After a joint trial, a jury found Angelo Commito guilty of offering and John LaFrance guilty of accepting a bribe to influence a Department of Health and Human Service grant decision. On appeal, LaFrance contends that the district court erred in refusing to grant his motion for separate trials, while both defendants argue that the district court erred in admitting unfairly prejudicial evidence. We reverse as to LaFrance, affirm as to Commito, and remand for further proceedings consistent with this opinion.

John LaFrance directed the Philadelphia regional office of the Division of Health Services Delivery, a subdivision of the Department of Health and Human Services. The subdivision was responsible for awarding grants to operate community health centers. Although the grant decision ultimately rested with LaFrance's superior, LaFrance would review the applications and recommend a particular bidder. Angelo Commito organized and administered prepaid health care plans for unions and other groups in the Chicago area. When Commito referred his groups' business to health care providers, he would typically receive a commission or broker's fee.

A jury found that Commito bribed LaFrance in order to steer a government grant to United Health Care, a health care provider which frequently contracted with organizations represented by Commito.1 The government's principal evidence against defendants was the testimony of Alan Cohn, a vice-president for United, who had been indicted for his involvement with Commito in another matter known as "the Munford deal." As part of his Munford deal plea agreement, Cohn agreed to testify against defendants regarding both the Munford deal and the facts underlying this case. The defendants' appeals concern the admission of the Munford evidence.

The Munford deal began when a Federal Bureau of Investigation agent posed as a representative of the Munford Corporation (a real company that cooperated with the FBI) and contacted Commito regarding optical care plans for Munford employees. Commito referred the FBI agent to Cohn, and Cohn and the FBI agent eventually reached an optical care agreement for Munford employees. The government claims that kickbacks for both Commito and the FBI agent were part of the deal and that Cohn knowingly facilitated the concealment of the kickbacks.2 The Munford sting operation incidentally uncovered the facts which led to the indictment in this case.

LaFrance argues that the district court's refusal to grant him a separate trial unfairly prejudiced him because of the Munford evidence introduced against Commito. We agree.

"Barring special circumstances, individuals indicted together should be tried together." United States v. Brugman, 655 F.2d 540, 542 (4th Cir. 1981) (citations omitted). To accommodate the "special circumstances" in which a joint trial would prejudice one of the defendants, Rule 14 of the Federal Rules of Criminal Procedure permits a severance of jointly indicted defendants.3 Here, LaFrance timely filed a Rule 14 motion for severance. "The grant or denial of a motion for severance under Rule 14 lies within the sound discretion of the trial court and its action on such a motion will be overturned only when there has been a clear abuse of such discretion ... [because] the denial of a severance deprive[d] the movant a fair trial and result[ed] in a miscarriage of justice." United States v. Santoni, 585 F.2d 667, 674 (4th Cir. 1978) (citations omitted).

LaFrance first moved for (and was refused) a separate trial in a pretrial motion. LaFrance renewed his motion for severance during the trial when it became increasingly evident that the government considered the Munford evidence against Commito a pillar of its case; it is undisputed that the Munford evidence could not be introduced against LaFrance. Although the district court instructed the jury to consider the Munford evidence only with respect to Commito,4 LaFrance persuasively argues that the cumulative effect of the Munford evidence made it nearly impossible for the jury to appraise LaFrance's culpability based solely upon the independent evidence introduced against him.

Cohn's testimony regarding LaFrance's acceptance of a bribe from Commito was the principal evidence offered at trial against LaFrance. Cohn, however, merely deduced that LaFrance accepted a bribe. Cohn never witnessed Commito bribing LaFrance nor was he directly told that LaFrance needed to be or had been bribed. Therefore, without the Munford evidence and its insinuations that Commito repeatedly engaged in illegalities and often did so in association with Cohn, the government's case against LaFrance is significantly weaker.

Further bolstering the government's relatively weak case against LaFrance were its occasional references to Commito's connections with organized crime. These references cropped up when the government presented its case against Commito, yet nonetheless implied to the jury that LaFrance associated with a reputed mobster.5 Without the joint trial, LaFrance's jury would not have received this information. Because the jury received highly prejudicial evidence, which although aimed at Commito undoubtedly prejudiced LaFrance as well, we conclude that the district court abused its discretion in not severing LaFrance's trial from that of Commito. See, e.g., United States v. Engleman, 648 F.2d 473, 481 (8th Cir. 1981) (reaching the same conclusion in light of similar facts).

Anticipating this conclusion, the government asserts in cursory fashion that the direct evidence against LaFrance was so substantial that the joint trial resulted in harmless error. See United States v. Lane, 474 U.S. 438, 450 (1986). The government raises and details this claim in a single sentence in its brief and then abandons this argument. Our review of the record indicates that the government principally relied on Cohn's testimony to incriminate LaFrance, and as mentioned above, Cohn's testimony was not conclusive on this issue. We are far from certain that the jury would have convicted LaFrance if it did not receive the Munford evidence. This uncertainty precludes a ruling that LaFrance's joint trial with Commito was harmless error. See United States v. Truslow, 530 F.2d 257, 261 (4th Cir. 1975) (severance required where court possessed "grave doubt" whether prejudicial evidentiary admissions "had substantial influence in bringing about a verdict").

Because the government concedes that evidence of the Munford deal cannot be introduced against LaFrance, we limit our review of the district court's admission of this evidence as it applied to Commito alone. Commito argues that the district court abused its discretion in admitting evidence of the Munford deal under Federal Rule of Evidence 404(b).6 Under Rule 404(b), "prior bad acts are admissible if they are (1) relevant to an issue other than character, (2) necessary, and (3) reliable." United States v. Rawle, 845 F.2d 1244, 1247 (4th Cir. 1988) (citations and footnotes omitted). Moreover, prior bad acts may be introduced to show proof of intent, plan, or knowledge. Fed. R. Evid. 404(b). Rule 404(b) admissions are, of course, subject to the general restrictions of Rule 403.7 Commito contends that even if the Munford evidence is admissible under Rule 404(b), the evidence remains unfairly prejudicial, and, as such, is barred from admission by Rule 403.

As announced in Rawle, to determine the admissibility of evidence under Rule 404(b), we must examine its relevancy, necessity, and reliability. If the Munford evidence satisfies these conditions, then we consider Commito's argument that its "probative value is substantially outweighed by the danger of unfair prejudice." Fed. R. Evid. 403.

"In order for evidence to be relevant, it must be sufficiently related to the charged offense." Rawle, 845 F.2d at 1247 n.3 (citation omitted). Putting the Munford evidence aside, the government's principal evidence against defendants was the testimony of Cohn. Cohn's testimony was based largely on inference and did not fully detail the conspiracy for which Commito was convicted. For instance, Cohn testified that Commito told him "we have to take care of the LaFrance boys,"8 and that he understood Commito to be saying that a bribe was in order. In response to this interpretation of his comment, Commito's defense was that he was a legitimate broker and that no bribe was offered. In contradiction to Commito's defense, Cohn further testified that when he balked at bribing LaFrance directly, Commito euphemistically suggested that the bribe money be taken out of the money that United Health Care would pay him for brokering the deal.

The Munford evidence established that in an identical situation, Commito responded the same way: During the Munford deal, when Cohn and Commito realized that the undercover agent could not receive a direct payment from United Health Care, Commito agreed to launder a bribe through himself. The Munford evidence explicated Cohn's testimony and negated Commito's defense by showing that what the government alleged in this case had happened before. Thus, admitting the Munford evidence clarified both Commito's intent and plan behind his cryptic statements to Cohn and illuminated Commito's and Cohn's business relationship. By providing these insights, the introduction of the Munford evidence enabled the jury to more accurately evaluate the competing stories offered by Cohn and Commito. In light of these considerations, we conclude that the Munford evidence was sufficiently related to the charged conduct so as to be relevant.

"[E]vidence is necessary and admissible where it is an essential part of the crimes on trial, see United States v. Masters, 622 F.2d 83, 86 (4th Cir. 1980), or where it 'furnishes part of the context of the crime.' United States v. Smith, 446 F.2d 200, 204 (4th Cir. 1971)." Rawle, 845 F.2d at 1247 n.4. Here, the introduction of the Munford evidence framed Cohn's and Commito's business relationship and enabled the jury to decipher Commito's euphemistic statements. Without this assistance, Commito's intent and plan in making his cryptic comments are masked. Moreover, by demonstrating that his cryptic statements had illegal intent and that he was fully aware that he was orchestrating a bribe, the Munford evidence impugned Commito's defense that he was an innocent broker. See United States v. Greenwood, 796 F.2d 49, 53 (4th Cir. 1986). Accordingly, the Munford evidence " 'furnishes part of the context of the crime,' " and therefore satisfies the necessity requirement established by this court. Rawle, 845 F.2d at 1247 n.4.

The government introduced the Munford evidence via a tape recording, documentary evidence, and the testimony of Alan Cohn and the undercover FBI agent. We consider this evidence highly reliable. See Rawle, 845 F.2d at 1248 (finding testimonial evidence sufficiently reliable, even though witnesses may have had incentives to cooperate with the government).

Commito focuses his appeal on this issue, arguing that the Munford evidence's real impact was to convince the jury that Commito was an unsavory character, who committed whatever crime charged by the government. We agree with the district court's opinion that determining whether the Munford evidence's probative value is substantially outweighed by the danger of unfair prejudice is a"close call" and "not an easy ruling." More importantly, however, we affirm the district court's decision to admit this evidence.

In a case like that here, where the probative value of the contested evidence is especially great, its prejudicial effect must be correspondingly greater in order to bar its admission. "The prejudice which rule is designed to prevent is jury emotionalism or irrationality." United States v. Greenwood, 769 F.2d at 53 (citation omitted). Bribes to facilitate health care deals "generally do not cause such unfair inflammatory reactions." Id. (citation omitted). We recognize that evidence suggesting that a defendant is involved with organized crime might excite a jury. We feel, however, that this excitement would not lead to irrationality. Because the district court's balancing of these considerations was not an arbitrary or irrational exercise of discretion, we affirm its decision. See United States v. Masters, 622 F.2d 83, 88 (4th Cir. 1980) (establishing that Rule 403 decisions will be disturbed only if the district court acted arbitrarily or irrationally).

We vacate the conviction of LaFrance, affirm Commito's conviction, and remand to the district court for further proceedings consistent with this opinion.

AFFIRMED IN PART, REVERSED IN PART, AND REMANDED

LUTTIG, Circuit Judge, concurring in part and dissenting in part:

I dissent from that portion of the majority opinion that holds that the district court "clearly abused its discretion," see ante at 4-5, in denying defendant John B. LaFrance's motion for severance. The court can reach the conclusion that it does only by ignoring the established presumption that juries follow their instructions, a presumption that the Supreme Court has admonished is subject to few exceptions not relevant in this case. The legal and practical effect of the court's holding, given the clarity and number of the trial court's limiting instructions and the relatively innocuous nature of the evidence introduced against John LaFrance's codefendant, is that severance will be required in virtually every case in which the government intends to introduce evidence of other bad acts against a codefendant. The decision thus seriously undermines the ability of the United States to conduct joint trials and disserves the "interests of justice" that the Supreme Court has recognized are served by such trials. See Richardson v. Marsh, 481 U.S. 200, 210 (1987). It appears, as well, that the majority's decision brings this circuit into square conflict with every other federal circuit in the country.

The trial court carefully instructed the jury in this case that the evidence of the so-called Munford deal (which did not even involve LaFrance) was relevant, if at all, to codefendant Commito's motive and intent, that it was introduced solely as to Commito's motive and intent, and that the evidence was not to be considered in their deliberations as to LaFrance's guilt. I would adhere to the"almost invariable assumption" that juries follow their instructions, id. at 206, and hold that, given these instructions, the district court did not abuse its discretion in denying the severance motion because there was no basis for believing that the jury would consider the Munford deal evidence in its deliberations with respect to LaFrance.

The consequences of the majority's holding can only be appreciated with an understanding of the factual context from which it emerges. The primary, but by no means the only, evidence against LaFrance and Commito, see discussion infra at 16, was the testimony of Alan Cohn regarding the Baltimore meeting. Cohn testified that he met for approximately two hours with Commito and Ed and John LaFrance on December 4, 1986, in Cohn's office at United Health Care in Baltimore. After about forty-five minutes, Commito asked the LaFrance brothers to excuse themselves from the office. See Tr. at 117 (Oct. 15, 1990). Commito then said to Cohn, "[w]e have to take care of the LaFrance boys." Id. Cohn testified that he took that to mean that Commito "wanted us to pay Eddie LaFrance money and go bribe John LaFrance." Id. at 118. Cohn initially refused to make the bribes, but then consented once Commito stated that he would take care of the payments out of money he supposedly owed United. See id. Commito then left the room to bring the LaFrance brothers back into the office. As the brothers reentered the room, John LaFrance looked at Cohn and said, "Thank you." Id. at 119. The government argued that an offer and acceptance of a bribe occurred at this meeting under a scheme whereby Commito, and apparently Ed LaFrance, would serve as the middlemen for any payments from Cohn to John LaFrance. Commito and LaFrance contended that nothing illegal occurred at that meeting and that Cohn's testimony was based upon mere speculation.

Following Cohn's testimony, the district court then admitted evidence of the Munford deal, a prior transaction involving Cohn and Commito. In that deal, Commito apparently structured an illegal kickback to an undercover agent who was posing as the representative of the Munford Corporation, which wanted to purchase an optical care plan for its employees. Cohn, Commito, and the agent worked out a scheme whereby United, which Cohn represented, would pay Commito one dollar per employee-member per month, and Commito would illegally pass on seventy-five cents of that payment to the agent.

Immediately after the Munford evidence was presented, the court instructed the jury,

I just want to point out to you that [the Munford evidence] has been offered in evidence here as evidence of[sic] relating to motive, intent and such things as that, relating to Mr. Commito in his dealings with Mr. Cohn. And I want to particularly emphasize to you that the Munford deal has no relationship to Mr. John LaFrance whatsoever, so it is in no way presented in-as relating to Mr. John LaFrance and is merely intended to focus on a similar act that might have bearing for your consideration if you ultimately so decide, relating to intent and motive of Mr. Commito in this matter.

J.A. at 120. In its instructions to the jury at the close of all of the evidence, the court again reaffirmed the limited purposes for which the evidence was admitted:

I want to remind you what I told you at some point in the trial and, namely, the evidence relating to the Munford transaction has nothing to do with the case against Mr. LaFrance and should be disregarded in your consideration of Mr. LaFrance in this matter.

Tr. at 218 (October 23, 1990).

The majority concedes that the Munford evidence tended to prove Commito's intent, motive, and plan at the Baltimore meeting at issue in this case. The majority holds, however, that admission of this evidence required the severance of LaFrance's trial from that of Commito because this evidence also tended to prove the existence of the illegal bribery scheme in this case.1 Says the court, "the cumulative effect of the Munford evidence made it nearly impossible for the jury to appraise LaFrance's culpability based solely upon the independent evidence introduced against him." Ante at 4-5. The court, in other words, concludes that this is an exceptional case in which the jury cannot be presumed to follow the court's instructions not to consider the Munford evidence in evaluating LaFrance's culpability.

The Supreme Court has long presumed the effectiveness of limiting instructions in compartmentalizing evidence for juries in joint trials and has instructed the courts of appeals to do the same. See, e.g., Opper v. United States, 348 U.S. 84, 95 (1954) ("To say that the jury might have been confused amounts to nothing more than an unfounded speculation that the jurors disregarded clear instructions of the court in arriving at their verdict. Our theory of trial relies upon the ability of a jury to follow instructions."); Lutwak v. United States, 344 U.S. 604, 618-19 (1953); United States v. Ball, 163 U.S. 662, 672 (1896). 2 There are limited exceptions to this general presumption, such as where a codefendant's confession incriminates the defendant. See Bruton v. United States, 391 U.S. 123, 135-37 (1968). However, as the Supreme Court recently reaffirmed, the presumption that juries follow their instructions is "almost invariable." Marsh, 481 U.S. at 206.3 The Court emphasized in Marsh that the cynical view of the jury expressed in Bruton is "a narrow exception" to the general presumption. Id. at 207; see also Francis v. Franklin, 471 U.S. 307, 325 n.9 (1985) ("Absent such extraordinary situations, how ever, we adhere to the crucial assumption underlying our constitutional system of trial by jury that jurors carefully follow instructions.").

While there doubtless will be exceptional cases where jury instructions will not be sufficient to cure the prejudice resulting from evidentiary admissions in a joint trial, this case, as the above recitation of the evidence confirms, is in no way exceptional. The Munford deal did not involve LaFrance, and thus the evidence of that deal did not directly incriminate him. This evidence at most tended to prove that Commito had attempted to bribe LaFrance at the Baltimore meeting. The jury would have had to engage in a chain of inferences from the Munford deal evidence to infer that LaFrance was guilty of the acceptance of a bribe in this entirely unrelated case. Thus, the "link" between Commito's and Cohn's activities in the Munford deal, and the possible acceptance of a bribe by LaFrance in this case simply was not so direct as to justify the essential conclusion that it was "overwhelming[ly] probab[le]" that the jury would be unable to obey the district court's instructions. See Marsh, 481 U.S. at 208.4

Given the rather unexceptional circumstances present in this case, the majority's rule (if it is followed in a principled way) will require severance in almost every case in which the government presents other bad act evidence against a codefendant. The rule thus threatens to undermine the critical role that joint trials play in our criminal justice system. See, e.g., id. at 209 (joint trials accounted for almost onethird of federal criminal trials from 1982-1987). As the Supreme Court has observed:

It would impair both the efficiency and the fairness of the criminal justice system to require ... that prosecutors bring separate proceedings, presenting the same evidence again and again, requiring victims and witnesses to repeat the inconvenience (and sometimes trauma) of testifying, and randomly favoring the last-tried defendants who have the advantage of knowing the prosecution's case beforehand. Joint trials generally serve the interests of justice by avoiding inconsistent verdicts and enabling more accurate assessment of relative culpability-advantages which sometimes operate to the defendant's benefit. Even apart from these tactical considerations, joint trials generally serve the inter ests of justice by avoiding the scandal and inequity of inconsistent verdicts.

Id. at 210.

Wholly apart from the consequences that the majority's rule will have for the government's ability to try jointly multiple codefendants, the court's rule appears to bring this circuit into square conflict with every other federal circuit in the nation. Every other circuit has affirmed a district court's refusal to sever despite the admission of Rule 404(b) other bad act evidence against a codefendant, so long as proper limiting instructions are given.5 Judge Heaney himself has held that limiting instructions can cure the prejudice to a defendant from the admission of other bad act evidence against a codefendant. United States v. Boone, 641 F.2d 609, 612-13 (8th Cir.) (Heaney, J.), cert. denied, 454 U.S. 831 (1981). Some circuits, in reliance upon Marsh, see supra note 3, have even held that there is no abuse of discretion where a district court denies severance under Fed. R. Crim. P. 14 in a case in which a codefendant's postconspiracy confession is admitted into evidence, provided that the district court properly instructs the jury and there is no reference to the defendant in the confession. See, e.g., United States v. Romero, 897 F.2d 47, 53 (2d Cir.), cert. denied, 110 S. Ct. 3253 (1990), and cert. denied, 111 S. Ct. 975 (1991); United States v. Benitez, 920 F.2d 1080, 1085-86 (2d Cir. 1990); United States v. Smith, 918 F.2d 1032, 1039 (2d Cir. 1990), cert. denied, 111 S. Ct. 1086 (1991); United States v. Soriano, 880 F.2d 192, 197 (9th Cir. 1989) (dictum). If it is not an abuse of discretion to deny severance where a codefendant's confession to the crime for which the codefendants are being tried is admitted into evidence, a fortiori it should not be an abuse of discretion to deny severance to prevent the jury from hearing the considerably more attenuated and significantly less prejudicial Munford evidence.

Because severance is committed to the sound discretion of the district court and no circumstances are present in this case that would warrant a departure from the presumption that jurors follow their instructions, I would hold that LaFrance was not unduly prejudiced as a result of the admission of the Munford evidence and, therefore, that the district court did not abuse its discretion in denying LaFrance's motion for severance.

Even were I to conclude that the district court abused its discretion, I still would not hold that reversal is required because I believe that any prejudice resulting from the court's failure to sever was harmless. The majority holds otherwise, reasoning that it is"far from certain that the jury would have convicted LaFrance if it did not receive the Munford evidence," ante at 6. However, the evidence simply does not support such a conclusion. The jury heard of LaFrance's breaches of agency regulations in providing improper assistance to Cohn over a long period of time. See Appellants' Br. at 5, 7-10. The jury also had before it Cohn's own testimony regarding the meeting in Baltimore. It learned that LaFrance and Cohn spoke repeatedly on the telephone after the meeting, used aliases and would not use their office telephones when speaking. See Appellee's Br. at 9-10. The jury heard a tape of a telephone conversation in which Ed LaFrance told John LaFrance not to sign any approval of the recommendation until "the eagle hits the nest." Appellee's Br. at 10-11. It also learned that when interrogated, John LaFrance denied the Baltimore meeting, despite the fact that the FBI saw him at the meeting through the windows of the office, he denied ever hearing the phrase, "when the eagle hits the nest," and he told the FBI that the name "Jim Lofton" that was in his personal telephone book next to Cohn's number was a person with the Occupational Safety and Health Administration. Tr. at 102-04, 109 (Oct. 17, 1990). This evidence, even without the corroborative Munford evidence, amply supports LaFrance's guilt and thus would require affirmance of the jury's verdict.

In conclusion, I believe that the district court was well within its broad discretion in denying LaFrance's motion for severance. Even if the court erred in denying the motion, I am convinced that any resulting prejudice was harmless. Accordingly, while I join all other portions of the majority's opinion, I respectfully dissent from that portion of the opinion reversing LaFrance's conviction.