A federal grand jury in the Eastern District of Michigan indicted James Thorpe for being a felon in possession of a firearm, in violation of 18 U.S.C. § 922(g). Thorpe, an African-American, moved to dismiss the indictment on the ground that he was being selectively prosecuted because of his race. After conducting a preliminary investigation in support of his claim, Thorpe formally moved the district court for discovery of all of the government's files regarding the Project Safe Neighborhoods (PSN) program under which Thorpe was being prosecuted.

The court granted Thorpe's motion, reasoning that Thorpe could not support his selective-prosecution claim without the requested materials and that the harm to the government, which had already disclosed some of the requested materials to another judge in a different case, would be minimal. When the government refused to fully comply with the district court's discovery order, the court dismissed with prejudice the indictment against Thorpe. For the reasons set forth below, we REVERSE the judgment of the district court granting Thorpe's discovery motion, REINSTATE the government's indictment against Thorpe, and REMAND the case to the district court for further proceedings consistent with this opinion.

At approximately 2:00 a.m. on March 31, 2003, officers from the Detroit Police Department responded to a dispatch reporting that there was an individual with a gun in a red Ford Taurus parked next to a gas pump at an Amoco station in Detroit. Shortly after their arrival at the station, the officers observed an African-American male lying asleep in the driver's seat of a red Ford Taurus that was parked next to a gas pump. Beside him was a .380 automatic handgun, which the officers saw on the car's front passenger seat inches from his hand. The individual was James Thorpe, whom the officers promptly arrested and placed in the back seat of the police car. Shortly thereafter, according to one of the officers, Thorpe voluntarily exclaimed, among other things, "I knew the gun was in there, but I just didn't take it out."

Following Thorpe's arrest by the Detroit police, he was prosecuted by the United States Attorney's Office in accordance with the PSN program. PSN is a Department of Justice initiative that encourages state and federal law enforcement, as well as other segments of the community, to collaborate in the reduction of "gun crime in America." Prosecutions under the PSN program, which occur at the federal level, typically arise by way of referrals from state law enforcement. According to the PSN web site, the U.S. Attorney in each participating federal district must have a strategic plan to attack gun crime, must constantly evaluate the plan's effectiveness, and must report semi-annually to the Department of Justice on several aspects of the district's PSN implementation efforts.

Thorpe filed a motion to dismiss the indictment in September of 2004. He alleged that the implementation of PSN in the Southern Division of the Eastern District of Michigan had resulted in selective prosecution on the basis of race, in violation of the Equal Protection Clause of the Fourteenth Amendment. On October 4, 2004, the district court authorized Thorpe, who is indigent, to expend one thousand dollars of government funds to investigate his claim. Following a hearing on Thorpe's motion to dismiss, but before the district court had reached a decision, Thorpe filed a separate motion for discovery regarding his claim of selective prosecution. Thorpe submitted the following statistical reports in support of his motion: (1) two 2002 reports from the U.S. Sentencing Commission's web site documenting the rates of prosecution of firearm offenses and the rates of imprisonment for those individuals convicted of firearm offenses in every district and circuit in the country, (2) an unattributed report documenting the racial composition of all of the counties in Michigan, and (3) two reports from local Federal Defender's Offices, one in Detroit and the other in Flint, documenting the race of the defendants in each office's "pending firearm cases with state origin."

The U.S. Sentencing Commission reports demonstrated that, at least in 2002, the number of defendants sentenced for federal firearms offenses in the Eastern District of Michigan as a percentage of all federal sentencings was approximately 60% greater than the national average (12.9% versus 8.1%). Moreover, whereas 89% of those individuals convicted of firearm offenses nationally were in prison in 2002, 97% of those individuals convicted of firearm offenses in the Eastern District of Michigan were in prison that year. No mention of PSN, however, appears in the U.S. Sentencing Commission reports. The unattributed report is devoid of crime statistics altogether. Instead, it appears to be a demographic report showing that the five counties in the Eastern District of Michigan with the highest percentage of African-Americans in their populations are Wayne (42.2%), Genessee (20.4%), Washtenaw (12.3%), Oakland (10.1%), and Jackson (7.9%). Finally, and most pertinent, the reports from the Federal Defender's Offices in Detroit and Flint revealed that of the 68 "pending firearm cases with state origin" in both offices combined, 60, or approximately 88%, involved African-American defendants. Thorpe also provided numerous printouts from both the local and national web sites for PSN, setting forth the information summarized in relevant part above.

Conceding that this information was not sufficient to establish his selective-prosecution claim, Thorpe then requested that the district court order the government to disclose its "entire records on Project Safe Neighborhood" for an in camera review. The district court granted Thorpe's discovery motion in March of 2005, while at the same time denying his motion to dismiss the indictment against him. Specifically, the court required the government to produce for an in camera review the following documents relating to PSN:

1) The criteria for cooperation and for prosecution or rejection of state cases;

2) The strategic plan regarding the Project;

3) A list of what documents and information the U.S. Attorney retains for cases prosecuted or rejected as part of Project Safe Neighborhoods;

4) Writings regarding or comprising the Memoranda of Understanding between the Federal Government and the Prosecutors' Offices in the Eastern District of Michigan;

5) Statistics distinguishing between cases resolved by plea in State Court and those referred for prosecution in Federal court; and

6) Any statistics on the cases the U.S. Attorney prosecutes or rejects, by race.

The district court stated that its decision was based on the fact that "[i]n United States v. Nixon, 316 [315] F.Supp.2d 876 (2004)[sic], the Honorable John Feikens ordered the Government to produce certain documents for in camera review relating to the Project Neighborhoods initiative."

The government initially refused to comply with the district court's order, citing "well established legal standards for discovery on a claim of selective prosecution" and "important institutional considerations." After the district court ordered the government to show cause why it should not be held in contempt for its failure to comply with the discovery order, however, the government relented, at least in part. In June of 2005, "[i]n view of the history" of Judge Feikens and the Nixon case, the government submitted "those same documents" that it had previously submitted in Nixon for in camera review by the district court. But the district court noted in its order dismissing the indictment that the government had failed to submit "an affidavit indicating which [of the other documents ordered for discovery] do not exist."

Because of the government's failure to fully comply with the discovery order, the district court, invoking its "supervisory powers" under McNabb v. United States, 318 U.S. 332, 340, 63 S. Ct. 608, 87 L. Ed. 819 (1943), dismissed with prejudice the government's indictment against Thorpe in July of 2005. The government timely appealed.

As an initial matter, we have jurisdiction over this appeal pursuant to 28 U.S.C. § 1291 and 18 U.S.C. § 3731. See United States v. Bass, 266 F.3d 532, 535-36 (6th Cir.2001) (finding that jurisdiction was proper because the district court's dismissal of the death penalty notice, like the dismissal of an indictment, constituted a "final, appealable order" for the purpose of 18 U.S.C. § 3731), rev'd on other grounds, 536 U.S. 862, 122 S. Ct. 2389, 153 L. Ed. 2d 769 (2002).

Notwithstanding the government's original argument for a de novo standard of review, we review a district court's decision to grant or to deny a motion for discovery regarding a selective-prosecution claim under the abuse-of-discretion standard. See United States v. Jones, 159 F.3d 969, 977-78 (6th Cir.1998), and cases cited therein; cf. United States v. Armstrong, 517 U.S. 456, 482, 116 S. Ct. 1480, 134 L. Ed. 2d 687 (1996) (Stevens, J., dissenting) ("Even if respondents failed to carry their burden of showing that there were individuals who were not black but who could have been prosecuted in federal court for the same offenses, it does not follow that the District Court abused its discretion in ordering discovery.") (emphasis added).

The government has since acknowledged the incorrectness of its argument, but we still pause to note that the de novo standard of review adopted by the Fourth and Tenth Circuits in United States v. Olvis, 97 F.3d 739 (4th Cir.1996), and United States v. James, 257 F.3d 1173 (10th Cir.2001), respectively, remains—by a considerable margin—the minority standard of review. See United States v. Sanders, 211 F.3d 711 (2d Cir.2000) (applying the abuse-of-discretion standard); United States v. Hedaithy, 392 F.3d 580 (3d Cir.2004) (same); United States v. Arenas-Ortiz, 339 F.3d 1066 (9th Cir.2003) (same); United States v. Quinn, 123 F.3d 1415 (11th Cir.1997) (same); see also, e.g., United States v. Penagaricano-Soler, 911 F.2d 833 (1st Cir.1990) (applying the abuse-of-discretion standard before Armstrong); United States v. Cooks, 52 F.3d 101 (5th Cir.1995) (same); United States v. Mitchell, 778 F.2d 1271 (7th Cir. 1985) (same); United States v. Hintzman, 806 F.2d 840 (8th Cir.1986) (same).

Both parties agree that Armstrong governs a district court's disposition of discovery requests related to selective-prosecution claims. As this court noted in Jones, the Armstrong Court held that "[b]ecause a selective prosecution claim is not a defense to the merits of a criminal charge but, instead, an independent assertion of misconduct, discovery is not available pursuant to Fed.R.Crim.P. 16." Jones, 159 F.3d at 975 n. 3 (citing Armstrong, 517 U.S. at 463, 116 S. Ct. 1480). A defendant hoping to obtain discovery must therefore make a showing of "some evidence tending to show the existence of the essential elements of the defense, discriminatory effect and discriminatory intent." Armstrong, 517 U.S. at 468, 116 S. Ct. 1480 (quotation marks omitted) (emphasis added). Although the "some evidence" standard is "rigorous," id., it is still relatively light, because "[o]bviously, a defendant need not prove his case in order to justify discovery on an issue." Jones, 159 F.3d at 978.

For the purpose of satisfying the discriminatory-effect prong, "some evidence" means "a credible showing" that "similarly situated individuals of a different race were not prosecuted." Armstrong, 517 U.S. at 465, 470, 116 S. Ct. 1480. Evidence "identify[ing] individuals who were not black and could have been prosecuted for the offenses for which respondents were charged, but were not so prosecuted," therefore suffices. Id. at 470, 116 S. Ct. 1480. Thorpe, however, has failed to meet this standard. This court's holding in Jones is instructive.

In Jones, the court reversed the district court's order denying discovery to Jones, an African-American man, on his selective-prosecution claim. The court reasoned that in addition to the flagrant and uncontroverted evidence of discriminatory intent, Jones had also satisfied the "some evidence" standard as to discriminatory effect by introducing evidence that "law enforcement referred only him and his co-defendant Billings for a federal prosecution that involved crack cocaine, and failed to refer for federal prosecution eight non-African-Americans who were arrested and prosecuted for crack cocaine." Jones, 159 F.3d at 978. Neither the small sample size of only 10 state prosecutions for crack cocaine nor the fact that Jones's codefendant was in fact Caucasian prevented this court from remanding Jones's case to the district court to compel discovery. Id.

The evidence that Thorpe presented to the district court in the present case, however, falls far short of the showing made in Jones. As Thorpe repeatedly concedes in his brief, the evidence that he submitted says nothing about "similarly situated" individuals. What Thorpe offered, in addition to the background statistics from the U.S. Sentencing Commission regarding rates of prosecution and imprisonment for firearm-related crimes, were the two reports from the Federal Public Defender's Office for the Eastern District of Michigan (FDO Reports). The FDO Reports demonstrated only that a large number of firearm-related prosecutions, presumably under the PSN program, had been pursued against African-Americans. Specifically, the reports showed that of the 68 "pending firearm cases with state origin" known to the FDOs, 60(88%) involved African-American defendants, 3(4%) involved Caucasian defendants, 3(4%) involved Hispanic defendants, and 2(3%) involved Native-American defendants.

The reports, however, said nothing about whether any of the "pending firearm cases" dealt with prosecutions for violations of 18 U.S.C. § 922(g), the statutory provision under which Thorpe was indicted. In other words, the two FDO Reports, as well as the more general data from the U.S. Sentencing Commission, constituted nothing more than "raw statistics regarding overall charges." See United States v. Bass, 536 U.S. 862, 863, 122 S. Ct. 2389, 153 L. Ed. 2d 769 (2002) (concluding that "nationwide statistics demonstrating that the United States charges blacks with a death-eligible offense more than twice as often as it charges whites" were insufficient for the purpose of satisfying the Armstrong standard) (quotation marks omitted). And in Bass, the Supreme Court held that such statistics "say nothing about charges against similarly situated defendants." Id. at 864, 122 S. Ct. 2389 (emphasis in original).

The evidence presented by Thorpe in support of his motion for discovery thus failed to demonstrate that the federal government had declined to prosecute other similarly situated individuals. Thorpe concedes this fact on appeal, arguing instead that it was and still is "impossible" for him or any other PSN defendant to know what constitutes a similarly situated individual for the purposes of PSN prosecutions. He notes that unlike the "clearly set forth" statutory crimes at issue in Armstrong and Bass, "the Government has never set forth the criteria for how a case is determined to qualify for PSN prosecution."

Thorpe's argument misses the point. Like the defendants in Armstrong and Bass, Thorpe was charged with the violation of a very specific federal statute, 18 U.S.C. § 922(g). He therefore cannot effectively distinguish those cases on the ground that there existed "a concrete definition of what would constitute a `similarly situated' defendant based on the plain language of each respective statute being enforced." Such a "concrete definition" exists in Thorpe's case as well. The PSN program, as the government properly notes, creates no new element or offense; it is simply a broad initiative designed to reduce certain types of independently defined crimes.

An individual similarly situated to Thorpe would therefore be a felon known to federal law enforcement officers who has been arrested for "possess[ing] in or affecting commerce, any firearm or ammunition." See 18 U.S.C. § 922(g). Thorpe does not need access to the government's "entire records on Project Safe Neighborhoods" to identify such individuals. In fact, precisely as the Supreme Court suggested to Armstrong, we suggest that Thorpe, too, "could have investigated whether similarly situated persons of other races were [arrested and/or prosecuted] by the State of [Michigan] and were known to federal law enforcement officers, but were not prosecuted in federal court." Armstrong, 517 U.S. at 470, 116 S. Ct. 1480. Instead, Thorpe appears to have made little effort to explore state-court records that were available to him as well as to every other member of the public.

Thorpe therefore failed to present even "some evidence" tending to show discriminatory effect despite the fact that it was not impossible for him to do so. Cf. Chavez v. Ill. State Police, 251 F.3d 612, 639-640 (7th Cir.2001) (noting that, unlike in a criminal case, "[i]n a civil racial profiling case, ... the similarly situated requirement might be impossible to prove" because the plaintiff "would not, barring some type of test operation, be able to provide the names of other similarly situated motorists who were not stopped"). To be sure, we do not mean to imply that Thorpe's discovery request constituted a "fishing expedition" undertaken in bad faith, but we do wish to point out that good faith alone cannot overcome legal insufficiency. Similarly, "[m]erely demonstrating that better evidence cannot be obtained without discovery does not suddenly render otherwise insufficient evidence sufficient." United States v. Arenas-Ortiz, 339 F.3d 1066, 1071 (9th Cir.2003).

Thorpe does refer in his brief to race-based statistics regarding PSN prosecutions that he obtained from the Washtenaw County Prosecutor's Office pursuant to a Freedom of Information Act request. But even if we assume that these statistics add weight to Thorpe's claim of selective prosecution, they are irrelevant for the purpose of this appeal because there is no indication in either the record or the briefs—nor was there any indication given by Thorpe's counsel during oral argument—that Thorpe presented these statistics to the district court before it ordered discovery.

Even if we assume for the sake of argument that Thorpe presented some evidence of discriminatory effect, reversal would still be called for because Thorpe, as the district court itself made clear, did not produce any evidence whatsoever of discriminatory intent. For the purpose of satisfying the discriminatory-intent prong, what "some evidence" means is not entirely clear. The Supreme Court has not had occasion to directly consider the issue, because the Court in both Armstrong and Bass found that the defendant had failed as an initial matter to satisfy the discriminatory-effect prong. Armstrong, 517 U.S. at 470, 116 S. Ct. 1480 ("The study failed to identify [similarly situated] individuals who . . . could have been prosecuted ... but were not.. . ."); Bass, 536 U.S. at 863, 122 S. Ct. 2389 ("We need go no further in the present case than consideration of the evidence supporting discriminatory effect.").

This circuit's opinion in Jones provides only slightly more guidance, because that case involved racial animus so blatant that the evidence of it, in the court's opinion, was sufficient to meet not only the "some evidence" discovery threshold but the higher prima facie merits standard as well. Jones, 159 F.3d at 977-78 (recounting how the officers who arrested Jones and his wife had, among other things, made and worn T-shirts depicting the two African-American suspects at the time of their arrest and later, while Jones was in jail awaiting trial, sent Jones a postcard of an African-American woman with bananas on her head).

Given this lack of clear guidance, a logical approach would be to apply the prevailing test for discriminatory intent that predated Armstrong, because Armstrong recognized that "[t]he requirements for a selective-prosecution claim draw on ordinary equal protection standards." 517 U.S. at 465, 116 S. Ct. 1480 (quotation marks omitted). In McCleskey v. Kemp, 481 U.S. 279, 292, 107 S. Ct. 1756, 95 L. Ed. 2d 262 (1987), the Supreme Court held that in all but "certain limited contexts," general statistics, without more, are insufficient to show discriminatory intent. Instead, a defendant hoping to prevail under the Equal Protection Clause "must prove that the decisionmakers in his case acted with discriminatory purpose." Id. (emphasis in original); see also Wayte v. United States, 470 U.S. 598, 610, 105 S. Ct. 1524, 84 L. Ed. 2d 547 (1985) ("Discriminatory purpose ... implies more than ... awareness of consequences. It implies that the decisionmaker ... selected or reaffirmed a particular course of action at least in part because of, not merely in spite of, its adverse effects upon an identifiable group.") (brackets and quotation marks omitted).

At least two federal courts, including this one, however, have questioned this approach in the discovery context. The first is United States v. Bradley, 880 F. Supp. 271 (M.D.Pa.1994). In Bradley, the court noted that "McCleskey did not address the significance of disparate impact [in] the context of a discovery request," where the standard is "more lax." Id. at 281, 107 S. Ct. 1756. Accordingly, the Bradley court determined that the McCleskey standard did not govern its determinations of evidentiary sufficiency at the discovery stage. The other court to express a similar view is our own. In United States v. Bass, 266 F.3d 532, 539-40 (6th Cir.2001), rev'd on other grounds, 536 U.S. 862, 122 S. Ct. 2389, 153 L. Ed. 2d 769 (2002), this court noted that although "McCleskey will certainly preclude Bass's selective prosecution claim if, at the end of discovery, he fails to show any additional evidence [of intentional discrimination, McCleskey] does not ... pose any bar to Bass at this preliminary stage."

Because the abuse-of-discretion standard of review governs, this uncertainty in the law reduces the effective range of the government's argument on appeal. A district court abuses its discretion when it "relies on erroneous findings of fact, applies the wrong legal standard, misapplies the correct legal standard when reaching a conclusion, or makes a clear error of judgment." Reeb v. Ohio Dep't of Rehab. & Corr., 435 F.3d 639, 644 (6th Cir.2006) (emphasis added). Precisely because the appellate courts, including this one, have had difficulty articulating a clear and uniform standard for what constitutes "some evidence" of discriminatory intent, the district court could not have known what the correct legal standard was when it ruled upon Thorpe's motion for discovery. That the district court "applie[d] the wrong legal standard" or "misapplie[d] the correct legal standard" in granting the motion is highly unlikely absent such knowledge. Id.

This limits the government to arguing that the district court either "relie[d] on erroneous findings of fact" or "ma[de] a clear error of judgment." Id. Fortunately for the government, the district court's order granting Thorpe's discovery motion was clearly erroneous on its face, as discussed further below.

Thorpe admits in his brief that he failed to provide the district court with any evidence of the government's discriminatory intent. He also concedes that "the original publicized plan for PSN obviously did not espouse any discriminatory purpose." Instead, his "evidence" of discriminatory intent amounts to an inference that he contends the government itself created:

Rather, the Government proved its discriminatory intent, based on the knowledge it necessarily had at the time of the indictment against Mr. Thorpe of the overwhelmingly racially discriminatory effect the prosecution choices the U.S. Attorney for the Eastern District of Michigan was making, by pursuing PSN prosecutions almost exclusively against non-Caucasian defendants.

(Emphasis in original.) Rephrased, Thorpe's argument appears to be that the government's pursuit of a program despite knowledge of that program's discriminatory effect is by itself "some evidence" of discriminatory intent.

The caselaw does not support Thorpe's argument. To be sure, the government exaggerates by implying that statistical evidence of discriminatory effect, without more, can never raise an inference of discriminatory intent. See United States v. Tuitt, 68 F. Supp. 2d 4, 10 (D.Mass.1999) ("A discriminatory effect which is severe enough can provide sufficient evidence of discriminatory purpose."). But such a finding remains the exceedingly rare exception to the general rule. Cf. Wayte, 470 U.S. at 610, 105 S. Ct. 1524 ("Discriminatory purpose . . . implies more than . . . awareness of consequences.").

On only a handful of occasions has the Supreme Court made such an exception. Most notable are the cases of Yick Wo v. Hopkins, 118 U.S. 356, 6 S. Ct. 1064, 30 L. Ed. 220 (1886), and, almost three-quarters of a century later, Gomillion v. Lightfoot, 364 U.S. 339, 81 S. Ct. 125, 5 L. Ed. 2d 110 (1960). In Yick Wo, the plaintiff presented statistics demonstrating that the City of San Francisco had applied an ordinance prohibiting the operation of laundries in wooden buildings against Chinese nationals only. The city had denied all 200 permit applications submitted by Chinese nationals, but had granted all 80 permit applications submitted by non-Chinese nationals. Yick Wo, 118 U.S. at 374, 6 S. Ct. 1064. The Court famously held:

Though the law itself be fair on its face, and impartial in appearance, yet, if it is applied and administered by public authority with an evil eye and an unequal hand, so as practically to make unjust and illegal discriminations between persons in similar circumstances, material to their rights, the denial of equal justice is still within the prohibition of the Constitution.

Id. at 373-74, 6 S. Ct. 1064.

In Gomillion, the plaintiffs alleged that the Alabama legislature's decision to alter the boundaries of the city of Tuskegee — "from a square to an uncouth twenty-eight-sided figure" — had resulted in the exclusion of all but 4 or 5 of 400(99%) possible black voters but not a single eligible white voter. Gomillion, 364 U.S. at 340-41, 81 S. Ct. 125. Echoing Yick Wo, the Court held that if these statistical allegations were to prove "uncontradicted or unqualified" at trial, "the conclusion would be irresistible, tantamount for all practical purposes to a mathematical demonstration, that the legislation is solely concerned with segregating white and colored voters by fencing Negro citizens out of town so as to deprive them of their pre-existing municipal vote." Id. at 341, 81 S. Ct. 125.

The statistical evidence that Thorpe presents — assuming, again, that it speaks to similarly situated persons — falls far short of approaching the disparities at issue in Yick Wo and Gomillion. First, the FDO Reports from Detroit and Flint, which document the cases of only 68 individuals, "are based on a statistically unimpressive number of federal defendants." See United States v. Turner, 104 F.3d 1180, 1185 (9th Cir.1997) (concluding that the district court had abused its discretion in granting discovery on the defendants' selective-prosecution claim). In addition, the very fact that the FDO Reports indicate that the government was pursuing similar prosecutions against Caucasian, Hispanic, and Native-American defendants, albeit in limited numbers, belies a claim of discriminatory intent towards African-American defendants specifically.

The case of United States v. Colbert, No. 04-80026 (E.D.Mich. Oct. 7, 2005), which Thorpe contends is directly on point, does not alter our analysis. In Colbert, the district court convened a hearing on Colbert's motion to dismiss the indictment against him on the ground of selective prosecution. The hearing occurred after the district court in the present case had ordered discovery and dismissed the indictment against Thorpe. At the Colbert hearing, the government insisted that Colbert's case was not a PSN prosecution, and the district court found this fact dispositive in denying Colbert's motion to dismiss. In its brief in the present case, however, the government lists Colbert as one of several PSN cases in the Eastern District of Michigan where the defendant has alleged selective prosecution. Thorpe argues that this change in the government's characterization of the Colbert case — first as a non-PSN case, and later as a PSN case — constitutes "a poor attempt to cover up what they know is an unconstitutional practice."

But even if we assume that the proceedings in Colbert add weight to Thorpe's similar claim of selective prosecution, those proceedings occurred after the district court in Thorpe's case had already granted discovery. The transcript of the Colbert hearing is therefore irrelevant to the question of whether the district court in this case abused its discretion in granting Thorpe discovery on his selective-prosecution claim. This does not mean, of course, that the Colbert case, or the Washtenaw County statistics, are of no use whatsoever to Thorpe. He still may use them to his advantage if he decides to further pursue his selective-prosecution claim on the merits despite our holding regarding his discovery request.

Notwithstanding the inapplicability of Colbert, Thorpe argues that requiring him to make a more affirmative showing of discriminatory intent without the benefit of discovery would be patently unfair:

As PSN is not a statutory scheme, and its guidelines for case selection for federal prosecution are unknown outside of the U.S. Attorney's Office, the Government has put all PSN defendants in a Catch-22: you have to prove the threshold with evidence of discriminatory intent to get discovery from the Government, but you cannot have access to any of the information necessary to prove (or to disprove) the threshold without discovery from the Government.

The Court in Armstrong, however, was well aware of precisely this conundrum, yet still found the "rigorous" standard for discovery to be justified: "The similarly situated requirement does not make a selective-prosecution claim impossible to prove." Armstrong, 517 U.S. at 466, 116 S. Ct. 1480. To be sure, the burden that Armstrong imposes on a defendant seeking discovery is still quite heavy, but it is not, in the Court's words, "insuperable." Id. at 470, 116 S. Ct. 1480.

In Armstrong's wake, numerous commentators and courts alike have questioned whether the Supreme Court's reassurance would prove hollow in practice. Most opined that it would, raising virtually the same arguments that Thorpe raises now on appeal. See, e.g., Anne Bowen Poulin, Prosecutorial Discretion and Selective Prosecution: Enforcing Protection After United States v. Armstrong, 34 Am. Crim. L.Rev. 1071, 1098 (1997) ("By requiring the defendant to produce specific evidence of an unprosecuted control group before granting discovery, the Court subjects the defendant to a `Catch 22': the defendant needs discovery to obtain the information necessary to entitle the defendant to discovery."); id. at 1073 ("[The Armstrong] holding creates a barrier few defendants are likely to surmount. Consequently, few selective prosecution claims will receive any meaningful judicial hearing."); Hubbard v. United States, No. 04-80321, 2006 WL 1374047, at *1 (E.D.Mich. May 17, 2006) ("The discovery rule established in Armstrong places a nearly impossible burden on defendants claiming selective prosecution since they cannot prove their claim without discovery and cannot obtain discovery without mustering some evidence in support of their claim.").

Despite this criticism, the Supreme Court in United States v. Bass, 536 U.S. 862, 122 S. Ct. 2389, 153 L. Ed. 2d 769 (2002), summarily and unequivocally dispelled any notions that it would consider, much less reduce, the threshold for discovery that it first set forth in Armstrong. Reversing this court's judgment that the district court had not abused its discretion in granting Bass's motion for discovery, the Supreme Court in Bass issued a brief per curiam opinion that minced no words as to Armstrong's enduring validity:

In [Armstrong], we held that a defendant who seeks discovery on a claim of selective prosecution must show some evidence of both discriminatory effect and discriminatory intent. . . . Under Armstrong, therefore, because respondent failed to submit relevant evidence that similarly situated persons were treated differently, he was not entitled to discovery. . . . The Sixth Circuit's decision is contrary to Armstrong and threatens the "performance of a core executive constitutional function." Armstrong, [517 U.S. at 465, 116 S. Ct. 1480]. For that reason, we reverse.

Bass, 536 U.S. at 863-64, 122 S. Ct. 2389. Thorpe's impossibility argument, accordingly, is misplaced.

We need only briefly consider Thorpe's further suggestion that we find "some evidence" of discriminatory intent simply because all of the PSN prosecutions in the Eastern District of Michigan arise from "a few select counties," or precisely those counties with the highest proportion of African-Americans in the population. The Ninth Circuit's opinion in United States v. Turner, 104 F.3d 1180 (9th Cir. 1997), is instructive. After staying the case pending the resolution of Armstrong in the Supreme Court, the Ninth Circuit held that the district court had abused its discretion in ordering discovery relating to the defendants' selective-prosecution claim. Similar to Thorpe in the present case, the Turner defendants had attempted to persuade the Ninth Circuit that "selection of a particular community for a particular enforcement operation constitutes racial discrimination if it is foreseeable that because of the ethnic composition of the community one race will necessarily provide most of the government's targets." Id. at 1185. The Ninth Circuit thoroughly—and, we think, correctly—rejected this line of reasoning:

The defendants' hypothetical has a superficial attraction but is seriously flawed. It is not entirely unnatural for an observer noting ethnic identity to come to the conclusion that when almost all the defendants charged with a particular offense have a certain skin color or ethnic identity that a racial or ethnic prejudice must be at work in the selection. Such an observer, however, must not be very familiar with the demographic and occupational patterns of the United States. Despite our reputation as a melting pot, different neighborhoods and different occupations often attract distinct racial or ethnic groups; so that if a particular kind of crime comes into vogue it may well be a feature of a neighborhood or of an occupation marked by one or another ethnic or racial characteristic.

In effect, as applied in this case, the defendants' hypothetical is an argument that the minorities of the inner city of Los Angeles must be denied the protection of law enforcement by the federal government because the likely suspects are overwhelmingly apt to be members of the minority living in that area. The defense is a grave perversion of proper sensitivity to the civil liberty of minorities. If any policy of government had a racially discriminatory effect, it would be to deny law enforcement on the grounds of a specious claim of racial discrimination.

Id. at 1185-86.

Thorpe's related argument regarding the racial composition of the counties selected by the U.S. Attorney for participation in the PSN initiative in the Eastern District of Michigan is similarly unavailing. He presented no evidence, in other words, that these particular counties were selected on the basis of their racial composition as opposed to being selected because of their level of gun violence.

In any event, the district court's order granting Thorpe's discovery motion was clearly erroneous on its face. Notwithstanding the fact that the court ostensibly analyzed Thorpe's motion under the prima facie standard that applies to the ultimate merits, rather than under the less onerous "some evidence" standard that applies to discovery, the court's analysis did not support and, in fact, could not have supported its conclusion. The district court, in assessing the evidence of discriminatory intent, stated in no uncertain terms:

Defendant has not addressed the second element—that the prosecution was initiated with a discriminatory purpose. . . . There is a strong presumption that the prosecutor acted in good faith, to overcome that presumption clear and convincing evidence to the contrary must be presented. . . . Defendant has not provided any evidence to overcome the presumption of good faith. The second element of the prima facie case was not established.

So even if the court had applied the correct "some evidence" standard for discovery purposes, Thorpe could not have prevailed.

The district court nonetheless granted discovery, relying solely on the fact that the government had produced some of the same documents requested by Thorpe for an in camera review in a different case (United States v. Nixon, 315 F. Supp. 2d 876 (E.D.Mich.2004) (opinion ordering discovery)), before a different judge (Judge Feikens), and regarding a different issue (whether any remedy existed in federal court for two PSN defendants whose counsel had rendered ineffective assistance relating to plea agreements offered to them in state court). Because Armstrong does not afford district courts the discretion to grant discovery in selective-prosecution cases where the rigorous "some evidence" standard has not been met, the granting of Thorpe's motion for discovery necessarily constituted an abuse of discretion. The district court's "supervisory powers" justification, set forth in its order dismissing the indictment against Thorpe, does not alter this conclusion for the related reason that validating it would effectively render Armstrong moot. Cf. Bank of Nova Scotia v. United States, 487 U.S. 250, 254, 108 S. Ct. 2369, 101 L. Ed. 2d 228 (1988) ("[I]t is well established that even a sensible and efficient use of the supervisory power . . . is invalid if it conflicts with constitutional or statutory provisions.") (citation and quotation marks omitted).

Finally, Thorpe argues that an in camera review of the materials ordered by the district court for discovery adequately addresses the government's concerns regarding the disclosure of confidential information relating to its charging decisions, thereby rendering the discovery order harmless. The pre-Armstrong caselaw provides a degree of support for this argument. See, e.g., United States v. Oaks, 508 F.2d 1403, 1405 (9th Cir.1974) (noting that the district court, in ordering discovery pursuant to a selective-prosecution claim, "can minimize the risk to the government by holding the proceeding in camera and issuing appropriate protective orders"); see also Ponte v. Real, 471 U.S. 491, 514-15, 105 S. Ct. 2192, 85 L. Ed. 2d 553 (1985) (Marshall, J., dissenting) ("The in camera solution has been widely recognized as the appropriate response to a variety of analogous disclosure clashes involving individual rights and government secrecy needs.").

In light of Armstrong and Bass, however, the in camera argument is insufficient to overcome the "some evidence" threshold for discovery in a selective-prosecution case. This court in Bass had upheld the district court's grant of discovery and remanded the case to allow the government to submit materials for the court's in camera review only. United States v. Bass, 266 F.3d 532, 540 (6th Cir.2001). The Supreme Court nonetheless reversed, making no mention of the in camera review requirement, much less conceding that such a review would be sufficient to mitigate the burden of discovery imposed on the government. See generally Bass, 536 U.S. 862, 122 S. Ct. 2389; cf. Armstrong, 517 U.S. at 469, 116 S. Ct. 1480 ("Discovery thus imposes many of the costs present when the Government must respond to a prima facie case of selective prosecution. It will divert prosecutors' resources and may disclose the Government's prosecutorial strategy."). Thorpe's in camera argument, therefore, does not overcome the district court's abuse of discretion in granting his discovery motion.

Thorpe also advances the related argument that the formal reporting requirements of the PSN program minimize whatever burden his discovery request might impose on the government. But the simple fact that local U.S. Attorneys in charge of implementing the PSN program report to the Department of Justice in Washington, D.C. on a regular basis does not by itself make the content of their reports public information that is accessible through court-ordered discovery.

Because the district court's discovery order was improper, the court a fortiori had no proper basis to dismiss the indictment. Cf. Armstrong, 517 U.S. at 461 n. 2, 116 S. Ct. 1480 ("We have never determined whether dismissal of the indictment, or some other sanction, is the proper remedy if a court determines that a defendant has been the victim of prosecution on the basis of his race. Here, it was the government itself that suggested dismissal of the indictments to the district court so that an appeal might lie.") (quotation marks omitted).

For all of the reasons set forth above, we REVERSE the judgment of the district court granting discovery, REINSTATE the indictment against the Thorpe, and REMAND the case to the district court for further proceedings consistent with this opinion.