concurring.
Although I concur in the result reached by the majority, I take advantage of this in banc opportunity to set forth my dissatisfaction with the rule both the majority and dissenting opinions apply. The rule fabricated by this court on previous occasions provides that when the NLRB proposes a Gissel bargaining order, it must articulate in detail the factors that justify its choice of this remedy rather than ordering a new election.1 See Hedstrom Co. v. NLRB, 629 F.2d 305 (3d Cir. 1980) (in banc) (Hedstrom II) cert. denied, 450 U.S. 996, 101 S.Ct. 1699, 68 L.Ed.2d 196 (1981); Hedstrom Co. v. NLRB, 558 F.2d 1137 (3d Cir. 1977) (Hedstrom I); NLRB v. Armcor Industries, Inc., 535 F.2d 239 (3d Cir. 1976). In my view, this procedural rule was promulgated without authority and rests on certain unwarranted assumptions about the factfinding competence of Administrative Law Judges never openly articulated in the opinions supporting the Armcor-Hedstrom rule. Moreover, our decisions indicate that rather than applying the rule in a neutral fashion, the panels of this court have manipulated it primarily to achieve the results a given panel majority desires.2
The rule has experienced a battered history in this court. The issue of what the ALJ and NLRB must find and say to impose a bargaining order has required three in banc sittings, NLRB v. Armcor Industries, Inc., No. 77-1495 (November 15, 1978) (Armcor II) (unreported per curiam),3 Hedstrom II, and the present case, and a number of panel opinions, e. g., NLRB v. K&K Gourmet Meats, Inc., 640 F.2d 460 (3d Cir. 1981); Electrical Products Division of Midland Ross Corp. v. NLRB, 617 F.2d 977 (3d Cir.), cert. denied, 449 U.S. 871, 101 S.Ct. 210, 66 L.Ed.2d 91 (1980); NLRB v. Garry Mfg. Co., 630 F.2d 934 (3d Cir. 1980); Rapid Mfg. Co. v. NLRB, 612 F.2d 144 (3d Cir. 1979). I joined the majority in Hedstrom II with the hope that the rule could be applied evenhandedly and because I felt that its purpose could be considered salutary: “to assist a reviewing court in determining whether the standards of Gissel have been satisfied.” Hedstrom II, 629 F.2d at 309. The rule, I thought, was intended to add predictability and stability to an important area of labor law.4 See majority op., at 519, 520. That the present case should be before us in banc so quickly after Hedstrom II eloquently attests to the utter failure of the rule to achieve its goal. We are in banc because the original panel disgorged three separate views on how to decide what I consider to be a very pedestrian case.
I.
In my view, this court’s first error is its wrongful assumption of authority to promulgate an essentially procedural rule *523for the NLRB. This is not the only occasion when this court has ventured into rule-writing escapades for a federal administrative agency. We began these independent intellectual frolics when we imposed ourselves as a special procedural rules committee for the Social Security Administration. See, e. g., Hargenrader v. Califano, 575 F.2d 434 (3d Cir. 1978) (requirement in Social Security disability claim that hearing officer articulate “a statement of subordinate factual foundations on which ultimate factual conclusions [sic] are based”5); Baerga v. Richardson, 500 F.2d 309 (3d Cir. 1974) (dicta), cert. denied, 420 U.S. 931, 95 S.Ct. 1133, 43 L.Ed.2d 403 (1975). Although such a requirement may be “helpful in the joint work of the administrative agency and the courts,” Baerga, 500 F.2d at 313, this observation alone cannot justify our intrusion into the procedures of a coordinate branch. This is not a federal common law area, nor an exercise of supervisory authority over courts inferior in the hierarchy. We must therefore look to Congress for any authority because, as I maintained in Hargenrader, 575 F.2d at 438-39 (dissenting opinion), federal courts of appeals were not given general supervisory powers over agencies of the executive branch. See also Cotter v. Harris, 642 F.2d 700, 708 (3d Cir. 1981) (Garth, J., dissenting). I found no Congressional authority in the Social Security cases; I find none here.
The judiciary’s relationship to the NLRB is defined by 29 U.S.C. § 160(e) & (f). We are instructed that “findings of fact if supported by substantial evidence on the record considered as a whole shall be conclusive,” 29 U.S.C. § 160(e). Congress has given little additional guidance for reviewing Board orders. When put to the task of defining the courts of appeals’ scope of review of the NLRB’s remedial orders, the Supreme Court has stressed our limitations:
A statute expressive of such large public policy as that on which the National Labor Relations Board is based must be broadly phrased and necessarily carries with it the task of administrative application. There is an area plainly covered by the language of the Act and an area no less plainly without it. But in the nature of things Congress could not catalogue all the devices and stratagems for circumventing the policies of the Act. Nor could it define the whole gamut of remedies to effectuate these policies in an infinite variety of specific situations. Congress met these difficulties by leaving the adaptation of means to end to the empiric process of administration. The exercise of the process was committed to the Board, subject to limited judicial review. Because the relation of remedy to policy is peculiarly a matter for administrative competence, courts must not enter the allowable area of the Board’s discretion and must guard against the danger of sliding unconsciously from the narrow confines of law into the more spacious domain of policy.
Phelps Dodge Corp. v. NLRB, 313 U.S. 177, 194, 61 S.Ct. 845, 852, 85 L.Ed. 1271 (1941). The Board’s remedial orders may not be set aside “unless it can be shown that the order is a patent attempt to achieve ends other than those which can fairly be said to effectuate the policies of the act.” Virginia Elec. & Power Co. v. NLRB, 319 U.S. 533, 540, 63 S.Ct. 1214, 1218, 87 L.Ed. 1568 (1943).
Contrary to the position of this court, review of a Gissel bargaining order has been dealt with no differently:
It is for the Board and not the courts ... to make [the] determination, based on its expert estimate as to the effects on the election process of unfair labor practices of varying intensity. In fashioning its remedies under the broad provisions of § 10(c) of the Act (29 U.S.C. § 160(c)), the Board draws on a fund of knowledge and expertise all its own, and its choice of remedy must therefore be given special respect by reviewing courts. See Fibreboard Paper Products Corp. v. NLRB, 379 U.S. 203 [85 S.Ct. 398, 13 L.Ed.2d 233] *524(1964). “[I]t is usually better to minimize the opportunity for reviewing courts to substitute their discretion for that of the agency.” Consolo v. FMC, 383 U.S. 607, 621, 86 S.Ct. 1018, 1027, 16 L.Ed.2d 131 (1966).
Gissel, 395 U.S. at 612 n.32, 89 S.Ct. at 1939. The Court could not have been more clear in its limitation of review, yet circuit judges continue to be so “profoundly convinced of their own expertise in such matters,” NLRB v. General Stencils, Inc., 472 F.2d 170, 176 (2d Cir. 1972) (Hays, J., dissenting), that they freely impose procedural rules on the Board when the remedy chosen is a bargaining order. Our review is much more limited: was there substantial evidence to support the factual findings? or, was the decision arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with the law?
Of course the Supreme Court does not require reviewing courts to rubber-stamp NLRB decisions. Rather, the Court has required the Board to provide some articulation of the reasons for its decisions. See NLRB v. Metropolitan Life Ins. Co., 380 U.S. 438, 443-44, 85 S.Ct. 1061, 1064, 13 L.Ed.2d 951 (1965); 6 see also SEC v. Chenery Corp., 332 U.S. 194, 195-96, 67 S.Ct. 1575, 1577, 91 L.Ed. 1995 (1947) (Chenery II) (administrative decision must reveal its basis). These precedents, however, do not authorize the meticulous inquiry this court undertakes in every Gissel case. They require no more than an indication that the agency has in fact exercised its discretion. See Metropolitan Life, 380 U.S. at 443, 85 S.Ct. at 1064. It is up to the agency to decide how detailed this explanation must be once it crosses the minimum threshold of clarity.7 If the record is insufficient to reveal the bases for the decision, the court can remand to the agency for “such additional explanation of the reasons for the agency decision as may prove necessary.” Camp v. Pitts, 411 U.S. 138, 143, 93 S.Ct. 1241, 1244, 36 L.Ed.2d 106 (1973). Even in conceding that a remand may at times be appropriate, however, the Court indicated that minimal explanations suffice. “The [agency’s] explanation may have been curt, but it surely indicated the determinative reason for the final action taken: the finding that a new bank was an uneconomic venture in light of the banking needs and the banking services already available in the surrounding community.” Id. Although it remanded the case to the court of appeals, it suggested that little else need be said for the court to review the decision. Nor do I understand the Supreme Court to require more exhaustive explanations when *525the NLRB chooses to impose a Gissel bargaining order simply because elections are a favored manner of selecting a bargaining representative. Cf. NLRB v. Walton Mfg. Co., 369 U.S. 404, 407-408, 82 S.Ct. 853, 854-855, 7 L.Ed.2d 829 (1962) (per curiam) (“There is no place .. . for one test of the substantiality of evidence in reinstatement cases and another test in other cases.”).
The rule applied in the present case goes far beyond the minimum threshold, and the dissent would require far more. As Justice Black observed in Chenery I, a judicial requirement of detailed findings as the price of court approval of administrative action will “bog the administrative power in a quagmire of minutiae.” 318 U.S. at 99, 63 S.Ct. at 464 (dissenting opinion). More important, “[h]ypercritical exactions as to findings can provide a handy but an almost invisible glideway enabling courts to pass ‘from the narrow confines of law into the more spacious domain of policy.’ ” Id. (quoting Phelps Dodge, 313 U.S. at 194, 61 S.Ct. at 852).
The Supreme Court has recently reemphasized that federal courts should not interfere with administrative action in areas in which Congress has charged an agency with primary responsibility for settling difficult and complex social problems. It acknowledged that agencies must grant a minimum threshold of procedural rights, but stressed that although “[a]gencies are free to grant additional procedural rights in the exercise of their discretion, . . . reviewing courts are generally not free to impose them if the agencies have not chosen to grant them.” Vermont Yankee Nuclear Power Corp. v. NRDC, 435 U.S. 519, 524, 98 S.Ct. 1197, 1202, 55 L.Ed.2d 460 (1978). The Court warned:
[T]his much is absolutely clear. Absent constitutional constraints or extremely compelling circumstances the “administrative agencies ‘should be free to fashion their own rules of procedure and to pursue methods of inquiry capable of permitting them to discharge their multitudinous duties.’ ” FCC v. Schreiber, 381 U.S. [279, 290, 85 S.Ct. 1459, 1467, 14 L.Ed.2d 383 (1965)], quoting from FCC v. Pottsville Broadcasting Co., 309 U.S. [134, 143, 60 S.Ct. 437, 441, 84 L.Ed. 656 (1940)]. Indeed, our cases could hardly be more explicit in this regard. .. . [T]he basic reason for [the] decision [in FCC v. Schreiber ] was the Court of Appeals’ serious departure from the very basic tenet of administrative law that agencies should be free to fashion their own rules of procedure.
We have continually repeated this theme through the years. . . . [W]hile a court may have occasion to remand an agency decision because of the inadequacy of the record, the agency should normally be allowed to “exercise its administrative discretion in deciding how, in light of internal organization considerations, it may best proceed to develop the needed evidence and how its prior decision should be modified in light of such evidence as develops.” [FPC v. Transcontinental Gas Pipe Line, 423 U.S. 326, 333, 96 S.Ct. 579, 583, 46 L.Ed.2d 533 (1976)].
435 U.S. at 543-44, 98 S.Ct. at 1211 (footnotes omitted); see also Chrysler Corp. v. Brown, 441 U.S. 281, 312-13, 99 S.Ct. 1705, 1723, 16 L.Ed.2d 208 (1979); Wallace Corp. v. NLRB, 323 U.S. 248, 253, 65 S.Ct. 238, 240, 89 L.Ed. 216 (1944). The Armcor rule simply cannot be justified on grounds that a Gissel bargaining order violates “constitutional constraints” or presents “extremely compelling” circumstances. Rather than being one of those “extremely rare” cases, the imposition of a bargaining order as a remedy is typical of the matters Congress delegated to agency discretion for the resolution of labor conflicts.
The rule announced in Armcor and applied here ignores the Supreme Court’s admonitions and leads to needless friction between the courts and agencies with, at best, marginal benefits for the litigants and employees. In outlining permissible Gissel orders, the Supreme Court recognized the complexity of the decision and clearly left it in the hands of the agency. See 395 U.S. at 612 n.32, 89 S.Ct. at 1939. The Board, in turn, recognizing the peculiarities of each Gissel case, has avoided the promulgation of *526precisely defined rules, instead treating each case on its facts. See, e. g., General Stencils, 195 NLRB 1109 (1972).
II.
Because I find no Congressional authority for the rule, I conclude that other reasons underlie its continued use. I perceive two reasons for the rule’s vitality even though in practice it yields unpredictable results: first, the rule sounds so obviously good that one opposes it only at the risk of inveighing against motherhood and apple pie, and second, the rule can be manipulated to achieve the subjective ends of the judges as they apply it more or less strictly. I will take the risk, and focus on the second reason. That the rule can be manipulated to produce desired results is evident from the cases this court has decided since Armcor.8 Judge Gibbons has already noted that it is “no secret” that a significant number of judges on this court have been signalling the Board that Gissel bargaining orders “are unwelcome in this circuit” while another group of judges is giving quite a different signal. NLRB v. K&K Gourmet Meats, Inc., 640 F.2d 460, 470-71 (3d Cir. 1981) (dissenting opinion) (comparing, inter alia, cases cited in note 8, supra). The Armcor line of decisions took the tack of refusing to enforce Gissel orders which recognized the union’s representation by concluding that the Board’s opinions were so opaque that a remand would be necessary. Like Judge Gibbons, I had hoped that the Board had hit upon a formula for stating its reasons satisfactorily. See K&K, 640 F.2d at 471-72. That the present unsophisticated case was voted in banc proves that my hopes were ill-founded.
A superficial reading of the opinions here suggests that the court is split over the application of a settled rule to a given state of facts. I doubt that many perceptive observers of this court will be deceived. The dispute here is not over application of a rule, but over the existence of an entirely different precept: whether the NLRB has the power to order an employer to bargain with a union that has not obtained representation status by an election because of an employer’s pervasive unfair labor practices. Although recognizing that Gissel is binding precedent, panels of this court have nonetheless tried to subvert it indirectly, focusing their denial of enforcement on the Board’s procedural failures. A rule that is manipulated to achieve desired results is not a “neutral principle,” and in my view the decisions applying it are not genuinely principled decisions because they do not rest on “analysis and reasons . . . transcending the immediate result . . .. ” Wechsler, Toward Neutral Principles of Constitutional Law, 73 Harv.L.Rev. 1, 15 (1959).
The panels and in banc courts will continue to divide as long as the opinions are cast as disputes over application of this dubious precept, settled though it may be. The precedential impact of cases at this level of dispute is severely limited by the facts presented. This in banc offering contains no assurance that panels will apply the rule more evenhandedly because each subsequent case will be factually distinguishable, and it may be that similar cases will be before us in banc (for who is to say that thrice is the limit stop of jural masochism?) as the particular labor-management philosophy of the randomly selected panel members continues to produce inconsistent results.9 In fine, the Armcor-Hedstrom rule *527permits a two-judge panel majority to foist on the Board its notions of our national labor policy in a given case or controversy, a practice directly contrary to the Congressional mandate which has lodged this responsibility in the NLRB.
III.
In addition to hostility to Gissel bargaining orders, I also detect in the Armcor-Hedstrom rule a distrust of the persons charged with making an initial determination of the necessity of a Gissel bargaining order, the administrative law judges. This view is confirmed by the Baerga-Hargenrader line of eases, in which this court has imposed increasingly rigorous factfinding requirements on Social Security Administration ALJs only when the decision of the ALJ is to deny social security disability benefits. See Cotter v. Harris, 642 F.2d 700, 708-712 (3d Cir. 1981) (Garth, J., dissenting). Such distrust, I suspect, stems from a belief that ALJs are second-class judges (if judges at all), marginally qualified, beholden to the employing agency, and, often, veterans of the agency whose cases they now decide. Many Article III judges, as well as scholarly critics and attorneys, consider the federal judiciary’s life tenure, employment by a separate branch, and other indicia of independence to be singular assurances of impartial judgment and easily scorn agency “employees” who perform judicial functions without the full investiture of Article III privileges.
Accepting for purposes of argument that to be impartial judges must be independent of all political or employment pressures, I submit that the view that the ALJs are not sufficiently independent or competent is now so shopworn as to be totally obsolete. To the contrary, ALJs, though not yet annointed with life tenure, enjoy an independence that in my view is plainly sufficient to satisfy reasonable doubts.10
For example, although ALJs are considered agency personnel, they are selected by the Office of Personnel Management (OPM) independently of agency recommendation or rating, 5 U.S.C. § 5362, and cannot be removed from office without a hearing establishing good cause before the Merit Systems Protection Board, 5 U.S.C. § 7521. Their pay is controlled by the Civil Service Commission. ALJs can be disqualified from a case only upon petition by either the agency or a private party. 5 U.S.C. § 556(b). Similarly, cases are assigned on a rotating basis so that the agency cannot “fix” the result by choice of judge. 5 U.S.C. § 3105. They are not subject to the whim of the agency. ALJs are strictly independent of investigative or prosecutorial personnel in the agency. 5 U.S.C. § 554(d).
Moreover, the selection process for ALJs should inspire more respect for this office than is generally extended by Article III judges; it is a process that requires rigorous inquiries into the background and competence of the candidates. Applicants must supply twenty professional references. A minimum of seven years of litigation experience is required to meet the threshold selection requirement. A test opinion must be drafted and evaluated on the basis of many factors including clarity and conciseness. See Office of Personnel Management, Announcement No. 318, Administrative Law Judge 13-16 (1979). Finally, after the various scores have been combined, applicants considered tentatively eligible are interviewed by a special panel usually composed of an OPM official, an attorney qualified in the field of administrative law, and an agency official. This committee submits a recommendation to the director of OPM who makes final eligibility determinations among qualified candidates. Once appointed to an agency, an ALJ is not subjected to the usual probationary employment period for agency employees, further insuring ALJ independence. 5 U.S.C. § 3321.
*528The Supreme Court recognized in Butz v. Economou, 438 U.S. 478, 98 S.Ct. 28947, 57 L.Ed.2d 895 (1978), that ALJs are “functionally comparable” to judges employed in the judicial branch and conferred on them absolute immunity for judicial acts. In doing so the Court observed:
There can be little doubt that the role of the modern federal hearing examiner or administrative law judge within this framework is “functionally comparable” to that of a judge. His powers are often, if not generally, comparable to those of a trial judge. . . . More importantly, the process of agency adjudication is currently structured so as to assure that the hearing examiner exercises his independent judgment on the evidence before him, free from pressures by the parties or other officials within the agency. Prior to the Administrative Procedure Act, there was considerable concern that persons hearing administrative cases at the trial level could not exercise independent judgment because they were required to perform prosecutorial and investigative functions as well as their judicial work . . . and because they were often subordinate to executive officials within the agency .... Since the securing of fair and competent hearing personnel was viewed as “the heart of formal administrative adjudication,” Final Report of the Attorney General’s Committee on Administrative Procedure 46 (1941), the Administrative Procedure Act contains a number of provisions designed to guarantee the independence of hearing examiners.
438 U.S. 513-14, 98 S.Ct. at 2914 (citations omitted).
The rigors of the selection procedure and the statutory protections of ALJ independence suggest to me that the federal judiciary need not look down its collective nose at ALJ decisions. Cf. Eastern Engineering & Elevator Co., Inc. v. NLRB, 637 F.2d 191 (3d Cir. 1980) (NLRB must respect special competence of ALJ who has the benefit of first-hand observation of witnesses). I am not proposing that the judiciary abandon all distinctions between ALJs and Article III judges when reviewing decisions, for Congress has clearly intended that review be on a different level. Of course, in the review of the choice and application of legal precepts, a judicial function dramatically different from reviewing factfinding, I have not been loath to criticize either the ALJ or the agency. See Allegheny General Hospital v. NLRB, 608 F.2d 965, 967-69 (3d Cir. 1979). But I perceive that under the legislative schema, at least in the context of factfinding, the difference may well be more tenuous than our decisions in the Gissel and Hargenrader context indicate.
IV.
I concur in the judgment enforcing the bargaining order in this case, but I reject the standard applied to the Board’s decision. I reject the standard because it represents unwarranted interference with choices committed by Congress to the agency and thereby impedes the development of a coherent national labor policy by the NLRB. I also reject the standard because it is a rule applied to achieve results not based on the reasons given. I am convinced that as long as the court continues to anatomize the Board’s reasons for imposing a bargaining order, these cases will continue to be self-inflicted wounds that will drain our energies; energies that should be husbanded, not dissipated.
GARTH, Circuit Judge, concurring in part and dissenting in part, with whom JAMES HUNTER, III, and WEIS, Circuit Judges, join:I.
I agree with the majority that there is substantial evidence to support the Board’s findings of unfair labor practices. I also agree with the majority’s reaffirmance of our precedents which require that the Board articulate its reasons for imposing a bargaining order. Maj. op. typescript at 11; Hedstrom Co. v. NLRB, 629 F.2d 305 (3d Cir. 1980) (en banc) (Hedstrom II); NLRB v. Armcor Indus., Inc., 535 F.2d 239 (3d Cir. 1976). I do not agree, however, that either the ALJ or the Board complied with our *529articulation requirement and thus I would deny enforcement of the bargaining order.
II.
Both the majority and I recognize this court’s insistence that the Board explain why it imposed a bargaining order rather than ordering a re-run election. In this respect we are all aware that the Supreme Court has admonished the Board that elections are the preferred way to ascertain whether a union has majority support. NLRB v. Gissel Packing Co., 395 U.S. 575, 602, 89 S.Ct. 1918, 1934, 23 L.Ed.2d 547 (1969). The Second Circuit has summarized the law governing the Board’s choice of remedies for employer unfair labor practices committed during a representation election as follows:
Where an employer’s misconduct taints a prior union election by adversely affecting the employees’ freedom of choice, the traditional remedy, frequently characterized as the “preferred” or “superior” remedy, see NLRB v. Jamaica Towing, Inc., 602 F.2d 1100, 1104 (2d Cir. 1979); Donn Products, Inc. v. NLRB, 613 F.2d 162, 165 (6th Cir. 1980); Peerless of America, Inc. v. NLRB, 484 F.2d 1108, 1122 (7th Cir. 1973), has been to (1) vacate the election, (2) enjoin the employer from engaging in such misbehavior, (3) require him to post “contrition” notices to his employees, disavowing any future interference, and (4) direct him to give union representatives reasonable access to the employees. This is then followed by a new Board-supervised election. The issuance by the Board of a bargaining order in lieu of a cease- and-desist order is only proper if, after an objective review of all of the relevant surrounding circumstances, including the nature of the employer’s misbehavior and any later events bearing on its impact on the employees, it may reasonably be concluded that the employees will be unable to exercise a free choice in an Board-supervised rerun election.
Thus, to deny an election, and by so doing, to disenfranchise the employees, the circumstances must be egregious, and the reasons for the denial must be explained.1 The only difference between the majority’s opinion and this opinion is that whereas the majority is satisfied with the Board’s “explanation”, I am not. In order to explain my position respecting the absence of explanation and articulation by the Board in this case, a brief review of the context in which the problem arises is necessary.
Any discussion of bargaining orders must start with Gissel which established two categories of cases in which bargaining orders may be imposed. I agree with the majority that here we are concerned with a Gissel II case. See maj. op., typescript at 11, n.l. In Gissel II cases — “less extraordinary cases marked by less pervasive practices which nevertheless still have the tendency to undermine majority strength and impede the election process”, NLRB v. Gissel Packing Co., 395 U.S. 575, 614, 89 S.Ct. 1918, 1940, 23 L.Ed.2d 547 (1969) — it first must be shown *530that at one point the union had a majority. The Board then should consider:
the extensiveness of an employer’s unfair practices in terms of their past effect on election conditions and the likelihood of their recurrence in the future. If the Board finds that the possibility of erasing the effects of past practices and of ensuring a fair election (or a fair rerun) by the use of traditional remedies, though present, is slight and that employee sentiment once expressed through cards would, on balance, be better protected by a bargaining order, then such an order should issue . . .
Id.
We have implemented this teaching by persistently requiring that the NLRB explicate the basis for the imposition of a bargaining order. See Kenworth Trucks v. NLRB, 580 F.2d 55, 59-60 (3d Cir. 1978); NLRB v. Craw, 565 F.2d 1267, 1271-72 (3d Cir. 1977); NLRB v. Eagle Material Handling, Inc., 558 F.2d 160, 166-68 (3d Cir. 1977); Hedstrom Co. v. NLRB, 558 F.2d 1137 (3d Cir. 1977) (Hedstrom I); NLRB v. Armcor Indus., Inc., 535 F.2d 239 (3d Cir. 1976). This requirement aids our review of a petition for enforcement by helping to ensure that the Board has given full consideration to the preferred alternative of a re-run election. See Armcor, supra at 245.
In the en banc decision of Hedstrom II, we wrote:
The primary purpose of this requirement is to assist a reviewing court in determining whether the standards of Gissel have been satisfied. Since the propriety of a bargaining order depends upon the existence of certain circumstances, it is fitting for the board to “explain with specificity the results of the unfair labor practices and, in particular, the unlikelihood of a fair election,” NLRB v. Craw, 565 F.2d 1267, 1272 (3d Cir. 1977), before seeking enforcement of such an order. This is not to say, of course, that a reasoned elaboration of the basis for a bargaining order by itself will assure that a careful weighing of the evidence has occurred, but only that it should help to further such an aim.
Id. at 309 (footnote omitted).
Our court held in Hedstrom II that the Board must provide a reasoned analysis, setting forth those factors justifying the imposition of a bargaining order. The analysis need not be elaborate, but the Board must:
“estimate the impact [of the unfair labor practices], taking into account the factors in the particular case which are indicative of actual effect or which plausibly, in the light of existing knowledge, would contribute to or detract from an actual impact,” and appraise “those factors which might reasonably have a bearing” on the likelihood of a fair rerun election.
Hedstrom II at 309 quoting Peerless of America, Inc. v. NLRB, 484 F.2d 1108, 1118 n. 16 (7th Cir. 1973). Peerless was also quoted in Kenworth Trucks of Philadelphia v. NLRB, 580 F.2d at 60, NLRB v. Craw, 565 F.2d at 1271, and NLRB v. Armcor Industries, Inc., 535 F.2d at 245.
Hedstrom I had resulted in a remand to the Board because no such analysis, articulation or explanation had been made by the Board in justifying the imposition of a bargaining order. Thereafter, in Hedstrom II, the Board, complying with our directive, explained that a bargaining order was necessary and that a rerun election could not fairly be ordered. On the record of that case the Board found that:
the total effect of the threats of closure, and the numerous other unfair labor practices against the background of the general awareness of the Fitchburg experience, was to instill in employees a strong fear of loss of employment that would continue to be operative even in the event of a second election. This fear could only have been exacerbated by the publication of the newspaper editorial.
We find that “the possibility of erasing the effects of past practices and of ensuring a fair election (or a fair rerun) by the use of traditional remedies, though present, is slight and that employee sentiment once expressed through cards *531would, on balance, be better protected by a bargaining order. 11 This conclusion is not significantly affected by the reviewing court’s reversal of our finding that President Ketcham violated Section 8(a)(1) by threatening to fire an employee. The reversal does not disturb any of the findings concerning the threats of plant closure, or the overall context in which those threats occurred, which form the basis for our conclusion that a bargaining order is warranted.
(Appendix Hedstrom II at 10a) (footnotes 10 and 12 omitted).
In further compliance with our requirements, the Board also reasoned and explained that the passage of time had not increased the prospects of a fair election:
the effect of Respondent’s threats of plant closure and its extensive unfair labor practices during the course of the campaign was to instill in employees the fear that they would lose their jobs through plant closure if the Union won the election. There is nothing to indicate that this fear has been erased during the past 3 years. To the contrary, recent events indicate that Respondent’s vigorous opposition to the Union continues to result in unfair labor practices, thereby sustaining, reinforcing, or increasing the fear that the Company might still carry out its earlier threat to close operations if the Union prevails.
(Appendix lla-12a). Finally, the Board reexamined our reversal of the Section 8(a)(5) violation and concluded that even absent that violation, a bargaining order was necessary.
A majority of this court held that the Board’s explanation satisfied our mandate of a “reasoned analysis.”
Our court is not alone in requiring the Board to explain its reasons for rejecting a re-run election in favor of imposing a bargaining order.2 The Seventh Circuit, whose opinion in Peerless of America Inc. v. NLRB, 484 F.2d 1108 (7th Cir. 1973) was relied upon by this court in Hedstrom II, Craw and Armcor, has again forcefully reiterated an identical instruction to the Board in Red Oaks Nursing Home v. NLRB, 633 F.2d 503 (7th Cir. 1980). The court in that case found the Board’s reasons for imposing a bargaining order insufficient. The ALJ, in an opinion adopted by the Board, had stated:
I do not, however, consider the unfair labor practices committed to be minor in view of their nature, number, and extent of probable impact throughout the bargaining unit. In my opinion, neither the factors recited above nor the Board’s traditional remedies for Section 8(a)(a) violations render it likely that another election would result in a truer measure of uninhibited employee choice than the authorization cards executed on or before August 22, 1977. I therefore in all the circumstances of this case find that the Respondent’s unfair labor practices have undermined the Union’s majority and made the holding of a fair election most unlikely, and recommend that the petition in Case 25-RC-6717 be dismissed and the Respondent be ordered to bargain on request.
Quoted in id. at 509.
In denying enforcement of the Board’s order, the Seventh Circuit wrote:
the opinion of the ALJ contains no consideration of the effectiveness of the ordinary remedies, the continuing impact of the unfair labor practices on the election process, and the probability of repeated violations, which are also prerequisites for a bargaining order. See Peerless, supra, 484 F.2d at 1118. By not stating separately why the circumstances of this case require a bargaining order, the *532Board appears in effect to be automatically issuing bargaining orders on the basis of virtually any unfair labor practice, a policy clearly not in accordance with Gissel. See NLRB v. Appletree Chevrolet, Inc., 608 F.2d 988, 998, 103 LRRM 2066 (4th Cir. 1979).
Id. at 509-510 (footnote omitted, emphasis in original).
It is thus evident that this Circuit is not the only circuit that requires the Board to provide an adequate articulation of reasons before it issues a bargaining order.
Requiring the Board to articulate its reasons for imposing a bargaining order does not represent an unwarranted judicial interference with administrative procedure. Indeed that requirement, as we have pointed out, stems from the Supreme Court’s instructions in Gissel, and from fundamental rules of administrative law. See, e. g., SEC v. Chenery Corp., 332 U.S. 194, 196-97, 67 S.Ct. 1575, 1577, 91 L.Ed. 1995 (1947); 5 U.S.C. § 557(c)(3)(A) (1976). The Supreme Court has consistently instructed that
[i]f the administrative action is to be tested by the basis upon which it purports to rest, that basis must be set forth with such clarity as to be understandable. It will not do for a court to be compelled to guess at the theory underlying the agency’s action; nor can a court be expected to chisel that which must be precise from what the agency has left vague and indecisive. In other words, “We must know what a decision means before the duty becomes ours to say whether it is right or wrong.” United States v. Chicago, M., St.P.& P.R. Co., 294 U.S. 499, 511, 55 S.Ct. 462, 467, 79 L.Ed. 1023.
Chenery at 196, 67 S.Ct. at 1577.
This basic requirement which focuses on effective judicial review cannot be deemed to constitute an undue burden on the Board, and does not intrude upon internal Board procedures in a manner proscribed by the principle expressed in Vermont Yankee Nuclear Power Corp. v. Natural Resources Defense Council, Inc., 435 U.S. 519, 98 S.Ct. 1197, 55 L.Ed.2d 460 (1978). See Kenworth Trucks v. NLRB, 580 F.2d 55, 61-63 (1978). Our requirement of reasoned articulation does no more than mandate that the Board inform us of the basis upon which its decision is predicated. “When the Board . . . exercises the discretion given to it by Congress, it must ‘disclose the basis of its order’ and ‘give clear indication that it has exercised the discretion with which Congress has empowered it.’ Phelps Dodge Corp. v. Labor Board, 313 U.S. 177, 197 [61 S.Ct. 845, 853, 85 L.Ed. 1271].” NLRB v. Metropolitan Insurance Co., 380 U.S. 438, 443, 85 S.Ct. 1061, 1064, 15 L.Ed.2d 951 (1965). The obligation of reasoned analysis is essential to perform our reviewing function and to guarantee the integrity of the administrative process.
III.
I cannot agree with the majority that the opinion of the administrative law judge adopted by the Board in this case, satisfies the standard set forth in Hedstrom II. When contrasted with the analysis and articulation of the Board in Hedstrom II, it is quite apparent that the conclusory discussion of the administrative law judge in this case is woefully deficient. It is even less detailed than the analysis rejected in Red Oaks. Virtually the entire thrust of the ALJ’s discussion here was focused not on the reasons why a fair re-election was impossible, but rather on the reasons why the initial election of December 30, 1977 had to be set aside. After summarizing the unfair labor practices found, he stated:
In sum, Respondent sought to impress upon the employees the futility of voting the Union in, and that there would be harsher dealing and reprisal if they did. The mass of violations of Section 8(a)(1), (3), and (4) of the Act were egregious unfair labor practices, which had the tendency to undermine, and undermined, the Union’s majority strength in the election of December 30,1977, and prevented the holding of a fair election.
By refusing to recognize and bargain with the Union, as requested on November 11,1977, and instead engaging in this course of unlawful conduct which under*533mined the Union’s majority status and prevented the holding of a fair election, Respondent also violated Section 8(a)(5) of the Act, Trading Port, Inc., 219 NLRB 298, 300-301 (1975).
Under the circumstances of this case, the sentiment of the unit employees expressed through the Union authorization cards is a more reliable measure of their desires on the issue of representation than the election held December 30, 1977. To remedy Respondent’s unfair labor practices, including its Section 8(a)(5) refusal to bargain, a bargaining order is necessary, whether the violations be viewed as category one unfair labor practices under Gissel, supra, see J.P. Stevens & Co., Inc., Gulistan Division v. N.L.R.B., 441 F.2d 514, 521-522 (C.A. 5, 1971), cert. denied 404 U.S. 830 [92 S.Ct. 69, 30 L.Ed.2d 59], or as category two unfair labor practices, see N.L.R.B. v. Kaiser Agricultural Chemicals etc., 473 F.2d 374, 382-383 (C.A. 5, 1973). The bargaining order will be issued as of the end of October 1977, when Respondent embarked on its course of unlawful conduct.
The reasoning of the ALJ was that since unfair labor practices had prevented the holding of a fair election and had undermined the union’s majority strength, a bargaining order was necessary. At no time did he satisfy the Hedstrom II obligation to estimate the impact of the various violations and appraise “those factors which might reasonably have a bearing on the likelihood of a fair re-run election.” Hedstrom II at 309. The ALJ never discussed in this section of his opinion why “the possibility of erasing the effects of past practices and of ensuring a fair election (or a fair rerun) by the use of traditional remedies, though present, is slight ...” Gissel 395 U.S. at 614, 89 S.Ct. at 1940. Only in the conclusions of law section of his opinion does the ALJ state, unsupported by any articulation of reasons, that “these [pre election] unfair labor practices cannot be corrected by conventional remedies, including a rerun election.” App. at 50. He failed completely to address, let alone answer, the crucial inquiry whether there was a likelihood that the unfair labor practices would recur, a consideration that Gissel indicates should be taken into account. Gissel at 614, 89 S.Ct. at 1940.
Moreover, the ALJ and the Board, in holding that a bargaining order was necessary under Gissel II, did not distinguish between pre- and post-card majority violations. See Rapid Mfg. Co. v. NLRB, 612 F.2d 144 (3d Cir. 1979). I agree with the majority that unfair labor practices that occur before a card majority is attained, may, under certain circumstances, support the imposition of a Gissel II bargaining order. Yet in my opinion, an inference is present that pre-card (as contrasted with post-card) majority unfair labor practices do not significantly undermine the union’s majority strength. This inference can be drawn because in Gissel II cases, where a majority is required, the union obviously gains adherents even after the occurrence of pre-card majority unlawful conduct. Normally such pre-majority violations would provide less support for the imposition of a bargaining order than would post-majority violations. Accordingly, when finding that the possibility for a fair re-run election is slight, the Board, if it relies on pre-card majority unfair labor practices, must carry a heavier burden in explaining why this conduct supports its imposition of a bargaining order. No such burden has been satisfied here by the Board. Indeed the ALJ and the Board failed even to distinguish between pre- and post-card violations. Thus I cannot agree with the majority that the AU’s opinion, which the Board adopted, satisfies the Gissel or Hedstrom II requirements.3
*534I do not believe that it is incumbent upon the members of this court to point out constantly and repetitively those factors which the Board must consider before it concludes that a fair re-run election cannot be held and that a bargaining order must be imposed. The Board, not the court, has the statutory burden of determining which remedy is appropriate. 29 U.S.C. § 160(c) (1976). See Gissel at 612 n. 32, 89 S.Ct. at 1939.4 Certainly this court alone has furnished more than ample guidance to the Board, as to what is expected in meeting this burden.
At the least, the Board in its review of the ALJ’s findings, should have considered: the closeness of the vote; the extent to which the impact of the unfair labor practices persisted in the minds of the employees and whether this impact likely would have tainted a future election; the change in the composition of the workforce; the status of the Mennen contract; the weight to be given pre-card majority violations; the effect of the passage of time; the probability of repeated violations by Permanent Label; and the effectiveness of preferred traditional remedies. There are undoubtedly other factors and facts that the Board should have considered, but since this court has repeatedly instructed the Board in various contexts as to what constitutes a proper analysis, see Hedstrom II at 311-12; Rapid Mfg. Co., supra; NLRB v. Garry Mfg. Co., 630 F.2d 934, 946 (3d Cir. 1980) (Weis, J., dissenting); Electrical Prods. Div. of Midland-Ross Corp. v. NLRB, 617 F.2d 977, 990 (3d Cir.) cert. denied, 449 U.S. 871, 101 S.Ct. 210, 66 L.Ed.2d 91 (1980) (Weis, J., dissenting), I am loathe to once again tread the same ground.
I do not believe that we as a court should be forced to infer from the ALJ’s opinion, which lacks a reasoned articulation, what the ALJ intended to explain to justify the bargaining order. The majority in its interpretation of the ALJ’s opinion, maj. op., at 520, has for the first time supplied the required articulation of reasons which neither the ALJ nor the Board furnished. It is possible that these are the same reasons which the Board would have articulated in explaining its denial of a re-run election. But then again, the possibility exists that the Board’s explanation may have differed from that now provided by the majority. It is for that very reason that the Supreme Court emphasized the following:
simple but fundamental rule of administrative law . . . that a reviewing court, in dealing with a determination or judgment which an administrative agency alone is authorized to make, must judge the propriety of such action solely by the *535grounds invoked by the agency. If those grounds are inadequate or improper, the court is powerless to affirm the administrative action by substituting what it considers to be a more adequate or proper basis. To do so would propel the court into the domain which Congress has set aside exclusively for the administrative agency.
Securities and Exchange Commission v. Chenery, 332 U.S. 194, 196, 67 S.Ct. 1575, 1577, 91 L.Ed. 1995 (1947).
Because in this case the Board has once again failed to comply with our oft-repeated instructions, I would deny enforcement to its order so that a re-run election could immediately be ordered5 or at the very least, I would remand for the explanation which this court requires. Thus, to the extent that the majority would give effect to the bargaining order, I respectfully dissent from the majority’s opinion.
. The Board may satisfy this requirement by specifically adopting the ALJ’s findings and reasoning without separate analysis. See Kenworth Trucks v. NLRB, 580 F.2d 55, 62-63 (3d Cir. 1978).
. Although 1 refer to the precept applied in this case as a rule, it fails to meet the criteria to deserve the name. Strictly speaking, a rule is a “precept attaching a definite, detailed legal consequence to a definite, detailed set of facts.” Pound, Hierarchy of Sources and Forms in Different Systems of Law 7 Tul.L.Rev. 475, 482 (1933). See United States v. Criden, 633 F.2d 346, 354 n.4 (3d Cir. 1980), cert. denied, 449 U.S. 1113, 101 S.Ct. 924, 66 L.Ed.2d 842 (1981). The hallmark of a rule is its predictable application. As discussed below, application of this rule yields anything but predictable results.
. Denying enforcement of a Gissel order by an equally divided court.
. I joined the court’s opinion in Hedstrom II, but I am now convinced that the full court erred in not administering the last rites to the Armcor innovation when we had the chance to do so then. In changing my view, I am mindful that “[w]isdom too often never comes, and so one ought not to reject it merely because it comes late.” Henslee v. Union Planters National Bank & Trust Co., 335 U.S. 595, 600, 69 S.Ct. 290, 293, 93 L.Ed. 259 (1949) (Frankfurter, J., dissenting). See also R. Aldisert, The Judicial Process 861 (1976).
. I attempted recently to clarify this impenetrable formulation of a standard. See Smith v. Harris, 644 F.2d 985, 990 (3d Cir. 1981) (concurring opinion).
. “[C]ourts may not accept appellate counsel’s post hoc rationalizations for agency action . . . . ” Burlington Truck Lines v. United States, [371 U.S. 156, 168, 83 S.Ct. 239, 245, 9 L.Ed.2d 207 (1962)]; see Securities & Exchange Comm’n v. Chenery Corp., 332 U.S. 194, 196, 67 S.Ct. 1575, 1577, 91 L.Ed. 1995 [1947], For reviewing courts to substitute counsel’s rationale or their discretion for that of the Board is incompatible with the orderly function of the process of judicial review. Such action would not vindicate, but would deprecate the administrative process for it would “propel the court into the domain which Congress has set aside exclusively for the administrative agency.” 380 U.S. at 444, 85 S.Ct. at 1064 (citation omitted).
. Justice Frankfurter, speaking for the Court in SEC v. Chenery, 318 U.S. 80, 95, 63 S.Ct. 454, 462, 87 L.Ed. 626 (1943) (Chenery I), indicated that the “articulation requirement” is quite minimal:
In finding that the Commission’s order cannot be sustained, we are not imposing any trammels on its powers. We are not enforcing formal requirements. We are not suggesting that the Commission must justify its exercise of administrative discretion in any particular manner or with artistic refinement. We are not sticking in the bark of words. We merely hold that an administrative order cannot be upheld unless the grounds upon which the agency acted in exercising its powers were those upon which its actions can be sustained.
There the considerations urged in support of the SEC’s order were not those on which the SEC relied. Thus the holding in Chenery I, reiterated in Chenery II, was nothing more than the familiar requirement that an administrative order cannot be sustained if based on a misperception of the law, even if it might be sustained on other grounds not articulated by the agency. To hold otherwise would allow a court to substitute its discretion for that of the agency, confusing administrative review with the kind of review applied to district court decisions.
. Compare NLRB v. K&K Gourmet Meats, Inc., 640 F.2d 460 (3d Cir. 1981); NLRB v. Craw, 565 F.2d 1267 (3d Cir. 1977); Hedstrom Co. v. NLRB, 558 F.2d 1137, 1148 (3d Cir. 1977) (Hedstrom 1); with NLRB v. Garry Mfg. Co., 630 F.2d 934 (3d Cir. 1980); Hedstrom Co. v. NLRB, 629 F.2d 305 (3d Cir. 1980) (in banc) (Hedstrom 17); NLRB v. Daybreak Lodge Nursing and Convalescent Home, Inc., 585 F.2d 79 (3d Cir. 1978); Kenworth Trucks of Philadelphia v. NLRB, 580 F.2d 55 (3d Cir. 1978).
. As a dispute over application of law to facts, this case does not meet our usual requirements for rehearing in banc. This court’s Internal Operating Procedure VIII B. provides: “This court does not ordinarily grant rehearing in banc where the panel’s statement of the law is correct and the controverted issue is solely the application of the law to the circumstances of the case.”
. See generally Lubbers, Federal Administrative Law Judges: A Focus on our Invisible Judiciary, 33 Admin.L.Rev. 109 (1981); Admin. Conference of the United States, Federal Administrative Law Judge Hearings 7-20 (1980). Cf. Nash v. Califano, 613 F.2d 10 (2d Cir. 1980) (ALJ has standing to challenge attempted interference with decisional independence by Social Security Administration).
. Judge Weis has expressed his views as to why bargaining orders should not be issued in an ordinary case:
Routine issuance of bargaining orders after a review of the facts and a conclusory finding that the unfair labor practices were “egregious,” “pervasive” or “chilling” is simply unacceptable. Bargaining orders have been defended as a necessary sanction to deter misconduct by an overreaching employer, but such orders do not affect the employer alone — they also fall on employees. It is they who lose the right to vote in favor of one union or another or none at all — rights that the National Labor Relations Act guarantees to them. Moreover, the Board’s remedies should right the wrong — not act as punitive measures. Republic Steel Corp. v. NLRB, 311 U.S. 7, 61 S.Ct. 77, 85 L.Ed. 6 (1940).
In imposing bargaining orders, the Board is subordinating the employee’s right to express his choice to its forecast as to who would have won the election if the employer had maintained laboratory conditions. When that prediction is little more than a guess, however, the Board’s paternalism does not justify depriving the employees of their vote. The extreme response of a bargaining order should be the exception and not the norm — a principle recognized by this court in Armcor and Hedstrom.
Electric Prods. Div. of Midland-Ross v. NLRB, 617 F.2d 977, 991 (3d Cir.), cert. denied, 449 U.S. 871, 101 S.Ct. 210, 66 L.Ed.2d 91 (1980) (Weis, J., dissenting) (footnotes omitted).
N.L.R.B. v. Gissel Packing Co., 395 U.S. at 614-615, 89 S.Ct. at 1940.
. See, e. g., NLRB v. Jamaica Towing, Inc., 632 F.2d 208, 211 (2d Cir. 1980); Chromalloy Mining and Minerals v. NLRB, 620 F.2d 1120, 1129-30 (5th Cir. 1980); NLRB v. Appletree Chevrolet, Inc., 608 F.2d 988, 998 (4th Cir. 1979); NLRB v. Pilgrim Foods, Inc., 591 F.2d 110, 119 (1st Cir. 1978); NLRB v. Pacific Southwest Airlines, 550 F.2d 1148, 1152 (9th Cir. 1977); Peerless of America, Inc. v. NLRB, 484 F.2d 1108 (7th Cir. 1973).
. I recognize that the “articulation” of the administrative law judge which was adopted by the Board in Kenworth Trucks v. NLRB, 580 F.2d 55 (3d Cir. 1978) could be equated in some respects to the articulation of the Board which we reject in this case. But I am satisfied that Kenworth does not compel the enforcement of the Board’s order here. Nowhere in Kenworth did the Court address the sufficiency of the ALJ’s explanation. Instead the Kenworth panel, as Chief Judge Seitz’s opinion recognizes, maj. op., at 521 n.2, focused on whether the *534Board itself, as distinct from the ALJ, was required to analyze and articulate the reasons for the issuance of a bargaining order. After first holding that the Board had an independent obligation to formulate the reasons for issuing a bargaining order, the court on rehearing held that a separate articulation was not necessary: the Board could simply adopt the ALJ’s explanation. After having decided the central issue before it, the court then proceeded to enforce the Board’s order, without assessing the sufficiency of the ALJ’s articulation of reasons.
On its face, the discussion implicitly accepted by the court in Kenworth is not dissimilar to the conclusory discussion found in this case. However, as I have pointed out, the court in Kenworth did not direct its attention to the adequacy of the ALJ’s analysis. Thus, Ken-worth may not be construed as holding that a conclusory analysis and conclusory articulation of reasons for the imposition of a bargaining order is all that is required. Even if Kenworth was construed as standing for such a proposition, our decision in Hedstrom II filed some two years after Kenworth and issuing from an en banc court clearly demonstrates that the type of conclusory statement found in Kenworth may not be relied upon to support a bargaining order.
Hence, whatever vitality may have once been attributed to such a construction of Kenworth, has been thoroughly undermined by Hedstrom II, and cannot fairly be read as diluting the Hedstrom II standard reaffirmed today.
. Cf. Red Oaks Nursing Home v. NLRB, 633 F.2d 503, 509 (7th Cir. 1980) (“typically, this court has reached the merits of the propriety of the order, in part to avoid further compromise of employee rights inherent in the delay of remand.... Thus, the effect of the Board’s continuing neglect to fashion standards for the exercise of its discretion is to place exclusive reliance on judicially prescribed standards for determining the need for bargaining orders, a practice not contemplated by section 10(c) of the Act.”)
. The longer a re-run election is delayed the less likely the union’s chances of reversing the outcome of the first election. Gissel at 611 n. 30, 89 S.Ct. at 1938. A remand also raises the possible need for further judicial review, and hence further delay.