dissenting.
Let me begin with the lesser of two concerns — what I consider to be the proper outcome of this appeal. The question before us is whether the arbitrator confined himself to matters within his jurisdiction. 45 U.S.C. § 153, First (q). The arbitrator had jurisdiction only to consider the parties’ agreement, and was charged with issuing a decision drawing its essence therefrom. Walsh v. Union Pac. R.R. Co., 803 F.2d 412, 414-15 (8th Cir.1986); Brotherhood of Ry., Airline and Steamship Clerks v. Kansas City Terminal Ry. Co., 587 F.2d 903, 906-07 (8th Cir.1978). It seems to me that a decision turning largely on a phantom provision, not actually within the parties’ agreement, can hardly be said to have hewed to the specifications of the contract. The district court was justified in vacating the award.
Given that the arbitrator exceeded his jurisdiction, the RLA gives us discretion as to the proper course of action. 45 U.S.C. § 153, First (q) (reviewing court has “jurisdiction to affirm the order of the division or to set it aside, in whole or in part, or it may remand the proceeding to the division for such further action as it may direct”). The court cites authorities supporting the proposition that in the rare case warranting reversal a remand might be appropriate, and chooses to do so in this instance. See, e.g., Union Pacific R.R. Co. v. United Transp. Union, 3 F.3d 255, 264 (8th Cir.*7861993). The equities, however, weigh heavily against the Union, for it was the Union which introduced the phantom provision and which cited it several times in its written submission to the arbitrator. The district court determined the evidence did not support a finding of fraud, a finding I respect. But the Union’s mistake, even if honest, ought to run against it. It should not be given a second bite at the proverbial apple.
The above, however, is but a quibble with the court over the proper disposition of the equities of this matter, something which ought not cause arbitration-law specialists much loss of sleep. The court’s opinion, however, innocuous as it seems, actually contains a much more worrisome problem. Specifically, the court fails to decide that which it is charged with deciding. The only basis for reversal implicated in this appeal is whether the arbitrator exceeded his jurisdiction. Rather than answer that question, the court simply ducks the issue, remanding the dispute on the grounds that the court is “unable to discern whether the arbitrator would have reached the same result” had he not relied on the phantom provision. However, we are not asked to discern what the arbitrator might have done under a different set of circumstances, but rather what the arbitrator did do under the facts presented.
I am uncertain as to where the court draws its authority to order a remand in the face of its stated uncertainty. Certainly it does not do so from the RLA. That statute does permit remands, granting courts “jurisdiction to affirm the order of the division or to set it aside, in whole or in part, or ... remand the proceeding to the division for such further action as it may direct.” 45 U.S.C. § 153, First (q). However, the statute then cabins that grant with the following language:
On such review, the findings and order of the division shall be conclusive on the parties, except that the order of the division may be set aside, in whole or in part, or remanded to the division, for failure of the division to comply with the requirements of this chapter, for failure of the order to conform, or confine itself, to matters within the scope of the division’s jurisdiction, or for fraud or corruption by a member of the division making the order.
45 U.S.C. § 153, First (q). Thus, a remand may occur only after a court has “set aside, in whole or in part,” the arbitrator’s decision, which in turn may occur only upon the statutory bases or upon such other bases as we have judicially created. See, e.g., Union Pacific R.R., 3 F.3d at 261-62 (setting award aside on basis of “well-defined and dominant” public policy). Here, the court punts the predicate question, whether the arbitrator’s award exceeded its jurisdiction, and reverses the district court’s conclusion to that effect. Having done so, the court lacks authority to remand under the RLA.
My major concern with the court’s course of action is that, despite using citations to language limiting our review, in the hands of future litigators the court’s decision may prove a precedential crowbar used to pry open judicial review of RLA awards. By declining to determine whether or not the arbitrator exceeded its jurisdiction, the court creates a third way for us to review arbitration rulings. In future cases we may simply recite the Supreme Court’s directives narrowing our review, but then send back arbitration decisions with which we disagree, on the grounds that we don’t understand what the arbitrator did, in order to give them another chance to “get it right.” If the court had vacated the arbitrator’s decision, as I think it should, then my disagreement with the court would simply be a dispute over the *787equities of the remand, a substantially less worrisome debate. As the opinion stands, however, I think it an imprudent and dangerous precedent.
Accordingly, I dissent.