dissenting.*
I dissent because I believe that the dispositions in both No. 94-8988, post, p. 193, and No. 94-9323, ante, p. 163, are improper extensions of our limited power to vacate without first finding error below.
It sometimes occurs that, after having considered the lower court decision and found error, an appellate court merely reverses or vacates and then remands — that is, it sets the judgment aside and sends the case back to the lower court for further proceedings, rather than entering or directing entry of judgment for the appellant or petitioner. That is the appropriate course whenever the finding of error does not automatically entitle the appellant or petitioner to judgment, and the appellate court cannot conduct (or chooses not to conduct) the further inquiry necessary to resolve the ques*178tions remaining in the litigation. Our books are full of such cases, from Glass v. Sloop Betsey, 3 Dall. 6 (1794), and Clarke v. Russel, 3 Dall. 415 (1799), to Vernonia School Dist. 47J v. Acton, 515 U. S. 646 (1995), and Tuggle v. Netherland, ante, p. 10.
What is at issue here, however, is a different sort of creature, which might be called “no-fault V&R”: vacation of a judgment and remand without any determination of error in the judgment below. In our discretionary certiorari system of review, such an order has acquired the acronym “GVR”— for the Court grants certiorari, vacates the judgment below, and remands for further proceedings.1 The question presented by today’s cases is whether there is any limitation (other than the mandate “do what is fair”) upon this practice. The Court’s per curiam opinions answer “no”; I disagree.
Title 28 U. S. C. § 2106 provides that “[t]he Supreme Court or any other court of appellate jurisdiction may affirm, modify, vacate, set aside or reverse any judgment, decree, or order of a court lawfully brought before it for review, and may remand the cause and direct the entry of such appropriate judgment, decree, or order, or require such further proceedings to be had as may be just under the circumstances.” This facially unlimited statutory text is subject to the implicit limitations imposed by traditional practice and by the nature of the appellate system created by the Constitution and laws of the United States. The inferior federal courts (to say nothing of state courts) are not the creatures *179and agents of this body — as are masters, whose work we may reject and send back for redoing at our own pleasure. Inferior courts are separately authorized in the Constitution, see Art. I, § 8; Art. Ill, § 1, created by Acts of Congress, see, e. g., Judiciary Act of 1789,1 Stat. 73; Evarts Act, Act of Mar. 3, 1891, 26 Stat. 826, and staffed by judges whose manner of appointment and tenure of office are the same as our own, see U. S. Const., Art. II, §2; Art. Ill, §1; 28 U. S. C. §§44, 133, 134. Despite the unqualified language of § 2106, we cannot, for example, “reverse” a judgment of one of these courts “and direct the entry” of a different judgment whenever we disagree with what has been done, but only when we can identify a controlling error of law. And I think we cannot “vacate” and “remand” in the circumstances here.
The Court today seeks to portray our no-fault V&R practice as traditionally covering a kaleidoscopic diversity of situations. See Lawrence v. Chater, ante, at 166-167. That is in my view a misportrayal; the practice has always been limited to a few discrete categories of cases. It began, apparently, in situations calling forth the special deference owed to state law and state courts in our system of federalism. In Missouri ex rel. Wabash R. Co. v. Public Serv. Comm’n, 273 U. S. 126 (1927), for example, rather than find error on the basis of the federal constitutional claims raised, this Court set aside the judgment of the Missouri Supreme Court and remanded the case to that court for further proceedings so that it could consider the meaning and effect of a state statute that had been enacted after its judgment had been entered. We reasoned that “[wjhile this Court may decide these [state-law] questions, it is not obliged to do so, and in view of their nature, we deem it appropriate to refer the determination to the state court.” Id., at 131. In other words, we left it to the state court to decide the effect of the intervening event, rather than follow our usual practice of deciding that question for ourselves, see, e. g., Steamship Co. v. Joliffe, 2 Wall. 450, 456-458 (1865). See generally United *180States v. Schooner Peggy, 1 Cranch 103, 110 (1801) (“If, subsequent to the judgment [entered by a lower court], and before the decision of the appellate court, a law intervenes and positively changes the rule which governs, the law must be obeyed, or its obligation denied.”). Later cases took the same deferential approach to state courts when the intervening event consisted of one of our own opinions. See, e. g., State Tax Comm’n v. Van Cott, 306 U. S. 511 (1939). By 1945, we could state that it was “[a] customary procedure” for the Court “to vacate the judgment of [a] state court where there has been a supervening event since its rendition which alters the basis upon which the judgment rests, and to remand the case so that the court from which it came might reconsider the question in light of the changed circumstances.” State Farm Mut. Automobile Ins. Co. v. Duel, 324 U. S. 154, 161 (1945). Similarly, where a federal court of appeals’ decision on a point of state law had been cast in doubt by an intervening state supreme court decision, it became our practice to vacate and remand so that the question could be decided by judges “familiar with the intricacies and trends of local law and practice.” Huddleston v. Dwyer, 322 U. S. 232, 237 (1944).
The “intervening event” branch of our no-fault V&R practice has been extended to the seemingly analogous situation (though not one implicating the special needs of federalism) in which an intervening event (ordinarily a postjudgment decision of this Court) has cast doubt on the judgment rendered by a lower federal court or a state court concerning a federal question. See, e. g., Amer v. Superior Court of Cal., County of Los Angeles, 334 U. S. 813 (1948); Goldbaum v. United States, 348 U. S. 905 (1955); Henry v. City of Rock Hill, 376 U. S. 776 (1964). This is undoubtedly the largest category of “GVRs” that now exists. See, e. g., Exxon Corp. v. Youell, post, p. 801; Kapoor v. United States, post, p. 801; Edmond v. United States, post, p. 802; Pacesetter Constr. Co. v. Carpenters 16 Northern Cal. Ctys. Conference Bd., post, p. 802; Doctor’s Associates, Inc. v. Casarotto, 515 U. S. 1129 *181(1995); Calamia v. Singletary, 514 U. S. 1124 (1995). We regularly hold cases that involve the same issue as a case on which certiorari has been granted and plenary review is being conducted in order that (if appropriate) they may be “GVR’d” when the case is decided. More recently, we have indulged in the practice of vacating and remanding in light of a decision of ours that preceded the judgment in question, but by so little time that the lower court might have been unaware of it. See, e. g., Grier v. United States, 419 U. S. 989 (1974). These applications of no-fault V&R have nothing to do with federalism, but they are appropriate to preserve the operational premise of a multitiered judicial system (viz., that lower courts will have the first opportunity to apply the governing law to the facts) and to avoid the unseemliness of holding judgments to be in error on the basis of law that did not exist when the judgments were rendered below. They thus serve the interests of efficiency and of concern for the dignity of state and lower federal tribunals.
An entirely separate branch of our no-fault V&R jurisprudence, but again one that originates in the special needs of federalism, pertains to decisions of state supreme courts that are ambiguous as to whether they rest on state-law or federal-law grounds. Rather than run the risk of improperly reversing a judgment based on state law, we adopted the practice of vacating and remanding so that the state court could make the reasons for its judgment clear. See, e. g., Minnesota v. National Tea Co., 309 U. S. 551 (1940); Department of Mental Hygiene of Cal. v. Kirchner, 380 U. S. 194 (1965).2
*182We have GVR’d with increasing frequency in recent years on the basis of suggestions or representations made by the Solicitor General. Some of these cases are nothing more than examples of the “intervening-event GVR” discussed above, the Solicitor General pointing out that a case or statute has intervened since the judgment below. See, e. g., Woods v. Durr, 336 U. S. 941 (1949); Altiere v. United States, 382 U. S. 367 (1966). We have also announced no-fault GVR’s, however, when there has been no intervening development other than the Solicitor General’s confession of error in the judgment. That is a relatively new practice. As recently as 1942 a unanimous Court (two Justices not participating) wrote the following:
“The public trust reposed in the law enforcement officers of the Government requires that they be quick to confess error when, in their opinion, a miscarriage of justice may result from their remaining silent. But such a confession does not relieve this Court of the performance of the judicial function. The considered judgment of the law enforcement officers that reversible error has been committed is entitled to great weight, but our judicial obligations compel us to examine independently the errors confessed. . . . Furthermore, our judgments are precedents, and the proper administration of the criminal law cannot be left merely to the stipulation of the parties. . . .” Young v. United States, 315 U. S. 257, 258-259 (1942).
Cf. U. S. Bancorp Mortgage Co. v. Bonner Mall Partnership, 513 U. S. 18 (1994) (setting aside of a valid judicial judgment should not turn upon agreement of the parties). Many of the early GVR’s based upon the Government’s confession of error appear not to have been no-fault V&R’s at all, but rather summary decisions on the merits, with remand for further proceedings. See, e. g., Chiarella v. United States, 341 U. S. 946 (1951) (“[u]pon consideration of the record and *183the confession of error by the Solicitor General,” remanding to the District Court for resentencing) (emphasis added); Penner v. United States, 399 U. S. 522 (1970) (“[o]n the basis of a confession of error by the Solicitor General and of an independent review of the record,” remanding to the District Court “with instructions to dismiss the indictment”).
Our recent practice, however, has been to remand in light of the confession of error without determining the merits, leaving it to the lower court to decide if the confession is correct. As late as 1981, the current Chief Justice, joined by Justice White, objected to this practice. See Mariscal v. United States, 449 U. S. 405, 407 (1981) (Rehnquist, J., dissenting) (“I harbor serious doubt that our adversary system of justice is well served by .. . routinely vacating judgments which the Solicitor General questions without any independent examination of the merits on our own”). I agree with that position. The practice is by now well entrenched, however. See, e. g., Reed v. United States, 510 U. S. 1188 (1994); Ramirez v. United States, 510 U. S. 1103 (1994). It may be considered a separate category of no-fault V&R.
Finally (and most questionably) we have in very recent years GVR’d where the Solicitor General has not conceded error in the judgment below, but has merely acknowledged that the ground, or one of the grounds, on which the lower court relied was mistaken. See, e. g., Alvarado v. United States, 497 U. S. 543 (1990); Chappell v. United States, 494 U. S. 1075 (1990). That is in my view a mistaken practice, since we should not assume that a court of appeals has adopted a legal position only because the Government supported it. Four Justices now sitting on the Court have disapproved this sort of GVR. See Alvarado, supra, at 545 (Rehnquist, C. J., joined by O’Connor, Scalia, and Kennedy, JJ., dissenting).3
*184Today’s cases come within none of these categories of no-fault V&R, not even the questionable last one. In Stutson v. United States, post, p. 193, the decision “in light of” which we vacate the judgment and remand, Pioneer Investment Services Co. v. Brunswick Associates Ltd. Partnership, 507 U. S. 380 (1993), had been on the books for well more than a year before the Eleventh Circuit announced the judgment under review, and for almost two years before that court denied rehearing. Moreover, the parties specifically argued to the Court of Appeals the- question whether Pioneer established the standard applicable to petitioner’s claim of “excusable neglect” under Federal Rule of Appellate Procedure 4(b), with the United States disagreeing with petitioner and taking the position that Pioneer was not controlling. The Eleventh Circuit ruled against petitioner on the merits of his claim; its one-sentence order contained neither a reference to Pioneer nor any suggestion that the court viewed the case as turning on which party’s proffered standard was applied.
The United States has now revised its legal position and— though it makes no suggestion that the Court of Appeals’ judgment was incorrect — is of the view that Pioneer does establish the standard governing petitioner’s claim. But the fact that the party who won below repudiates on certiorari its position on a particular point of law does not give rise to any “intervening,” postjudgment factor that must be considered. The law is the law, whatever the parties, including the United States, may have argued. As described above, *185we have sometimes GVR’d where the Government has, while still supporting the judgment in its favor, conceded the error of a legal point on which the lower court explicitly relied. As I have explained, see supra, at 183, in my view even that practice denies valid judgments the respect to which they are entitled. But the GVR in the present case goes still further. We do not know in this case whether the Eleventh Circuit even agreed with the Government’s position that has now been repudiated; for all we know, the court applied Pioneer and found against petitioner under that standard. The judgment is declared invalid because the Eleventh Circuit might (or might not) have relied on a standard (non-Pioneer) that might (or might not) be wrong, that might (or might not) have affected the outcome, and that the Eleventh Circuit might (or might not) abandon (whether or not it is wrong) because the Government has now abandoned it. This seems to me beyond all reason.
The Court justifies its setting aside of the judgment on the ground that “we [do not] place an excessive burden on [the Eleventh Circuit], relative to [petitioner’s] liberty and due process interests, by inviting it to clarify its ambiguous ruling.” Stutson, post, at 196. Vacating for ambiguity may be justifiable, as I have noted, when the ambiguity calls into question our very power to take and decide the case, see supra, at 181, and n. 2. But where that power is (as it is here) beyond doubt, it seems to me quite improper to vacate merely in order to get a better idea of whether the case is “worth” granting full review. If this is appropriate with respect to court of appeals’ summary dispositions of criminal cases, I see no reason why it is not appropriate with respect to criminal dispositions accompanied by opinions as well. Or, for that matter, why it is not appropriate for civil cases. “GVR’d for clarification of —” should become a common form of order, drastically altering the role of this Court. In my view we have no power to make such a tutelary remand, *186as to a schoolboy made to do his homework again.4 The Court insists that declining to remand for clarification would risk “immunizing summary dispositions ... from our review,” Stutson, post, at 196. That is not so. It is fully within our power to review this case, and any other case summarily decided below, by granting certiorari and proceeding to consider the merits; or indeed, where the circumstances warrant, to summarily reverse. Cf. Heilman, “Granted, Vacated, and Remanded” — Shedding Light on a Dark Corner of Supreme Court Practice, 67 Judicature 389, 391-392 (1984) (noting that in the 1970’s as the Court’s GVR practice “increased far beyond what it had been in earlier years,” its use of summary reversal based on intervening precedents decreased dramatically).
In No. 94-9323, the Court again GVR’s because the Government has changed a legal position: The Commissioner of Social Security informs us that she now agrees with petitioner on a preliminary point of law that the Court of Appeals found in the Government’s favor. And here again, respondent does not concede that the judgment below was in error, for she “ha[s] not. . . reached a firm conclusion” as to her position on the subsequent point of law that will (if her recantation on the preliminary point is accepted) control the outcome of the case. Brief for Respondent in No. 94-9323, p. 13.5 There is, however, a special factor in this second case: Respondent is an agency head, whose view on the legal point in question is in some circumstances entitled to defer*187ence, see Chevron U. S. A. Inc. v. Natural Resources Defense Council, Inc., 467 U. S. 837 (1984). If it were clear that respondent’s change in position were entitled to deference, I would have no problem with the GVR; the new position would then constitute an intervening postjudgment factor whose effect the Court of Appeals should be allowed to consider. But even if we allow deference to an agency view first expressed in pending litigation (as some think we should not, see Anthony, Which Agency Interpretations Should Bind Citizens and the Courts?, 7 Yale J. Reg. 1, 60-61 (1990); cf. Merrill, Judicial Deference to Executive Precedent, 101 Yale L. J. 969, 1023 (1992)), surely a decent concern for those litigating against the Government and for our lower court judges should induce us to disregard, for Chevron purposes, a litigating position first expressed at the certiorari stage. The United States is the most frequent, and hence the most calculating, of our litigants. If we accord deference in the circumstances here, we can expect the Government to take full advantage of the opportunity to wash out, on certiorari, disadvantageous positions it has embraced below; and we can expect it to focus less of its energy upon getting its position “right” in the courts of appeals.
The Court, however, thinks it unnecessary to decide the deference question. It is enough, as the Court sees it, that its summary review has led it to “believe that [the] agency interpretation is reasonably probably entitled to deference and potentially determinative.” Ante, at 172. I do not agree. It seems to me our “intervening-event GVR’s” should not be extended to the situation where (1) the intervening event consists of a party’s going back on what it argued to the court of appeals, and (2) it is not even certain that the change in position is legally cognizable. That seems to me to accord inadequate respect to the work of our colleagues below. Moreover, it is not clear to me that the question before us (should an agency change of position at the certiorari stage be accorded deference?) can even be reached *188by the Court of Appeals. Surely we do not expect the Court of Appeals to declare our vacation and remand invalid. Thus, the Court of Appeals will have before it the somewhat different question whether the agency change of position before it is entitled to deference. I suppose it may conclude that, since a change of position on certiorari is not entitled to deference, a change of position on a remand triggered by change of position on certiorari is not entitled to deference— but that would assuredly be a convoluted holding. The question of what is permissible on certiorari seems to me peculiarly within the domain of this Court. Since we are in doubt on the deference point in the present case, we should either deny the petition, or grant it and have the deference point argued.
The Court’s failure to comprehend why it should make any difference that the Government’s changed litigating position may not be entitled to deference, see ante, at 172-173, displays a lamentable lack of appreciation of the concept of adding insult to injury. It is disrespectful enough of a lower court to set its considered judgment aside because the Government has altered the playing field on appeal; it is downright insulting to do so when the Government’s bait- and-switch performance has not for a certainty altered any factor relevant to the decision. In that situation, at least, we should let the Government live with the consequences of its fickleness or inattention. The Court claims that it would “defeat the purpose of GVR’ing” to determine the deference issue on the merits, since that issue is “based on a circumstance ... that will not be present in any other case brought under the statute at issue.” Ibid. That is true enough (barring the unlikely event that the Government in a later case under this very statute again switches its position at the certiorari stage). But the issue of whether Chevron deference should be accorded to a certiorari-stage switch of litigating position is not at all unique to the individual case or bound up with the underlying statute. It always arises, of *189course, in an individual case involving a particular statute, as do most questions of law. But the issue itself is thoroughly generalizable, and of general importance. In any event, I do not urge that we determine the deference issue on the merits; my vote in these cases is to deny the petitions. Finally, I must remark upon the Court’s assertion that we issued “just such a GVR order last Term, without recorded dissent,” ante, at 173, citing Schmidt v. Espy, 513 U. S. 801 (1994): It is not customary, but quite rare, to record dissents from grants of certiorari, including GVR’s. It would be wrong to conclude from the unsigned order in Schmidt that the vote to GVR was unanimous, or even close to unanimous. Thus, Schmidt does not demonstrate that bait-and-switch-deference GVR’s are an accepted practice; but the fact that Schmidt was apparently the first-ever such GVR, combined with the fact that the Government is back one Term later for another helping, demonstrates the accuracy of my prediction that the Solicitor General will be quick to take advantage of this new indulgence.
What is more momentous than the Court’s judgments in the particular cases before us — each of which extends our prior practice just a little bit — is its expansive expression of the authority that supports those judgments. It acknowledges, to begin with, no constitutional limitation on our power to vacate lower court orders properly brought before us. Ante, at 166. This presumably means that the constitutional grant of “appellate Jurisdiction” over “Cases . . . arising under [the] Constitution [and] Laws of the United States,” Art. Ill, §2, empowers the Court to vacate a state supreme court judgment, and remand the case, because it finds the opinion, though arguably correct, incomplete and unworkmanlike; or because it observes that there has been a postjudgment change in the personnel of the state supreme court, and wishes to give the new state justices a shot at the case. I think that is not so. When the Constitution divides our jurisdiction into “original Jurisdiction” and “appellate *190Jurisdiction,” I think it conveys, with respect to the latter, the traditional accoutrements of appellate power. There doubtless is room for some innovation, particularly such as may be necessary to adapt to a novel system of federalism; but the innovation cannot be limitless without altering the nature of the power conferred.
Not only does the Court reject any constitutional limitation upon its power to vacate; it is unwilling to submit to any prudential constraint as well. Even while acknowledging the potential for “unfair[ness] or manipulation]” and professing to agree that “our GVR power should be exercised sparingly,” ante, at 168, 173, the Court commits to no standard that will control that power, other than that cloak for all excesses, “the equities,” ante, at 168; see ante, at 173, 174, 175; post, at 196. We may, as the Court now pronounces, set aside valid judgments not merely when they are wrong, not merely when intervening events require that someone (either the lower court or we) reconsider them on new facts or under new legal criteria, not merely when it is ambiguous whether we have power to review them, not merely when the United States concedes that the judgment below (or one of the points of law relied upon below, or even one of the points of law possibly relied upon below) is wrong; but whenever there is “a reasonable probability that the decision below rests upon a premise that the lower court would reject if given the opportunity for further consideration.” Ante, at 167. The power to “revis[e] and correc[t]” for error, Marbury v. Madison, 1 Cranch 137, 175 (1803), has become a power to void for suspicion. Comparing the modest origins of the Court’s no-fault V&R policy with today’s expansive dénouement should make even the most Polly-annish reformer believe in camel’s noses, wedges, and slippery slopes.
The Court justifies its approach on the ground that it “alleviates the potential for unequal treatment that is inherent in our inability to grant plenary review of all pending cases *191raising similar issues.” Ante, at 167 (internal quotation marks omitted). I do not see how it can promote equal treatment to announce a practice that we cannot possibly pursue in every case. If we were to plumb the “equities” and ponder the “errors” for all the petitions that come before us — if we were to conduct, for example, in all cases involving summary decisions, today’s balancing of the “burden” to the Court of Appeals against the litigant’s “interests” in having clarification of the ruling, see Stutson, post, at 196, or today’s calculation of “the overall probabilities and equities,” ante, at 173 — we would have no time left for the cases we grant to consider on the merits. Of course we do not purport to conduct such inquiries, not even the basic one of whether the decision below is probably in “error” — which is why we insist that our denial of certiorari does not suggest a view on the merits, see, e. g., Teague v. Lane, 489 U. S. 288, 296 (1989); Singleton v. Commissioner, 439 U. S. 942 (1978) (Stevens, J., respecting denial of petition for writ of certiorari). Moreover, even if we tried applying the Court’s “totality-of-the-circumstanees” evaluation to all the petitions coming before us, we would be unlikely to achieve equal treatment. Such a plastic criterion is liable to produce inconsistent results in any series of decisions; it is virtually guaranteed to do so in a series of decisions made without benefit of adversary presentation (whether we should GVR is rarely briefed, much less argued — as it has not been here) and announced without accompaniment of a judicial opinion (we almost never give reasons as the Court has done today). The need to afford equal treatment argues precisely against the “totality-of-the-circumstances” approach embraced by the Court, and in favor of a more modest but standardized GVR practice.
Henceforth, I shall vote for an order granting certiorari, vacating the judgment below without determination of the merits, and remanding for further consideration, only (1) where an intervening factor has arisen that has a legal bear*192ing upon the decision, (2) where, in a context not governed by Michigan v. Long, 463 U. S. 1032 (1983), clarification of the opinion below is needed to assure our jurisdiction, and (3) (in acknowledgment of established practice, though not necessarily in agreement with its validity) where the respondent or appellee confesses error in the judgment below. (I shall not necessarily note my dissent from GVR’s where those conditions do not exist.) As I have discussed, neither of the present cases meets these standards. Accordingly, I respectfully dissent from today’s orders and would deny both petitions.
[This opinion applies also to No. 94-8988, Stutson v. United States, post, p. 193.]
I emphasize that what is at issue here is our power to set aside a valid judgment — not, as Justice Stevens’ concurrence would have it, “our discretionary authority to manage our certiorari docket.” Ante, at 175. We do the latter by accepting or declining review. But “[w]henever this Court grants certiorari and vacates a court of appeals judgment in order to allow that court to reconsider its decision . . . , the Court is acting on the merits.” Board of Trustees of Keene State College v. Sweeney, 439 U. S. 24, 25-26 (1978) (Stevens, J., dissenting) (emphasis added). Thus, today’s orders go far beyond what Justice Stevens now refers to as “administration of [our certiorari] docket.” Ante, at 176.
In Michigan v. Long, 463 U. S. 1032 (1983), we largely supplanted this policy with the rule that state-court decisions discussing federal law will be presumed to be based on federal law unless the contrary is clear from the face of the opinion. Id., at 1037-1044; see also Arizona v. Evans, 514 U. S. 1, 6-9 (1995) (reaffirming this approach). But cf. Capital Cities Media; Inc. v. Toole, 466 U. S. 378 (1984) (post-Long decision vacating and remanding for clarification of state supreme court decision rendered without opinion).
The Court misdescribes my position when it states that I would limit GVR’s “based on confessions of error that do not purport to concede the whole case” to “eases in which the confession of error concerns a ‘legal *184point on which the lower court explicitly relied.’” Ante, at 171 (quoting infra, at 185). Both the text above and the sentence immediately following the phrase that the Court quotes from my dissent, see ibid., make my position clear. The line of distinction I would draw — and the one long established in our practice — is between a respondent’s concession of error in the lower court’s judgment and a respondent’s concession of error that goes not to the judgment but merely to an aspect of the reasoning below or of respondent’s argument below.
Netherland v. Tuggle, 515 U. S. 951 (1995), upon which the Court relies, see ante, at 170, is not to the contrary. That was not a “no-fault V&R,” but a reversal of the lower court for abuse of discretion in its entry of a stay order.
Because the Commissioner is not prepared to say that she disagrees with petitioner as to the proper disposition of this case, it is questionable whether any case or controversy subsists. Quite apart from the other difficulties with the course the Court has chosen, it seems to me we should not permit the Commissioner to trouble the Fourth Circuit again until she makes up her mind on this issue.