Skip to main content


Reset A A Font size: Print

United States Court of Appeals, Sixth Circuit.

UNITED STATES of America, Plaintiff-Appellee, v. David Jamar ARMSTRONG, Defendant-Appellant.

No. 18-5079

Decided: April 03, 2019

Before: DONALD, LARSEN, and NALBANDIAN, Circuit Judges. ON BRIEF: Anne Buckleitner, SMIETANKA, BUCKLEITNER, STEFFES & GEZON, Grandville, Michigan, for Appellant. Charles P. Wisdom, UNITED STATES ATTORNEY'S OFFICE, Lexington, Kentucky, Anthony J. Bracke, UNITED STATES ATTORNEY'S OFFICE, Ft. Mitchell, Kentucky, for Appellee.


David Armstrong sold a confidential informant about three grams of heroin during three controlled buys. He pleaded guilty to one count of distribution, and the district court sentenced him to thirty-seven months in prison. That sentence was based, in part, on the district court's finding that he sold around seventy grams of heroin to the informant over the course of two years. He contends that the district court's finding was erroneous. We affirm.


After conducting several staged drug deals with an informant, the Government charged David Armstrong with three counts of distributing controlled substances in violation of 21 U.S.C. § 841(a)(1). The informant paid Armstrong $ 140 for about one gram of heroin during each deal. Armstrong eventually pleaded guilty to one of the three distribution charges.

At sentencing, the parties disagreed over the quantity of heroin that the court should use to calculate Armstrong's sentencing range. Under the Sentencing Guidelines, district courts must consider the defendant's entire relevant conduct beyond the scope of the conviction. See U.S.S.G. § 1B1.3(a)(2). That means a defendant convicted of selling one gram of heroin as part of a larger drug-trafficking operation faces a higher sentence than someone who—though convicted of the same offense—had no other criminal activity. Here, the informant claimed she purchased about one gram of heroin from Armstrong 70 times over eighteen to twenty-four months. So the Government argued that the court should calculate his sentencing range based on 70 grams of heroin. See U.S.S.G. § 2D1.1(c)(10). The probation office agreed. But Armstrong claimed he only sold the informant heroin a few times and in much smaller quantities. He asked for an offense level based on fewer than ten grams. See U.S.S.G. § 2D1.1(c)(14).

The district court took evidence to resolve the dispute. Neither Armstrong nor the informant testified, but the court considered their out-of-court statements to make its findings. On top of that, Joe Schulkens, one of the two officers who handled the informant, testified about his conversations with her before setting up the controlled buys.

Schulkens handled the informant for the second and third buys. He came in after the first purchase because the officer originally handling the matter left the force. When Schulkens took over, he interviewed the informant about her history with Armstrong. The informant told him she “had been buying heroin from him for a year and a half to two years.” During that period, she purchased a gram of heroin about 70 times. Schulkens explained that they based the controlled purchases off this information to avoid drawing suspicion—asking an informant to purchase an unusually large quantity of drugs might tip the suspect off. This apparently worked. The informant completed two more purchases of heroin (one week apart), each for about a gram.

On cross, Armstrong's counsel tried to undermine the credibility of Schulkens's claim that he patterned the controlled buys off the informant's history. Armstrong stated in a letter to the court that he sold the informant heroin only a few times before the controlled buys, and that each time she bought much less than one gram. But on these three occasions, Armstrong said, the informant told him she was buying for a friend. That lessened any suspicion he might otherwise have had. Schulkens admitted that he did not know what the informant told Armstrong and it “could very well be” that she told him this story. R. 33, Sentencing Hr'g Tr. at 17, PageID 181.

The district court judge then made a credibility determination. He found that the informant's out-of-court statements were more reliable than Armstrong's. Several factors went into this decision, which the judge stated on the record. First, unlike Armstrong, the informant had no motive to exaggerate the number of prior transactions. Schulkens testified that the informant was “working off charges” and that she received no additional benefit from inflating her history with Armstrong.1 Id. at 14, PageID 178. Armstrong, on the other hand, had a strong incentive to lie to reduce his Guidelines range. Second, the police corroborated the informant's claim that she purchased heroin in one-gram quantities by conducting three transactions for the same amount. Based on that corroboration and the informant having no reason to lie, the court found her statements were sufficiently reliable to establish Armstrong's relevant conduct for sentencing.

The district court then sentenced Armstrong to a within-Guidelines term of imprisonment of 37 months. The only issue he raises on appeal is the judge's decision to rely on the informant's out-of-court statement that she purchased about 70 grams of heroin from Armstrong over two years. Because those statements were unreliable, he argues, the district court improperly inflated his sentencing range under the Guidelines.


District courts routinely rely on hearsay for the factfinding part of a sentencing decision. So long as the information has “some evidentiary basis” to satisfy a “minimal indicium of reliability,” the district court can consider it without regard for the rules of evidence. See United States v. Silverman, 976 F.2d 1502, 1504 (6th Cir. 1992) (quoting United States v. Smith, 887 F.2d 104, 108–09 (6th Cir. 1989)); accord U.S.S.G. § 6A1.3(a). The indicia-of-reliability standard is a “relatively low hurdle.” United States v. Moncivais, 492 F.3d 652, 659 (6th Cir. 2007). It allows courts to consider “[a]ny information” that may be reliable. See U.S.S.G. § 6A1.3 cmt. (2016). And on top of that, we review those reliability decisions under the highly deferential, clearly erroneous standard. See United States v. Gibson, 985 F.2d 860, 864 (6th Cir. 1993). That means we will reverse the district court's finding of reliability only if it leaves us “with the definite and firm conviction that a mistake has been committed.” United States v. Darwich, 337 F.3d 645, 663 (6th Cir. 2003) (quoting United States v. Latouf, 132 F.3d 320, 331 (6th Cir. 1997)).2

Those rules change slightly when the hearsay comes from a confidential informant—a not unusual occurrence. District courts can rely on hearsay from an “unidentified informant” only if “there is good cause for the non-disclosure of the informant's identity and there is sufficient corroboration by other means.” Silverman, 976 F.2d at 1504 (quoting Smith, 887 F.2d at 108–09); accord U.S.S.G. § 6A1.3(a) cmt. This raises the burden a bit for the Government if it wants to prove sentencing facts through a confidential informant. There must be good reason for not revealing the informant's identity, and there must be “sufficient corroboration” of the informant's statements.

Even under the (slightly) elevated standard for relying on hearsay from a confidential informant, there was no error here. The informant told Schulkens that she had known Armstrong for several years and she purchased a gram of heroin from him about 70 times. Schulkens and the other officer corroborated that information by arranging three transactions, each for about one gram. Each purchase went off without a hitch and exactly how the informant described. That sufficiently corroborates her out-of-court statements for U.S.S.G. § 6A1.3(a).

Armstrong argues that the statements were unreliable because the Government failed to corroborate the informant's specific claim that she made about 70 purchases over two years. That slices the meaning of “sufficient corroboration” a little too thin. “Sufficient” means just that: sufficient. It does not mean exhaustive. It does not mean perfect. It's enough corroboration to satisfy the indicia-of-reliability standard, a standard lower than even a preponderance of evidence. See Black's Law Dictionary 1661 (10th ed. 2014) (defining “sufficient” as “[a]dequate” or “of such quality ․ as is necessary for a given purpose”); Moncivais, 492 F.3d at 658–59. And an unidentified informant's claims meet this standard when the district court can reasonably corroborate enough of the information to make a general finding about the informant's reliability. United States v. Rogers, 1 F.3d 341, 344 (5th Cir. 1993).3 Particular corroboration for each claim is not required.

Rogers is particularly persuasive on this point. There, the defendant made the same argument Armstrong does today. Several unidentified informants provided information about the scope of the defendant's drug-trafficking activities. Some of that information turned out to be true, some of it not true. But overall, the district court found the unidentified informants credible and relied on their uncorroborated claims about drug quantities. The Fifth Circuit affirmed, holding that “sufficient indicia of reliability accompanied the [informants'] reports that the district court was justified in relying on them to determine the quantity of drugs with which Rogers had been associated without corroboration of the specific amounts alleged.” Rogers, 1 F.3d at 344 (emphasis added).

The same is true here. The informant provided several pieces of information that both Armstrong and the controlled purchases corroborated. Based on that, the district court was well within its discretion to credit the reliability of the informant's statements, including portions that lacked particularized corroboration. By doing so, we cannot say the court committed clear error.4

Although couched as an argument about inadequate corroboration, much of Armstrong's argument really focuses on whether the informant's claims are credible. He emphasizes the dispute over whether she told Armstrong that she was purchasing for a friend. He also speculates that she might be motivated to exaggerate the claim for personal reasons. And he says that her claims were “impossible” to rebut because they contained no allegation of a “specific context as to time and place.” Appellant's Br. at 14. All these issues go to the informant's credibility, not whether her claims were sufficiently corroborated.

But tellingly, Armstrong chose not to call the informant to testify at the hearing. He knew who she was, and his counsel told the court that he made the strategic decision not to have her testify. So while Armstrong bemoans the fact that “he never had the opportunity” to cross-examine the informant because she “never appeared in court,” Appellant's Br. at 27, he leaves out that he made no effort to do so. Armstrong intentionally gave up the right to cross examination—“the greatest legal engine ever invented for uncovering the truth.” Doe v. Baum, 903 F.3d 575, 581 (6th Cir. 2018) (internal quotation marks omitted). And by doing so, he abandoned his right to cry foul over his inability to rebut her claims.

Armstrong also argues that the district court erred by failing to provide “good cause” for the non-disclosure of the informant's identity. The record lacks any indication that Armstrong raised this issue below, so our review is limited to plain error. See United States v. Barnett, 398 F.3d 516, 525 (6th Cir. 2005). We find none here. Plain error requires a showing of error “so obvious or clear” that it “affected the fairness, integrity, or public reputation of the judicial proceedings.” United States v. Vonner, 516 F.3d 382, 386 (6th Cir. 2008) (en banc). Even if the district court should have made a finding of good cause for not revealing the informant's identity, that error is not so “obvious or clear” when the defendant had admitted that he knew who she was.

All this dovetails nicely to our final point. Armstrong's appeal centers on the non-disclosure of the informant's identity. He calls her “unidentified” or “nondisclosed” throughout his briefs, suggesting that the Government's secrecy prejudiced him. See, e.g., Appellant's Br. at 14, 27. And the elevated standard that requires “sufficient corroboration” to rely on hearsay statements arises only when the informant's identity has not been disclosed. See Silverman, 976 F.2d at 1504. Yet Armstrong told the court that he knew her identity and deliberately chose not to have her testify. He now tries to sandbag the Government (and the district court) by arguing that it was unfair he could not cross examine her—a decision he made.

Confidential informants pose a unique problem for district courts at sentencing. When the Government asks for a higher sentence based on information provided by an unidentified informant, the defendant might reasonably object: Who is this person, your honor? How can you sentence me based on uncorroborated claims from someone without telling me who it is? That's the scenario we must guard against by requiring sufficient corroboration and good reason for the non-disclosure of an informant's identity. But that's not the scenario here. Armstrong was not worried about accusations from an unimpeachable mystery witness. He knew who she was, and he admitted he sold her drugs on previous occasions. Whatever concerns over reliability we might ordinarily have with confidential informants, they are greatly diminished—if not extinguished—here. So while we are satisfied that the Government sufficiently corroborated the informant's claims, we are skeptical that such heightened evidence of reliability was even necessary.

* * *

We affirm.


1.   Nor would it have affected the charges that the Government could bring against Armstrong. Schulkens testified that they would typically stage enough controlled buys to reach the total quantity of drugs to bring the charges. So if an informant purchased 0.20 grams in each buy, they would arrange additional transactions until they reached the desired amount.

2.   Some decisions from this circuit have inconsistently applied the clear-error standard for these kinds of decisions. See United States v. Santana, 723 F. App'x 331, 337–38 (6th Cir. 2018) (collecting cases). But that does not mean we are free to adopt whichever standard we find the most persuasive. Gibson appears to be the first published decision after Silverman, and so we must follow it. See Gaddis ex rel. Gaddis v. Redford Twp., 364 F.3d 763, 770 (6th Cir. 2004); see also Santana, 723 F. App'x at 338.

3.   The Guidelines cites Rogers as the first case for supporting the “sufficient corroboration” requirement. See U.S.S.G. § 6A1.3 cmt.

4.   Our decision in United States v. Gibbs, 182 F.3d 408 (6th Cir. 1999), does not alter the analysis here. Gibbs did not involve the reliability of an unidentified informant's hearsay testimony, so the court never discussed the meaning of “sufficient corroboration” in this context. But more importantly, the court reversed the district court because the hearsay statements were based on conjecture, not because they lacked sufficient corroboration. The Government provided no information about how the informant knew the amount of drugs the defendant trafficked. So we reversed with instructions to “elicit more information with respect to [the informant's] original statements and the source of [his] knowledge ․” Id. at 442. Unlike Gibbs, we know the source of the informant's knowledge here—she participated in the sales and has direct knowledge about the transactions. While district courts cannot rely on speculation from a witness to establish sentencing facts, the informant's claims here are far from that.

NALBANDIAN, Circuit Judge.

Copied to clipboard