binding on employees up to the limits of their economic dependence on the employer.

In that imagined world, the employer could prohibit employees from speaking out contrary to the employer’s interests and views—from disclosing dangerous and illegal activities of the employer, or opposing the employer’s favoured political views and candidates—all on pain of discharge. Indeed, the employer could compel employees to actively support the employer’s political preferences. And unless we import into this imagined world the secret ballot (which gained currency in the United States only in the late nineteenth century), employers could dictate their employees’ actual votes if they chose to do so. Employees could be dragooned—to whatever extent they needed or wanted to keep their job—into amplifying the political voices of their employers and of the monied classes instead of expressing and promoting their own opinions and interests.

If we extend these conditions into government employment, it would mean that incumbent elected officials and their appointees could prescribe whatever conditions they wished on public employment. The employer could demand employees’ silence about political matters and abstention from political activity, both off-duty and on-duty, or they could demand employees’ strict loyalty to and active political support of the officials or party in power, in either case backed by the threat of discharge. Incumbents could thus entrench their own power and insulate themselves from electoral challenge—not absolutely but by a margin that would grow with the public payroll. And much information and informed opinion about the operations of government, and about corruption, deception, or abuse of power, could be bottled up inside government offices by a code of silence.

In this imagined world, employees would still be formally free to speak their minds and vote their own preferences—as long as they were willing to quit or court dismissal. That freedom would be largely illusory, however, for those who lacked the resources to survive without a job and who could not readily find another employer that did not impose the same restrictions. Not all employers would exploit their power over employees; but the latters’ freedom and political independence would be tenuous and subject to their employers’ whims. Political equality among citizens would be a formalist fiction. At least if a large share of the citizenry were economically dependent on employment, it is debatable whether this imagined society could fairly call itself either free or democratic.

Is the problem of employee unfreedom in this imagined society dissolved by the fact that it arises within a voluntary contractual relationship that the employee freely enters and can freely leave?2 On a traditional liberal understanding of political freedom, one might be untroubled by non-state restrictions to which one has consented in the contractual sense; after all, the freedom of contract is itself an essential aspect of liberal freedom.3 For others in the republican tradition, the problem of employer domination is more serious; a free and self-governing society must ensure citizens’ freedom from domination by others, not just the state.4 Neo-republicans, like their forebears, are more sensitive than classical liberals to the problem of domination within non-state or ‘horizontal’ relationships, and less beguiled by the magic of contractual consent. But thoughtful liberals and republicans alike would be concerned by the spillover effects of employer domination in our imagined system of public discourse, political equality, and democratic accountability.

One need not simply imagine such a system, for something much like it prevailed in the United States for much of its history. It was against that background that the modern US law of employee speech rights evolved. In that evolution one can discern a gradual and partial shift from a starkly liberal conception towards a more republican conception of what kind of freedom matters in a democratic society. Let us turn first to the highly articulated body of law that evolved in public sector employment under the First Amendment.

22.2 Public Employee Speech Rights under the First Amendment

Until the late 1920s, the First Amendment did not amount to much, at least in the courts. It did not even entail a right to speak in the public square, for that was the government’s property to govern as it wished.5 By 1940, the Supreme Court had begun deploying the First Amendment to open up public spaces for expression and to limit the power of the government-as-sovereign to regulate citizens’ speech. For several more decades, however, the government-as-employer remained unconstrained by the First Amendment. Although action by a government employer was ‘state action’, the so-called ‘rights-privileges’ doctrine blocked employees’ constitutional claims: public employment was not a right but a privilege that the government could withhold or withdraw at its whim—for good reason, bad reason, or no reason at all.6 Nothing in the Constitution thus constrained government censorship or punishment of its employees’ speech and associations. As Justice Holmes memorably put it: ‘[t]he petitioner may have a constitutional right to talk politics, but he has no constitutional right to be a policeman’.7

In the constitutional vacuum created by the rights-privileges doctrine, various government bodies pursued diametrically opposed conceptions of the proper relationship between public employment and partisan electoral politics. Where political patronage, or the ‘spoils system’, held sway, public employees could get or keep their jobs only by joining and actively supporting the party in power. Elsewhere, and especially in the federal government, a progressive movement in favour of merit-based civil service reached its apotheosis with the Hatch Act, which prohibited most federal employees from actively participating in partisan politics. Both systems deprived public employees of political freedoms that the First Amendment guaranteed to other citizens, but both were constitutionally permissible conditions on the ‘privilege’ of public employment.

The rights-privileges distinction came under growing pressure during the 1950s as the hydra-headed anti-communist crusade gained steam. Legal elites, including those on the Supreme Court, came to recognize that the government could command ideological conformity and quash dissent, not only by exercising its coercive powers as sovereign but also by conditioning citizens’ access to professional licences, public contracts, public employment, and other economic goods that the government controlled. Accordingly, rights-privileges gave way to the doctrine of ‘unconstitutional conditions’,8 and some government entitlements, including some jobs, were deemed ‘property’ protected by due process.9 Under those doctrines, even if the government is not required to grant individuals professional licences, jobs, or welfare benefits, it may not deny those valuable benefits for unconstitutional reasons (or in some cases without due process of law). That reframing of the nature of government employment teed up the question whether the government may condition a job on the sacrifice of the freedom of speech protected by the First Amendment. In other words, could the government fire Holmes’s policeman for talking politics?

In the public sector, First Amendment questions might arise either over the discharge or discipline of an individual employee based on expressive activity or over broad policies restricting public employees’ beliefs, associations, and expression (like the Hatch Act). Individual personnel actions have generated the bulk of the litigation and a body of doctrine that has exercised a certain gravitational pull on the Court’s treatment of systemic policies that regulate the speech of large groups of public employees.10

22.2.1 The First Amendment in Individual Disciplinary Decisions

The First Amendment made its official debut within public employment in Pickering v Board of Education,11 which overturned the discharge of a teacher who had written a letter to the editor criticizing the school board’s budgeting process and its allocation of funds between educational and athletic programmes. The Court emphasized that the subject of Pickering’s letter was ‘a matter of legitimate public concern’ as to which ‘free and open debate is vital to informed decision-making by the electorate’.12 Because teachers like Pickering were especially ‘likely to have informed and definite opinions’ on such matters within the educational system, it was ‘essential that they be able to speak out freely on such questions’.13 The interest of the public in hearing what they had to say thus bolstered public employees’ interest in speaking freely in the public sphere.

In opening up government workplaces to First Amendment scrutiny, Pickering opened up a raft of questions about how that would work. One thing was clear from the outset: the Court had no intention of importing into government workplaces the ‘uninhibited, robust, and wide-open’ debate of the press and the public square.14 As the Court explained in Pickering:

Thus was born the Pickering balancing test, which seemed capable of handling the full gamut of free speech controversies that could arise out of public employment. Yet the sheer number of potential controversies, coupled with the indeterminacy of the test, seemed to ensure a steady flow of those claims into the federal courts.

The Supreme Court tightened the spigot on those claims in Connick v Myers.16 Sheila Myers had been fired for circulating a questionnaire among her co-workers on issues such as employees’ morale, confidence in supervisors, and fairness in transfer and promotion decisions. Averring that ‘government offices could not function if every employment decision became a constitutional matter’,17 the Court held that only speech on ‘matters of public concern’ was protected under Pickering.18 Connick picked up the key phrase from Pickering, where it appeared to function as one dimension of the overall balancing test, and promoted it to a threshold test. Although it had long been clear that expression on public issues was at the core of First Amendment values, Connick broke new ground in requiring the speaker to bring her speech within that core as a threshold matter.19 Was it not for the public to decide what it was concerned about? Now it was for the courts to define the contours of the category of speech on matters of public concern. And it did not generally include speech on ‘internal office affairs even within an important public office’.20

What was now the two-step Connick–Pickering test cut out a good deal of what employees might want to talk about with each other. But what about ‘water-cooler’ speech among co-workers on other topics? That scenario might test the joints of Pickering’s rationale, for casual conversations among co-workers are of no immediate value to voters. Yet those conversations surely play a role in the formation of public opinion; they may also contribute more diffusely to civic life by cultivating bonds across lines of identity that separate citizens in most social settings.21 And individuals’ freedom to converse with their co-workers during the work day is surely an aspect of personal liberty.

The Court gave at least a tacit nod to those free speech values when the water-cooler scenario arose in Rankin v McPherson.22 McPherson, a clerical employee in a county sheriff’s office, had been overheard chatting with a co-worker about the assassination attempt on President Reagan; she was fired for saying, ‘If they go for him again, I hope they get him’. The Court held that the comment was on matters of public concern given its context—a conversation about Reagan’s policies and their impact on racial minorities—despite its provocative tinge.23 Proceeding to the balancing test, the Court held that McPherson’s interest in freely speaking on such matters at work outweighed the negligible impact of her comment on legitimate managerial interests.24 Rankin represented a qualified victory for the freedom of speech among co-workers, albeit only as to speech on public issues.

Until the mid-2000s, the key public employee speech cases in the Supreme Court involved speech that either took place at work (as in Rankin), or was about the employer (as in Pickering), or both (as in Connick). The full spectrum of public employee speech controversies included speech that was either more or less connected to their work—speech that was either part of the employee’s job, or was wholly unrelated to the job. Those scenarios probed the scope of the Connick–Pickering test.

The Court confronted the problem of speech-that-is-the-job in Garcetti v Ceballos.25 Assistant district attorney Ceballos claimed that he had suffered reprisals after opining in a memo, and later testifying, that an arresting police officer may have lied in an affidavit that was the basis for an arrest warrant. Ceballos believed he was legally and professionally obligated to disclose exculpatory information about possible police misconduct. But the Supreme Court held that the First Amendment offered him no refuge because the memo and the testimony were within his job duties as supervising attorney in the underlying prosecution:

Recall Pickering’s reference to ‘the interests of the [employee], as a citizen, in commenting upon matters of public concern’. Just as Connick transformed the last phrase into the threshold ‘public concern’ test, Garcetti gave independent threshold significance to the phrase ‘as a citizen’. So while Ceballos had plainly spoken on matters of public concern, he lost because he ‘did not act as a citizen when he went about conducting his daily professional activities, such as supervising attorneys, investigating charges, and preparing filings’.27 In effect, the speech through which an employee performs the job belongs to the government employer, and is subject to its plenary control. If the legislature or the agency decides to protect employees who report official mistakes, misconduct, or abuse, (for example, through a ‘whistle-blower’ statute), so be it. But the employer is otherwise free to punish the employee without any burden of justification.

The problem of speech-that-is-the-job poses a serious dilemma. Much of the actual work of public employees is performed through speech or writing; and much of that expression is inescapably on matters of public concern, sometimes including official misconduct. No doubt it would be a difficult and sensitive judicial task to balance the individual and public interests in airing crucial matters of public concern against managers’ ability to evaluate employees’ job performance without judicial second-guessing. Perhaps the employee should have to make some heightened showing of the public importance and accuracy of the speech.28 Or perhaps the solution lies in the environs of ‘First Amendment due process’.29 But Garcetti gave a categorical verdict in favour of managerial freedom, however unjustified, over employee freedom of expression, however valuable.

At the opposite end of the spectrum of job relatedness—that is, off-duty speech that is unrelated to the job or agency—one might look for a similarly categorical approach, this time favouring free speech interests. That is one reading of United States v National Treasury Employees’ Union (NTEU),30 which struck down a statutory ban on federal employees’ accepting any compensation for off-duty speeches or writings, even those with no connection to the employee’s official duties. The opinion was a bit opaque about why the speech was protected: was it because it all fell within the ‘protected category of citizen comment on matters of public concern’, or because it was off-duty and ‘unrelated to the employment’?31 We will return to NTEU for its analysis of broad policies restricting employee speech; but its import was unclear for challenges to individual discharge or discipline based on expression that was unrelated to the work.32

That issue seemed to arise in City of San Diego Police Department v Roe,33 which upheld the discharge of a police officer for making and selling sexually explicit videos in which he appeared in (and then partly out of) a generic police uniform while issuing and revoking traffic tickets. City of San Diego, a per curiam decision, delivered both good news and bad news for public employee speech rights. The good news lay in the Court’s reading of NTEU on a crucial point: an employee’s speech that is unrelated to the employment—that is, unrelated in time (off-duty), place (outside the workplace), and subject matter—enjoys stronger First Amendment protection than is afforded by Connick and Pickering. In short, when the nexus between the speech and the work becomes too attenuated, the government’s heightened power to regulate its employees’ speech runs out, and employees recover something like their full freedom as a citizen vis-à-vis the government. That good news for public employees is tempered, however, by the Court’s narrow view of what speech counted as ‘unrelated to the employment’. It did not include Roe’s p*rnographic videos because his use of a generic police uniform ‘brought the mission of the employer and the professionalism of its officers into serious disrepute’.34 Roe was thus relegated to the Connick–Pickering test, which he failed at the threshold: His videos were not on matters of public concern.35

As the doctrine now stands, individual public employee speech cases fall into four categories, as depicted in Figure 22.1. ‘Garcetti speech’ that is part of the employee’s job performance is wholly unprotected by the First Amendment. Speech that is not part of the job itself but is somehow related to the employment (in time, place, or subject matter) comes in two flavours: ‘Pickering speech’ on matters of public concern, protected under a balancing test, and ‘Connick speech’ that is not on matters of public concern, and is unprotected. Speech that is ‘unrelated to the employment’ in time, place, and subject matter (call it ‘NTEU speech’) is constitutionally protected to some greater but still-uncertain degree, whether or not it is not on matters of public concern.

One question, to which I will shortly turn, is whether this doctrinal matrix, developed to resolve quotidian conflicts between government managers and individual employees, applies to broader rules and policies affecting public employee speech rights. The former tend to fall into recurring patterns, while the latter are more likely to be sui generis. Moreover, some of the concerns that shadow individual disputes—concerns about micro-managing government offices and unleashing a flood of litigation—fall away in the context of large-scale rules and policies. Indeed, in that context, some of the arguments for affording the government greater power to control the speech of its employees than the speech of citizens at large might be misplaced.

Consider the Supreme Court’s own answer to this question: ‘[w]hat is it about the government’s role as employer that gives it a freer hand in regulating the speech of its employees than it has in regulating the speech of the public at large?’:36

This explanation meshes well with the Pickering balancing test, though somewhat less well with Connick and Garcetti, which allow the government employer to restrain some employee speech with no showing that it ‘detract[ed] from the agency’s effective operation’. All in all, however, the Waters explanation seems keyed to the quotidian context of personnel management. And it hints at the larger stakes in those controversies for democratic governance.

Clearly democracy requires informed public debate about public affairs, the performance of public officials, and the workings of government; and public employees are crucial sources of information and informed opinions on these matters. But free public discourse and democratic processes for electing officials and expressing policy preferences do not translate into democratic governance unless elected officials can actually implement those policies. And they can do that only by effectively managing government workplaces and public employees. Democratic governance depends on both open and informed public debate about governance and public officials’ ability to manage employees and evaluate their performance. That is the central dilemma at the heart of the doctrine governing public employee speech rights. But that dilemma comes into play quite differently, if at all, when the government enacts broad speech-restrictive rules for its employees. Let us now turn to that branch of public employee speech doctrine.

22.2.2 Systemic Regulation of Public Employee Speech and Association

The entry of the First Amendment into public employment put at risk some deeply entrenched systemic features of public employment that had grown up in the era of rights-privileges. In particular, the First Amendment now cast a shadow over both the Hatch Act ban on partisan political activity by civil service employees and the still-extant state and local patronage systems that compelled partisan activity by public employees. But the two systems met different fates in the immediate wake of Pickering.

The Hatch Act ban on active partisan involvement in political campaigns survived its post-Pickering test in Civil Service Commission v National Association of Letter Carriers.38 The Act plainly banned forms of political expression that are normally highly protected under the First Amendment. Citing Pickering’s paeon to balancing of interests, the Court turned to the arguments on the other side. First, the Court agreed with the Act’s supporters that, paradoxically, it actually protected most employees’ political freedom—that prohibiting employees from engaging in partisan activity was necessary to protect them from being coerced to do so by their bosses.39 But the main defence of the Act was on systemic political process grounds: Congress had long ago determined that ‘the rapidly expanding Government workforce should not be employed to build a powerful, invincible, and perhaps corrupt political machine’.40 Beyond concerns about corruption and entrenchment of incumbents, Congress feared that a bureaucracy staffed with partisans entailed ‘hazards to fair and effective government’, including the risk of ‘political justice’.41 The Court found in these arguments a powerful justification for the Act’s infringement on public employees’ rights to participate actively in partisan politics.42

Letter Carriers and its disquisition on the evils of political patronage cast a shadow over the patronage systems still in place in state and local governments. The Court took the next step in Elrod v Burns, overturning the dismissal of several Cook County sheriff’s employees ‘solely because of their partisan political affiliation or nonaffiliation’.43 The decision was straightforward as a matter of First Amendment doctrine: ‘The right to associate with the political party of one’s choice is an integral part of this basic constitutional freedom’.44 But the Court also reprised the political process arguments that had won the day in Letter Carriers: political patronage impairs the ‘free functioning of the electoral process’, ‘starve[s] political opposition by commanding partisan support’, and ‘tips the electoral process in favor of the incumbent party’.45

The defenders of patronage had their own political process argument: political patronage helped to support the vitality of political parties and thus of the democratic process. According to the Elrod petitioners, ‘we have contrived no system for the support of party that does not place considerable reliance on patronage. The party organization makes a democratic government work and charges a price for its services’.46 The Court demurred, for while patronage supplied a dubious economic motivation for partisan involvement of a kind, it also impeded individuals’ more genuine support for political parties without jobs to distribute.47

Letter Carriers and Elrod together put a strong stamp of judicial approval on a particular conception of the proper relationship between public employment and partisan politics: non-partisan merit-based selection for public employees, but for a handful of politically-accountable policy-making officials. Good-government reformers had made headway over the last century in promoting the civil service system over the political spoils system. There is much to be said for that system, including the incidental protection of freedom of expression off the job that follows from requiring a job-related basis for discipline and discharge. Now the First Amendment was squarely on the side of the reformers, and in the hands of judges who were notably distant from, if not disdainful of, the grubby world of partisan politics. The subsequent collapse of the major political parties as grassroots organizations and orchestrators of electoral politics might suggest that there was something in the petitioners’ warning after all.48 Still, it is hard to see how the practice of assigning rank-and-file jobs based on party membership and participation could survive once the First Amendment had entered the scene.

Although both Letter Carriers and Elrod reviewed broad policies affecting public employee speech rights, it was only later in NTEU that the Court explicitly addressed the distinction between individual disciplinary actions based on speech and ‘a sweeping statutory impediment to speech’. Recall that NTEU struck down a statute banning federal employees’ acceptance of outside compensation, or honoraria, for nearly all speeches and writings, including those with no ‘connection with the employee’s official duties’.49 On the one hand, the Court observed, legislative, and especially congressional, policy judgments were generally entitled to greater deference than ‘an individual executive’s disciplinary action’.50 On the other hand, ‘the widespread impact’ of the statutory ban, along with its tendency to ‘chill[] potential speech before it happens’ cut in favour of closer scrutiny.51 The Court held that the Pickering balancing test still applied to the ban, but that ‘the Government’s burden is greater…than with respect to an isolated disciplinary action’. Specifically:

The government failed to meet that burden with regard to the honoraria ban.

NTEU suggested a possible template—a more speech-protective version of Pickering—for First Amendment review of broad policies regulating public employees’ speech, or at least their off-duty speech unrelated to the employment. The NTEU version of Pickering still recognized that the government’s ‘interests as an employer in regulating the speech of its employees’ differed from its interests as sovereign in ‘regulat[ing] the speech of the citizenry in general’. But the difference appears to be narrower, and the protection of speech greater, when the government is exercising regulatory authority across broad swaths of its workforce than when it is making individual personnel decisions based on speech with some bearing on the job.

22.2.3 Janus and the Double-Edged First Amendment

The latest First Amendment challenge to systemic policies governing public employees was Janus v AFSCME, which overturned a long line of cases on the constitutionality of compulsory union fees in the public sector. Janus requires a brief detour into the field of labour law.

Under US collective bargaining statutes, public and private, a union chosen by a majority of employees in a bargaining unit is the ‘exclusive’ representative of all employees, and is duty-bound to fairly represent all of them, whether union members or not. Inherent in such systems is a serious free rider problem: if employees get the benefit of union representation without paying for it through union dues, then union members must pay higher dues to subsidize the representation of non-members, which might lead some members to quit and become free riders. Accordingly, unions have long bargained for ‘union security’ provisions to compel all employees to pay dues or fees to the union as a condition of employment. To unions’ opponents, requiring individuals to pay money to an organization they oppose is a form of compelled speech; and when the state is doing the compelling, it violates the First Amendment. That claim has fuelled a tireless and well-funded battle against union security provisions on multiple fronts, political and judicial, since the New Deal.53

Over forty years ago in Abood v Detroit Board of Education,54 the Supreme Court established a compromise on the constitutionality of compulsory union fees in the public sector: individuals could not be compelled to support the political and ideological activities of unions, but they could be compelled to bear their fair share of the costs of negotiating and administering the collective agreement through a reduced ‘agency fee’. Although the latter ‘has an impact upon [objectors’] First Amendment interests’,55 the impact was modest:

A public employee who believes that a union representing him is urging a course that is unwise as a matter of public policy is not barred from expressing his viewpoint. Besides voting in accordance with his convictions, every public employee is largely free to express his views, in public or private, orally or in writing.56

That modest burden was outweighed by the government’s interest in maintaining a system of exclusive representation—versus ‘the confusion and conflict that could arise if rival…unions, holding quite different views [about terms and conditions of employment] each sought to obtain the employer’s agreement’57—and the obvious risk of free riding in that system. The Abood compromise was repeatedly reaffirmed in succeeding decades.

Janus reversed Abood. For the Janus majority, the exaction of agency fees was no minor imposition on free speech; it was akin to ‘compelling individuals to mouth support for views they find objectionable’.58 From that tendentious premise, it followed that agency fees must be subjected to ‘exacting scrutiny’.59 And while the majority grudgingly accepted the government’s interest in a system of exclusive representation, it denied that agency fees were necessary to support that system. The free rider problem attending exclusive union representation was no more serious, opined the majority, than that facing any other ‘private groups [that] speak out with the objective of obtaining government action that will have the effect of benefiting non-members’,60 and could not justify the burden on objectors’ free speech rights.

Janus is willfully obtuse about the distinctive free rider problem facing unions under US labour laws.61 Moreover, by forcing union members to subsidize the representation of anti-union (or merely opportunistic) free riders, Janus arguably creates a compelled speech problem much like the one it purports to solve.62 But the linchpin of Janus is the disingenuous equation of the compelled payment of small sums of money with the compelled expression of views.63 Abood and its precursors had opened that door by characterizing mandatory fees as a form of compelled speech; but that characterization was tempered by the recognition that the agency fee does nothing to impair the right of ‘every public employee…to express his views, in public or private, orally or in writing’, and to oppose positions taken by the union in politics or collective bargaining. Abood’s disapproval of compelled financial support for unions’ ideological activities was one half of a compromise that sought to respect both the free speech interests of objectors and the practical imperatives of a system of exclusive representation. But the Abood compromise proved to be highly vulnerable to the uncompromising version of the First Amendment wielded by the Janus majority.

The main rationale of Janus is grounded not in the jurisprudence of public employee speech rights surveyed above, but in a libertarian strain of First Amendment law that is becoming a useful bludgeon against the regulatory state.64 Justice Kagan in dissent called out the majority for ‘weaponizing the First Amendment, in a way that unleashes judges…to intervene in economic and regulatory policy’.65 The First Amendment is a protean weapon in that regard because much of the activity that modern states regulate—indeed, much regulation itself—can be portrayed as expressing something, verbally or otherwise. That includes the conduct of business or labour relations, the payment of health insurance premiums, and cake-baking.66 Merely complying with duly-enacted laws that one opposes, or paying fees to a duly-elected union that one opposes as in Janus, may be seen as compelled speech triggering ‘exacting’ First Amendment scrutiny.

The body blow that Janus struck against public sector unions is of a piece with the larger campaign against the regulatory state, for Janus depicts public sector unions as leading culprits in the growth of government and state budgets.67 More than was necessary to uphold the rights of objectors, the majority thus signalled its agreement with their opposition to public sector unions. Barely off-stage is public sector unions’ role in the electoral arena. Abood long ago ended objectors’ forced contribution to unions’ partisan electoral activities. But at least for anti-union activists, it was a more-than-incidental benefit of Janus that unions would now be strapped for cash, and less able to participate in electoral politics. Unions will have to spend more of their members’ dues representing a growing number of non-members, and perhaps redirecting some of the money those members would have chosen to spend on advancing their own shared political views through the union. That is another free speech problem that Janus creates.68

While the majority conspicuously ignored the jurisprudence of public employee free speech rights in its main argument, it managed to unsettle some of that jurisprudence in addressing what it called the ‘alternative justifications’ for Abood. Abood’s defenders had argued that the line it drew between unions’ political activities and their collective bargaining activities was compatible with Connick’s line between speech on matters of public concern and speech on ordinary terms and conditions of employment. Just as actual speech on the latter was unprotected, they argued, ‘compelled speech’ on such matters in the form of the agency fee raised no First Amendment concerns. But even if collective-bargaining related speech was on matters of public concern, its compelled support, like its suppression, should be subject to Pickering-type balancing, in which the government’s interest in defending exclusive representation from the threat of free riding should outweigh the limited burden of the agency fee on dissenters’ free speech interests. The Janus majority rejected every part of that argument, and in the process reconfigured some of its doctrinal elements.

First, the majority cast doubt on whether Pickering was applicable at all to ‘general rules that affect broad categories of employees’, and denied its relevance to ‘compel[led] speech or speech subsidies’.69 The paradoxical upshot is that the outright suppression or punishment of public employees’ speech on matters of public concern appears to trigger a lower level of scrutiny than does the exaction of a small fee for representational services that one opposes.

Second, in applying the law of public employee speech, arguendo as it were, the Court recast the Connick category of speech on matters of public concern: a single employee’s demand for ‘a 5% raise…would likely constitute a matter of only private concern…But a public-sector union’s demand for a 5% raise for the many thousands of employees it represents would be another matter entirely’, given its impact on the public fisc.70 Of course, the first might lead to the second; that is how speech works. But the Court deemed it obvious that, ‘[w]hen a large number of employees speak through their union, the category of speech that is of public concern is greatly enlarged, and the category of speech that is of only private concern is substantially shrunk’.71 That sounds right—and it even sounds like potential good news for employee speech rights.72 Janus’s ‘enlarged’ view of public concern speech could work in favour of unions and their members in future cases—for example, in challenging restrictions on union organizing or bargaining rights. It might even unsettle the disfavoured constitutional treatment of labour picketing versus picketing on ‘public issues’.73 Then again, maybe not, for Janus suggests that actual state suppression of speech is not as bad as state-compelled financial support of speech—once the latter is equated to forcing employees to ‘recite words with which they disagree’.74 Chances are slim, at least while this majority holds sway, that either Janus’s higher level of scrutiny for broad policies affecting employee speech or its expansion of the category of speech on matters of public concern will end up protecting the actual speech of either individual employees or unions.75

The Court sought support from elsewhere in the public employee speech canon, albeit only in the course of refuting the stare decisis argument: ‘[b]y overruling Abood, we end the oddity of privileging compelled union support over compelled party support’, struck down in Elrod.76 The ‘oddity’ was not so odd, however, given some crucial differences between political patronage and the agency fee: First, patronage compelled actual party membership and active involvement in party activities; the agency fee compels neither union membership nor union activity. Second, political patronage required support of partisan political speech; agency fees require support of only unions’ representational functions. Third, the union speech supported by agency fees yields tangible economic gains to employees that nearly always exceed the burden of the agency fee by a large margin; there was no parallel in the patronage context. Fourth, political patronage strongly discouraged individuals from joining or supporting competing organizations or causes; the agency fee does nothing of the sort.

In one sense, however, the analogy to Elrod holds: like Elrod, though for far flimsier reasons, Janus rejects a considered state policy judgment about the appropriate relationship between the government and its employees. Agency fees are one component of the policy of allowing public employees, by majority rule, to choose collective representation and bargaining on behalf of shared goals. In undercutting both the state policy and the collective choice of employees, Janus is doubly anti-democratic. To be sure, democratic self-governance is often in tension with individual liberty, and that makes the First Amendment a potential double-edged sword for public employees. Janus sharpened the edge that cuts against those who seek to amplify their collective voice in both the workplace and the political process through democratic institutions. Although Janus’s blow against public sector unionism is less dispositive than Elrod’s rejection of political patronage, it may lead to much the same end as individuals take up the invitation to exit from collective institutions that they no longer have to pay for.

22.3 Non-Constitutional Speech Rights in the Private Sector Workplace

For public employees, the First Amendment has become a weapon—double-edged and of fluctuating efficacy—against state policies, managerial judgments, and collective decisions that impinge on individual freedom of expression. Free speech claims in private employment have followed a very different course, for the state action barrier continues to block constitutional claims against private sector employers. Indeed, in the hands of the Lochner Court, the Constitution forbade the legislature from protecting employee speech and associations—specifically, union activity and membership—against employer reprisals: Any legislative restriction on employers’ freedom to fire employees was fatally at odds with the constitutional ‘liberty of contract’.77 But once the Lochner era gave way to the New Deal, law-makers began to carve out exceptions to employment-at-will and to restrain employer power over employee speech.

First and still foremost, the National Labor Relations Act (NLRA) of 1935 prohibits employers from interfering with employees’ union activity and ‘other concerted activity for…mutual aid or protection’.78 Even with no union in sight, employers are generally barred from punishing employees for joining together to protest poor working conditions or unfair employer actions (even if they protest by walking off the job),79 or for discussing their salaries with each other,80 or for publicly airing shared work-related complaints.81 Employees have the right to communicate with their co-workers about such matters at the workplace during non-work time as well as off-duty and on social media.82

The NLRA creates a rudimentary analogue to the First Amendment for the private sector workplace, complete with its own ‘public forum’ doctrine.83 But its protections are in some ways the reverse of the actual First Amendment in public sector employment. The NLRA protects speech about working conditions, but not speech on matters of public concern unrelated to the interests of employees as such. That is not because the values underlying the NLRA are at odds with the First Amendment values of freedom, democracy, and active citizenship; it is because the NLRA sought to pursue those values within the polity of the workplace. It aimed to empower workers vis-à-vis employer sovereignty, to render them citizens versus mere subjects of the workplace, and to bring a measure of democracy to the private sector workplace.

The NLRA protected employees’ freedom of speech and association both for its intrinsic value and for its role in promoting ‘industrial democracy’ in the form of union representation and collective bargaining. In that respect federal labour law mirrored central goals of the First Amendment within the larger polity. Indeed, the First Amendment might have been put to the service of those goals of workplace democracy within the public sector workplace. But as we have seen, the doctrine has gone in quite the opposite direction: Connick denied First Amendment protection in the public sector for speech on shared workplace concerns that would have been protected by the NLRA in the private sector. And then Janus turned the First Amendment into a cudgel against the collective bargaining systems, modelled on the NLRA, by which legislatures have sought to enhance public employees’ voice at work. Anti-union advocates are surely plotting to extend Janus into the private sector, though that effort faces a formidable state action hurdle.84

The NLRA protects a wide swathe of private sector employee speech in and about the private sector workplace.85 But it fails to protect much valuable speech, including speech on matters of public concern other than workplace issues and purely individual speech that has no element of ‘concerted activity’. Some of that speech finds refuge under a motley patchwork of federal, state, and local statutes and state common law doctrines that is impossible to summarize and synthesize here. But it is worth noting one important strain of that law, which protects ‘whistle-blowers’—employees who disclose harmful or illegal activity within the organization to the public, the government, or sometimes higher-ups in the organization.86 Crucially, whistle-blower laws are widespread in the public sector, where they supplement the First Amendment, and fill some of its gaps. But in the private sector, there is no constitutional protection to supplement; whistle-blower protections are islands of protection in a sea of employer discretion.

In some ways, whistle-blower protections track the First Amendment as it operates in public sector employment; but the divergences are revealing. For example, most whistle-blower laws protect employees only when they report or complain of statutory violations.87 In some ways that is narrower than Connick’s category of speech on matters of public concern. But in other ways it is broader, for many laws regulate terms and conditions of employment, and prohibit employer retaliation against employees who report or complain about violations. For another example, some early common law decisions denied protection to those who reported wrong-doing internally as part of their job.88 Increasingly, however, such reporting is within the zone of protected activity, contra Garcetti.89

Obviously, there is nothing shady or surprising in the fact that statutory and common law doctrines protect some speech that the Constitution does not. While Connick and Garcetti deny First Amendment protection to some public employee speech, they do not preclude legislatures or state courts from protecting that speech (as some public sector labour laws and whistle-blower laws do). And while the state action doctrine blocks constitutional claims against employer censorship and punishment of speech in the private sector, it poses no hurdle to legislation or common law doctrine that protects that speech. Still, it is noteworthy that legislatures and common law courts have often discerned significant value in employee speech that the Supreme Court has excluded from First Amendment protection even in the public sector.

What this very short and partial survey of private sector speech protections has left out thus far are the spaces in between those protections—spaces that, under the background rule of employment-at-will, are subject to plenary managerial power. Even relative to the limited First Amendment protection of public employee speech, the non-constitutional protection of employee speech in the private sector is spotty and sparse.90 Private sector employees are largely unprotected in their ‘water-cooler speech’ (except for that which is covered by the NLRA), in their off-duty political associations and expression, and in their refusal to support the employer’s political agenda.91 If there is any such thing as an exception that proves the rule, it might be the Third Circuit’s singular decision in Novosel v Nationwide Insurance Co.92 Novosel held that an employee who was fired for refusing to support his employer’s lobbying position had stated a claim in tort for wrongful discharge in violation of public policy—that is, the public policy embodied in the First Amendment. The case is widely admired by employment law scholars, but widely criticized or ignored in the courts.93

In the few remaining outposts of job security in the private sector, mostly in the small and shrinking domain of union representation, the employer’s burden of justifying discharge and serious discipline provides indirect protection of speech.94 Such ‘just-cause’ protections are pervasive elsewhere in the developed world.95 But in the at-will world in which most US employees work, that indirect protection of speech is lacking, too. What is left is employees’ right of exit, and whatever reputational pressures might face an employer that sought too openly to dictate or suppress employees’ off-duty, non-work related speech. That is not nothing, but it is nothing like what public employees, and most employees in the rest of the developed world, enjoy by way of freedom from employer censorship and repression of speech.

22.4 Conclusion

Woven through US history is an abiding attachment to the liberal idea that the freedom to choose whether to enter or exit an employment relationship was all the freedom employees needed vis-à-vis their employer. Over the last century, US courts and legislatures have carved employee rights, including free speech rights, out of the broad power that employers otherwise wield under their side of employment-at-will, that is, their power over continued employment. Even today, employment-at-will is the norm, and employee rights and freedoms—beyond the right to refuse or quit employment—are exceptional. There are by now quite a lot of exceptions, and together they carry the whiff of a competing idea—one that resonates with neo-republican thought—that a free society should protect employees from employer domination beyond ensuring their freedom to exit the employment relationship.96 That includes a measure of protection for employees’ freedom of expression.

But how much and what kind of protection? Specifically, how much and what kind of freedom of expression is compatible with the control that employers need (or demand) over those they hire to do a job? The Supreme Court set out to answer a version of that question for government employers and employees under the First Amendment. The resulting body of law might have offered a model for how legislatures and common law courts should balance free speech rights and employer needs in the private sector as well. But it has not turned out that way. As a descriptive matter, a comparison between the two bodies of law reveals as many paradoxes as parallels. The main currents of the law of private employee speech rights, far from tracking the Court’s First Amendment doctrines in the public sector, protect both more and less speech, and very different kinds of speech, than the latter. As a normative matter, neither body of law responds adequately to the problem that employer domination poses for employees’ freedom of expression in a democratic and capitalist society.

In the public sector, the doctrine’s categorical exclusions from the First Amendment leave employees vulnerable to unjustified employer domination, and leave a great deal of valuable speech unprotected. But a combination of statutory speech protections and civil service protections have done much to supplement those constitutional protections. In the private sector, employees are relegated to an incoherent collection of statutory and common law protections that leave too much room for employers to leverage their economic power into power over employees’ speech and associations. Perhaps employers have been too successful in persuading law-makers and judges of the necessity of managerial discretion and the adequacy of exit in curbing abuse. One way or another, the beguiling promise of consent through contract continues to exercise a strong gravitational pull on the law of employee speech rights.