delivered the opinion of the Court.

We decide today whether, at the end of a strike, an employer is required by the Railway Labor Act (RLA or Act), 44 Stat. 577, as amended, 45 U. S. C. § 151 et seq., to displace employees who worked during the strike in order to reinstate striking employees with greater seniority.

HH

In March 1984, Trans World Airlines, Inc. (TWA), and the Independent Federation of Flight Attendants (IFFA or *429Union) began negotiations pursuant to §6 of the RLA, 45 U. S. C. § 156, on a new collective bargaining agreement to replace their prior agreement due to expire on July 31, 1984. The existing collective bargaining agreement created a complex system of bidding the general effect of which was to insure that those flight attendants with the greatest seniority would have the best opportunity to obtain their preferred job assignments, flight schedules, and bases of operation as vacancies appeared, and to insure that senior flight attendants would be least affected by the periodic furloughs endemic to the airline industry. Thus, for example, should a job vacancy appear at the highly desirable Los Angeles or San Francisco bases of operation or “domiciles,” the most senior qualified flight attendant who bid on such a vacancy would be entitled to it. Conversely, should a reduction in force eliminate a position in the Los Angeles domicile, the furloughed flight attendant could opt to displace the most junior attendant of equal rank in the entire system or the most junior attendant of lower rank either at the same domicile or in the entire system. 1981-1984 TWA/IFFA Collective Bargaining Agreement, Arts. 12-13, 18-A, 18-B, reprinted in App. 31-62.

For two years TWA and the Union unsuccessfully bargained over wages and working conditions not including the seniority bidding system. They pursued all the required dispute resolution mechanisms of the RLA, including direct negotiation, 45 U. S. C. § 152 Second, mediation, 45 U. S. C. § 155 First, and the final 30-day “cooling off” period. Ibid. By early 1986 a strike seemed imminent, and on March 7, 1986, the Union went out on strike.

TWA informed its flight attendants before and during the strike that it would continue operations by hiring permanent replacements for striking flight attendants, by continuing to employ any flight attendant who chose not to strike, and by rehiring any striker who abandoned the strike and made an unconditional offer to return to any available vacancies. *430TWA also informed its flight attendants that any vacancies created as a result of the strike would be filled by application of the seniority bidding system to all working flight attendants and that such job and domicile assignments would remain effective after the strike ended. App. 120-122, 132-134, 137-139. Thus, at the conclusion of the strike, senior full-term strikers would not be permitted to displace permanent replacements or junior nonstriking flight attendants and could be left without an opportunity to return to work. TWA’s promise not to displace working flight attendants after the strike created two incentives specifically linked to the seniority bidding system: it gave senior flight attendants an incentive to remain at, or return to, work in order to retain their prior jobs and domicile assignments; it gave junior flight attendants an incentive to remain at, or return to, work in order to obtain job and domicile assignments that were previously occupied by more senior, striking flight attendants.

As promised, TWA continued its operations during the 72-day strike by utilizing approximately 1,280 flight attendants who either did not strike or returned to work before the end of the strike and by hiring and fully training approximately 2,350 new flight attendants, some 1,220 of whom were hired during the first few days of the strike. On May 17, 1986, the Union made an unconditional offer to TWA on behalf of the approximately 5,000 flight attendants who had remained on strike to return to work. TWA accepted the offer but refused the Union’s May 27th demand that TWA displace those prestrike employees who were working as of May 17th (“crossover” employees). Accordingly, TWA initially recalled only the 197 most senior full-term strikers to fill available job and domicile vacancies. By the terms of a poststrike arbitral agreement, these strikers and all subsequently reinstated full-term strikers returned to work as vacancies arose and with precisely the seniority they would have had if no strike *431had occurred. In May 1988, more than 1,100 full-term strikers had been reinstated with full seniority.

In an effort to reinstate all the full-term strikers by displacing the newly hired flight attendants and less senior crossover employees, the Union proceeded on two fronts. First, it brought an injunction action alleging that the full-term strikers were not “economic strikers” but “unfair labor practice strikers” entitled to reinstatement by application of principles this Court has developed in interpreting the National Labor Relations Act (NLRA). 29 U. S. C. § 151 et seq. See Mastro Plastics Corp. v. NLRB, 350 U. S. 270 (1956). The District Court ultimately ruled against the Union on this claim. Independent Federation of Flight Attendants v. Trans World Airlines, Inc., 682 F. Supp. 1003 (WD Mo. 1988), appeal pending, No. 88-1984M (CA8). At the same time, the Union filed the instant action contending that, even assuming the strike was economic, the full-term strikers were entitled to reinstatement either under the terms of the prestrike collective bargaining agreement or under the RLA itself. On cross motions for partial summary judgment, the District Court held that the full-term strikers were not entitled to displace either the junior crossovers or the 1,220 new hires employed by TWA immediately after the strike commenced. (The motions did not require the District Court to rule on the status of the remaining new hires.) The District Court also held that 463 new hires not fully trained by the end of the strike could be displaced by full-term strikers. Independent Federation of Flight Attendants v. Trans World Airlines, Inc., 643 F. Supp. 470 (WD Mo. 1986).

Meanwhile, TWA sought a declaratory judgment that the union security clause of the prestrike collective bargaining agreement containing provisions for the checkoff of union dues and a requirement that new hires join the Union did not survive the self-help period after the parties had bargained to impasse. On cross motions for summary judgment, the same District Court ruled that, because the union security clause *432was not part of the prestrike negotiations, it had survived the strike. Trans World Airlines, Inc. v. Independent Federation of Flight Attendants, 640 F. Supp. 1108 (WD Mo. 1986).

Appeals were taken from both judgments. The Court of Appeals affirmed the District Court’s ruling that the union security clause had survived the period of self-help. Trans World Airlines, Inc. v. Independent Federation of Flight Attendants, 809 F. 2d 483 (CA8 1987). In a separate opinion, the same panel also affirmed the District Court’s ruling that full-term strikers could not displace the 1,220 fully trained new hires but could displace the 463 untrained new hires. Independent Federation of Flight Attendants v. Trans World Airlines, Inc., 819 F. 2d 839 (CA8 1987). The Court of Appeals, however, reversed the District Court’s ruling that more senior full-term strikers could not displace junior crossovers. In so holding, the court relied primarily on its reading of the union security clause of the prestrike collective bargaining agreement and, secondarily, on judicial interpretations of the NLRA. Id., at 843-845.

We granted petitions for writs of certiorari in both cases. Trans World Airlines, Inc. v. Flight Attendants, 482 U. S. 913 (1987) (TWA I); Flight Attendants v. Trans World Airlines, Inc., 485 U. S. 958 (1988) (TWA II) (certiorari granted only to consider displacement of crossovers). Last Term, we affirmed by an equally divided Court the judgment of the Court of Appeals in TWA I that the union security clause survived the strike. 485 U. S. 175 (1988). Today, we reverse the Court of Appeals in TWA II and hold that an employer is not required by the RLA to lay off junior crossovers in order to reinstate more senior full-term strikers at the conclusion of a strike.

II

We have observed in the past that carefully drawn analogies from the federal common labor law developed under the NLRA may be helpful in deciding cases under the RLA. Trainmen v. Jacksonville Terminal Co., 394 U. S. 369, 377 *433(1969). Thus, as in this case, those lower courts that have examined the reinstatement rights of strikers under the RLA have turned to NLRA precedents for guidance. E. g., Air Line Pilots Assn. International v. United Air Lines, Inc., 614 F. Supp. 1020, 1041, 1045-1046 (ND Ill. 1985), aff’d in part and rev’d in part on other grounds, 802 F. 2d 886 (CA7 1986), cert. denied, 480 U. S. 946 (1987); National Airlines, Inc. v. International Assn. of Machinists & Aerospace Workers, 416 F. 2d 998, 1004-1006 (CA5 1969).

We first considered the reinstatement rights of strikers under the NLRA in NLRB v. Mackay Radio & Telegraph Co., 304 U. S. 333 (1938). In Mackay Radio, radio and telegraph operators working in the San Francisco offices of a national telecommunications firm went on strike. In order to continue operations, the employer brought employees from its other offices to fill the strikers’ places. At the conclusion of the strike, the striking operators sought to displace their replacements in order to return to work. We held that it was not an unfair labor practice under § 8 of the NLRA for the employer to have replaced the striking employees with others “in an effort to carry on the business,” or to have refused to discharge the replacements in order to make room for the strikers at the conclusion of the strike. Id., at 345-346. As we there observed, “[t]he assurance by [the employer] to those who accepted employment during the strike that if they so desired their places might be permanent was not an unfair labor practice nor was it such to reinstate only so many of the strikers as there were vacant places to be filled.” Id., at 346. On various occasions we have reaffirmed the holding of Mackay Radio. See NLRB v. Erie Resistor Corp., 373 U. S. 221, 232 (1963) (“We have no intention of questioning the continuing vitality of the Mackay rule . . .”); NLRB v. Fleetwood Trailer Co., 389 U. S. 375, 379 (1967) (Employers have “ ‘legitimate and substantial business justifications’ for refusing to reinstate employees who engaged in an economic strike . . . when the jobs claimed by the *434strikers are occupied by workers hired as permanent replacements during the strike in order to continue operations”); Belknap, Inc. v. Hale, 463 U. S. 491, 604, n. 8 (1983) (“The refusal to fire permanent replacements because of commitments made to them in the course of an economic strike satisfies the requirement. . . that the employer have a ‘legitimate and substantial justification’ for its refusal to reinstate strikers”).

TWA asks us to apply this line of cases decided under the NLRA to determine the status under the RLA of those prestrike flight attendants who were working at the conclusion of the strike. TWA argues that it would be completely anomalous to hold that full-term strikers may displace junior crossovers when, as the Union has conceded, they may not displace newly hired permanent replacements under either statute. The Union, by contrast, argues that the rule of Mackay Radio is inapplicable to junior crossovers because of differences between the RLA and the NLRA and because, even under the NLRA, junior crossovers would be treated differently from newly hired permanent replacements.1

The Union relies on Erie Resistor, supra, to distinguish junior crossovers from new hires under the NLRA. In Erie Resistor we struck down an employer’s award of 20 years’ superseniority to new hires and crossovers as an unfair labor practice within the meaning of § 8(a)(1) and § 8(a)(3) of the NLRA. 29 U. S. C. §§ 158(a)(1), 158(a)(3). We observed:

“. . . Super-seniority affects the tenure of all strikers whereas permanent replacement, proper under Mackay, affects only those who are, in actuality, replaced. It is *435one thing to say that a striker is subject to loss of his job at the strike’s end but quite another to hold that in addition to the threat of replacement, all strikers will at best return to their jobs with seniority inferior to that of the replacements and of those who left the strike.

“. . . Unlike the replacement granted in Mackay which ceases to be an issue once the strike is over, the [superseniority] plan here creates a cleavage in the plant continuing long after the strike is ended. Employees are henceforth divided into two camps: those who stayed with the union and those who returned before the end of the strike and thereby gained extra seniority. This breach is reemphasized with each subsequent layoff and stands as an ever-present reminder of the dangers connected with striking and with union activities in general.” 373 U. S., at 230-231.

The Union does not and cannot contend that reinstated full-term strikers have less seniority relative to new hires and junior crossovers than they would have had if they had not remained on strike. It is clear that reinstated full-term strikers lost no seniority either in absolute or relative terms. Thus, unlike the situation in Erie Resistor, any future reductions in force at TWA will permit reinstated full-term strikers to displace junior flight attendants exactly as would have been the case in the absence of any strike. Similarly, should any vacancies develop in desirable job assignments or domiciles, reinstated full-term strikers who have bid on those vacancies will maintain their priority over junior flight attendants, whether they are new hires, crossovers, or full-term strikers. In the same vein, periodic bids on job scheduling will find senior reinstated full-term strikers maintaining their priority over all their junior colleagues. In short, once reinstated, the seniority of full-term strikers is in no way affected by their decision to strike.

*436Nevertheless, IFFA argues that TWA’s refusal to displace junior crossovers will create a “cleavage” between junior crossovers and reinstated full-term strikers at TWA “long after the strike is ended.” Id., at 231. This is the case because desirable job assignments and domiciles that would have been occupied by the most senior flight attendants had there been no strike will continue to be held by those who did not see the strike through to its conclusion. For example, the senior full-term striker who worked in the Los Angeles domicile before the strike may have been replaced by a junior crossover. As poststrike vacancies develop in TWA’s work force, permitting reinstatement of full-term strikers, they are not likely to occur in the most desirable domiciles. Thus, it is unlikely that the senior full-term striker would be reinstated back to her preferred domicile. Resentful rifts among employees will also persist after the strike, the Union argues, because TWA’s prestrike assurance of nondisplacement to junior crossovers, unlike the same assurance to new hires, “set up- a competition among those individuals who participated in the original decision to strike, and thereby undermined the group’s ability to take the collective action that it is the very purpose of the [RLA] to protect.” Brief for Respondent 36-37.

We reject this effort to expand Erie Resistor. Both the RLA and the NLRA protect an employee’s right to choose not to strike. 45 U. S. C. § 152 Fourth; 29 U. S. C. § 157, and, thereby, protect employees’ rights to “the benefit of their individual decisions not to strike . . . .” Post, at 448, n. 4 (Brennan, J., dissenting).2 Accordingly, in virtually *437every strike situation there will be some employees who disagree with their union’s decision to strike and who cannot be required to abide by that decision. It is the inevitable effect of an employer’s use of the economic weapons available during a period of self-help that these differences will be exacerbated and that poststrike resentments may be created. Thus, for example, the employer’s right to hire permanent replacements in order to continue operations will inevitably also have the effect of dividing striking employees between those who, fearful of permanently losing their jobs, return to work and those who remain stalwart in the strike. In such a situation, apart from the “pressure on the strikers as a group to abandon the strike,” to which the dissent refers, post, at 449 (Brennan, J., dissenting), a “competition” may arise among the striking employees to return to work in order to avoid being displaced by a permanent replacement. Similarly, employee awareness that an employer may decide to transfer working employees to necessary positions previously occupied by more senior striking employees will isolate employees fearful of losing those positions and employees coveting those positions from employees more committed to the strike. Conversely, a policy such as TWA employed here, in creating the incentive for individual strikers to return to work, also “puts pressure on the strikers as a group to abandon the strike,” ibid., in the same manner that the hiring of permanent replacements does.

None of these scenarios, however, present the prospect of a continuing diminution of seniority upon reinstatement at the end of the strike that was central to our decision in Erie Resistor. All that has occurred is that the employer has filled vacancies created by striking employees. Some of these vacancies will be filled by newly hired employees, others by doubtless more experienced and therefore more needed employees who either refused to strike or abandoned the strike. The dissent’s observation that, “at the conclusion of the strike,” discrimination in the filling of “available *438positions” based on union activity is impermissible is beside the point. See post, at 450 (Brennan, J., dissenting). The positions occupied by newly hired replacements, employees who refused to strike, and employees who abandoned the strike are simply not “available positions” to be filled. As noted above, those positions that were available at the conclusion of the strike were filled “according to some principle, such as seniority, that is neutral. . . .” Ibid. (Brennan, J., dissenting). That the prospect of a reduction in available positions may divide employees and create incentives among them to remain at work or abandon a strike before its conclusion is a secondary effect fairly within the arsenal of economic weapons available to employers during a period of self-help.

To distinguish crossovers from new hires in the manner IFFA proposes would have the effect of penalizing those who decided not to strike in order to benefit those who did. Because permanent replacements need not be discharged at the conclusion of a strike in which the union has been unsuccessful, a certain number of prestrike employees will find themselves without work. We see no reason why those employees who chose not to gamble on the success of the strike should suffer the consequences when the gamble proves unsuccessful. Requiring junior crossovers, who cannot themselves displace the newly hired permanent replacements, and “who rank lowest in seniority,” post, at 447 (Brennan, J., dissenting), to be displaced by more senior full-term strikers is precisely to visit the consequences of the lost gamble on those who refused to take the risk. While the employer and union in many circumstances may reach a back-to-work agreement that would displace crossovers and new hires or an employer may unilaterally decide to permit such displacement, nothing in the NLRA or the federal common law we have developed under that statute requires such a result. That such agreements are typically one mark of a successful strike is yet another indication that crossovers opted not to *439gamble; if the strike was successful the advantage gained by declining to strike disappears.

J — I HH

The Union argues, however, that whether or not the NLRA prohibits a crossover policy such as TWA’s, the statutory framework of the RLA forbids such a policy.

Although we have observed that the NLRA may provide useful analogies for interpreting the RLA, we have also emphasized that the NLRA “cannot be imported wholesale into the railway labor arena. Even rough analogies must be drawn circumspectly with due regard for the many differences between the statutory schemes.” Trainmen v. Jacksonville Terminal, 394 U. S., at 383. Thus, in Trainmen itself we declined to examine the “panoply of detailed law developed” under the NLRA to determine what kind of secondary picketing in a railway dispute may be enjoined by state courts. Rather, we held that Congress had entirely withdrawn such injunctive power from the States: “[Pjarties who have unsuccessfully exhausted the Railway Labor Act’s procedures for resolution of a major dispute . . . [may] employ the full range of whatever peaceful economic power they can muster, so long as its use conflicts with no other obligation imposed by federal law.” Id., at 391-392. Similarly, two Terms ago in Burlington Northern R. Co. v. Maintenance of Way Employes, 481 U. S. 429 (1987), we declined to find in the RLA an implied limit on a union’s resort to secondary activity by analogy to the NLRA. These cases have read the RLA to provide greater avenues of self-help to parties that have exhausted the statute’s “virtually endless,” id., at 444, dispute resolution mechanisms than would be available under the NLRA. Nevertheless, they provide the backdrop for the Union’s contention that, in this case, we should understand provisions of the RLA to limit “the full range of whatever peaceful economic power [the parties] can *440muster,” Trainmen, supra, at 392, beyond the limitations even imposed by the NLRA. This we decline to do.

The Union points to § 2 Fourth of the RLA as the source of this limitation on the use of the employer’s economic power. The section provides, in pertinent part:

“No carrier, its officers or agents shall deny or in any way question the right of its employees to join, organize, or assist in organizing the labor organization of their choice, and it shall be unlawful for any carrier to interfere in any way with the organization of its employees, ... or to influence or coerce employees in an effort to induce them to join or remain or not to join or remain members of any labor organization . . . .” 45 U. S. C. § 152 Fourth.

The Union argues that TWA’s crossover policy, which created an incentive for flight attendants either not to join or to abandon the strike, constituted influence or coercion in an effort to induce the flight attendants not to remain members of IFFA and was, therefore, impermissible under §2 Fourth.

Section 2 Fourth was enacted as part of the 1934 amendments to the RLA. 48 Stat. 1185. From the time of our very first opportunity to interpret the 1934 amendments, we have viewed them as addressing primarily the precertification rights and freedoms of unorganized employees. In Virginian R. Co. v. Railway Employees, 300 U. S. 515 (1937), we observed that the employees’ freedom “to organize and to make choice of their representatives without the ‘coercive interference’ and ‘pressure’ of a company union . . . was continued and made more explicit by the amendment of 1934.” Id., at 543, citing § 2 Third, § 2 Fourth, and Texas & N. O. R. Co. v. Railway Clerks, 281 U. S. 548 (1930). In Switchmen v. National Mediation Bd., 320 U. S. 297 (1943), the Court divided over whether the federal courts have jurisdiction under § 2 Fourth to review a certification of union representatives for collective bargaining by the National Mediation Board acting under § 2 Ninth of the RLA as amended in 1934. Both the majority and the dissent agreed, however, that *441“[t]he 1934 Act was directed particularly at control over the initial step in collective bargaining — the determination of the employees’ representatives.” Id., at 317 (Reed, J., dissenting); see also id., at 302 (opinion of the Court); Machinists v. Street, 367 U. S. 740, 759 (1961).

The explanation for the precertification focus of the 1934 amendments is clear. The RLA provides an exhaustively detailed procedural framework “to facilitate the voluntary settlement of major disputes.” Trainmen v. Jacksonville Terminal, supra, at 378. The effectiveness of these private dispute resolution procedures depends on the initial assurance that the employees’ putative representative is not subject to control by the employer and on the subsequent assurance that neither party will be able to enlist the courts to further its own partisan ends. See Chicago & N. W. R. Co. v. Transportation Union, 402 U. S. 570, 596-597 (1971) (Brennan, J., dissenting) (the duty to exhaust the dispute resolution procedures “does not contemplate that governmental power should, after failure of the parties to reach accord, be added to the scales in favor of either party and thus compel the other to agree upon the aided party’s terms. Rather, at that point, impasse was to free both parties to resort to self-help”); Burlington Northern, supra, at 451-452 (the availability of self-help measures rather than judicial remedies “may increase the effectiveness of the RLA in settling major disputes by creating an incentive for the parties to settle prior to exhaustion of the statutory procedures”). Thus, we have understood judicial intervention in RLA procedures to be limited to those cases where “but for the general jurisdiction of the federal courts there would be no remedy to enforce the statutory commands which Congress had written into the Railway Labor Act.” Switchmen, supra, at 300; Chicago & N. W. R. Co., supra, at 595 (Brennan, J., dissenting) (“The underlying cohesiveness of the decisions [permitting judicial interference] lies in the fact that in each instance the scheme of the Railway Labor Act could not begin to work without judicial involvement”).

*442Here, TWA and the Union followed without interference the scheme of the RLA to an unsuccessful conclusion and then turned to self-help. We have more than once observed that, at this final stage of a labor dispute regulated by the RLA, “the Act is wholly inexplicit as to the scope of allowable self-help.” Trainmen, 394 U. S., at 391; Burlington Northern, 481 U. S., at 447-448. Such silence does not amount to a congressional imprimatur on all forms of postnegotiation self-help. It does, however, indicate that we' should hesitate to imply limitations on all but those forms of self-help that strike a fundamental blow to union or employer activity and the collective bargaining process itself. Accordingly, just as we saw no statutory basis for limiting the secondary activities of unions during a period of self-help in Trainmen and Burlington Northern, we see no basis in §2 Fourth for prohibiting the crossover policy employed by TWA once bargaining had reached an impasse. Both self-help measures fall squarely within the “full range of whatever peaceful economic power [the parties] can muster” once they have “unsuccessfully exhausted the Railway Labor Act’s procedures for resolution of a major dispute . . . .” Trainmen, supra, at 392. Neither measure prevented the scheme of the RLA from working; neither measure was inherently destructive of union or employer activity. Similarly, we see nothing in Railway Clerks v. Florida East Coast R. Co., 384 U. S. 238 (1966), so heavily relied upon by the Union, that is to the contrary. In Florida East Coast we recognized a carrier’s ability to depart from the terms of an existing collective bargaining agreement when reasonably necessary to operate during a strike. As the Union itself concedes, see n. 1, supra, nothing in the collective bargaining agreement or any poststrike agreement between TWA and IFFA prohibits the crossover policy adopted by TWA. Thus, there was no departure from the collective bargaining agreement that would require an examination of reasonable necessity.

*443IV

Neither the RLA itself nor any analogies to the NLRA indicate that the crossover policy adopted by TWA during the period of self-help was unlawful. Rather, the decision to guarantee to crossovers the same protections lawfully applied to new hires was a simple decision to apply the preexisting seniority terms of the collective bargaining agreement uniformly to all working employees. That this decision had the effect of encouraging prestrike workers to remain on the job during the strike or to abandon the strike and return to work before all vacancies were filled was an effect of the exercise of TWA’s peaceful economic power, a power that the company was legally free to deploy once the parties had exhausted the private dispute resolution mechanisms of the RLA. Accordingly, the judgment of the Court of Appeals is

Reversed.

with whom Justice Marshall joins, dissenting.

The issue in this case is whether under the Railway Labor Act (RLA) an employer, in allocating available jobs among members of a bargaining unit at the conclusion of a strike, may discriminate against full-term strikers by giving preference to employees who crossed the picket line to return to work before the strike was over. Because I conclude that such discrimination on the basis of union activity is “inherently destructive” of the right to strike, as guaranteed by both the RLA and the National Labor Relations Act (NLRA), I dissent.

I

Notwithstanding the Court’s suggestion that the portion of the RLA at issue here addresses “primarily” the precerti-fication context, ante, at 440, it should be clear that under the RLA an employee’s right to strike is protected against coercion by her employer. The Court relies in part on Trainmen *444v. Jacksonville Terminal Co., 394 U. S. 369 (1969), but it overlooks the clear teaching of that case:

“[EJmployees subject to the Railway Labor Act enjoy the right to engage in primary strikes over major disputes. . . . Whether the source of this right be found in a particular provision of the Railway Labor Act or in the scheme as a whole, it is integral to the Act.” Id., at 384-385 (footnote omitted).

The “particular provision,” we made clear, was §2 Fourth. Id., at 385, n. 20. While the issue in Jacksonville Terminal was the extent of a state court’s power to issue an antistrike injunction, we emphasized that the RLA’s guarantee of the right to strike was not limited to the context of interference by the State: “However, §2 Fourth of the RLA, added in 1934, was designed primarily, if not exclusively to prohibit coercive employer practices.” Ibid, (emphasis added). Whatever may have been the “primary” purpose of § 2 Fourth, it is too late in the day to suggest that this provision, at least when read in the context of the entire RLA, does not prohibit employer coercion of the right to strike.

The Court compounds its error in regard to the reach of § 2 Fourth with a more fundamental mistake when it appears to assume that the employer’s action in this case is sanctioned by the mere fact that it occurred during the “self-help” stage of the dispute. Ante, at 440-442. Clearly this cannot be the case. I am confident that the Court would agree, for example, that an employer could not legally discharge striking employees under the RLA. But if this is so, it must be because the RLA contains some injunction against employer interference with the right to strike, even when that interference consists of actions taken during the period of permissible self-help. Thus, the question is not whether the RLA protects the right to strike against employer coercion — for it surely does — but whether that protection goes so far as to prohibit the specific employer practice at issue here.

*445The key to this case is a fundamental command of the RLA and the NLRA alike, which in the case of the RLA is textually anchored in § 2 Fourth: the employer may not engage in discrimination among its employees — whether at the precer-tification stage, the bargaining stage, or during or after a strike — on the basis of their degree of involvement in protected union activity such as a strike.1 This case thus falls within the class of cases in which judicial intervention to enforce the right at issue is justified because “the scheme of the Railway Labor Act could not begin to work without judicial involvement.” Chicago & N. W. R. Co. v. Transportation Union, 402 U. S. 570, 595 (1971) (Brennan, J., dissenting). The “central theme” of the RLA is, of course, “to bring about voluntary settlement.” Ibid. But “unless the unions fairly represented all of their employees; unless the employer bargained with the certified representative of the employees; unless the status quo was maintained during the entire range of bargaining, the statutory mechanism could not hope to induce a negotiated settlement.” Ibid. The same is true here: the statutory scheme would be just as incapable of bringing about a negotiated settlement if the employer, in the name of “self-help,” impermissibly retaliated against employees because of their exercise of their right under the RLA to engage in protected union activity such as a strike.2

II

A

That the RLA broadly enjoins discrimination against strikers does not necessarily settle the issue, of course. In the context of the NLRA we have on occasion found reason to *446make an exception to that statute’s nondiscrimination provision in the name of the employer’s “necessity.” See NLRB v. Mackay Radio & Telegraph Co., 304 U. S. 333 (1938). The RLA itself provides little guidance as to whether the employer is in any way privileged, in allocating jobs at the end of a strike, to give preference to bargaining unit members who crossed the picket line to return to work. As we have previously noted, “the Act is wholly inexplicit as to the scope of allowable self-help.” Jacksonville Terminal, 394 U. S., at 391.

While of course “the National Labor Relations Act cannot be imported wholesale into the railway labor arena,” id., at 383, we have frequently “referred to the NLRA for assistance in construing the Railway Labor Act.” Ibid. Given the paucity of RLA precedent on the specific issue before us, the Court quite properly looks to the NLRA for guidance. Ante, at 432-439. It arrives at an incorrect conclusion, however, because it mischaracterizes the employer’s action and because it appears unwilling to take seriously the protection Congress has seen fit to afford to the right to strike.

The Court’s conception of this case is most clearly expressed in a key paragraph that summarizes its discussion of the NLRA case law:

“To distinguish crossovers from new hires in the manner IFFA proposes would have the effect of penalizing those who decided not to strike in order to benefit those who did. . . . We see no reason why those employees who chose not to gamble on the success of the strike should suffer the consequences when the gamble proves unsuccessful. Requiring junior crossovers ... to be displaced by more senior full-term strikers is precisely to visit the consequences of the lost gamble on those who refused to take the risk.” Ante, at 438.

This understanding of the Union’s position contains a factual and a legal error, both of which infect the Court’s analysis of the case.

*447In the first place, refusing to discriminate in favor of crossovers is not to visit the consequences of the lost strike on “those who refused to take the risk,” but rather on those who rank lowest in seniority. Whether a given flight attendant chose to take the risk of the strike or not is wholly immaterial. Rather — as is virtually universally the case when work-force reductions are necessary for whatever reason in a unionized enterprise — it is the most junior employees, whether strikers or crossovers, who are most vulnerable. This is precisely the point of seniority.

More fundamental, I fear, is the legal mistake inherent in the Court’s objection to “penalizing those who decided not to strike in order to benefit those who did.” The Court, of course, does precisely the opposite: it allows TWA to single out for penalty precisely those employees who were faithful to the strike until the end, in order to benefit those who abandoned it. What is unarticulated is the Court’s basis for choosing one position over the other. If indeed one group or the other is to be “penalized,”3 what basis does the Court have for determining that it should be those who remained on strike rather than those who returned to work? I see none, unless it is perhaps an unarticulated hostility toward strikes. In any case the NLRA does provide a basis for resolving this question. It requires simply that in making poststrike reinstatements an employer may not discriminate among its employees on account of their union activity. That, in fact, is the holding of NLRB v. Mackay Radio, supra, at 346 — the more familiar teaching as to the employer’s right to hire permanent replacements having been dictum. If an employer may not discriminate — in either direction — on the basis of the employee’s strike activity, then it follows that the employer must make decisions about which employees to reinstate on *448the basis of some neutral criterion, such as seniority. That is precisely what the Union asks.4

B

We have recognized only a narrow exception to the general principle prohibiting discrimination against employees for exercising their right to strike. Since Mackay Radio it has been accepted that an employer may hire “permanent replacements” in order to maintain operations during a strike, and that these replacements need not be displaced to make room for returning strikers. The question here is whether the Mackay exception should be expanded to cover the present case, involving as it does members of the striking bargaining unit who have crossed the picket lines, rather than new hires from outside the bargaining unit. Despite the superficial similarity between the two situations, strong reasons counsel against applying the Mackay rule to crossover employees.

The employer’s promise to members of the bargaining unit that they will not be displaced at the end of a strike if they *449cross the picket lines addresses a far different incentive to the bargaining-unit members than does the employer’s promise of permanence to new hires. The employer’s threat to hire permanent replacements from outside the existing work force puts pressure on the strikers as a group to abandon the strike before their positions are filled by others. But the employer’s promise to members of the striking bargaining unit that if they abandon the strike (or refuse to join it at the outset) they will retain their jobs at strike’s end in preference to more senior workers who remain on strike produces an additional dynamic: now there is also an incentive for individual workers to seek to save (or improve) their own positions at the expense of other members of the striking bargaining unit. We have previously observed that offers of “individual benefits to the strikers to'induce them to abandon the strike . . . could be expected to undermine the strikers’ mutual interest and place the entire strike effort in jeopardy.” NLRB v. Erie Resistor Corp., 373 U. S. 221, 230-231 (1963). Such a “divide and conquer” tactic thus “strike[s] a fundamental blow to union . . . activity and the collective bargaining process itself.” Ante, at 442.

In Erie Resistor we found the employer’s offer of super-seniority to new hires and crossovers to be “inherently destructive” of the right to strike and therefore in contravention of §§ 8(a)(1) and (a)(3) of the NLRA. 373 U. S., at 231-232. In my view the same conclusion should apply here. Beyond its specific holding outlawing superseniority, I read Erie Resistor to stand for the principle that there are certain tools an employer may not use, even in the interest of continued operations during a strike, and that the permissibility of discriminatory measures taken for that purpose must be evaluated by weighing the “necessity” of the employer’s action (i. e., its interest in maintaining operations during the strike) against its prejudice to the employees’ right to strike.5 It *450seems clear to me that in this case the result of such an analysis should be to forbid the employer to give preferential treatment to crossovers, because of the destructive impact of such an action on the strikers’ mutual interest. Thus, when an employer recalls workers to fill the available positions at the conclusion of a strike, it may not discriminate against either the strikers or the crossovers. Rather it must proceed according to some principle, such as seniority, that is neutral as between them.6 ' That TWA failed to do.7

*451III

Precedent under the NLRA clearly forbids an employer to burden the right to strike in the manner TWA has done in this case, and I see no reason why that conclusion should not apply equally under the RLA.

In a case like this it is not difficult to conjure up a parade of horribles to support either position. Forbidding an employer to discriminate in favor of crossovers, as I would do, makes it impossible for a junior employee who does not want to strike, and who is unable to persuade a majority of her colleagues to adopt that stance, to be sure that she can save her job. But that employee is in the same position she would be in if a layoff were necessary for other reasons beyond her control, such as an economic downturn. The principle of seniority is based on the notion that it is those employees who have worked longest in an enterprise and therefore have most at stake whose jobs should be most protected. Permitting the employer to give preference to crossovers, as the Court today does, will mean that an employee of only six months’ experience, who abandoned the strike one day before it ended, could displace a 20-year veteran who chose to remain faithful to the decision made collectively with her fellow workers until the group as a whole decided to end the strike. Unfortunately there will be individual injustices whichever *452rule we adopt. I would favor — and I believe Congress has provided for — the rule that errs on the side of preferring solidarity and seniority, rather than a rule that would permit the employer to discriminate on the basis of protected union activity.

with whom Justice Brennan joins as to Parts I and II, dissenting.

The central question in this Railway Labor Act (RLA) case is whether it is unlawful for a carrier to refuse to reinstate employees who supported a strike until its end (“full-term strikers”) solely because the carrier chooses to retain in its active work force employees who returned to work before the strike’s conclusion (“crossovers”).1

The Court today answers that question in the negative, concluding that such conduct never violates the RLA, regardless of whether business necessity dictated the carrier’s course of action. In dissent, Justice Brennan takes the diametrically opposite view, in agreement with the Court of Appeals. Justice Brennan finds such conduct “inherently destructive,” ante, at 443, of the right to strike and violative of the RLA regardless of any proffered business justification. In my view, neither of these positions accurately captures the delicate balance our RLA precedents have attempted to achieve between the public’s dual interests in the maintenance of transportation service during labor disputes and in the long-term stability of labor relations in the rail and airline industries.

*453My differences with Justice Brennan are limited in scope. Concisely stated, I give greater weight than he does to the RLA’s policy in favor of continued operations, and accordingly conclude that this case should be remanded to permit TWA to make a factual showing that its crossover policy truly was necessary for that purpose. The Court’s opinion presents far greater concerns, as much because of the false' assumptions that underlie the Court’s analysis as because of its erroneous result.

I

The threshold question is whether the provisions and policies of the RLA place any limit on a carrier’s exercise of self-help during a strike. The Court acknowledges that the RLA does contemplate such a limit. Indeed, there would be little need to distinguish, see ante, at 436, TWA’s crossover policy from the superseniority policy in NLRB v. Erie Resistor Corp., 373 U. S. 221 (1963), if the RLA had no relevance to the legality of grants of superseniority, or to other, even more egregious, discriminatory, and coercive employer practices. But the Court adopts a stingy interpretation of the RLA, reserving the RLA’s protective force for only the most extraordinary circumstances. In so doing, the Court uses language which suggests that any limit on employer self-help must be “implied],” ante, at 442, which in turn suggests that the Court finds no express limit in the text of the RLA. I find no basis for that view, a view which does not sit comfortably with the Court’s opinion read as a whole and which results in a far too restrictive reading of the RLA.

When the Court addressed the permissible scope of employer self-help under the RLA in Trainmen v. Jacksonville Terminal Co., 394 U. S. 369 (1969), it held that the RLA permits “parties who have unsuccessfully exhausted the-Railway Labor Act’s procedures for resolution! of a major dispute to employ the full range of whatever peaceful economic power they can muster, so long as its use conflicts ivith no- other obligation imposed by federal law Id., at 392 (emphasis *454added). In applying that holding to the facts of this case, the Court rejects the proposition that § 2 Fourth of the RLA, 44 Stat. 577, as amended, 45 U. S. C. § 152 Fourth, creates a relevant conflicting federal obligation.

The Court’s stated reason for rejecting the applicability of § 2 Fourth sweeps too broadly. The Court places great emphasis on the fact that the 1934 amendments which introduced §2 Fourth had a “precertification focus.” Ante, at 441. It should be clear, however, that a precertiflcation focus is not the same as a postcertification blindspot. In 1934, Congress was faced with evidence that railroad employees’ efforts at self-organization had been thwarted by coercive employer tactics, including the support of employer-dominated company unions. See Machinists v. Street, 367 U. S. 740, 759 (1961). Certainly, Congress had cause for concern: unless each side is free to choose its own bargaining representative, there can be no legitimate bargaining relationship. There is no indication, however, that Congress’ concern in enacting § 2 Fourth is satisfied at the moment of a union’s certification. Congress aimed to protect the employee’s right to organize and join unions “with a view to asserting himself as to hours, conditions, and wages,” 78 Cong. Ree. 11720 (1934) (remarks of Rep. Monaghan) — not as an end in itself. This Court long has recognized that a “primary purpose of the major revisions made in 1934 was to strengthen the position of the labor organizations vis-a-vis the carriers, to the end of furthering the success of the basic congressional policy of self-adjustment of the industry’s labor problems.” Machinists v. Street, 367 U. S., at 759.

Indeed, the Court today acknowledges that, precertification focus notwithstanding, §2 Fourth has relevance to the right of employees to decide whether to assist in postcerti-fication union activities, free from employer coercion. The Court places substantial reliance on § 2 Fourth as the source of “an employee’s right to choose not to strike,” ante, at 436, *455a right relevant to this case only if it applies to posicertification strike activity.

Finding § 2 Fourth to be a source of the right not to strike is entirely proper. In Radio Officers v. NLRB, 347 U. S. 17 (1954), the Court held that the protection § 8(a)(3) of the NLRA affords against employer discrimination “to . . . discourage membership in any labor organization,” 29 U. S. C. § 158(a)(3), extends to “discrimination to discourage participation in union activities as well as to discourage adhesion to union membership.” 347 U. S., at 40. I see no reason why similar language in §2 Fourth, i. e., its protection of employees’ right to “join or remain or not to join or remain members of any labor organization,” should not be read in a similar fashion. Cf. Trainmen v. Jacksonville Terminal Co., 394 U. S., at 385, n. 20. Neither, apparently, does the Court. And if §2 Fourth bars discrimination or retaliation against employees who choose not to strike, the same must be true of discrimination or retaliation against employees who choose to strike. See Railway Labor Executives’ Assn. v. Boston & Maine Corp., 808 F. 2d 150, 158 (CA1 1986), cert. denied, 484 U. S. 830 (1987); Air Line Pilots Assn. v. United Air Lines, Inc., 802 F. 2d 886, 897 (CA7 1986), cert. denied, 480 U. S. 946 (1987).

In contrast, the Court’s suggestion that the RLA provides employees no express protection against discrimination on the basis of levels of support for union activities leads the Court to limit the RLA’s force to whatever protections this Court is willing to “imply” from the RLA’s general policies. This uncertainty carries with it the danger of undermining the stability of labor relations under the RLA. Under this Court’s longstanding RLA jurisprudence, a strike that takes place after the RLA’s dispute resolution mechanisms have failed “represents only an interruption in the continuity of the relation” between employer and union, not an invitation for “labor-management relations [to] revert to the jungle.” Railway Clerks v. Florida East Coast R. Co., 384 U. S. 238, *456246-247 (1966). Stated otherwise, a strike under the RLA is a “bounded conflict.” Cf. Estreicher, Strikers and Replacements, 38 Lab. L. J. 287, 288 (1987). Contract negotiations are limited in scope to the matters raised by the parties’ bargaining notices, see 45 U. S. C. § 156; both during and after strikes that occur following unsuccessful mediation, the union often will maintain its status as exclusive bargaining representative. See, e. g., Trans World Airlines, Inc. v. Independent Federation of Flight Attendants, 809 F. 2d 483, 492 (CA8 1987), aff’d by equally divided Court, 485 U. S. 175 (1988). The long-term stability of labor relations thus will depend upon the maintenance of the working relationship between the union and the employer. This Court has been aware in the past that one party’s power of self-help cannot be permitted effectively to negate the other’s, lest “the right of self-help . . . become unilateral,” Railway Clerks v. Florida East Coast R. Co., 384 U. S. at 246, and that a carrier cannot be permitted to reap rewards from a strike so much in excess of the rewards of negotiation that it will “have a strong reason to prolong the strike and even break the union.” Id., at 247. The central emphasis of the RLA on continuity of labor relations requires courts to take the long view. See, e. g., Empresa Ecuatoriana de Aviacion v. District Lodge No. 100, 690 F. 2d 838, 845 (CA11 1982), cert. dism’d, 463 U. S. 1250 (1983); National Airlines, Inc. v. International Assn. of Machinists & Aerospace Workers, 416 F. 2d 998, 1006 (CA5 1969).

The Court’s position leaves far too little room for these concerns. By interpreting the RLA as affording protection to striking employees only in the most unusual circumstances, the Court encourages employers to test the limits, knowing that the burden will fall on the employees to demonstrate that the employer’s conduct has crossed an artificially high barrier of “implied” tolerance for employer coercion. The Court thus needlessly creates incentives to undermine long-*457term labor stability and to expand labor conflicts beyond their natural bounds.

In sum, this Court consistently has recognized that there is a difference between traditional self-help economic pressure and coercion or discrimination in derogation of federal law. The Court today continues to recognize this principle, and is willing to “imply” protection in extraordinary circumstances. But Congress did not leave the protection of employee rights to this Court’s selective “implication.” I reject this Court’s failure to give full force to §2 Fourth, the RLA’s express statutory prohibition of coercive and discriminatory employer conduct.

II

Even under the standards the Court articulates today, the result it reaches in this case cannot stand. The Court’s conclusion that TWA’s conduct cannot be said to violate the statutory rights (implied or otherwise) of full-term strikers fails to take seriously the significant discriminatory impact of TWA’s refusal to reinstate full-term strikers. That failure rests on two assumptions that are patently inconsistent with central tenets of federal labor law.

First, the Court appears to suggest that because there were no “vacancies” for the full-term strikers to fill, employer “discrimination” cannot have been a factor in the final allocation of poststrike positions in the active work force. Contrary to this view, this Court long has held that the mere fact that a particular employee occupies a job at the conclusion of a strike does not entitle the employee to retain that job. This is illustrated by our NLRA precedents. Under NLRB v. Mackay Radio & Telegraph Co., 304 U. S. 333, 347 (1938), an employer subject to the NLRA is “not bound to displace men hired to take the strikers’ places in order to provide positions for them” if the employer has found it necessary to promise the replacements permanent employment in *458order to operate during the strike.2 In contrast, positions occupied by new hires to whom no promise of “permanent replacement” status is made are as good as “vacancies” from the full-term strikers’ point of view. The employer’s legal right to resist a union demand for reinstatement flows from the necessity of the offer of permanence; absent such necessity, the employer may be required to furlough (or discharge) the replacements to make room for the strikers’ return. See NLRB v. Fleetwood Trailer Co., 389 U. S. 375, 378-379 (1967); Belknap, Inc. v. Hale, 463 U. S. 491, 514, 517 (1983) (opinion concurring in judgment). The poststrike situation is not, in short, a game of musical chairs: it is governed not by the rule of capture, but by conflicting claims of legal entitlement.

Second, and in tacit recognition that the poststrike situation is governed by law rather than by force or happenstance, the Court elevates the rights of crossovers to the preeminent position, a position which in the Court’s view flows naturally from the RLA’s and NLRA’s protection of “an employee’s right to choose not to strike.” Ante, at 436. From the fact that some employees will disagree with the union’s decision to strike, the Court deduces the proposition that “employees who chose not to gamble on the success of the strike” should not “suffer the consequences when the gamble proves unsuccessful.” Ante, at 438.

The Court’s analysis entirely ignores, and threatens to vitiate, the “‘majority-rule concept [that] is today unquestionably at the center of our federal labor policy.’” NLRB v. Allis-Chalmers Mfg. Co., 388 U. S. 175, 180 (1967), quoting Wellington, Union Democracy and Fair Representation: Federal Responsibility in a Federal System, 67 Yale L. J. 1327, 1333 (1958). “Congress has seen fit to clothe the bargaining representative with powers comparable to those possessed *459by a legislative body both to create and restrict the rights of those whom it represents.” Steele v. Louisville & Nashville R. Co., 323 U. S. 192, 202 (1944) (discussing the duty of fair representation). What the Court characterizes as “their union’s decision to strike,” ante, at 437 (emphasis added), is the decision reached by the majority of the members of the bargaining unit through democratic processes. The right to remain a member of the collectivity but to opt out of the consequences of particular collective decisions when the going gets rough is not a normal incident of participation in the democratic process.

The Court also overlooks the long-recognized fact that the benefits of successful union activity flow to all members of the bargaining unit regardless of their personal support for the union. See Railway Employees v. Hanson, 351 U. S. 225, 238 (1956); Machinists v. Street, 367 U. S., at 762. By elevating the right of crossovers to be “free rider[s],” id.., at 762-763, nn. 13 and 14, to the status of a first principle of labor law, the Court forgets that, by definition, the benefits and burdens of collective action are borne collectively.

This newly asserted statutory right of dissidents to be free from the consequences of collective action buckles under the heavy load the Court asks it to bear. As TWA concedes and the Court recognizes, employers and unions often lawfully agree to displace crossovers through poststrike back-to-work agreements, and employers may unilaterally decide to permit such displacement. See Brief for Petitioner 29; see also Copaz Packing Corp., 115 LRRM 1008, 1008 (1983) (NLRB General Counsel Advice Memorandum) (employers are “privileged to enter into a strike settlement which providefs] that . . . crossovers and strikers who remained on strike until the settlement would be treated equally for recall purposes”); Florida East Coast R. Co., 41 Lab. Arb. 1001, 1006-1007 (1963) (recommendation of Presidential Emergency Board that the carrier replace crossovers and new-hire replacements, who were the poststrike “occupants of the jobs cov*460ered by agreements between the Carrier and the organizations with striking employees to the extent necessary to permit these jobs to be filled on the basis of seniority”). If the right of dissidents to be free of the economic consequences of strikes is so central, it is difficult to see why the union has the power to bargain it away or why the employer has the power to ignore it.

In sum, the Court concludes that TWA’s conduct was lawful on the basis of two assumptions: that the resulting job distribution is justified by the absence of “vacancies” for the returning strikers, and that TWA’s acts were a necessary consequence of its duty to respect the crossovers’ statutory right not to strike. The Court allows these assumptions to stand in the way of considering the adverse impact of TWA’s actions on the full-term strikers’ statutory rights. But I find these assumptions to be without foundation, and thus turn to the question the Court fails to reach.

l — t h — I I — I

A

At the conclusion of the strike, TWA refused to reinstate full-term strikers to positions then occupied by crossovers. In analyzing the lawfulness of TWA’s conduct, certain NLRA principles provide a useful starting point. This Court has recognized under the NLRA that an employer’s refusal to reinstate striking employees discourages employees from exercising their right to organize and to strike, NLRB v. Fleetwood Trailer Co., 389 U. S., at 378, and violates the statutory prohibition against discrimination “unless the employer . . . can show that his action was due to ‘legitimate and substantial business justifications.’” Ibid., quoting NLRB v. Great Dane Trailers, Inc., 388 U. S. 26, 34 (1967). If the employer fails to meet this burden, the inquiry is at an end. Furthermore, in certain circumstances, “the Board can find an unfair labor practice even if the employer introduces evidence that the conduct was motivated by business consider*461ations,” id., at 34, by striking “‘the proper balance between the asserted business justifications and the invasion of employee rights.’” Metropolitan Edison Co. v. NLRB, 460 U. S. 693, 703 (1983), quoting Great Dane, 388 U. S., at 33-34.3

These basic principles are consistent with our RLA precedents. In Railway Clerks v. Florida East Coast R. Co., 384 U. S. 238 (1966), the carrier, during a strike, resorted to self-help in facial violation of § 2 Seventh of the RLA, which prohibits unilateral changes in terms and conditions of employment embodied in collective agreements. The Court held that the carrier could not fulfill its duty to the public to make reasonable efforts to maintain service during the strike if § 2 Seventh were applied with full force during strikes. To accommodate the public interest in continued service, it inter*462preted the RLA as granting the carrier a “closely confined and supervised” power to alter the terms of the agreement during a strike in order to continue service under the particular strike conditions presented by that case. 384 U. S., at 246. The appropriate standard for reviewing a carrier’s alteration of an agreement, the Court concluded, was adequately captured by the words “reasonably necessary,” “provided that ‘reasonably necessary’ is construed strictly” to mean “only such changes as are truly necessary ... for the continued operation” of the carrier. Id., at 248.

In this case, we address conduct that facially violates a different provision of the RLA: § 2 Fourth’s bar against conduct by a carrier which, by its natural tendency, induces or influences employees in their decisions to support or refrain from supporting union activities. The logic of Florida East Coast R. Co., however, is equally applicable here and suggests that a carrier’s refusal to reinstate strikers — conduct which, on its face, violates § 2 Fourth because of its tendency to influence adversely employees’ willingness to support strikes — is unlawful if the refusal was not truly necessary for the continued operation of the carrier during the strike.

In my view, there is no basis under the RLA for a presumption that offers of permanence are necessary in order to induce crossovers and outside replacements to work during a strike. Cf. Hot Shoppes, Inc., 146 N. L. R. B. 802, 805 (1964); Belknap, Inc. v. Hale, 463 U. S., at 504, n. 8 (discussing presumptive necessity of offers of permanence to outside replacements under the NLRA). The Court recognized in Florida East Coast R. Co., 384 U. S., at 246, that a carrier may have need to “improvis[e] and emplo[y] an emergency labor force” in order to continue operations. Under the RLA, as under the NLRA, in short, the Court has recognized that the employer has “the right to protect and continue his business by supplying places left vacant by strikers.” NLRB v. Mackay Radio & Telegraph Co., 304 U. S., at 345. The Union does not here dispute that proposition, *463nor does it question that RLA employers may offer new hires “permanent” status. Cf. id.., at 346. But this Court has also recognized that the public has an interest in the long-term stability of labor relations in industries governed by the RLA. See Virginian R. Co. v. Railway Employees, 300 U. S. 515, 552 (1937). A rule that presumes that replacements and crossovers must be offered permanence would needlessly infringe on that interest in stability.

“There may be some who will. . . argu[e] that employees must take their chances on being permanently replaced when they elect to go on strike. There is little doubt that striking employees have lost their jobs in many firms through the application of this principle. On the other hand, we are concerned in this case not with an ordinary private business but with a common carrier in an industry vital to the public. . . . Experience suggests that the prospects for achieving a ‘peaceable settlement’ of this dispute will remain in jeopardy so long as the striking employees are prevented from working by the presence of the newly-hired replacements. While this situation persists, the organizations can be expected to employ every legitimate means to put pressure on the company to reinstate the strikers. Controversy of this kind may interfere with the legitimate needs of passengers and shippers .... Moreover, other railroads may be tempted to follow the example of this carrier, thus provoking bitter and disruptive disputes in other sections of the country.” Florida East Coast R. Co., 41 Lab. Arb., at 1006-1007.

This risk should be taken only if absolutely necessary to the carrier’s continued operations. Presuming the need does gratuitous damage to significant statutory interests.4

*464B

In his dissent, Justice Brennan does not reach the question whether a carrier who offers permanence to replacements and crossovers is entitled to a presumption of business necessity. Indeed, he would not even permit TWA to make a case-specific showing that its crossover policy was necessary for its continued operation during the strike. Here, our positions differ: I would require the carrier to prove the business necessity of offering permanence to replacements and crossovers on the facts of each case.5

*465Justice Brennan rests his contrary position on NLRB v. Erie Resistor Corp., 373 U. S. 221 (1963). In that case, an employer granted 20 years’ superseniority to employees (new hires and crossovers) who had worked during a strike, which later placed reinstated full-term strikers at a substantial and long-term risk of layoff. There, the NLRB found, and this Court agreed, that “the employer’s insistence that its overriding purpose in granting super-seniority was to keep its plant open and that business necessity justified its conduct was unacceptable since ‘to excuse such conduct would greatly diminish, if not destroy, the right to strike guaranteed by the Act.’” Id., at 225-226 (quoting Erie Resistor Corp., 132 N. L. R. B. 621, 630 (1961)). Because the Court concluded that the stated business justification would not outweigh the asserted interest in continued operation, no factual inquiry into whether the employer’s claim that he could not otherwise have operated during the strike was held to be necessary.

Two considerations cause me to part ways with Justice Brennan’s conclusion. First, it is not so clear to me as it is to Justice Brennan, ante at 449, that TWA’s conduct in this case is sufficiently egregious for its destructive impact to outweigh the interest in maintaining operations during the strike. In Erie Resistor, this Court identified a number of factors that made grants of superseniority particularly harmful to employee rights. Several, but not all, of those factors are present in this case. TWA’s conduct, like the conduct at issue in Erie Resistor, induces employees to abandon the strike and particularly harms full-term strikers. See 373 U. S., at 230-231. But in Erie Resistor the Court stressed the fact that, for years after the strike, reinstated strikers would face a greater risk of layoff because of the additional seniority given to those who worked during the strike. Although Erie Resistor does not suggest an overarching principle identifying which factors are dispositive, the absence of a *466similar continued threat of loss of employment suggests to me that the crossover policy at issue here is not so destructive of employee rights as was the superseniority policy at issue in Erie Resistor. The fact that the Court struck the balance against the employer in Erie Resistor is thus not dis-positive of this case.

Second, and more generally, I am concerned that a standard that permits courts to balance employer and employee interests in the abstract, without a concrete evidentiary record, will lead to erroneous results that endanger the unique statutory interests embodied in the RLA. In the past we have recognized that the public has a significant interest in the continuity of transportation services during labor disputes, and that the RLA protects that interest. Railroad and airline industry employers, we have held, must make “reasonable efforts to maintain the public service at all times, even when beset by labor-management controversies.” Florida East Coast R. Co., 384 U. S., at 245. I recognize that we have stopped short of holding that federal law imposes an absolute duty to operate during strikes, see id., at 250 (White, J., dissenting), and thus have never held that the interest in continued operation cannot be outweighed by other concerns. In my view, however, the balance should be struck on a case-by-case basis and upon a factual record. I expect that it will be a rare case in which gravely destructive carrier conduct will be proved necessary to continued operation under the strict standard of necessity established by Florida East Coast. The ultimate question as to which interest should prevail in such a case is one we can afford to leave unanswered until it is presented on proper facts.

IV

Because the Court of Appeals found TWA’s conduct unlawful without considering whether TWA’s crossover policy was “truly necessary” for continued operations during the strike, I would vacate the judgment of the Court of Appeals and di*467rect that court to remand the case for consideration of that issue.6 Inasmuch as this Court is now reversing outright, I dissent.