dissenting:
The majority correctly notes that this court has prescribed a very narrow standard for reviewing National Mediation Board (“NMB” or “Board”) decisions whether to terminate mediation and proffer arbitration. But a narrow standard of review should not be the functional equivalent of no review. Unfortunately, the majority opinion accomplishes that result.
In International Association of Machinists and Aerospace Workers, AFL-CIO v. National Mediation Board et al., 425 F.2d 527 (D.C.Cir.1970), Judge Leventhal announced that
[t]he Railway Labor Act taken as a whole does not fairly require the conclusion that the courts are without jurisdiction to provide a remedy if the Board continues mediation on a basis that is completely and patently arbitrary and for a period that is completely and patently unreasonable, notwithstanding the lack of genuine hope or expectation that the parties will arrive at an agreement.
Id. at 537. The following year this court revisited the mediation issue in Delaware & Hudson Railway v. United Transportation Union, 450 F.2d 603 (D.C.Cir.1971), in which Judge Leventhal indicated. that in this circumstance
[t]he role of the Government officials is considered so sensitive and critical that *247ordinary doctrines permitting judicial action on a complaint of arbitrary delay and protraction are so drastically restricted that court relief from continuation of the process will be available, if at all, only in a most extraordinary situation bordering on patent official bad. faith.
Id. at 608.
Under the majority’s reading and application of the foregoing language, the NMB’s decision to continue mediation is beyond judicial review except where a party seeking relief can demonstrate patent official bad faith. Such a conclusion is unwarranted from our precedent and unsatisfactory. First, by seizing upon the “patent official bad faith” language quoted above, the majority ignores the thrust of Judge Leventhal’s decision in Machinists—that the NMB’s decision whether to continue mediation must be reviewable — and effectively removes such decisions from judicial review. Second, the majority fails to recognize the disparity between the facts in Machinists that gave rise to the standard it applies and the facts of the case sub judice. Finally, the majority’s approach renders superfluous the detailed Section 9A procedures which Congress thought important enough to construct and include in the 1981 amendment of the RLA.
A. Judge Leventhal’s Balance
Anyone reading only the majority’s excerpts from Machinists would likely conclude that Judge Leventhal’s sole focus was to justify excluding meddlesome courts from the delicate mediatory processes established by the RLA. In fact, Judge Leventhal announced the above-quoted standard of reviewing the NMB’s termination decision only after explicitly rejecting the Board’s argument that its decision was unreviewable. See Machinists, 425 F.2d at 533-37. Noting that the rule of law creates a strong presumption that agency action is reviewable by a court, that this time-tested doctrine has been reinforced by the provisions of the Administrative Procedure Act, and that the Union’s right to self-help has strong constitutional and common-law underpinnings, Judge Leventhal concluded that the Board’s refusal to terminate mediation was reviewable. See id. Significantly, in reaching its decision, the court in Machinists distinguished Switchmen’s Union of North America v. National Mediation Board, 320 U.S. 297, 64 S.Ct. 95, 88 L.Ed. 61 (1943), in which the Supreme Court held that a district court was without jurisdiction to review a Board decision certifying representatives of a carrier’s employees for collective bargaining. 425 F.2d at 535-36. The majority fails to recognize this distinction and mistakenly cites Switchmen’s Union and other certification cases to support its position regarding review of NMB decisions not to terminate mediation. See Maj. Op. at 1433.
While the Machinists court recognized that the RLA procedures were designed to be drawn out, Judge Leventhal also stated that
[tjhis perspective does not lead, however, to the conclusion that the process is indefinitely interminable, on the say-so of the agency alone, at least where the court may, as we think it can, reconcile both in theory and in practice a limited power of scrutiny that avoids the threat of administrative absolutism, but does not abort or trammel the administrative process.
425 F.2d at 537-38.
The majority’s approach ineluctably leads to the administrative absolutism that Judge Leventhal sought to avoid. First, the majority states that a court is powerless to order the NMB to terminate mediation unless the Board’s refusal to terminate is “patent official bad faith.” Maj. Op. at 1434. We are instructed further that a reviewing court may not probe the reasons behind a mediator’s decision to continue mediating despite apparent failure to advance toward settlement. Maj. Op. at 1435. Presumably, then, we are left to discern “patent official bad faith” from objective data like the length of the mediation in question. The majority appears to accept this, cautioning that “[o]nly in the most extreme circumstances will a court find a period to be completely and patently unreasonable so as to indicate patent offi*248cial bad faith.” Maj. Op. at 1440-1441. However, earlier in its opinion, the majority states that “[t]he proper time for a dispute to remain in mediation is not susceptible of any objective or statistical estimation by the court.” Maj. Op. at 1438 (emphasis added). How we are to assess whether the length of mediation has exceeded reasonable bounds remains a mystery. I am left to conclude that absent some clear evidence of fraud or corruption, the NMB’s refusal to terminate mediation is effectively beyond a reviewing court’s remedial powers. Instead, under the majority’s scheme, a dispute may be kept in mediation interminably on the “say-so” of the Board alone. Judge Leventhal’s balance is gone: the Board’s decision is effectively unreviewable; the parties’ constitutionally-rooted right to engage in self-help is rendered inaccessible; administrative absolutism reigns.
B. Factual Disparity
The majority employs the standard of review developed in Machinists without regard for the factual differences between that case and the case sub judice. While I acknowledge that language — and particularly a “test” or a “standard”- — often takes on an existence independent of the factual situation that generated it, I think that it is particularly useful to examine the events to which Judge Leventhal was responding in Machinists. Machinists involved a dispute between the International Association of Machinists .and Aerospace Workers (the “Union”) and National Airlines, Inc. (“National”). The parties exchanged notices of proposed changes in the collective bargaining agreement between the Union and National on October 31, 1968. After initial conferences between the parties failed to produce a settlement, National appealed for the assistance of the NMB on December 16, 1968. Before the NMB had assigned a mediator to the dispute, both parties engaged in self-help. On February 27, 1969, the Union began an action in federal district court, seeking an injunction compelling the NMB to assign a mediator. Before the court could rule on this motion, the Board assigned a mediator to the dispute. After the Board replaced the mediator and several mediation sessions had been held, the Union amended its complaint to seek an injunction directing the Board to proffer arbitration. On August 4,1969, the district court ordered the Board to proffer arbitration. The NMB appealed the district court’s order and this court reversed. See Machinists, 425 F.2d at 531-33.
In reversing, the Machinists court emphasized certain factors supporting its decision. In particular, the court noted that although the time spent in mediation (6 months) was “more than customary,” this was excusable because “this dispute involved a large number of issues, some of them complex, and ... the time required to mediate successfully other difficult cases had been even greater.” 425 F.2d at 541. The court stressed that of those issues to which substantial time had been given in mediation, “approximately 40% had been resolved.” Id. Finally, the court noted that the Board’s inability to resolve more of the issues might have been attributable to the short time that the parties had attempted to resolve the dispute in conference before seeking the Board’s mediatory assistance. Id. When these factors are combined with the fact that the entire mediatory process had been conducted “under the shadow of litigation,” id. at 542, it is no wonder that this court was extremely reluctant to interfere with this ongoing effort at settlement. The court was correct to note that the “failure or inconclusiveness [of the mediation] does not demonstrate similar lack of success for a mediation process conducted without the artificial pressure and distortion of an overhanging litigation.” Id.
The facts of the case sub judice are very different. If we deduct the time that this case has been in litigation, the parties were in mediation from October 29,1986 through June 24, 1988—twenty months free from the pressures of litigation that might explain failure to make any progress. Since February 28, 1988, there have been no joint mediation sessions. Further, unlike Machinists, the parties in this case conferenced for almost ten months (August 28, *2491985 through June 10, 1986) before requesting the NMB’s assistance. During this entire time, not one of the issues confronting the parties has been resolved, and the last offer made by Metro-North before litigation began was regressive from the previous offer that the Union’s membership had rejected overwhelmingly five weeks earlier. In short, none of the objective indicia of progress that counseled restraint in Machinists is present in this case.
In contrast, this case presents several compelling reasons for affirming the district court’s decision. First, this case has been in mediation for longer than all but one other Section 9A dispute. Although the majority dismisses any distinction between Section 9A disputes and other disputes under the RLA, I find cogent the district judge’s reasoning for comparing the length of mediation in this case only with the mediation periods of other Section 9A cases. The Board has acknowledged that one of the benefits of an indefinite period of mediation is that it can use its power to terminate mediation in two ways. By withholding termination, the Board can encourage the party seeking change to compromise, as the status quo obtains during the mediation process. By threatening termination of mediation, the Board can pressure the non-moving party to bargain, lest the moving party be allowed immediate resort to self-help. In non-Section 9A cases, this creates a dynamic in which the Board’s power to terminate mediation weighs upon both parties to the negotiations. The statutory framework under Section 9A, however, necessarily diminishes much of the leverage that the Board has in encouraging a non-moving party to reach agreement. By its terms, Section 9A allows this party to extend the status quo, even after the mediation process has failed, by an additional eight months. Where resort to self-help does not pose the same risks of immediacy, the dynamics of negotiation — including when the Board deems it reasonable to terminate mediation — logically v/ould be altered. Comparisons between mediation periods are meaningful only to the extent that the dynamics of the mediations in question are reasonably parallel. Therefore, the district judge’s decision to compare the length of time this case has been in mediation with the mediation periods of other Section 9A cases seems appropriate.
The factors indicating that this mediation has no hope of success, that its continuation uniquely burdens the Union, and that it extends an unintended windfall to Metro-North are numerous: this is one of the longest Section 9A mediations on record; Board Member Wallace made statements to the Union indicating that the Board did not favor Section 9A procedures; this dispute has lasted longer than the collective bargaining agreement that gave rise to this action; in almost two years of pre-litigation mediation and ten months of pre-mediation conferences, not a single issue separating the parties has been resolved; and Metro-North’s final offer was regressive. Individually, any one of these factors fails to move this dispute into the admittedly narrow class of cases in which judicial intervention would be appropriate. Taken together, these individual indicators of mediatory failure coalesce into an unmistakable message to the Board and to this court: it is time to move on to the elaborate procedures that Congress has provided in Section 9A for cases in which mediation has not worked and manifests no reasonable possibility of success. This case is not Machinists.
C. Statutory Scheme
My final and most serious objection to the majority’s approach in this case is that it renders superfluous the elaborate procedures that Congress enacted in Section 9A. Congress, in the RLA, created a very thorough procedure to avoid, alleviate, or at least delay the hardships and economic chaos that could result from strikes in the transportation industry. The majority opinion points out that both the railroads and their unions were very deeply involved in the drafting of the statute. The RLA was aimed at major disputes, the kind that both sides dug in on, as opposed to disputes about the interpretation or administration *250of contractual terms to which the parties had agreed.
In 1981, Congress amended the RLA to provide additional procedures for resolving major disputes within publicly funded commuter railroads. This amendment permits either party to the dispute to extend the status quo period under the statute by as much as 8 additional months by invoking not one, but two presidential emergency boards once the ordinary means for resolving major disputes under the RLA has been exhausted. As a result of this congressionally ordained legislative framework, major disputes between a railroad and its union are sequenced very specifically as to time and procedure — as the majority opinion freely acknowledges.
When it comes to applying this time-specific plan to the instant dispute, however, the court cuts up the procedure in a way totally unblessed by Congress in the statute. Under the court’s reading of the statutory scheme, the parties must follow the initial procedures very meticulously: notices must be served; the disputants must make a show of being in a mediative mode. But then, if either or both parties become intransigent (as often happens in major labor disputes), the majority vacates all of the remaining means of resolution envisioned by Congress — the proffer of arbitration, the availability of the two emergency boards, and the selection of the most reasonable settlement offer by the second emergency board — and converts the mediation board into a Board of the Status Quo. This Status Quo Board can hold the moving party hostage apparently forever — subject only to review for fraud or corruption. This is the procedure the majority finds Congress intended and, presumably, the procedure to which the union and employer factions agreed. One can only imagine the outcry that would have been heard throughout Congress if any Member had proposed to write into the statute the specific procedure framed by the majority. I have no doubt that strong opposition would have come from the railroads as well, as there is no certainty that the NMB would always use its permanent status quo power to benefit management.
Logic dictates that Congress would not have bothered to design the comprehensive and intricate post-mediation procedures of Section 9A if it had perceived that the NMB could keep a dispute in mediation forever absent fraud or corruption. If Congress’ intention in enacting Section 9A was to prolong the time before either party could resort to self-help, as the majority suggests, the finite extension of the status quo that Section 9A permits is at most superfluous if the Board could keep the dispute in mediation interminably on its own say-so. Because we must credit Congress with some reasoned basis for enacting Section 9A, we are compelled to conclude that Congress did not view the Board’s power to retain a dispute in mediation to be virtually unlimited.
D. Conclusion
All agree that courts should not freely intervene in a Board decision to hold a dispute in mediation. But, consistent with Machinists and the procedures of the RLA as enacted by Congress, a trial judge must have discretion in measuring the reasonableness of the Board’s delay in acknowledging mediatory failure and moving on to arbitration and the other procedures mandated by Section 9A. Many times both sides will want to mediate. Many times one side or the other will fear the next phase of dispute resolution provided for in the RLA’s detailed sequence. Many times the resisting party will not be capable of self-help. In any event, it is not for this court to rewrite a statute and impose our values favoring unlimited mediation on this thorny and contentious field. Congress developed and mandated the statutory sequence; there is no place for imposing a half-time delay of unknown and unknowable duration in a statutory scheme that Congress worked out in such detail. Where, as here, the trial court legitimately concluded that there was no evidence of progress toward resolving a protracted dispute and that prolonging the status quo weighed unconscionably upon one party to *251this dispute, the clear consequences of the statute ought to be invoked. I dissent.