Glacier Northwest v. Teamsters/Opinion of Justice Jackson

From Wikisource
Jump to navigation Jump to search
Glacier Northwest, Inc., doing business as CalPortland, v. International Brotherhood of Teamsters Local Union No. 174
Supreme Court of the United States
4237187Glacier Northwest, Inc., doing business as CalPortland, v. International Brotherhood of Teamsters Local Union No. 174Supreme Court of the United States

SUPREME COURT OF THE UNITED STATES


No. 21–1449


GLACIER NORTHWEST, INC., DBA CALPORTLAND, PETITIONER v. INTERNATIONAL BROTHERHOOD OF TEAMSTERS LOCAL UNION NO. 174
ON WRIT OF CERTIORARI TO THE SUPREME COURT OF WASHINGTON
[June 1, 2023]

Justice Jackson, dissenting.

The right to strike is fundamental to American labor law. Congress enshrined that right in the National Labor Relations Act (NLRA) and simultaneously established the National Labor Relations Board to adjudicate disputes that arise between workers and management. That decision reflected Congress’s judgment that an agency with specialized expertise should develop and enforce national labor law in a uniform manner, through case-by-case adjudication. For its part, this Court has scrupulously guarded the Board’s authority for more than half a century. See San Diego Building Trades Council v. Garmon, 359 U. S. 236 (1959). Under Garmon, and as relevant here, a court presented with a tort suit based on strike conduct generally must pause proceedings and permit the Board to determine in the first instance whether the union’s conduct is lawful if the conduct at issue is even “arguably” protected by the NLRA. Id., at 245.

Today, the Court falters. As the majority acknowledges, the Board’s General Counsel has filed a complaint with the Board after a thorough factual investigation, and that complaint alleges that the NLRA protects the strike conduct at the center of this state-court tort suit. The logical implication of a General Counsel complaint under Garmon is that the union’s conduct is at least arguably protected by the NLRA. Consequently, where (as here) there is a General Counsel complaint pending before the Board, courts—including this Court—should suspend their examination. Garmon makes clear that we have no business delving into this particular labor dispute at this time.

But instead of modestly standing down, the majority eagerly inserts itself into this conflict, proceeding to opine on the propriety of the union’s strike activity based on the facts alleged in the employer’s state-court complaint. As part of this mistaken expedition, the majority tries its own hand at applying the Board’s decisions to a relatively novel scenario that poses difficult line-drawing questions—fact-sensitive issues that Congress plainly intended for the Board to address after an investigation. And in the course of inappropriately weighing in on the merits of those questions at this stage, the majority also misapplies the Board’s cases in a manner that threatens to both impede the Board’s uniform development of labor law and erode the right to strike.

In my view, today’s misguided foray underscores the wisdom of Congress’s decision to create an agency that is uniquely positioned to evaluate the facts and apply the law in cases such as this one. This case is Exhibit A as to why the Board—and not the courts—should ordinarily take the first crack at resolving contentious, fact-bound labor disputes of this nature. Because the majority’s ruling suggests otherwise, I respectfully dissent.

I

The majority’s brief opinion quotes Garmon’s “arguably protected” test and endeavors to apply it. Ante, at 3–4, 6–11. But the opinion devotes relatively little space to the origins and purpose of that longstanding precedent. That omission is telling. A proper understanding of Garmon’s foundation sheds considerable light on the majority’s sequential missteps in this case.

A

Congress’s passage of the NLRA “marked a fundamental change in the Nation’s labor policies.” Sears, Roebuck & Co. v. Carpenters, 436 U. S. 180, 190 (1978). Prior to that point, union activity had been viewed as “a species of ‘conspiracy,’ ” prompting substantial conflict between labor and management. Ibid. With the enactment of the NLRA in 1935, “Congress expressly recognized that collective organization of segments of the labor force into bargaining units capable of exercising economic power comparable to that possessed by employers may produce benefits for the entire economy in the form of higher wages, job security, and improved working conditions.” Ibid.

The heart of the NLRA is §7, which safeguards workers’ rights “to self-organization, to form, join, or assist labor organizations, to bargain collectively through representatives of their own choosing, and to engage in other concerted activities for the purpose of collective bargaining or other mutual aid or protection.” 29 U. S. C. §157. Among the “ ‘concerted activities’ ” that the Act unquestionably protects is “the vital, economic instrumen[t] of the strike.” Garmon, 359 U. S., at 241; see §163.

Section 8 of the NLRA provides a list of “unfair labor practice[s]” that employers and unions are prohibited from engaging in. §158. For example, it is an unfair labor practice for an employer to “interfere with, restrain, or coerce employees in the exercise of” their §7 rights, including the right to strike. §158(a)(1). And it is an unfair labor practice for a union to “refuse to bargain collectively with an employer.” §158(b)(3). Taken together, §7 and §8 establish certain conduct that Congress has deemed protected (§7) and prohibited (§8).

B

Congress could have stopped there. But “Congress did not merely lay down a substantive rule of law to be enforced by any tribunal competent to apply law generally to the parties.” Garner v. Teamsters, 346 U. S. 485, 490 (1953). Rather, Congress “went on to confide primary interpretation and application of its rules to a specific and specially constituted tribunal”: the National Labor Relations Board. Ibid.; see generally §§153–156.

By statutory mandate, the Board is composed of five members who are appointed by the President with the advice and consent of the Senate. §153(a). Congress also provided for an independent General Counsel, who is likewise presidentially appointed and Senate confirmed. §153(d); see NLRB v. Food & Commercial Workers, 484 U. S. 112, 117–118 (1987). The General Counsel conducts investigations into unfair labor practices and brings complaints before the Board through a “particular procedure” that Congress has prescribed “for investigation, complaint and notice, and hearing and decision, including judicial relief pending a final administrative order” from the Board. Garner, 346 U. S., at 490; see §§153, 160.

The Board has fleshed out this process via rulemaking authority that Congress has delegated. §156. If a person believes that an employer or union has committed an unfair labor practice, the person may file a charge with a regional director, who acts on behalf of the General Counsel. 29 CFR §101.2 (2022). The regional director investigates the charge. §101.4. If “the charge appears to have merit and efforts to dispose of it by informal adjustment are unsuccessful,” the regional director issues a complaint on behalf of the General Counsel. §101.8. When a General Counsel’s complaint issues, an administrative law judge (ALJ) holds a hearing and issues a decision, which the Board reviews if any party files an exception. §§101.8–101.12. If the Board finds that a party has engaged in an unfair labor practice, it must order the party to “cease and desist” and to take “such affirmative action … as will effectuate the policies” of the NLRA. 29 U. S. C. §160(c).

C

The history and structure of the NLRA make clear that Congress “entrusted administration of the labor policy for the Nation to a centralized administrative agency”—the Board—“armed with its own procedures, and equipped with its specialized knowledge and cumulative experience.” Garmon, 359 U. S., at 242. Congress thought the Board’s primary role was “necessary to obtain uniform application of [the NLRA’s] substantive rules and to avoid th[e] diversities and conflicts likely to result from a variety of local procedures and attitudes toward labor controversies.” Garner, 346 U. S., at 490. That judgment makes perfect sense. The NLRA’s substantive principles are intrinsically broad and potentially conflicting, leaving much for future articulation through case-by-case adjudication. Drawing the line between activities that constitute a protected strike, on the one hand, and unprotected actions for which employers may validly discipline employees, on the other, is a legally and factually complex task. Moreover, that task implicates important economic policy considerations about the relative bargaining power of labor and management that affect not only the parties to a particular labor dispute but also our broader national economy.

To effect Congress’s intent, this Court has consistently recognized that “courts are not primary tribunals to adjudicate [these] issues.” Garmon, 359 U. S., at 244. Rather, “it is to the Board that Congress entrusted the task of applying the Act’s general … language in the light of the infinite combinations of events which might be charged as violative of its terms.” Beth Israel Hospital v. NLRB, 437 U. S. 483, 500–501 (1978) (internal quotation marks omitted). And the Board, “if it is to accomplish the task which Congress set for it, necessarily must have authority … to fill the interstices of the broad statutory provisions.” Id., at 501. So, while the Board’s decision “is not the last word” on these complex matters—given that its decisions are subject to review in federal court—“it must assuredly be the first.” Marine Engineers v. Interlake S. S. Co., 370 U. S. 173, 185 (1962) (emphasis added).

For that reason, this Court has long held that courts presented with claims arising out of a labor dispute must sometimes pause their proceedings to permit the Board to consider the dispute in the first instance. As relevant here, we have held that if §7—including its protection of the right to strike—“arguably” protects the conduct at issue in a state-court suit, then the court must await the Board’s word as to whether the conduct is, in fact, protected. Garmon, 359 U. S., at 245.

To determine whether conduct is “arguably protected,” a state court examines the showing of the party invoking Garmon and seeking to pause the litigation. The court asks whether that party has (1) “advance[d] an interpretation of the [NLRA] that is not plainly contrary to its language and that has not been ‘authoritatively rejected’ by the courts or the Board,” and (2) “put forth enough evidence to enable the court to find that the Board reasonably could uphold a claim based on such an interpretation.” Longshoremen v. Davis, 476 U. S. 380, 395 (1986). If so, the state court must pause proceedings to allow the Board to consider the complex legal and factual contours of the question whether the union’s conduct is actually protected by the NLRA.

The majority refers to this as “Garmon preemption,” in keeping with historical practice. Ante, at 3. But the term “preemption” is something of a misnomer. Rather than entirely and automatically precluding the state-court suit, the rule instead requires state courts to take a “jurisdictional hiatus” while the Board considers the dispute in the first instance. Sears, Roebuck & Co., 436 U. S., at 203. If the Board determines (subject to judicial review) that §7 protects the union’s conduct, normal conflict preemption kicks in: A state court may not hold a union liable on state-law claims for conduct that is protected by the NLRA. See Brown v. Hotel Employees, 468 U. S. 491, 503 (1984). But “if the Board decides that the conduct is not protected,” the state court may proceed to “entertain the litigation.” Davis, 476 U. S., at 397.[1]

With these general principles in mind, I now turn to the particulars of this case.

II

This suit arises out of a union-organized strike. Petitioner Glacier Northwest is a concrete-delivery company, and respondent International Brotherhood of Teamsters Local Union No. 174 (Union) represents Glacier’s concrete-delivery truckdrivers. After the drivers went on strike, Glacier sent disciplinary letters to some of the drivers. The Union filed an unfair labor practice charge with the Board, alleging that the disciplinary letters were unlawful retaliation against the drivers for engaging in strike conduct that is protected by the NLRA.

Glacier then filed a complaint in Washington state court, alleging that the Union engaged in tortious conduct when it instructed the drivers to strike at a time when there was wet concrete in some of the company’s delivery trucks. In response, the Union filed another Board charge, maintaining that Glacier’s lawsuit constituted additional unlawful retaliation.

With respect to Glacier’s tort suit, the Washington courts engaged in the standard Garmon inquiry, ultimately resulting in a determination by the Washington Supreme Court that the lawsuit could not proceed because the Union’s strike conduct was arguably protected by the NLRA. Glacier sought, and we granted, certiorari to review that decision. Notably, however, after the Washington Supreme Court issued its decision, the regional director acting on behalf of the Board’s General Counsel filed an administrative complaint against Glacier. In my view, for the reasons explained below, that subsequent event has greatly simplified the Garmon question.

A

The filing of the General Counsel’s administrative complaint necessarily suffices to establish that the Union’s strike conduct is “arguably protected” within the meaning of Garmon. Thus, the General Counsel’s complaint should have marked the end of any court involvement in this matter at this time.

The General Counsel’s complaint alleges that Glacier interfered with strike conduct protected by §7 when it disciplined its drivers for walking off the job and when it filed this tort suit. That complaint represents the General Counsel’s conclusion—reached after an extensive independent investigation involving collecting testimony and other evidence, and after careful consideration of the competing legal principles and policy concerns—that the Union’s claim that its strike conduct was protected “appears to have merit.” 29 CFR §§101.4, 101.8. One “cannot credibly contend that a claim that makes it through this gauntlet does not concern conduct ‘arguably’ protected by the NLRA.” Davis Supermarkets, Inc. v. NLRB, 2 F. 3d 1162, 1179 (CADC 1993); accord, Makro, Inc., 305 N. L. R. B. 663, 670 (1991).

A court presented with a General Counsel complaint should therefore find Garmon inherently satisfied. This is so because the entire point of Garmon’s arguably-protected test is to permit the court to assess the facts and relevant labor law in service of a gatekeeping function. The answer to the Garmon question simply (and solely) establishes whether the court can continue to entertain a lawsuit that relates to the challenged strike conduct, or whether the legal action must be suspended to allow the Board to make an initial assessment of the matter. The court evaluates the existing evidence and the law for a specific reason: to determine whether the lawsuit attacks arguably-protected conduct such that entertaining the legal action will interfere with the Board’s prerogative to develop the facts and adjudicate the merits of the dispute as part of the Board’s broader authority to develop national labor law.

If the General Counsel investigates the matter and files a complaint with the Board alleging that the union’s conduct is protected, it becomes indisputable that the pending legal action might interfere with the Board’s authority. Thus, a General Counsel complaint relieves the court of the burden of having to make the arguably-protected assessment based on its own understanding of the evidence and labor law—it is “arguable” that the union’s conduct is protected because the General Counsel is arguing just that. To be sure, we have said that the arguably-protected test is “not without substance” and is “not satisfied by a conclusory assertion of pre-emption.” Davis, 476 U. S., at 394. But an allegation from the Board’s General Counsel after a thorough investigation is a far cry from a “conclusory assertion” of protection.[2]

What is more, by virtue of the General Counsel’s complaint, the Board is, at this very moment, exercising its authority to adjudicate the merits of this dispute. On January 12, 2023, an ALJ denied Glacier’s motion to postpone the ALJ hearing on the General Counsel’s complaint pending this Court’s decision in this case. As the ALJ explained, the General Counsel’s pleading “constituted a determination that the strikers’ conduct was at least arguably protected by [the NLRA] and that this agency became the exclusive forum for adjudicating whether the strikers’ conduct was protected.”[3] A nine-day hearing ensued, and the parties completed posthearing briefing last week. We have said that “the need for protecting the exclusivity of [the Board’s] jurisdiction is obviously greatest when the precise issue brought before a court is in the process of litigation through procedures originating in the Board.” Marine Engineers, 370 U. S., at 185. That is exactly the situation here.

For these reasons, I believe that the filing of the General Counsel’s complaint is more than sufficient to trigger Garmon’s pause, and that it must be so if consistency with Congress’s intent to give the Board primary authority to interpret and enforce the NLRA is to be maintained. In circumstances like these, “the States as well as the federal courts must defer to the exclusive competence of the National Labor Relations Board.” Garmon, 359 U. S., at 245. And this Court is no exception. Because the General Counsel has now filed a complaint with the Board concerning the labor dispute at issue in this case, all courts—including this one—should stand down.

B

The majority does not take issue with my conclusion that the General Counsel’s complaint triggers a Garmon hiatus; instead, it takes no position on the matter, leaving the question open for the Washington courts to decide on remand. Ante, at 11, n. 3.

The majority’s reason for declining to address this argument is noteworthy. It explains that, because the General Counsel’s complaint was filed after the Washington Supreme Court had affirmed the dismissal of Glacier’s complaint on Garmon grounds, “[t]he lower courts have not addressed the significance, if any, of the Board’s complaint with respect to Garmon preemption.” Ante, at 11, n. 3. And since we are “ ‘a court of review, not of first view,’ ” ibid. (quoting Cutter v. Wilkinson, 544 U. S. 709, 718, n. 7 (2005)), the majority declines to “do so in the first instance.” Ante, at 11, n. 3.

This rationale is inconsistent with the broader approach that the majority takes in this case. It would be one thing if the Court simply noted the filing of the General Counsel’s complaint and authorized the lower courts to evaluate the impact of that complaint on the Garmon question in the first instance. But it goes further: The majority also inserts itself into the midst of this labor dispute now (despite the General Counsel’s complaint), proceeding to apply the Board’s cases to novel and difficult line-drawing questions and ultimately concluding that the strike conduct alleged in Glacier’s complaint is not even arguably protected.

The majority cannot have it both ways. A concern about the Court’s institutional role justifies, at most, vacating the judgment below and remanding for the lower court to consider the import of the General Counsel’s complaint. The same observation that compels the majority to allow for such lower-court consideration—that we are “ ‘a court of review, not of first view,’ ” ante, at 11, n. 3—should have likewise led it to decline to intrude into this labor dispute while it is pending before the Board.

III

For the reasons discussed above, I would have vacated the Washington Supreme Court’s judgment and remanded with directions to stay proceedings or dismiss Glacier’s complaint without prejudice, on the straightforward ground that the General Counsel’s complaint triggers the jurisdictional hiatus that Garmon requires.[4]

The majority sidesteps my preferred resolution of this matter and instead proceeds to engage in Garmon’s “arguably protected” test by applying a series of fact-intensive Board decisions to the bare allegations in Glacier’s state-court complaint. To do this, the majority invokes the Board’s “reasonable precautions” principle. Ante, at 6–8. That principle (discussed in Part IV, infra) is derived from the Board’s determination that striking workers must take reasonable precautions to protect persons, the employer’s premises, and its equipment from foreseeable, aggravated, and imminent harm due to the sudden cessation of work. The majority has taken it upon itself to apply the Board’s reasonable-precautions principle to the factual allegations about the Union’s conduct that Glacier alleges in this lawsuit, and it thereby concludes that the drivers’ conduct is not even arguably protected by the NLRA.

This course of action (which is already confounding given that the Board itself is currently considering the challenged strike conduct with the benefit of developed facts and labor law expertise) reflects an analytical approach to the issues presented that cannot be squared with Garmon.

A

Whether the NLRA protects particular strike conduct often turns on subtle factual disputes and nuanced legal distinctions. Here, for example, whether the Union’s strike conduct is protected or unprotected might well depend on whether the drivers left the concrete-delivery trucks’ revolving drums turning when they walked off the job. So, too, might it depend on fine legal gradations concerning how imminent or how aggravated the risk of harm must be to trigger the duty to take reasonable precautions. These kinds of determinations cry out for evidentiary hearings, and in this highly fact-sensitive area of the law, which generally develops on a case-by-case basis, the scope of NLRA protection in a given set of circumstances is typically determined once the facts have been established—through discovery, debate, and sometimes the tedious work of making contentious credibility determinations.

Fortunately, in this regard, Congress has gifted our legal system with an expert agency that thoroughly investigates what happened—i.e., the facts of strike-related labor disputes—and then engages in the initial task of answering the sometimes complex, always fact-bound question whether the NLRA protects the strike conduct at issue. Meanwhile, a court that is undertaking Garmon’s arguably-protected analysis is engaged in a fundamentally different inquiry. As explained in Part II–A, supra, while the court is most certainly considering strike conduct arising from a labor dispute, it is not meant to address the merits of these complex questions. Under the NLRA and Garmon, courts must take as a given that the Board is the entity to which Congress has assigned responsibility for initially determining what happened and taking the first crack at deciding whether the NLRA protects the union’s conduct. And far from usurping that Board function, Garmon tasks the court with merely conducting a threshold, gatekeeping assessment of whether the lawsuit before it must be paused, or whether the suit can proceed because it is not even arguable that the conduct at issue in the lawsuit is protected by the NLRA.

To avoid veering into the Board’s assigned territory, it is crucial that the courts have a clear understanding of the nature of the Garmon assessment and what it requires. The court asks, first of all, whether the party invoking Garmon has “advance[d] an interpretation of the [NLRA] that is not plainly contrary to its language and that has not been ‘authoritatively rejected’ by the courts or the Board.” Davis, 476 U. S., at 395. This inquiry involves merely comparing the union’s claim about the scope of its protection to the broad protective language of the statute and deciding whether the union’s interpretation has already been definitively rejected either by courts or by the Board.

The second task is to determine whether the party invoking Garmon has “put forth enough evidence to enable the court to find that the Board reasonably could uphold a claim based on such an interpretation.” Davis, 476 U. S., at 395. Again, this is not an invitation to supplant the Board’s fact-finding role or to usurp the authority that Congress has given the Board to make the initial underlying protected-or-unprotected determination. Rather, the point of this part of the Garmon assessment is simply to determine whether it is arguable that the Board—in the exercise of its discretion to develop labor law and aided by its investigation into the facts—could conclude that the strike conduct at issue is protected by the NLRA. See 359 U. S., at 245.

Thus, consistent with a statutory scheme that gives primacy to the agency’s expertise, a court’s task under Garmon is unmistakably modest. It must merely assess whether, in light of existing law and the evidence that has been amassed related to this strike, it is possible that the union could prevail before the Board. Put another way, instead of stepping into the Board’s shoes as primary factfinder, or even prognosticating about what the Board is likely to decide concerning the extent of NLRA coverage, a court that stands down upon a proper Garmon analysis has simply determined (1) that existing law does not plainly and authoritatively prohibit the strike conduct at issue, and (2) that evidence exists concerning how the strike was conducted that might ultimately favor the union, such that the lawsuit should pause to allow the Board to gather the facts and apply its expertise to determine whether the strike was lawful.

B

The majority seems to misunderstand all this in the context of this case. It correctly concludes that the Union has carried its burden of “advancing an interpretation of the [NLRA] that is not plainly contrary to its language and that has not been authoritatively rejected by the courts or the Board.” Ante, at 6 (internal quotation marks omitted). But it finds that the Union has failed to satisfy the second Garmon step, and it does so after undertaking its own assessment of the facts alleged in Glacier’s complaint and endeavoring to apply the Board’s fact-bound reasonable-precautions precedents. See, e.g., ante, at 7 (determining, based on alleged facts, that “[t]he drivers engaged in a sudden cessation of work that put Glacier’s property in foreseeable and imminent danger” and that the risk of harm to the concrete-delivery trucks was “both foreseeable and serious”); ibid. (concluding that “[t]he Union failed to ‘take reasonable precautions,’ ” after hypothesizing various steps that, according to the majority, the Union should have taken but did not).

Given what I have already said about Garmon’s purpose and what it calls for, the majority’s error in proceeding in this fashion is obvious. To my mind, if a court that is evaluating what to do per Garmon finds itself weighing in on such fact-bound matters as whether the strike posed a risk of harm that was aggravated enough or imminent enough to remove NLRA protection, or starts contemplating whether the precautions that the striking employees took to address any such risk were reasonable enough to allow them to retain the right to strike, it has unwittingly wandered into a domain that Congress intentionally assigned to the Board to address in the first instance.[5]

It is clear to me that Congress plainly intended for the Board’s factfinding function to be at the forefront of this kind of legal evaluation. Thus, in my view, when a court undertakes the Garmon analysis in a context such as this one, it should take care to limit itself to its own assigned responsibility: the mere determination of whether, given the union’s evidence and legal interpretation, the Board could possibly conclude that the union had taken reasonable precautions. If yes, the court should suspend the pending legal action to let the Board decide the question. To conclude no, given the fact-bound nature of the reasonable-precautions analysis, a court in all but the most exceptional circumstances will need to be able to point to a reasonable-precautions case from the Board that is on all fours with the facts of the case before it and that found the conduct unprotected. In that circumstance, the court can proceed with the suit, without breaking new legal ground on the scope of the right to strike.

In all events, then, courts can properly decide the Garmon issue without making law in this area, precisely as Congress intended. Indeed, I think we best respect congressional intent regarding the Board’s authority to develop uniform labor law by leaving the application of the Board’s reasonable-precautions principle to the Board itself. The majority’s contrary approach opens up the possibility that courts around the country will now act on bare allegations to generate conflicting results about the contours of the venerated right to strike, which, ironically, was the primary concern that motivated Congress to create the Board in the first place.

IV

For what it’s worth, even if the majority’s approach to deciding the Garmon question were the correct one, the majority misapplies the reasonable-precautions principle to the allegations here in a manner that threatens to impinge on the right to strike and on the orderly development of labor law.

A
1

A strike, by definition, is a “concerted stoppage of work by employees,” or “any concerted slowdown or other concerted interruption of operations by employees.” §142(2). When employees stop working, production may halt, deliveries may be delayed, and services may be canceled. At the risk of stating the obvious, this means that the workers’ right to strike inherently includes the right to impose economic harm on their employer.

Congress was well aware that organized labor’s exercise of the right to strike risks harm to an employer’s economic interests. See §151; NLRB v. Erie Resistor Corp., 373 U. S. 221, 234 (1963) (Congress’s protection of the right to strike reflects its understanding that strikes are authorized “economic weapon[s]”). Yet, Congress protected that right anyway. In fact, the threat of economic harm posed by the right to strike is a feature, not a bug, of the NLRA. The potential pain of a work stoppage is a powerful tool, and one that unquestionably advances Congress’s codified goal of achieving “equality of bargaining power between employers and employees.” §151. Unions leverage a strike’s economic harm (or the threat of it) into bargaining power, and then wield that power to demand improvement of employees’ wages and working conditions—goals that, according to Congress, benefit the economy writ large. See Sears, Roebuck & Co., 436 U. S., at 190.

Still, the right to strike is, of course, not unlimited. But when “Congress chose to qualify the use of the strike, it did so by prescribing the limits and conditions of the abridgment in exacting detail.” Erie Resistor, 373 U. S., at 234. Section 8 enumerates several limitations. For example, a union must notify an employer that it intends to terminate or modify its contract—and thus that a strike is possible—at least 60 days before striking. §158(d). A union cannot strike for unlawful purposes, such as putting economic pressure on parties other than the primary employer. §158(b)(4)(i)(B). And, in certain healthcare settings, unions must provide at least 10 days’ notice of the precise date and time of a strike. §158(g).

Additionally, §163 of the NLRA (which Congress added via the 1947 Taft-Hartley Amendments, 61 Stat. 151) states that “nothing in this subchapter, except as specifically provided for herein, shall be construed so as either to interfere with or impede or diminish in any way the right to strike, or to affect the limitations or qualifications on that right.”

Thus, the text of the NLRA allows for only two kinds of limitations on the right to strike: those enumerated in the Act itself, and the “limitations or qualifications” on the right that existed when the Taft-Hartley Amendments were enacted. See NLRB v. Drivers, 362 U. S. 274, 281–282 (1960). The only relevant limitation here is the one set out in NLRB v. Fansteel Metallurgical Corp., 306 U. S. 240 (1939).[6]

Our Fansteel decision stands for the principle that “employees ha[ve] the right to strike but they ha[ve] no license to commit acts of violence or to seize their employer’s plant.” Id., at 253. The facts of that case involved 95 striking employees who effected a “sit-down strike by taking over and holding two of [their employer’s] key buildings.” Id., at 248 (internal quotation marks omitted). The employees subsequently engaged in “a pitched battle” in which they “resisted the attempt by the sheriff to evict and arrest them.” Id., at 249. We held that the NLRA did not condone this conduct, which would “put a premium on resort to force” and would “subvert the principles of law and order which lie at the foundations of society.” Id., at 253.

Congress’s incorporation of Fansteel’s limitation into the NLRA establishes that, while employees have the right to withhold their labor peaceably, subsequent affirmative acts of violence, or seizure of an employer’s premises, are not protected labor practices.

2

As a general matter, the dispute in this case is over whether employees can withhold their labor if doing so risks damage to their employer’s property. As explained above, by carefully restricting limitations on the right to strike in the NLRA itself, Congress has indicated that the act of peacefully walking off the job is protected strike conduct even if economic harm incidentally results. What is not protected is any subsequent affirmative step to destroy or seize the employer’s property. This is the statutory backdrop against which the Board has developed the narrow requirement that striking employees must take reasonable precautions before or when they strike in order to forestall or address foreseeable, imminent, and aggravated injury to persons, premises, and equipment that might otherwise be caused by their sudden cessation of work.

The Board first applied this “reasonable precautions” principle to rank-and-file employees in Marshall Car Wheel & Foundry Co., Inc., 107 N. L. R. B. 314, 315 (1953), enf. denied on other grounds, 218 F. 2d 409 (CA5 1955). There, employees at a foundry walked off the job at a time when the foundry’s furnace was full of hot molten iron, threatening severe damage to the employer’s plant and equipment. 107 N. L. R. B., at 315. The Board concluded that the employees’ strike conduct was not protected by the NLRA, because the employees had a “duty to take reasonable precautions to protect the employer’s physical plant from such imminent damage as for[e]seeably would result from their sudden cessation of work.” Ibid.

The Board has also applied this principle in other similar cases. It determined, for example, that strikers who walked out of a certain kind of chemical plant—a plant that handled “extremely hazardous” chemicals that were “a hazard not only to employees but also to individuals living in the vicinity”—without shutting down the equipment had engaged in unprotected conduct. General Chemical Corp., 290 N. L. R. B. 76, 77, 83 (1988). Similarly, the Board held that the strike conduct of security guards whose walkout exposed a federal building’s occupants to “imminent” danger was not protected by the NLRA. International Protective Servs., Inc., 339 N. L. R. B. 701, 703 (2003).

But the narrow duty that Marshall Car Wheel and its progeny impose does not—and cannot—displace the general rule that labor strikes are protected even when the workers’ withdrawal of their labor inflicts economic harm on the employer. So the Board has also repeatedly held that employees have no duty to prevent the loss of perishable goods caused by their sudden cessation of work.

In a leading case, employees at a raw poultry plant decided to walk out at 8 a.m. “because by that time all employees would have reported to work and [the employer] would be in full operation with its largest number of chickens on the line.” Lumbee Farms Co-op., 285 N. L. R. B. 497, 503 (1987). The Board affirmed the ALJ’s reasoning that “[t]he fact that the strike occurred during the workday when chickens were on the line and vulnerable to loss does not mean employees automatically lost protection under the Act,” because “[s]trikers are not required under the Act to institute the strike at a specific time of day.” Id., at 506. Indeed, it is “[n]orma[l]” for “planned employee strikes [to be] timed to ensure the greatest impact on an employer.” Ibid.

The Board has applied this same reasoning in cases involving, for example, cheese and milk. See Leprino Cheese Co., 170 N. L. R. B. 601, 605 (1968); Central Okla. Milk Producers Assn., 125 N. L. R. B. 419, 435 (1959). In those cases, the Board also explained that the reasonable-precautions principle is “limited to situations involving a danger of ‘aggravated’ injury to persons or premises”—a danger “[o]bviously” not posed by the loss of, for example, cheese. Leprino Cheese, 170 N. L. R. B., at 607 (emphasis added). The Board has consistently reiterated that “[l]oss is not uncommon when a strike occurs.” Central Okla. Milk Producers, 125 N. L. R. B., at 435.

In short, it is indisputable that workers have a statutory right to strike despite the fact that exercising that right risks economic harm to employers. Congress has, in effect, drawn a line between those economic harms that are inherent in the act of peacefully walking off the job (which do not render the strike unprotected), and those that result from workers taking subsequent affirmative steps to seize the employer’s premises or engage in acts of violence (strike conduct that is not protected by the NLRA). The Board has further recognized a narrow duty that arises if a sudden cessation of work risks foreseeable, imminent, and aggravated harm to persons, premises, or equipment. Beyond this narrow reasonable-precautions requirement, however, employees have no obligation to protect their employer’s economic interests when they exercise the right to withhold their labor.

B

Glacier does not allege that the cement truckdrivers committed acts of violence or seized its plant or property as part of the strike the Union orchestrated. Instead, the thrust of its complaint is that the Union was aware of “the perishable nature of batched concrete,” App. 9, and that the drivers’ walkout was intentionally timed so as to risk harm to that product. See id., at 10 (alleging “sabotage, ruination and destruction of Glacier’s batched concrete”).

I agree with the majority that the risk of losing the batched concrete alone would not be sufficient to divest the striking drivers of statutory protection. As Glacier acknowledges, wet concrete is a perishable good. Ibid. And the Board has repeatedly reaffirmed that the loss of such perishable goods due to a mere work stoppage does not render a strike unprotected.

There is also no duty to take reasonable precautions to prevent this kind of economic loss, which—standing alone—posed no risk to persons, premises, or equipment, let alone a risk of aggravated harm. While it seems that the drivers were in a position to save the batched concrete that was inside their trucks when the strike was called (by, for instance, continuing to deliver it to the intended customers), that is beside the point. Employees have a protected right to withhold their labor. And it would undercut that right if they could be held liable for the incidental loss of the perishable goods (which includes concrete no less than raw poultry, cheese, or milk) that they tend to as part of their job.[7]

Where I disagree with the majority is the conclusion it draws from the fact that the batched concrete also risked harm to the drivers’ trucks, at least as alleged in Glacier’s complaint. The majority repeatedly ties the loss of the concrete—in particular, the risk that it would harden in the trucks—to the alleged risk of harm to the delivery trucks themselves.[8] But, to me, the alleged risk of harm to Glacier’s trucks involves a relatively complex factual analysis under the Board’s reasonable-precautions principle.

Glacier alleges that, “[o]nce at rest, concrete begins hardening immediately, and depending on the mix can begin to set within 20 to 30 minutes.” Id., at 8. Its complaint also asserts that “[i]f batched concrete remains in the revolving drum of the ready-mix truck beyond its useful life span, the batched concrete is certain or substantially certain to harden in the revolving drum and cause significant damage to the concrete ready-mix truck.” Id., at 9. But Glacier’s own submissions in Washington state court suggest that the Union instructed the drivers to return their trucks to Glacier’s yard after the strike began and to keep the ready-mix trucks running. See id., at 34, 77. Glacier’s submissions also suggest that those precautions actually provided the company’s managers and nonstriking employees with sufficient time to decide how to address the situation to prevent any harm to the trucks. See id., at 13, 72, 77, 82–83.

Was any risk of harm to the trucks here “imminent,” given the allegation that the Union instructed the drivers to keep the trucks running? Is the risk of concrete hardening in a delivery truck “aggravated,” in the way Marshall Car Wheel contemplates? Was returning the trucks to the employer’s premises and leaving them running a sufficient “reasonable” precaution, because it gave the employer sufficient time to address any risk of harm? Making the call about whether the NLRA protects the Union’s conduct raises these questions and others. Importantly, these kinds of questions not only involve making nuanced factual distinctions but also demonstrate that applying the Board’s reasonable-precautions precedents is, at bottom, a line-drawing exercise. Under circumstances like these, a court can confidently declare that a union’s conduct is not even arguably protected for Garmon purposes only where the allegations make out a clear Fansteel claim or where the alleged facts implicate a reasonable-precautions case that is directly on point. Because neither is true here, the Court should have concluded that the Union’s conduct was at least arguably protected.

Even if the Court’s task under Garmon were to apply the Board’s reasonable-precautions principle to the allegations of Glacier’s complaint and decide whether or not the Union engaged in unprotected conduct (to reiterate: that is not the assignment, see Part III–A, supra), I cannot agree with the majority’s conclusion that the risk to the trucks rendered the drivers’ strike unprotected by the NLRA. Instead, I would have credited Glacier’s own account, and thus would have concluded that the Union took reasonable precautions when it instructed the drivers to return the trucks and leave them running to avoid the concrete hardening imminently in the drums. The majority reaches the opposite conclusion by giving far too little weight to the allegation that the drivers returned the trucks, and also by substantially discounting the allegations that support the Union’s claim that the drivers left their trucks and revolving drums running. See ante, at 11.

Fortunately, the pending Board determination of what actually happened in connection with this particular strike will establish—as a matter of fact and not mere allegation—what precautions (if any) the drivers actually took and what harm (if any) the Union’s conduct actually posed to Glacier’s trucks.[9] But our different takes on these allegations only underscore the potential for variable outcomes when courts apply the Board’s fact-dependent principles to bare assertions.

To the extent that the majority’s conclusion rests on the alleged fact that “by reporting for duty and pretending as if they would deliver the concrete, the drivers prompted the creation of the perishable product” that “put Glacier’s trucks in harm’s way,” ante, at 10, I see nothing aggravated or even untoward about that conduct. Glacier is a concrete-delivery company whose drivers are responsible for delivering wet concrete, so it is unremarkable that the drivers struck at a time when there was concrete in the trucks. While selling perishable products may be risky business, the perishable nature of Glacier’s concrete did not impose some obligation on the drivers to strike in the middle of the night or before the next day’s jobs had started. To the contrary, it was entirely lawful for the drivers to start their workday per usual, and for the Union to time the strike to put “maximum pressure on the employer at minimum economic cost to the union.” NLRB v. Insurance Agents, 361 U. S. 477, 496 (1960); see also Lumbee Farms Co-op., 285 N. L. R. B., at 506.

Nor was the onus of protecting Glacier’s economic interests if a strike was called in the middle of the day on the drivers—it was, instead, on Glacier, which could have taken any number of prophylactic, mitigating measures.[10] What Glacier seeks to do here is to shift the duty of protecting an employer’s property from damage or loss incident to a strike onto the striking workers, beyond what the Board has already permitted via the reasonable-precautions principle. In my view, doing that places a significant burden on the employees’ exercise of their statutory right to strike, unjustifiably undermining Congress’s intent. Workers are not indentured servants, bound to continue laboring until any planned work stoppage would be as painless as possible for their master. They are employees whose collective and peaceful decision to withhold their labor is protected by the NLRA even if economic injury results. *** Today, the majority fails, in multiple respects, to heed Congress’s intent with respect to the Board’s primary role in adjudicating labor disputes, despite ostensibly applying Garmon, the bedrock case on that issue. The Court’s ruling is likely to cause considerable confusion among the lower courts about what Garmon requires. And any such confusion not only threatens to encroach upon the Board’s prerogatives, as Congress has assigned them, but also risks erosion of the right to strike.

Yet, the posture of this case provides an opportunity to mitigate the results of the majority’s errors. On remand, the state court should dismiss Glacier’s complaint without prejudice or stay its proceedings in view of the General Counsel’s complaint. Meanwhile, the Board—which is not bound by the allegations in Glacier’s complaint when making its assessment, and is well equipped to make findings of fact concerning the strike conduct at issue—should proceed to determine whether Glacier has interfered with strike conduct that is protected by the NLRA, as alleged by the General Counsel.


  1. Justice Thomas seeks to undercut our Garmon precedent by describing it as “od[d]” and “strang[e]” relative to “ ‘the usual preemption rule.’ ” Ante, at 1, 3 (opinion concurring in judgment). But, as discussed, the Garmon rule is not a standard preemption doctrine; it is different because it is doing different work. Garmon protects Congress’s judgment that the Board, not state or federal courts, should be generally responsible for the development of our Nation’s labor law. The required pause when Garmon’s “arguably protected” test is satisfied allows for efficient resolution of the dispute prior to the expenditure of state judicial resources, and the temporary nature of the pause makes it narrower, not broader, in effect than ordinary preemption.
  2. This is not to suggest that the General Counsel’s complaint is the end of the story, as the Board may ultimately disagree with the factual or legal basis of that pleading. But the complaint is surely sufficient to establish arguable protection, such that a court should stay its hand.
  3. Order Denying Motion for Postponement of Hearing in Glacier Northwest, Inc., Nos. 19–CA–203068, 19–CA–211776, p. 7.
  4. The Washington Supreme Court affirmed the dismissal of Glacier’s claims. Because only a pause of the state-court litigation is necessary under Garmon, the proper disposition is either a stay of proceedings or dismissal without prejudice.
  5. Justice Blackmun warned that a formulation of the Garmon test that directs attention to a party’s evidence might lead some courts to make such an errant assessment, “under the guise of weighing the sufficiency of the evidence.” Longshoremen v. Davis, 476 U. S. 380, 404 (1986) (opinion concurring in part and dissenting in part) (expressing the concern that a future court might misunderstand its role and mistakenly undertake “[to] mak[e] precisely the determination that Garmon makes clear is for the Board, and only the Board, to make”). That warning was prescient.
  6. The Senate Report accompanying the Taft-Hartley Amendments explained the four kinds of pre-existing “limitations or qualifications” on the right to strike that Congress had in mind in §163, which were drawn from decisions of the Board and this Court. See S. Rep. No. 105, 80th Cong., 1st Sess., 28 (1947); Drivers, 362 U. S., at 281–282. The three other exceptions concern strikes for illegal objectives, strikes in breach of contract, and strikes in breach of other federal law. See S. Rep. No. 105, at 28.
  7. Justice Alito, relying on the rule from NLRB v. Fansteel Metallurgical Corp., 306 U. S. 240 (1939), gleans more from the loss of concrete than either the majority or I do. He concludes that the NLRA’s right to strike does not protect the drivers’ alleged conduct because Glacier has alleged that the drivers purposefully caused the batched concrete to be destroyed. In my view, that approach fails to appreciate the distinction Fansteel drew between purposefully but peacefully stopping work (and the economic consequences that flow from that decision), which is protected, and taking subsequent, affirmative steps of violence or property seizure, which is unprotected. To be sure, Fansteel would have rendered the drivers’ actions here patently unprotected if they had taken the affirmative steps of stealing the trucks, slashing the trucks’ tires, or dumping out the concrete after they went on strike. But nothing like that is alleged in Glacier’s complaint.
  8. See, e.g., ante, at 8 (“[T]he Union executed the strike in a manner designed to compromise the safety of Glacier’s trucks and destroy its concrete”); ante, at 10 (“[The drivers] not only destroyed the concrete but also put Glacier’s trucks in harm’s way. This case therefore involves much more than ‘a work stoppage at a time when the loss of perishable products is foreseeable’ ”); ante, at 12 (“The Union’s actions not only resulted in the destruction of all the concrete Glacier had prepared that day; they also posed a risk of foreseeable, aggravated, and imminent harm to Glacier’s trucks”).
  9. For the same reason, the state court would not be bound by the majority’s recitation of the facts at this motion-to-dismiss stage in any future proceedings on this matter in state court.
  10. For example, Glacier could have instituted a lockout, see American Ship Building Co. v. NLRB, 380 U. S. 300, 310 (1965), used nonstriking employees to deliver the batched concrete, or had temporary replacement drivers lined up and ready to go. Glacier was on notice that a strike was possible because the Union was statutorily required to give 60-days advance notice of the proposed termination or modification of the collective-bargaining agreement, §158(d), and because negotiations had broken down.