United States v. Gary Wyche

SRINIVASAN, Circuit Judge,

concurring in part and dissenting in part:

I concur in the court’s opinion insofar as it affirms as to appellant Wyche’s sentence. I respectfully dissent from Part II.B of the court’s opinion, however, in which the court affirms as to appellant Smith’s sentence on harmless-error grounds even though the government has made no request that we do so.

As the Court explains, a district court’s drug quantity finding in a sentence modification proceeding under 18 U.S.C. § 3582 (c)(2) “cannot be inconsistent with factual determinations made by the original sentencing court.” Ante, at 1293; accord United States v. Kennedy, 722 F.3d 439, 442 (D.C.Cir.2013). The same principle necessarily governs in the context of successive § 3582(c)(2) motions: a district court’s quantity finding in the later sentence modification proceeding cannot be inconsistent with factual determinations made by the court that decided the defendant’s earlier motion. Here, when the district court modified Smith’s sentence in 2008, the court assigned Smith a base offense level of 34, which at the time corresponded to offenses involving at least 500 grams but less than 1.5 kilograms of crack. Ante, at 1291, 1295-96 & n. 12; see U.S.S.G. § 2D1.1(c)(3) (2007). But when the district court in 2012 later denied Smith’s subsequent § 3582(c)(2) motion, the court held that Smith’s base offense level is 38, which now corresponds to offenses involving 8.4 kilograms of crack or more. See U.S.S.G. § 2Dl.l(c)(l) (2011). The 2012 determination that Smith’s base offense level is 38 cannot be squared with the 2008 determination that his base offense level is 34. The 2012 district court’s order therefore should be vacated, and the case should be remanded so that the district court can recalculate Smith’s base offense level consistent with the 2008 decision.

The majority understandably does not defend the 2012 court’s disregard of the drug quantity range established by the 2008 court. Instead, the majority affirms the decision on harmless-error grounds, reasoning as follows. Under the 2011 crack cocaine amendments, the base offense level of 34 now corresponds to at least 840 grams but less than 2.8 kilograms of crack, while the base offense level of 32 corresponds to at least 280 grams but less than 840 grams of crack. Smith qualifies for a sentence reduction only if his new base offense level is 32. Even though the 2012 district court could not attribute more *1299 than 1.5 kilograms of crack to Smith, the court still could have denied Smith’s motion if it had held him responsible for anywhere between 840 grams and 1.5 kilograms of crack. And because the 2012 court concluded that Smith was responsible for more than 8.4 kilograms of crack, the court necessarily would have held him responsible for at least 840 grams of crack. See ante, at 1296-97.

Significantly, the government did not advance that (or any) harmless-error argument in its submissions to this court. And although we have “discretion to determine sua sponte whether an error is harmless,” United States v. Davis, 596 F.3d 852, 861 (D.C.Cir.2010), the exercise of that discretion should be carefully circumscribed, see United States v. Pryce, 938 F.2d 1343, 1348 (D.C.Cir.1991) (opinion of Williams, J.). In my view, the harmlessness of the error in this case is not sufficiently obvious or readily discernible from the record to warrant a sua sponte harmless-error determination.

The 2012 district court did not consider the implications of the 2008 district court’s decision to sentence Smith based on a drug quantity range of 500 grams to 1.5 kilograms. The majority assumes that, even if the 2012 district court had considered the 2008 court’s decision, the 2012 court would still have held Smith responsible for a disqualifying quantity of drugs. I am not so certain. The judge who granted Smith a sentence modification in 2008 was intimately familiar with Smith’s case, having presided over it for nine years and having issued two previous decisions concerning Smith’s sentence. That judge and the Probation Office (which recommended the reduction) both concluded in 2008 that Smith’s sentence should be modified on the basis of an offense level corresponding to 500 grams to 1.5 kilograms of crack. The United States Attorney’s Office, in deeid-ing not to oppose Smith’s 2008 motion for a sentence reduction, evidently thought the same. The 2012 district court might have scrutinized the record to assess why all of those actors assumed that Smith should be held responsible only for a limited quantity of crack cocaine rather than for more than 8.4 kilograms. Cf. United States v. Singleton, 759 F.2d 176, 178 (D.C.Cir.1985) (noting that under law-of-the-case doctrine, a subsequent court generally “refuse[s] to reopen what has been decided” in the same case, including “questions decided by necessary implication” (internal quotation marks omitted)).

The 2012 district court’s failure to consider the implications of the 2008 court’s decision is especially significant because the 2012 court’s drug-quantity finding already rests on an uncertain foundation. In affirming the 2012 district court’s attribution to Smith of over 8.4 kilograms of crack, the majority states that the “[t]he original sentencing court [in 1989] found that the conspiracy handled roughly 4 kilograms of cocaine base per month during Smith’s five-plus months’ participation.” Ante, at 1297. And the 2012 district court, the majority reasons, was “entitled to rely” on that figure. Id. at 1297. If the original sentencing court had indeed made a four-kilograms-a-month finding, and if the court had further found that all of the crack distributed by the conspiracy during those months should be attributed to Smith, this might well be the sort of case in which a sua sponte harmless-error inquiry would be appropriate. Cf. United States v. Stover, 329 F.3d 859, 874 (D.C.Cir.2003) (per curiam) (“The District Court must make individualized findings linking each appellant’s scope of participation in the conspiracy with the quantum of drugs attributed to him.” (alterations and internal quotation marks omitted)). But the 1989 sentencing court did not make such a finding. Rather, the sentencing judge stated: “I don’t *1300 know exactly how much crack was distributed during the period that Mr. Smith was a participant, but we have heard about 4 kilos a month ...” (emphasis added). The judge then explained that, because four kilograms a month is “so much more” than 500 grams, “I find that [Smith] did distribute more than 500 grams of crack during the period when he was a member.”

To be sure, the original sentencing court might have found Smith responsible for distributing more than 840 grams if it had been called upon to make such a determination. But it was not. The court found only that Smith distributed more than 500 grams, expressly declining to reach any conclusion concerning the amount of crack distributed by the conspiracy (and also making no finding about Smith’s own responsibility for drugs distributed by the conspiracy). The 2012 district court therefore was mistaken when it said that the 1989 court had “estimated that the conspiracy handled roughly 4 kilograms a month.” And the 2012 district court otherwise gave no explanation of the basis for its conclusion that the drug conspiracy handled more than 8.4 kilograms of crack while Smith was a member (or, if it did, why Smith should be held responsible for the full quantity).

In my view, none of the other portions of the record identified by the majority demonstrate Smith’s responsibility for 840 grams of crack with sufficient clarity to justify a sua sponte harmless-error affir-mance. While the 1989 sentencing court found “substantial evidence” that Smith played a managerial role in the drug enterprise, see ante, at 1297, that does not speak directly to the question of drug quantity. The trial testimony concerning Smith’s role in the drug conspiracy, see id. at 1297-98, likewise fails to establish Smith’s responsibility for at least 840 grams of crack. One witness says she saw Smith counting eight or nine large ziplock bags containing an unidentified white substance. Another witness describes a time when her apartment was “like wall-to-wall cocaine,” but she also suggests that Smith had already left the apartment by that time. A third witness says that Smith personally gave him $3,000 worth of crack, and a fourth witness says that Smith was present while another member of the drug enterprise counted $26,000 of cash, but neither witness links the dollar amount with a specific drug quantity. Our opinions in other cases describe roughly contemporaneous crack transactions in Washington, D.C., at prices well over $100 per gram, e.g., United States v. Brown, 16 F.3d 423, 425 (D.C.Cir.1994); United States v. Beckham, 968 F.2d 47, 53 (D.C.Cir.1992), so even if Smith could be held responsible for $3,000 worth of crack (or $26,000 worth of crack), that would not necessarily demonstrate his responsibility for 840 grams of crack or more.

None of this is intended to suggest that, if the case were remanded, the district court would be likely to hold Smith responsible for less than 840 grams of crack. Indeed, it may well be that the district court would examine the entire record and prior course of proceedings — including the 2008 district court’s decision — and decide to attribute to Smith more than (perhaps far more than) 840 grams. But I do not consider the matter sufficiently free from doubt to justify a sua sponte harmless-error determination. We ordinarily leave it to district courts in the first instance to scrutinize the record and make drug quantity findings. And we ordinarily would be reluctant to canvass the record ourselves to affirm a decision on harmless-error grounds when the government has made no argument asking us to do so. Respect *1301 fully, I believe the ordinary course is the appropriate one here.