concurring in part, and dissenting in part:
I agree with the conclusion in Parts I and II of the majority opinion that the acquittal of a principal does not, as a matter of statutory interpretation, necessarily preclude the conviction of one charged with aiding and abetting. I cannot, however, accept the majority’s reluctance to acknowledge a role for non-mutual collateral estop-pel in criminal cases. Moreover, I am persuaded that this case may well be an appropriate instance in which to apply it.1 Indeed, the majority’s statutory analysis, eliminating the distinction between principals and aiders and abettors, and thus treating Standefer as if he were Niederberger, logically points to the application in Stande-fer’s favor of the collateral estoppel that would be available to Niederberger.
In Hubbard v. Hatrak, 588 F.2d 414 (3d Cir. 1978), this court held that non-mutual collateral estoppel was not required in criminal cases by the Due Process Clause of the Constitution. Accordingly, we noted, in the absence of state law to the contrary, appellant Hubbard could not attack his conviction for felony-murder on the ground that his accomplice had previously been convicted simply of second-degree murder. Nevertheless, we recognized that there is a significant interest in the appearance of fairness, an interest which may often be vindicated by a policy of collateral estoppel.
There is much to be said, in the criminal law context, for associating the doctrine of collateral estoppel with the principles of due process. Plainly, the appearance of evenhandedness in the administration of justice weighs heavily among our jurisprudential concerns, and estoppel is directly addressed to that appearance. For that reason thoughtful observers probably will applaud the introduction of non-mutual collateral estoppel into New Jersey criminal law.
588 F.2d at 417-18. Thus, while we could not say that the “appearance of evenhand-edness in the administration of justice” was of constitutional dimensions, we acknowledged that it “weighs heavily among our jurisprudential concerns.”
A concern for the appearance of even-handedness in the administration of justice persuades me that where the government has truly had a full and fair opportunity to prove its case, any fact which was necessarily determined against it should be applied in a second defendant’s favor in a new trial.2 The majority, on the other hand, seems to use a different scale in gauging the weight of this fairness concern.
*1109We agree that confidence in the integrity of the criminal justice system is, to some extent, diminished when, one criminal escapes punishment while another — particularly one charged as a result of acts performed by the acquitted defendant — is convicted and sentenced.
Majority Opinion at 1093. Nevertheless, it adds, “we do not believe the ‘appearance of evenhandedness’ is so overriding a concern that it should, as a general matter, warrant the same type of protection as is mandated by the Constitution.” Id. Many of the reasons marshalled by the majority for relegating collateral estoppel virtually to a legal oblivion are beside the point. First, the opinion notes, “a large number of acquittals result from the enforcement of evidentiary and constitutional rules that prohibit the prosecution from proving as much as it might were it completely unfettered.” Id. at 1094. This, however, is not an argument against the application of non-mutual estoppel in all criminal cases; rather, it is a reason why great care must be exercised in deciding whether the government has truly had a full and fair opportunity to prove its case, and whether a particular fact was necessarily determined against it in the prior litigation. Some, but not all, constitutional and evidentiary rules do deprive the government of a “full and fair” opportunity to prove its case. Where, for example, the defendant in the first trial claimed the protections of the fifth amendment,' the government has been foreclosed from establishing as much as it legitimately might in the absence of such a claim. The fifth amendment vindicates a uniquely personal right of privacy, see generally Note, Formalism, Legal Realism, and Constitutionally Proteeted Privacy Under the Fourth and Fifth Amendments, 90 Harv.L.Rev. 945 (1977), and thus a second defendant is not entitled to the collateral estoppel benefits of the first defendant’s personal right. On the other hand, where the first acquittal was the result of the government’s inability to introduce unlawfully seized evidence, a second defendant, as to whom that evidence would also be relevant — and perhaps even independently admissible — might be permitted to assert an estoppel defense. The issue, of course, is not free from doubt and I leave for another time the question of whether fourth amendment requirements, together with other constitutional and evi-dentiary restrictions, effectively preclude the government from a full and fair opportunity to prove its case. That is, unless the majority’s apparent aversion to collateral estoppel forecloses further consideration of these issues.
Second, the majority opposes non-mutual collateral estoppel because it suspects that the government may refuse "to prosecute some co-defendants at all for fear of jeopardizing stronger cases yet to be tried, or convictions already obtained.” Majority Opinion at 1094. If subsequent cases are indeed stronger, because of different procedural opportunities or newly available evidence, collateral estoppel treatment need not be given to the first decision.3 It is only where the government has had a truly full and fair opportunity to press its position that a subsequent bite at the apple is precluded. Moreover, adherence to the policies of collateral estoppel does not require a court to overturn previous convictions because of subsequent acquittals. While the principle of evenhandedness is undoubtedly *1110implicated in such inconsistent verdicts, a respect for the finality of judgments— equally essential to the doctrine of collateral estoppel — requires leaving well enough alone. I have no trouble endorsing a future-focused policy of estoppel, barring the government from repeated efforts at establishing the same fact or facts, while at the same time opposing a backward-focused policy which would make a nullity of prior judgments. The issue is whether the government, having lost once, should be given a second opportunity. The majority’s discussion of retroactivity, introduced for the purpose of injecting the emotional issue of collateral attacks on criminal judgments, is a straw man argument.
The majority is also uncomfortable with non-mutual collateral estoppel in criminal cases because it might allow a defendant to plead as an established fact what may have been the result of doubt, confusion or compromise. But this argument surely proves too much. As the Second Circuit recently observed:
[T]he Government suggests that the jury may have compromised, and hence what was actually decided can never be known. But this assertion is made in the teeth of clear case law that the possibility that the jury acquitted out of a desire to compromise or to show mercy, or from “simple frustration after hours of tedious debate,” United States v. Flowers, 255 F.Supp. 485, 487 (E.D.N.C.1966); see, e. g., Cosgrove v. United States, 224 F.2d 146, 154 (9th Cir. 1955), is not a basis for refusing to apply collateral estoppel. A contrary rule would, of course, eviscerate the doctrine altogether, for no one who is not present during the jury’s deliberations can ever know precisely how the jury reached its verdict.
United States v. Mespoulede, 597 F.2d 329 (2d Cir. 1979). Moreover, the majority’s contention — that the previous verdict may have been compromised or confused — would seem to bar estoppel, not only when it is a second defendant who seeks the benefits of that verdict, but also in instances where the defendant in the first trial is tried on a new count or related charge. Yet, collateral es-toppel is plainly available — indeed constitutionally required — in these latter cases, despite the possibility of compromise or confusion. See Ashe v. Swenson, 397 U.S. 436, 90 S.Ct. 1189, 25 L.Ed.2d 469 (1970); United States v. Venable, 585 F.2d 71 (3d Cir. 1978).
The majority also suggests that the anti-harassment aspect of double jeopardy disappears in the non-mutual context because Standefer has not been tried before. Majority Opinion at 1092. The difference between mutual and non-mutual collateral estoppel here, however, is one of degree. To suggest that Standefer, though not a party to the Niederberger trial, was not subjected to an accusation that he bribed Niederberger, is to elevate form over substance. Standefer’s alleged conduct certainly was exposed to public scrutiny and his reputation tarnished.
Finally, the majority notes that “[t]he precise determination of what a jury decided and why can be particularly difficult. In non-jury cases an appellate court is presented with the decision of a trial judge and an explanation of his reasoning. But non-jury verdicts are the exception rather than the rule.” Majority Opinion at 1095. Once again, this argument, if pursued to its logical extreme, would prohibit collateral estoppel, not only in cases like this one, but also in double jeopardy cases. Yet, the Supreme Court has, in the latter context at least, expressly enjoined trial courts to “ ‘examine the record of a prior proceeding, taking into account the pleadings, evidence, charge, and other relevant matter, and conclude whether a rational jury could have grounded its verdict upon an issue other than that which the defendant seeks to foreclose from consideration.’ ” Ashe v. Swenson, 397 U.S. at 444, 90 S.Ct. at 1194 (quoting Mayers & Yarbrough, Bis Vexari: New Trials and Successive Prosecutions, 74 Harv.L.Rev. 1, 38-39 (1960)). In so doing, courts are not to be bound by the “hyper-technical and archaic approach of a 19th century pleading book, but with realism and rationality.” Ashe, 397 U.S. at 444, 90 S.Ct. at 1194. Surely the capacity of courts to *1111pursue the inquiry into what was finally determined by a prior verdict does not dissolve outside the narrow range of double jeopardy cases. We do it in civil cases tried to a jury with relatively little difficulty.
In the final analysis, I suppose I simply take more seriously than the majority the perception of unfairness inherent in the disparate treatment of similarly situated defendants. I do not think the public is oblivious to the marked differences in legal and financial resources that exist between separate defendants and which often result, for all intents and purposes, in an acquittal for one and a conviction for another. And I agree with Judge Aldisert that “the public will not readily understand the court’s holding that a revenue agent did not receive a gift but a private citizen helped him receive the gift he did not receive.” Concurring and Dissenting Opinion at 1105. Even-handedness cannot, of course, be pursued in isolation from the legal obstacles which preclude a full and fair hearing for the government; neither can it blind us to the fact that, very often, it will be impossible to discern what a given adjudication finally determined. However, evenhandedness counts for something, and the majority cannot wish it away on the strength of rationales that suggest abrogation of the doctrine in the double jeopardy, and in many civil contexts as well.4
Believing as I do that non-mutual collateral estoppel should occasionally be available in criminal cases, it seems to me useful to sketch briefly how the doctrine should operate procedurally. In United States v. Inmon, 568 F.2d 326, 330 (3d Cir. 1977), we noted that double jeopardy is a defense which must be pleaded by the defendant. Since a defendant’s interest in avoiding double jeopardy is a matter of constitutional right, it follows that where double jeopardy is not involved, and the defendant’s interest in collateral estoppel is strictly non-constitutional, the defendant must also come forward and put in issue the question of estoppel. The government may then present evidence tending to show that it lacked a full and fair opportunity to prove its case the first time, or that the fact which the defendant seeks to foreclose from relitigation was not actually determined in the first proceeding. The ultimate burden of establishing that the fact in question was actually determined in the first litigation rests with the defendant, who seeks the protections of estoppel. See, e. g., United States v. Seijo, 537 F.2d 694, 697 (2d Cir. 1976), cert. denied, 429 U.S. 1043, 97 S.Ct. 745, 50 L.Ed.2d 756 (1977); United States v. Davis, 460 F.2d 792, 796 (4th Cir. 1972); United States v. Smith, 446 F.2d 200, 202-OS (4th Cir. 1971). The government, on the other hand, retains the burden of establishing that it did not receive a full and fair opportunity to litigate the issue sought to be foreclosed.
Once the pleadings are before the trial court, that court must decide whether the issue was actually determined in the first trial, and whether the government had a fair opportunity to establish its version. In discharging this function, trial courts would, pursuant to the Ashe Court’s injunction, survey the record of the first trial and determine whether a rational jury could have predicated its decision on a basis other than that which the defendant seeks to foreclose. The trial judge is invested with great discretion in performing this task. He must, in addition, consider any argu-*1112merits offered by the government which suggest why a full and fair opportunity to establish the facts was not available in the first trial. I have suggested above how some of these latter concerns might be resolved.5
In the instant case, Standefer did come forward and put in issue the estoppel effect of Niederberger’s acquittal on several counts. See Appendix at 17a — 19a (motion to dismiss). While his contention could have been more artfully presented, I think, at the very least, he has preserved the estoppel question for purposes of review. Because Standefer’s estoppel claim was imprecise, the trial court did not rule on the motion. Thus, to the extent that estoppel is arguably available in this case, I believe a remand to the trial court is required.
Standefer contended that a judgment of acquittal should have been entered on COUNTS 1, 3, and 5 of the indictment against him. Count 1 charged a violation of 26 U.S.C. § 7214(a)(2) in connection with the alleged Pompano Beach gratuity; Count 3 charged a violation of the same statute in connection with the Doral Country Club; Count 5 charged a violation of the same statute in connection with the Seaview Country Club. Because Niederberger had been acquitted of these same charges, Standefer sought to foreclose relitigation of the charges in his trial. As I noted above, the trial court did not reach the estoppel question. As to Counts 3 and 5, however, this was harmless error. While the first jury did acquit Niederberger under 26 U.S.C. § 7214(a)(2) of the Doral and Seaview offenses, it nevertheless convicted him of those same transactions under 18 U.S.C. § 201(g). These inconsistent verdicts perhaps reflected a misconception on the jury’s part of the appropriate legal standards for liability under the two statutes. Whatever the reason for the jury’s confusion, however, such inconsistent verdicts cannot serve as the basis for collateral estoppel. See, e. g., Harary v. Blumenthal, 555 F.2d 1113, 1116-17 (2d Cir. 1977).
The Pompano Beach violation in Count 1, however, cannot be similarly dispatched. Niederberger was acquitted under both statutes for the Pompano Beach violation. As the majority notes, Niederberger testified in his own behalf, and there is no showing on this record why collateral estop-pel is not at least arguably appropriate. Accordingly, I would remand for district court findings on this Count.
The majority, in its effort to play both sides against the middle, contends that even if non-mutual collateral estoppel may occasionally be recognized, this is not the case in which to do so. The majority reasons that the inconsistent verdicts effectively undermine the preclusive effect of the entire Niederberger verdict, even with respect to the Pompano Beach Count. I disagree. Whatever inconsistencies were inherent in the other charges against Niederberger, the findings with respect to Pompano Beach were thoroughly consistent. In Rogers v. LaVallee, 517 F.2d 1330 (2d Cir. 1975), cert. denied, 423 U.S. 1078, 96 S.Ct. 866, 47 L.Ed.2d 89 (1976), the Second Circuit considered an analogous issue. A defendant had been acquitted on charges of first and second degree kidnapping under one section of a kidnapping statute; the jury was hung, however, on first and second degree charges under another section of the statute. The defendant was retried on the first degree charge and was convicted, despite the fact that he had already been acquitted of second degree kidnapping, a necessary element in the first degree charge. The district court rejected defendant’s habeas corpus motion, in which he had urged collateral estoppel based on double jeopardy. The court reasoned that the apparent inconsistencies in the verdict as a whole precluded the court’s finding that any binding determination in defendant’s favor had been made. The Second Circuit rejected that view, however. In reversing and granting the writ of habeas corpus, the court declared:
The difficulty with the district court’s point of view, however, is that it rests on the proposition that “viewing the verdict herein as an entity and as a whole, I *1113cannot find that the jury herein necessarily acquitted petitioner of kidnapping 2nd degree.” In fact, however, the jury did just that, however erroneously and no matter how confused. The verdict was an express acquittal of kidnapping in the second degree. It was rendered without correction by the trial court and without objection to the inconsistency by the prosecution.
517 F.2d at 1334 (footnote omitted). Admittedly, this was a double jeopardy case; yet, the principle of evenhandedness also requires courts to give full effect to final judgments, without ascribing to those judgments the inconsistencies inherent in other parts of a jury’s determinations.
The majority also rejects the application of estoppel in this case because, it reasons, the evidence in Standefer’s trial was different from that introduced against Nieder-berger. Specifically, the majority notes, Standefer testified at his trial and did not contest the receipt of the trips. But that objection misses the point. If estoppel had been granted at the outset, Standefer may never have had to testify or present a defense. Because Standefer’s testimony might never have been offered had his es-toppel claim been granted, it cannot be said that the admittedly damaging evidence elicited from him renders the estoppel question harmless or, in the majority’s view, irrelevant.
The majority straddles an uneasy line in this case. It refuses to endorse non-mutual collateral estoppel in criminal. eases while balking at rejecting the doctrine outright. In its labored effort to avoid any resolution of the issue, it fails adequately to suggest why, if estoppel is ever to be applied, this is not an appropriate occasion to do so. But I reserve for last a more important criticism than any having to do with the particular result reached by the majority, a result which, I concede, is a respectably arguable one. Specifically, I object to the premises upon which the majority in this case has acted. The majority argues:
The public interest in the accuracy and justice of criminal results is greater than the concern for judicial economy professed in civil cases and we are thus inclined to reject, at least as a general matter, a rule that would spread the effect of an erroneous acquittal to all those who participated in a particular criminal transaction.
Majority Opinion at 1093. In the first place, the argument in favor of non-mutual collateral estoppel in criminal cases is not, as the quoted language at last tacitly suggests, based upon concerns for judicial economy. I agree that concerns for judicial economy should play a much smaller role in criminal justice decisionmaking than they do.
Next, I am less confident than the judges in the majority in my capacity, or theirs, to know what is an “erroneous acquittal.” The institution of jury trial supplanted trial by ordeal, not because juries were thought to be more likely to achieve objective truth than were druids or soothsayers, but because the participation in the fact-finding process of a cross-section of the community was deemed to lend a sufficient legitimacy to that imperfect process to make publicly acceptable the imposition of sanctions. In the criminal context, when the sovereign fails to convince that ad hoc tribunal, sanctions are not imposed. But the application or withholding of sanctions results not because the verdict is either correct or incorrect in any objective sense, but because under our legal system our government has chosen not to sanction when it fails to convince beyond a reasonable doubt. Despite their confidence, neither I nor the judges in the majority can really determine that the Niederberger acquittal was erroneous. Indeed, the odds are exactly even that when the government is given a second chance to convince a new jury of a given set of facts the second verdict, rather than the first, will be objectively incorrect. The issue is not whether either verdict is correct, but whether the government, having gone to the well once, should be given a second opportunity. When the defendant is the same, the double jeopardy clause prohibits the second trip to the well'without regard to the correctness or incorrectness of the initial verdict. When the defendant is, by virtue of 18 U.S.C. § 2, treated as if he is the same, it is hard to identify the justifica*1114tion for a different outcome. And, while non-mutual collateral estoppel is not constitutionally mandated, I submit that its desirability as a non-constitutional rule of criminal law should also be determined without engaging in speculation as to which of two successive verdicts on the same alleged facts is more likely to be correct.
In addition, the majority’s statement that “the purpose of a criminal court is . to vindicate the public interest in the enforcement of the criminal law,” Majority Opinion at 1093, discloses an erroneous view both of the public good and of the role of an independent judiciary in a sanctioning system. Sanctioning, ultimately, depends upon the availability of physical power. A sovereignty, having a practical monopoly upon the application of physical power, could sanction merely by the application of that power. Resort to the ritual of courts and trials, however, reflects a desire to gain public acceptance for the imposition of sanctions, thereby hopefully reducing the incidence of resistance and the need to resort to force. Our democratic legal system is predicated upon the assumption that such ends are desirable. Accordingly, our rules for criminal justice sanctioning have been carefully designed so as to maximize both the fact and appearance of fairness, by deliberately tilting many rules in favor of the defendant and against the sovereign. In a given instance the result of that system may be to deprive the sovereign of the opportunity to punish a malefactor. But that result can be deemed a public harm only if one loses sight of the long-range benefits of a system deliberately designed to tolerate such results. Their occasional occurrence is no public harm but rather a public good, in that they represent the system’s functioning as it was designed to function. Moreover, judges do not possess physical coercive power, but serve only to legitimate its imposition. Since legitimation is the essential contribution of the judicial system, judges act consistently with their function, and thus in the public interest, only when they adopt rules maximizing fairness to defendants. If those rules occasionally produce results which temporary majorities abhor, that should be of little concern to the courts, for our perspective must be determined by the long-range benefits accruing from our efforts to secure the appearance of fairness. By pursuing fairness, rather than acting as law enforcement agents, we achieve, in the long-run, the willing consent of the governed. The republic will not fall because the court concludes that the government should, when it loses before one jury, be collaterally es-topped from proceeding on the same issues before another. But it will be endangered if judges conclude that they must “vindicate the public interest in the enforcement of the criminal law,” rather than develop rules which satisfy the claims of individual fairness.
Because I accept a role, albeit limited, for collateral estoppel in criminal cases, and because the trial court has not yet made the findings sufficient to show why estoppel is inappropriate in this case, I respectfully dissent from the affirmance of Standerfer’s conviction on the Pompano Beach count. I concur, however, in the affirmance of the Doral and Seaview convictions.
. By urging a consideration of non-mutual collateral estoppel in this case, rather than contending that the acquittal of a principal requires the acquittal of an aider and abettor, I do not retreat from the views I expressed in my dissenting opinion in United States v. Bryan, 483 F.2d 88 (3d Cir. 1973) (en banc). In Bryan I stated that because Bryan was explicitly indicted as an aider and abettor of a named principal, he could not be convicted “on the basis of evidence relevant only to a charge of being a principal.” 483 F.2d at 98. To do so would be to convict Bryan of “a charge for which he was not indicted.” Id. Because of this variance from the indictment, I thought Bryan’s conviction could not stand. No issue of variance is presented on this appeal.
. See Restatement (Second) of Judgments § 88 (Tent. Draft No. 2 1975), quoted with approval in State v. Gonzalez, 75 N.J. 181, 380 A.2d 1128, 1132 n.5 (1977).
. See, e. g., Restatement (Second) of Judgments § 88(2), quoted with approval in State v. Gonzalez, 380 A.2d, at 1132 n.5:
§ 88. Issue Preclusion in Subsequent Litigation with Others.
A party precluded from relitigating an issue with an opposing party, in accordance with §§ 68 and 68.1, is also precluded from doing so with another person unless he lacked full and fair opportunity to litigate the issue in the first action or unless other circumstances justify according him an opportunity to relitigate the issue. The circumstances to which consideration should be given include those enumerated in § 68.1 and also whether:
(2) The forum in the second action affords the party against whom preclusion is asserted procedural opportunities in the presentation and determination of the issue that were not available in the first action and that might likely result in the issue’s being differently determined. .
. The majority cites several cases which have rejected the application of non-mutual collateral estoppel to criminal trials. It is true that the policy has not thus far achieved extensive support. There are, of course, a number of courts who have accepted such estoppel claims. See, e. g., United States v. Shuford, 454 F.2d 772 (4th Cir. 1971) (acquittal of principal would require acquittal of aider and abettor); United States v. Prince, 430 F.2d 1324 (4th Cir. 1970) (per curiam) (same); United States v. Bruno, 333 F.Supp. 570 (E.D.Pa.1971); State v. Gonzalez, 75 N.J. 181, 380 A.2d 1128 (1977); People v. Taylor, 12 Cal.3d 686, 117 Cal.Rptr. 70, 527 P.2d 622 (1974). Cf. also United States v. Cas-per, 541 F.2d 1275, 1278-79 (8th Cir. 1976) (apparently accepting the reasoning of the trial court that the lack of privity between the defendants in the first and second trials did not bar the application of collateral estoppel). Nevertheless, as the majority itself observes, “the apparent novelty” of non-mutual collateral estoppel in criminal cases “does not necessarily make it inappropriate or unwise.” Majority Opinion at 1092.
. See p. 1109 supra.