Securities Exchange Commission v. Monarch Funding Corporation, Leo M. Eisenberg, Steven R. Cloyes, and Richard M. Cannistraro, Richard O. Bertoli

McLAUGHLIN, Circuit Judge: The question before us on this appeal, one of first impression, is whether findings made in a criminal sentencing proceeding may preclude relitigation of an issue in a subsequent civil case. The SEC sued Richard Bertoli in the United States District Court for the Southern District of New York (Sand, J.) seeking injunctive relief based on his alleged violations of the federal securities laws. Those same alleged violations formed the predicate acts for a parallel criminal prosecution of Bertoli under the Racketeer Influenced and Corrupt Organizations Act (“RICO”), 18 U.S.C. § 1961, et seq., in the United States District Court for the District of New Jersey (Lechner, J.). In the criminal case, a jury acquitted Bertoli of the RICO charges but convicted him on related obstruction of justice charges. Following the conviction, Judge Lechner enhanced Bertoli’s sentence after finding that he had committed securities fraud and had engaged in -an eight-year conspiracy to cover it up. The SEC then moved for summary judgment in the civil proceeding, arguing that Bertoli should be collaterally estopped from denying his securities fraud liability by virtue of Judge Lechner’s sentencing findings. Judge Sand agreed and granted the SEC a permanent injunction enjoining Bertoli from future violations of Section 10(b) of the Securities Exchange Act of 1934 (the “1934 Act”) together with Rule 10b-5 promulgated thereunder, and Section 17(a) of the Securities Act of 1933 (the “1933 Act”). Bertoli now appeals. Supported by various amici, he maintains that sentencing findings should never be given preclusive effect in civil litigation. Alternatively, Bertoli argues that even if collateral estop-pel could extend to sentencing findings, application of the doctrine in this case was inappropriate. While we decline to adopt the sweeping per se prohibition urged by Bertoli, we conclude that the application of collateral estoppel in this case was improper. Accordingly, we vacate and remand for further proceedings. BACKGROUND We recount only the facts necessary to this appeal. 1 The SEC originally filed this civil suit in September 1985 in the Southern District of New York (Sand, J.). See S.E.C. v. Monarch Funding Corp., 85 Civ. 7072 (S.D.N.Y.) (LBS). The Commission alleged that Bertoli and others violated Section 10(b) of the 1934 Act, Rule 10b-5, and Section 17(a)(1)-(3) of the 1933 Act in connection with his involvement in the activities of Monarch Funding Corporation (“Monarch”), a securities brokerage firm in New York City. The SEC sought disgorgement of Bertoli’s ill-gotten gains and a permanent injunction enjoining Bertoli from future violations of those laws. *299 The civil action was placed on Judge Sand’s suspense calendar pending the outcome of a related criminal prosecution in the United States District Court for the District of New Jersey (Lechner J.). See United States v. Richard O. Bertoli, 854 F.Supp. 975 (D.N.J.1989). That action was prosecuted by the New Jersey United States Attorney’s Office with “a high degree of consultation and coordination” with the SEC. S.E.C. v. Monarch Funding Corp., 983 F.Supp. 442, 455 (S.D.N.Y.1997). In the criminal action, Bertoli and his two co-defendants, Richard Cannistraro and Leo Eisenberg, faced a multiple count indictment. Of particular relevance to this appeal are Counts One and Two, charging Bertoli with RICO violations based on a pattern of racketeering activity involving, inter alia, predicate violations of Section 10(b) and Rule 10b-5. Also relevant to this appeal are Count Three, which charged Bertoli with conspiracy to obstruct justice in violation of 18 U.S.C. § 371, based on his attempts to obstruct the various civil, grand jury and criminal proceedings arising from his alleged fraud, and Count Six accusing him of obstructing - justice in violation of 18 U.S.C. §§ 1502 and 1503 by moving certain proceeds of his racketeering activities from the Cayman Islands to the principality of Andorra in Europe. After Eisenberg and Cannistraro pled guilty, Bertoli was tried before a jury in the summer of 1993. Following a three-month trial, Bertoli was convicted of the obstruction of justice charges contained in Counts Three and Six. He was acquitted, however, on all other charges, including Counts One and Two, which charged the predicate securities fraud violations. A. The Original Sentencing Following Bertoli’s conviction, the parties engaged in extensive litigation over various post-trial and sentencing issues. On March 28, 1994, Judge Lechner sentenced Bertoli, issuing a 189-page opinion two days later to, inter alia, “clarify and amplify the rulings made during ... sentencing.” United States v. Bertoli, 854 F.Supp. 975, 1164 (D.N.J.1994) (“Bertoli I”). Judge Lechner began the sentencing portions of that opinion by recounting Ber-toli’s long and diverse history of legal troubles. Included in those troubles were several run-ins with the SEC in the 1970s, which resulted in injunctions against Bertoli. See Bertoli I, 854 F.Supp. at 1127. Also noteworthy were Bertoli’s efforts to thwart a creditor through the fraudulent transfer of assets to his brother and minor children, see id. at 1127-28, as well as his ensuing personal bankruptcy, see id. at 1060. Next, Judge Lechner turned to the calculation of Bertoli’s sentence, using the 1993 version of the United States Sentencing Guidelines and accompanying Commentary. Specifically, Judge Lechner applied § 2J1.2, the Guideline applicable to obstruction of justice. That provision provides, in pertinent part that: “[i]f the offense involved obstructing the investigation or prosecution of a criminal offense, apply § 2X3.1....” Section 2X3.1, in turn, provides that , the offense level is to be calculated on the basis of the criminal conduct underlying the investigation obstructed by the defendant. And, according to Judge Lechner, § 2X3.1 applied even if the defendant was not actually convicted of that underlying criminal conduct. See Bertoli I, 854 F.Supp. at 1144-46. Judge Lechner found that the underlying criminal conduct that Bertoli had attempted to conceal was securities fraud— specifically, Bertoli’s orchestration of what Judge Lechner -termed the “Stock Manipulation Schemes.” To summarize, Judge Lechner found that in or about 1982 or 1983, Bertoli began to work at Monarch, the securities brokerage firm. See id. at 1128. Eisenberg, Monarch’s owner and president, permitted Bertoli to use the firm’s offices to promote and arrange secu *300 rities transactions on a “behind the scenes” basis. Id. Among the deals arranged by Bertoli was Monarch’s underwriting of the initial public offerings (“IPOs”) of three new issuers: Liquidation Control, Inc. (“LCI”), Toxic, Waste Containment, Inc. (“Toxic”), and High Technology Capital Corp. (“High Tech”). See id. In trading immediately following the IPOs, Bertoli orchestrated an artificial increase in the price of each stock through a series of controlled trades at successively higher prices. See id. at 1129. After the prices had risen to a certain level, unwitting members of the investing public were brought in to purchase the stocks. Cannistraro, an analyst at a New York City brokerage, helped recruit outside buyers by writing favorable research reports about the IPOs. See id. Unsurprisingly, the true nature of this fraudulent scheme was not disclosed in the public documents filed and issued in connection with the IPOs. See id. at 1129 n. 237. The scheme was hugely successful. Within a few months of their issuance, the shares of LCI, Toxic, and High Tech were trading at many times their offering prices. See id. at 1129. Eventually, Ber-toli and his cohorts pricked the bubble at the inflated prices, selling out for millions in profit. The stock prices then plummeted leaving the public investors with large losses. See id. Bertoli, Cannistraro, and Eisenberg initially stashed their profits in Cayman Island accounts, either in their own names or those of various companies they created to serve as repositories. Later, however, Bertoli transferred those profits to the tiny principality of Andorra. See id. at 1130, 1135. According to Judge Lechner, these findings established that Bertoli had committed securities fraud “by at least a preponderance of the evidence.” Id. at 1139. As a result, directed by the cross reference provision of § 2X3.1, Judge Lechner calculated Bertoli’s sentence using § 2F1.1—the Guideline applicable to fraud. Application of that Guideline yielded a total offense level of 28 and a Guideline range of 87 to 108 months. Judge Lechner sentenced Bertoli to two 100-month concurrent prison terms and three years supervised release. Judge Lechner also imposed, without explanation, a special condition of supervised release barring Bertoli from associating with any person involved in the securities industry. On appeal, the Third Circuit affirmed Bertoli’s conviction but vacated his sentence. See United States v. Bertoli, 40 F.3d 1384, 1401-11 (3d Cir.1994) (“Bertoli II”). The Third Circuit determined that by using the 1993 version of the Guidelines and Commentary, Judge Lechner imposed a sentence in excess of what was permissible under the 1989 version applicable, to Bertoli’s conduct, thereby violating his right against ex post facto punishment. See Bertoli II, 40 F.3d at 1407. B. The Resentencing On remand, the government moved for an enhancement of Bertoli’s sentence under Guideline § 2J1.2(b)(2), which Judge Lechner had been prevented from applying at the original sentencing because he had chosen to apply the mutually exclusive § 2X3.1 provision. Section 2J1.2(b)(2) allows for a three-level increase for obstruction of justice convictions if the offense “resulted in substantial interference with the administration of justice.” After holding a hearing, Judge Lechner issued an opinion expressly incorporating by reference his factual findings from Bertoli I. See United States v. Bertoli, 89 CR 218 (D.N.J. March 22, 1995) (“Bertoli III”). In the second opinion, Judge Lechner found that Bertoli moved assets, induced an important witness to evade service of process, filed false affidavits, and shredded documents, all in an effort to interfere with the government’s investigation into Monarch’s activities. Judge Lechner further found that Bertoli had directed two key witnesses to lie to SEC investigators. More specifically, Judge Lechner found *301 that Bertoli caused Robert Cooper, a stock broker and a player in the fraudulent LCI IPO, to claim falsely that: (1) he, not Cannistraro, wrote a favorable research report on LCI based on analysis of the company and discussions with its officers; and (2) Bertoli had never discussed LCI with him and had no involvement in the drafting of the research report. See Bertoli III at 19-21 (referring to Bertoli I, 854 F.Supp. at 1131-32). Similarly, Judge Lechner found that Bertoli instructed Ei-senberg to “do whatever he had to do” to protect himself and Bertoli, and that Ei-senberg responded by lying to the SEC regarding Monarch’s trading in LCI and Toxic securities. Id. (referring to Bertoli I, 854 F.Supp. at 1131-32). These activities were, according to Judge Lechner, “critical components of Bertoli’s eight year conspiracy to cover up and conceal his illegal fraudulent trading schemes.” Id. at 19. Based on these findings, Judge Lechner concluded that Bertoli had substantially interfered with justice under § 2J1.2(b)(2) with respect to each of three sentencing “Groups” he had previously established for purposes of calculating Bertoli’s offense level. See Bertoli I, 854 F.Supp. at 1138-1144 (explaining “grouping” methodology under U.S.S.G. §§ 1B1.2(d), 3D1.1, and 3D1.2). Imposing the § 2J1.2(b)(2) enhancement within each sentencing Group yielded a total offense level of 25, which provided for a Guidelines range of 63 to 78 months imprisonment. Judge Lechner sentenced Bertoli to 78 months. In addition, Judge Lechner reimposed, again without explanation, the special condition of supervised release forbidding Bertoli from associating with any person involved in the securities industry. The Third Circuit affirmed Bertoli’s re-sentencing by summary order. C. Judge Sand’s Summary Judgment Rulings 1. The June 1996 decision Following Bertoli’s resentencing, the SEC moved for summary judgment in the pending civil action, arguing that Bertoli should be collaterally estopped by virtue of Judge Lechner’s sentencing findings from denying that he violated the federal securities laws. See S.E.C. v. Monarch Funding Corp., No. 85 Civ. 7072, 1996 WL 348209 at *3 (S.D.N.Y. June 24, 1996) (“Monarch I”). Judge Sand agreed with the SEC, holding that the elements of civil liability under Section 10(b), Rule 10b-5, and Section 17(a)(1)-(3) had been established by Judge Lechner’s sentencing findings. Accordingly, the court granted the SEC’s request for a permanent injunction enjoining Bertoli from committing future violations of those laws. Bertoli appealed. On the SEC’s motion, this Court remanded to allow the district court to consider Judge Lechner’s opinion in Bertoli III, which the SEC had failed to submit on the original motion for summary judgment. 2. The October 1997 decision On remand, the SEC renewed its motion for summary judgment on collateral estop-pel grounds. Bertoli responded by arguing that: (1) sentencing findings should never be given preclusive effect in subsequent civil litigation; and (2) even if they could, applying collateral estoppel in this case would be inappropriate. In a published opinion, Judge Sand rejected Bertoli’s arguments for a per se rule. See S.E.C. v. Monarch Funding Corp., 983 F.Supp. 442, 448 (S.D.N.Y.1997) (“Monarch II”). The court recognized the “potential dangers underlying the extension of collateral estoppel to sentencing findings.” Id. at 447. Judge Sand reasoned, however, that such dangers should not be dispositive so long as, after “close scrutiny” of the sentencing proceeding, it is clear that the “traditional safeguards” of the collateral estoppel doctrine as set forth by the Supreme Court and by this Court have been met. Id. at 446 (citing Park *302 lane Hosier Co. v. Shore, 439 U.S. 322, 99 S.Ct. 645, 58 L.Ed.2d 552 (1979), and Gelb v. Royal Globe Ins. Co., 798 F.2d 38 (2d Cir.1986)). Applying these traditional safeguards, the district court rejected Bertoli’s contention that he had been deprived of an adequate opportunity to litigate his securities fraud liability at sentencing. While acknowledging that the typical sentencing proceeding may afford inadequate procedural opportunities for purposes of collateral estoppel, Judge Sand found that the Bertoli sentencing was “anything but a garden-variety sentencing.” Id. at 458. In the nearly two-year period between Bertoli’s conviction and his ultimate sentencing, the parties made voluminous submissions and participated in at least two hearings. See id. at 449-50. These “protracted proceedings,” Judge Sand concluded, afforded Bertoli ample opportunity to challenge the government’s evidence supporting Judge Lechner’s sentencing findings. Id. at 458. Moreover, the court found that because the civil action antedated the criminal action, and in fact was suspended for the express purpose of allowing the criminal action to be decided first, Bertoli could not claim that the use of collateral estoppel was unforeseeable. See id. at 450, 458. Finally, Judge Sand rejected Bertoli’s complaints that Judge Lechner made his sentencing findings after considering two hearsay documents — namely, an affidavit of a government investigator, Michael J. Cahill; and the plea allocution of co-conspirator Can-nistraro. Judge Sand found that Judge Lechner’s consideration of these documents presented no obstacle to estoppel because, prior to resentencing, Judge Lechner held a hearing at which Bertoli actually called Cahill as a witness and could have called Cannistraro. Judge Sand also concluded that the findings of securities fraud were necessary to Bertoli’s sentencing. In reaching this conclusion, Judge Sand declined to rely on the findings made for Bertoli’s original sentencing because they had been deprived of all preclusive effect by virtue of the Third Circuit’s reversal in Bertoli II. See Monarch II, 983 F.Supp. at 451 & n. 9 (citing Buck v. United States Aviation Underwriters, Inc., 763 F.2d 224, 226-27 (6th Cir.1985)). Instead, Judge Sand relied on the resentencing findings made for purposes of the § 2J1.2(b)(2) enhancement for substantial interference with justice, concluding that they were necessary to Bertoli’s ultimate sentence on essentially two discrete theories. First, Judge Sand reasoned that Judge Lechner’s resentencing findings only made sense in the context of his prior finding that Bertoli had committed securities fraud. To Judge Sand, the findings made for purposes of the § 2J1.2(b)(2) enhancement were “significant only” because of the prior finding that Bertoli orchestrated the Stock Manipulation Schemes. Id. at 454. As such, Judge Sand concluded that “the adjustments to Groups One, Two and Three each presuppose the validity of factual findings that establish Bertoli’s liability for securities fraud.” Id. For his second necessity rationale, Judge Sand relied on the proposition that inferential determinations may provide the basis for collateral estoppel. See id. at 453 & n. 14. Applying that proposition, Judge Sand reasoned that the resentencing findings Judge Lechner necessarily made for purposes of the § 2J1.2(b)(2) enhancement, “also establish Bertoli’s liability for securities fraud.” Id. at 453 (citing Monarch I, 1996 WL 348209 at *7-8). As examples, Judge Sand pointed to Judge Lechner’s findings that Bertoli told Cooper and Eisenberg to lie about his involvement in the LCI report and Monarch’s trading activities. These findings were, according to Judge Sand, “necessary” because absent a finding that Bertoli was in fact involved in preparing the LCI report and Monarch’s activities, “Judge Lechner could not have properly concluded that Cooper and Eisenberg’s statements were mistruths *303 warranting an adjustment under § 2J1.2(b)(2).” Id. Based on this analysis, Judge Sand reaffirmed his original determination that: (1) Bertoli was collaterally estopped by Judge Lechner’s sentencing findings from denying liability under Section 10(b), Rule lob-5, and Section 17(a) of the securities laws; and (2) the SEC was entitled to a permanent injunction enjoining Bertoli from future violations of those laws. However, Judge Sand found that collateral estoppel did not apply to the SEC’s claim for disgorgement, and the SEC abandoned that claim. A final judgment was entered and Bertoli now appeals. DISCUSSION I. Jurisdiction Before reaching the substantive aspects of this appeal, we briefly address a jurisdictional challenge by Bertoli. He contends that after the SEC dropped its claim for disgorgement, the district court lost subject matter jurisdiction to issue an injunction in the civil action because that injunction would merely duplicate judgments already rendered against Bertoli in earlier proceedings. Therefore, he argues,there was no “case or controversy” as required by Article III, Section 2 of the United States Constitution. We disagree. Article III limits subject matter jurisdiction to “cases” and “controversies.” United States v. Probber, 170 F.3d 345, 347 (2d Cir.1999). Thus, an action which promises to merely duplicate the relief afforded by prior proceedings may, indeed, be non-justiciable. See Marshel v. AFW Fabric Corp., 552 F.2d 471, 472 (2d Cir.1977) (per curiam). As a factual matter, however, there is simply no such duplication in this case. The prior injunctions entered against Bertoli were: (1) a 1975 order enjoining him from violating the record-keeping provisions of Section 17(a) of the 1934 Act; (2) a 1979 SEC order barring him from associating with any broker dealers; and (3) Judge Lechner’s condition of supervised release that he not associate with anybody in the securities industry. None of these judgments duplicate the specific injunctive relief granted by Judge Sand in this case — namely, the permanent injunction enjoining Bertoli from violating the anti-fraud provisions of Section 10(b), Rule 10b-5, and Section, 17(a) of the 1933 Act. Judge Sand had jurisdiction to issue that injunction; and we have jurisdiction over this appeal. II. Offensive Collateral Estoppel Bertoli’s position is that: (1) sentencing findings should never merit preclusive effect; and (2) even if they should in a particular case, this is not that case. We decline to adopt a per se rule against extending the doctrine of offensive collateral estoppel to sentencing findings. We agree with Bertoli, however, that applying the doctrine in this case was inappropriate. We review a district court’s grant of summary judgment based on the doctrine of collateral estoppel de novo, construing the record in the light most favorable to the non-moving party and drawing all inferences in that party’s favor. See Boguslavsky v. Kaplan, 159 F.3d 715, 719 (2d Cir.1998). Under the doctrine of offensive collateral estoppel (more recently called offensive issue preclusion), a plaintiff may foreclose a defendant from relitigating an issue the defendant has previously litigated but lost against another plaintiff. See Parklane Hosiery Co. v. Shore, 439 U.S. 322, 329, 99 S.Ct. 645, 58 L.Ed.2d 552 (1979). The principal virtue of collateral estoppel is self-evident: it promotes judicial economy by reducing the burdens associated with revisiting an issue already decided. See id. at 326, 99 S.Ct. 645; Gelb, 798 F.2d at 44. Its virtues do not come without a price, however. Just as occasionally “the race is not to the swift, nor the battle to the strong ... but time and chance happeneth to them all,” Eccle *304 siastes 9:11 (King James ed.), so too the results of an earlier resolution of an issue may simply be wrong. See Johnson v. Watkins, 101 F.3d 792, 795 (2d Cir.1996) (citing 18 Charles A. Wright, Arthur R. Miller & Edward H. Cooper, Federal Practice and Procedure § 4416, at 142 (1981)). Ultimately, in allowing collateral estop-pel, courts have decided that “the occasional permanent encapsulation of a wrong result is a price worth paying to promote the worthy goals of ending disputes and avoiding repetitive litigation.” Johnson, 101 F.3d at 795; see Gelb, 798 F.2d at 44. When the efficiency rationale for collateral estoppel fails, however, courts have understandably declined to apply the doctrine. See, e.g., Davis v. West Community Hosp., 786 F.2d 677, 682 (5th Cir.1986) (district court properly declined to apply collateral estoppel because precluding relitigation “would not have furthered one of the important underlying purposes of this doctrine: promotion of judicial economy”); Swineford v. Snyder Co., 15 F.3d 1258, 1269 (3d Cir.1994). Although collateral estoppel jurisprudence generally places termination of litigation ahead of a correct result, there are some circumstances that so undermine confidence in the validity of an original determination as to render application of the doctrine impermissibly “unfair” to a defendant. Parklane, 439 U.S. at 330, 99 S.Ct. 645; see Remington Rand Corp. v. Amsterdam-Rotterdam Bank, N.V., 68 F.3d 1478, 1487 (2d Cir.1995). One such circumstance arises where the second action affords “procedural opportunities unavailable in the first action that could readily cause a different result.” Parklane, 439 U.S. at 330-31, 99 S.Ct. 645. Another exists where a defendant has little incentive to litigate the relevant issue vigorously in the original action, particularly if the second action is not foreseeable. See id. at 330, 99 S.Ct. 645. And of course, collateral estoppel may have a devastating impact on a civil litigant’s constitutional right to a jury trial. True, the Seventh Amendment poses no insurmountable barrier to applying collateral estoppel. See id. at 336, 99 S.Ct. 645 (holding that once a common factual issue has been resolved in a previous action (by a jury or otherwise), “there is no further factfinding function for the jury to perform”). However, this is true only where the issue was “fully and fairly ” adjudicated in the prior proceeding. Id. at 325, 99 S.Ct. 645 (emphasis added) (internal quotation marks omitted). To strike an appropriate balance between the competing concerns for fairness on the one hand, and efficiency on the other, courts have imposed a number of prerequisites to assure that the precluded issue, whether or not correctly resolved, was at least carefully considered in the first proceeding. See Gelb, 798 F.2d at 44; Johnson, 101 F.3d at 795. For the bar to apply: (1) the issues in both proceedings must be identical; (2) the issue in the prior proceeding must have been actually litigated and actually decided; (3) there must have been a full and fair opportunity for litigation in the prior proceeding; and (4) the issue previously litigated must have been necessary to support a valid and final judgment on the merits. See Gelb, 798 F.2d at 44; Boguslavsky, 159 F.3d at 720. A. Application of Collateral Estoppel to Sentencing Findings As the district court acknowledged, the application of ’ collateral estoppel to sentencing findings presents a novel issue. While a few cases have brushed against the question, they have done so without much elaboration leaving the darkness unobscured. See Allen v. City of Los Angeles, 92 F.3d 842, 850 (9th Cir.1996), overruled on other grounds by Acri v. Varian Assocs., Inc., 114 F.3d 999 (9th Cir.1997) (en banc); United States v. Montes, 976 F.2d 235, 239 (5th Cir.1992); M. Prusman Ltd. v. Ariel Maritime Chroup, Inc., 781 F.Supp. 248, 250, 252 (S.D.N.Y.1991). *305 Not surprisingly, the parties disagree on how this nettlesome issue should be resolved. For its part, the SEC assumes that collateral estoppel should, at the very least, be presumptively available in the sentencing context so long as the traditional safeguards of the doctrine are met. On the other hand, Bertoli and the various amici assert that sentencing findings should never be given preclusive effect. We chart a middle course. The arguments for a per se prohibition are attractive. As Bertoli maintains, such a rule is warranted because allowing sentencing findings to control subsequent litigation is simply “unfair” — as that term is used in Parklane — in two fundamental respects. First, a plenary civil trial affords a defendant procedural opportunities that are unavailable at sentencing and that could command a different result. Unlike a civil litigant, a criminal defendant’s opportunities to take discovery may be limited for sentencing purposes. See Wade v. United States, 504 U.S. 181, 186, 112 S.Ct. 1840, 118 L.Ed.2d 524 (1992). Although whenever a factor important to sentencing is reasonably in dispute, a defendant must be given an adequate opportunity to present information on that factor, see U.S.S.G. § 6A1.3(a), a defendant has no absolute right either to present his own witnesses or to receive a full-blown evidentiary hearing. See United States v. Prescott, 920 F.2d 139, 143 (2d Cir.1990). Moreover, the evidence admissible against a civil litigant must survive challenge under the Federal Rules of Evidence. By contrast, a sentencing judge “is largely unlimited either as to the kind of information [he] may consider, or the source from which it may come,” United States v. Sisti, 91 F.3d 305, 312 (2d Cir.1996) (internal quotation marks omitted) (alteration in original), so long as the information has sufficient indicia of reliability to support its probable accuracy, see U.S.S.G. § 6A1.3(a). Second, the incentive to litigate a sentencing finding is frequently less intense, and certainly more fraught with risk, than it would be for a full-blown civil trial. As Judge Sand acknowledged, a criminal defendant will often choose not to challenge sensitive issues during sentencing for any number of reasons, including a belief, or at least a hope, that the sentencing court will grant a prosecutorial downward departure motion or other recommendation. In addition, sometimes sentencing procedures may deter a defendant from raising certain issues. See Monarch II, 983 F.Supp. at 448. For instance, given that a sentencing judge may make a finding based on evidence inadmissible in a civil trial without even giving the defendant an evidentiary hearing to challenge that evidence, see Prescott, 920 F.2d at 144, a defendant may choose to let sleeping dogs lie. Finally, a defendant, though uniquely knowledgeable about underlying events, may be reluctant to testify during sentencing. Again, there are a number of reasons for such reticence — not the least of which is that if the defendant is disbelieved, his sentence may be enhanced under Guideline § 3C1.1. In light of these legitimate concerns over potential unfairness, it is important to reflect upon whether the efficiency rationale for collateral estoppel would be advanced or hindered, were the doctrine to be freely available in the sentencing context. After all, it makes little sense to forego the opportunity to reexamine a potentially wrong decision if the economies achieved by doing so are slight or nonexistent. See Johnson, 101 F.3d at 795. In our view, several considerations suggest that, despite its theoretical economies, applying collateral estoppel in the sentencing context will just as often multiply, rather than reduce, total litigation. First, where a civil action is pending or just beyond the horizon, allowing sentencing findings to earn collateral estoppel respect may greatly increase the stakes at sentencing, producing more exhaustive litigation over matters of only tangential importance to the criminal case. This risk is *306 exacerbated by the procedural looseness of sentencing litigation. There is virtually no limit on the extent and character of the evidence that the government may seek to introduce at sentencing. See Sisti, 91 F.3d at 312. Even if the sentencing judge is able to fend off all attempts to introduce gratuitous material, the additional burdens that such an approach would impose on our already over-worked district courts would change the essential nature of sentencing proceedings. They are not supposed to be “mini-trials.” United States v. Pugliese, 805 F.2d 1117, 1123 (2d Cir.1986); see Prescott, 920 F.2d at 144. Second, while a permissive approach to collateral estoppel will probably lead to sentencing proceedings of mushrooming complexity, ironically, there is no guarantee that subsequent civil actions will be made proportionately simpler. As Judge Sand acknowledged, the concerns of unfairness raised by Bertoli require a more cautious approach in the civil case—in his words a “close scrutiny” and “searching examination”—than would be otherwise necessary. See Monarch II, 983 F.Supp. at 446, 448. These practical considerations undermine a primary rationale for allowing collateral estoppel in the first place. If the economies achieved by applying collateral estoppel are not readily apparent, why risk the permanent encapsulation of a wrong result? See Johnson, 101 F.3d at 795. And that risk is especially acute with respect to sentencing proceedings, in which, for the reasons already mentioned, the government’s factual assertions may well be subjected to less adversarial testing. While extending collateral estoppel effect to sentencing findings may, in a given case, threaten fairness and/or judicial efficiency, these factors do not justify the blanket prohibition urged by Bertoli. Generally speaking, the same concerns of unfairness and inefficiency were considered by the Parklane Court and were rejected as justifications for a total ban on the use of offensive collateral estoppel. See 439 U.S. at 330-31, 99 S.Ct. 645. Instead, the Court held that the “the preferable approach” is to entrust the task of minimizing such dangers to the discretion of our district courts. Id. at 331, 99 S.Ct. 645. And although such dangers may be more pronounced in the sentencing context, we are confident that our trial judges will be able to limit their impact on a case-by-case basis. So long as the threat to fairness and/or efficiency has been minimized, we see no need to entirely foreclose application of the doctrine. Indeed, precluding relitigation of findings made during sentencing may promote an institutional goal of particular importance to the criminal process, namely, “preserving] the integrity of the judicial system by eliminating inconsistent results.” Johnson, 101 F.3d at 795; see 18 Federal Practice and Procedure § 4416, at 139 (1981) (“preclusion prevents the risk of the clearest and most embarrassing inconsistencies”). For these reasons, we reject the per se rule urged by Bertoli and the amici On the other hand, we cannot accept the SEC’s position that collateral estoppel should presumptively extend to sentencing findings on the same basis as in other contexts. To the contrary, we conclude that precluding relitigation on the basis of such findings should be presumed improper. While we do not foreclose application of the doctrine in all sentencing cases, we caution that it should be applied only in those circumstances where it is clearly fair and efficient to do so. And the burden should be on the plaintiff in the civil case to prove these elements. B. The Instant Case In this case, the SEC has failed to show that preclusion was fair, as that concept is refined in Gelb. See 798 F.2d at 44-45. In addition, the record reveals that application of collateral estoppel here was simply not efficient. For these reasons, we conclude that Judge Sand erred in precluding Bertoli from litigating his securities fraud liability afresh. *307 As noted above, to minimize the potential unfairness caused by applying collateral estoppel to a finding that may be wrong, it is essential that the finding have been “necessary” to the judgment in the first action. . See id. As explained by Judge Learned Hand, the necessity requirement “proteet[s]” against unfairness, by ensuring that the issue will be “really disputed and that the loser will have put out his best efforts.” The Evergreens v. Nunan, 141 F.2d 927, 929 (2d Cir.1944). We recognize that in other contexts there is authority for relaxing the necessity requirement. As Judge Sand pointed out, determinations that can be inferred from necessary findings have sometimes provided a basis for estoppel. See Monarch II, 983 F.Supp. at 454 & 453 n. 14 (citing In re Mirulla, 163 B.R. 912, 916-17 (Bankr.D.N.H.1994) (applying collateral estoppel where “[although it is true that ‘malice’ was not per se found by the state court, the written decision of the state court nevertheless is replete with factual findings from which it can be inferred”); and Greenblatt v. Drexel Burnham Lambert, Inc., 763 F.2d 1352, 1361-62 (11th Cir.1985)). However, given the potential for unfairness associated with applying collateral estoppel based on sentencing findings, we hold that only the stricter approach remains appropriate in the sentencing context. Hence, regardless of how carefully considered an issue may have been during the process leading up to decision, and regardless of what may be inferred from that decision, estoppel does not apply to a finding that was not legally necessary to the final sentence. See 18 Charles A. Wright, Arthur R. Miller & Edward H. Cooper, Federal Practice and Procedure § 4421, at 196 (1981). Throughout proceedings in this civil case, Bertoli’s primary argument has been that the finding that he committed securities fraud was not necessary to’ the ultimate sentence imposed in Bertoli III. Three separate bases for necessity have been offered in response to this contention. First, there are Judge Lechner’s findings of securities fraud in Bertoli I, which were expressly incorporated by reference into Bertoli III. According to the SEC, these prior fraud findings were necessary to Bertoli III because the § 2J1.2(b)(2) enhancement for substantial interference with the administration of justice, “necessarily rested” on such findings. Second, there are the obstructive act findings that were necessary to the § 2J1.2(b)(2) enhancement imposed in Bertoli III. According to Judge Sand, these same obstructive act findings also established Bertoli’s liability for securities fraud. Third, the SEC contends that the securities fraud findings were necessary to the imposition of the special condition of Bertoli’s supervised release that he was not to associate with any persons involved in the securities industry. We are unpersuaded by any of these responses. The SEC’s assertion that the § 2J1.2(b)(2) enhancement “necessarily rested” on the prior finding of securities fraud from Bertoli I, is in essence a contention that Bertoli could not have interfered with the investigation into his securities fraud, unless there was such a fraud to cover up in the first place. As the SEC puts it, “it was not thé obstructive acts themselves, but Bertoli’s attempt through those obstructive acts to cover up his role in the fraud that warranted the enhancement for substantial interference with the administration of justice.” This argument improperly confuses why Bertoli may have felt it necessary to -interfere with the administration of justice, with what was legally necessary to find before concluding that he did so. To start with, concealment of his alleged securities fraud need not have been the motive behind Bertoli’s various acts of obstruction at all. There is a host of competing explanations as to why a former bankrupt would attempt to hide assets, or why a man subject to an order barring him from associating with any broker dealer would seek to shield his involvement with *308 Monarch. These plausible alternative explanations for Bertoli’s obstructive acts reinforce why, as a matter of fairness, it is so important to focus not on necessity from the defendant’s subjective perspective, but on the legal necessity of a finding to the ultimate judgment. And in this case, it was not legally necessary for Judge Lechner to find that Ber-toli committed securities fraud to impose the § 2J1.2(b)(2) enhancement. It is true that in Bertoli I Judge Lechner found that Bertoli committed securities fraud by orchestrating the Stock Manipulation Schemes, and that that finding was expressly incorporated by reference into the Bertoli III decision. It is also true that in Bertoli III, Judge Lechner found that Bertoli “substantially interfered with the administration of justice” under § 2J1.2(b)(2), to “cover up” his securities fraud. This does not mean, however, that it is legally necessary for a defendant to be found guilty of securities fraud to impose the §• 2J1.2(b)(2) enhancement. That obviously cannot be so, for one could be held responsible for substantially interfering with an investigation. within the meaning of § 2J1.2(b)(2), without having been involved in the conduct investigated at all. Thus, while Judge Lechner’s findings that Bertoli induced perjury, attempted to conceal assets and shredded documents were necessary to the § 2J1.2(b)(2) enhancement, the finding that these nefarious activities were part of a securities fraud was not. Nor do we agree with Judge Sand that the same obstructive act findings that were necessary to the § 2J1.2(b)(2) enhancement, “also establish Bertoli’s liability for securities fraud.” See Monarch II, 983 F.Supp. at 453. The problem with this reasoning is that it infers one congeries of legal determinations—that Bertoli violated Section 10b, Rule 10b-5, and Section 17(a)—from facts that were necessarily found to support another—that Bertoli interfered with the administration of justice under § 2J1.2(b)(2). This case provides a textbook illustration of how such a process of inference can produce a potentially unfair result. To have violated Section 10(b) and Rule 10b-5, Bertoli must have: (1) made a material misrepresentation or a material omission as to which he had a duty to speak, or used a fraudulent device; (2) with scienter; (3) in connection with the purchase or sale of securities. See S.E.C. v. First Jersey Securities, Inc., 101 F.3d 1450, 1467 (2d Cir.1996). Essentially the same elements are required under Section 17(a)(1)-(3) in connection with the offer or sale of a security, though no showing of scienter is required for the SEC to obtain an injunction under subsections (a)(2) or (a)(3). See id. (citing Aaron v. SEC, 446 U.S. 680, 701-02, 100 S.Ct. 1945, 64 L.Ed.2d 611 (1980)). While Judge Lechner’s findings regarding Bertoli’s various obstructive acts may have been necessary to the § 2J1.2(b)(2) enhancement, contrary to Judge Sand’s conclusion, none of those findings establish the elements of securities fraud. Bertoli’s various efforts to hide assets clearly do not do so. Similarly, while Judge Lechner found that Bertoli told Eisenberg to lie about his involvement with Monarch, that finding alone does not establish all the elements of a Section 10b, Rule 10b-5 or Section 17(a) violation. And although Ber-toli may have overseen the drafting of the LCI report, that finding does not establish that the LCI report was materially misleading. See First Jersey Securities, 101 F.3d at 1467. We also reject the third necessity argument advanced in this case—the SEC’s contention that the securities fraud findings were necessary to the special condition of Bertoli’s release that he not associate with any persons involved in the securities industry. As a threshold matter, we note that the SEC failed to raise this argument in the district court. The general rule is that “a federal appellate court does not consider an issue not passed upon below.” *309 Austin v. Healey, 5 F.3d 598, 601 (2d Cir.1993) (internal quotation marks omitted) (quoting Singleton v. Wulff, 428 U.S. 106, 120, 96 S.Ct. 2868, 49 L.Ed.2d 826 (1976)). We see no reason to eschew application of that general rule in this case. In any event, even if we were to exercise our considerable discretion and consider this argument, we would find it to be unpersuasive. For a finding to merit estoppel effect it must not only be necessary to the final judgment, but must also have been actually litigated and actually decided in the initial action. See Gelb, 798 F.2d at 44. Again, the actual litigation and actual decision prerequisites help ensure that a finding was carefully considered in the first action, and that it therefore may serve as a fair basis for estoppel. See id.; Johnson, 101 F.3d at 795. A corollary of the actual litigation and decision requirements is that “[w]hen a court cannot ascertain what was litigated and decided, issue preclusion cannot operate.” 18 James W. Moore et al, Moore’s Federal Practice, § 132.03[2][g] (3d ed.1998) (citing Mitchell v. Humana Hosp.-Shoals, 942 F.2d 1581, 1583-84 (11th Cir.1991)); see also id. § 132.03[3][d], In light of these principles, we need not address the potentially far reaching issue of whether it was necessary for Judge Lechner to find that Bertoli committed securities fraud before imposing the special condition. Throughout sentencing proceedings, the special condition of supervised release received little attention from either the parties or the court. Indeed, the prospect that Bertoli would be banned from further participation in the securities industry is not even mentioned in either of Judge Lechner’s lengthy sentencing opinions, raising the question of whether it was ever really litigated. Moreover, Judge Lechner never specifically ruled that he was banning Bertoli from the securities industry because Ber-toli had violated the securities laws. • Indeed, there may have been alternative reasons for that ban which did not hinge on the finding that Bertoli committed securities fraud. After all, Bertoli was convicted of obstruction of justice, not securities fraud, and it is possible that Judge Lech-ner imposed the ban on the basis of the obstruction of justice alone. See U.S.S.G. § 5F1.5(a) (court may impose occupational restrictions only if “a reasonably direct relationship existed between the defendant’s occupation ... and the conduct relevant to the offense of conviction; ” and (2) the “restriction is reasonably necessary to protect the public [from continued] unlawful conduct similar to that for which the defendant was convicted.” (emphasis added)). In light of the fact that Bertoli’s conviction for obstruction of justice arose from his interference with investigations focusing on securities fraud, Judge Lech-ner may have believed that banning him from the securities industry was “reasonably necessary to protect the public,” id., even if Bertoli himself had not committed securities fraud, simply because he had obstructed the government’s efforts to police that industry. We cannot rule that option out because Judge Lechner gave no specific reasons for imposing the ban. And in the absence of clarity on this issue, applying collateral estoppel on the basis of the special condition would be improper. See Mitchell, 942 F.2d at 1584 (refusing to apply collateral estoppel “[b]ecause the state court did not specify the reasons for its decision, [and] we cannot be certain if the court actually decided” the issue on the ground that would allow preclusion); Fay v. South Colonie Cent. Sch. Dist., 802 F.2d 21, 30 (2d Cir.1986) (denying preclusion where underlying decision was “vague”). In sum, application of collateral estoppel in this case failed to satisfy the prerequisites to preclusion that are designed to ensure fairness. In addition, we conclude that precluding relitigation in this case failed to promote the primary rationale for collateral estop-pel, namely: judicial economy. The sentencing proceedings in the criminal action were exhaustive. As noted, there was a *310 high degree of cooperation between the SEC and the New Jersey United States Attorney’s Office in the criminal action. Given that cooperation, we cannot say that the criminal action was not complicated by efforts to have express findings made on tangential issues at sentencing in order to give the SEC a chance to invoke collateral estoppel in the subsequent civil action. More importantly, however, collateral es-toppel did not do much to simplify the civil action. Even discounting the added complexities associated with the novelty of the issue, the district court’s “close scrutiny” and “searching examination” in this case required considerable effort—in all probability more effort than would have been required for a summary adjudication under Federal Rule of Civil Procedure 56(c), or even for a trial. In light of such efforts, it made little sense to bar Bertoli from litigating his securities fraud liability ab initio. See Davis, 786 F.2d at 682. We raise this point to make clear that in determining whether to apply collateral estoppel to sentencing findings in the future, district courts should start by making a threshold assessment of whether it will be efficient to do so. Given the potential unfairness associated with extending collateral estoppel to sentencing findings generally, if the court reasonably determines that the doctrine will not promote efficiency, it should feel free to deny preclusion for that reason alone. CONCLUSION The decision of the district court granting summary judgment based on the doctrine of offensive collateral estoppel is vacated. This case is remanded for further proceedings. Footnotes [1] For a comprehensive account of the prior proceedings in this case, as well as the related criminal cases, see S.E.C. v. Monarch Funding Corp., 983 F.Supp. 442 (S.D.N.Y.1997); United States v. Bertoli, 854 F.Supp. 975 (D.N.J.), aff'd in part, vacated in part, 40 F.3d 1384 (3d Cir.1994); United States v. Cannistraro, 800 F.Supp. 30 (D.N.J.1992); United States v. Eisenberg, 773 F.Supp. 662 (D.N.J.1991).