concurring.
I concur in the dismissal of the interlocutory appeal and denial of the petition for writ of mandamus. I write separately, however, to address whether a district court may order that a percentage of settlement monies paid in pending multidistrict litigation (“MDL”) eases be set aside to fund possible fee awards to counsel whose efforts confer a common benefit when motions to remand those cases remain undecided in the district court.
This question arises frequently in MDLs, often affecting hundreds or thousands of plaintiffs. It is important and, as it has not been addressed by any circuit, novel. I therefore have concluded that the issue should be addressed head-on.
Facts
Zyprexa is an anti-psychotic drug manufactured by Eli Lilly and Company (“Lilly”). It has been prescribed for the treatment of schizophrenia and bipolar disorder in over twelve million patients. The litigation below involved over 30,000 individuals who were prescribed and took Zyprexa and claimed that they consequently suffered injuries.1
Many of these individuals filed personal injury law suits in state courts across the country. Lilly removed most of the cases to federal district courts. Some Zyprexa plaintiffs moved to remand in the districts to which the cases were removed, arguing that the federal courts lacked subject matter jurisdiction.
Beginning in 2004, the Judicial Panel on Multidistrict Litigation (“JPML”) transferred thousands of the Zyprexa cases from federal district courts throughout the United States to the Honorable Jack B. Weinstein in the Eastern District of New York for coordinated or consolidated pretrial proceedings.2 Any pending motions, including remand motions filed in the transferor districts, accompanied the transferred cases.3
Soon thereafter, Judge Weinstein appointed a Plaintiffs’ Steering Committee (“PSC I”) composed of attorneys from thirteen law firms who represented various individual plaintiffs. PSC I was responsible for, among other things, coordinating and conducting pretrial discovery for all plaintiffs, speaking for all plaintiffs in pretrial proceedings, making and arguing any motions, and pursuing settlement.4
*120In November 2005, virtually all of the cases then pending in the MDL were settled.5 The settlement contemplated the creation of a claims administration process pursuant to which settling cases were divided into three tracks, attorneys fees were capped for cases in each track, and special masters were appointed to oversee and apply that process.6
In early 2006, the district court entered an order which, as far as is relevant here, provided that “the costs, disbursements and fees of the plaintiffs’ steering committee shall be paid out of the general settlement fund” to the extent approved by the special masters.7 This was implemented in part by setting aside one percent of the gross settlement amount payable to each settling plaintiff in an escrow fund.8
The district court then appointed a second Plaintiffs’ Steering Committee (“PSC II”) to coordinate pretrial proceedings in the few non-settling early cases and in the thousands of incoming and remaining cases, most of which had been filed, removed to federal court, and subsequently transferred by the JPML following the original settlement.9 Lilly and PSC II moved for the establishment of a fund to compensate the attorneys who had worked and would continue to work for the common benefit of all Zyprexa plaintiffs after the original settlement. The district court granted the motion and provided for the establishment of a common benefit fund with certain details to be worked out in subsequent proceedings.10
In due course, the district court ordered that three percent of each remaining plaintiffs gross recovery, if any, whether from a settlement or judgment, be placed in an escrow account.11 Half the money was to come from each plaintiffs recovery and half from fees otherwise payable to his or her attorney.12 Any lawyer who worked for the common benefit of all federal Zyprexa plaintiffs, including but not limited to members of PSC II, were eligible to apply for compensation from this account. Compensation would be granted only on a showing that the attorney “provided significant assistance to all plaintiffs.”13 If any money remained in the fund after all claims on it settled, it would be distributed on a pro rata basis to those who had paid portions of their recoveries into the account.14
Petitioner-appellant the Mulligan Law Firm (“Mulligan”) represents over two thousand plaintiffs who originally filed their actions in state and federal courts in seven states and whose cases against Lilly *121are pending in the MDL.15 Motions to remand sixty-one of these cases to state courts remain pending in the district court.
In December 2006, Mulligan and Lilly agreed in principle to settle all of Mulligan’s Zyprexa cases, although the settlement had not yet been implemented as of the date of argument, because the settling plaintiffs had not fully satisfied the preconditions to consummation.16 Hence, it was not then certain that the settlement funds in fact will be distributed to the Mulligan plaintiffs.17
Anticipating that the settlement would be finalized, the district court enjoined Mulligan from distributing settlement funds to itself or its clients until the fund administrator certified that three percent of the funds designated for each settling plaintiff had been set aside.18 It specified that “[pjayments may be made to individual plaintiffs as soon as that can be done while ensuring that sufficient funds have been held back to comply with the three percent common benefit assessment.”19 As of the date of oral argument, however, the court below had not awarded any fees or costs out of the fund. Mulligan acknowledges that it will have the opportunity to oppose any award of fees in the event fee applications are filed.20
Plaintiffs appeal the district court’s order enjoining them from distributing any settlement funds without a certification that the set-aside had been made.21 Alternatively, they seek a writ of mandamus ordering the district court to vacate the injunction.
Discussion
I. Mandamus
A. The Traditional Mandamus Standard
Mandamus and other prerogative writs “are reserved for really extraordinary causes.”22 ' Mandamus is “not to be used as a substitute for appeal, even though hardship may result from delay and [a] perhaps unnecessary trial.”23 Indeed, it *122often has been said that mandamus is not available unless the petitioner’s right to relief is “clear and indisputable”24 and a direct appeal from a final judgment would not be an adequate remedy.25 Moreover, while traditional formulations of the standard have spoken of using the writ to confine a lower court to the exercise of its proper jurisdiction,26 it long has been clear that mandamus will not lie to review a claim of mere error in a lower court’s jurisdictional determination.27 Indeed, any different view would conflict with the requirement that the petitioner’s right to relief be clear and indisputable.
Mulligan argues that this Court may grant its petition under the traditional mandamus standard because the district court lacked jurisdiction to issue the set aside order in sixty-one of the cases Mulligan originally filed in state court. At least until the district court actually awards fees out of the common benefit fund, however, Mulligan cannot demonstrate a need for traditional mandamus because it has not yet suffered any injury. In any case, it lacks a “clear and indisputable” right to relief28 even assuming that the set aside order were erroneous. But traditional *123mandamus is not the only potentially relevant standard.
B. Advisory Mandamus
The stringencies of traditional mandamus standards admit of a limited exception for what has been termed “advisory mandamus” — the use of mandamus to provide guidance on a novel question of general or exceptional importance to the administration of justice that should not await review by appeal from a final judgment.
In Schlagenhauf v. Holder,29 the Supreme Court “departed in some degree” from the traditional mandamus standards to countenance the use of the writ for such advisory purposes.30 There, the Court indicated that mandamus may be used to settle important questions of first impression where there is a “substantial allegation of usurpation of power” by the district court.31 Thus, courts of appeals have some scope for use of mandamus to settle novel questions. On the other hand, as this Court has written, “[n]o one could reasonably suppose that the [Supreme] Court has now subscribed ... to a doctrine that any non-frivolous claim of error in a decision is a claim of ‘usurpation of power’ on the theory that courts are bound to decide all issues correctly.”32
The Second Circuit’s decisions subsequent to Schlagenhauf do not resolve entirely the uncertainty as to the criteria dividing appropriate from inappropriate uses of advisory mandamus. The cases that bear most strongly on the facts at bar are In re International Business Machines Corp.33 and United States v. Amante.34 In IBM, the Circuit granted *124mandamus to halt the district court’s consideration of whether the Tunney Act requirement of judicial approval of consent decrees in government antitrust cases35 applied to stipulations of dismissal under Federal Rule of Civil Procedure 41(a)(1).36 It did so, however, not simply because it disagreed with the district court’s retention of jurisdiction to consider the issue, but for three other reasons. First, both parties were united in their desire that the case be dismissed outright.37 Second, the case already had gone on for thirteen years, tying up “one of the nation’s largest industrial concerns,” during which time “many of the underlying products [had become] commercially outmoded.”38 Finally, there was no serious basis for the assertion that the Tunney Act applied to a Rule 41(a)(1) dismissal in light of the plain language of the statute and “the clear and indisputable legislative history » 39
Amante concerned a different question, viz. whether a district court had erred in bifurcating a simple felon-in-possession firearm case to require first a trial on the issue of possession and then a separate trial on the issue of the prior felony conviction, the latter of which would take place only after a jury finding of possession. In granting mandamus to vacate the bifurcation order, the Circuit relied upon its conclusions that (1) the issue was significant because “[i]ts resolution will affect numerous defendants in similar trials and will aid in the administration of criminal justice,” and (2) “no remedy [wa]s available to the government other than mandamus.”40
Both IBM and Amante demonstrate that this Court has granted mandamus for advisory purposes when presented with questions of significance and the district court had committed a clear abuse of discretion and usurpation of power.41 In addition, both cases reached the merits of the issues presented by the petitions as, of course, the Court was compelled to explain the district court’s error. But they did not involve the question whether this Court may reach the merits for advisory purposes to deny a petition for mandamus where the issue and need for review are sufficiently pressing.
This question was answered in the affirmative in Kaufman v. Edelstein.42 There, two expert witnesses whom the government had subpoenaed to testify in a civil antitrust case petitioned for a writ of man*125damus directing the district court to quash the subpoenas. Observing that the district court’s denial of the motions to quash “clearly was no 'usurpation of power’ ” nor a clear abuse of discretion, the Court, speaking through Judge Friendly, stated that the only argument in favor of granting the writ was that the case presented a “novel and important” issue.43 The Court then discussed the merits of the case at some length, concluding that, although there existed no privilege protecting expert witnesses from being called to testify, courts in their discretion could protect expert witnesses when good cause was shown.44 Judge Friendly listed a number of factors pertinent to deciding whether a particular expert witness merited the protection of the court.45
In sum, we have held that it is appropriate for us to discuss the merits of a challenged district court action to deny mandamus when the question is of sufficient importance and there is a dearth of guidance on the issue.
C. Application To This Petition
The district court’s set aside order, in my view, raises just such a question. The issue never has been addressed by this Circuit. Yet it arises frequently in MDLs,46 many encompassing hundreds and sometimes thousands of cases and affecting hundreds or thousands of plaintiffs.47 A resolution of the issue here would be “applicable to [this] entire class of cases.”48 A decision on the merits thus would offer guidance to both courts and litigants and aid the efficient administration of justice.
II. The District Court’s Order
A Subject Matter Jurisdiction
The fundamental question here is whether a district court has subject matter jurisdiction — power—to order that settlement funds be set aside to fund possible fee awards to counsel in MDL cases that have been removed to federal court and in which motions to remand remain undecided in the district court. In essence, Mulligan argues that it does not and that we therefore either should (1) decide the jurisdictional issues that were not passed upon below and exempt from the set aside order any cases in which removal was improper or, alternatively, (2) exempt all of these cases from the set aside order on the *126theory that the entry of the set aside order in advance of determination of the remand motions was erroneous. These contentions are unpersuasive.
The starting point for this analysis is the proposition that a federal court has jurisdiction to determine its own jurisdiction.49 While it must make this determination before it reaches the merits of the case before it,50 it is empowered to issue nondispositive orders between the filing of the action and its ultimate determination on the merits.51
Mulligan nevertheless argues that federal courts are permitted, prior to a ruling disposing of any subject matter jurisdiction challenge, to take only such actions as may aid in determining their jurisdiction and that actions reaching the merits are void.52 The district court’s set aside order, in its view, reached the merits and thus went beyond the limits of the court’s authority. Mulligan relies for this proposition on United States Catholic Conference v. Abortion Rights Mobilization, Inc.53
Catholic Conference involved an appeal from an order holding two non-party witnesses in civil contempt, over their objection that the district court lacked subject matter jurisdiction, for refusing to comply with a subpoena. The Supreme Court held that a non-party witness can challenge the subject matter jurisdiction of a court in defense of a civil contempt citation.54 It stated further that the contempt orders would be void if the district court lacked subject matter jurisdiction. Mulligan argues that Catholic Conference therefore stands for the proposition that federal courts are permitted only to “take actions in aid of determining [their] jurisdiction.” 55 But he overreads the decision.
In Willy v. Coastal Corp.,56 the Supreme Court upheld a district court’s Rule 11 sanctions order notwithstanding a subsequent determination that the district court had lacked subject matter jurisdiction, and it rejected the petitioner’s contention that this result was precluded by Catholic Conference. As the Court explained:
“[a] final determination of lack of subject-matter jurisdiction of a case in federal court, of course, precludes further *127adjudication of it. But such a determination does not automatically wipe out all proceedings had in the district court at a time when the district court operated under the misapprehension that it had jurisdiction.”57
Thus, a district court’s order that is “collateral to the merits” may be upheld despite a later conclusion that the court lacked subject matter jurisdiction over the litigation.58
In this case, the set aside order simply imposed an assessment in order to create a fund that could be used to compensate attorneys who demonstrate that their efforts conferred a benefit on Zyprexa plaintiffs generally.59 It is even less related to the ultimate merits than orders awarding attorney’s fees, which are collateral matters over which a court retains jurisdiction even if it ultimately is determined to lack subject matter jurisdiction.60 It does not preclude the Mulligan plaintiffs from challenging any proposed distribution of funds on the ground that they did not benefit from the efforts of counsel to whom a distribution might be made. At least in these circumstances, the set aside order was collateral to the merits and well within the district court’s power.
The district court’s determination to defer consideration of the jurisdictional challenges also was comfortably within the bounds of its discretion. While it usually is advisable for district courts to rule on any challenge to subject matter jurisdiction early in a lawsuit, district courts have broad scope to manage their own dockets in light of considerations of “economy of time and effort for itself, for counsel, and for litigants.”61 The impact of such considerations in this case, and in some other MDL and mass tort cases, often may suggest a different course.
Alleged mass torts, as the term implies, spawn masses of litigation. Plaintiffs understandably sue in fora that they regard as convenient or likely to be favorable including, in many instances, state courts. Indeed, some frame their complaints in ways that raise issues concerning removability. Defendants, moved by similar considerations, frequently remove many of these cases to federal courts. These tactical steps in turn often provoke large numbers of remand motions, as occurred in this MDL.62 Indeed, Mulligan alone has *128filed motions to remand in over 50 of its Zyprexa cases.63
A district judge managing such a complex situation reasonably may conclude that the court’s time, especially early in the litigation, is better spent on activities other than deciding scores or hundreds of individual remand motions. It often would be appropriate to conclude that many of the cases in which such motions are filed are likely to settle and that the expenditure of judicial resources on remand motions likely would be wasted. Moreover, even if an MDL or comparable court dealt immediately with remand motions pending at a given moment, new cases and new remand motions are likely to be filed throughout the pendency of the litigation.64 Deciding each motion prior to addressing other issues common to all or most cases therefore could require a significant expenditure of judicial time and resources, derailing the progress of the litigation as a whole. Delays of this nature could well frustrate the purpose of MDLs, which is “to promote just and efficient conduct” of actions involving common questions of fact.65
While every case turns on its own facts, I cannot say that the district court’s decision to give priority to moving thousands of Zyprexa cases towards settlement or other disposition over focusing promptly on the hundreds of pending remand motions was an improvident exercise of the broad discretion that inheres in its docket management function.
B. The Common Fund Doctrine
The district court’s set aside order raises also the question whether a district court in an MDL or other litigation involving a large number of separately represented individual claimants, in which any recovery will be unique to each plaintiff, has the authority to create a fund out of those recoveries for the purpose of financing an award of fees and expenses to those counsel whose work benefits the entire group of plaintiffs. The answer to this question lies in the common fund or common benefit doctrine.
The so-called “American rule” is that the prevailing litigant ordinarily must bear the litigant’s own attorney’s fees.66 Since the nineteenth century, however, the Supreme Court has recognized an equitable exception to this rule67 — known as the common fund or common benefit doctrine — that permits litigants or lawyers who recover a common fund for the benefit of persons other than themselves to obtain reasonable attorney’s fees out of the fund, thus spreading the cost of the litigation to its beneficiaries.68 The doctrine “reflects the traditional practice in courts of equity,”69 which recognized “that persons who obtain the benefit of a lawsuit without eontribut*129ing to its cost are unjustly enriched at the successful litigant’s expense.”70 A court acting in equity therefore may assess attorney’s fees proportionally among those benefitted by the suit.71
In Alyeska Pipeline, the Supreme Court identified several characteristics of cases in which the application of the common benefit doctrine may be appropriate. These include the ease in identifying the persons, or classes of persons, benefitted by the recovery, ease in tracing the benefit flow from the fund to those persons, and confidence that the costs of litigation can be shifted with some exactitude to those benefitted by the litigation.72 Courts routinely employ the common benefit doctrine in awarding fees to counsel in class actions 73 which, when successful, often result in lump sum recoveries that benefit identifiable classes. Class actions present also the free-rider problems the common fund doctrine was designed to remedy, as class members who do not hire counsel nonetheless benefit from any recovery. The doctrine thus prevents the unjust enrichment of these class members at the expense of class counsel by compensating counsel in proportion to the benefit they have obtained for the entire class, rather than just the named class representatives with whom they contracted. Likewise, it prevents the unjust enrichment of class members at the expense of the class representatives who contracted with attorneys to prosecute the action.74
Application of the doctrine to class actions is straightforward. The class generally is represented by counsel. Any proceeds from the litigation are awarded to the class as a whole or by a formula that permits determination of the amount of the aggregate benefit conferred on the class. Courts may set some percentage of the lump sum recovery as the fee or, using the “lodestar approach,” ascertain the number of hours reasonably billed to the class, multiply it by an appropriate hourly rate, and deduct the product from any recovery.75
The situation is somewhat different with respect to MDLs consisting of individual cases prosecuted by individual plaintiffs, sometimes numbering in the thousands,76 and other litigation involving large numbers of separately represented claimants. Unlike most class actions, recoveries by individual plaintiffs or groups of plaintiffs *130in such matters may occur at different times, and individual plaintiffs or groups of plaintiffs, unlike most individual class members, usually are represented by individual counsel. Nevertheless, there are substantial similarities to class actions as well. As an initial matter, the efficient handling of such cases demands a similar approach to case management. District courts typically appoint a lead counsel or plaintiffs’ steering committee to coordinate and conduct pretrial proceedings on behalf of all plaintiffs in order to avoid what otherwise might well become chaotic.77 Moreover, while individual plaintiffs are separately represented, they typically benefit also — often predominantly — from the work of the lead counsel or committee.
The same equitable considerations that warrant payment of class counsel out of common funds generated by their efforts apply in these circumstances as well. The desirability — indeed, the compelling need — to have pretrial proceedings managed or at least coordinated by lead counsel or a steering or executive committee demands the existence of a source of compensation for their efforts on behalf of all. The logical, and a most equitable, source of that compensation is recoveries of individual plaintiffs who benefit from that work. Indeed, foreclosing those recoveries as a source of funding for the common benefit work would enrich the non-contributing individual plaintiffs unjustly at the expense of either or both of the lead counsel and any contributing individual plaintiffs.78 The district court thus acted within the scope of its discretion when it established an account to compensate counsel for common benefit work funded by a set aside of future Zyprexa MDL recoveries.79
Conclusion
For the foregoing reasons, I conclude that the district court did not err as a matter of law or abuse its discretion in ordering a set aside of settlement monies in a common benefit fund. I therefore concur in the denial of the petition and dismissal of the appeal.
. In re Zyprexa Prods. Liab. Litig., 549 F.Supp.2d 496, 498-99 (E.D.N.Y.2008).
. See In re Zyprexa Prods. Liab. Litig., 314 F.Supp.2d 1380, 1382-83 (J.P.M.L.2004).
. See Federal Judicial Center, Manual for Complex Litigation Fourth § 22.35, at 371 (2004) (hereinafter “Manual”).
. See In re Zyprexa Prods. Liab. Litig., No. 04 MDL 1596, 2004 WL 3520245, at *2-3 (E.D.N.Y. Jun. 17, 2004).
. See In re Zyprexa Prods. Liab. Litig., 467 F.Supp.2d 256, 261 (E.D.N.Y.2006).
. In re Zyprexa Prods. Liab. Litig., 424 F.Supp.2d 488, 490 (E.D.N.Y.2006).
. Id. at 497.
. In re Zyprexa Prods. Liab. Litig., 467 F.Supp.2d at 263.
. See In re Zyprexa Prods. Liab. Litig., 467 F.Supp.2d at 261.
. Id. at 275.
. In re Zyprexa Prod. Liab. Litig., No. 04 MD 1596, 2007 WL 2340790, at *1 (E.D.N.Y. Aug. 17, 2007). An earlier order imposed a set aside of a percentage of plaintiffs' attorney's fees, rather than of each plaintiff's gross recovery. Following hearings before a special master and Judge Weinstein, the nature of the set aside was modified. Compare id. with In re Zyprexa Prod. Liab. Litig., 467 F.Supp.2d at 266.
. In re Zyprexa Prod. Liab. Litig., 2007 WL 2340790, at *1.
. In re Zyprexa Prod. Liab. Litig., 467 F.Supp.2d at 266.
. Id.
. A-319, 349, 393, 589.
. In re Zyprexa Prod. Liab. Litig., No. 04 MD 1596, 2007 WL 2340791, at *1 (E.D.N.Y. Aug. 17, 2007).
. Oral Arg. 11:34-36.
. Id.
. Id.
. Oral Arg. 11:04-05; 11:24-27.
. A-816.
. Ex parte Fahey, 332 U.S. 258, 260, 67 S.Ct. 1558, 91 L.Ed. 2041 (1947); see also Aref v. United States, 452 F.3d 202, 206 (2d Cir.2006) ("Mandamus is an extraordinary remedy, available only in extraordinary circumstances." (citing In re United States, 10 F.3d 931, 933 (2d Cir.1993))); In re Nagy, 89 F.3d 115, 117 (2d Cir.1996) (requiring an "extraordinary showing" to get mandamus review in lieu of appeal from final judgment (citing In re Drexel Burnham Lambert Inc., 861 F.2d 1307, 1312 (2d Cir.1988))); In re Repetitive Stress Injury Litig., 11 F.3d 368, 373 (2d Cir.1993) ("The granting of a writ of mandamus is an extraordinary measure and should be done sparingly.”); United States v. Victoria—21, 3 F.3d 571, 575 (2d Cir.1993) ("Such writs may not be used as a mere substitute for an appeal, but rather the power should only be exercised in extraordinary situations.”) (internal quotations omitted).
. Schlagenhauf v. Holder, 379 U.S. 104, 110, 85 S.Ct. 234, 13 L.Ed.2d 152 (1964) (citations omitted); accord Bankers Life & Cas. Co. v. Holland, 346 U.S. 379, 383, 74 S.Ct. 145, 98 L.Ed. 106 (1953); Roche v. Evaporated Milk Ass’n, 319 U.S. 21, 30, 63 S.Ct. 938, 87 L.Ed. 1185 (1943); In re Ivy, 901 F.2d 7, 10 (2d Cir.1990) (quoting Roche); cf. In re Traffic Executive Ass'n-E. R.R.s, 627 F.2d 631, 634 (2d Cir.1980) (denying writ where district court disapproved settlement agreement, noting that "[t]he likelihood that class members will find it tedious and time consuming to *122prove their losses does not make this an out-of-the-ordinary case”).
. E.g., Moses H. Cone Mem’l Hosp. v. Mercury Constr. Corp., 460 U.S. 1, 18, 103 S.Ct. 927, 74 L.Ed.2d 765 (1983); Bankers Life, 346 U.S. at 384, 74 S.Ct. 145; In re Certain Underwriter, 294 F.3d 297, 302 (2d Cir.2002); In re Aguinda, 241 F.3d 194, 202 (2d Cir.2001); Drexel, 861 F.2d at 1312.
. E.g., United States v. Giffen, 473 F.3d 30, 38 n. 7 (2d Cir.2006); United States v. Amante, 418 F.3d 220, 222 (2d Cir.2005); In re SEC ex rel. Glotzer, 374 F.3d 184, 187-88 (2d Cir.2004); United States v. Coppa, 267 F.3d 132, 138-39 (2d Cir.2001); In re von Bulow, 828 F.2d 94, 98-99 (2d Cir.1987).
. See, e.g., Parr v. United States, 351 U.S. 513, 520-21, 76 S.Ct. 912, 100 L.Ed. 1377 (1956); Bankers Life, 346 U.S. at 382-83, 74 S.Ct. 145; Roche, 319 U.S. at 26, 63 S.Ct. 938; Hong Mai Sa v. Doe, 406 F.3d 155, 159 (2d Cir.2005) (citing Richardson Greenshields Secs., Inc. v. Lau, 825 F.2d 647, 652 (2d Cir.1987)).
. See, e.g., Ex parte Chicago, R.I. & Pac. Ry. Co., 255 U.S. 273, 275-76, 41 S.Ct. 288, 65 L.Ed. 631 (1921) (Brandéis, J.) (mandamus does not lie to review doubtful district court jurisdictional determination); In re Ivy, 901 F.2d at 10 (”[T]he writ will not issue to review an order overruling a plea to the jurisdiction ....” (quoting Roche, 319 U.S. at 31, 63 S.Ct. 938)); Ward Baking Co. v. Holtzoff, 164 F.2d 34, 36 (2d Cir.1947) (denying writ where ”[w]e are not satisfied that the district court 'clearly' lacks jurisdiction”); see also United States v. DiStefano, 464 F.2d 845, 850 (2d Cir.1972) (Friendly, J.) ("[M]ere error, even gross error in a particular case, as distinguished from a calculated and repeated disregard of governing rules, does not suffice to support issuance of the writ.”); accord, e.g., Nat'l Right to Work Legal Def. v. Richey, 510 F.2d 1239, 1242 (D.C.Cir.1975) (‘‘If the lower court is clearly without jurisdiction, the writ will ordinarily be granted. If, however, the jurisdiction of the lower court is doubtful ... the writ will ordinarily be denied.” (quoting Ex parte Chicago, 255 U.S. at 275, 41 S.Ct. 288)) (citations omitted); Comfort Equip. Co. v. Steckler, 212 F.2d 371, 373 (7th Cir.1954) ("If a rational and substantial legal argument can be made in support of the questioned jurisdictional ruling the case is not appropriate for mandamus or prohibition even though on normal appeal a reviewing court might find reversible error.” (quoting Am. Airlines v. Forman, 204 F.2d 230, 232 (3d Cir.1953))).
. Moses H. Cone Mem’l Hosp., 460 U.S. at 18, 103 S.Ct. 927; see also In re Vazquez-Botet, 464 F.3d 54, 57 (1st Cir.2006) (" '[E]ven where the merits clearly favor the petitioner, relief may be withheld for lack of irreparable injury!.]’ " (citing In re Martinez-Catala, 129 F.3d 213, 217 (1st Cir.1997))); In re Nat’l Presto Indus., Inc., 347 F.3d 662, 663 (7th Cir.2003) (requiring showing of irreparable harm before granting mandamus relief); United States v. Wexler, 31 F.3d 117, 128 (3rd Cir.1994) (same).
. 379 U.S. 104, 85 S.Ct. 234, 13 L.Ed.2d 152.
. Kaufman v. Edelstein, 539 F.2d 811, 817 (2d Cir.1976) (Friendly, J.). See generally Note, Supervisory and Advisory Mandamus Under the All Writs Act, 86 Harv. L.Rev. 595 (1973) (hereinafter "Harvard Note”).
The Supreme Court has recognized a second limited exception to traditional mandamus standards for the exercise of "supervisory control of the District Courts by the Courts of Appeals.” La Buy v. Howes Leather Co., 352 U.S. 249, 259-60, 77 S.Ct. 309, 1 L.Ed.2d 290 (1957). In La Buy, the district court, notwithstanding prior admonitions by the Seventh Circuit against excessive use of special masters, referred two antitrust cases to special masters on the ground that its docket was too congested. It did so in line with a prior practice of too frequent references to special masters that the Supreme Court characterized as "little less than an abdication of the judicial function.” 352 U.S. at 256, 258, 77 S.Ct. 309. In those circumstances, the Court held that supervisory mandamus was appropriate. Id. at 259-60, 77 S.Ct. 309. The case therefore appears to authorize "supervisory” mandamus "when a plausible case could be made for the danger of frequent recurrence” of a “probably erroneous practice.” See Harvard Note, 86 Harv. L.Rev. at 609-10; see also, e.g., United States v. Yemitan, 70 F.3d 746, 748 (2d Cir.1995) ("[A]n arbitrary practice of sentencing without prof-erred [sic] reasons would amount to an abdication of judicial responsibility subject to mandamus.”) (dictum).
As demonstrated by the discussion below, because the practice challenged here was not "probably erroneous,” much less an "abdication of the judicial function,” supervisory mandamus has no proper role in this case.
. 379 U.S. at 110-11, 85 S.Ct. 234.
. Kaufman, 539 F.2d at 819 (citations omitted).
. 687 F.2d 591 (2d Cir.1982).
. 418 F.3d 220.
Additionally, this Court has read Schlagenhauf as creating an "escape hatch” from the finality rule in the context of discovery, when the "question is of extraordinary significance or there is extreme need for reversal of the district court's mandate before the case goes to judgment.” Am. Express Warehousing, Ltd. v. Transamerica Ins. Co., 380 F.2d 277, 282 (2d Cir.1967). Subse*124quent discovery cases also have invoked this standard. See, e.g., In re Attorney Gen. of the United States, 596 F.2d 58, 63 (2d Cir.1979); Nat’l Super Spuds, Inc. v. N.Y. Mercantile Exch., 591 F.2d 174, 181 (2d Cir.1979); see also Glotzer, 374 F.3d at 187 (quoting Coppa, 267 F.3d at 137-38 (applying same standard, but without characterizing mandamus review as "advisory”)). These cases establish that mandamus sometimes may be granted to reverse an erroneous discovery ruling where the ruling orders discovery that could not be undone on appeal from a final judgment because the allegedly undiscoverable information will have been disclosed. But the discovery cases shed little light on the appropriate course in this case, in part because Mulligan would suffer no cognizable harm if review were denied unless and until there is a final order to disburse money it contributes to the set aside.
. 15U.S.C. § 16(e)(1).
. The district court had heard argument on the issue but had not yet issued a decision. 687 F.2d at 596.
. Id. at 600.
. Id. at 599-600.
. Id. at 600-02.
. 418 F.3d at 222.
. Id. at 222; see also IBM, 687 F.2d at 603.
. 539 F.2d 811.
. Id. at 819.
. Id. at 819, 821.
. Id. at 822; see also Stans v. Gagliardi, 485 F.2d 1290, 1292 (2d Cir.1973).
. The Manual for Complex Litigation instructs courts "[ejarly in the litigation” to “determine the method of compensation” for lead and liaison counsel “and establish arrangements for their compensation, including setting up a fund to which designated parties should contribute in specified proportions.” manual § 14.215, at 202.
. See, e.g., In re Wilson, 451 F.3d 161, 163 (3d Cir.2006) (30,000-35,000 plaintiffs with suits pending before MDL court); In re Diet Drugs, 369 F.3d 293, 298 (3d Cir.2004) (MDL involved 18,000 individual lawsuits and over 100 putative class actions); In re Orthopedic Bone Screw Prods. Liab. Litig., 193 F.3d 781, 784 (3d Cir.1999) (“thousands of plaintiffs” in MDL); In re San Juan Dupont Plaza Hotel Fire Litig., 111 F.3d 220, 222 (1st Cir.1997) ("hundreds of claims” filed by “thousands of plaintiffs” consolidated in MDL); In re Dow Corning Corp., 86 F.3d 482, 485 (6th Cir.1996) (440, 000 plaintiffs elected to become class members for purposes of settlement in MDL); In re Food Lion, Inc. Fair Labor Standards Act "Effective Scheduling” Litig., 73 F.3d 528, 531 (4th Cir.1996) (nearly 1,000 plaintiffs in MDL); In re Agent Orange Prod. Liab. Litig., 818 F.2d 145, 165-66 (2d Cir.1987) (240,000 claimants in MDL).
. SEC v. Stewart, 476 F.2d 755, 758 (2d Cir.1973) (discussing standard for supervisory mandamus).
. United States v. United Mine Workers of Am., 330 U.S. 258, 291, 67 S.Ct. 677, 91 L.Ed. 884 (1947); Kuhali v. Reno, 266 F.3d 93, 100-01 (2d Cir.2001).
. Sinochem Int'l Co., Ltd. v. Malaysia Int’l Shipping Corp., 549 U.S. 422, 431, 127 S.Ct. 1184, 167 L.Ed.2d 15 (2007) (citing Ruhrgas AG v. Marathon Oil Co., 526 U.S. 574, 585, 119 S.Ct. 1563, 143 L.Ed.2d 760 (1999) and Steel Co. v. Citizens for a Better Env’t, 523 U.S. 83, 100-01 & n. 3, 118 S.Ct. 1003, 140 L.Ed.2d 210 (1998)).
. See Sinochem, 549 U.S. at 431, 127 S.Ct. 1184 (recognizing that federal court "has leeway ‘to choose among threshold grounds for denying audience to a case on the merits’ ” before determining jurisdiction and noting permissibility of dismissal of state law claims on discretionary grounds, dismissal based on Younger abstention, and dismissal under Totten v. United States, 92 U.S. 105, 23 L.Ed. 605 (1876), all before considering jurisdictional questions); United Mine Workers, 330 U.S. at 290, 67 S.Ct. 677 (district court "unquestionably had the power to issue a restraining order ... pending a decision upon its own jurisdiction”); see also id. at 295, 67 S.Ct. 677 (noting Supreme Court would affirm district court's judgment of criminal contempt as validly punishing violations of order then outstanding, even if it was later determined that the district court lacked jurisdiction).
. Pet. Br. at 32-34.
. 487 U.S. 72, 108 S.Ct. 2268, 101 L.Ed.2d 69 (1988).
. Id. at 76, 108 S.Ct. 2268.
. Pet. Br. at 32.
. 503 U.S. 131, 112 S.Ct. 1076, 117 L.Ed.2d 280 (2001).
. Id. at 137, 112 S.Ct. 1076.
. Id.
. In re Zyprexa Prod. Liab. Litig., 467 F.Supp.2d at 266.
. See, e.g., Cooter & Gell v. Hartmarx Corp., 496 U.S. 384, 395, 110 S.Ct. 2447, 110 L.Ed.2d 359 (1990) ("It is well established that a federal court may consider collateral issues after an action is no longer pending. For example, district courts may award costs after an action is dismissed for want of jurisdiction.”); Moore v. Permanente Med. Group, Inc., 981 F.2d 443, 445 (9th Cir.2003) ("It is clear that an award of attorney's fees is a collateral matter over which a court normally retains jurisdiction even after being divested of jurisdiction on the merits.”); see also In re Nineteen Appeals, 982 F.2d at 609 (“[A]n order which definitively resolves claims for attorney's fees and expenses payable out of a common fund is severable from the decision on the merits.”) (citing Trustees v. Greenough, 105 U.S. 527, 531, 26 L.Ed. 1157 (1881)); Overseas Dev. Disc Corp. v. Sangamo Constr. Co., 840 F.2d 1319, 1324 (7th Cir.1988) (declaring it to be "settled that a decision awarding or denying attorney’s fees and expenses from a common fund ... is severable from the decision on the merits”); Memphis Sheraton Corp. v. Kirkley, 614 F.2d 131, 133 (6th Cir.1980) (award of attorney’s fees from common fund "is collateral” to the merits (citing Sprague, 59 S.Ct. at 780)).
. Landis v. North Am. Co., 299 U.S. 248, 254-55, 57 S.Ct. 163, 81 L.Ed. 153 (1936).
. See, e.g., Galati v. Eli Lilly & Co., No. 05-4338-CV-C-NKL, 2005 WL 3533387, at *1 (W.D.Mo. Dec. 22, 2005) (hundreds of remand motions pending in Zyprexa MDL).
. A-814-15.
. See, e.g., In re Zyprexa Prods. Liab. Litig., 467 F.Supp.2d at 261-62.
. 28 U.S.C. § 1407.
. See, e.g., Alyeska Pipeline Serv. Co. v. Wilderness Soc’y, 421 U.S. 240, 247, 95 S.Ct. 1612, 44 L.Ed.2d 141 (1975).
. Boeing Co. v. Van Gemert, 444 U.S. 472, 478, 100 S.Ct. 745, 62 L.Ed.2d 676 (1980) (citing Cent. R.R. & Banking Co. v. Pettus, 113 U.S. 116, 5 S.Ct. 387, 28 L.Ed. 915 (1885) and Trustees v. Greenough, 105 U.S. 527, 26 L.Ed. 1157 (1882)).
. Id.; Alyeska, 421 U.S. at 245, 95 S.Ct. 1612; Mills v. Elec. Auto-Lite Co., 396 U.S. 375, 392-92, 90 S.Ct. 616, 24 L.Ed.2d 593 (1970); Sprague v. Ticonic Nat’l Bank, 307 U.S. 161, 164-65, 59 S.Ct. 777, 83 L.Ed. 1184 (1939); Cent. R.R., 113 U.S. at 124-25, 5 S.Ct. 387; Greenough, 105 U.S. at 532.
. Boeing, 444 U.S. at 478, 100 S.Ct. 745 (citing Greenough, 105 U.S. at 532-37).
. Id. (citing Mills, 396 U.S. at 392, 90 S.Ct. 616).
. Id. (citing Mills, 396 U.S. at 394, 90 S.Ct. 616).
. Alyeska, 421 U.S. at 264 n. 39, 95 S.Ct. 1612; see Boeing, 444 U.S. at 478-79, 100 S.Ct. 745; In re Nineteen Appeals Arising Out of San Juan Dupont Plaza Hotel Fire Litig., 982 F.2d 603, 607 (1st Cir.1992).
. See, e.g., Boeing, 444 U.S. at 479-81, 100 S.Ct. 745; In re Agent Orange Prod. Liab. Litig., 818 F.2d 226, 237 (2d Cir.1987).
. Boeing, 444 U.S. at 477, 100 S.Ct. 745; Lindy Bros. Builders, Inc. v. Am. Radiator & Standard Sanitary Corp., 487 F.2d 161, 165 (3d Cir.1973); 4 Newberg on Class Actions § 14:6 (4th ed.2002).
. Goldberger v. Integrated Res., Inc., 209 F.3d 43, 47 (2d Cir.2000).
. For example, in many MDLs involving personal injury claims, courts have denied class certification pursuant to Fed.R.Civ.P. 23(b)(3) based on findings that questions affecting individual members of the putative class regarding injury, causation and/or damages predominate over questions of fact or law common to the putative class. See, e.g., In re Rezulin Prods. Liab. Litig., 210 F.R.D. 61, 66-68, 71 (S.D.N.Y.2002); In re Phenylpropanolamine Prods. Liab. Litig., 208 F.R.D. 625, 633 (W.D.Wash.2002); see also In re Prempro, 230 F.R.D. 555, 566-67 (E.D.Ark.2005) (consumer fraud claim).
. See, e.g., Smiley v. Sincoff, 958 F.2d 498, 499 (2d Cir.1992) (district court did not abuse its discretion in requiring that attorney fee paid to plaintiffs’ committee be distributed pro rata among committee members); Vincent v. Hughes Air West, Inc., 557 F.2d 759, 774-75 (9th Cir.1977) (district court had authority to direct appointment of committee of lead counsel); MacAlister v. Guterma, 263 F.2d 65, 68 (2d Cir.1958) (approving appointment of "general counsel” and consolidation of actions for pretrial purposes to avoid unnecessary expenses, duplication and delay in three stockholders' derivative actions); manual § 10.221 at 24-25.
. See In re Diet Drugs, 582 F.3d 524, 546 (3d Cir.2009) (approving district court's apportionment of settlement funds among class counsel in MDL pursuant to common benefit doctrine); In re Air Crash Disaster, 549 F.2d 1006, 1016, 1018-19 (5th Cir.1977) (holding that district court properly exercised its authority to award fees to MDL lead counsel committee paid by other plaintiffs’ counsel pursuant to common benefit doctrine); see also Sprague, 307 U.S. at 166, 59 S.Ct. 777 ("That the party ... neither purported to sue for a class nor formally established by litigation a fund available to the class, does not seem to be a differentiating factor so far as it affects the source of the recognized power of equity to grant reimbursements.”); Smiley, 958 F.2d at 499, 501 (holding that district court did not abuse its discretion in establishing fee structure whereby non-committee member attorneys placed percentage of their fees in escrow for pro rata distribution among committee members).
. Zyprexa, 467 F.Supp.2d at 265-67.