dissenting:
With respect for my brethren of the majority, I nevertheless dissent from their holding that the present order of remand by the Benefits Board of Review is a “final order” and, thus, subject to judicial review at this time. 33 U.S.C. § 921(c). In a commendable effort to expedite judicial review in the present ease, the majority has adopted a principle of case-by-case determination of “finality” that will, however, unsettle the relative certainty that was previously attached to that concept. While the benefits of expeditious review in the present case may be apparent, the general rule enunciated will require, as its consequence, many premature appeals in the future and much additional resultant delay before the final award or rejection of compensation claims.
In the present case, the principal (but not the only) issue before the Administrative Law Judge and the Board was whether the injured employees should be awarded compensation based upon § 10(a) (the employee’s wages during the prior year), or upon § 10(b) (prior years’ wages of coworkers), or upon § 10(c) (an alternative, when either of the former methods “cannot reasonably be applied” to arrive at the annual earnings of the injured employee) of the Act, 33 U.S.C. § 910(a), (b), or (c). Affirming the ALJ, the Board held that § 10(c) applied to the facts of this case, but it remanded for additional findings of fact, with leave to reopen the record to take additional evidence, for a redetermination of the claimant’s weekly wage, as well as for consideration of other issues, such as attorney’s fees. Independent of whether we ultimately decide whether the Board was correct in finding § 10(c) to provide the proper method for computing the award of compensation,1 we must first decide whether this order was “final” so as to permit us to exercise appellate jurisdiction to review it.
*753The majority notes, correctly, the settled jurisprudence to the effect that the “final order” appealability-requirement of 33 U.S.C. § 921(c), with regard to Board decisions, is essentially determined by the same criteria that apply to the “final judgments” appealability-requirement of 28 U.S.C. § 1291 with regard to district court decisions.2 As to the latter, the Supreme Court recently reiterated the test of finality as being “ ‘a decision by the District Court that “ends the litigation on the merits and leaves nothing for the court to do but execute the judgment.” ’ ” Firestone Tire & Rubber Company v. Risjord, 449 U.S. 368, 373, 101 S.Ct. 669, 673, 66 L.Ed.2d 571 (1981). The Court emphasized several of the important purposes served by the requirement of finality: the avoidance of piecemeal appeals and of the obstruction to just claims by permitting a succession of separate appeals from the various rulings to which litigation may give rise, as well as the promotion of efficient judicial administration. Id. As this court itself has reiterated,
[The finality] requirement has the support of considerations generally applicable to good judicial administration. It avoids the mischief of economic waste and of delayed justice. Only in very few situations, where intermediate rulings may carry serious public consequences, has there been a departure from this requirement of finality for federal appellate jurisdiction.
In re Corrugated Container Antitrust Litigation, 611 F.2d 86, 89 (5th Cir.1980).
The present Board order, although determining a central issue of the claim — the method of calculating weekly compensation due — , remanded the proceedings to the Administrative Law Judge for further findings as to the actual weekly rate to be awarded and as to attorney’s fees. It did not end the litigation on the merits, and no final administrative order awarding weekly compensation and attorney’s fees would be entered until after the remand.
Under such circumstances, where in effect a central issue of liability has been determined to administrative finality by the Board but remand is ordered to calculate and make a specific monetary award, the decisions until the present have uniformly refused to consider the Board order of liability as “final” for purposes of judicial reviewability at that time, both in this circuit, United Fruit Company v. Director, Office of Workers’ Compensation Programs, 546 F.2d 1224 (5th Cir.1977), and in other circuits, Director, Office of Worker’s Compensation Programs v. Brodka, 643 F.2d 159 (3d Cir.1981); National Steel and Shipbuilding Company v. Director, Office of Workers’ Compensation Programs, 626 F.2d 106 (9th Cir.1980); Newport News Shipbuilding Corporation v. Director, Office of Workers’ Compensation Programs, 590 F.2d 1267 (4th Cir.1978). As stated in one of the lead decisions applying this rule, “[i]t is a well-established rule of appellate jurisdiction . . . that where liability has been decided but the extent of damages remains undetermined, there is no final order.” Sun Shipbuilding & Dry Dock Company v. Benefits Review Board, 535 F.2d 758, 760 (3d Cir.1976).
The virtues of this rule are apparent. It avoids dual appeals on both liability and award, with the consequent doubled judicial appellate delays, in favor of a single review of the final administrative award. An injured worker, whose award is vacated by the Board for recalculation as to the properly exact amount of weekly compensation due, does not suffer the prolonged denial of compensation that will occur as his employer seeks judicial review of the singled issue of liability (which, moreover, in the prepon*754derant number of cases will be affirmed) decided by a Board opinion that affirms liability but vacates and remands for exact calculation. The certainty of the rule of nonfinality in such instances permits all parties to rely on judicial review only after administrative remand and final calculation of the actual award, without being forced to take a precautionary appeal to the courts lest a failure to do so result in the loss of judicial review of the liability issue as determined by the Board’s remand order, if it is subsequently characterized as “final” as to the liability decision.
In failing to follow this (until now) well-settled jurisprudential rule, the majority analyzes the facts and issues of this particular case, and decides that because in this particular ease the appellate court decision now will determine the proper legal standard to be applied on the remand, and minimize the risk of a wasted agency hearing and later appeal, an appeal at this stage of the proceedings in this case will be more efficient and will not prejudice the parties (because the proper award of compensation cannot be made until the conclusion of the litigation). The majority concludes that we will “recognize appellate jurisdiction in this case• where [1] the substantive legal issue is clearly posed and [2] all that will remain after it is decided is use of the record already completed to calculate the wage base for recovery.”
Before noting some of the practical consequences to this newly created exception from the “finality” requirement for review-ability of administrative orders, I point out the tenuous authority upon which creation of this exception is based. Unlike all of the previously cited decisions, the majority does not advert to the traditional and quite limited exceptions to the finality rule rarely permitted. See, e.g., Brodka, supra, 643 F.2d at 163 & n. 9 (collateral order final in nature and severable from merits; irreparable injury; unresolved issues purely ministerial in nature); 16 Wright, Miller, Cooper and Gressman, Federal Practice and Procedure, § 3942, see esp. at p. 314 (1977); 15 Wright, Miller and Cooper, Federal Practice and Procedure, § 1910-13 (1976). Rather, the majority purports to find in Ingalls Shipbuilding Division v. White, 681 F.2d 275 (1982), a rather free-wheeling authority to dispense with the finality requirement in any individual case where the spectre of piecemeal review is not raised because in a given case the record is complete, the issue presented is a legal standard rather than a factual dispute, and the same legal issue (perhaps incorrectly decided initially by the Board) would recur on the subsequent resort to judicial review after the remand.
Ingalls, however, is a case presenting exceptional and distinguishable facts. The sole issue there concerned whether an administrative law judge had the power to approve a settlement, with an important collateral question as to the reviewability or not of a settlement so approved, as against contentions (ultimately upheld by.us) that the judge was without authority to do so and that the Director, Office of Workers’ Compensation Programs, did have standing to appeal the ALJ’s order approving the compromise. The Board had upheld such power in the ALJ, but had remanded for his reconsideration of certain factors. Balancing competing considerations of piecemeal review and the danger of denying justice by delay in the decision of this crucial issue in the administration of the compensation act, Ingalls concluded that “the facts here fall within the unique situation that is established as an exception to technical finality in Gillespie [v. United States Steel Corp., 379 U.S. 148, 85 S.Ct. 308, 13 L.Ed.2d 199 (1964).]” 681 F.2d at 279 (emphasis added).
Ingalls did not explicitly specify the unique and exceptional facts. However, its reference to Gillespie — where, as in Ingalls, there was a close or at least unresolved issue as to the appealability itself of the basic order (in Ingalls, on whether any order of an ALJ approving a settlement was reviewable at the instance of the Director) —would indicate to me that no blanket exception to finality was intended whenever, in a panel’s view, the spectre of piecemeal review was not raised by deciding the initial appeal. That the “pragmatic finali*755ty” exception to finality recognized by Gillespie applies only in extremely limited and extraordinary circumstances may be recognized by the Supreme Court’s recent characterization of that decision’s holding:
In Gillespie, the Court upheld an exercise of appellate jurisdiction of what it considered a marginally final order that disposed of an unsettled issue of national significance because review of that issue unquestionably ‘implemented the same policy Congress sought to promote in § 1292(b),’ id., at 154, 85 S.Ct. at 312, and the arguable finality issue had not been presented to this Court until argument on the merits, thereby ensuring that none of the policies of judicial economy served by the finality requirement would be achieved were the case sent back with the important issue undecided.
Cooper & Lybrand v. Livesay, 437 U.S. 463, 477 n. 30, 98 S.Ct. 2454, 2462 n. 30, 57 L.Ed.2d 351 (1978).
The present petition for judicial review of a Board order presents no unusual or unique circumstances such as are implicated in Gillespie or in the quite limited exceptions to finality previously recognized. Absent the present decision by the majority, there was no issue of marginal appealability that justified this court’s immediate determination of the appealability issue and of the correctness of the underlying order. The issue decided, although important, is not of such national significance to the administration of the Act as to require its decision on this (premature) appeal rather than by the appeal from the final order that awards compensation and resolves all unsettled issues. Nor, for reasons to be set forth, does the new general exception to finality created by the majority avoid thwarting the purposes of the finality rule.
Rather, by importing into the finality concept a case-by-case measure of exception, in general application the newly created judicial exception substantially impedes these purposes of the finality concept and must inevitably result in multiple appeals, additional delays before final decision, and new issues of appellate jurisdiction that will impede the decision by judicial review of the merits of administrative orders. This may be illustrated by reference to the present proceedings.
The majority’s exception permits immediate reviewability, despite technical non-finality, where the record is complete as to factual matters and where only a legal standard is at issue, immediate decision of which will facilitate the ultimate decision after remand. With regard to the present case, I note:
1. The majority assumes the record is complete, probably subconsciously influenced by its ultimate determination that § 10(c) does not apply and that the record now contains all facts needed to decide under § 10(b). Absent that determination, the record is not complete, as the Board recognized in vacating and remanding for additional findings of fact, including the taking of evidence if necessary, to make a proper weekly award under § 10(c). If we had affirmed (instead of reversed) the Board’s determination that § 10(c) furnishes the proper measure of weekly disability compensation, the effect of our decision retaining appellate jurisdiction would be to remand (as the Board ordered on June 10, 1981, some eighteen months ago) to the Administrative Law Judge to calculate the weekly award, following which the employer (or the claimant) would be entitled once again to appeal this now-final order. Whether the majority had affirmed or reversed the Board’s non-final order, however, in either event the majority’s new exception rule builds into the administrative compensation-awarding processes an additional delay, in this case extending more than a year and a half (so far), by allowing an initial appeal as to the liability-standard, to be followed (quite possibly) by yet a second resort to the review process after, on remand, the ALJ fixes the rate of compensation and decides the subsidiary undecided issues, such as attorney’s fees.
2. The employer sought judicial review of the Board’s remand order, until now considered a non-reviewable non-final order. Assume that instead the parties had *756followed the previously accepted methodology, applying for judicial review only after a final order, and thus had complied with the remand and secured an expeditious administrative determination of unresolved issues, including the compensation, to be incorporated in a technically as well as actually final order. Upon judicial review from the true final order, the parties are faced with the possibility that, under the majority’s new “pragmatic finality” exception, their failure to appeal the initial “final” order (determining only the standard of liability) forecloses subsequent review of the initial liability decision.3 What must the parties do to avoid this possibility? Obviously, to avoid this possibility, a party must seek initial review of the initial Board remand-order, although it involves considerable additional appellate delays (here, eighteen months so far), and although the court may decide that, after all, the appeal is premature because the order is not “final” under the flexible finality rule envisioned by the majority’s new exception.
3. If the previously accepted concept of “finality” for purposes of judicial review had been followed in this case, then on the single judicial review the courts would be concerned only with the merits of the litigation. (If a party appealed despite the accepted rule, the usual practice in this court would probably result in a dismissal by summary order of an administrative panel, since the rule of finality previously followed was certain and was easily administerable.) Under the new rule of appealability sanctioned by the majority, a new adjective issue of appealability is injected into every appeal from a Board remand-order or from an order by the Board following an unappealed Board remand-order.4 Is or was the initial order “pragmatically” final? This will involve individual analysis in each case to determine whether where what the Board decided was a “legal standard” (instead of the factual application of a legal standard, a mixed law-fact question, for example) and whether the initial record was “complete” such that immediate judicial review of the initial order would add to the efficiency of disposition and would theoretically avoid the spectre of piecemeal judicial review. To replace the former certain definition of administrative finality, the majority substitutes a rule that requires case-by-case analysis in each individual appeal, that in many instances will result in the prolongation of judicial review by the frequent necessity of oral argument and collegiate *757interchange as to the variables now injected into this adjective issue of appealability per se — and one as to the application of which, I predict, reasonable appellate minds will differ. For the previously certain rule of “technical” as well as actual finality, the majority substitutes a flexible “pragmatic” finality rule of uncertain case-by-case application that must inevitably encourage precautionary interlocutory appeals, and their consequent delay, and that may in an appreciable number of eases delay the decision of the merits by injecting a complicated threshold adjective issue of appealability.
4. And what about the claimant (such as poor Roundtree, here), as the reviewing court entertains the initial appeal as to the legal standard of liability, despite a remand by the Board to calculate the weekly compensation. Here, back in June 1981, the Board vacated the provision of the order of the Administrative Law Judge awarding Roundtree a weekly compensation award, see Record Excerpts, p. 43, and remanded to that judge for him to make further findings and to take further evidence, if necessary, in order to calculate the proper weekly amount. We are informed that here the employer has continued to pay the claimant weekly compensation.
However, when the Board vacates the only order that requires an employer to pay weekly compensation, then — in the absence of a subsequent order by the administrative judge on the remand renewing the employer’s obligation to pay weekly compensation — I would suppose that, as the Director in brief suggests, it is at least arguable whether the employer was obliged to continue to make compensation payments to its disabled employee before there was an administrative order fixing their actual amount. Should in such circumstances an employer terminate weekly compensation until an order as to its weekly amount is finally issued, the availability of interlocutory review of the Board’s remand order (deciding the measure of liability, but remanding for calculation of the amount due) will add a considerable period of delay during which the disabled employee will receive no weekly compensation (here, more than eighteen months, so far), solely for the purpose of judicial review of the Board’s interlocutory order. Only following conclusion of that interim lengthy judicial review of the measure-of-liability portion of the Board’s remand-order, will the proceedings finally be remanded to the Administrative Law Judge to issue an order awarding weekly compensation in calculated amount — as would have occurred, absent interim judicial review, many months earlier. I cannot believe that either the purposes or intent of the compensation act are served by an interpretation of the judicially-reviewable “finality” provision of 33 U.S.C. § 921(c), that will insert such additional interlocutory-review delays that may disrupt for an appreciable time the continued payment of weekly compensation to an admittedly disabled employee solely for the purpose of affording interlocutory judicial review.
I sympathize with my brothers of the majority in their commendable desire to decide now a central issue of this proceeding, since the appeal is already here, because in this case they view it to be more efficient to do so, under the particular issue and record presented to us. I share their general view that formal procedural rules should be applied in the light of their function, and that a practical rather than a technical construction is often suitable to advance the end purpose of a procedural rule. Here, however, the very purpose of the finality rule is thwarted by the practical construction accorded it; although it may advance the efficient decision of this particular case, the flexible rule of pragmatic finality adopted by the majority will impede the efficient and expeditious decisions of many, many appeals from administrative remand-orders in the future. At the same time, the newly created exception to finality replaces with large elements of uncertain administration the previously certain and efficient rule of finality as accorded to Board orders. The need for clear and easily administered rules as to the threshold adjective issue of reviewability overrides any values of flexible case-by-case administra*758lion. As has been stated in a related context:
Although well-established rules of appealability might at times cause an action to be determined unjustly, slowly, and expensively, they have nonetheless the great virtue of forestalling the delay, harassment, expense, and duplication that could result from multiple or ill-timed appeals. The great value of the final judgment rule may well be that it combines generally effective review with guides sufficiently clear to prevent most of the great waste that could result from protective appeals and litigation over appellate jurisdiction. Earnest pursuit of a “practical approach” could quickly destroy this accomplishment.
15 Wright, Miller, and Cooper, supra, § 3913 at p. 523.
I therefore must respectfully dissent from the majority’s opinion.
. Here, for instance, the majority decides that instead, § 10(b) is the proper formula, so that more efficient non-piecemeal review is appropriate. However, had the majority decided that the Board correctly applied § 10(c), the result of the majority rationale would permit (or require, see text below) two judicial reviews of the same award — the first (the present) as to the method of calculating the compensation benefits, and the second, if one of the parties disagreed as to the actual dollar rate of the actual award or as to the attorney’s fees or other remaining issues in the case.
I may add that I entertain very serious reservations as to the correctness of the majority’s conclusion that the Board improperly used § 10(c) under the Board’s determination that neither § 10(a) nor § 10(b) provided a reasonable and fair method of determining the “annual earning capacity”, § 10(c), of this particular employee, injured on the first day of his new employment after having left much more highly remunerative work of the same nature. Cf. also § 8(h), defining “wage-earning capacity” and providing that “actual earnings” measure if they “fairly and reasonably represent his earning capacity.”
I read § 10(c) as a Congressional grant of administrative discretion, subject to review only as an abuse, to determine the annual earning capacity of an injured employee where § 10(a) and § 10(b) do not reasonably and fairly result in a calculation of the annual earning capacity of the particular employee injured. See, to same effect, 1A Benedict on Admiralty, § 74 at pp. 4-41 through 44 (7th ed. 1982). Thus, § 10(a) and § 10(b) were not legislatively intended to be a statutory straight jacket mechanically applied. In the present case, had the claimant Roundtree worked for the present employer (instead of for himself, doing the same type of work) for the preceding year, it would be obvious that his previous year’s earning capacity and annual wages were greater than those of his present co-employees who had worked “in the same or in similar employment in the same or a neighboring place”, § 10(b), during the preceding year (upon which compensation is based, rather than on the increased pay such employees were now paid during the year of Roundtree’s injury). In terms, § 10(b) seems addressed to an injured employee who had not worked for the employer defendant “during substantially the whole of such [preceding] year,” in which event his average annual wage is calculated as if he had worked the whole rather than only part of the year for that employer. However, since I do not believe we have jurisdiction to review this interlocutory order of the Board, I will comment no further as to the majority’s ruling on the merits.
. Particular requirements of an administrative review scheme may sometimes impose “special considerations” on the determination of finality that vary from the “classical jurisdictional requirements” applied to appeals from district courts. Sun Shipbuilding and Dry Dock Company v. Benefits Review Board, 535 F.2d 758, 760 (3d Cir. 1976), see Weinberger v. Salfi, 426 U.S. 749, 764-67, 95 S.Ct. 2457, 2566-67, 45 L.Ed.2d 522 (1975). No such particular requirements or special considerations are relied upon or appear with regard to the present clearly delineated administrative and judicial review scheme.
. This type of problem is illustrated by Croker v. Boeing Co., 662 F.2d 975 (3d Cir.1981) (en banc). The district court entered judgment on the merits against the defendant in October 1979, leaving for determination only the amount of attorney’s fees. In March 1980, the district court set the amount of attorney’s fees, and the employees timely appealed within thirty days of the latter order. The defendant moved to dismiss the appeal by the plaintiffs as to certain merit-rulings in the October 1979 judgment. The defendant contended that the 1979 judgment was final, since it had decided all issues in the appeal. The court of appeal rejected this contention, holding that the 1979 judgment was not “final” under the general test that it did not terminate the litigation and leave nothing to be done to enforce the judgment.
The majority, rather unsuccessfully in my view, distinguishes this decision and numerous others to same effect by noting that the record at the time of the initial order did not contain all the factual matter necessary for the final determination of the amount of attorney’s fees, so the initial order was not “final” and was not reviewable at that time. I am unable to see how the present record differs' — even aside from the lack of complete evidence to decide the § 10(c) issue. In the present case also, there was not a complete factual record necessary to fix the amount of attorney’s fees due, an issue that the Board had remanded to the ALJ. This circumstance may indicate some of the difficulty and uncertainty in determining whether there is a “complete record” so as to permit a reviewing court to consider the order under review as “final”.
Perhaps some of the obvious possibilities of injustice and the necessity for precautionary appeal could be avoided by holding that, when an appeal is perfected, the order may be pragmatically “final” enough to be reviewable, but that a failure to appeal at that time will not preclude subsequent review of the technically (and actually) ■ “final” order. However, such flexible interpretation would thwart the purposes of the finality concept, aside from raising fundamental issues of judicial disregard of the legislative final-order basis of appellate jurisdiction.
. See note 3 supra.