Harvard Law Review Harvard Law Review Harvard Law Review

Harvard Law Review Forum

Can Unions Be Sued for Following the Law?

Responding to William Baude & Eugene Volokh, The Supreme Court, 2017 Term — Comment: Compelled Subsidies and the First Amendment

The full text of this Response may also be found by clicking on the PDF link to the left.

Here is a short summary of the right-to-work movement’s legal strategy in the aftermath of its victory in Janus v. AFSCME1×1. 138 S. Ct. 2448 (2018). : If you can’t kick a man when he’s down, when can you kick him? For within weeks of Janus’s pronouncement that the First Amendment forbids public sector unions to collect agency fees from objecting employees,2×2. Id. at 2486. right-to-work groups filed a flood of class action lawsuits seeking the refund of millions of dollars’ worth of fees that were paid in the years before Janus was even decided,3×3. See Noam Scheiber, Trump Nominee Is Mastermind of Anti-Union Legal Campaign, N.Y. Times (July 18, 2018), https://nyti.ms/2JESrtC [https://perma.cc/9GN7-BTVZ] (describing refund lawsuits filed in Washington, California, New Jersey, New York, Pennsylvania, Minnesota, and Ohio that could “cost unions hundreds of millions of dollars”). when such fees were indisputably lawful.4×4. See sources cited infra note 8. Commentators have observed that these retroactive refund suits threaten to bankrupt unions around the nation.5×5. See, e.g., Moshe Z. Marvit, For 60 Years, This Powerful Conservative Group Has Worked to Crush Labor, The Nation (July 5, 2018), https://www.thenation.com/article/group-turned-right-work-crusade-crush-labor/ [https://perma.cc/S5UR-GASJ] (arguing that “now that the [National Right to Work Legal Defense Foundation] . . . has succeeded in crippling unions’ ability to collect future dues, the next move will be to bleed them for past dues”).

In Compelled Subsidies and the First Amendment,6×6. William Baude & Eugene Volokh, The Supreme Court, 2017 Term — Comment: Compelled Subsidies and the First Amendment, 132 Harv. L. Rev. 171 (2018). Professors William Baude and Eugene Volokh argue that “Janus makes it likely” that public sector unions will indeed be liable under 42 U.S.C. § 1983 for refunds of money they collected in years before Janus was even issued.7×7. Id. at 201. At least so long as the fees were collected during the applicable statute of limitations. See id. at 202. We think otherwise, and this Response explains why.

We start in Part I by presenting a vision of the world as it would exist if Baude and Volokh are right. It turns out that imposing financial liability on public sector unions for conduct that was perfectly lawful when it took place (because both state law and judicial precedent authorized the unions to collect fair-share fees8×8. See Abood v. Detroit Bd. of Educ., 431 U.S. 209, 235–36 (1977) (upholding fair-share fees against First Amendment challenge); Aaron Tang, Life After Janus, 119 Colum. L. Rev. (forthcoming 2019) (manuscript at 11 nn.54–57) (on file with the Harvard Law Library) (collecting state laws that authorized collection of fair share fees). ) is a kind of maneuver that cannot be neatly confined to the context of union fee refunds.

In Part II, we explain why this unsavory state of affairs is hardly necessary. In fact, the law requires otherwise. In particular, we describe three legal arguments that should stop the union-refund suits from getting off the ground: careful application of the doctrine of civil retroactivity; defenses that were available against the most closely analogous tort at common law, including that unions acted in good faith reliance on existing law; and ordinary principles of class action certification.

I. What if Unions Can Be Sued for Following the Law?

The consequences of retroactive reparations in the wake of newly announced rules of constitutional law cannot be confined to the unions. Under existing law, a number of persons within the meaning of § 1983 do not receive qualified immunity.9×9. See, e.g., Richardson v. McKnight, 521 U.S. 399, 401 (1997) (holding that private prison guards are not entitled to qualified immunity); Wyatt v. Cole, 504 U.S. 158, 159 (1992) (holding that private parties who had invoked “state replevin, garnishment, and attachment statutes later declared unconstitutional” were not entitled to immunity); Owen v. City of Independence, 445 U.S. 622, 657 (1980) (holding that municipalities are not entitled to qualified immunity). If courts are required to award retroactive damages for breaches of new rules of constitutional law, these actors, too, could be successfully sued for damages awards simply because they followed state laws — even if those laws are endorsed by binding United States Supreme Court precedent.

Consider private prison guards. The Court held in Richardson v. McKnight10×10. 521 U.S. 399. that they are not entitled to qualified immunity, reasoning that private prison guards were not similarly situated to state employees in light of what the Court called “competitive market pressures.”11×11. Id. at 409; see also id. (“Competitive pressures mean not only that a firm whose guards are too aggressive will face damages that raise costs, thereby threatening its replacement, but also that a firm whose guards are too timid will face threats of replacement by other firms with records that demonstrate their ability to do both a safer and a more effective job.”). The Court has also rejected qualified immunity for local governments.12×12. Owen, 445 U.S. at 657. To be sure, local governments do receive tremendous insulation from liability for constitutional violations.13×13. See generally Fred Smith, Local Sovereign Immunity, 116 Colum. L. Rev. 409 (2016). Among other things, negligence by high-ranking officials does not trigger liability, and local governments are not liable for the acts of their employees under the theory of respondeat superior liability. Id. at 414–16, 431–33. Still, when a municipal ordinance violates the Federal Constitution, local governments may not rely on qualified immunity as an additional layer of protection from liability.

It is one thing to say that defendants such as private prison guards and local governments are liable for violating the Constitution, even if the precise contours of that right were not clearly established at the time of the violation. It is quite another, however, to conclude that these types of defendants are liable for following state law in the wake of new rules of constitutional law. Consider the context of same-sex marriage, for example. In the 1972 case of Baker v. Nelson,14×14. 409 U.S. 810 (1972) (mem.), overruled by Obergefell v. Hodges, 135 S. Ct. 2584 (2015). the Supreme Court summarily dismissed an appeal from a state court ruling that rejected a gay couple’s contention that they had a constitutionally protected right to marry.15×15. See id. at 810. It was not until 2015, in Obergefell v. Hodges,16×16. 135 S. Ct. 2584. that the Supreme Court reached the opposite conclusion, holding that same-sex couples must be permitted to marry.17×17. Id. at 2608. And at that time, many states still had laws on the books that prohibited same-sex marriage. Under Baude and Volokh’s view of the law, thousands of same-sex couples could potentially sue local governments for retroactive damages suffered during the years that they could not get married.

Another context that helps make plain the consequences of retroactive damages for new constitutional rules is the Supreme Court’s Eighth Amendment jurisprudence, which depends heavily on “evolving standards of decency.”18×18. Estelle v. Gamble, 429 U.S. 97, 106 (1976). In Miller v. Alabama,19×19. 567 U.S. 460 (2012). the Supreme Court held for the first time that it violates the Eighth Amendment to impose on juveniles mandatory sentences of life without the possibility of parole.20×20. Id. at 465. Is it the law that those who received such sentences could seek reparations from local governments where local district attorneys enforced life without parole sentences, even if those sentences were entirely consistent with state law and federal precedent at the time? Is it the law that private prison guards who imprisoned these juvenile offenders must pay thousands of dollars in reparations as well? Or suppose the Supreme Court were to rule at some future point that the death penalty is unconstitutional. Could reparations be sought from municipalities where local district attorneys successfully sought the death penalty, a practice then consistent with most states’ laws21×21. Robert J. Smith & Zoë Robinson, Constitutional Liberty and the Progression of Punishment, 102 Cornell L. Rev. 413, 476 (2017) (noting that “thirty-one states and the federal government authorize” the death penalty). and extant Supreme Court precedent?22×22. See Gregg v. Georgia, 428 U.S. 153, 207 (1976) (opinion of Stewart, Powell, and Stevens, JJ.).

These concerns are even more salient at this current moment in the life of the doctrine of qualified immunity. Indeed, Baude has written a highly persuasive and influential account of why the doctrine of qualified immunity, as it exists today, likely cannot be reconciled with § 1983’s text and history.23×23. See William Baude, Is Qualified Immunity Unlawful?, 106 Calif. L. Rev. 45, 47 (2018). Jurists and scholars alike are wrestling with that claim, and a cloud of doubt hangs over the doctrine.24×24. See Ziglar v. Abbasi, 137 S. Ct. 1843, 1870 (2017) (Thomas, J., concurring in part and concurring in the judgment) (noting his “growing concern with [the Court’s] qualified immunity jurisprudence”). If qualified immunity were found to be unlawful, the consequences of retroactive damages for new constitutional rights would be all the more explosive, as the class of defendants who could be sued for such damages would grow exponentially. Police officers, municipal clerks, government-employed prison officials, and others would all be on the hook to pay costly damage awards out of pocket for engaging in perfectly lawful conduct, simply because they failed to predict that the Supreme Court would (at some unknown future point) change the law.

An illustration may help drive home the point. Until the Supreme Court’s holding in Carpenter v. United States25×25. 138 S. Ct. 2206 (2018). that the collection of cell site location data from private cell phone providers constitutes a search under the Fourth Amendment,26×26. Id. at 2223. thousands of police officers investigated criminal suspects by taking the commonplace step of collecting and using this data without a warrant. Prevailing law in as many as thirty-three states allowed this investigative practice, so officers in these states had little reason to fear a lawsuit.27×27. See Charles Blain, Police Could Get Your Location Data Without a Warrant. That Has to End, Wired (Feb. 2, 2017, 7:00 AM), https://www.wired.com/2017/02/police-get-location-data-without-warrant-end/ [https://perma.cc/XKG7-Z3JK]. Nonetheless, if the Court pares back qualified immunity and Baude and Volokh are correct about retroactive liability, all of these officers would find themselves personally liable for thousands of dollars in damages. On Baude and Volokh’s view, the fact that the officers were merely following existing state law — and the fact that they obviously could not predict the Supreme Court’s later decision to overturn that law — would be no defense.

To hold that damages are available under these circumstances would, then, have consequences for local governments, private entities, and potentially enormous numbers of government officials all acting under the color of state law. Such a holding would also raise important questions about notice and fairness. But that is not all. The availability of such damages might also have consequences for future plaintiffs who are seeking to vindicate constitutional rights. Scholars like Professors Richard Fallon and John Jeffries have discussed that the creation of overly broad constitutional remedies can chill the development of constitutional rights, as judges attempt to calibrate the effect of their decisions on competing constitutional values.28×28. See Richard H. Fallon, Jr., The Linkage Between Justiciability and Remedies — And Their Connections to Substantive Rights, 92 Va. L. Rev. 633, 684–85 (2006); John C. Jeffries, Jr., The Right-Remedy Gap in Constitutional Law, 109 Yale L.J. 87 (1999). Permitting retroactive damages for altogether new rules of constitutional law is a blunt tool, and it is conceivable that if this tool lurked behind new constitutional rights, fewer such rights would be identified. It is easy to imagine some Justices getting cold feet in Carpenter, for example, if recognizing a reasonable expectation of privacy in cell phone location data necessitated the imposition of personal liability on tens of thousands of police officers for conduct that was entirely lawful when undertaken.

II. Why the Union-Refund Suits Should Fail

We suspect persons of all political orientations would be troubled if private and government actors alike could be sued personally for damages because they followed prevailing law at Time One without predicting that the Supreme Court would later change the law at Time Two. Fortunately, the law does not require that outcome. In the context of the union-refund suits in particular, three legal principles operate to cut off the massive refund liability sought by right-to-work groups: the Supreme Court’s civil retroactivity doctrine, context-specific defenses that are available to bar § 1983 liability, and Rule 23 class action requirements.

A. Civil Retroactivity

The foundational premise on which the union-refund lawsuits rest is that “under standard civil retroactivity doctrine, Supreme Court decisions supposedly state the true law as it has always been, rather than changing the law.”29×29. Baude & Volokh, supra note 6, at 201. Baude and Volokh argue that the upshot of this rule is that “courts must treat the involuntary collection of agency fees before Janus as unconstitutional.”30×30. Id. However, a close examination of the rationale underlying civil retroactivity doctrine — and a close reading of Janus itself — reveals a strong argument that Janus’s rule should not be applied to pre-Janus conduct to begin with.

As Baude and Volokh correctly observe,31×31. Id. the Supreme Court has established the general presumption that “[w]hen this Court applies a rule of federal law to the parties before it, that rule is the controlling interpretation of federal law and must be given full retroactive effect.”32×32. Harper v. Va. Dep’t of Taxation, 509 U.S. 86, 97 (1993). The reason for this rule is a basic concern for treating like persons alike: if the Court were to give the litigants in the case before it the retroactive benefit of a new rule, then fairness requires giving the same benefit to all similarly situated persons.33×33. See id. (“[T]he Court has no more constitutional authority in civil cases than in criminal cases to . . . treat similarly situated litigants differently.”) (quoting Am. Trucking Ass’ns v. Smith, 496 U.S. 167, 214 (1990) (Stevens, J., dissenting)). To use the union fee setting as an example, it would seem unfair for the Court to “selective[ly] appl[y]”34×34. Id. (quoting Griffith v. Kentucky, 479 U.S. 314, 323 (1987)). Janus’s new rule so as to award Mark Janus himself a refund for fees he paid before June 27, 2018, only to deny the same relief to other similarly situated objecting workers on the ground that the law does not apply retroactively as to them.

All of this is easy enough to understand. But the Court’s endorsement of full civil retroactivity is only a presumption: as Harper v. Virginia Department of Taxation35×35. 509 U.S. 86 (1993). itself held, the critical condition precedent to retroactivity is that the Court must first apply the new rule of federal law “to the parties before it.”36×36. Id. at 97–98 (quoting James B. Beam Distilling Co. v. Georgia, 501 U.S. 529, 539 (1991) (opinion of Souter, J.)). A leading treatise accordingly explains that Harper “did not hold that all decisions of federal law must necessarily be applied retroactively” and that “the Court has not renounced the power to make its decisions entirely prospective, so that they do not apply even to the parties before it.”37×37. 1 Laurence H. Tribe, American Constitutional Law § 3-3, at 226 (3d ed. 2000). Justice Kennedy made a similar observation in a concurring opinion just two years after Harper.38×38. Reynoldsville Casket Co. v. Hyde, 514 U.S. 749, 761 (1995) (Kennedy, J., concurring in the judgment) (“We do not read today’s opinion to surrender in advance our authority to decide that in some exceptional cases, courts may shape relief in light of disruption of important reliance interests or the unfairness caused by unexpected judicial decisions.”).

The possibility of a prospective-only application of Janus — which would stop the refund class actions in their tracks — is one of the most surprising aspects of Justice Alito’s majority opinion in the case. For despite having a chance to make the new rule fully retroactive simply by holding that Janus himself should get a refund for fees he previously paid, the majority steered clear from awarding any backward-looking remedy.39×39. True, the original complaint in the case apparently only sought prospective declaratory relief. See Janus v. AFSCME, 138 S. Ct. 2448, 2462 (2018) (“The Governor commenced an action in federal court, asking that the law be declared unconstitutional. . . .”). But the operative complaint by the time the case made it before the Supreme Court was filed by interveners — including Mark Janus himself. Id. (“[T]he case proceeded on the basis of [Janus’s] new complaint.”). And that complaint explicitly requested “compensatory damages from AFSCME . . . for all compulsory fees seized from [Janus] . . . from the beginning of the applicable statute of limitations. . . .”). Plaintiffs’ Second Amended Complaint at 16, Rauner v. AFSCME, No. 15-CV-01235, 2015 WL 2385698 (N.D. Ill. July 21, 2016), aff’d sub nom. Janus v. AFSCME, 851 F.3d 746 (7th Cir. 2017), rev’d, 138 S. Ct. 2448 (2018). The Supreme Court’s refusal to award backward-looking relief in the face of a clear request for it is thus one reason to hold that Janus has no retroactive application to begin with.

If anything, the opinion is most fairly read to suggest that the new rule forbidding collection of agency fees applies only in the future. Immediately after observing that “[i]t is hard to estimate how many billions of dollars have been taken from nonmembers and transferred to public-sector unions in violation of the First Amendment,” the Court did not say anything about the union’s obligation to pay back those fees.40×40. Janus, 138 S. Ct. at 2486. Instead, the Court framed its ruling as a forward-looking one: “Those unconstitutional exactions cannot be allowed to continue indefinitely.”41×41. Id. Then, in the concluding section of the opinion, the majority again framed its holding as a prospective rule instead of a retroactive one: “For these reasons, States and public-sector unions may no longer extract agency fees from nonconsenting employees.”42×42. Id. Given the Supreme Court’s seemingly intentional refusal to announce a retrospective rule in Janus itself, lower courts considering post-Janus refund claims would be well within reason to dismiss them on the ground that Janus established a uniform forward-looking rule of conduct only.

B. The Good Faith (and Other Common Law) Defense(s)

If courts nevertheless decide to apply the new rule announced in Janus retroactively, the unions should still prevail due to the availability of certain defenses, most notably their good faith reliance on existing state and constitutional law. To their credit, Baude and Volokh acknowledge this possibility, admitting that two lower courts relied on the good faith defense to dismiss retroactive union-refund claims brought in the aftermath of Harris v. Quinn,43×43. 134 S. Ct. 2618 (2014). a predecessor case to Janus involving home health care workers.44×44. See id. at 2623–25; Baude & Volokh, supra note 6, at 203 (first citing Jarvis v. Cuomo, 660 F. App’x 72, 75–76 (2d Cir. 2016); then citing Winner v. Rauner, No. 15 CV 7213, 2016 WL 7374258, at *5–6 (N.D. Ill. Dec. 20, 2016), appeal dismissed (7th Cir. Aug. 18, 2017)). Baude and Volokh are skeptical of these rulings, however, on the ground that they lack “clear authority” or “endorse[ment] [from] the Supreme Court.”45×45. Baude & Volokh, supra note 6, at 203.

It is true that the Supreme Court has never formally recognized a good faith defense that private defendants may raise against § 1983 actions. However, neither has the Court rejected such a defense; it has instead reserved the question three times.46×46. Richardson v. McKnight, 521 U.S. 399, 413–14 (1997) (leaving for “another day” the question “whether or not the private defendants . . . might assert, not immunity, but a special ‘good-faith’ defense,” id. at 413); Wyatt v. Cole, 504 U.S. 158, 169 (1992) (“[W]e do not foreclose the possibility that private defendants faced with § 1983 liability . . . could be entitled to an affirmative defense based on good faith. . . .”); Lugar v. Edmondson Oil Co., 457 U.S. 922, 942 n.23 (1982) (“We need not reach the question of the availability of [a good faith] defense to private individuals at this juncture.”). Moreover, five Justices expressed support for the private-party good faith defense in Wyatt v. Cole47×47. 504 U.S. 158. without any disagreement from the majority,48×48. See id. at 174 (Kennedy, J., concurring) (“[T]here is support in the common law for the proposition that a private individual’s reliance on a statute, prior to a judicial determination of unconstitutionality, is considered reasonable as a matter of law. . . .” (citing Birdsall v. Smith, 122 N.W. 626, 627 (Mich. 1909))); id. at 177 (Rehnquist, C.J., dissenting) (agreeing “that a good-faith defense will be available for respondents to assert on remand”); id. at 169 (majority opinion) (leaving open the possibility that private defendants “could be entitled to an affirmative defense based on good faith”). and every court of appeals to consider the question has concluded that it exists.49×49. See Clement v. City of Glendale, 518 F.3d 1090, 1097 (9th Cir. 2008); Pinsky v. Duncan, 79 F.3d 306, 311–12 (2d Cir. 1996); Vector Research, Inc. v. Howard & Howard Attorneys P.C., 76 F.3d 692, 699 (6th Cir. 1996); Jordan v. Fox, Rothschild, O’Brien & Frankel, 20 F.3d 1250, 1276–77 (3d Cir. 1994); Wyatt v. Cole, 994 F.2d 1113, 1118 (5th Cir. 1993).

Baude and Volokh suggest that the lower court decisions recognizing the good faith defense have nonetheless erred by fashioning the defense as a bar to liability that is categorically available to all private defendants in every § 1983 claim,50×50. See, e.g., Winner v. Rauner, No. 15 CV 7213, 2016 WL 7374258, at *5 (N.D. Ill. Dec. 20, 2016), appeal dismissed (7th Cir. Aug. 18, 2017) (agreeing with “[e]very federal appellate court that has considered the good-faith defense [and that] has found that it exists for private parties”); Jarvis v. Cuomo, 660 F. App’x 72, 75 (2d Cir. 2016) (rejecting argument that good faith defense is inapplicable to First Amendment claims because the defense “need not relate to or rebut specific elements of an underlying claim”). rather than a defense whose availability depends on the particular nature of the underlying constitutional violation.51×51. Baude & Volokh, supra note 6, at 203 (asserting that unions “should not be too confident that they will have [a good faith] defense against Janus suits” even if “the Court turns to private law analogues [to find] such a defense”). As Baude has written elsewhere, the Supreme Court’s recognition of § 1983 defenses originally turned on a comparison to “the individual elements of particular common-law torts,” not the broader notion that a defense available in one tort setting could serve as “evidence of a more general [defense] . . . export[able] to other claims.”52×52. Baude, supra note 23, at 59. Note that there is some dispute over whether good faith is more properly viewed as an element of the claim that the plaintiff must show, or whether it is an affirmative defense that the defendant must raise. See Wyatt, 504 U.S. at 176 at n.1 (Rehnquist, C.J., dissenting) (“Describing the common law as providing a ‘defense’ is something of a misnomer — under the common law it was plaintiff’s burden to establish as elements of the tort both that the defendant acted with malice and without probable cause. Referring to the defendant as having a good-faith defense is a useful shorthand for capturing plaintiff’s burden and the related notion that a defendant could avoid liability by establishing either a lack of malice or the presence of probable cause.” (citation omitted)). We do not take that distinction to be meaningful here, however, given that there is no allegation that public sector unions collected fair share fees for a bad faith motive — that is, a motive other than to provide collective bargaining-related services to agency-fee payers and full members. Or as the Court framed the inquiry in Wyatt: if the defendants were “shielded from tort liability when Congress enacted [§ 1983,] we infer from legislative silence that Congress did not intend to abrogate” the defenses that would have been available to them against “the most closely analogous torts.”53×53. 504 U.S. at 164. Technically, Wyatt looked to analogous torts for the purpose of determining if qualified immunity would be supported. Id. But the Court recognized that the same analogy could support a good faith defense if immunity were not available. Id. at 165, 169; see also, e.g., Pierson v. Ray, 386 U.S. 547, 556–57 (1967) (“[Section] 1983 ‘should be read against the background of tort liability,’” and “[p]art of the background of tort liability, in the case of police officers making an arrest, is the defense of good faith.” (quoting Monroe v. Pape, 365 U.S. 167, 187 (1961))). Note that recognizing the availability of a good faith defense in the union-refund context, because such a defense would have been available against the most analogous tort at common law, could lead in theory to similar defenses in the police officer, municipal clerk, and prison official examples listed above. The answer would depend on a close analysis of the most analogous tort in each setting. Alternatively, Congress might wish to enact a more generally applicable good faith defense to all § 1983 claims. We do not take a position on the desirability of such legislation in this Response.

Still, in light of the uniform consensus in the courts of appeals and the support of at least five Justices in Wyatt, district courts may find themselves bound to rule that the unions enjoy a general good faith defense without regard to the most closely analogous tort. Nor would the outcome change even if a court were to conclude otherwise,54×54. One might conclude as much, for example, if one agrees that the entire doctrine of qualified immunity as it currently exists is itself unlawful, such that all § 1983 defendants — whether private or governmental — must identify a defense available to the most closely analogous common law tort. See Baude, supra note 23. because public sector unions would have enjoyed a good faith defense against the most analogous tort at common law. In fact, a close examination of common law tort analogues reveals that unions should enjoy an additional defense that independently precludes liability: the fact that they did not act maliciously in the course of relying on state agency fee laws.

The critical first step in this analysis is to identify the common law tort that is most analogous to the constitutional violation alleged by the objecting workers who seek a refund. That violation, of course, is the one announced in Janus, so it makes sense to look to Janus’s own characterization of the unlawful conduct as a point of reference when searching for the analogous tort. And once one does so, it is apparent that the defining feature of this particular constitutional violation is that a private party (the union) has utilized a state law “procedure [that] violates the First Amendment” — that is, the private party has used a state-created legal process to “automatically deduct[]” a portion of each “nonmember’s wages” and transfer it to the union.55×55. Janus v. AFSCME, 138 S. Ct. 2448, 2486 (2018); see also id. at 2461 (describing the process under Illinois law by which “a union categorizes its expenditures as chargeable or nonchargeable and thus determines a nonmember’s ‘proportionate share,’” which is audited and “certified to the employer,” which “automatically deducts that amount from the nonmembers’ wages”).

Indeed, it is this very fact — the unions’ reliance on a state statute authorizing collection of agency fees — that enables objecting union workers to sue under § 1983 at all. Baude and Volokh put it best: “[t]he state statutes authorizing the collection of agency fees are” what amount to the “unconstitutional state action” that is challengeable under § 1983.56×56. Baude & Volokh, supra note 6, at 201. It is thus critical to treat these statutes as the focal point of a § 1983 inquiry into the most analogous tort, for if the statutes didn’t exist, no § 1983 claim would either.

With the role of the state’s statutorily prescribed procedure for obtaining objecting workers’ wages front and center, it becomes clear that the most analogous tort against the unions is the common law tort of abuse of process, which is a “cause[] of action against private defendants for unjustified harm arising out of the misuse of governmental processes.”57×57. Wyatt, 504 U.S. at 164. In fact, there is a close parallel between the constitutional tort at issue in the Janus refund claims and the claim in Wyatt, which also relied on the abuse of process tort analogy.58×58. See id. (recognizing “abuse of process” as “the most closely analogous tort[] . . . in this case”). In Wyatt, the plaintiff filed a § 1983 claim against a private-party defendant for utilizing a state law replevin procedure that resulted in the state wrongfully transferring the plaintiff’s property to the defendant. Just so, here: the plaintiffs are objecting workers who seek relief against a private party defendant (their union) because the union utilized a state law procedure to wrongfully transfer the workers’ property. In both situations, the party seeking to obtain property belonging to another was automatically entitled to such property as a matter of law upon the completion of certain basic administrative requirements.59×59. Compare id. at 160 (explaining how an individual acting under Mississippi’s replevin statute could “obtain a court order for seizure of property possessed by another” by swearing that “the applicant was entitled to that property,” and how judges had “no discretion to deny a writ of replevin”), with Cal. Gov’t Code § 3546(a) (West Supp. 2017) (explaining how public sector unions were automatically entitled to agency fee deductions under state law upon meeting certain requirements, such as informing a public employer that they were the certified exclusive representative of a bargaining unit).

To be sure, abuse of process tort claims typically involve allegations of misuse of judicial procedures (as in Wyatt) rather than statutory procedures. But that distinction is of no consequence for purposes of identifying the proper analogy. The canonical definition of the tort at common law entails the “unjustifiable employment of the processes of the law.”60×60. Joel Prentiss Bishop, Commentaries on the Non-Contract Law § 220, at 88 (Chicago, T.H. Flood & Co. 1889) (emphasis added); see also Francis M. Burdick, The Law of Torts § 4, at 298 (3d ed. 1913) (“It is well settled that an action lies for the malicious abuse of lawful process. . . .”). A statute creating a process by which party X acquires the legal right to party Y’s property constitutes a “process of law” whether party X is required to obtain court approval for the property transfer or whether party X is entitled to the transfer so long as it follows administrative procedures set forth in the statute.61×61. See, e.g., Cal. Gov’t Code § 3546(a) (West Supp. 2017) (“[U]pon receiving notice from the exclusive representative of a public school employee who is in a unit for which an exclusive representative has been selected pursuant to this chapter, the [government] employer shall deduct the amount of the fair share service fee authorized by this section from the wages and salary of the employee and pay that amount to the employee organization.”). Indeed, many cases explicitly recognize the availability of an abuse of process claim arising out of misuse of administrative (as opposed to judicial) process.62×62. See, e.g., Melvin v. Pence, 130 F.2d 423, 426–27 (D.C. Cir. 1942); cf. Restatement (Second) of Torts § 680 (Am. Law Inst. 1965) (recognizing abuse of process claims in administrative proceedings).

Baude and Volokh argue that the most analogous tort may sound in restitution or unjust enrichment rather than abuse of process.63×63. See Baude & Volokh, supra note 6, at 203 (“[I]f the Court turns to private law analogues for [the good faith] defense, it might find that restitution and unjust enrichment provide the better analogue.”). They analogize the unions’ receipt of agency fees to a situation where a government illegally assesses a tax on a citizen; in that circumstance, the taxpayer is often entitled to restitution of her tax payment.64×64. See id. Baude and Volokh also mention restitution of “payments under a judicial order that has since been reversed.” Id. But of course what is different about that example is the lack of a final judgment creating a legal expectation that a party may rely on. In the context of the public sector unions, Abood was not some lower court decision awaiting appeal; it was a binding Supreme Court decision for more than forty years. But that unremarkable rule seems to miss the most salient aspect of the objecting workers’ refund claim: that some other private party, not the government, has used a state-sanctioned process to wrongfully obtain their property.

This distinction between private and government beneficiaries is not a mere formality. One can see the equity in eliminating the government’s ability to claim good faith reliance on an illegal tax law of its own creation as an excuse to keep unlawfully withheld tax receipts. The Restatement explains, after all, that “one of the cornerstones of the law of restitution and unjust enrichment” is that a defendant “is not permitted to profit by his own wrong.”65×65. Restatement (Third) of Restitution and Unjust Enrichment § 3 (Am. Law Inst. 2011). But that rationale is not implicated when a private party relies on a law that was passed by the people’s chosen representatives (who were in turn acting in reliance on an on-point Supreme Court decision).

Alternatively, objecting workers may argue that the more analogous tort is conversion. This, too, is an uncomfortable fit. To start, there is disagreement over whether money is even a proper subject of a conversion action if it is not specifically earmarked or differentiated.66×66. See H.D. Warren, Annotation, Nature of Property or Rights Other Than Tangible Chattels Which May Be Subject of Conversion, 44 A.L.R. 2d 927, § 7[a] (1955) (collecting conflicting sources on whether an action for conversion will lie if money is not specifically earmarked). More glaringly, the tort of conversion fails to include as an element the crucial feature of the First Amendment violation identified in Janus (the very feature that enables objecting workers to sue in light of § 1983’s state action requirement): the existence of a state law process to transfer the plaintiff’s property to the defendant. In fact, conversion is often defined by reference to unlawful, wrongful, or unauthorized conduct67×67. See, e.g., France v. Gibson, 101 S.W. 536, 536 (Tex. Civ. App. 1907) (affirming the trial court’s definition of conversion as “[t]he unlawful and wrongful exercise of dominion, ownership, or control by one person over the property of another”); Hill v. Campbell Comm’n Co., 74 N.W. 388, 389 (Neb. 1898) (“[C]onversion is any unauthorized act which deprives the owner of his property permanently or for an indefinite time.”).  — an element that would seem lacking by definition where the defendants receive property only in reliance on state agency fee laws.68×68. Unsurprisingly, then, there is authority for the application of a good faith defense to conversion claims, too. See Restatement (Second) of Torts § 222A(2)(c) (Am. Law Inst. 1965) (recognizing that one factor in “determining the seriousness of the interference and the justice of requiring the actor to pay the full value” is the defendant’s “good faith”); Jordan v. Wilhelm, 770 P.2d 74, 76 (Or. Ct. App. 1989) (upholding trial court ruling that defendants did not commit conversion when they seized plaintiff’s property “in good faith and believed that they were acting within their ‘legal rights’”).

If abuse of process is indeed the best tort law analog, then the second step in the analysis is to determine whether the defenses available to an abuse of process action would benefit the unions here. At common law, an abuse of process claim would fail if the plaintiff could not show both malice and the lack of probable cause.69×69. Bishop, supra note 60, § 225, at 90 (“There must be both malice and the want of probable cause combining.”); Thomas M. Cooley, A Treatise on the Law of Torts 220 n.5 (2d ed., Chicago, Callaghan & Co. 1888) (“To put into force the process of the law maliciously, and without any reasonable or probable cause, is wrongful . . . .” (quoting Churchill v. Siggers, 3 El. & Bl. 929, 937 (1854))). So unions can defeat liability if either element is not satisfied.

We start with the probable cause element because that has been the subject of lower court rulings thus far.70×70. See cases cited supra notes 44, 50. At common law, probable cause was present where the “combination of facts and proofs [would] fairly lead the reasonable mind to the belief” that the defendant’s action was proper.71×71. Bishop, supra note 60, § 239, at 95. This meant that liability was foreclosed when a defendant acted with a good faith belief in the lawfulness of a statute.72×72. See Birdsall v. Smith, 122 N.W. 626, 627 (Mich. 1909) (“Every statute should be considered valid until there is a judicial determination to the contrary, and these defendants had a right to act upon such assumption. The question whether they had probable cause for making this complaint cannot depend on what may be held as to the validity of the law.”). And in the agency fee context, a reasonable person had every reason to believe that the collection of such fees was lawful before Janus, given that state law authorized it and the Supreme Court had upheld the practice decades earlier in Abood v. Detroit Board of Education.73×73. 431 U.S. 209, 222 (1977).

Baude and Volokh respond that perhaps this reliance was unfounded. Because a pair of earlier Supreme Court opinions questioned the continued vitality of state agency fee laws, they argue, “the courts may well conclude that unions were knowingly gambling on the continued validity of Abood.”74×74. Baude & Volokh, supra note 6, at 204. But expecting the unions to predict when, exactly, they could no longer rely in good faith on Abood’s settled rule is a dangerous exercise. True, the Court referred to Abood as an “anomaly” in its 2012 decision in Knox v. SEIU, Local 1000,75×75. 567 U.S. 298, 311 (2012). but it also said in that case that Abood’s permission of agency fees was “justified by the interest in furthering ‘labor peace’” — hardly the kind of language that would inform private actors that the case was no longer worthy of reliance.76×76. Id. (quoting Chi. Teachers Union v. Hudson, Local No. 1, 475 U.S. 292, 303 (1986)). The Court then had the opportunity to overrule Abood in its 2014 decision in Harris v. Quinn, but chose not to.77×77. 134 S. Ct. 2618 (2014). Does that mean continuing to collect agency fees was in good faith or bad faith? And does it matter that for a period of several months after Justice Scalia’s death in February 2016, public sector unions might have reasonably anticipated a friendly fifth vote to uphold Abood in the form of Judge Garland? Asking private actors to engage in all of this guesswork before following the existing law is surely a worse approach than a bright-line rule that says it is always good faith to rely on a law that is explicitly blessed by a controlling Supreme Court decision — unless and until that decision is overturned.78×78. See opinions cited supra note 48.

In any case, even if one accepted the argument that unions somehow acted in bad faith when they followed decades-old state law and Supreme Court precedent, that would prove at most the lack-of-probable-cause element of the abuse of process tort. But once one acknowledges that § 1983 defendants should be entitled to raise the defenses that were available as against the most analogous common law tort,79×79. See supra notes 52–53 and accompanying text; see also Will Baude, More on Suits Against Unions for Janus Violations: A Response to Aaron Tang, Volokh Conspiracy: Reason (July 26, 2018, 8:03 AM), https://reason.com/volokh/2018/07/26/more-on-suits-against-unions-for-janus-v [https://perma.cc/3EA7-JXN3] (recognizing the possibility of defenses from the most analogous common law tort). it becomes clear that the union-refund plaintiffs must also show that the unions acted with malice in the course of collecting agency fees.80×80. See supra note 69 and accompanying text (defining elements of abuse of process tort); see also Wyatt v. Cole, 504 U.S. 158, 172 (1992) (Kennedy, J., concurring) (explaining how in the “common-law action[]” of abuse of process, “it was essential for the plaintiff to prove that the wrongdoer acted with malice and without probable cause”). After all, prior to § 1983’s enactment, a plaintiff seeking to sue a private defendant for wrongfully using a legal process to obtain the plaintiff’s property would lose unless the defendant acted with malice, a concept then understood to mean acting with an “evil or unlawful purpose.”81×81. Bishop, supra note 60, § 232, at 92; see also Restatement (Second) of Torts § 682 cmt. a (Am. Law Inst. 1965) (“The gravamen of the misconduct for which the liability [for abuse of process] is imposed . . . is the misuse of process, no matter how properly obtained, for any purpose other than that which it was designed to accomplish.”); 2 C.G. Addison & H.G. Wood, A Treatise on the Law of Torts 69 (New York, James Cockcroft & Co. 1876) (“Proof of the absence of belief in the truth of the charge by the person making it . . . is almost always involved in the proof of malice.”). Like the lack-of-probable-cause requirement, this malice element was so “well established” that “Congress would have specifically so provided had it wished to abolish the doctrine.”82×82. Pierson v. Ray, 386 U.S. 547, 554–55 (1967) (applying the same theory to judicial immunity).

Yet there is no allegation in the union-refund suits that the public sector union defendants used the state agency fee procedure for some evil ulterior purpose — for example, to punish or coerce nonmembers based on their views, as opposed to collecting fees for the purpose of bargaining on behalf of all workers (nonmembers included83×83. See Vaca v. Sipes, 386 U.S. 171, 181–82 (1967) (describing duty of fair representation under the National Labor Relations Act, which governs private sector labor relations, but which has been recognized in state public sector labor law as well). ) and representing them in grievance procedures against the employer.84×84. The malice requirement serves important objectives on first principles. For when a private party is sued for the simple act of following a statutory procedure that has been duly enacted by elected officials, something more than the mere act alone must be shown to trigger liability, lest countless private citizens be punished for ordinary, law-abiding conduct. The malice element thus sensibly requires something more — an action with a subjectively-held, wrongful ulterior purpose. See supra note 81 and accompanying text. Of course, once a law is struck down, any subsequent conduct relying on it would no longer be lawful, which is why no one disputes that public sector unions would have to refund any agency fees collected after Janus was decided. But the pre-Janus collection of fees would have been actionable at common law only if the plaintiff could establish the parallel elements of the abuse of process tort — malice and lack of probable cause. Courts should accordingly dismiss the refund actions for this independent reason, separate and apart from any good faith defense.

C. Rule 23 Class Action Requirements

Section 1983 suits seeking massive classwide refund liability should fail for another reason: they do not satisfy the commonality and predominance requirements for certification as a class action under Rule 23 of the Federal Rules of Civil Procedure.85×85. Fed. R. Civ. P. 23(a)(2), (b)(3). The arguments that follow take the Supreme Court’s Rule 23 class action precedents as a given; this Response takes no position on the propriety or desirability of the Court’s various rulings in this area.

Recall the Supreme Court’s explanation that defendants in § 1983 claims should be able to raise the same defenses as would have been available under the analogous tort at common law.86×86. See supra note 53 and accompanying text. We have argued that the proper analogy is to the abuse of process tort, to which the good faith defense (and a lack of malice defense) was widely recognized at common law. However, should courts agree with Baude and Volokh’s view that these § 1983 claims are not sufficiently analogous to tort claims that invite such a defense, unions can still invoke the class action certification requirements of Federal Rule of Civil Procedure 23 to limit retroactive liability.

Suppose, for example, that a court were to conclude that unjust enrichment is the proper tort law analogue and that a good faith defense is not available as a result. Even then, it would remain the case that other unjust enrichment–specific defenses would be available. To that point, the very first defense that the Restatement identifies as against an unjust enrichment claim is that the recipient of the payment at issue was not unjustly enriched.87×87. Restatement (Third) of Restitution & Unjust Enrichment § 62 (Am. Law Inst. 2011). As the Restatement explains, “the recipient may defend by showing that some or all of the benefit conferred did not unjustly enrich the recipient when the challenged transaction is viewed in the context of the parties’ further obligations to each other.”88×88. Id.

Public sector unions will have strong arguments on the merits that they were not unjustly enriched by objecting workers’ agency fee payments given two “further obligations” that the unions owe to those workers in exchange for those payments: the duty to bargain on their behalf and the duty to represent them in grievance procedures. The former duty results in what is on average an 11% wage premium for unionized workers, an amount that far exceeds (and thus renders “just”) the typical 2% agency fee.89×89. Jill Manzo et al., The State of the Unions 2016: A Profile of Unionization in the Twin Cities, in Minnesota, and in America 14–15 (2016) (estimating an 11.1% union wage premium in 2015). In addition, the latter duty can be viewed as a kind of litigation insurance that obligates the union to defend each objecting worker against any employer action that violates the worker’s rights under the collective bargaining agreement.

To overcome this defense, an objecting worker would need to show both that she would have earned at least the same net wage premium in the absence of a union funded by agency fees and that she did not benefit from the union’s duty to represent her in employer grievances. Those propositions, however, are highly fact bound with respect to each objecting worker. While some objecting workers may have been high performers who could have individually negotiated an equivalent wage premium, others might have earned less in the absence of an adequately funded union to bargain collectively on their behalf.90×90. It is not enough for the objecting workers to argue that the union would have provided the identical wage premium anyway because of its duty to fairly represent all workers. See Vaca v. Sipes, 386 U.S. 171 (1967) (describing the duty unions have under labor law to treat members and nonmembers fairly). For one thing, once objecting workers are free to opt out of agency fees, the union’s power to negotiate for better wages is invariably compromised. And just as significantly, if the objecting workers had not been required to pay agency fees in the past, state law may well have been amended to eliminate the union’s duty to bargain on behalf of nonmembers to begin with. See, e.g., Dave Umhoefer & Sarah Hauer, From Teacher “Free Agency” to Merit Pay, the Uproar Over Act 10 Turns Into Upheaval in Wisconsin Schools, Milwaukee J. Sentinel (Oct. 9, 2016), https://projects.jsonline.com/news/2016/10/9/from-teacher-free-agency-to-merit-pay-the-uproar-over-act-10.html [https://perma.cc/BT72-JF74] (finding that a majority of teachers experienced reduced or stagnant wages after Wisconsin’s transition to right to work, while a small group of high-performing teachers received salary increases). Likewise, some objecting workers may have received valuable representation directly from the union in disputes with their employer; others may have benefitted indirectly from the union’s representation of a similarly situated worker.

Rule 23’s commonality and predominance requirements exist precisely to keep these kinds of fact-dependent, case-by-case inquiries out of class action litigation. As to commonality, the Supreme Court has explained that “[w]hat matters . . . is not the raising of common ‘questions’ — even in droves — but, rather, the capacity of a class-wide proceeding to generate common answers apt to drive the resolution of the litigation.”91×91. Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 350 (2011) (quoting Richard A. Nagareda, Class Certification in the Age of Aggregate Proof, 84 N.Y.U. L. Rev. 97, 132 (2009)). Assuming that the unjust enrichment analogue is correct, litigating objecting workers’ refund allegations as a class will do nothing to answer the crucial liability question of whether a union actually provided offsetting value to any individual worker via a union-negotiated wage premium or grievance representation. The Court’s observation in Wal-Mart Stores, Inc. v. Dukes92×92. 564 U.S. 338. thus seems apt here, too: “Dissimilarities within the proposed class are what have the potential to impede the generation of common answers.”93×93. Id. at 350 (quoting Nagareda, supra note 91, at 132).

The problem is even clearer with respect to the predominance inquiry.94×94. The objecting workers’ class is certifiable, if at all, under Rule 23(b)(3)’s predominance test because their actions involve monetary claims and do not raise the risk of incompatible standards for the union or unfairness to other objecting workers. See Fed. R. Civ. P. 23(b). The Supreme Court has held that this requirement involves asking whether “the common, aggregation-enabling, issues in the case are more prevalent or important than the non-common, aggregation-defeating, individual issues.”95×95. Tyson Foods, Inc. v. Bouaphakeo, 136 S. Ct. 1036, 1045 (2016) (quoting William B. Rubenstein, 2 Newberg on Class Actions § 4.49, at 195–96 (5th ed. 2012)). By “individual issues,” Rule 23 refers to issues for which “members of a proposed class will need to present evidence that varies from member to member.”96×96. Id. (quoting Rubenstein, supra note 95, § 4.50, at 196–97). That, of course, is precisely what characterizes the crux of unjust enrichment liability, as each member of the proposed class would need to present evidence regarding how much she would have earned individually in the absence of an adequately funded union (including the possibility that she would have had to negotiate her salary individually based on her performance97×97. See, e.g., Matt Murphy, Union Divide Contributed to Janus Bill’s Demise, WBUR News (Aug. 7, 2018), http://www.wbur.org/news/2018/08/07/union-bill-janus-demise [https://perma.cc/CE6Q-N3C4] (noting support from the Teamsters for a bill that would make clear that, after Janus, “contracts negotiated by unions only apply to union members”). ) as well as evidence proving that she did not benefit from any of the representation services provided by the union to her or a similarly situated employee. In the absence of such proof for each and every proposed class member, there is simply no way to know if agency fees collected by the union were “unjust” or not.

Similar problems exist for certifying a § 1983 class action if a court were somehow to deem conversion the best analog. As noted earlier, liability for the tort of conversion is precluded unless the plaintiff can establish that the defendant “wrongfully” obtained the plaintiff’s property, a showing that at a minimum involves proof that the defendant’s interference was “unwarranted.”98×98. Gruber v. Pac. States Sav. & Loan Co., 88 P.2d 137, 139 (Cal. 1939). But even assuming that it can be “wrongful” or “unwarranted” for a union to follow a state law agency fee procedure, how is a court to know if the collection of agency fees from every member of the proposed class of objecting workers was “wrongful” or “unwarranted” without knowing whether the objecting workers actually benefitted from the union’s representation? It seems highly inequitable to hold a defendant liable for converting a plaintiff’s money if the defendant actually provided substantial remunerative benefits (such as a much larger wage increase or legal representation in a dispute with the employer) to the plaintiff in exchange. The point is not to suggest that all objecting workers were better off because of the union, but rather that there is no way to know without precisely the kind of individualized factfinding and evidentiary inquiry that deprive the class action device of its utility.


There is blood in the water, and right-to-work groups are closing in. Classwide demands for the refund of upwards of eight figures’ worth of agency fees collected before Janus was decided could be a deathly blow to organized labor. As we have explained, however, these suits should fail due to the best reading of Janus as itself establishing a prospective-only rule, the unions’ good faith conduct and lack of malice, and the lack of class action commonality and predominance. Such an outcome would be more than legally correct; it would help prevent the sweeping consequences of imposing liability on a host of other defendants — municipal clerks, prison officials, police officers, and others — who simply followed the law on the books. 99×99. Recommended Citation: Aaron Tang & Fred O. Smith Jr., Can Unions Be Sued for Following the Law?, 132 Harv. L. Rev. F. 24 (2018), https://harvardlawreview.org/2018/11/can-unions-be-sued-for-following-the-law/.


* Acting Professor of Law, University of California, Davis, School of Law.

** Associate Professor of Law, Emory University School of Law; Visiting Associate Professor of Law, University of Chicago Law School.

Recommended Citation: Aaron Tang & Fred O. Smith Jr., Can Unions Be Sued for Following the Law?, 132 Harv. L. Rev. F. 24 (2018), https://harvardlawreview.org/2018/11/can-unions-be-sued-for-following-the-law/.