The Failed Promise of Statutory Protection

The subject of the legal regulation of labor is one of great complexity. Up to the present time a priori objections to such regulations have delayed their introduction, and only gradually, as experience has demonstrated their usefulness, have they been extended to situations which seem to require them. In … the United States the notion that the legislative power should not be used … to regulate conditions of employment has been abandoned by most thoughtful persons, but the prejudice against interference … is as strong as ever.

Henry R. Seager, Economics, 1904, p. 431

Following a period of legislative inaction, selective statutory restrictions on the right to dismiss came into existence largely as a byproduct of labor legislation of the late 1920s and early 1930s. The introduction of limitations to the at-will rule within the NLRA framework, in particular, marked the long overdue recognition that, as long as employers had the right to dismiss employees, at-will public policy goals, such as industrial peace and the extension of orderly collective bargaining, were unattainable.

Following a roughly historical chronology, this chapter explores how, from the 1920s onwards, restrictions on dismissals were constructed around notions of “orderly” collective bargaining. Thematically, the focus of the chapter is on the creation of new institutional structures and their impact on the status of workers in terms of job security. Underlying this analysis is the tentative hypothesis that the NLRA, and the practices which evolved from it, provided unions and their members with a sense of control over dismissal rights which was largely illusionary. This mistaken sense of control, in turn, encouraged unions to put efforts into job security enhancing measures at the plant and company level which ultimately did not constrain managerial prerogatives effectively. This lack of real control became apparent in the mid 1960s, when the Supreme Court handed down several decisions which reaffirmed the right of management to close branches and discharge employees without union interference. Apart from excluding non-unionized workers, the NLRA system, perhaps against the intentions of its original sponsors, ultimately came to severely circumscribe the right of unions to bargain over job security at the very time when such protection was needed.

The Promised Lands of Protected Bargaining

At the turn of the century, many US industrial relations scholars questioned the assumption that injustices in the labor market could be remedied through legislative acts and/or, more generally, via a strengthening of individual employment rights. Opposition to legislative approaches was grounded primarily in the belief that solutions to the “labor problems of industrial societies” could be created more easily by strengthening the standing of organized labor as collective bargaining agent rather than by creating a host of specific employment regulations.[1] Accordingly, in 1911, the Harvard economist Taussig suggested that the most urgent task in reforming US employment relations was not detailed new legislation per se, but rather the protection of bargaining representatives:[2]

The workmen clearly gain by having their case in charge of chosen representatives, whether or not these be fellow employees; and collective bargaining and unionization up to this point surely bring no offsetting disadvantages to society. As to the immediate employees, there is often a real danger that he who presents a demand, or a grievance, will be “victimized.” He will be discharged and perhaps blacklisted; very likely on some pretext, but in fact because “he has made trouble.”

In the 1930s, Taylor’s influential Labor Problems and Labor Law argued, very much along the lines of earlier reform advocates, that individual workers had been deprived of their ability to bargain primarily because of the expansion and centralization of management.[3] To remedy this situation, Taylor argued, the state had to enable workers to bargain collectively, both for wages and for the protection of their jobs. Said Taylor:[4]

Legally free to dispose of his services at any price he deems just, immediate necessity in the face of an oversupply of labor reduces that freedom to empty words. His [meaning the worker’s] inferior bargaining position is not wholly due to economic inequality, but in part to a lack of knowledge of labor conditions, and a bargaining skill less effective than that of his employer. The injustices growing out of the individual bargaining burden affect not only the individual worker but the entire group to which he belongs. Unregulated competition resulting from individual bargaining tends to pull down the terms of employment to the level of the weakest employer…

Taylor’s notion that inequalities of labor were due to the exposure of workers to individual rather than collective bargaining echoed the opinions of some of the nation’s leading judges of the time. Judges Holmes and Field had earlier opposed bans on union activity on account of the fact that union activity merely counterbalanced the combination of capitalists.[5] Despite the gradual acknowledgement of the legitimacy of strike action by some courts, up until the 1920s, few judges had been willing to offer protection to those workers who were discharged for union membership or strike activity. In theory, collective bargaining could serve to limit the power disequilibrium between the employer, who, as Holmes says “is free to discharge the worker, and the worker who depends on his job for his livelihood.”[6] In practice, however, the relationship between job security and collective action had remained largely antonymous. Post World War I, workers who participated in collective action, be it as organizers or as strike participants, were likely to face retaliatory discharges or even blacklisting.[7] Industrial actions in which in excess of 1,000 workers were permanently dismissed included the Homestead strike of 1892, the Pullman strike of 1894, and the steel strike of 1919-20, which involved approximately 365,000 workers and resulted in over 10,000 permanent discharges. In the Boston police strike of 1919, in which the policemen struck for the right to organize with an AFL affiliate, meanwhile, more than one third of the police force were permanently discharged.

The first congressional statute addressing issues of dismissal and organizing activity, the Erdman Act, had attempted to prohibit the retaliatory discharge of union members working on the railroads; at a time when the railroads were the only area where the Federal Government had the authority to regulate such matters. Passed by Congress in 1898, Section 10 of the Erdman Act made it an offense to threaten an employee “with discharge” or to blacklist the employee after a discharge because of membership in a labor organization. Specifically the Act read: [8]

That any employer subject to the provisions of this act and any officer, agent or receiver of such employer who shall require any employee, or any person seeking employment, as a condition of such employment, to enter into an agreement, either written or verbal, not to become or remain a member of any labor corporation, association, or organization; or shall threaten any employee with loss of employment, or shall unjustly discriminate against any employee because of his membership … or who shall, after having discharges an employee, attempt or conspire to prevent such employee from obtaining employment or who shall after the quitting of an employee, attempt or conspire to prevent such employee from obtaining employment, is hereby declared to be guilty of a misdemeanor, and … shall be punished for such offense by a fine of not less than one hundred dollars and not more than one thousand dollars.

In 1908, section 10 of the Erdman Act was declared in violation of the Fifth Amendment by the Supreme Court in Adair v. United States. This rather predictable decision again rendered members of labor organizations unprotected from retaliatory discharges.[9]

Unionized workers were given some support by the courts in the Brandeis and Holmes Supreme Court decisions of the 1920s.[10] Explicit legislative protection of those engaging in organizing activity however commenced as late as 1926 with the passage of the Railroad Labor Act (RLA), which, apart from requiring employers to bargain with unions, prohibited employers from discriminating against union members.[11] The RLA applied originally to interstate railroads and related undertakings, but was later amended to include airlines engaged in interstate commerce. The Norris La Guardia Act (NLGA) of 1932 gave some federal sanction to the right of labor unions to organize and strike.[12] Implicitly, it also limited the ability of federal courts to enforce “yellow dog contracts,” under which workers promised not to join a union or promised to discontinue union membership.[13] The National Industrial Recovery Act (NRA) of 1933, the predecessor of the National Labor Relations Act, introduced the idea of codes of “fair competition” which fixed wages and hours in certain industries. Title I of the Act, which was declared unconstitutional in 1935, guarantied the right of employees to collective bargaining without interference or coercion (which included the dismissal of employees). [14]

The National Labor Relations Act (NLRA) of 1935, or Wagner Act, included some previously invalidated labor sections of the NRA, as well as a number of additions. Primarily concerned with restricting employer activities against union organizing and bargaining efforts, the NLRA prohibited employers from, firstly, “dominating or otherwise interfering with the formation of labor unions”; secondly, “interfering or restraining employees engaged in exercising their rights to organize and bargain collectively; and, thirdly, from “refusing to bargain collectively with unions representing a company’s employees.” In doing so, sections 7 and 8 of the NLRA effectively tied the legal protection of employees from retaliatory discharges to the right of employees to organize collectively. The Act stated to this effect that:[15]

Sec. 7. Employees shall have the right to self-organization, to form, join, or assist labor organizations, to bargain collectively through representatives of their own choosing, and to engage in concerted activities, for the purpose of collective bargaining or other mutual aid or protection.

Sec. 8. It shall be an unfair practice for an employer—

(1) To interfere with, restrain, or coerce employees in the exercise of the rights guaranteed in section 7.

(2) To dominate or interfere with the formation or administration of any labor organization or contribute financial or other support to it…

(3) By discrimination in regard to hire or tenure of employment or any term or condition of employment to encourage or discourage membership in any labor organization…

(4) To discharge or otherwise discriminate against an employee because he had filed charges or given testimony under this act.

(5) To refuse to bargain collectively with the representatives of his employees…

Under the NLRA regime, employers were required “not to refuse to bargain collectively with the representatives of his employees” with regard to “rates of pay, wages and hours of employment, or other conditions of employment.”[16] While the Act had made it clear that retaliatory dismissals of union members were illegal, it gave no guidance on the question of whether bargaining over “other conditions of employment,” included issues relating to job security.[17] Moreover, despite the appearance of sweeping legislation, coverage under the NLRA’s protective umbrella was narrow. Public employees at the federal, state, and local level, agricultural workers, domestic workers, and supervisory employees all were excluded.[18] Nonetheless, for those covered by the Act, statutory dismissal protection was available in connection with established categories of protected activity the courts had created. This included dismissals for strike action, union membership and related activities.

Indeed, at its outset, the NLRB rulings allowed significant numbers of dismissed employees to gain reinstatement. From the appointment of the Board in the Fall of 1935 until March 1939, the Board handled a total of 20,192 cases involving over 4.5 million workers. Of these cases 19,018 or four fifths were closed. Of the total cases closed, about 52% were decided by agreements, while the remainder were dismissed, withdrawn or closed in some other way before coming to the Board. About two thousand cases were strike cases, involving 356 thousand workers, of which 75% were settled and in which 227 thousand workers had to be re-employed. An additional 15 thousand cases were decided in favor of workers alleging non-strike related discriminatory discharges, and resulted in the reinstatement of the respective workers. Between January 1 of 1938 and April 1 of 1939 alone, the Board heard 1,675 cases alleging discriminatory discharges and ordered the reinstatement and/or compensation of 1,022 workers.[19]

In theory, there was a potential for collective bargaining agreements to include job security guarantees of some form. Given existing cultural pre-dispositions, both amongst the judiciary and managers, however, the possibility of partial union control over personnel and investment decisions was remote. Judicial support for the right to manage had a strong pedigree and its influence would not wane quickly. In the 1890s already, some state courts had felt the need to defend the right to manage. In the view of most courts this right was as much a part of the free labor creed as was the right to work. “Free labor” required that both employers and individual workers were fully responsible for their decisions. Permitting workers to organize and successively influence managerial decisions was viewed as a danger to free economic competition. In State v. Glidden, an outraged Connecticut judge stated, that once workers could influence managerial decision, no longer would the heads of industrial and commercial enterprises rise from the “ranks of the toilers, no longer could self-reliant ambitious men push to the fore.”[20] Unable to manage as they saw fit, businessmen would stop risking their capital, time and experience. “At best, the nation’s business would be conducted by paternalistic enterprises, at worst anarchy pure and simple …” would prevail.

At the turn of the century, Taussig had already predicted that union demands for job security would clash with managers’ insistence on “the right to manage.” His Principles of Economics stated to this effect that:[21]

Private ownership carries with it the seeds of conflict–the inevitable clash between those who employ and who are employed. Disguise it as we may, smooth over to our utmost, adjust where we can, there the conflict is, ever liable to break out. … The private employer … regards his business as his own, its methods of management as subject to his own judgment. It is almost invariably urged by him and his spokesman that the effective working of the business machine depends above all on unfettered freedom in the selection and tenure of employees. So long as this attitude prevails, the workman will feel in turn that he must retain his weapon of defense, the strike, even though it entail injury to a wide circle of persons. … Even if employers were to consent to restrictions on their power of discharge, contests would remain, strikes would brew. And on the other hand discharge is but one of the matters in which employers absolute rule is to be questioned. Discharge is conspicuous because it is the outstanding weapon.

As long as unions and their members had little formal protection through the law, management had been able to assert its dominance with relative ease, if only by dismissing those who questioned it. Once NLRA legislation protected concerted action, this situation had changed radically, and conflicts between unions and management over dismissal rights were pre-destined.

When President Truman called the second National Labor Management Conference in 1945, labor and management representatives found themselves unable to agree on the boundaries of collective bargaining. Disagreement had arisen particularly with regard to management’s right to make workers redundant, close and/or relocate branches. The statement of the management representative at the conference expressed the employers dismay over this matter:[22]

Labor members of the Committee on Management’s Rights to manage have been unwilling to any listing of specific management functions. Management members of the Committee conclude … therefore, that the labor members are convinced that the field of collective bargaining will, in all probability, continue to expand into the field of management.

The only possible end of such a philosophy would be the joint management of the enterprise. To this management members naturally cannot agree. Management has functions that must not and cannot be compromised to the public interest. If labor disputes are to be minimized, … labor must agree that certain specific functions and responsibilities of management are not subject to collective bargaining.

In theory, the evolving conflict about the appropriate limits of collective bargaining, and particularly the rights of labor to interfere with management’s redundancy and dismissal decisions, was resolved by reference to new management concepts such as the residual rights doctrine. In practice, a set of employer friendly court decisions and the decline of unions in the US settled the issue, first, in rough terms, during the first decade of NLRA rule, and then, in greater detail, over the following three decades.

The notion of residual rights, which deserves a passing mention in this context, developed from the 1940s onwards to become a prominent feature of the management of industrial relations in the 1960s and 1970s. The residual rights doctrine postulated that management rights were the result of an evolutionary process, whereby initially management possessed total freedom in ordering the affairs of the enterprise. This included freedoms with regard to whom to hire and dismiss and when to do so. Union demands and labor legislation encroached on this freedom. It followed that every time a manager made a contractual concession, and/or every time a labor law restricted management options, the original rights of management were reduced. What remained then were the residual rights, not specifically renounced by management or restricted by law.[23] If, for instance, management renounced the right to dismiss according to productivity or any other performance criterion and agreed to dismiss according to seniority, seniority replaced management’s previous decision criteria. Meanwhile other issues, such as how many workers could be dismissed in a specific time period, remained within the exclusive sphere of managerial decision making.[24]

Adopting this view, many arbitration decisions applied a two-stage approach to questions about the appropriate bargaining remit of a union. If union representatives and management disagreed on whether an issue was a legitimate bargaining item, previous contractual agreements as well as legal requirements had to be investigated. If no explicit statement restricting management’s rights in the respective matter could be found in these sources, the issue typically had to be considered as falling within management’s remit. Since explicit renunciations of the rights to dismiss were typically rare, management usually maintained broad discretion over dismissals, which fell outwith causes covered explicitly by just-cause rules.

Because existing practices and informal agreements had little legal bearing on conflicts over the interpretation of the NLRA, the residual rights doctrine offered almost no guidance to the courts in evaluating the legitimacy of union involvement in termination decisions. Here an alternative, and in many ways even more restrictive approach, evolved over time. While the NLRB of the early years generally looked favorably upon workers whose discharge could in some way be linked to union activity, it also condoned a wide set of permissible grounds for dismissal. In this context, several NLRB decisions early on vindicated traditional assumptions about managerial prerogatives. Discharges were sustained by the NLRB in cases involving gross inefficiency of a worker, incompetence, change in equipment, “ruckus and horseplay”, absenteeism, brawling, cursing of the boss, and the violation of company rules.[25] Most importantly, discharges in the absence of employee misconduct were frequently declared permissible if there was no evidence for anti-union activity. This included discharges for lack of work, which were generally approved by the Board even in absence of union consultation, as long as anti-union bias could not be proven. In its Seagrave decision of 1938, for instance, the Board set a precedent for the preservation of employment-at-will within collective bargaining.[26] Seagrave, an automotive equipment plant had discharged an employee three weeks after he got his job. The foreman testified to the fact that the employee’s work was satisfactory. The worker, a CIO member, had previously been arrested for disorderly conduct during a strike and alleged that he was fired because of this previous involvement, and, more specifically, because his foreman had received a blacklist showing his name. The spokesman of the company explained that the polisher was hired because of a temporary emergency arising from the receipt of a special order, and that he was dismissed when the work on that order let up. The Board found no evidence for anti-union activity and declared the dismissal legal.

In the case of Sheba Ann Frocks (1938), similarly, thirty employees, who had been dropped from the payroll of the Sheba garment plant, complained to the Board alleging that their discharge was based on their CIO membership.[27] Company officials testified that the layoffs took place because of a lack of work at the end of the regular production season. The Board accepted this explanation because the company retained over half of its CIO employees and discharged non-union employees as well, although not proportionally. In its conclusion the Board stated that, in the case of a dismissal for legitimate business reasons, such as slack work, no consultation with union members was required.

While NLRB decisions of the late 1930s, such as Seagrave and Sheba, delineated the space between dismissal protection and managerial prerogatives more or less by default, several court decisions attempted to give guidance which was general enough to be applied to other contexts. This tendency towards establishing a formula which ringfenced managerial decision making from union intrusion could already be detected in the Supreme Court’s ruling on NLRB v. Jones & Laughlin Steel, the landmark case better known for its acceptance of the NLRA. In Jones, the Supreme Court stressed that although the Act required bargaining, it did not “compel” agreement.[28] For the Supreme Court, in other words, the NLRA was legal because, and only because, the Act did not interfere with “the normal exercise of the right of the employer to select employees or to discharge them.”[29] That, in defining normal rights, the Supreme Court emphasised the right to discharge workers did not bode well for those who expected the Act to significantly reduce arbitrary dismissals. With Jones, the court had indicated that outwith matters directly related to collective bargaining, employment-at-will was still very much in place, with restrictions only affecting those discharges which were explicitly declared illegal in the NLRA. More importantly, it had implied that would be difficult to create an agreement sanctioned and protected by the Act which would eliminate the right of employers to discharge workers for “legitimate” reasons.

In NLRB v. Sands Manufacturing (1938), a federal appeals court was even more explicit in affirming management’s freedom to dismiss workers.[30] In Sands, a collective agreement between the company and MESA, a labor union, was broken by the union. The company apparently bargained collectively with MESA. After two months, the company signed an agreement with another union, some of whose members were employed in order to replace MESA members. The NLRB ordered reinstatement of the MESA employees and requested the circuit court to enforce its order. The 6th circuit set aside the order and dismissed the petition to enforce. With respect to the termination of the employer-employee relationship the court stated that:[31]

The statute [meaning the NLRA] does not interfere with the normal right of the employer to select or discharge his employees … If employees violate their contract they may be discharged for that reason and this does not constitute a discrimination in regard to tenure of employment nor an unfair labor practice, nor does it continue a discharge because the employees are members of a union. … [T]he statute does not provide that the relationship held in status quo under Title 29, Section 152(3) [meaning the prohibition of dismissals during strikes] shall continue in absence of wrongful conduct on the part of the employer and of rightful conduct on the part of the employees. If such were its meaning, the right of the employer to select, and discharge his employees … would be cut off.

The Sands decision was in many regards more radical than previous rulings. In Sands, the court had concluded that, provided the employer had engaged in bargaining, NLRA legislation had to be interpreted so as not to otherwise constrain the employers’ rights to select and discharge employees. In other words, the court indicated that any action which would effectively restrict the right of employers to discharge, after basic bargaining obligations were met, could be struck down.

While both the Jones & Laughlin Steel and the Sands cases redefined space for at-will discharges relatively broadly, the Supreme Court’s 1942 Montgomery Ward decision attempted to give a comprehensive definition of management’s rights which gave managers broad control over discharge decisions.[32] In its Montgomery Ward decision, the 9th Circuit excluded from arbitrable grievances:[33]

… changes in business practice, the opening and closing of new units, the choice of personnel (subject, however to the seniority provision), the choice of merchandise to be sold, and other questions of a like nature not having to do directly and primarily with the day-to-day life of the employees and their relations with supervisors.

Although Montgomery Ward supported traditional concepts of management rights with respect to day-to-day arbitration, it left open a number of important questions with regard to dismissals arising as a consequence of longer term strategic decisions. This included questions relating to the dividing line between a rational business decision to relocate a plant, and one involving, for example, the elimination of a unionized plant–an illegal antiunion activity. Moreover, the Court’s decision to exclude changes in business practice from arbitrable grievances, merely prohibited unions from insisting on arbitration in these matters; and hence relieved management from the legal duty to discuss these matters in good faith. This did neither mean that union representatives could not bargain about these issues when contracts were negotiated, nor did it imply that once management conceded to union involvement in these matters, this involvement was illegal or unenforceable.

The latter issue of bargaining about alleged management prerogatives was addressed first in 1952 in NLRB v. American National Insurance Group.[34] In American National, the Supreme Court held that management could enforce limits to bargaining on the basis of a management prerogative clause, under which the union was ousted from involvement in certain matters. American National’s management prerogative clause included issues of discipline and work schedules; that is, statutory rights with respect to mandatory bargaining. The court, nonetheless, rejected the Board’s position that employers were obligated to establish ongoing bargaining during the terms of the collective agreement on issues subject to defined managerial prerogatives.

While in American National the company had attempted to impose broad limitations on bargaining rights, many companies insisted “only” on the type of management prerogatives listed in the Montgomery case, such as the freedom to decide on the closure of units. In the mid-1950s, Haber and Levison reported that over 80% of the contracts signed in the building industries contained one or another form of a managerial rights clause. Many of these clauses explicitly prohibited bargaining over issues of job security.[35] The management literature, meanwhile, welcomed American National because companies were now less likely to face NLRA proceedings if they refused to discuss issues of employment security. This was the case, not only where companies had gained past assurances that union representatives would respect managerial prerogatives, but also where such clauses could be “inferred” from existing bargaining agreements.[36]

Management rights in matters of dismissals and layoffs were “clarified” further in the 1958 Supreme Court decision on Borg-Warner. In NLRB v. Wooster Division of Borg-Warner the Court held that there were three subjects of bargaining: mandatory, nonmandatory, and illegal.[37] The obligation to bargain, as specified in the NLRA, applied only to mandatory subjects. A nonmandatory subject was “permissive,” meaning that it could be raised by either party. However, when a party insisted on a position regarding such an area to the point of impasse, it was acting illegally under the provisions of the Act.[38] Since the law had defined the mandatory subjects of bargaining, Borg-Warner played an important role in the preservation of managerial prerogatives with regard to redundancies and dismissals. Under Borg-Warner, union demands for job security or employment guarantees could be rejected, as they could not be reasonably classified as mandatory bargaining items.[39]

When determining what were mandatory and nonmandatory bargaining subjects, the NLRB and the courts of the 1950s and 1960s typically referred to the relevant NLRA section 9(a) which mandated bargaining for pay, wages, hours of employment, and other conditions of employment. Given these specifications, any issue involving pay and hours was obviously a mandatory bargaining item, requiring both parties to bargain in good faith or face sanctions through NLRB proceedings. More problematic was the clause including, “other conditions of employment.” When issues like redundancies, mass layoffs and mass discharges were at stake, the courts and the Board usually interpreted “other conditions of employment” to mean that union involvement in decisions about which workers were to be laid off or made redundant, was mandatory. To this effect union representatives were to be informed about planned manpower reductions. Union representatives were free to address issues related to discharges, make suggestions with regard to manpower relocation, or suggest alternative ways of cutting costs. If the company refused, unions, however, could not insist on a settlement of the issue. While strike action relating to these matters was not per se illegal, any protracted industrial action on non-mandatory manpower issues was likely to be declared an unfair labor practice by the NLRB or the courts.[40] This approach, needless to say, gave unions with little power to influence a company’s manpower decisions even in industries where levels of organization were high. Since it was often difficult to link a redundancy decision to union avoidance or to invoke contractual clauses which