The Good-Faith Doubt Test and the Revival of Joy Silk Bargaining Orders


The last fifty-two years have borne witness to the swift degradation and virtual irrelevance of the bargaining order. By the end of the twentieth century even pro-enforcement officials in the NLRB were acknowledging the difficulty of obtaining an enforceable bargaining order, and the remedy rarely appears these days in the agency’s published decisions.

This is not the product of the usual economic or political factors cited as reasons for the labor movement’s and its attendant regulating schema’s diminishment. Rather, the decline of the bargaining order can be explained almost entirely by the disappearance of the so-called Joy Silk doctrine from the labor law landscape in 1969. Under Joy Silk, the NLRB would order an employer to recognize and bargain with a labor union if the union represented a majority of the employees in an appropriate bargaining unit at the time it requested recognition, and the employer denied the request while lacking a good faith doubt as to the union’s majority status and took action calculated to dissipate that majority status. Under closer examination, it has become clear that the NLRB’s abandonment of the good-faith doubt test in favor of the misconduct-centric analysis enunciated in Gissel Packing is intimately connected to the agency’s inability to enforce the law and, as a result, fulfill its statutory mission of encouraging collective bargaining.

This Article addresses the primary objections to the NLRB’s use of the good-faith doubt test in the realm of union requests for recognition that were raised in its heyday and elucidates their lack of historical, legal, and practical foundations. The NLRB’s inquiry into motive in the pre-recognition context was statutorily permissible, logically consistent, and effectively deterrent. Reviving the good-faith doubt test of Joy Silk and enforcing Section 8(a)(5) of the NLRA as written would better encapsulate the bargaining orders envisioned in past Supreme Court precedent and equip the NLRB with a tool historically proven to prevent misconduct in elections.


The National Labor Relations Act is considered a broken statute. Decades of criticism of the United States’ keystone federal labor law have produced a rock-solid consensus among the field’s scholars: the law fails to prevent the wrongdoing for which it was designed to address, and it cannot adequately remedy those wrongs once committed.1Pulling no punches, one of the world’s leading labor economists has declared that the statute “no longer fits American economic reality and has become an anachronism irrelevant for most workers and firms.” Richard B. Freeman, What Can We Learn from the NLRA to Create Labor Law for the Twenty-First Century?, 26 A.B.A. J. Lab. & Emp. L. 327, 330 (2011). Legislative attempts to strengthen the statute have perennially failed, regardless of the scope of the changes sought,2Cynthia L. Estlund, The Ossification of American Labor Law, 102 Colum. L. Rev. 1527, 1532–44 (2002). and administrative efforts have proven either fleeting or ineffective.3See Catherine L. Fisk & Deborah C. Malamud, The NLRB in Administrative Law Exile: Problems with Its Structure and Function and Suggestions for Reform, 58 Duke L.J. 2013, 2043–57 (2009). What remains is a legal apparatus that relies largely upon voluntary compliance, armed with little more than the bare threat of litigation costs and reputational harm.4See H.R. Rep. No. 116-347, at 20–21 (2019); Lynn Rhinehart & Celine McNicholas, Shortchanged—weak anti-retaliation provisions in the National Labor Relations Act cost workers billions, Econ. Pol’y Inst. (Apr. 22, 2021), [].

This precarious state of affairs could suggest that the statute’s path to impotence was inevitable due to fatal flaws in its design. Indeed, many eminent commentators have argued just that.5See Matthew Dimick, Counterfeit Liberty, 3 Catalyst 47 (2019); James Gray Pope, The Thirteenth Amendment versus the Commerce Clause: Labor and the Shaping of American Constitutional Law, 1921-1957, 102 Colum. L. Rev. 1, 112–22 (2002); Christopher L. Tomlins, The State and the Unions: Labor Relations, Law, and the Organized Labor Movement in America, 1880-1960, at 317–28 (1985). But it is possible that this historical prognostication is overstated. Perhaps the law retains much of its past vitality but—like other pioneering progressive legislation—has fallen victim to anti-enforcement impulses amongst American regulators over the last half-century.6See, e.g., Luke Herrine, The Folklore of Unfairness, 96 N.Y.U. L. Rev. 431, 514–22 (2021) (chronicling the entrenchment of anti-enforcement sentiment at the Federal Trade Commission).

Early expressions of the law’s potency are not difficult to find. Consider this excerpt from one of the first labor law cases to reach the Supreme Court:

It is for the Board not the courts to determine how the effect of prior unfair labor practices may be expunged. That determination the Board has made in this case and in similar cases by adopting a form of remedy which requires that an employer bargain exclusively with the particular union which represented a majority of the employees at the time of the wrongful refusal to bargain despite that union’s subsequent failure to retain its majority. The Board might well think that, were it not to adopt this type of remedy, but instead order elections upon every claim that a shift in union membership had occurred during proceedings occasioned by an employer’s wrongful refusal to bargain, recalcitrant employers might be able by continued opposition to union membership indefinitely to postpone performance of their statutory obligation. In the Board’s view, procedural delays necessary fairly to determine charges of unfair labor practices might in this way be made the occasion for further procedural delays in connection with repeated requests for elections, thus providing employers a chance to profit from a stubborn refusal to abide by the law. That the Board was within its statutory authority in adopting the remedy which it has adopted to foreclose the probability of such frustrations of the Act seems too plain for anything but statement.7Frank Bros. Co. v. NLRB, 321 U.S. 702, 704–05 (1944) (internal citations and quotations omitted).

This passage, written by Justice Hugo Black in a unanimous opinion enforcing an order by the National Labor Relations Board (NLRB, Labor Board, or the Board), is one of the clearest endorsements of the federal government’s authority to remedy violations of labor law. The Court deferentially appraised the NLRB’s expertise in the fact-intensive world of industrial relations and accurately delineated the agency’s reasoning for adopting the specific remedy at issue, leaving no room for ambiguity as to why the Court was approving the government’s position. The decision, barely a thousand words in length when cleaved of citations, was evidently not a difficult one.

A new student of federal labor law as developed under the National Labor Relations Act (NLRA or the Act) and its amendments829 U.S.C. § 151 et seq. (2018). could thus hardly be faulted for reading Justice Black’s language and presuming that the remedy the Court canonized—an order for an employer to bargain with the majority representative of its employees despite that majority status not having been certified by means of a secret-ballot election, more colloquially known as a bargaining order—would become an indispensable weapon in the NLRB’s administrative arsenal for deterring employer wrongdoing. While this held true for the quarter-century of case law that followed the Frank Bros. decision, the last fifty-three years have borne witness to the swift degradation and virtual irrelevance of the bargaining order. By the end of the twentieth century even pro-enforcement officials in the agency were acknowledging the difficulty of obtaining an enforceable bargaining order,9Memorandum GC 99-7 from Fred Feinstein, General Counsel of the National Labor Relations Board to All Regional Directors, Officers-in Charge and Resident Officers, Guideline Memorandum Concerning Gissel, 1 (Nov. 10, 1999) [] (noting difficulty of enforcing bargaining orders in the circuit courts of appeal) [hereinafter Gissel Memo]. For an overview of then-General Counsel Fred Feinstein’s pro-enforcement reputation, see Steven Greenhouse, Beleaguered Labor Board Counsel is Bowing Out, N.Y. Times, Mar. 6, 1998, at A1. and today the remedy rarely appears in the Labor Board’s published decisions.

The under-utilization of this remedy is not the product of the usual economic or political factors cited as reasons for the labor movement and its attendant regulating schema’s diminishment. Rather, the decline of the bargaining order can be explained almost entirely by the disappearance of the so-called Joy Silk doctrine from the labor law landscape in 1969.10Joy Silk Mills, Inc., 85 N.L.R.B. 1263 (1949), enforced as modified, 185 F.2d 732 (D.C. Cir. 1950). cert. denied, 341 U.S. 914 (1951). Under the Joy Silk doctrine, the NLRB would order an employer to recognize and bargain with a labor union if the union represented a majority of the employees in an appropriate bargaining unit at the time it requested recognition, and the employer denied the request while lacking a good faith doubt as to the union’s majority status and took action calculated to dissipate that majority status.11Joy Silk Mills, Inc. v. NLRB, 185 F.2d 732, 741 (D.C. Cir 1950). The union’s claim of majority status was almost always proffered through a showing of signed authorization cards. This more or less described the Labor Board’s practice for the first two decades of the Taft-Hartley Act, until the good-faith doubt test was abandoned (in a manner of speaking)12See infra Part II.E. by the Board at oral argument of the Gissel Packing case.13NLRB v. Gissel Packing Co., Inc., 395 U.S. 575, 594 (1969). Shedding the “thicket” of any inquiry into employers’ state of mind, the resulting Gissel brand of bargaining orders generally are only issued where an employer’s unlawful conduct is so “outrageous” and “pervasive” so as to make the holding of a fair election impossible.14Id. at 613–14. This shift in the law, vociferously lobbied for by many pillars of the mid-century labor bar, resulted in an explosion of employer unfair labor practices and widescale desertion of the NLRB’s election machinery.15. Brian J. Petruska, Adding Joy Silk to Labor’s Reform Agenda, 57 Santa Clara L. Rev. 97, 110–33 (2017). Notwithstanding this inversion of its statutory mission,16Id. at 138–43. no serious push was made by the agency to resuscitate the good-faith doubt test for employer refusals to bargain with unions that had requested recognition but not obtained certification through a Board election. Gissel remained frozen as precedent regardless of its efficacy, having almost immediately crystallized bargaining orders as an “extraordinary” remedy of last resort.17See NLRB v. American Cable Systems, Inc., 427 F.2d 446, 448 (5th Cir. 1970).

The ice may have finally begun to thaw on August 12, 2021 when Jennifer Abruzzo, the newly-confirmed General Counsel nominated to the position by President Joseph Biden Jr., released a memorandum which, among many other agenda items, cited Joy Silk as an “initiative” she would “carefully examine” and directed the NLRB’s regional staff to seek out cases for agency coordination which fit the traditional good-faith doubt fact pattern.18Memorandum GC 21-04 from Jennifer A. Abruzzo, General Counsel of the National Labor Relations Board, to All Regional Directors, Officers-in Charge and Resident Officers (Aug. 12, 2021) [] [hereinafter Abruzzo Memo]. Given the historical context of General Counsels’ debut memoranda, which typically serve as a platform of sorts for the policies to be pursued and the decisions to be overturned during the life of their respective terms,19See, e.g., Memorandum GC 18-02 from Peter R. Robb, General Counsel of the National Labor Relations Board to All Regional Directors, Officers-in-Charge and Resident Officers
(Dec. 1, 2017) [].
it is safe to assume that Joy Silk is once again the official prosecuting policy of the General Counsel’s office and, by extension, the Labor Board’s regional offices. This would represent a seismic shift in the law governing the customs of union recognition requests to be felt by parties and counsel alike; few if any labor lawyers today have ever practiced under the Joy Silk doctrine.20General Counsel Abruzzo has recently filed a brief in a case pending before the Board which requests that the Joy Silk doctrine be reinstated in its original form. Brief in Support of General Counsel’s Exceptions to the Administrative Law Judge’s Decision at 36–45, Cemex Construction Materials Pacific, LLC, 28-CA-230115 (Apr. 11, 2022).

Of course, this General Counsel will not be writing on a clean slate. The good-faith doubt test was immensely controversial by the time it was being vigorously enforced in the 1960s. Commentary published in law reviews, trade journals, and congressional hearings almost unanimously derided the Kennedy Board’s handling of the case law as subjective, simplistic, naïve, and worse. Indeed, it is difficult to find a single contemporaneous scholar or practitioner who defended the Joy Silk doctrine on its merits. This is an astonishing fact of history—not only because agreement is famously rare amongst the warring factions of labor law,21See Bertrand P. Pogrebin, NLRB Bargaining Orders Since Gissel: Wandering from a Landmark, 46 St. John’s L. Rev. 193, 193 (1971) (“The NLRB’s abandonment of the good-faith doubt test, in refusal to bargain cases, should have been one of those rare moments in the development of the labor law when both labor and management could applaud—and for the same reasons.”). but because good faith doubt was always the more coherent framework for bargaining order remedies and faithful application of the NLRA’s text than the favored standard which ultimately culminated in the Supreme Court’s Gissel decision.

This Article, then, is the first attempt to respond to the critical scholarship at length and defend the Joy Silk doctrine on its substantive laurels.22As will be discussed, Brian Petruska has done yeoman’s work in retraining scholarly attention to the history of Joy Silk bargaining orders and demonstrating their worth as a remedial tool. Petruska, supra note 15. However, Petruska mostly focused his analysis towards demonstrating the quantitative impact of Joy Silk on the NLRB’s ability to deter unfair labor practices rather than addressing and rebutting the contemporaneous criticism that ultimately led the doctrine to fall out of favor with key agency officials. See discussion infra Part II.B. This Article can thus be viewed as an attempt to supply the normative counterpart to Petruska’s positive argument in favor of the NLRB’s revival of the Joy Silk doctrine. Section I recounts the history of the bargaining order remedy from the earliest days of the Wagner Act through its maturation under the Taft-Hartley Act in the form of the good-faith doubt test, until the test’s unceremonious departure without the deliberation of administrative adjudication. Section II explores the modern ecosystem of bargaining orders in the wake of the stricter Gissel test, noting the endangered status of this remedy and its deleterious effects on the NLRB’s twin responsibilities of remedying unfair labor practices and holding elections that “determine the uninhibited desires” of employees “under conditions as nearly ideal as possible.”23Gen. Shoe Corp., 77 N.L.R.B. 124, 126–27 (1948). Section III recapitulates the various objections levied against the Joy Silk doctrine during the late 1960s debate over whether the Labor Board could compel an employer to recognize and bargain with a union based solely on authorization cards and rejects them as legally and logically insufficient, asserting that the good-faith doubt test is the superior vehicle for enforcing federal labor law compared to the bargaining order-analysis which followed. The Article concludes by welcoming the long-overdue revival of the Joy Silk doctrine and situating the remedy within the modern legal landscape.

I. The History of NLRB Bargaining Orders and the Good-Faith Doubt Test

A. The Origins and Establishment of the Bargaining Order Remedy

The NLRA speaks in remarkably broad terms regarding an employer’s duty to bargain with a union. Section 8(a)(5) of the Act makes it an unfair labor practice for an employer “to refuse to bargain collectively with the representatives of his employees, subject to the provisions of section 9(a).”24National Labor Relations Act § 8(a)(5), 29 U.S.C. § 158(a)(5) (2018). Section 9(a) in turn defines a “representative” as one “designated or selected for the purposes of collective bargaining by the majority of the employees in a unit appropriate for such purposes[.]”25National Labor Relations Act § 9(a), 29 U.S.C. § 159(a) (2018). This language has remained substantively unaltered since the Act’s original passage in 193526The lone caveat being that the original Section 8(5) of the Wagner Act became Section 8(a)(5) in 1947 as a result of the Taft-Hartley Act’s addition of union unfair labor practices by way of Section 8(b). and is buttressed by the statute’s explicit encouragement of collective bargaining as an engine of national labor policy.27National Labor Relations Act § 1, 29 U.S.C. § 151 (2018) (declaring that it is “the policy of the United States to . . . encourag[e] the practice and procedure of collective bargaining”). The statute thus makes clear that an employer, when presented with a demand for collective bargaining by the majority representative of its employees, must bargain with that representative, but left the NLRB and federal courts responsible to give effect to key components of this duty. Chief among these unfilled gaps was when—if ever—an employer would be obligated to bargain with a union that claimed to possess a majority of employees but had not demonstrated that support through an election, the Act’s only specified method for certification.28The Wagner Act originally provided that the NLRB could certify representatives through “a secret ballot of employees, or utilize any other suitable method to ascertain such representatives.” See Craig Becker, Democracy in the Workplace: Union Representation Elections and Federal Labor Law, 77 Minn. L. Rev. 495, 505–08 (1993).

The NLRB would answer this question by regularly certifying unions through means other than a secret-ballot election in its first years of operation, sanctioning tallies of signed authorization cards, membership applications, strike and picket-line participation, and other vectors of majority support.29Id. at 508. This practice prompted fierce opposition from employers, political conservatives, a skeptical press, and even many established unions,30Id. at 508–10. persuading the Board in Cudahy Packing to forswear any path for certification other than elections.3113 N.L.R.B. 526, 531–32 (1939) (“Although in the past we have certified representatives without an election upon a showing of the sort here made, we are persuaded by our experience that the policies of the Act will best be effectuated if the question of representation which has arisen is resolved by secret ballot.”); Becker, supra note 28, at 510–11. The Taft-Hartley Congress, intent on making this policy permanent, amended Section 9(c) to render elections the sole option for Board certification.32Becker, supra note 28, at 511–12.

But the NLRA did not speak exclusively in terms of certification. Sections 8(a)(5) and 9(a) referred to representatives that were “designated or selected” by a majority, conspicuously opening the door for Board-ordered bargaining in resolutions of unfair labor practice cases where unions could “produce satisfactory evidence of a majority.”331 NLRB Ann. Rep. 84–85, 89 (1936). Voluntary bargaining with a noncertified union was also plainly lawful under this interpretation. That is to say, the statute from its inception appeared to contemplate the possibility that a union could obtain recognition where wronged through an employer’s unfair labor practices, even if it had not demonstrated majority support by winning a secret-ballot election prior to adjudication of the charges.

The Cudahy Packing decision was silent on this language, but the Supreme Court shortly thereafter approved the practice of Board-ordered recognition of unions that were victims of unlawful refusals to bargain. In NLRB v. Bradford Dyeing Association,34310 U.S. 318 (1940). the Court upheld the NLRB’s findings that the employer unlawfully dissipated the charging union’s majority by firing union supporters and creating a company-dominated labor organization. The Labor Board had restored the union’s majority and ordered the employer to recognize and bargain with the union, rejecting the argument that its order was forcing a bargaining representative on an unwilling majority.35Bradford Dyeing Ass’n, 4 N.L.R.B. 604, 616–17 (1937) (“[T]he record is clear that, had it not been for the unfair labor practices of the respondent . . . the respondent’s employees would have remained members . . . . The unfair labor practices of the respondent cannot operate to change the bargaining representative previously selected by the untrammelled [sic] will of the majority.” (emphasis added)). In enforcing the Board’s order in full, the Court clearly established the agency’s authority to issue bargaining orders in cases where the employer’s unfair labor practices had destroyed a union’s majority. It is worth noting that majority status in this case had been denoted by membership application cards, and no issue of certification (versus “designation” or “selection”) was raised.36Bradford Dyeing Ass’n, 310 U.S. at 339. While the NLRB’s order was issued over a year-and-a-half before the implementation of the Cudahy Packing rule—thus giving the employer room to challenge its duty to bargain with a union that had not prevailed in a secret-ballot election—the Court’s decision came almost a year after the rule was announced and raised no quarrel with the card-based majority.

Consistent with its deferential posture to the NLRB’s early crafting of industrial policy,37The NLRB famously went undefeated in the first sixteen Board decisions appealed to the Supreme Court. James A. Gross, The Reshaping of the National Labor Relations Board 17 (1981). But see Karl Klare, Judicial Deradicalization of the Wagner Act and the Origins of Modern Legal Consciousness, 1937–1941, 62 Minn. L. Rev. 265 (1978) (criticizing the grounds on which several Wagner-Act era NLRB orders were enforced or denied by the Supreme Court). the Court gave the agency great leeway in determining when, if ever, elections were required to establish majority status in the face of unremedied unfair labor practices. In NLRB v. P. Lorillard Co.,38314 U.S. 512 (1942) (per curiam). the Court summarily reversed a judgment from the Sixth Circuit which held that the passage of time and emergence of a rival labor organization since the Labor Board first ordered the employer to bargain with the charging union necessitated an election to reassess the latter’s majority status.39The Board had denied the employer’s motion for a rehearing on these matters and rejected its petition for an election. P. Lorillard Co., 16 N.L.R.B. 684, 687 (1939). Chiding the lower court that it was “for the Board to determine” what remedy of the employer’s refusal to bargain would best effectuate the policies of the NLRA—citing Bradford Dyeing Ass’n as the lead case for this proposition—the Justices nixed the Sixth Circuit’s modification of the Board’s order and enforced the original order in full.40P. Lorillard Co., 314 U.S. at 513. Once again, the union’s majority status was based upon membership application cards,41See P. Lorillard Co., 16 N.L.R.B. at 690–91. which the Court deemed sufficient to constitute a “duly selected bargaining representative” of the employees in question.42P. Lorillard Co., 314 U.S. at 512.

However, the question of whether an employer must recognize and bargain with a union which had lost its majority status for reasons ostensibly unrelated to the employer’s initial refusal to recognize was left unresolved. The Supreme Court’s solution through Frank Bros. Co. v. NLRB43321 U.S. 702 (1944). served as a template for the “modern” bargaining order. In that case, the charging union obtained signed cards from forty-five of the eighty employees working in the employer’s factory and requested recognition and bargaining.44See id. at 702. The Court decided these cards showed the employees had “designated” the union as their bargaining representative. Id. After its overtures were denied the union filed a petition with the NLRB for an election; the employer subsequently engaged in an “aggressive” anti-union campaign which included threats of plant closure if the union won the election.45Id. at 702–03. The union withdrew its election petition and instead filed unfair labor practice charges, which, after seven months, resulted in the Labor Board issuing a complaint against the employer which alleged an unlawful refusal to bargain.46Id. at 703. In challenging the appropriateness of the proposed bargaining order remedy, the employer proffered evidence that the union no longer possessed a majority because thirteen of the original supporters had left the factory and were replaced by new hires after the filing of the charges.47Id. Unlike the facts of Bradford Dyeing and P. Lorillard Co., this loss of majority was not alleged to be the direct product of unlawful discharges or other employer coercion; the Court apparently assumed arguendo that these resignations occurred “in the normal course of business.”48See id.

Justice Black, enforcing the NLRB’s order for the employer to recognize and bargain with the union, emphatically defended the agency’s decision to peg the union’s majority status to the date of the employer’s unlawful refusal to bargain.49See id. at 704–05. Rejecting the employer’s attempts to distinguish the employee resignations from the employer’s rejection of the collective bargaining principle, Justice Black invoked the Labor Board’s tenet that an “unlawful refusal of an employer to bargain collectively with its employees’ chosen representatives disrupts the employees’ morale, deters their organizational activities, and discourages their membership in unions.”50Id. at 704. Justice Black dispensed with any notion that the remedy visited an injustice upon the employees “who may wish to substitute for the particular union some other bargaining agent or arrangement” by pointing out that a bargaining order does not “fix a permanent bargaining relationship” upon the shop.51Id. at 705.

Thus, before the end of World War II, the Supreme Court issued three unanimous decisions within a four-year span enforcing NLRB bargaining orders in which signatures alone served as evidence of employees having designated or selected unions as their respective bargaining representatives. These cases definitively established that under the Wagner Act, the Labor Board possessed the statutory authority to compel an employer to recognize and bargain with a union that had not demonstrated its majority status in a secret-ballot election, even if the union had lost its majority since the time of the initial recognition request either because of the employer’s independent unfair labor practices or simply from administrative inertia. It remained less clear, however, under what conditions an employer would be required to recognize a union at the time it claimed support from a majority of the employer’s employees. It was also uncertain what events would preclude an employer who escaped this initial obligation from relying on an election to test the union’s majority. It would fall on the NLRB to weave these seemingly disparate threads into the nascent patchwork of national labor policy.

B. Constructing and Institutionalizing the Good-Faith Doubt Test

The NLRB did not have to look far to find a suitable application of the bargaining order remedy. While the good-faith doubt test would become synonymous with the Joy Silk decision,52. See infra notes 66–76 and accompanying text. the Labor Board applied primitive versions of it throughout the 1940s. And like so many aspects of foundational American law, the test’s origins and eventual popularization can be fairly credited to Learned Hand’s judicial reasoning. Tasked with substantiating the agency’s rather conclusory finding of the employer’s unlawful refusal to bargain in the famous Remington Rand case,53See Remington Rand, Inc., 2 N.L.R.B. 626 (1937). The NLRB followed nearly 105 pages of factual findings in its decision with a one-sentence analysis upholding the union’s charge that the employer violated Section 8(5) of the Wagner Act. Id. at 731. But this short treatment of the record was defensible as the employer publicly and emphatically rejected its duty to bargain with the union and claimed the Act was unconstitutional. Id. at 731 n.100. For a detailed recounting of Remington Rand as it pertained to early defiance of federal labor law, see Ahmed A. White, Industrial Terrorism and the Unmaking of New Deal Labor Law, 11 Nev. L.J. 561, 574–78 (2011). For a historical account of employers’ coordinated constitutional challenges to the NLRA, see James A. Gross, The Making of the National Labor Relations Board 171–73, 204–11 (1974). the Second Circuit enforced the Board’s order in full and lent intellectual support to the broad condemnations of industrial obstinacy.54NLRB v. Remington Rand, Inc., 94 F.2d 862 (2d Cir. 1938). Judge Hand elaborated upon an employer’s duty to bargain at great length:

The [NLRB] was certainly free to find that the respondent had been guilty of “unfair labor practices,” for it obviously meant not to confer with the [union] . . . . The respondent answers that it had no official or conclusive information that the [union] was the duly accredited bargaining representative of the men, and that it could not have had until the Labor Board had itself so decided. The Labor Board does indeed have that power, section 9(c), 29 U.S.C.A. § 159(c), and when there is a real doubt, we may assume arguendo that the employer need not decide the issue at his peril; faced by two sets of putative representatives, each claiming to be the properly accredited one, it would seem fairly plain that he need not choose at his peril, especially if he is not allowed to take a vote himself. The same is equally true, though only one set makes the claim; he may be in genuine doubt how many it represents. If he cannot satisfy himself of their credentials, and if he cannot by informal appeal to the Labor Board invoke its power, it would certainly seem that he should be free not to recognize either; but from that immunity it does not in the least follow that he need be satisfied with no evidence except the Board’s certificate; it may be entirely apparent from other sources that one set really represents the majority. In the case at bar even though the respondent were in doubt as to the [union]’s authority, that doubt did not excuse it; for it is quite plain that its position was not based upon any doubt, but upon its unwillingness to treat with “outside” representatives of its employees; that is to say, to recognize the solidarity of the craft as such. The greater included the less, and having taken that position, it may not now say that it could not know whether the [union] was properly accredited. Had that been the real reason for its refusal, presumably it would have been persuaded by the evidence which the [union] could have presented. It made no effort to learn the facts and took the chance of what they might be.55Id. at 868–69 (emphasis added).

In this passage, Judge Hand determined that an employer presented with a request for recognition from an organizing union had no automatic escape from bargaining through the NLRB’s election machinery (“the Board’s certificate”), presciently tracing the same path the Supreme Court would trek in its Bradford Dyeing, P. Lorillard Co., and Frank Bros. trio of decisions. Judge Hand allowed only two exceptions to this general duty: (1) where an employer was faced with competing claims of majority status from multiple unions;56This was such a frequent phenomenon in the Wagner Act era that Congress felt compelled to grant employers the right to file an election petition when confronted by demands for recognition from more than one union. This provision, Section 9(c)(1)(B), was later amended to encompass a demand for recognition from a single union. See Becker, supra note 28, at 519. or (2) where an employer held a “genuine” or “real” doubt of the organizing union’s majority status such that “he cannot satisfy himself of their credentials.” Because the employer “made no effort learn the facts” of the union’s majority—for example, by reviewing the veracity of the union’s cards—and instead refused to deal with the union under any circumstances, the employer was guilty of an unlawful refusal to bargain. Post-hoc demands for the Labor Board’s certification process would not free him of his obligation.

The next major development came again from the Second Circuit. In NLRB v. Dahlstrom Metallic Door Co., Judge Charles Edwards Clark upped Judge Hand’s ante and castigated the certification-only bargaining argument as “frivolous,” for “[a]n employer is under a duty to bargain as soon as the union representative presents convincing evidence of majority support.”57112 F.2d 756, 757 (2d Cir. 1940). Judge Clark’s language is startling when surveying the law’s later development in Linden Lumber Div. v. NLRB, 419 U.S. 301 (1974), discussed infra. Citing Remington Rand, Judge Clark held that the charging union’s tender of authorization cards from 295 employees of the 310-person bargaining unit constituted “ample evidence” of majority status and the employer “could not in good faith” ignore the union’s request for recognition.58Dahlstrom Metallic Door, 112 F.2d at 757 (emphasis added). The Second Circuit unanimously upheld the NLRB’s finding of an unlawful refusal to bargain and enforced the bargaining order remedy despite the lack of any independent unfair labor practices committed by the employer following the union’s initial request for recognition.59Id. at 758; see Transcript of Oral Argument at 44, Linden Lumber Div., Summer & Co. v. NLRB, 419 U.S. 301 (1974) (No. 73-1231).

The Second Circuit’s ruminations on an employer’s motive in the recognition-request context quickly percolated through the NLRB’s decisions. The Labor Board would variously describe an employer’s necessary mindset as one of “genuine”,60Consolidated Machine Tool Corp., 67 N.L.R.B. 737, 740 (1946); The Nubone Co., Inc., 62 N.L.R.B. 322, 326 (1945). “honest”,61Roanoke Pub. Warehouse, 72 N.L.R.B. 1281, 1282 (1947); Consolidated Machine Tool Corp., 67 N.L.R.B. at 740; The Nubone Co., 62 N.L.R.B. at 326; Lennox Furnace Co., Inc., 20 N.L.R.B. 962, 980 (1940). “reasonable”,62Frank Bros. Co., 44 N.L.R.B. 898, 911 (1942); Burnside Steel Foundry Co., 7 N.L.R.B. 714, 723 (1938). or even a “bona fide” doubt63Artcraft Hosiery Co., 78 N.L.R.B. 333, 334 (1948). when presented with a union’s claim of majority status, and the judiciary responded in kind.64See, e.g., NLRB v. James Thompson & Co., 208 F.2d 743, 746 (2d Cir. 1953) (honest doubt); NLRB v. Franks Bros. Co., 137 F.2d 989, 991 (1st Cir. 1943) (reasonable doubt); NLRB v. Clinton E. Hobbs Co., 132 F.2d 249, 251 (1st Cir. 1942) (genuine doubt); NLRB v. Harris-Woodson Co., 179 F.2d 720, 723 (4th Cir. 1950) (bona fide doubt). But in the latter years of the 1940s, and notably following the Taft-Hartley Act’s codification of the parameters of collective bargaining in Section 8(d),65Section 8(d) provides that “to bargain collectively is the performance of the mutual obligation of the employer and the representative of the employees to meet at reasonable times and confer in good faith with respect to wages, hours, and other terms and conditions of employment . . .” Labor Management Relations Act, 1947 (Taft-Hartley) § 8(d), 29 U.S.C. § 158(d) (2018) (emphasis added). the NLRB coalesced towards “good faith” as the benchmark for a permissible level of suspicion.66See Wilson & Co., Inc., 77 N.L.R.B. 959, 960 (1948); R.J. Lovvorn, 76 N.L.R.B. 84, 86 (1948); Chamberlain Corp., 75 N.L.R.B. 1188, 1190 (1948). The “good faith” language had previously been utilized by multiple circuit courts. E.g., NLRB v. Crown Can Co., 138 F.2d 263, 266–67 (8th Cir. 1943); Lebanon Steel Foundry v. NLRB, 130 F.2d 404, 409 (D.C. Cir. 1942); NLRB v. Chi. Apparatus Co., 116 F.2d 753, 758 (7th Cir. 1940). This was cemented as the agency’s test for bargaining order remedies in 1949 when, by way of the Joy Silk case, the Labor Board finally corralled its case law:

We have previously held that an employer may in good faith insist on a Board election as proof of the Union’s majority but that it “unlawfully refuses to bargain if its insistence on such an election is motivated, not by any bona fide doubt as to the union’s majority, but rather by a rejection of the collective bargaining principle or by a desire to gain time within which to undermine the union.” In cases of this type the question of whether an employer is acting in good or bad faith at the time of the refusal is, of course, one which of necessity must be determined in the light of all relevant facts in the case, including any unlawful conduct of the employer, the sequence of events, and the time lapse between the refusal and the unlawful conduct.67Joy Silk Mills, Inc., 85 N.L.R.B. 1263, 1264 (1949) (quoting Artcraft Hosiery, 78 N.L.R.B. at 334).

This synthesis of the good-faith doubt test would receive fulsome approval from the reviewing District of Columbia Circuit.68Joy Silk Mills, Inc. v. NLRB, 185 F.2d 732, 741–42 (D.C. Cir. 1950). While stipulating that the NLRB’s election processes may provide the good-faith employer “a determination of whether or not the union does in fact have a majority and is therefore the appropriate agent with which to bargain,” the D.C. Circuit—citing Remington Rand—endorsed the Board’s chief policy argument: “it is not one of the purposes of the election provisions to supply an employer with a procedural device by which he may secure the time necessary to defeat efforts toward organization being made by a union.”69Id. at 741. The court proceeded to enforce the Labor Board’s bargaining order despite the union having lost a secret-ballot election, holding that the remedy in this context was “amply sustained by precedent” in the Bradford Dyeing, P. Lorillard Co., and Frank Bros. trilogy of cases.70Id. at 744–45 (citing all three cases and quoting Frank Bros. at length). Moreover, the court determined there was substantial evidence in the record to support the Board’s finding that the employer lacked good faith doubt of the union’s majority at the time of the original request for recognition and thus had unlawfully refused to bargain under Section 8(a)(5).71Id. at 741–42. The NLRB’s findings were based upon the employer’s violation of Section 8(a)(1) when—12 days after rejecting a proffering of authorization cards from the union indicating support from 38 of the mill’s 52 employees, and five days after negotiating a consent election with the union—a supervisor unlawfully promised that the employer would implement a compensated rest period and shift rotation in a pair of captive audience meetings, amidst more muted anti-union campaigning. Id. at 736–37, 739–41. Given the abbreviated time lapse of these events, the D.C. Circuit held it was reasonable for the NLRB to conclude that “the employer did not suddenly suffer a change of heart.” Id. at 742. Those only familiar with the Gissel brand of bargaining orders will likely be surprised to see such an order sustained on the back of a comparatively light infraction: a single supervisor’s unlawful promise of benefits.

The D.C. Circuit’s ratification of the good-faith doubt test spread like wildfire across its sister circuits. Such Joy Silk bargaining orders, as they soon became known, would be enforced in every extant federal court of appeals within a decade, usually in perfunctory fashion.72See, e.g., NLRB v. Lunder Shoe Corp., 211 F.2d 284, 288 (1st Cir. 1954); NLRB v. Pyne Molding Corp., 226 F.2d 818, 821 (2d Cir. 1955); NLRB v. Epstein, 203 F.2d 482, 484 (3d Cir. 1953); NLRB v. Inter-City Advert. Co., 190 F.2d 420, 422 (4th Cir. 1951); NLRB v. Stewart, 207 F.2d 8, 13 (5th Cir. 1953); NLRB v. Model Mill Co., 210 F.2d 829, 829–30 (6th Cir. 1953); NLRB v. Taitel, 261 F.2d 1, 4–5 (7th Cir. 1958); NLRB v. Wheeling Pipe Line, Inc., 229 F.2d 391, 393 (8th Cir. 1956); NLRB v. W. T. Grant Co., 199 F.2d 711, 712 (9th Cir. 1952); NLRB v. Burton-Dixie Corp., 210 F.2d 199, 201 (10th Cir. 1954). The core tenets of the Joy Silk doctrine would even gain tacit acceptance in the Supreme Court in 1956 when, in United Mine Workers v. Arkansas Oak Flooring Co., a seven-Justice majority cited a host of good-faith doubt cases for the proposition that “[a] Board election is not the only method by which an employer may satisfy itself as to the union’s majority status.”73351 U.S. 62, 72 n.8 (1956). The Court also more elaborately tied the “designated or selected” language in Section 9(a) to the nonelection context than it had in its previous bargaining order cases, finding that the union’s cards signed by 179 employees in the 225-person unit constituted a “conceded majority designation of the union.”74Id. at 68 n.2, 75. Crucially, if the union in the case had been eligible under the Labor Board’s unfair labor practice jurisdiction, “the absence of any bona fide dispute as to the existence of the required majority of eligible employees” meant that “the employer’s denial of recognition of the union would have violated § 8 (a)(5) of the Act.”75Id. at 69 (emphasis added). The union in Arkansas Oak Flooring had refused to file the anti-communist affidavits necessary at the time to take advantage of the NLRB’s processes. Id. at 66. See also Brooks v. NLRB, 348 U.S. 96, 103–104 (1954) (stating that “a bargaining agency may be ascertained by methods less formal than a supervised election”).

By the end of the 1950s the good-faith doubt test had been approved by every court possessing the plenary power to review it, an incredible achievement for the NLRB given the mixed success of the agency’s remedial experiments under Section 10(c).76National Labor Relations Act § 10(c), 29 U.S.C. § 160(c) (2018) (empowering the NLRB “to take such affirmative action . . . as will effectuate the policies of this Act”). See generally Michael Weiner, Comment, Can the NLRB Deter Unfair Labor Practices? Reassessing the Punitive-Remedial Distinction in Labor Law Enforcement, 52 UCLA L. Rev. 1579, 1590–1617 (2005). Whatever one thinks regarding the merits of the test, it is clear that this was not a hastily constructed remedy foisted unilaterally upon the labor-management community by a zealous government agency. Instead, the good-faith doubt test was the consequence of years of deliberation between the Labor Board and the courts, spearheaded by one of the most esteemed jurists in all of American law77For the quintessential Judge Hand biography, see generally Gerald Gunther, Learned Hand: The Man and the Judge (1994). and bolstered by several unanimous Supreme Court decisions. The Joy Silk doctrine had been delivered into the world, and by any reasonable measure, it was a celebrated birth.

C. The Kennedy Board Fortifies the Joy Silk Doctrine

Though the good-faith doubt test encountered no discernible pushback in its early years, Joy Silk became limited in its usefulness as an organizing tactic not long into its lifespan. In what would become known as the Aiello doctrine, the Eisenhower Board prohibited the application of bargaining orders in cases where unions lost an election after previously alleging an unlawful refusal to recognize under Section 8(a)(5), restricting the remedy in such scenarios to a rerun of an election that had been set aside due to employer misconduct.78Aiello Dairy Farms, 110 N.L.R.B. 1365 (1954). Aiello itself overruled contrary precedent in The M. H. Davidson Co., 94 N.L.R.B. 142 (1951). The inevitable result was that a union whose request for recognition had been rejected by an employer lacking a good faith doubt in the union’s majority status was forced to choose between two unenviable options: (1) forego its unfair labor practice charge and risk the union’s majority in a potentially tainted election,79Of course, the lack of a bargaining order remedy in these cases greatly incentivizes employer misconduct during the election process. See infra Section II.B. or (2) cease its organizing efforts and proceed to trial, effectively putting its fate in the hands of NLRB decisionmakers amidst the often years-long process of unfair labor practice litigation.

While some justified Aiello under the logic of depriving unions a second bite at the Board’s administrative apple,80See Harry L. Browne, Labor Board Unsettles the Scales, 42 Notre Dame L. Rev. 133, 145–46 (1967). But see Howard Lesnick, Establishment of Bargaining Rights Without an NLRB Election, 65 Mich. L. Rev. 851, 856 n.22 (1967) (“Objections about giving the union ‘two bites at the apple’ hardly seem to prove much; after all, when the other fellow has put a worm in the apple, it is hardly going very far to allow a second bite.”). the decision to arbitrarily withdraw Section 8(a)(5)’s protections from unions based upon their decision to engage with the agency’s election machinery makes little sense as a companion to Joy Silk, which rests upon the foundation that an employer’s conduct following its refusal to recognize an organizing union may shed light on the nature of the refusal itself. Under this view, any required waiver of a meritorious Section 8(a)(5) charge derogated from the NLRB’s duty to enforce public policy in the form of good-faith bargaining.81Aiello Dairy Farms, 110 N.L.R.B. at 1374–75 (Member Peterson, dissenting). This matter was wisely resolved during labor law’s stage of infancy:The Board asserts a public right vested in it as a public body, charged in the public interest with the duty of preventing unfair labor practices. The public right and the duty extend not only to the prevention of unfair labor practices by the employer in the future, but to the prevention of his enjoyment of any advantage which he has gained by violation of the [NLRA] . . . as the means of defeating the statutory policy and purpose.
National Licorice Co. v. NLRB, 309 U.S. 350, 364 (1940).
Aiello remained a hitch in Joy Silk’s gait from 1954 until 1964, when the Kennedy Board removed these self-imposed limitations in the case of Bernel Foam Products Co.82146 N.L.R.B. 1277 (1964). The Bernel Foam majority’s dissatisfaction with the rerun election remedy was supported by a recent study which found that 69% of rerun elections were won by the party whose misconduct interfered with the original election. Daniel H. Pollitt, NLRB Re-Run Elections: A Study, 41 N.C. L. Rev. 209, 212 (1963). The restoration of both the representation and unfair labor practice route as navigable channels for bargaining orders, previously urged by Congress,83A special House of Representatives labor subcommittee, chaired by Representative Roman Pucinski, recommended early in the Kennedy administration that the NLRB “give serious consideration to revitalizing the so-called Joy Silk doctrine and ordering certification and recognition of a union as an appropriate remedy to dissipate the adverse effects on union majority status of employer interference with the right of self-organization.” H.R. Rep. No. 87-75142, at 24 (1961); see also James. A. Gross, Broken Promise: The Subversion of U.S. Labor Relations Policy, 19471994, at 18384 (1995). Chairman Frank W. McCullough would later credit the Pucinski Committee for giving the NLRB “useful nudges” in the direction of “revitalizing the Joy Silk doctrine.” Oversight Hearings on the National Labor Relations Board: Hearings Before the Subcomm. on Labor-Management Relations of the H. Comm. on Education and Labor, 94th Cong. 256 (1975). naturally increased the prominence of the good-faith doubt test in the labor organizing terrain.84In something of a house-keeping role, the Kennedy Board quickly clarified that Joy Silk bargaining orders could only issue on behalf of unions which had lost an election if that election had been set aside on the basis of meritorious objections. Kolpin Bros. Co., 149 N.L.R.B. 1378, 1379–80 (1964); Irving Air Chute Co., Inc., 149 N.L.R.B. 627, 629–30 (1964).

Two other developments in the early 1960s functioned as legal incubators for Joy Silk cases. First, the Kennedy Board developed the “independent knowledge” variant of the good-faith doubt analysis in Snow & Sons, which held that a bargaining order would issue even without the presence of independent unfair labor practices calculated to dissipate the union’s majority if the employer had confirmed the union possessed majority support but still refused recognition.85134 N.L.R.B. 709 (1961), enforced, 308 F.2d 687 (9th Cir. 1962). The NLRB determined that the employer discriminatorily refused to reinstate nine of its employees which had gone on strike upon the refusal to bargain, reversing the trial examiner on this Section 8(a)(3) allegation, but that finding was extraneous to the issuance of the bargaining order under Section 8(a)(5). Id. at 711. There, the employer reneged on an agreed-upon check of authorization cards after a third party authenticated the union’s majority status and insisted upon testing the results in an NLRB election, even though the employer did not doubt the accuracy or propriety of the card check.86Id. at 710. While the Labor Board had never explicitly confined Joy Silk to cases in which separate violations of the NLRA were present, the practical reality of litigation meant that it was difficult for Board attorneys to prove that an employer lacked good faith doubt of a union’s majority status without the commission of unfair labor practices following the refusal to recognize.87A court will defer to the NLRB’s findings of fact only if they are supported by “substantial evidence on the record considered as a whole.” Labor Management Relations Act, 1947 (Taft-Hartley) § 10(e), 29 U.S.C. § 160(e) (2018); accord Universal Camera Corp. v. NLRB., 340 U.S. 474 (1951). Snow & Sons thus represented the Board’s perhaps mechanistic commitment to prosecuting the rare case in which an employer, by its own actions, vitiated itself of asserting any defense of uncertainty regarding its employees’ sentiments without engaging in conduct proscribed under Section 8(a)(1) or (3) of the Act.88Snow & Sons would soon serve as the basis for the frequent complaint that “an employer could close his eyes” and “had no obligation to look around him. But if, by chance, he opened his eyes and found out something, then he could have an obligation to bargain[.]” Transcript of Oral Argument at 56–57, Linden Lumber Div., Summer & Co. v. NLRB, 419 U.S. 301 (1974) (No. 73-1231).

Second, the Kennedy Board moved to reinstate authorization cards as a generally valid barometer of employees’ designation or selection of their bargaining representative for purposes of Section 9(a). Despite the Supreme Court’s unflinching approval of a card-based bargaining order in Frank Bros. and the prominent view that certification-only bargaining was a “frivolous” interpretation of the NLRA,89See, e.g., Dahlstrom Metallic Door Co., 112 F.2d 756, 757 (2d Cir. 1940). the Eisenhower Board had deemed authorization cards “unreliable as evidencing the employees’ considered desires” and strove to hold elections barring “extraordinary circumstances.”90Se. Rubber Mfg. Co., 106 N.L.R.B. 989, 994 (1953) (Member Farmer, dissenting); accord A. L. Gilbert Co., 110 N.L.R.B. 2067, 2069 (1954) (declaring that “the employer may refuse to rely upon evidence of representation in the form of authorization cards signed by employees and insist that the union prove its majority in a secret election” unless the employer ran afoul of Joy Silk). The Eisenhower Board minority considered this an unstated limitation of the Joy Silk doctrine. A. L. Gilbert Co., 110 N.L.R.B. at 2078 (Members Murdock and Peterson, concurring in part and dissenting in part); see also Sunset Lumber Prods., 113 N.L.R.B. 1172, 1179 (1955) (Members Murdock and Peterson, concurring in part and dissenting in part). For an analysis concluding that the Eisenhower Board majority narrowed the good-faith doubt test through its skepticism of authorization cards, see Helen F. Humphrey, The Duty to Bargain, 16 Ohio St. L.J. 403, 409–414 (1955). The Kennedy Board not only rejected this rather pessimistic view of the average employee’s intelligence but also presumed that workers made rational and informed choices when affixing their name to an explicit request for collective bargaining.91See Congressional Oversight of Administrative Agencies (National Labor Relations Board): Hearings Before the Subcomm. on Separation of Powers of the S. Comm. on the Judiciary, 90th Cong. 1669 (1968) (memorandum from Chairman Frank McCulloch) [hereinafter 1968 Hearings]. In Cumberland Shoe Corp., the NLRB held that an unambiguous card designating the union as the signed employee’s bargaining representative will be counted as a valid designation unless it is proven that the employee was told that the card’s sole purpose was to be used to obtain a secret-ballot election.92144 N.L.R.B. 1268 (1963), enforced, 351 F.2d 917 (6th Cir. 1965). The attractiveness of the Cumberland Shoe doctrine—which broadly permitted the campaign tactics unions had already engaged in for decades—was mostly in its practicality; it avoided the card-by-card adjudication of probing into the subjective intent of each signer which threatened to bog down trials held months or even years after the initial gathering of the card.93Cumberland Shoe Corp., 351 F.2d at 919 n.3. The Kennedy Board’s decision again hewed faithfully to the Joy Silk doctrine, in which the D.C. Circuit stated that “an employee’s thoughts (or afterthoughts) as to why he signed a union card, and what he thought that card meant, cannot negative the overt action of having signed a card designating a union as bargaining agent.” 185 F.2d 732, 743 (D.C. Cir. 1950).

The combined ramification of the Kennedy Board’s Bernel Foam, Cumberland Shoe, and Snow & Sons decisions was that a union could obtain a bargaining order on the basis of signed authorization cards even if it had lost an election or, in the exceptional case, the employer had not committed independent unfair labor practices—in other words, exactly where the law stood in the early years of the Taft-Hartley Act when the good-faith doubt test was forged.94See Peoples Motor Express, Inc. v. NLRB, 165 F.2d 903 (4th Cir. 1948) (enforcing a bargaining order under Section 8(a)(5) based upon a card majority despite the union having lost an election); NLRB v. Cullen, 201 F.2d 369 (10th Cir. 1953) (per curiam) (enforcing a bargaining under Section 8(a)(5) despite no independent unfair labor practices committed by the employer). But in the years between Joy Silk’s inception and its restoration, the labor movement had become unified and refocused,95See Gross, supra note 83, at 13739 (discussing the AFL-CIO’s merger in terms of the shared goal of favorable labor law reform). aware of the legal advantages it had lost during the Eisenhower administration and willing to press its rights before the NLRB. Unions increasingly organized around a recognition-request strategy, and Joy Silk-related litigation ate an increasing share of the Board’s docket throughout the 1960s: card-based bargaining orders increased from 35 in 1962 to 107 in 1967,96Id. at 359 n.146. while the portion of Board-mandated bargaining relationships originating in unfair labor practice cases compared to secret-ballot elections more than tripled in this span.97Id. at 359 n.145. The strain of this increasing workload was also felt in the NLRB’s operations. As General Counsel Arnold Ordman reported in 1967:The processing of Section 8(a)(5) cases represents a substantial workload factor since they normally involve complex fact and legal issues and require more Regional Office time and manpower to investigate and process to completion. This is particularly true because the court-approved Joy Silk and Bernel Foam doctrines have given rise to an increasing number of cases requiring proof of majority through authorization cards. Investigation of the union’s majority in such cases has imposed further manpower requirements and increased the time to investigate such cases.
Office of the General Counsel, National Labor Relations Board: Summary of Operations 15 (1967).
It is against this backdrop that the ensuing criticism of the Joy Silk doctrine was framed.

D. Criticism and Modification of the Good-Faith Doubt Test

The Kennedy Board’s reinvigoration of the bargaining order remedy created a firestorm of controversy within labor law discourse. The Bernel Foam and Cumberland Shoe doctrines were especially savaged by management-side practitioners, who hurled accusations at the Board of pro-union bias and anti-election subterfuge.98See, e.g., Robert Lewis, The Use and Abuse of Authorization Cards in Determining Union Majority, 16 Lab. L.J. 434 (1965); Louis Shuman, Requiring a Union to Demonstrate Its Majority Status by Means of An Election Becomes Riskier, 16 Lab. L.J. 426 (1965). Derek Bok, a respected professor from Harvard Law School, published an influential article on NLRB election law which questioned the use of bargaining orders in all but the most flagrant cases of employer lawlessness, opining that “[u]nions should not be permitted to seize upon trivial violations in order to obtain bargaining rights without an election.”99Derek C. Bok, The Regulation of Campaign Tactics in Representation Elections Under the National Labor Relations Act, 78 Harv. L. Rev. 38, 137–38 (1964). A management attorney by the name of Woodrow Sandler placed a widely-read opinion piece in the U.S. News & World Report which featured a laundry list of complaints with the Kennedy Board’s application of the Joy Silk doctrine and called for legislation to restrict management’s duty to bargain to unions that obtained certification through a secret-ballot election.100Woodrow J. Sandler, Another Worry for Employers, U.S. News & World Report, March 15, 1965, reprinted in To Repeal Section 14(b) of the National Labor Relations Act: Hearings on S. 256 Before the Subcomm. on Labor of the S. Comm. on Labor and Public Welfare, 89th Cong. 190–94 (1965) [hereinafter Section 14(b) Hearings]. By the summer of 1965, the legislative campaign to repeal Taft-Hartley’s Section 14(b) “right to work” provision was being hijacked by the introduction of legislation intended to limit the Labor Board’s ability to issue Joy Silk bargaining orders.101See 111 Cong. Rec. 20052 (daily ed. Aug. 11, 1965) (statement of Sen. Javits); 111 Cong. Rec. 15124 (daily ed. June 29, 1965) (statement of Sen. Fannin). Chairman Frank McCulloch was grilled by Senators Fannin and Javitz just a few weeks prior on the state of the Joy Silk doctrine during his reappointment hearing. Nomination of Frank W. McCulloch, of Illinois, to be a Member of the National Labor Relations Board (Reappointment): Hearing Before the S. Comm. on Labor and Public Welfare, 89th Cong. 4–16, 36–41 (1965).

The NLRB, seemingly caught off guard by this backlash, made its first concession. In Hammond & Irving, decided mere months after introduction of that hostile legislation, the Labor Board refused to issue a bargaining order in a case where the employer unlawfully interrogated six employees out of the 110-person unit.102Hammond & Irving, Inc., 154 N.L.R.B. 1071, 1072 (1964). The Board determined that the employer committed its unfair labor practice less than a month after it refused to bargain with the union’s card-based majority and in the final days before the scheduled election.103Id. Even though these factors generally weighed heavily in the Joy Silk analysis,104Joy Silk Mills, Inc., 85 N.L.R.B. 1263, 1264 (1949) (“the question of whether an employer is acting in good or bad faith at the time of the refusal is, of course, one which of necessity must be determined in the light of all relevant facts in the case, including any unlawful conduct of the employer, the sequence of events, and the time lapse between the refusal and the unlawful conduct.” (emphasis added)). the Board reasoned that “not every act of misconduct necessarily vitiates the respondent’s good faith” and “[t]his interrogation, while unlawful, was not so flagrant that it must necessarily have had the object of destroying the Union’s majority status.”105Hammond & Irving, 154 N.L.R.B. at 1073. For this proposition, the NLRB curiously cited to a footnote in the trial examiner’s decision of a case in which a bargaining order had issued. Cosmodyne Mfg. Co., 150 N.L.R.B. 96, 104 n.29 (1965). The good-faith doubt test, which had more or less operated as a per se rule for unfair labor practices in non-Eisenhower years, suddenly was seeing shades of gray.

The NLRB went even further a month later in deciding John P. Serpa, Inc.106155 N.L.R.B. 99 (1965). Here, the employer did not respond to the union’s request for recognition; when prodded again by the union, the employer simply stated it had no intention of recognizing the union, drawing a refusal to bargain charge.107Id. at 100. The employer did not petition the NLRB for an election until weeks after the Section 8(a)(5) charge was filed. Id. at 103. While the employer did not commit any independent unfair labor practices, its failure to affirmatively express any doubt about the union’s majority status or file for an election suggested that a bargaining order was due under Snow & Sons.108Id. at 103. Instead, the Labor Board held that Section 8(a)(5) did not reach the described conduct because there was no evidence in the record “indicating that respondent has completely rejected the collective-bargaining principle” or sought “to gain time within which to undermine the union and dissipate its majority.”109Id. at 101. More, it was now the Board attorneys’ burden to prove bad faith on the part of the employer; the mere fact that the union had presented its cards to the employer for inspection could not alone create the obligation to bargain or establish the employer’s bad faith.110Id. The NLRB, in distinguishing Snow & Sons, appeared to isolate its applicability going forward to cases in which an employer rejected a previously agreed-upon means of determining the union’s majority status. Id. at 101 n.4.

Despite the understated tone of the John P. Serpa decision, this act of burden-shifting represented a tectonic shift in Board law. It had been well understood since the Dahlstrom case that “adequate proof” of a card-based majority “could not in good faith be ignored[,]” and that an employer which rejected such presentations and “made no effort to learn the facts” of the union’s majority was guilty of an unlawful refusal to bargain.111NLRB v. Dahlstrom Metallic Door Co., 112 F.2d 756, 757 (2d Cir. 1940) (citing NLRB v. Remington Rand, Inc., 94 F.2d 862, 869 (2d Cir. 1938)). It is thus hardly surprising that the reviewing court of the John P. Serpa decision reversed the Board’s dismissal of the complaint and instructed the NLRB to issue a bargaining order, citing its own enforcement of Snow & Sons as controlling. Retail Clerks Union, etc. v. NLRB, 376 F.2d 186, 190 (9th Cir. 1967). The Ninth Circuit ruled that the employer had an affirmative duty to respond to the union’s recognition request after consulting with its attorney; the decision to silently wait out the union represented a “desire to gain time and to take action to dissipate the union’s majority” under Joy Silk. Id. at 191. Perhaps realizing that a more definitive pronouncement in this area was required, the NLRB moved to codify its new framework in the Aaron Bros. case:

While an employer’s right to a Board election is not absolute, it has long been established Board policy that an employer may refuse to bargain and insist upon such an election as proof of a union’s majority unless its refusal and insistence were not made with a good-faith doubt of the union’s majority. . . . Absent an affirmative showing of bad faith, an employer, presented with a majority card showing and a bargaining request, will not be held to have violated his bargaining obligation under the law simply because he refuses to rely upon cards, rather than an election, as the method for determining the union’s majority. . . . [I]t is the General Counsel who must come forward with evidence and affirmatively establish the existence of such bad faith.112Aaron Bros. Co., 158 N.L.R.B. 1077, 1078–79 (1966).

The Labor Board further raised the standard for violations which could garner a bargaining order; the General Counsel would now have to demonstrate that the employer engaged in either “substantial unfair labor practices calculated to dissipate union support” or in the rare Snow & Sons-type situation involving a bad-faith takeback.113Id. at 1079 (emphasis added).

Although the NLRB professed loyalty to the Joy Silk doctrine in transferring the burden of proof from employers to the General Counsel, consider the nuts-and-bolts impact of this modification for Board prosecutors. Where a Board attorney could previously set forth the facts of the employer’s obstinate conduct and wait to rebut any assertions of good faith doubt, they were now required to present evidence of the employer’s lack of good faith up front. This essentially limited the cases in which the Board could make a prima facie case at trial for a Joy Silk bargaining order to Snow & Sons scenarios, Section 8(a)(3) discriminatory actions, and serious Section 8(a)(1) violations.114The Aaron Bros. decision made clear that the “substantial” language would eliminate at least some Section 8(a)(1) unfair labor practices as criteria for bargaining orders: “this does not mean that any employer conduct found violative of Section 8(a)(1) of the Act, regardless of its nature or gravity, will necessarily support a refusal-to-bargain finding.” Id. (emphasis added). Now an employer would no longer violate Section 8(a)(5) merely because it refused to rely on authorization cards, and it would not even need to advance any reasons for rejecting a union’s request for recognition. The subjective-motivation analysis intrinsic in Joy Silk thus threatened to collapse in on itself where the evidentiary burden was placed on the party having to prove the other’s state of mind. Board Member Howard Jenkins perceptively observed in his concurrence that the effect of the majority’s decision was to require an affirmative showing of bad faith rather than merely the absence of good faith.115Id. at 1081 (Member Jenkins, concurring).

Despite this acquiescence, commentators continued to excoriate the good-faith doubt test as administered by the NLRB, coming in the form of law review articles,116See Pogrebin, supra note 21, at 193–98; Andrea Svanoe Christensen & Thomas G. S. Christensen, Gissel Packing and “Good Faith Doubt”: The Gestalt of Required Recognition of Unions under the NLRA, 37 U. Chi. L. Rev. 411, 433–49 (1970); Note, NLRB v. Gissel Packing Co.: Bargaining Orders and Employee Free Choice, 45 N.Y.U. L. Rev. 318, 321–32 (1970); Comment, Labor Law – Authorization Cards Held to Be a Legitimate Alternative to an NLRB Election When a Fair Election is Improbable Because of Employer’s Unfair Labor Practices, 2 Rutgers Camden L.J. 201, 208–09 (1970) [hereinafter Rutgers Comment]; Melvin J. Welles, The Obligation to Bargain on the Basis of a Card Majority, 3 Ga. L. Rev. 349, 355–57 (1969); Harry L. Browne, Obligation to Bargain on Basis of Card Majority, 3 Ga. L. Rev. 334, 340–48 (1969); Jacob Sheinkman, Recognition of Unions Through Authorization Cards, 3 Ga. L. Rev. 319, 328–29 (1969); Robert J. Affeldt, Bargaining Orders Without An Election: The National Labor Relations Board’s “Final Solution”, 57 Ky. L.J. 151, 204–14 (1968); Comment, Employer “Good Faith Doubt”, 116 U. Pa. L. Rev. 709 (1968) [hereinafter Pennsylvania Comment]; Frederic T. Spindel, Comment, Union Authorization Cards: A Reliable Basis for an NLRB Order to Bargain, 47 Tex. L. Rev. 87, 103–04 (1968); Browne, supra note 80; L. Edwin Greer, Comment, Labor Law – Employer’s Good Faith Doubt of Union’s Majority, 27 La. L. Rev. 564, 57376 (1967); Lesnick, supra note 80, at 858–60; Robert Lewis, NLRB again Stretches Joy Silk’s Tenuous Fabric, 18 Lab. L.J. 222, 222–23 (1967); Comment, Refusal-To-Recognize Charges Under Section 8(a)(5) of the NLRA: Card Checks and Employee Free Choice, 33 U. Chi. L. Rev. 387, 397–405 (1966) [hereinafter Chicago Comment]; Note, Union Authorization Cards, 75 Yale L.J. 805, 828–36 (1966) [hereinafter Yale Note]; Robert P. Duvin, The Duty to Bargain: Law in Search of Policy, 64 Colum. L. Rev. 248, 267–68 (1964). This footnote constitutes my best attempt at harvesting all contemporary published skepticism or outright denunciations of the good-faith doubt test as articulated in Joy Silk. congressional testimony,117E.g., 1968 Hearings, supra note 91, at 79–80, 172–78, 614–620, 730–32, 1025–27, 1070–79, 1162–65. and—for the first time since the Joy Silk doctrine gained widespread acceptance in the courts—a spate of judicial opinions, especially from the once sympathetic Second Circuit.118E.g., NLRB v. United Min. & Chem. Corp., 391 F.2d 829, 837–38 (2d Cir. 1968); NLRB v. S. S. Logan Packing Co., 386 F.2d 562, 569–71 (4th Cir. 1967); NLRB v. River Togs, Inc., 382 F.2d 198, 206–208 (2d Cir. 1967); NLRB v. S.E. Nichols Co., 380 F.2d 438, 441–45 (2d Cir. 1967); Peoples Service Drug Stores, Inc. v. NLRB, 375 F.2d 551, 556–58 (6th Cir. 1967); NLRB v. Gotham Shoe Mfg. Co., 359 F.2d 684, 699–703 (2d Cir. 1966) (Timbers, J., dissenting); NLRB v. Flomatic Corp., 347 F.2d 74, 77–80 (2d Cir. 1965). To be sure, criticism of the good-faith doubt test did not originate in the Kennedy Board’s Bernel Foam and Cumberland Shoe doctrines. E.g., NLRB v. Dan River Mills, Inc., 274 F.2d 381, 386–89 (5th Cir. 1960); NLRB v. Hannaford Bros., 261 F.2d 638, 640–41 (1st Cir. 1958); NLRB v. James Thompson & Co., 208 F.2d 743, 745–46 (2d Cir. 1953) (expressing early skepticism of the Joy Silk framework). Much of the venom directed at Joy Silk stemmed from its role in facilitating the use of the “dual-purpose” authorization cards sanctioned in Cumberland Shoe. The Fourth Circuit’s sudden broadside against the Labor Board’s treatment of card-based bargaining orders, commenced by Judge Clement Haynsworth’s opinion in the infamous Logan Packing case, served as the clarion call which would soon necessitate Supreme Court review:

It would be difficult to imagine a more unreliable method of ascertaining the real wishes of employees than a “card check,” unless it were an employer’s request for an open show of hands. The one is no more reliable than the other. No thoughtful person has attributed reliability to such card checks. . . . Though the card be an unequivocal authorization of representation, its unsupervised solicitation may be accompanied by all sorts of representations. “We need these cards to get an election. You believe in the democratic process, don’t you? Do you want to deny people the right to vote? Isn’t it our American way to resolve questions at the polls? Do you want to deprive us of that right? Are you a Hitler or something?” . . . The unreliability of the cards . . . . is inherent . . . in the absence of secrecy and in the natural inclination of most people to avoid stands which appear to be nonconformist and antagonistic to friends and fellow employees.119S. S. Logan Packing Co., 386 F.2d at 565–66. Included in this list of so-called thoughtless people were, apparently, many Justices of the Supreme Court of the United States, which had unanimously enforced several bargaining orders that relied on card-based majorities. See supra Part I.A.

However, significant criticism was levied against the substantive assumptions undergirding the NLRB’s view of good faith, independent of the issues assailing Cumberland Shoe. Judge Haynsworth argued in Logan Packing that an employer’s violations, even if “extreme conduct,” are merely “a neutral item of evidence” in an inquiry into its motivation, for “subsequent unfair labor practices have a tendency to prove only the employer’s opposition to the union’s organizational effort; they throw no light on his belief or disbelief of the union’s claim of majority status.”120S. S. Logan Packing Co., 386 F.2d at 568. He further interpreted the Taft-Hartley amendments to somehow prevent the issuance of a bargaining order under Section 8(a)(5) alone (as opposed to remedying “outrageous and pervasive” unfair labor practices under Section 8(a)(1) or (3)), despite Congress having altered neither Section 8(a)(5) nor the “designated or selected” language of Section 9(a) in 1947.121Id. at 569–70. Judge Haynsworth’s reasoning was likely already foreclosed by the majority opinion in United Mine Workers v. Arkansas Oak Flooring Co., 351 U.S. 62 (1956), see supra notes 73–75 and accompanying text, but it was explicitly rejected by the Supreme Court in NLRB v. Gissel Packing Co., Inc., 395 U.S. 575 (1969), infra notes 132–136 and accompanying text. Snow & Sons bargaining orders would thus not be enforced in the Fourth Circuit, and the barrier for independent unfair labor practices was raised above even the “substantial” intensifier in Aaron Bros.

There were courts that made Judge Haynsworth’s argument more persuasively and without succumbing to his invective.122It is not a coincidence that Judge Haynsworth’s nomination to the Supreme Court in 1969 failed in large part due to vociferous opposition by organized labor, which called attention to his perfect record of reversals in the Supreme Court on all seven of his labor law opinions that were granted certiorari. His trio of 1968 opinions denying enforcement of Joy Silk bargaining orders represented the most recent three of this resume. See 115 Cong. Rec. 34428 (daily ed. Nov. 17, 1969) (statement of Sen. Metcalf) (citing General Steel Prods., Inc., v. NLRB, 398 F. 2d 339 (4th Cir. 1968); NLRB v. Heck’s, Inc., 398 F. 2d 338 (4th Cir. 1968); NLRB v. Gissel Packing Co., 398 F.2d 336 (4th Cir. 1968)). The eminent Judge Henry Friendly arguably did more than anyone to change the narrative surrounding the good-faith doubt test, sowing the fertile grounds for revision from which many scholarly attacks would seed. In NLRB v. River Togs, Judge Friendly lambasted the “cursory” treatment the Labor Board gave to good-faith doubt allegations and moved to sever the link between an employer’s misconduct and its concomitant duty to bargain:

[W]e see no logical basis for the view that substantial evidence of good faith doubt is negated solely by an employer’s desire to thwart unionization whether by proper or even by improper means. . . . [the employer’s] efforts to counter the Union . . . were “as consistent with a desire to prevent the acquisition of majority status as with a purpose to destroy an existing majority.”123NLRB v. River Togs, Inc., 382 F.2d 198, 206–07 (2d Cir. 1967) (quoting Lesnick, supra note 80, at 855).

Judge Friendly astutely observed that while the NLRB continued to speak in terms of good-faith doubt, the agency was actually conditioning bargaining orders on whether the employer’s unlawful conduct “made a fair election impossible.”124Id. at 207. The Aaron Bros. decision demonstrated the Labor Board’s retreat from its role in championing an employer’s bargaining obligations, now simply guaranteeing “the most reliable means available to ascertain the true desires of employees with respect to the selection of a collective-bargaining representative”; like a sliding scale, an employer’s unfair labor practices cautioned the Board against operating its election machinery, but only if the employer’s practices were of sufficient severity.125Id. at 207–208 (quoting Aaron Bros. Co., 158 N.L.R.B. 1077, 1079 n.10 (1966)). Put another way, violations of the NLRA now shed light only on the pertinence of the government’s procedures rather than what duties an employer owed its employees at the time the violations were committed.126Id. at 208 (“[U]nfair labor practices directed at undermining the union will be given effect not so much as evidence effectively contradicting existence of the doubt, as to which their probative force is often slight, but rather as a violation . . . so serious as to make a bargaining order an appropriate remedy.”). Good faith doubt emerges as little more than a fig leaf from Judge Friendly’s review.

This development of the law tracked the overwhelming consensus of the labor bar. Virtually every legal commentator who addressed the matter of card-based bargaining relationships in the 1960s urged the NLRB to abandon its good-faith doubt test and resist issuing bargaining orders except where an employer’s unfair labor practices were of the kind that preclude a fair election, including law professors,127Affeldt, supra note 116, at 164; Bok, supra note 99; Christensen & Christensen, supra note 116, at 447–49; Lesnick, supra note 80, at 860. law students,128Chicago Comment, supra note 116, at 400–401; Greer, supra note 116, at 573–74; Pennsylvania Comment, supra note 116, at 729–31; Spindel, supra note 116; Yale Note, supra note 116, at 841. NLRB personnel,129Welles, supra note 116, at 357–59; Piggly Wiggly El Dorado Co., 154 N.L.R.B. 445, 445 n.1 (1965) (rebuking Trial Examiner’s criticism of NLRB case law as “unwarranted” and “extraneous”); Laura J. Cooper & Dennis R. Nolan, The Story of NLRB v. Gissel Packing: The Practical Limits of Paternalism, in Labor Law Stories 207 (Catherine Fisk & Laura J. Cooper eds., 2005) (describing an NLRB attorney’s desire for agency to adopt theory that card-based bargaining orders could only issue “in cases in which employers engaged in such ‘flagrant coercive conduct’ that a fair election became impossible”). management-side attorneys,130Browne, supra note 80, at 146; Lewis, supra note 116; Sandler, supra note 100, at 193; 1968 Hearings, supra note 91, at 1074 (statement of Gerald C. Smetana). and even a union-side attorney.131Sheinkman, supra note 116, at 329. The merits of that position will be considered in Part III of this Article, but it is sufficient to say that the sheer volume of opprobrium flowing from law school libraries, law firm offices, federal court chambers, and even the halls of Congress suggested that Joy Silk’s status as a doctrinal linchpin had reached its breaking point. But the good-faith doubt test would have a last chance at immortality when, in late 1968, the Supreme Court granted certiorari on four bargaining-order cases and consolidated them for oral argument, three being Fourth Circuit denials of enforcement that directly challenged the NLRB’s authority to issue bargaining orders based on authorization cards.132NLRB v. Gissel Packing Co., Inc., 395 U.S. 575, 575 (1969). The fourth case included in the consolidation, The Sinclair Co. v. NLRB, 397 F.2d 157 (1st Cir. 1968), was another bargaining-order appeal that presented the additional issue of to what extent employer statements that unionization could result in plant closure are protected under Section 8(c) of the Taft-Hartley Act or the First Amendment to the United States Constitution. Labor Management Relations Act, 1947 (Taft-Hartley) § 8(c), 29 U.S.C. § 158(c) (2018); U.S. Const. amend. I. That issue, ultimately resolved against the petitioning employer, is beyond the scope of this Article.

E. The NLRB Abandons the Good-Faith Doubt Test in Gissel

At first glance, the Supreme Court’s decision in Gissel appeared to be a tremendous victory for the NLRB.133Gissel Packing, 395 U.S. 575. In a unanimous opinion written by Chief Justice Earl Warren, the Court upheld the Labor Board’s use of authorization cards in bargaining orders under the Taft-Hartley amendments, rejecting any contrary readings of Sections 8(a)(5) and 9(a);134Id. at 595–600. The Court found it dispositive that the Senate had rejected a House-proposed amendment which would have explicitly restricted Section 8(a)(5)’s reach to unions that had either been voluntarily recognized by the employer or certified by means of a secret-ballot election. Id. at 599. Having dispatched with the employers’ chief statutory argument, the Court added that Section 9(c)(1)(B) of the Taft-Hartley Act did not grant employers an absolute right to an election at any time. Id. at 591, 600. approved the Bernel Foam doctrine’s framework of post-election litigation for employer refusals to bargain;135See id. at 615 n.34. and endorsed the Cumberland Shoe doctrine as a proper safeguard of union solicitation.136Id. at 601–610. The academic and ideological attacks on authorization cards had failed; the Court’s proclamation that cards “can adequately reflect employee sentiment” did not even require an “extended discussion.”137Id. at 603. The decision vindicated the Kennedy Board’s most controversial forays into union representation law and Section 10(c) remedies.

Amidst this administrative triumph, one might also have expected the Court to ratify the NLRB’s twenty-year-old criteria for evaluating the issuance of bargaining orders. Incredibly, despite all parties having briefed the Joy Silk doctrine and its Aaron Bros. modification as controlling law,138Cooper & Nolan, supra note 129, at 207–208, 214. the Court noted early in its analysis that because the Labor Board “announced at oral argument that it had virtually abandoned the Joy Silk doctrine altogether[,]” “an employer’s good faith doubt is largely irrelevant.”139Gissel Packing, 395 U.S. at 594. This was a stunning error by either the agency’s representative at oral argument or the Court in its interpretation of the briefs, for the NLRB certainly had not “abandoned” the good-faith doubt test by way of adjudication as of the day it argued Gissel on March 26, 1969. Indeed, the Board issued a bargaining order just 23 days prior on the basis of the employer lacking a good faith doubt as to the union’s majority status,140Heck’s, Inc., 174 N.L.R.B. 951, 952 (1969). and it employed the test again on June 16, 1969, the very day that the Court decided Gissel.141Clay City Beverages, Inc., 176 N.L.R.B. 681, 682 (1969). The NLRB in this case even cited Joy Silk as authority. Id. at 682 n.3. The presenter of the NLRB’s position, Associate General Counsel Dominick L. Manoli, had worked at the agency for 28 years and was considered a skilled oral advocate,142See Cooper & Nolan, supra note 129, at 212–13. “(Between his first certiorari petition in 1945 and his last oral argument in 1971, Manoli’s name appeared on 646 documents filed in the Supreme Court.”). arguing many major cases of mid-century labor law.143E.g., Textile Workers Union v. Darlington Mfg. Co., 380 U.S. 263 (1965); NLRB v. Exchange Parts Co., 375 U.S. 405 (1964); NLRB v. Wash. Aluminum Co., 370 U.S. 9 (1962); Int’l Ladies’ Garment Workers’ Union v. NLRB (Bernhard-Altmann Tex. Corp.), 366 U.S. 731 (1961); NLRB v. Insurance Agents’ Int’l Union, 361 U.S. 477 (1960); NLRB v. United Steelworkers of America (Nutone & Avondale), 357 U.S. 357 (1958); NLRB v. Borg-Warner Corp., 356 U.S. 342 (1958); NLRB v. Truck Drivers Local 449 (Buffalo Linen Supply Co.), 353 U.S. 87 (1957); NLRB v. Babcock & Wilcox Co., 351 U.S. 105 (1956); Mastro Plastics Corp. v. NLRB, 350 U.S. 270 (1956). It is inconceivable that Manoli was confused or mistaken on this simple matter of black letter law, especially in preparation for a case of Gissel’s magnitude. What could explain his erroneous assertion to the Justices that the Labor Board no longer looked to an employer’s good faith doubt?

After extensively researching the NLRB’s files and interviewing living witnesses with knowledge of the Gissel case prep, Professors Laura J. Cooper and Dennis R. Nolan have presented compelling evidence that Manoli deliberately misled the Court as to the state of Board law.144Cooper & Nolan, supra note 129, at 219–22. In addition to being acutely aware of the labor bar’s desire for the agency to discard the good-faith doubt test in favor of an analysis relying strictly on the severity of employer misconduct,145Id. at 220 (discussing Manoli’s consumption of Welles, supra note 116). NLRB personnel were concerned how best to answer the Court’s inevitable questions regarding an employer’s simple preference for an election:

Board attorney Thomas E. Silfen, who had written the first draft of the Board’s brief in Gissel, was present on a day shortly before the oral argument when Norton Come did a mock argument with Manoli. In the practice session, Manoli was presented with a hypothetical of an employer who hadn’t committed unfair labor practices and who also had no doubt about the validity of the authorization cards. The hypothetical employer told the union he wouldn’t recognize the union solely on the basis of authorization cards and wanted an election. The lawyer pretending to be a Supreme Court justice asked Manoli whether this employer would have committed an unfair labor practice [under Section 8(a)(5)]. Manoli replied that this wasn’t an unfair labor practice. The mock justice then reminded Manoli that in the hypothetical the employer had no good faith doubt. According to Silfen, Manoli “was left speechless, without a good answer.146Cooper & Nolan, supra note 129, at 220–21.

This issue was a conceptual obstacle in the NLRB’s path to victory, but Joy Silk did not place it there. Under the pre-modified “pure” version of the Joy Silk doctrine, the hypothetical employer’s inability to proffer a good faith doubt as to the union’s majority status would have resulted in the finding of an unfair labor practice under Section 8(a)(5), consistent with the NLRB’s then-interpretation of what the duty to bargain required when confronted with a duly designated bargaining representative. The Labor Board’s choice in Aaron Bros. to emphasize the extent of unlawful subsequent conduct yet simultaneously retain the inquiry into an employer’s state of mind at the time it refused recognition created the exact dilemma that perplexed Manoli. That is, if an employer rejected a union’s request for recognition but openly accepted the validity of its card majority, and (1) these cards were not inherently unreliable demonstrations of employee choice (as the Board argued contra the Fourth Circuit), and (2) the issuance of a bargaining order still ostensibly hinged on whether the employer possessed a good faith doubt as to the union’s level of support at the time it made its request, then how does that employer escape a meritorious refusal to bargain charge if it insists upon proceeding to an election?147Recall that under Aaron Bros. an employer could insist upon an election absent an affirmative showing of bad faith by the General Counsel, to be demonstrated by substantial independent unfair labor practices or other conduct that demonstrated its lack of good faith, such as a Snow & Sons situation of repudiating an agreed-upon card check. Aaron Bros. Co., 158 N.L.R.B. 1077, 1078–79 (1966). But this safe harbor presumably did not extend to situations where an employer insisted upon an election despite stating it did not have a good faith doubt about the union’s majority status.

Manoli’s solution was to pretend this dilemma didn’t exist under Aaron Bros. and unilaterally substitute the labor bar’s preferred test for bargaining orders in place of NLRB case law. To the great confusion of the Justices and his fellow advocates in the courtroom, Manoli repeatedly claimed throughout his argument that an employer could automatically reject a union’s showing of cards and require the union to certify its majority through a secret-ballot election so long as the employer did not commit unfair labor practices that precluded a fair election.148Cooper & Nolan, supra note 129, at 220–21; Transcript of Oral Argument at 3–28, NLRB v. Gissel Packing Co., 395 U.S. 575 (1969) (No. 573) (argument of Dominick L. Manoli). Though skeptical, the Court accepted Manoli’s version of events and jettisoned any contemplation of good faith doubt in reviewing the various employers’ conduct; in its place, the Labor Board could now only “take into consideration the extensiveness of an employer’s unfair practices in terms of their past effect on election conditions and the likelihood of their recurrence in the future.”149Gissel Packing, 395 U.S. at 614. The Court proceeded to distinguish between employers’ unfair labor practices predicated upon their impact on the agency’s election processes and the range of the appropriate remedy:

Gissel I

  • In “exceptional” cases marked by “outrageous” and “pervasive” unfair labor practices, a “fair and reliable election cannot be had” and a bargaining order should issue;150Id. at 613–14.

Gissel II

  • In cases of “less pervasive practices which nonetheless still have the tendency to undermine majority strength and impede the election process[,]” the NLRB should consider “the extensiveness of an employer’s unfair practices in terms of past effect on election conditions and the likelihood of their recurrence in the future” before issuing a bargaining order;151Id. at 614–15. All four of the bargaining orders on appeal from the First and Fourth Circuits fell into this “Gissel II” category. They were remanded to the NLRB for reconsideration consistent with the Court’s analysis, to be made without findings of the employers’ good or bad faith. Id. at 616. and

Gissel III

  • In cases of “minor or less extensive unfair labor practices,” a bargaining order should not issue “because of [the practices’] minimal impact on the election machinery[.]”152Id. at 615.

With good faith doubt deemed abandoned by the NLRB—regardless of whether its demise came by more proper means of adjudicative deliberation by the Board itself or the rogue machinations of one of its attorneys153Cooper and Nolan never go so far as to conclude that Manoli’s actions were unethical, but they concede that “Manoli surely knew that he could not personally create Board law.” Cooper & Nolan, supra note 129, at 221.—the sliding scale of Aaron Bros. was elevated as the agency’s sole concern when appraising a bargaining order. In just five years, a rare coalition of labor law academics and practitioners had rebelled against the Kennedy Board’s enforcement of the Joy Silk doctrine and successfully lobbied for its replacement. Given the unanimity and efficacy of this crusade, it is surprising that these experts’ endeavors would almost immediately prove disastrous to the NLRA’s fragile ecosystem of enforcement.

F. The Last Gasp of Good Faith Doubt in Linden Lumber

It was not long before the matter of union recognition requests and employers’ duty to bargain was back before the Supreme Court. The Gissel decision, soon criticized for having posed as many questions as it answered,154See Christensen & Christensen, supra note 116, at 421–23. purposefully refused to resolve the issue of when, if ever, an employer that had committed no independent unfair labor practices had a duty to recognize a union asserting its majority status.155NLRB v. Gissel Packing Co., 395 U.S. 575, 594 (1969) (“[W]e need not decide whether a bargaining order is ever appropriate in cases where there is no interference with the election processes.”). Indeed, the hypothetical question that had stumped Manoli in his argument preparation was never actually relevant to the case, as all of the employers involved were found to have committed unfair labor practices. Cooper & Nolan, supra note 129, at 221–22. Given Manoli’s desertion of the good-faith doubt test in Gissel, the NLRB was left to craft its own standard in addition to deciding whether to maintain the two exceptions Manoli gave to the rule that an employer can insist upon an election absent the commission of serious unfair labor practices: (1) that an employer could not refuse to bargain if it had independent knowledge that a majority of its employees supported the union, and (2) that an employer could not refuse recognition initially by objecting to the appropriateness of the unit and then later claim that it doubted the union’s majority status.156Gissel Packing, 395 U.S. at 594. The first exception referred to the Snow & Sons doctrine. See supra notes 85–88 and accompanying text. The second exception referred to United Aircraft Corp., 144 N.L.R.B. 492 (1963), another Joy Silk-era case. While neither exception explicitly relied upon the Joy Silk doctrine for its own vitality, each involved employer-motive examination that fit poorly with the agency’s transition from good faith doubt to employer misconduct as the trigger for a bargaining order.157See Christensen & Christensen, supra note 116, at 433–44.

There was more at stake than just the NLRB’s juggling of rationales in the fallout of a Supreme Court decision. Though Gissel did not offer much guidance to the Labor Board in how to address card-based recognition requests going forward, its expansive interpretation of the statutory language confirmed that the agency possessed the authority to compel an employer to recognize and bargain with a majority-supported union under virtually any circumstances. While most of the labor bar viewed the NLRB’s options as either promoting observance of its election machinery or tolerating a lesser form of union organizing, the Labor Board faced a more basic decision in terms of its administrative mission: whether to continue to enforce Section 8(a)(5) of the NLRA with respect to unions that had been “designated or selected” by a majority of employees as it did with respect to unions that had been certified by the agency or voluntarily recognized by the employer.

After initially opting to maintain the status quo,158Wilder Mfg. Co., Inc. (Wilder I), 185 N.L.R.B. 176 (1970) (finding violation of Section 8(a)(5) and issuing bargaining order due to employer’s independent knowledge of union’s majority status); Pacific Abrasive Supply Co., 182 N.L.R.B. 329 (1970) (same). the newly-constituted Nixon Board abruptly chose to stop enforcing the law. In Linden Lumber Division, Summer & Co., the NLRB addressed the issue of “whether, absent election interference, an employer who insists on an election must initiate the election by his own petition.”159190 N.L.R.B. 718, 720 (1971). This was one of the matters the Supreme Court pondered but specifically declined to resolve. Gissel Packing, 395 U.S. at 601 n.18. A three-member majority determined that employer intransigence does not violate Section 8(a)(5), effectively placing the burden on unions to file an election petition once their request for recognition failed.160Linden Lumber, 190 N.L.R.B. at 721. The Labor Board also took the opportunity to confirm Joy Silk’s demise, having never held so prior to Manoli’s maneuver:

The facts of the present case have caused us to reassess the wisdom of attempting to divine, in retrospect, the state of employer (a) knowledge and (b) intent at the time he refuses to accede to a union demand for recognition. Unless, as in Snow & Sons, the employer has agreed to let its “knowledge” of majority status be established through a means other than a Board election, how are we to evaluate whether it “knows” or whether it “doubts” majority status? And if we are to let our decisions turn on an employer’s “willingness” to have majority status determined by an election, how are we to judge “willingness” if the record is silent . . . or doubtful . . . as to just how “willing” the Respondent is in fact? We decline, in summary, to reenter the “good-faith” thicket of Joy Silk, which we announced to the Supreme Court in Gissel we had “virtually abandoned . . . altogether[.]”161Id. at 720–21 (internal citations omitted).

The Nixon Board further elaborated on its petition-filing stance in “Wilder II”, a reconsideration of an earlier proceeding in which the Board applied the independent knowledge doctrine post-Gissel.162Wilder Mfg. Co., Inc., 198 N.L.R.B. 998 (1972). On remand, the majority appeared to concede that its position was entirely a matter of preference, but insisted “that the proper course in such cases is for the union, on behalf of the employees, to invoke our election processes.”163Id. at 999. What is more, the decision to allow employers to refuse recognition under virtually any circumstances and face no obligation to engage the Labor Board’s services was justified as the pro-collective bargaining position. This incentivized unions to prove their majority through elections “in a matter of weeks” instead of by “litigation which would take us nearly a year to reach” or more.164Id. The majority emphatically rejected the Joy Silk era, which was said to have “encourage[d] conflict, strikes, and contested litigation before this Board as a means of establishing a shaky foundation for future bargaining[.]”165Id. Although the answer as to whether this new policy actually encouraged the practice of collective bargaining will be made clear in Part II of this Article, it is enough to say that Board Member John Fanning’s dissent, which predicted increased delays and decreased voluntarism between parties in derogation of the NLRA’s policies, proved quite prescient. Id. at 1002 (Member Fanning, dissenting).

The Linden Lumber and Wilder cases eventually made their way to the Supreme Court after being consolidated on appeal to the D.C. Circuit.166Truck Drivers Local 413 v. NLRB, 487 F.2d 1099 (D.C. Cir. 1973). The D.C. Circuit reversed the Nixon Board’s postulation that “the decision to recognize a union on the basis of cards is entirely within the control of the employer[,]” finding that Sections 8(a)(5), 9(a), and 9(c)(1)(B) combined to “require an employer asserting a doubt of majority status to resolve the possibility through a petition for an election.”167Id. at 1106, 1111. Some form of good faith doubt thus was compelled by the statute. The NLRB could presumably abandon Joy Silk and any of its variants through administrative discretion,168See id. at 1108 n.36 (“It is one thing to abandon a subjective standard used to determine when employers are required to recognize on the basis of cards and replace it by the alternative, and arguably more objective, standard of “independent knowledge”. . . . It is another thing to abandon any standard.”). but it could not forego any requirement at all that an employer demonstrate its “willingness to have majority status determined by an election.”169Id. at 1105–11. It was inconceivable to the D.C. Circuit that the NLRA could be read to grant the employer an absolute right to “avoid both any duty to bargain and any inquiry into the actuality of his doubt” when presented with a majority of signed authorization cards if it also had no duty to at least begin the process of resolving the representation question through the Board’s processes.170Id. If nothing else, the Act required the employer to do something.

The Supreme Court disagreed. In a 5-4 decision authored by Justice William Douglas, the majority determined that the NLRB’s allotment of priorities within its election processes was neither arbitrary and capricious nor an abuse of discretion.171Linden Lumber Div., Summer & Co. v. NLRB, 419 U.S. 301, 309–310 (1974). Mining the NLRA for supporting authority, Justice Douglas could point only to the lack of any language in Section 9(c)(1)(B) that compelled employers to file a petition for an election.172See id. at 307–08. The Court’s analysis notably declined to contemplate how Section 9(c)(1)(B) interacted with any other provisions of the statute, including Sections 8(a)(5) and 9(a). The rest of the opinion can at best be described as policy justifications for the Labor Board’s decision, primarily focusing on the length of time spent in elections versus bargaining order litigation and the allegedly negligible difference in delays stemming from a union-filed petition.173Id. at 306–07, 309. The union parties in Gissel and Linden Lumber vigorously disagreed that employer-filed and union-filed petitions were one in the same, and the D.C. Circuit accepted the distinction. Truck Drivers Local 413, 487 F.2d at 1111–12 n.48. The Court offered a parting shot for the good-faith doubt test while sending it to its grave:

An employer concededly may have valid objections to recognizing a union on that basis. His objection to cards may, of course, mask his opposition to unions. On the other hand he may have rational, good-faith grounds for distrusting authorization cards in a given situation. He may be convinced that the fact that a majority of the employees strike and picket does not necessarily establish that they desire the particular union as their representative. Fear may indeed prevent some from crossing a picket line; or sympathy for strikers, not the desire to have the particular union in the saddle, may influence others. These factors make difficult an examination of the employer’s motive to ascertain whether it was in good faith. To enter that domain is to reject the approval by Gissel of the retreat which the Board took from its “good faith” inquiries.174Linden Lumber, 419 U.S. at 306 (emphasis added).

It is interesting that this dicta from Justice Douglas made its way into the majority opinion in Linden Lumber whereas no such commentary was offered in Gissel, the case in which the Joy Silk doctrine actually influenced the lower proceedings.

Justice Potter Stewart provided a powerful rejoinder on behalf of the four dissenting Justices. He explained that the “absolute right” position staked out by the NLRB disregarded the clear language of Sections 8(a)(5) and 9(a), which “intended to impose upon an employer the duty to bargain with a union that has presented convincing evidence of majority support, even though the union has not petitioned for and won a Board-supervised election.”175Id. at 310, 313–14, 317 (Potter, J., dissenting). While the agency was not confined to administering “a subjective test of the employer’s good faith in refusing to bargain with the union” as it did prior to Gissel, the Labor Board could not “adopt a rule that avoids subjective inquiries by eliminating entirely all inquiries into an employer’s obligation to bargain with a noncertified union selected by a majority of his employees.”176Id. at 314–15 (emphasis added). Section 9(c)(1)(B), originally intended as an act of self-protection for the employer which mirrored the union’s right to file a petition for an election, was now somehow interpreted by the majority as imposing an obligation upon unions to seek certification before obtaining the protections of Sections 8(a)(5) and 9(a).177See id. at 311–12 (majority opinion). Justice Stewart tellingly cited to the Arkansas Oak Flooring case for the proposition that the NLRA’s provisions could not be read as so exclusive. Id. at 311 n.2.

It is worth contextualizing the degree to which the last five years had warped the law from its relatively settled state. Where once the notion that an employer need only bargain with a certified union was widely considered “frivolous,”178E.g., NLRB v. Dahlstrom Metallic Door Co., 112 F.2d 756, 757 (2d Cir. 1940); NLRB v. Gissel Packing Co., 395 U.S. 575, 596 n.9 (1969) (citing Dahlstrom). the root of this concept was now deemed a permissible reading of the NLRA, as Linden Lumber had validated the exact sort of employer obduracy that had inspired the NLRB and the courts to construct the good-faith doubt test in the first place. By looking only to an employer’s post-request conduct as cause for issuing a bargaining order, the labor bar had essentially amended Section 8(a)(5) to activate only in the pre-certification process, when an employer independently violated the other provisions of Section 8(a). Neither the statute’s language nor its legislative history supports such a view.

II. The Post-Joy Silk Landscape of Bargaining Orders

The NLRB’s about-face in Linden Lumber marked the beginning of a 50-year period in which the agency has refused to enforce Section 8(a)(5) of the NLRA in cases where the charging party is a union that has been “designated or selected” as the majority representative rather than certified. While many scholars have examined the decline of collective bargaining and the diminished deterrence of federal labor law enforcement,179See infra notes 188–91 and accompanying text. few have tied these phenomena to the Labor Board’s ceding of jurisdiction over what was previously a rich environment for refusal to bargain charges. Upon closer examination, it becomes clear that the NLRB’s abandonment of the good-faith doubt test in favor of the misconduct-centric analysis enunciated in Gissel is intimately connected to the agency’s inability to enforce the law and, as a result, fulfill its statutory mission of encouraging collective bargaining.

A. The Enervation of Bargaining Orders Under the Gissel Framework

Despite the labor bar’s effusive lobbying for an objective standard in issuing bargaining orders and the broad approval granted to the remedy by the Supreme Court, the NLRB’s first wave of cases applying this standard were treated with great skepticism by the courts. Within two years of the Gissel decision, a majority of the circuits had rejected enforcement of such bargaining orders and criticized the Labor Board’s failure to heed the heavy evidentiary burden required to sustain the “extraordinary” remedy.180Petruska, supra note 15, at 114 n.85 (gathering cases). The Gissel remedy was first branded as “extraordinary” just nine months after the Supreme Court’s decision. NLRB v. American Cable Sys., Inc., 427 F.2d 446, 448 (5th Cir. 1970). The Board was accused of simply reciting the unfair labor practice findings and announcing the need for a bargaining order à la Joy Silk,181E.g., Pogrebin, supra note 21, at 201. and courts pounced on the Board for neglecting a rapidly proliferating set of factors that judges read into Gissel’s mandate when determining whether a fair holding of an election—or, more commonly, a fair rerun of a tainted election—was truly impossible.182See e.g., NLRB v. Jamaica Towing Co., 632 F.2d 208, 212–13 (2d Cir. 1980) (articulating theory of “hallmark” violations for review of bargaining orders). Judge Friendly, the frequent foe of the Kennedy Board’s prosecution of the good-faith doubt test, set the tone early for the judiciary’s hostility to Gissel:

[W]e do not believe the Court thought that all the Board needed to do in deciding whether to dispense with the admittedly superior method of the election process to determine employee sentiment was, in Judge Goldberg’s apt phrase, to use “a litany, reciting conclusions by rote without factual explication[.]” . . . [T]here could be an opinion by the full Board illuminating how it meant to apply its Gissel-given authority—a course particularly important for an agency that is forced by the press of business so often to delegate its authority to three-member panels. Failing that, the Board should explain in each case just what it considers to have precluded a fair election and why, and in what respects the case differs from others where it has reached an opposite conclusion.183NLRB v. Gen. Stencils, Inc., 438 F.2d 894, 901–02 (2d Cir. 1971) (internal citations omitted).

The opposing view was articulated most clearly by Judge Friendly’s colleague on the Second Circuit, Judge Paul Hays. Judge Hays had spoken of the need for judicial deference to the NLRB’s expertise during the Joy Silk era,184See NLRB v. Gotham Shoe Mfg. Co., 359 F.2d 684, 687 (2d Cir. 1966); NLRB v. Flomatic Corp., 347 F.2d 74, 80–81 (2d Cir. 1965) (Hays, J., dissenting). and he hit the same notes in his review of Gissel bargaining orders:

In warning the Courts of Appeals that it is for the Board not the courts to determine whether a bargaining order is justified, the Supreme Court could hardly have been more emphatic than it was in Gissel. Yet circuit judges continue to be so profoundly convinced of their own expertise in such matters that they do not hesitate to disregard the Supreme Court’s admonitions. Surely such matters as the effect on employees of threats to close the plant, to discharge for tardiness and to deprive them of plant privileges, are exactly the areas in which the Board rather than the judges have a special “fund of knowledge and expertise.”185NLRB v. Gen. Stencils, Inc., 472 F.2d 170, 176 (2d Cir. 1972) (Hays, J., dissenting).

But Judge Friendly’s vision carried the day, as judicial hostility to Gissel bargaining orders soon surpassed the acrimony which provoked the Supreme Court’s review of the good-faith doubt test.186See Petruska, supra note 15, at 112–15. The D.C. Circuit would arguably downplay the state of the law when, many years later, it acknowledged that courts have “long viewed [bargaining orders] with suspicion.”187Lee Lumber & Bldg. Material Corp. v. NLRB, 117 F.3d 1454, 1461 (D.C. Cir. 1997). These developments stand in stark contrast to the labor bar’s predictions of the 1960s which foresaw smoother sailing for the NLRB in federal courts upon abandoning the good-faith doubt test. See Welles, supra note 116, at 355.

Indeed, the NLRB’s difficulties in enforcing Gissel bargaining orders—and the agency’s eventual retreat from enforcing the remedy at any significant volume—are well documented. One study found that the average enforcement rate in the circuit courts was barely seventy percent.188See Terry A. Bethel & Catherine A. Melfi, Judicial Enforcement of NLRB Bargaining Orders: What Influences the Courts?, 22 U.C. Davis L. Rev. 139, 161 (1988). Another observed that the Labor Board’s issuance of bargaining orders had drastically decreased over time; whereas 176 bargaining orders were issued by the agency between 1979 and 1982,189Terry A. Bethel & Catherine Melfi, The Failure of Gissel Bargaining Orders, 14 Hofstra Lab. L.J. 423, 429, 431 (1997). an average of just 9.7 such orders were issued annually from 1987 to 1996.190Peter J. Leff, Failing to Give the Board Its Due: The Lack of Deference Afforded by the Appellate Courts in Gissel Bargaining Order Cases, 18 Lab. Law. 93, 109 (2002). From 1994–1996, only half of the twenty bargaining orders reviewed by the circuit courts in those years were enforced. Id. at 110–11. Common sense dictates that a remedy which is barely deployed and inconsistently enforced will have little deterrent effect, so it is little wonder that every quantitative analysis of Gissel bargaining orders concluded that they have failed in their stated goal of remedying unlawful employer conduct in the organizing context.191See, e.g., id. at 119–20; Bethel & Melfi, supra note 189, at 451–52; Benjamin W. Wolkinson, Nancy Briggs Hanslowe, & Shlomo Sperka, The Remedial Efficacy of Gissel Bargaining Orders, 10 Indus. Rels. L.J. 509, 521 (1989); Bethel & Melfi, supra note 188, at 174–75; Rosemary O’Shea, Gissel Bargaining Orders (1982) (unpublished LL.M. thesis, Harvard University) (cited in Paul Weiler, Promises to Keep: Securing Workers’ Rights to Self-Organization Under the NLRA, 96 Harv. L. Rev. 1769, 1795 nn.94–95 (1983)).

Perhaps the greatest challenge facing the consistent enforcement of Gissel bargaining orders is the passage of time attendant with administrative delays. Circuit courts have routinely refused to enforce bargaining orders given “subsequent events bearing upon employee choice, including changes in the management as well as in the work force and the passage of time.”192E.g., NLRB v. Jamaica Towing, Inc., 632 F.2d 208, 214–15 (2d Cir. 1980). But this concept of “changed circumstances” is not another example of the judiciary’s overreach; the problem inheres in Gissel’s urging that the NLRB—when weighing the extensiveness of an employer’s unfair labor practices—may contemplate the “past effect on election conditions and the likelihood of their recurrence in the future.”193NLRB v. Gissel Packing Co., 395 U.S. 575, 614 (1969) (emphasis added). While the Labor Board initially ignored the circuit courts’ demands that it consider the passage of time in its calculations, the Board symbolically acquiesced in 1999, acknowledging the D.C. Circuit’s critical stance on the matter.194See Cooper Hand Tools, 328 N.L.R.B. 145, 146 (1999) (citing Flamingo Hilton-Laughlin v. NLRB, 148 F.3d 1166, 1171 (D.C. Cir. 1996); Charlotte Amphitheater Corp. v. NLRB, 82 F.3d 1074, 1078 (D.C. Cir. 1996)). The NLRB did not hold that it was compelled to take changed circumstances into consideration, but suggested it could discretionarily refuse to issue bargaining orders on this basis. Gil A. Abramson, The Uncertain Fate of Gissel Bargaining Orders in the Circuit Courts of Appeal, 18 Lab. Law. 121, 134 (2002). This appeared to be part of a broader agency-wide reckoning after 30 years of the remedy’s dwindling success.195See generally Gissel Memo, supra note 9.

A recent application of the changed circumstances phenomenon goes far in demonstrating the Gissel doctrine’s internal flaws. In Sysco Grand Rapids, LLC, the NLRB exercised its discretion to reject an administrative law judge’s (“ALJ”) recommendation for a Gissel bargaining order solely on the belief that the order would not be enforced on appeal due to an excessive passage of time.196367 NLRB No. 111, slip op. at 2 (Apr. 4, 2019). The union in that case began organizing the employer’s workplace in late 2014 and had accrued a card majority by December 18 of that year.197Id. at 1–2. The union would eventually obtain cards from 99 of the 162 employees in the shop. Id. at 6 nn.4, 6. The employer thereafter waged an aggressive antiunion campaign which included several “hallmark” violations of the NLRA, including “threats of loss of wages, benefits, and jobs; threat of plant closure; the pretextual discharge of a top union supporter in the workplace; and the solicitation of grievances with a responsive grant of benefits.”198Id. at 1, 30–31. The union requested recognition on March 6, 2015 and, upon rejection by the employer, filed a petition for an election on March 11.199See id. at 30. The union lost the election on May 7 by a vote of 71 to 82 and proceeded to file a slew of unfair labor practice charges against the employer’s pre- and post-petition conduct.200Id. at 23. The ALJ found merit in virtually all of the charges and recommended that the election be set aside, the petition dismissed, and the employer be ordered to recognize and bargain with the union due the employer’s “severe and continuous” conduct unlawfully dissipating the union’s card-based majority and preventing the holding of a fair rerun election.201Id. at 23–33. The misconduct culminated in the issuance of a “Gissel II” bargaining order. Id. at 28­­–31.

The ALJ’s decision was not published until March 2, 2017. Over 200 exhibits were introduced and over 2,000 pages of testimony were elicited at trial, for which hearings were conducted over the course of fourteen days spread across four-and-a-half months.202Brief of the Charging Party Union/Petitioner at 3, Sysco Grand Rapids, LLC, 07-CA-146820 (filed Dec. 30, 2016). The employer had successfully moved to postpone the trial date twice and filed other customary pleadings, including motions for a bill of particulars and to revoke a subpoena.203This information can be gleaned from the NLRB’s public dockets of the underlying charges in the case, 07-CA-146820 et al., and the original petition, 07-RC-147973. But the delays did not stop at the trial level. After the case was transferred to the Board for review, the employer filed three motions to re-open the hearing seeking to introduce evidence of alleged changed circumstances, all of which prompted objections, responses, and reply memoranda, in addition to the standard exception-based briefing to the ALJ’s decision. Somewhat comically, the employer even challenged a typographical correction issued by the ALJ regarding the charging party’s title, asserting that the erroneous local union number provided in the decision retroactively deprived the ALJ of jurisdiction over the case.204See Respondent’s Exceptions to the Administrative Law Judge’s Decision at 2, Sysco Grand Rapids, LLC, 07-CA-146820 (filed May 8, 2018).

The full Board in Washington, D.C. finally handed down its own decision on April 4, 2019, over four and a half years since the union began organizing the employer and almost four years after the election was held. The majority held that this passage of time, accompanied by a 30% turnover in the shop since the union obtained its majority, significantly decreased the likelihood of enforcing the Gissel bargaining order on appeal, even though the “severity of the Respondent’s unfair labor practices” would normally require the issuing of one.205Sysco Grand Rapids, slip op. at 2. Not wanting to “engender further litigation and delay over the propriety of a bargaining order,” the majority rejected the ALJ’s recommended remedy and instead ordered a rerun of the May 2015 election.206Id.

Still the matter persisted. The NLRB denied the employer’s motion for reconsideration on November 14, 2019 and briefing on the employer’s appeal of the unfair labor practices involved multiple extensions of time by both parties. The Sixth Circuit issued a decision upholding the meat of the Labor Board’s unfair labor practice findings on September 4, 2020.207See Sysco Grand Rapids, LLC v. NLRB, 825 F. App’x 348, 352–56 (6th Cir. 2020). The ballots of the rerun election were tallied on January 22, 2021, followed by a nearly five-month-long process resolving various objections to the election. On June 14, 2021, nearly seven years after the original organizing campaign began, the results of the election were certified.208In a result that labor lawyers of all stripes would regard as something of a miracle, the union had won 62 to 61.

Somewhere amidst this administrative nightmare, the majority’s decision in Sysco Grand Rapids drew a spirited dissent from then-Board Member (and current Chairman) Lauren McFerran, who championed the bargaining order’s place within the NLRA’s remedial weaponry.209Sysco Grand Rapids, slip op. at 5–10 (Member McFerran, concurring in part and dissenting in part). While not necessarily refuting the majority’s prognostication of the Gissel order’s sure death on appeal, McFerran instead made a stand against unilateral disarmament by the agency:

[E]ven at the risk of nonenforcement, we must stand for the [National Labor Relations] Act, not only because the Board is primarily responsible for effectuating the purposes and policies of the Act, but also because the Board must set national labor policy and cannot effectively do so by reverting to the least protective approach from a patchwork of views among the federal courts of appeals.210Id. at 7. But cf. Donald L. Dotson & Charles M. Williamson, NLRB v. the Courts: The Need for an Acquiescence Policy at the NLRB, 22 Wake Forest L. Rev. 739 (1987).

While cathartic, the reality remains that the Gissel framework is terminally vulnerable to legal stalling tactics and the sometimes-leisurely pace of federal bureaucracy.211It is worth wondering how much of the delays in NLRB enforcement is owed to the decades-long decline of the agency’s staffing numbers in its regional offices. See U.S. Gov’t Accountability Off., GAO-21-242, National Labor Relations Board: Meaningful Performance Measures Could Help Improve Case Quality, Organizational Excellence, and Resource Management 13–18 (2021); Robert Iafolla & Hassan A. Kanu, NLRB ‘Staffing Crisis’ in the Field May Get Worse, Union Says, Bloomberg Law (Mar. 27, 2019, 2:53 PM), []. Because the test is predicated upon the NLRB’s ability to administer its election procedures for a hypothetically untainted electorate, the composition of the workforce rationally can be considered relevant to the remedy. This strongly incentivizes employers to delay the outcome of Labor Board litigation to achieve as much employee turnover as possible. One quickly sees why General Counsel Abruzzo appears to be targeting Sysco Grand Rapids for reversal expressly on the theory that employer delays contributed to the Board’s failure to issue a bargaining order.212See Abruzzo Memo, supra note 18, at 6.

Instead of exploring substantive alternatives to Gissel, the only creative energy expended by the agency in recent decades towards strengthening its bargaining order remedy has been the prioritization of injunctive relief under Section 10(j).213See Gissel Memo, supra note 9, at 17 (mandating submission of any potential Gissel case to NLRB’s Injunction Litigation Branch). Section 10(j) provides, in relevant part:The Board shall have power, upon issuance of a complaint . . . charging that any person has engaged in or is engaging in an unfair labor practice, to petition any United States district court, within any district wherein the unfair labor practice in question is alleged to have occurred . . . for appropriate temporary relief or restraining order. Upon the filing of any such petition the court . . . shall have jurisdiction to grant to the Board such temporary relief or restraining order as it deems just and proper.
Labor Management Relations Act, 1947 (Taft-Hartley) § 10(j), 29 U.S.C. § 160(j).
True, litigating the union’s unlawful loss of its majority as soon as the field investigation is concluded does much to address the lag involved in most NLRB litigation, and some courts have accepted this proactive remedy.214See, e.g., Overstreet v. NP Red Rock, LLC, 2021 U.S. Dist. LEXIS 134802 (D. Nev. July 20, 2021); accord Scott v. Stephen Dunn & 8 Assocs., 241 F.3d 652 (9th Cir. 2001); NLRB v. ElectroVoice, Inc., 83 F.3d 1559 (7th Cir. 1996); Asseo v. Pan American Grain Co., 805 F.2d 23 (1st Cir. 1986); Seeler v. Trading Port, Inc., 517 F.2d 33 (2d Cir. 1975); cf. Boire v. Pilot Freight Carriers, Inc., 515 F.2d 1185, 1194 (5th Cir. 1975) (upholding district court’s refusal to issue interim bargaining order where circuit court was “not convinced that a continuation of the non-bargaining status will so deleteriously affect the union that it cannot recover”). But there is no indication that the Labor Board seeks and obtains 10(j)-variety Gissel orders at any meaningful volume,215General Counsel Leonard Page reported in 1999 that the NLRB had sought an “interim” Gissel bargaining order in just 68 cases from 1990 to 1998. See supra Gissel Memo, note 9, at 15. Of course, the Labor Board itself issued just 119 decisions during this period involving a request for a “full” Gissel order. Id. In comparison, the Board issued 107 Joy Silk bargaining orders in 1967 alone. Gross, supra note 83, at 359 n.146. and as the next subpart of this Article will demonstrate, it is clear that this initiative has not noticeably deterred the historical trends of employer misconduct. Moreover, Section 10(j) relief can have the novel consequence of later harming the Board’s ability to enforce a bargaining order where the court believes an injunction already dispelled the sort of misconduct that Gissel is premised upon remedying.216See Novelis Corp. v. NLRB, 885 F.3d 100, 110 (2d Cir. 2018) (rejecting a Gissel bargaining order under belief of “the unlikelihood of reoccurring unfair practices”; court was “not prepared to presume that a company would commit new violations when faced with an existing federal court order and potential contempt proceedings”). It is thus doubtful that the agency can bootstrap Gissel into a state of potency merely by shuffling the order in which courts are presented with its precepts. The problem remains the precepts themselves.217See discussion supra Part III.B.

The story of Gissel inverts the history of Joy Silk. Whereas Joy Silk was developed through years of dialogue between the NLRB and the courts before being universally accepted as national labor policy, only to later fall victim to backlash from more robust enforcement, Gissel was foisted upon labor law overnight and immediately resisted by the federal judiciary to the point where its relevance has linearly waned. Gissel has nevertheless survived its criticism, lumbering on in obscurity.

B. The Explosion of Employer Misconduct Following the Retreat from Joy Silk

While outwardly serving as remedies in individual cases, both the Joy Silk and Gissel breeds of bargaining orders function more broadly as deterrents against what the NLRB perceives as widescale employer wrongdoing; the former by condemning bad-faith rejections of “the collective bargaining principle” mandated by Section 8(a)(5) and the latter as a tripwire once misconduct has reached a certain level.218To be sure, Supreme Court case law forbids the NLRB from imposing punitive remedies despite the expansive language of Section 10(c). See, e.g., Local 60, United Brotherhood of Carpenters and Joiners of America v. NLRB, 365 U.S. 651, 653–56 (1961) (rejecting the Eisenhower Board’s Brown-Olds remedy requiring the reimbursement of union dues collected from an unlawful closed shop arrangement, even if no specific evidence of coercion was presented). But while some critics of the bargaining order remedy once decried it as punitive, see 1968 Hearings, supra note 91, at 1045 (statement of National Retail Merchants Association); Yale Note, supra note 116, at 830, the Court closed the door on any such argument in Gissel. See NLRB v. American Cable Systems, Inc., 427 F.2d 446, 448 (5th Cir. 1970) (“The [bargaining] order is not a traditional punitive remedy, but is a therapeutic one.”). The doctrines’ worth as backstops of federal labor law enforcement must therefore be measured by their ability to prevent the type of activity for which they were originally conceived. Here, a clear winner emerges.

Until recently, no one had compared the competing effects on labor policy of the Joy Silk and Gissel eras of bargaining orders, even those who have written negatively of Gissel’s preventative and restorative effects.219See supra notes 188–91 and accompanying text. Arguably the most glaring omission here is by Professor Wolkinson, who early in his career analyzed the Joy Silk doctrine at some length. Benjamin Wolkinson, Remedial Efficacy of NLRB Remedies in Joy Silk Cases, 55 Cornell L. Rev. 1 (1969). In fairness, Wolkinson’s focus in both his Joy Silk and Gissel studies was on unions’ likelihood of achieving a first collective bargaining agreement after being on the receiving end of a bargaining order, rather than analyzing the doctrines’ influence upon employer misconduct. Brian Petruska filled this void in a prescient and powerful article which examined the twin occurrences of (1) the explosion in employer unfair labor practices during the 1970s and (2) the simultaneous drop in NLRB-conducted elections, and argued that the disappearance of Joy Silk from Board case law explains much of these phenomena.220Petruska, supra note 15. While it is concededly counterintuitive that a little known doctrine which is almost completely excluded from modern labor law curriculum could be responsible for the fall of the mid-century legal regime, Petruska deftly presents and navigates the data while dispatching with competing theories.

According to Petruska, after Joy Silk was deemed abandoned by the Supreme Court in Gissel, the number of unfair labor practice charges alleging unlawful employer intimidation under Section 8(a)(1) or unlawful employee discharges under Section 8(a)(3) saw a “nearly instantaneous” spike.221Id. at 116–27, 132. Unlawful discharge allegations increased 125% from 1969 to 1981 after rising only slightly from 1959 to 1969; unlawful intimidation charges increased 525% from 1969 to 1981 despite having remained static the decade prior.222Id. at 116 tbl.2, 117. Meanwhile, elections had risen steadily during the 1960s but began to flatline in the 70s, eventually cratering in a direction that has never been reversed.223Id. While discharge and intimidation charges also fell in terms of raw numbers after 1980, the drop in elections did not lead to a decrease in the rate at which employers were accused of misconduct in the few elections that were actually held.224See id. at 120 tbl.4, 121. In fact, the number of discharge and intimidation charges filed per election continued to rise through the twenty-first century.225Id. at 122 tbl.5. For example, where just one discharge allegation was filed for every election held from 1959 to 1969, an average of 3.2 such charges per election were filed from 1981 to 2009.226Id. at 123.

Suffice it to say, something happened in (or around) 1969 to dramatically affect employers’ willingness to unlawfully interfere in elections.227The increase in charges cannot be explained simply as a strategic decision by unions to file more frequently with the NLRB regardless of the merit of the charges. As Petruska points out, the agency’s “merit factor”—i.e., the rate at which an unfair labor practice charge will result in a decision by the investigating regional office to issue a complaint—has increased every decade after 1970. Id. at 118. Petruska points to the Joy SilkGissel exchange as the culprit.228Id. at 127–32. Under the Joy Silk doctrine, employers did not have an incentive to commit unfair labor practices in the run-up to an election, as most Section 8(a)(1) or (3) violations could prevent an employer from arguing it had a good-faith doubt of the union’s majority status and result in Board-ordered bargaining. Gissel substantially raised a union’s burden for proving the need for a bargaining order by making only “outrageous” and “pervasive” unfair labor practices worthy of bypassing the NLRB’s election machinery, meaning that many—if not most—violations could escape scrutiny under this test. The internal analysis for an employer thus changed to how much misconduct it could get away with, balanced against the electoral benefits it received from intimidating its workforce. Given Gissel’s stated preference for elections, many employers were emboldened to explore the doctrine’s boundaries.

While Petruska’s argument runs counter to the prevailing theory of labor history that management opposition to unionism spread and intensified as a culture in the 1970s and was dramatically punctuated by President Ronald Reagan’s firing of the Professional Air Traffic Controllers Organization strikers in 1981,229Petruska cites Professor Richard Freeman’s work as the archetype for this viewpoint. Id. at 132, 157–59. the latter explanation is lacking. Individual employers have overwhelmingly opposed unionization throughout American history and chosen to litigate and lobby in favor of the nonunion default, contrary to claims which suggest that mid-century management ideology generally accepted labor unions as a fact of industrial life.230Id. at 158 n.241 (citing Thomas Kochan, Collective Bargaining and Industrial Relations 182–83 (1980)). The prominence of the Joy Silk doctrine in mid-century labor law reflects the routine fashion in which employers refused union requests for recognition, and the immense controversy surrounding the Kennedy Board’s application of the good-faith doubt test in the 1960s demonstrates that the unanimity of management opposition to unionism almost certainly predated the point at which unfair labor practice charges exploded in number. Instead of the result of a spontaneous cultural consensus, the trend of the 1970s “is better explained by a management culture already hostile to unions, with a sudden change in the legal regime that reduced deterrence.”231Id. at 132 (emphasis added).

By the late 1970s, the NLRB was processing three times more unfair labor practice charges than election petitions.232Id. at 142 tbl.11. This ratio only expanded in the 1980s and 90s as petitions sharply dropped in number.233Id. The Labor Board’s primary duty thus increasingly became remedying unfair labor practices rather than resolving questions of representation, a striking subversion of Congress’s original vision of the agency as one encouraging the practice and procedure of collective bargaining.234Id. at 142–43. The inevitable inference is that employer misconduct had made secret-ballot elections an increasingly difficult and time-consuming endeavor for unions, leading them to abandon the certification process en masse in pursuit of more fruitful means of organizing.235See James J. Brudney, Neutrality Agreements and Card Check Recognition: Prospects for Changing Paradigms, 90 Iowa L. Rev. 819, 828 (2005) (finding that more than 80% of all private-sector workers unionized from 1998 to 2003 had organized outside of the NLRB election process). In one of the great ironies of American labor law, the near-apocalyptic pronouncements of Joy Silk’s coming evisceration of elections instead were realized under the prophets’ preferred test in Gissel.236Petruska, supra note 15, at 138–39.

Before concluding this Article’s historical discussion, it is worth revisiting the pivotal Logan Packing’s final words. As part of a general diatribe against the NLRB’s approach to remedying misconduct in the pre-recognition context, Judge Haynsworth laid forth his narrow vision for what place bargaining orders should occupy in the industrial order:

In those exceptional cases where the employer’s unfair labor practices are so outrageous and pervasive and of such a nature that their coercive effects cannot be eliminated by the application of traditional remedies, with the result that a fair and reliable election cannot be had, the Board may have the power to impose a bargaining order as an appropriate remedy for those unfair practices. . . . The remedy is an extraordinary one, however, and, in light of the guaranty of § 7 of employees’ rights not to be represented, its use, if ever appropriate, must be reserved for extraordinary cases.237NLRB v. S. S. Logan Packing Co., 386 F.2d 562, 570–71 (4th Cir. 1967).

This passage represents, for all intents and purposes, the test mostly adopted in Gissel and carried forth by the circuit courts in the decades after.238It is significant that the Supreme Court went out of its way in Gissel to minimize its “actual area of disagreement” with the Fourth Circuit regarding the contour of bargaining orders. NLRB v. Gissel Packing Co., Inc., 395 U.S. 575, 613 (1969). While Gissel was originally interpreted as a victory for unions over the recalcitrant views of the Fourth Circuit, the most important component of the case was convincingly decided in the employers’ favor. That so few scholars have traced this thread back to the Joy Silk era of bargaining orders—even when critiquing the Gissel structure—thus requires a fulsome defense of the logic underlying the remedy’s glory days.

III. Restoring The Good-Faith Doubt Test as the Lodestar of Bargaining Orders

A. The Elite Consensus Against Considering Subjective Motivation for
Refusals to Recognize

As discussed earlier, by the late 1960s the good-faith doubt test was under fire by an impressive armada of labor law experts in academia, government, and private practice.239See discussion supra Part I.D. The NLRB has historically been susceptible to pressure from public and private sources,240See supra notes 29–32 and accompanying text; Gross, supra note 37, at 232–36 (describing the NLRB’s retreat to a more restrictive determination of bargaining units following criticism by Congress and the American Federation of Labor); William B. Gould IV, Politics and the Effect on the National Labor Relations Board’s Adjudicative and Rulemaking Processes, 64 Emory L. J. 1501, 1517–18 (2015) (describing congressional and industry backlash to the Clinton Board’s attempted rulemaking regarding the so-called “single unit” rule). and the Kennedy Board—more controversial than any variation of the Board since its original Wagner Act composition241See, e.g., Browne, supra note 80, at 135 (arguing that the Kennedy Board was guilty of “a recurring failure to carry out congressional policy”); Arthur Krock, Free Enterprise at Stake Before the Court, N.Y. Times, Oct. 23, 1964, at 38 (decrying the Kennedy Board’s decisions in Darlington Mfg. Co., 139 N.L.R.B. 241 (1962) and Fibreboard Paper Prods. Corp., 138 N.L.R.B. 550 (1962) as “radically restrict[ing]” free enterprise); Kenneth C. McGuiness, The New Frontier NLRB (1963) (criticizing virtually every decision of note by the Kennedy Board as contravening the statutory intent of the Taft-Hartley Act).—proved to be no exception when it emaciated the Joy Silk doctrine into its Aaron Bros. skeleton.242See supra notes 111–115 and accompanying text. This modification did nothing to assuage the many opponents of Joy Silk, as they continued to call for the Labor Board to abandon the good-faith doubt test through the pivotal year of 1969.243Most notable in this year was the publication of a symposium held at the University of Georgia Law School regarding the NLRB’s treatment of card-based bargaining orders, featuring articles written by a leading Labor Board, management, and union attorney. Cooper & Nolan, supra note 129, at 220. All three participating practitioners criticized the agency’s pre-Gissel approach. See Browne, supra note 116, at 340–48; Sheinkman, supra note 116, at 328–29; Welles, supra note 116, at 355–57.

The Supreme Court in Gissel undercut critics’ most forceful argument when it approved the NLRB’s use of authorization cards in the context of union recognition. The Fourth Circuit had denounced cards as an inherently unreliable expression of employee choice244NLRB v. S. S. Logan Packing Co., 386 F.2d 562, 565–66 (4th Cir. 1967). and some commentators opined that cards should never be used by the Labor Board to establish majority status.245In one extreme example, an influential student note argued that a rerun election was the only suitable remedy for an employer’s unlawful interference with its employees’ free choice, even if the election requires “two or three or ten” replications. Yale Note, supra note 116, at 818–19. The note was cited by the Supreme Court in Gissel and in several of the major cases in the lower courts. See NLRB v. Gissel Packing Co., Inc., 395 U.S. 575, 602 n.19 (1969); S. S. Logan Packing Co., 386 F.2d at 566 n.17; NLRB v. River Togs, Inc., 382 F.2d 198, 205 n.10 (2d Cir. 1967). Alternatively, the Cumberland Shoe doctrine came under attack for its deregulatory approach to the agency’s monitoring of card solicitation.246See generally Christensen & Christensen, supra note 116, at 415 (describing this period as the “War of the Authorization Card”). While Gissel clearly subordinated card-based majorities below those established by elections, the Court’s attachment of at least some prestige to cards eliminated the most direct attacks against the Board’s use of bargaining orders.

But substantive objections to Joy Silk remained. Opponents raised three major complaints about the doctrine, all concerning the NLRB’s conception and application of good faith doubt. In the first and most basic complaint, commentators questioned the Labor Board’s search into employer motive in analyzing potential Section 8(a)(5) violations where a Section 9(a) relationship had not previously existed (the “Statutory Objection”).247See id. at 413–15; Welles, supra note 116, at 355–56. Neither Sections 8(a)(5) or 8(b)(3)248Labor Management Relations Act, 1947 (Taft-Hartley) § 8(b)(3), 29 U.S.C. § 158(b)(3) (2018) (providing that it is an unfair labor practice for a union “to refuse to bargain collectively with an employer, provided it is the representative of his employees subject to the provisions of section 9(a)”). make any mention of good or bad faith, and Section 8(d) appears to speak of the duty to bargain only in terms of when a representative has already been ascertained.249Christensen & Christensen, supra note 116, at 413, 434. While the Supreme Court in Arkansas Oak Flooring deigned the pivotal inquiry in the recognition context to be whether there was an “absence of any bona fide dispute as to the existence of the required majority[,]”250United Mine Workers v. Ark. Oak Flooring Co., 351 U.S. 62, 69 (1956). some suggested that this did not beg the question of the employer’s state of mind; rather, a bona fide dispute could be read to exist only in the presence of more objective, “convincing support” than the mere proffering of authorization cards, such as “by a strike or by a vote to strike.”251Christensen & Christensen, supra note 116, at 434–35. The agency’s blending of pre-recognition events with post-recognition duties was thus an “artificial” departure from the obligations imposed by the NLRA.252Duvin, supra note 116, at 268.

Second, critics interrogated the logic of the NLRB’s assumptions in applying the good-faith doubt test (the “Logical Objection”). It was frequently averred that an employer’s commission of an unfair labor practice sometime following its rejection of a union’s request for recognition could not retroactively prove that the employer lacked a good faith doubt of the union’s assertion of majority status.253See, e.g., Yale Note, supra note 116, at 828–29. Critics argued that an employer could in good faith believe that the union did not constitute a majority of the workforce at the time of its request, and later violate the law to prevent the coalescing of such support.254Id. at 829. Further, an employer could simply doubt the legitimacy of a card-based majority until it was established through an election based upon the company’s past organizing experiences.255This was the employer’s situation in NLRB v. S. S. Logan Packing Co., 386 F.2d 562, 567 (4th Cir. 1967). The Labor Board’s subjective practice of “retrospective divination” required more specific evidence than “the recitation of ancient and unreasoned precedent.”256Pogrebin, supra note 21, at 193; Yale Note, supra note 116, at 823. This was the dreaded “thicket,” denounced in Linden Lumber, that Gissel would supposedly lead the Board out of.

Finally, opponents charged the good-faith doubt test with perverting the policies of the Taft-Hartley Act (the “Policy Objection”). Even if card-based bargaining orders had survived the 1947 amendments, Congress’s inclusion of Section 8(c)257This Section protects “the expressing of any views, arguments, or opinion” from an unfair labor practice finding so long as the “expression contains no threat of reprisal or force or promise of benefit.” Labor Management Relations Act, 1947 (Taft-Hartley) § 8(c), 29 U.S.C. § 158(c) (2018). Although its language does not speak in terms of employers or unions, Section 8(c) was expressly included as a repudiation of NLRB case law which restricted most employer speech against union organizing efforts. See Gross, supra note 83, at 104–107. evinced a clear sponsorship of employers’ right to campaign against unionization.258See Yale Note, supra note 116, at 831. Joy Silk not only threatened to deprive employers of the ability to lawfully persuade their employees but also punished employers for being truthful and not masking their motives.259See id. at 832. Professor Howard Lesnick articulated this sentiment in a celebrated passage:

I must confess to a total inability to understand in what sense relevant here an employer is obliged to accept the collective bargaining principle. Of course he is enjoined to bargain in good faith (when he is obliged to bargain at all), and he must refrain from discrimination or coercion affecting employee attitudes toward collective bargaining. But what has that got to do with his state of mind or motivation when it comes to his response to an initial demand for recognition? It certainly cannot be unlawful for him to want to defeat the union at the polls and thereby obtain a lawful ground for rejecting collective bargaining. If an employer is to be permitted, first, to insist on an election, and then to campaign against the union in that election, it is perfectly clear that he is being permitted to reject the collective bargaining principle so long as his employees do not, by voting for the union, oblige him to accept it. It simply encourages hypocrisy to permit an employer to acknowledge his doubts, but not his hopes.260Lesnick, supra note 80, at 858–59. Commentators additionally argued that card-based bargaining orders harmed employees by preventing them from hearing both sides of the debate. See, e.g., Chicago Comment, supra note 116, at 390. Given that the employer almost never lost its right to campaign against the union unless it engaged in unfair labor practices, Lesnick expressed much less tolerance for this argument. Lesnick, supra note 80, at 863–68.

The Labor Board’s requirement under the “pure” Joy Silk test that the employer not only possess a good faith doubt of the union’s majority but that it espouse such doubt at the time it rejected the union’s recognition request and make some effort to ascertain the veracity of the union’s claim led one judge to dramatically describe the employer’s position as resembling that of Odysseus “as he makes the perilous passage between Scylla and Charybdis.”261NLRB v. Dan River Mills, Inc., 274 F.2d 381, 388 (5th Cir. 1960). The employer’s dilemma was supposedly great: it could (1) insufficiently state its doubt, breaching Joy Silk if the union in fact had a majority; (2) wade too far into the waters of workplace polling by unlawfully interrogating its employees, in contravention of Struksnes Construction Co., Inc.;262Struksnes Constr. Co., Inc., 165 N.L.R.B. 1062 (1967). (3) if successful in verifying the union’s majority support, either by lawfully polling the employees or some other means of “independent knowledge,” ensnare itself in a Snow & Sons situation by not immediately recognizing the union; or, most speculatively, (4) unwittingly violate Section 8(a)(2) by recognizing a union that did not in fact possess majority support.263Int’l Ladies’ Garment Workers’ Union v. NLRB (Bernhard-Altmann Tex. Corp.), 366 U.S. 731 (1961). But see NLRB v. Air-Master Corp., 339 F.2d 553, 556–57 (3d Cir. 1964); NLRB v. Signal Oil & Gas Co., 303 F.2d 785, 788 n.3 (5th Cir. 1962); NLRB v. Whitelight Prods., 298 F.2d 12, 15 (1st Cir. 1962). The result was to elevate “form over substance,”264Welles, supra note 116, at 356. privileging employers with sophisticated legal counsel over those who failed to speak the “magic words” required by the Labor Board.265Cooper & Nolan, supra note 129, at 220–21; Pogrebin, supra note 21, at 195; Rutgers Comment, supra note 116, at 208. One leading attorney described his view of the mid-century employer’s predicament in striking terms: “he embraces the union either as a sweetheart or as a Sabine Woman relaxed because rape is inevitable.” Christensen & Christensen, supra note 116, at 430 (quoting Joseph Barbash, Authorization Card Problems and Their Cure, 22 N.Y.U. Ann. Conf. Lab. 143 (1969)).

These assertions were never adequately rebutted. Only three defenses of the good-faith doubt test of any prominence were published prior to the Gissel decision: a short memorandum sent from Secretary of Labor W. Willard Wirtz to the Senate labor subcommittee in 1965;266Section 14(b) Hearings, supra note 100, at 19–26 (reprinting Secretary Wirtz’s memo in full). a speech delivered by an Associate General Counsel of the NLRB to members of the labor bar in 1968;267H. Stephan Gordon, Union Authorization Cards and the Duty to Bargain, 19 Lab. L.J. 201 (1968). and a memorandum sent from Chairman Frank McCulloch to the Senate judiciary committee in that same year.2681968 Hearings, supra note 91, at 1665–70. Congressman Frank Thompson, Jr. provided a fulsome defense of the good-faith doubt test on the House floor in August 1969 after Gissel’s publication, but he did so under the mistaken belief that the decision somehow rested upon the laurels of Joy Silk. 115 Cong. Rec. 23973–74 (daily ed. Aug. 13, 1969) (statement of Rep. Thompson, Jr.) [hereinafter Thompson, Jr. Statement]. None of the government officials addressed the substantive objections to Joy Silk highlighted above, instead concerning themselves with the alleged excesses of the doctrine by comparing the small number of card-based bargaining orders to the agency’s voluminous election docket.269Section 14(b) Hearings, supra note 100, at 19–20; Gordon, supra note 267, at 202–03, 222; 1968 Hearings, supra note 91, at 1667. Moreover, Gordon’s and McCulloch’s defenses came two years after the NLRB had already significantly modified the Joy Silk doctrine with its Aaron Brothers decision. With Joy Silk back on the labor-management community’s radar for the first time in decades,270See supra notes 18, 20 and accompanying text. it is well past time that the doctrine be vindicated.

B. The Good-Faith Doubt Test as the Superior Vehicle of Labor Law Enforcement

This final portion of the Article will address the primary objections to the NLRB’s use of the good-faith doubt test in the realm of union requests for recognition identified in the previous subsection, and elucidate their lack of historical, legal, and practical foundations.

1. The Statutory Objection

Arguments that attempt to distinguish the law of recognition from the law of bargaining in terms of when an employer’s motive should matter ignore the legislative history of the NLRA and the Act’s early developments by the NLRB. Senator Robert Wagner—who almost singlehandedly shepherded the NLRA through the upper chamber271 Philip Ross, The Government as a Source of Union Power 97 (1965). after delegating the drafting of the Act to his top assistant272See Kenneth M. Casebeer, Holder of the Pen: An Interview with Leon Keyserling on Drafting the Wagner Act, 42 U. Mia. L. Rev. 285 (1987).—made clear that a duty of recognition meant nothing without a corresponding duty to bargain, and vice versa.273 Ross, supra note 271, at 81. This sentiment, fleshed out in the seminal Houde case from the “old” National Labor Relations Board,274Houde Eng’g Corp., 1 N.L.R.B. 35 (1934). was clearly expressed in the Senate committee report regarding the inclusion of Section 8(5):

But after deliberation, the committee has concluded that this fifth unfair labor practice should be inserted in the bill. It seems clear that a guarantee of the right of employees to bargain collectively through representatives of their own choosing is a mere delusion if it is not accompanied by the correlative duty on the part of the other party to recognize such representatives as they have been designated . . . and to negotiate with them in a bona fide effort to arrive at a collective bargaining agreement.275 Ross, supra note 271, at 78 (emphasis added).

Importantly, Wagner considered an employer’s attempt to deal with its employees “otherwise than through representatives they have named” for the purposes of collective negotiations “would be the clearest interference with the right to bargain collectively.”276Id. at 72. The NLRB would soon agree. See Griswold Mfg. Co., 6 N.L.R.B. 298, 307 (1938) (“To meet and negotiate with a committee of employees while deliberately withholding union recognition does not satisfy the requirements of the Act. The paramount importance of the fact of union recognition alone in securing collective bargaining has been asserted repeatedly in our decisions.”).

Of course, two concepts can be symbiotic without being equivalent, but the Statutory Objection is further undermined by the simultaneous development of the good faith standard in both realms of Section 8(5). When the NLRB and the courts were crafting the concept of employer intent in the recognition arena,277See supra Part II.B. they were simultaneously clarifying that fulfillment of the duty to bargain was dependent on the employer’s good or bad faith posture at the negotiating table.278 Ross, supra note 271, at 108–12. Some of the Labor Board’s earliest cases instructed that an employer must possess a “bona fide intent” to reach a collective bargaining agreement with its employees,279Atlas Mills, Inc., 3 N.L.R.B. 10, 21 (1937). or else the duty would produce only a “barren right of discussion”;280St. Joseph Stock Yards Co., 2 N.L.R.B. 39, 53 (1936). stalling tactics served as a “substitute for non-recognition of the employees’ representatives[.]”281S. L. Allen & Co., 1 N.L.R.B. 714, 727 (1936). An employer was further obligated to “accept the procedure of collective bargaining in good faith . . . in light of the prevailing practice . . . and the spirit and purpose of the Act[.]”282Inland Steel Co., 9 N.L.R.B. 783, 803 (1938). Tellingly, an employer’s lack of good faith could be inferred from its subsequent conduct, including discharges, delays, and inconsistent negotiation positions.283 Ross, supra note 271, at 110–11. These pronouncements should sound familiar to the language utilized in Joy Silk, which considered an employer’s refusal of a union’s lawful request for recognition to lack good faith if the employer was motivated “by a rejection of the collective bargaining principle or by a desire to gain time within which to undermine the union.”284Joy Silk Mills, Inc., 85 N.L.R.B. 1263, 1264 (1949) (quoting Artcraft Hosiery Co., 78 N.L.R.B. 333, 334 (1948)). The Supreme Court approved the good faith standard for post-recognition violations of the duty to bargain in 1941,285H. J. Heinz Co. v. NLRB, 311 U.S. 514, 525–26 (1941). and Congress moved to codify the Board’s case law in this area by way of Section 8(d).286 Ross, supra note 271, at 148. For an example of the NLRB’s pre-Taft-Hartley analysis which contained language foreshadowing Section 8(d), see P. Lorillard Co., 16 N.L.R.B. 684, 696 (1939).

Employing a motive-based analysis in both the pre- and post-recognition contexts is not an “artificial” amalgamation of disparate doctrines. Rather, splitting the two despite their identical statutory origins, oft-intertwined factual scenarios, and overlapping interpretative timelines would be nothing less than arbitrary. The Statutory Objection declines to investigate why every relevant tribune of federal labor law adopted an implied inquiry into motive for employer refusals to bargain almost immediately after passage of the Wagner Act, or why a slew of unanimous Supreme Court opinions in the early 1940s upheld the NLRB’s use of bargaining orders to remedy such violations. Indeed, the only artificial departure from case law here appears to be the employer’s election privilege announced in Linden Lumber, which ignored virtually the entirety of Section 8(a)(5) case law from the preceding decades to arrive at a preferred policy outcome regarding the Labor Board’s administration of its unfair labor practice and election mechanisms.287See Charles J. Morris, Undercutting Linden Lumber: How a Union Can Achieve Majority-Status Bargaining without an Election, 35 Hofstra Lab. & Emp. L.J. 1, 1–10 (2017). The Board’s abrupt cessation of jurisdiction over pre-recognition refusal to bargain charges becomes almost impossible to defend upon realizing that for 35 years, the Board, the courts, and necessary majorities of Congress all declined to extinguish this power.288See 1968 Hearings, supra note 91, at 1666–67 (observing that both the Fannin and Javits bills were rejected by Congress).

It is true that state-sponsored searches into persons’ thoughts and rationales can pose difficulties for administrative agencies that must enforce their laws through litigation because of the requisite due-process rights afforded to prosecuted parties. But the NLRA, when boiled down to its essentials, was in effect a regulation passed in condemnation of the prevailing employer mindset—that of reflexive and spirited resistance to unionization.289See generally National Labor Relations Act § 1, 29 U.S.C. § 151 (2018) (condemning “[t]he denial by some employers of the right of employees to organize and the refusal by some employers to accept the procedure of collective bargaining”). The NLRB was constructed to remedy “bona fide dispute[s]” of employee representation,290United Mine Workers v. Ark. Oak Flooring Co., 351 U.S. 62, 69 (1956). and it is entirely permissible (if not necessary) for the Labor Board to filter the meanings of “bona fide” and “dispute” through the lens of scienter. In this telling of history, “good faith” is not a straitjacket for the Board that was shed after Gissel; instead, it breathed life into the duty to bargain and gave voice to the law’s moral component.291This Article proposes that the Gissel brand of the bargaining order—which will only issue in the presence of severe violations of Section 8(a)(1) and (3)—makes little practical sense as the companion to a remedy which derives its existence from Section 8(a)(5). The complaint that Joy Silk constituted an unlawful punitive remedy certainly seems to have more rhetorical heft against Gissel. See supra note 218.

2. The Logical Objection

The clearest consequence of the Supreme Court’s decision in Gissel was that the NLRB could continue ordering employers to bargain with unions based on a majority of signed authorization cards. While the Court deemed elections to be the “preferred” method of ascertaining majority support,292NLRB v. Gissel Packing Co., Inc., 395 U.S. 575, 602 (1969). it ultimately delegated to the Labor Board the ultimate determination of what value to ascribe to cards.293The Court dismissed the Fourth Circuit’s argument that cards are “totally invalid” for assuring employee choice, and in its most famous holding, reasoned that cards likely become “the most effective” method in the face of “conduct disruptive of the election process[.]” Id. So long as the Board does not prioritize cards in a way that subsumes elections,294Presumably, this would include any attempt by the NLRB to establish an automatic card check method of union designation through either the agency’s adjudicative or rulemaking processes. But see Halima Woodhead, Can Card-Check Be Unilaterally Imposed by the NLRB?, 1 Am. U. Lab. & Emp. L.F. 85, 95–98 (2011) (arguing otherwise based upon a literal reading of Sections 8(a)(5) and 9(a)). the Board is within its power to consider employees’ signatures as a generally valid designation of a majority representative.295Gissel implicitly rejected any reliance on Sunbeam Corp., 99 N.L.R.B. 546, 550 (1952), which was frequently quoted by Joy Silk’s critics for the proposition that authorization cards are a “notoriously unreliable method of determining majority status of a union[.]” See 1968 Hearings, supra note 91, at 174, 1163, 1391. As Chairman McCulloch pointed out, these citations ignored that Sunbeam was a case involving competing unions, which poses far different issues for the vitality of authorization cards. Id. at 797, 851, 1669. The full quote from the case—conveniently cropped by critics—makes this damningly clear: “This Board has also long recognized that authorization cards are a notoriously unreliable method of determining majority status of a union as a basis for making a contract where competing unions are soliciting cards, because of the duplications which then occur.” Sunbeam, 99 N.L.R.B. at 550–51 (emphasis added); accord Brown Truck and Trailer Mfg. Co., Inc., 106 N.L.R.B. 999, 1001 (1953).

This is not some empty truism. Whereas the labor bar frequently claimed pre-Gissel that an employer that committed an unfair labor practice after rejecting a union’s request for recognition may have in good faith believed the union did not yet possess a majority, an employer that was presented with evidence of a valid card count can no longer claim to have doubted the desires of his workforce at the time the request was made. The employer may reasonably believe that a sufficient number of these designations were made hastily or lazily and thus are capable of persuading against unionization, but any unfair labor practices committed in place of lawful campaign tactics can fairly be inferred as bad-faith efforts by the employer to unlawfully dissipate the union’s majority status after being placed on notice.

Even if one denies any causal link between employer misconduct and a lack of good faith doubt, subjective motivation is regularly dispositive of cases under the NLRA even in the absence of specific evidence proving a party’s unlawful intent. As mentioned, intent is paramount in post-recognition refusals to bargain under Section 8(a)(5).296See NLRB v. Ins. Agents’ Int’l Union, 361 U.S. 477, 508 (1960) (Frankfurter, J., concurring) (“The state of mind with which the party charged with a refusal to bargain entered into and participated in the bargaining process is the ultimate issue upon which alone the Board must act in each case . . .”); cf. NLRB v. Katz, 369 U.S. 736, 742–43 (1962) (finding violation of Section 8(a)(5) from employer’s unilateral change in conditions of employment even in absence of overall subjective bad faith). Motive of a “reliable manifestation” must normally be established by circumstantial evidence297Ins. Agents’, 361 U.S. at 505. gleaned from the totality of the circumstances “both at and away from the bargaining table.”298Cent. Mgmt. Co., 314 N.L.R.B. 763, 770 (1994) (citing Overnite Transp. Co., 296 N.L.R.B. 669, 671 (1989)). The employer’s bad faith would otherwise be virtually impossible to prove without a smoking-gun statement indicating its deception.299Irving Abramson, The Anatomy of Boulwarism with a Discussion of Forkosch, 19 Cath. U. L. Rev. 459, 480 (1970); see also Shattuck Denn Mining Corp. v. NLRB, 362 F.2d 466, 470 (9th Cir. 1966) (“Actual motive, a state of mind, being the question, it is seldom that direct evidence will be available that is not also self-serving.”).

Labor law’s willingness to forego direct evidence in unfair labor practice analysis is even more accentuated in Section 8(a)(3) cases.300Professor Archibald Cox declared that in these cases, “[t]he reason the employer’s motive is decisive is plain[,]” for where “the aggressor is not actuated by a desire to protect a recognized interest, the basis for his excuse disappears.” Archibald Cox, Some Aspects of the Labor Management Relations Act, 1947, 61 Harv. L. Rev. 1, 20–21 (1947). In Radio Officers’ Union v. NLRB, the Supreme Court held that motive of unlawful discrimination could be inferred by the Labor Board from the mere effect of the employer’s actions.301Radio Officers’ Union v. NLRB, 347 U.S. 17, 42–48 (1954). The employers in the consolidated Radio Officers case had undertaken their discriminatory actions upon request of the respective unions, in derogation of Section 8(b)(2). Labor Management Relations Act, 1947 (Taft-Hartley) § 8(b)(2), 29 U.S.C. § 158(b)(2) (2018). In Justice Stanley Reed’s words, “specific evidence of intent to encourage or discourage is not an indispensable element of proof of violation” of Section 8(a)(3).302Radio Officers’, 347 U.S. at 44. Channeling “the common law rule that a man is held to intend the foreseeable consequences of his conduct[,]” Justice Reed reasoned that “employer conduct inherently encourages or discourages union membership” if “a natural consequence of his action was such encouragement or discouragement.”303Id. at 45. Thus, if the Board has concluded “that encouragement or discouragement will result, it is presumed that he intended such consequence.”304Id. (emphasis added). Crucially, the agency could draw upon the “common experiences” of industrial relations to infer unlawful motivation, even where no “independent proof” of it existed in the record.305Id. at 48, 51.

The Court would prominently apply Radio Officers in a pair of landmark cases in the 1960s. In NLRB v. Erie Resistor Corp., the Court found that a “superseniority” plan offered to strike replacements and crossovers was unlawful because it was “inherently destructive” of the right to strike and that no specific evidence of the employer’s motivation was necessary to sustain the Labor Board’s finding of a Section 8(a)(3) violation.306NLRB v. Erie Resistor Corp., 373 U.S. 221 (1963). The employer had grounded its decision to implement the plan upon a discriminatory albeit facially lawful purpose—i.e., the need to attract workers to keep the plant running during the strike—and claimed that the Board must demonstrate “specific illegal intent” to override the employer’s business justification.307Id. at 222–27. The Court rejected that the proffering of a legitimate business reason was an automatic defense to a Section 8(a)(3) charge where direct evidence was absent; the destruction of Section 7 rights remain a “foreseeable consequence” of the challenged conduct, so the agency must weigh the employees’ interests against the interest of the employer in determining the conduct’s legality.308Id. at 228–29. “[B]ecause the Board’s judgment was that the claimed business purpose would not outweigh the necessary harm to employee rights . . . it could properly put aside evidence of respondent’s motive and decline to find whether the conduct was or was not prompted by the claimed business purpose.” Id. at 236–37. The Great Dane Trailers case attempted to distill Erie Resistor’s reasoning into a workable test.309NLRB v. Great Dane Trailers, Inc., 388 U.S. 26 (1967). Where discriminatory conduct is “‘inherently destructive’ of important employee rights[,]” no proof of anti-union motive is necessary and the conduct violates Section 8(a)(3) even if the employer introduces affirmative evidence of business justifications.310Id. at 34. But in cases where “the adverse effect of the discriminatory conduct on employee rights is ‘comparatively slight’” and “the employer has come forward with evidence of legitimate and substantial business justifications for the conduct[,]” the NLRB must produce direct evidence of discriminatory motive.311Id.

Given the rather amorphous concepts of “inherently destructive” conduct versus that which is only “comparatively slight,” the Great Dane Trailers doctrine’s workability was almost immediately called into question.312See, e.g., Leonard S. Janofsky, New Concepts in Interference and Discrimination Under the NLRA: The Legacy of American Ship Building and Great Dane Trailers, 70 Colum. L. Rev. 81, 96–100 (1970). But the arguably more impactful aspect of the case—its treatment of the burden of proof—has endured. In either an inherently destructive or comparatively slight context, the General Counsel’s presentation of a circumstantial case in which an employer’s discriminatory conduct could have affected employees’ Section 7 rights shifts the burden to the employer “to establish that he was motivated by legitimate objectives since proof of motivation is most accessible to him.”313Great Dane Trailers, 388 U.S. at 34. The employer also bears the burden in more commonly litigated individual discrimination cases. Under the Labor Board’s Wright Line “causation” test, the General Counsel bears the initial burden of putting forth evidence sufficient to support an inference that an employee’s protected conduct was a motivating factor in an employer’s decision to discipline.314251 N.L.R.B. 1083, 1089 (1980). Once this prima facie case has been made, the employer bears the burden of demonstrating “that the same action would have taken place even in the absence of the protected conduct.”315Id. The Supreme Court approved of the Board’s causation test in 1983 as a permissible interpretation of the NLRA, rejecting the lower court’s reasoning that Section 10(c) forbade the agency from allocating the burden of persuasion to the accused party.316NLRB v. Transp. Mgmt. Corp., 462 U.S. 393, 400–04 (1983). Justice White was emphatic: “It is fair that [the employer] bear the risk that the influence of legal and illegal motives cannot be separated, because he knowingly created the risk and because the risk was created not by innocent activity but by his own wrongdoing.” Id. at 403.

What connection does this Section 8(a)(3) case law have to the merits of the good-faith doubt test in pre-recognition refusals to bargain? Compare what is required of the NLRB in prosecuting a case under a Great Dane Trailers theory and what is required of it under Joy Silk. Under Great Dane Trailers, the General Counsel must only point to the employer’s conduct and argue it is inherently destructive of employee rights. This is a high threshold, but it is one that can be met in the briefs. At trial, the General Counsel need not present any direct evidence that the employer intended to retaliate against protected activity or that the conduct actually discouraged union membership. The mere occurrence of the conduct means that a discriminatory inference can be drawn as to the employer’s motive. Under Joy Silk, the General Counsel must only point to the employer’s conduct and argue that it either dissipated the union’s majority status or rejected the principles of collective bargaining. Direct evidence establishing the employer’s intent to accomplish these results is not necessary, as the inference fairly follows from the employer’s willingness to commit an unfair labor practice after learning of the union’s majority through its request for recognition.317The “common experiences” of modern labor relations further counsel towards the assumption that subsequent misconduct in the election context was done deliberately to dissipate a union’s majority. See supra notes 219–24 and accompanying text; Anna Stansbury, Do US Firms Have an Incentive to Comply with the FLSA and the NLRA?, Working Paper No. 21-9, 2021, Peterson Inst. for Int’l Econ., 22–30 (concluding that modern application of the NLRA’s remedies is so weak that “it is extremely unlikely that the costs of noncompliance will outweigh the benefits for employers.”); Kate Bronfenbrenner, No Holds Barred: The Intensification of Employer Opposition to Organizing, EPI Briefing Paper No. 235, Econ. Policy Inst., 6, 15 (2009) (finding that 75% of employers in the study’s sample of 1,004 NLRB elections from 1999 through 2003 were “alleged to have committed at least one illegal action” and 40% of these elections involved the filing of at least one unfair labor practice charge); Jon Schmitt & Ben Zipperer, Dropping the Ax: Illegal Firings During Union Election Campaigns, Ctr. for Econ. and Pol’y Rsch., 1 (2009) (finding that by 2005 the probability that a pro-union worker would be fired in an NLRB election campaign was roughly one in fifty-two). In both cases, circumstantial evidence alone requires the employer to affirmatively prove its clear conscience—in Great Dane Trailers, by proffering legitimate business reasons for its discriminatory conduct, and in Joy Silk, by demonstrating its good faith doubt that the union possessed majority support.

To be sure, some commentators have criticized the NLRB’s preoccupation with parties’ motive in adjudicating cases, especially where it relies on inferences of a mindset that is not otherwise in the record.318See Thomas G. Christensen & Andrea H. Svanoe, Motive and Intent in the Commission of Unfair Labor Practices: The Supreme Court and the Fictive Formality, 77 Yale L.J. 1269 (1968). But it is apparent from this quick canvas of hornbook law that independent evidence of an employer’s subjective motivation is not a precondition to sustaining an unfair labor practice, and the good-faith doubt test fits squarely within this paradigm.319It is somewhat ironic that at the same time the good-faith doubt test was falling out of favor within labor law, the “inherently destructive” doctrine was being formulated as a major recalibration of Section 8(a)(3) precedent. Furthermore, a non-circumstantial test will not involve less conjecture. As Petruska has persuasively argued, a fatal flaw of Gissel is that it requires judges to speculate about future events—specifically, the NLRB’s ability to administer a fair election if only traditional remedies are applied.320Petruska, supra note 15, at 112–15. With changed circumstances often present by the time a case has reached the courts321See supra notes 209–12 and accompanying text. and the omnipresent temptation under Gissel to defer to the “preferred” system of elections, there is no reason why its forward-looking test would create more determinacy for bargaining orders than the backward-looking test of Joy Silk, which at least resembles the standard format for remedial analysis under Section 8(a)(5). If Joy Silk is a logical thicket, then Gissel is a veritable briar patch, a territory that wrongdoers have learned to nimbly navigate.

3. The Policy Objection

Finally, there is no basis for the complaint that the good-faith doubt test placed employers in an impossible quandary. The grand metaphor of Scylla and Charybdis must not be conceded, for the law of recognition under the “pure” Joy Silk doctrine did not resemble the trials and tribulations found in The Odyssey; in reality, it was a largely intuitive process. In his much-discussed memorandum to the Senate labor subcommittee written in defense of the Kennedy Board, Secretary Wirtz described the employer’s journey with decidedly less flourish than Homer:

[U]nder the Board’s practices and policies, an employer who questions whether cards reflect the true choice of his employees does not have to abide by those cards. Nor does he have to seek out evidence that the cards do not reflect the employees’ true choice. All he has to do is file a petition for an election, if the union does not, refrain from unfair labor practices, and then wait for the Board to determine the representation issue through a secret-ballot election. Of course, during the preelection period, the law allows an employer to express to employees his views, arguments, or opinions on the subject of unionization, if he makes no threat of reprisal or force or promise of benefit and does not otherwise improperly interfere with the free choice of his employees. “Where an employer departs from these standards or demonstrably has no doubt of the union’s majority, the Board’s reliance upon card-established majorities seems clearly warranted and is approved in the courts.”3221968 Hearings, supra note 91, at 1416–17.

The accuracy of Secretary Wirtz’s synopsis can be demonstrated through a counterfactual. If an employer’s duty to recognize a union was a legal quagmire in the 1960s and capable of significant exploitation by unions, we would expect to see litigation on this issue at a volume that threatened to subsume the NLRB’s election processes. After all, why would a union that can so easily trap an employer into a card check through manipulation of favorable case law ever willingly risk its majority in an election? But statistics from this era show that Joy Silk saw very little use compared to the Labor Board’s election machinery. While card-based bargaining orders increased in magnitude under the Kennedy Board,323See Gross, supra notes 96–97 and accompanying text. the 107 such cases in fiscal year 1967 consumed barely one percent of the total number of situations in which the Board confronted a question of representation, as 8,193 secret-ballot elections were conducted in that year alone.324Gordon, supra note 267, at 203–04. To paint an even clearer picture, Gordon reported that just 365 of the NLRB’s 43,247 cases involving representation in the five-and-a-half years preceding his article resulted in a bargaining order premised on a card majority. Id. at 222. The remaining total, comprising over 99% of the sample, were resolved through elections. Id. Moreover, Secretary Wirtz unearthed only ten total cases since the creation of the Joy Silk doctrine in which the agency ordered bargaining solely based on an employer’s lack of good faith doubt in the union’s majority status without the presence of independent unfair labor practices—in other words, the Snow & Sons line of cases.3251968 Hearings, supra note 91, at 1415–16. The Secretary was thus on solid ground when he confidently declared that “[e]lections are the rule. Card checks are rare exceptions.”326Id. at 1412. See also Thompson, Jr. Statement, supra note 268, at 23973 (“the Board conducts 49 or 50 secret ballot elections for every ‘Joy Silk’ bargaining order that it enters. The secret ballot NLRB election remains the standard statutory method for registering the worker’s choice whether he wishes to be represented for purposes of collective bargaining.”). Secretary Wirtz, writing in criticism of the Eisenhower Board’s perceived weakening of the Joy Silk doctrine, once flipped the “magic words” attack on its head by pointing out that an employer could escape its statutory duty to recognize even a “clearly selected employee representative” with the right response to the union’s request. W. Willard Wirtz, Two Years of the New N.L.R.B., 1955 A.B.A. Sec. Lab. Rel. L. Proc. 4, 11 (1955).

It defies belief that roughly 99% of employers presented with a request for recognition in this time frame already possessed able counsel or received sound advice from attorneys regarding their legal options, so one must conclude that either unions were consciously foregoing an ostensibly guaranteed route to exclusive representation and bargaining rights, or these seas were much more navigable for employers than suggested.327One must also consider Professor Wolkinson’s suggestion that voluntary recognition in potential Joy Silk fact patterns kept at least some cases from becoming litigation. Wolkinson, supra note 219, at 22–25. But assuming that voluntary recognition was not so widespread as to significantly alter the ratios on record, any incentive provided by Joy Silk towards the private creation of collective bargaining relationships is simply another reason why the NLRB should consider re-adopting its good-faith doubt test. Here, Joy Silk’s sparse utilization versus the NLRB’s election machinery belies a massive influence on the latter’s components. The reduced rates at which employers drew unfair labor practice charges when the good-faith doubt test was controlling law provided fewer opportunities for the Board to consider a bargaining order.328See discussion supra Part II.B. By “fewer opportunities,” I mean on average. The NLRB conducted just 1,619 elections involving 96,964 employee voters in fiscal year 2009. 74 NLRB Ann. Rep. 1 (2009). In fiscal year 1967, these numbers were 8,183 and 628,730, respectively. 32 NLRB Ann. Rep. 1 (1967). Given the proliferation of unfair labor practice charges that followed immediately after the NLRB relinquished the good-faith doubt test, it appears there was a widespread understanding among employers in the 1950s and ‘60s that violating the NLRA in the election context could have grave (and expensive329Wolkinson, supra note 219.) consequences.

Therein lies the most compelling policy reason for resurrecting the Joy Silk doctrine and the best response to Professor Lesnick’s befuddlement regarding the Labor Board’s monitoring of employers’ good faith.330See Lesnick, supra note 80, at 858–59. Under its General Shoe doctrine, the Board requires that elections be conducted under a standard of “laboratory conditions” that “enable[s] employees to register a free and untrammeled choice for or against a bargaining representative.”331General Shoe Corp., 77 N.L.R.B. 124, 126–27 (1948). The Gissel test, which intrinsically restricts the volume of bargaining orders through classification as an “extraordinary” remedy, places an inexorable strain on General Shoe by “routinizing a baseline level of [unfair labor practices] that will be deemed ordinary, or at least not extraordinary, in the administration of the Act.”332Petruska, supra note 15, at 140–41. Joy Silk’s restriction of election misconduct resulted in a more proper calibration of laboratory conditions and was, therefore, the more faithful application of the NLRA.333See id. at 138–43. The utter collapse of the bargaining order remedy despite the persistence of the Bernel Foam and Cumberland Shoe doctrines post-Gissel further illustrates that good faith doubt is the missing pillar of the Board’s structure, and the foundation upon which any restoration effort must be built.334See Air 2, LLC, 341 N.L.R.B. 176, 189 (2004) (reaffirming Bernel Foam); DTR Indus., 311 N.L.R.B. 833, 838–40 (1993) (reaffirming Cumberland Shoe). As Representative Thompson, Jr. eulogized, the Joy Silk doctrine was a “finger-in-the-dike” for the country’s labor relations, “not to be measured by the mere handful of cases in which it is applied, but by the countless hundreds of cases in which employers hesitate to coerce and threaten their employees because of the knowledge that this remedy stands as a legal safeguard of employee rights.” Thompson, Jr. Statement, supra note 268, at 23974.

The NLRB’s inquiry into motive in the pre-recognition context was statutorily permissible, logically consistent, and effectively deterrent. With this history in mind, the question becomes why would the Board not return to an analysis of good faith doubt as refuge from the current failed regime? Perhaps Lesnick put it best when he pondered—in rhetorical fashion—why the Board would voluntarily abandon its “winning formula.”335Lesnick, supra note 80, at 858.


In lieu of excessive drumbeating, this Article will conclude with two brief comments on ancillary issues. First, it is indisputable that the NLRB may discard the case law carved out in Gissel and Linden Lumber to re-adopt the Joy Silk doctrine. Neither case proposes to resolve what the Labor Board must do in the realm of recognition requests.336See Morris, supra note 287, at 10–15. The Supreme Court expressly refused to consider the matter of good faith in Gissel on the assurances of the agency’s advocate,337NLRB v. Gissel Packing Co., Inc., 395 U.S. 575, 594 (1969). and Linden Lumber merely approved of the Board’s rule as a permissible reading of the NLRA in the most reserved language possible.338Linden Lumber Div., Summer & Co. v. NLRB, 419 U.S. 301, 309–10 (1974) (“In light of the statutory scheme and the practical administrative procedural questions involved, we cannot say that the Board’s decision that the union should go forward and ask for an election on the employer’s refusal to recognize the authorization cards was arbitrary and capricious or an abuse of discretion.”); accord Florian Bartosic, The Supreme Court, 1974 Term: The Allocation of Power in Deciding Labor Law Policy, 62 Va. L. Rev. 533, 589 (1976) (“In view of the very narrow standard of review applied by the Supreme Court majority in Linden . . . a newly constituted Board could in the near future . . . return to the Kennedy Board’s independent knowledge test, or for that matter, to the Joy Silk ‘good faith doubt’ doctrine.”). The Board has successfully overruled precedent that previously earned Supreme Court approval,339Petruska, supra note 15, at 106 (noting the NLRB’s replacement of R.J. Smith Construction Co., 191 N.L.R.B. 693 (1971) with John Deklewa & Sons, 282 N.L.R.B. 1375 (1987)). and the Court’s acknowledgment of the good-faith doubt test in Arkansas Oak Flooring can pave the road for the attorneys who litigate the doctrinal transition. The administrative argument should be relatively simple: the Board is entitled to deference in its formulations of national labor policy,340Fall River Dyeing & Finishing v. NLRB, 482 U.S. 27, 37 (1987). this deference extends even to departures from prior policy,341NLRB v. Curtin Matheson Sci., Inc., 494 U.S. 775, 787 (1990). and Joy Silk itself represented a permissible interpretation of the statute.342United Steelworkers of Am. v. NLRB, 376 F.2d 770, 773 (D.C. Cir 1967). There is abundant evidence for the Board to determine “in light of changing industrial practices” and the agency’s “cumulative experience in dealing with labor-management relations” that the need exists to reclaim the good-faith doubt test and again enforce Section 8(a)(5) in the pre-recognition context.343NLRB v. J. Weingarten, Inc., 420 U.S. 251, 266 (1975).

Second, Joy Silk was not a silver bullet in its time and thus cannot be one today. While the good-faith doubt test was adept in preventing unfair labor practices and establishing collective bargaining relationships, the remedy had much less influence in securing first contracts between unions and employers. One contemporary study found that while 49% of the unions that benefited from a Joy Silk bargaining order were able to eventually obtain a collective bargaining agreement, a relatively encouraging number,344Wolkinson, supra note 219, at 21–22. this figure fell to 23% in cases where employers exhausted their appeals in litigation.345Id. at 22. The author of that study recommended that the NLRB adopt a “make-whole” compensatory remedy for any losses employees sustain as a result of their employer’s failure to bargain in good faith, which was then under consideration by the agency.346Id. at 39–41; see also Paul R. Baier, Rights Under a Collective Bargaining Non-Agreement: The Question of Monetary Compensation for a Refusal to Bargain, 47 J. Urb. L. 253, 262–63 (1969); Note, Monetary Compensation as a Remedy for Employer Refusal to Bargain, 56 Geo. L. J. 474, 501 (1968). Despite the D.C. Circuit determining that such relief in Section 8(a)(5) cases is “necessary to implement the goals of the Act,”347United Auto Workers v. NLRB, 449 F.2d 1046, 1049 (D.C. Cir. 1971); accord Int’l Union of Elec. Workers v. NLRB, 426 F.2d 1243, 1248–53 (D.C. Cir. 1970). the Labor Board has refused to exercise this power.348Ex-Cell-O Corp., 185 N.L.R.B. 107 (1970). For brief histories of the Ex-Cell-O litigation, see Gross, supra note 83, at 224–25; Kenneth C. Hunt, Note, NLRB Remedies—Attorney’s Fees in Refusal-to-Bargain Cases, 1975 Duke L.J. 352, 355–58 (1975). General Counsel Abruzzo, likely realizing the natural pairing of this remedy with Joy Silk, included such make-whole relief in her debut memorandum.349Abruzzo Memo, supra note 18, at 8.

This Article will end where it began: by admiring the language of Frank Bros. The Supreme Court could not have been clearer in its unanimous decision that the NLRB possessed the authority to issue bargaining orders and was to be given maximum flexibility in honing the remedy to counter “frustrations of the Act” and “expunge[]” the unlawful conduct of “recalcitrant” employers.350Frank Bros. Co. v. NLRB, 321 U.S. 702, 704–05 (1944). The economic and legal climate affecting the core concern of Wagner Act-era industrial relations—that is, the unwillingness of employers to collectively bargain with the majority representative of their employees—has not changed precipitously since the time these words were written. Indeed, defiance of the NLRA remains robust, and private-sector bargaining relationships are at their most endangered since the statute was passed. The concessions made to the labor bar in the late 1960s described in this Article proved unworkable in practice and placed the Labor Board in a permanent defensive posture regarding its bargaining order remedy. Reviving the good-faith doubt test of Joy Silk and enforcing Section 8(a)(5) as written would better encapsulate the bargaining orders envisioned in Frank Bros. and equip the Board with a tool historically proven to prevent misconduct in elections. Few initiatives in General Counsel Abruzzo’s agenda could prove more effective in carrying forth the Board’s statutory mission.

*J.D. 2018, University of Kentucky College of Law. Member, Indiana Bar. In the course of writing this Article, the author was hired as a field attorney in the National Labor Relations Board’s Indianapolis regional office. The following analysis represents the opinions and views of the author alone, and does not constitute, nor should it be construed as, representing the views of the National Labor Relations Board, its General Counsel, or any of its Regional offices.