United States v. Flemmi

245 F.3d 24 (2001) | Cited 21 times | First Circuit | March 30, 2001


[Hon. Mark L. Wolf, U.S. District Judge]

Stephen J. Flemmi, one of Boston's most notorious gangsters, served surreptitiously as an informant for the Federal Bureau of Investigation (FBI). In the end, however, there was a falling-out and the government indicted him. This appeal, brought pursuant to 18 U.S.C. § 3731 (a statute that allows the United States, before trial, to appeal orders "suppressing or excluding evidence"), follows on the heels of a district court order barring the government from introducing certain evidence at Flemmi's trial. 1 The court based the suppression order on its conclusion that the government, in obtaining the evidence, had abused the grand jury process. United States v. Flemmi, 108 F. Supp. 2d 39, 43 (D. Mass. 2000). We reverse.


The district court has done a significant public service by bringing to light the tangled relationship between Flemmi and the FBI, and the details of that unholy alliance make for fascinating reading. But those facts are by now old hat, e.g., United States v. Flemmi, 225 F.3d 78, 80-82 (1st Cir. 2000); United States v. Salemme, 91 F. Supp. 2d 141, 148-315 (D. Mass. 1999), and it would serve no useful purpose to rehearse them today. This appeal requires only that we limn the procedural history antecedent to the suppression order.

The grand jury originally indicted a single defendant on October 25, 1994. It subsequently broadened its horizons, naming several additional defendants, including Flemmi, in a superseding indictment returned on January 10, 1995. In that indictment, the grand jury charged Flemmi with, among other things, racketeering and racketeering conspiracy under the Racketeer Influenced and Corrupt Organizations Act (RICO). See 18 U.S.C. § 1962(c), (d). To make these charges stick, the prosecution had to prove that Flemmi engaged in a "pattern of racketeering activity," id. § 1962(c), by participating in the commission of no fewer than two predicate acts within a ten-year time frame, id. § 1961(5). Those predicate acts had to constitute crimes of the type specifically enumerated in the statute. Id. § 1961(1).

To pave the way for this showing vis-à-vis Flemmi, the first superseding indictment designated fifteen predicate acts: one that involved attempted murder, one that involved suborning perjury, and several others that involved gambling or extortion. This specification was augmented when the grand jury returned a second superseding indictment on August 1, 1995. 2 The third superseding indictment, however, was a horse of a different hue. Because the grand jury proceedings leading up to this indictment lie at the heart of the district court's suppression order, we rehearse the relevant facts.

Even before the grand jury handed up the second superseding indictment, the government had heard whispers that Flemmi, in conducting his racketeering enterprise, may have participated in four murders (dating back to 1967). The government subsequently contacted Hugh "Sonny" Shields (who had been acquitted, along with Flemmi, in an earlier state court case involving one of the murders). Although Shields persuaded the prosecutors that he possessed salient information about Flemmi's role in the slayings, he refused to testify unless he received immunity.

The government arranged for a grant of use immunity and hauled Shields before the grand jury in October of 1995. His testimony not only implicated Flemmi in all four homicides but also pointed prosecutors to another potential witness, referred to as "John Doe No. 2." The government brought Doe No. 2, duly immunized, before the grand jury in November of 1995. His testimony likewise inculpated Flemmi in respect to the murders.

Word of the grand jury's renewed activity apparently leaked, and Flemmi moved to dismiss the indictment on the ground that the government was playing fast and loose by employing the grand jury as a vehicle for trial preparation. Before this motion could be adjudicated, the grand jury returned a third superseding indictment. This indictment, handed up on May 21, 1996, added five new predicate acts to the racketeering charges against Flemmi: one dealing with the rigging of horse races (a charge not at issue here) and the other four dealing with the murders. 3 Flemmi responded by moving to suppress the testimony of Shields and Doe No. 2, as well as any evidence derived therefrom.

The district court quite sensibly treated Flemmi's motions as a unit and heard oral arguments late in 1996. The court thereupon took the matter under advisement until July 5, 2000. At that time, the court concluded that the only tangible work product of the challenged grand jury sessions - the third superseding indictment - did not alter the fundamental character of the crimes charged because the added materials did not accuse Flemmi of having committed any new federal crime, but merely attributed more predicate acts to him. Flemmi, 108 F. Supp. 2d at 41-43. This rendered unavailable a safe harbor that the government had sought to reach and set the stage for further inquiry. The court conducted that further inquiry and found that the government had used the grand jury process in the fall of 1995 and thereafter principally for trial preparation, that is, as a means to "compel and freeze the otherwise unavailable testimony" of Shields and Doe No. 2. Id. at 42.

In reaching the conclusion that the safe harbor for "new charges" was unavailable, the court relied on a double jeopardy analysis that indicated, to its satisfaction, that the offenses charged in the second and third superseding indictments were precisely the same. Id. at 57-60. This meant, the court reasoned, that the embellishments to the indictment were no more than additional evidence of the felonies with which Flemmi already had been charged. Id. at 60. Deploying the grand jury as a mechanism for collecting such information, the court ruled, constituted trial preparation (and, accordingly, sufficed to ground a finding of abuse). Id. at 62 (suggesting that the inclusion of additional predicate acts did no more than "impermissibly strengthen[] already-existing charges"). Deeming suppression a condign remedy, the court granted Flemmi's motion to exclude the evidence gleaned from Shields and Doe No. 2.


Claims of grand jury abuse raise a unique set of concerns. The relevant inquiry, strictly speaking, is neither a pure question of fact nor a pure question of law. In re Grand Jury Proceedings (Fernandez Diamante), 814 F.2d 61, 71 (1st Cir. 1987). To the contrary, the inquiry most often comprises a hybrid in that it typically involves an "application of a legal standard designed to ensure that the grand jury, a body operating peculiarly under court supervision, is not misused by the prosecutor . . . ." Id. (quoting In re Grand Jury Subpoena Duces Tecum Dated Jan. 2, 1985 (Simels), 767 F.2d 26, 29 (2d Cir. 1985)).

Given this reality, appellate tribunals have crafted an intermediate standard of review for evaluating district court orders accepting or rejecting claims of grand jury abuse. Under that standard, we accord respect to the lower court's findings, but scrutinize them somewhat less deferentially than we would if either the traditional "abuse of discretion" or "clearly erroneous" rubric applied. See United States v. Leung, 40 F.3d 577, 581 (2d Cir. 1994); Fernandez Diamante, 814 F.2d at 71. This intermediate level of appellate scrutiny is akin to what we have in other contexts termed "independent review." E.g., United States v. Tortora, 922 F.2d 880, 882-83 (1st Cir. 1990) (describing independent review as "an intermediate level of scrutiny, more rigorous than the abuse-of-discretion or clear-error standards, but stopping short of plenary or de novo review," and deeming such review appropriate for appellate oversight of pretrial detention orders).


Although the grand jury operates under judicial supervision, it is essentially an independent institution. In recognition of this status, courts afford grand jury proceedings a presumption of regularity. United States v. Johnson, 319 U.S. 503, 513 (1943). This presumption attaches even after the grand jury has returned an initial indictment. After all, superseding indictments setting forth new charges or adding new defendants are familiar fare. E.g., United States v. Melendez, 228 F.3d 19, 21 (1st Cir. 2000) (superseding indictment added two new defendants); United States v. Peña-Lora, 225 F.3d 17, 23 (1st Cir. 2000) (superseding indictment set forth new charges); United States v. Bender, 221 F.3d 265, 267 (1st Cir. 2000) (superseding indictment added two counts); United States v. Li, 206 F.3d 56, 59 (1st Cir. 2000) (en banc) (superseding indictment added four new defendants). It follows logically that, as a general rule, "evidence obtained pursuant to [an ongoing grand jury] investigation may be offered at the trial on the initial charges." Leung, 40 F.3d at 581.

Notwithstanding the presumption of regularity, prosecutors do not have carte blanche in grand jury matters. However, a party asserting a claim of grand jury abuse must shoulder a heavy burden. See id.; United States v. Badger, 983 F.2d 1443, 1458 (7th Cir. 1993); United States v. Jenkins, 904 F.2d 549, 559 (10th Cir. 1990). One way to carry this burden is to show that the government used the grand jury principally to prepare pending charges for trial. See Fernandez Diamante, 814 F.2d at 70 (explaining "that a grand jury may not conduct an investigation for the primary purpose of helping the prosecution prepare indictments for trial").

This proposition is more simply stated than applied. While it is easy to say that the court's inquiry must focus on the primary purpose underlying the grand jury's involvement, there is a fine line between an improper "trial preparation" use of a grand jury and a proper "continuing investigation" use. This fine line is difficult to plot and, in most instances, determining whether a prosecutor has overstepped it will depend on the facts and circumstances of the particular case.

To assist in the inquiry, courts have devised certain proxies. Thus, if a grand jury's continuing indagation results in the indictment of parties not previously charged, the presumption of regularity generally persists. United States v. Gibbons, 607 F.2d 1320, 1328-29 (10th Cir. 1979). So too when the grand jury's investigation leads to the filing of additional charges against previously indicted defendants. In re Grand Jury Proceedings (Johanson), 632 F.2d 1033, 1041 (3d Cir. 1980). These are purposes befitting the accepted institutional objectives of the grand jury, and their presence bears convincing witness to the propriety of the prosecutor's stewardship. See United States v. Sasso, 59 F.3d 341, 351-52 (2d Cir. 1995); In re Maury Santiago, 533 F.2d 727, 730 (1st Cir. 1976); United States v. Dardi, 330 F.2d 316, 336 (2d Cir. 1964).

Cognizant of this line of cases, the court below concentrated on whether the third superseding indictment - which admittedly haled no new parties into court - altered the nature of the charges previously lodged against Flemmi. The court concluded that the indictment charged no new crime. Flemmi, 108 F. Supp. 2d at 42. Laying that potential proxy to one side, the court then found that the raison d'ître for the ongoing grand jury investigation was trial preparation, specifically, what the court believed was the government's desire to bolster its existing case by memorializing the testimony of Shields and Doe No. 2. Id.

The district court's finding of improper purpose flowed from, and depended upon, its finding that the third superseding indictment did not charge a new offense. See id. at 62. But the court based the underlying finding on a double jeopardy analysis. See id. at 57-60. It asked, in effect, whether Flemmi would be protected from prosecution under the charges laid in the third superseding indictment had he previously been tried under the second. See id. at 61-62. Answering that query affirmatively, the court concluded that the two indictments necessarily charged the same offenses. See id.

This approach is innovative, but unsound. The Double Jeopardy Clause "embodies a triumvirate of safeguards: It protects against a second prosecution for the same offense after acquittal. It protects against a second prosecution for the same offense after conviction. And it protects against multiple punishments for the same offense." United States v. Ortiz-Alarcon, 917 F.2d 651, 653 (1st Cir. 1990) (citation omitted). When, as now, the question is whether the prosecutor's evidence-gathering constituted an abuse of the grand jury, none of these concerns is implicated. The jurisprudence of the Double Jeopardy Clause is therefore inapposite, and we eschew the lower court's misplaced reliance on it. The appropriate method of analysis is more straightforward.

This case turns on whether the facts, viewed objectively, reveal a proper justification for the government's continued resort to the grand jury. In the circumstances at hand, that analysis, as the district court recognized, centers on the grand jury's work product and requires a frank comparison of the charges contained in the third superseding indictment and those contained in the immediately preceding indictment. At this juncture, however, we part company with the lower court; that contrast should take place not by means of a mechanical invocation of double jeopardy principles, but with an eye toward determining whether the new matter contained in the later indictment, assayed in a practical, commonsense manner, demonstrates that the government's ongoing use of the grand jury was primarily for a proper purpose.

Here, the results of the comparison are telling. As said, accusations of grand jury abuse can be conclusively rebuffed by a showing that the challenged proceedings led to the joinder of new defendants or the inclusion of new charges. E.g., Johanson, 632 F.2d at 1041; Gibbons, 607 F.2d at 1328-29. Here, the resumed grand jury proceedings led to the inclusion in the indictment of material that both added new RICO predicate acts and increased the maximum penalty to which Flemmi was exposed. This is analogous to a superseding indictment that adds a new charge - a permissible use of a grand jury that yields a sufficiently substantial change to defeat an accusation of grand jury abuse. We explain briefly.

Adding the four predicate acts of murder to the RICO counts fundamentally altered the character of the indictment not only because their inclusion contemplated new proof but also because their inclusion increased the maximum sentence that could be imposed upon Flemmi in the event of a guilty verdict. A person convicted of a RICO violation ordinarily "shall be . . . imprisoned not more than 20 years." 18 U.S.C. § 1963(a). If, however, "the violation is based on a racketeering activity for which the maximum penalty includes life imprisonment," then the maximum available penalty stretches to life. See id. The measurement is restricted to those predicate acts charged in the body of the indictment. See United States v. Carrozza, 4 F.3d 70, 81 (1st Cir. 1993). 4

In this case, the previous versions of the indictment specified no predicate act that carried a potential sentence of life imprisonment. A murder committed in Massachusetts at the relevant time (and today, for that matter) carries such a penalty. See Mass. Gen. Laws ch. 265, §§ 1, 2 (1959). Thus, the insertion of the murders as predicate acts in the third superseding indictment effectively raised the stakes by increasing the statutory maximum applicable to the existing RICO charges against Flemmi from twenty years to life. To that extent, the indictment entailed greater jeopardy. Cf. Apprendi v. New Jersey, 120 S. Ct. 2348, 2365 n.19 (2000) (declaring in a related, but not identical, context that any fact that increases the defendant's exposure beyond the prescribed statutory maximum "is the functional equivalent of an element of a greater offense").

This set of circumstances puts to rest any notion that the government was abusing the grand jury process. Since the third superseding indictment contained charges analogous to a new offense, the investigation leading to it constituted a proper use of the grand jury. 5 Any different result would unfairly hamstring the government in its pursuit of legitimate law enforcement objectives.

Let us be perfectly clear. We agree with Flemmi that the appropriate inquiry is a matter of substance, not form. A prosecutor's renewed resort to the grand jury for evidence-gathering purposes cannot be validated simply by having the grand jury return any old superseding indictment. If, say, a superseding indictment merely corrects peripheral details or adds something trivial to the pending charges, an inquiring court has every right to be skeptical. But when the new indictment charges new crimes, adds new defendants, or otherwise works a major change in the prior indictment that is sufficiently analogous, for these purposes, to charging new crimes or adding new defendants, it adequately evinces the propriety of the prosecutor's purpose and thus becomes a safe harbor for the government.


We need go no further. Because the third superseding indictment was sufficiently analogous to the lodging of a new criminal charge, we conclude that no abuse of the grand jury process occurred. Since the district court's decision suppressing the testimony of Shields and Doe No. 2 (and the fruits thereof) rests on a contrary premise, it cannot stand.


1. This is the second interlocutory appeal in this case. On an earlier occasion, we reversed a different suppression order. See United States v. Flemmi, 225 F.3d 78 (1st Cir. 2000).

2. This indictment dropped two defendants, added a new defendant, left intact the charges previously lodged against Flemmi, and tagged him with three more predicate acts based on alleged extortions.

3. The government did not seek to indict Flemmi for the crime of murder because there is no federal statute that can be applied to the 1967 slayings without violating the Ex Post Facto Clause. This fact, however, does not prohibit reference to the slayings as predicate acts in connection with the RICO counts. See United States v. Brown, 555 F.2d 407, 416-17 (5th Cir. 1977) (upholding, against constitutional challenge, government's use of predicate acts occurring prior to RICO's effective date in conjunction with predicate acts occurring after that date).

4. To be sure, a sentencing judge may consider uncharged predicate acts in a RICO case, e.g., Carrozza, 4 F.3d at 80, but the judge nonetheless must stay below the maximum penalty allowed under the charges delineated in the indictment and submitted to the jury. See Apprendi v. New Jersey, 120 S. Ct. 2348, 2362-63 (2000); United States v. Robinson, ___ F.3d ___, ___ (1st Cir. 2001) [No. 00-1674, slip op. at 11-12]; see also Carrozza, 4 F.3d at 81 (acknowledging that "the statutory maximum sentence must be determined by the conduct alleged within the four corners of the indictment").

5. Although the district court rejected this line of reasoning, see Flemmi, 108 F. Supp. 2d at 60-61, it did so without any party having cited to it the Supreme Court's hot-off-the-presses Apprendi decision - and Apprendi undermines the district court's rationale for rejection. See Apprendi, 120 S. Ct. at 2354.

Back to top