United States v. Brent Brewbaker

87 F.4th 563

Opinion

PUBLISHED

UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT

No. 22-4544

UNITED STATES OF AMERICA, Plaintiff - Appellee,

v.

BRENT BREWBAKER, Defendant - Appellant.

Appeal from the United States District Court for the Eastern District of North Carolina, at Raleigh. Louise W. Flanagan, District Judge. (5:20-cr-00481-FL)

Argued: September 22, 2023 Decided: December 1, 2023

Before GREGORY and RICHARDSON, Circuit Judges, and Patricia Tolliver GILES, United States District Judge for the Eastern District of Virginia, sitting by designation.

Reversed in part, affirmed in part, and remanded by published opinion. Judge Richardson wrote the opinion, in which Judges Gregory and Giles joined.

ARGUED: Elliot Sol Abrams, CHESHIRE, PARKER, SCHNEIDER, PLLC, Raleigh, North Carolina, for Appellant. Peter Matthew Bozzo, UNITED STATES DEPARTMENT OF JUSTICE, Washington, D.C., for Appellee. ON BRIEF: Jonathan S. Kanter, Assistant Attorney General, Doha G. Mekki, Principal Deputy Assistant Attorney General, Maggie Goodlander, Deputy Assistant Attorney General, Adam Ptashkin, Rachel Kroll, Alison Friberg, Daniel E. Haar, Stratton C. Strand, Scott McAbee, Patrick M. Kuhlmann, Antitrust Division, UNITED STATES DEPARTMENT OF JUSTICE, Washington, D.C., for Appellee.

2

RICHARDSON, Circuit Judge:

Brent Brewbaker appeals from his conviction of a per se antitrust violation under § 1 of the Sherman Act, as well as five counts of mail and wire fraud. Before his five-day trial, Brewbaker asked the district court to dismiss the Sherman Act count for failing to state an offense. Fed. R. Crim. P. 12(b)(3)(B)(v). The district court didn’t. But it should have—caselaw and economics show that the indictment failed to state a per se antitrust offense as it purported to do. So we reverse Brewbaker’s Sherman Act conviction. But we affirm his fraud convictions and remand for resentencing. I. Background Contech Engineering Solutions manufactured and sold corrugated steel and aluminum pipe and plate. Starting in 1988, Contech relied on its distributor and exclusive dealer in North Carolina, Pomona Pipe Products, for one way to sell its goods.

One element of Contech and Pomona’s manufacturer-distributor relationship was their involvement in North Carolina Department of Transit (“NCDOT”) aluminum- structure projects. 1 These projects, scattered throughout North Carolina, involved installing aluminum structures to prevent flooding. To award these projects, NCDOT used a bidding process. There were only three consistent bidders: Contech, Pomona, and Lane Enterprises.

But the apparent contest between Contech and Pomona was really a win-win for both companies. When Pomona won a NCDOT project, it would complete the required

1

It’s unclear when Contech and Pomona both started bidding on NCDOT projects.

But it was by 2007 at the latest.

3

services using Contech’s aluminum. See J.A. 1843 (aluminum from Contech accounted for around 75% of Pomana’s bid). And if Contech won, the opposite was true—it’d supply the aluminum, but Pomona would provide the necessary services. So in the end, as long as one of them won, both companies got paid. And they often won, as Lane’s bids were consistently higher than either Contech’s or Pomona’s.

One consequence of Contech and Pomona’s win-win situation was that they had to communicate to calculate their bids. Neither company could submit a bid otherwise; Contech couldn’t come up with its bid price without knowing how much Pomona would charge for its services, just as Pomona couldn’t come up with its bid price without knowing how much Contech would charge for the aluminum. Thus, up until 2009, this communication was the norm.

In 2009, however, the norm changed. That year, Brewbaker—then a sales manager—was put in charge of Contech’s NCDOT bids. And when he took charge, he saw an opportunity to strengthen Contech’s relationship with its long-time distributor by ensuring Pomona won the NCDOT projects.

For Pomona to win, Brewbaker had to make sure Contech lost. So, when he calculated Contech’s bid price, Brewbaker didn’t just ask Pomona what it’d charge for its services. Instead, he—or another Contech employee at his direction—would ask Pomona for its total bid price. Then, Contech would add a small percentage to Pomona’s number to arrive at Contech’s own bid. This ensured that Pomona’s bid was always lower than Contech’s. And because Lane’s bids were nearly always higher than both Pomona’s and Contech’s, Pomona would generally win.

4

Beyond pleasing Pomona, Brewbaker saw that submitting losing bids had two other perks. First, it allowed Contech to stay on NCDOT’s “emergency bid list” that would qualify Contech for additional aluminum supply business if it came up. Second, it would allow Contech’s losing bids to serve as backups—if Pomona lost a bid for some technical reason, Contech would still get the project and still get paid.

Naturally, Pomona was all for winning the NCDOT bids, so it went along with Brewbaker’s plan. Thus, starting around 2009, Pomona routinely shared its NCDOT bid prices with Contech, and Contech used the bids to calculate its own, higher bids. All the while, Contech and Pomona were submitting certifications along with their bids that stated the bids were “submitted competitively and without collusion.” E.g., J.A. 685.

Also during this time, Brewbaker tried to cover his tracks. He deleted conversations between Pomona and Contech employees, otherwise opted for phone calls over digital paper trails, and made sure that the percent he added to Pomona’s bid varied to avoid raising “red flag[s]” to NCDOT. J.A. 2315. This may have stemmed from Contech’s antitrust training, which cautioned against getting information from competitors.

Despite Brewbaker’s efforts, the FBI and the Department of Justice’s Antitrust Division eventually caught up with him. In October 2020, a grand jury indicted both him and Contech on six counts. Count One alleged a per se violation of the Sherman Act’s § 1, 15 U.S.C. § 1, while Counts Two through Six alleged federal mail- and wire-fraud violations, 18 U.S.C. §§ 1341, 1343.

To support the Sherman Act count, the indictment alleged that Contech and Brewbaker “rig[ged] bids.” E.g., J.A. 50. The speaking indictment specified:

5

• Contech “ma[de] products such as . . . aluminum pipe and fittings,” J.A. 45;

• Pomona 2 was “an aluminum structure design and installation company” that also “served as a dealer for” Contech, J.A. 46;

• Contech “regularly sold aluminum pieces” to Pomona which Pomona “used . . . to complete work on behalf of NCDOT, including for aluminum structure projects,” J.A. 46;

• Contech and Pomona (among others) submitted bids for NCDOT aluminum structure projects; and

• Under an agreement between Contech and Pomona, Contech and Brewbaker obtained Pomona’s bid price and added a nominal amount to create Contech’s own, intentionally losing bid.

According to the indictment, these allegations showed Contech and Brewbaker’s agreement with Pomona “was a per se unlawful, and thus unreasonable, restraint of interstate trade and commerce.” J.A. 50.

As for the fraud counts, the indictment alleged that Contech and Brewbaker misled NCDOT by submitting intentionally losing bids and by falsely certifying that the bids were submitted competitively and without collusion. The indictment asserted that these certifications were false and fraudulent because Contech colluded with Pomona on the bid price and submitted a non-competitive bid that was intentionally higher than Pomona’s. As alleged, Contech “held itself out as a competitor to” Pomona when submitting bids,

The indictment didn’t refer to Pomona by name. Instead, it called Pomona

2

“Company A.” But, at trial, Company A’s identity was revealed.

6

even though Contech “also benefitted when [Pomona] won . . . because it supplied aluminum pieces to [Pomona] for use in” the projects. J.A. 56.

In December 2020, Contech moved “to apply the rule of reason” under Federal Rule of Criminal Procedure Rule 12(a)(1). J.A. 63–64. Brewbaker joined the motion. Contech and Brewbaker argued that the indictment merely alleged that Contech “submitted an additional direct bid that would not undercut its dealer’s price.” J.A. 75. This, according to Contech, wasn’t a per se § 1 violation but a business practice that should be analyzed under the rule of reason.

Contech also explained that the indictment didn’t allege a horizontal restraint, but a vertical one. A horizontal restraint is one between competitors, while a vertical restraint is one between firms at different levels of distribution. 3 See Ohio v. Am. Express Co., 138 S. Ct. 2274, 2283 (2018). And the Supreme Court has held that only some horizontal restraints are subject to the per se rule. Vertical restraints, on the other hand, are subject to the rule of reason. See Leegin Creative Leather Prods., Inc. v. PSKS, Inc., 551 U.S. 877, 886, 907 (2007). From Contech and Brewbaker’s perspective, the indictment alleged a restraint between Contech as supplier and Pomona as distributor, so the rule of reason should apply.

3

To illustrate the horizontal-vertical distinction, consider the sale of Nike shoes.

Foot Locker and Dick’s Sporting Goods compete to sell Nike shoes to consumers. So an agreement between them to set the price of shoes would be a horizontal restraint. But an agreement between Nike and Foot Locker to set the price of shoes that Nike supplies to Foot Locker for sale to consumers would be vertical.

7

In support of its motion, Contech submitted various exhibits. Some were affidavits that addressed the factual underpinnings of the case against Contech and Brewbaker. But there was also an affidavit by antitrust professor Dr. Kenneth G. Elzinga. In that affidavit, Dr. Elzinga described the economics behind Contech and Pomona’s relationship. In short, he explained that it was an example of a so-called “dual distribution” arrangement, in which “a manufacturer and its distributor both offer prices for the manufacturer’s products.” J.A. 108. And the alleged bid rigging within this arrangement, Dr. Elzinga concluded, wouldn’t “always or almost always” hurt competition. J.A. 108.

The district court, however, denied Contech and Brewbaker’s motion without considering its exhibits. Treating the motion as a motion to dismiss Count One for failure to state an offense, Fed. R. Crim. P. 12(b)(3)(B)(v), the district court determined it was prohibited from looking at such “extrinsic evidence,” J.A. 968–69. Then, it concluded that the indictment alleged on its face a horizontal bid-rigging restraint between competitors subject to the per se rule. So it denied the motion.

Soon after, Contech pleaded guilty to Counts One and Two. But Brewbaker proceeded to trial. During Brewbaker’s trial, the jury heard testimony that established the facts as described above (e.g., Brewbaker’s concoction and submission of intentionally losing NCDOT bids, his efforts to conceal the scheme, his certifications that the bids were submitted competitively and without collusion, etc.). They didn’t hear evidence, however, as to the procompetitive intent or effects of Contech and Pomona’s particular setup. That evidence was irrelevant once the district court applied the per se rule because a restraint subject to the per se rule is necessarily anticompetitive. See United States v. W.F. Brinkley

8

& Son Constr. Co., 783 F.2d 1157, 1162 (4th Cir. 1986). And the jury was instructed that they “need not be concerned with whether the agreement was reasonable or unreasonable, the justifications for the agreement, or the harm, if any done by it,” when determining Brewbaker’s guilt under the Sherman Act. J.A. 2593.

In the end, the jury found Brewbaker guilty on all counts. He was sentenced to 18 months’ imprisonment. This timely appeal followed. II. Discussion On appeal, Brewbaker advances several arguments against his Sherman Act conviction. One is that the indictment should have been dismissed because it did not state a per se Sherman Act offense. We agree and therefore reverse his Sherman Act conviction. But the Sherman Act jury instructions didn’t so infect the jury’s consideration of the mail- and wire-fraud counts as to require their reversal. So we affirm those convictions.

A. The Sherman Act count should have been dismissed.

A criminal indictment—like a civil complaint—should be dismissed for failing “to state an offense.” Fed. R. Crim. P. 12(b)(3)(B)(v); cf. Fed. R. Civ. P. 12(b)(6). And an indictment—much like a civil complaint—must contain “a plain, concise, and definite written statement of the essential facts constituting the offense charged.” Fed. R. Crim. P. 7(c)(1); cf. Fed. R. Civ. P. 8(a)(2). Despite the similarity in the criminal and civil rules, we rarely see district courts dismiss indictments for the failure to state an offense. See James M. Burnham, Why Don’t Courts Dismiss Indictments?, 18 Green Bag 2d 347 (2015). But

9

district courts have as much of a responsibility to police criminal indictments as they do civil complaints.

Whether the district court grants or denies a motion to dismiss an indictment, we review the court’s legal conclusions de novo and any factual findings for clear error. United States v. Perry, 757 F.3d 166, 171 (4th Cir. 2014). An indictment may legally fail to state an offense by omitting a necessary element. United States v. Hooker, 841 F.2d 1225, 1227– 28 (4th Cir. 1988) (en banc). But it may also fail to state an offense if “the allegations therein, even if true, would not state an offense.” United States v. Thomas, 367 F.3d 194, 197 (4th Cir. 2004). And whether the allegations fail to state an offense is a legal question that we review de novo. United States v. Good, 326 F.3d 589, 591–92 (4th Cir. 2003).

To state an offense under § 1 of the Sherman Act, an indictment must allege the defendant (1) knowingly entered (2) an agreement (3) that imposed an unreasonable restraint of trade (4) in interstate or foreign commerce. See Dickson v. Microsoft Corp., 309 F.3d 193, 202 (4th Cir. 2002); W.F. Brinkley, 783 F.2d at 1162. 4 The indictment here alleged that the agreement was an “unreasonable restraint” because it fell within the class of agreements that are per se unreasonable under the Sherman Act. See United States v. Portsmouth Paving Corp., 694 F.2d 312, 317 (4th Cir. 1982).

4

Section 1 of the Sherman Act declares “[e]very contract, combination . . . , or conspiracy, in restraint of trade or commerce . . . to be illegal.” 15 U.S.C. § 1. While the plain language might broadly cover any agreement to restrain trade, the Supreme Court has repeatedly instructed that “Congress intended to outlaw only unreasonable restraints.” Texaco Inc. v. Dagher, 547 U.S. 1, 5 (2006) (quoting State Oil Co. v. Khan, 522 U.S. 3, 10 (1997)).

10

As an initial matter, the district court properly treated Brewbaker’s “Motion to Apply the Rule of Reason” as a motion to dismiss the indictment for failure to state an offense constituting a per se Sherman Act violation. J.A. 968. The indictment only alleged that the restraint was a per se violation. To permit conviction under the rule of reason— i.e., another method of establishing § 1’s “unreasonable” element—would constructively amend the indictment. Cf. Stirone v. United States, 361 U.S. 212, 217 (1960). A constructive amendment occurs whenever “the government or the court broadens the possible bases for conviction beyond those included in the indictment” by, for example, asserting a specific legal theory in an indictment but then relying on a different theory at trial. United States v. Ellis, 121 F.3d 908, 923 (4th Cir. 1997). And that’s impermissible. Stirone, 361 U.S. at 217; Russell v. United States, 369 U.S. 749, 770 (1962).

So now we must consider whether the district court should have dismissed the indictment for failure to state a per se offense. We think so.

1. The factual allegations in the indictment did not state a per se violation of the Sherman Act.

Contrary to the district court’s conclusion, the indictment did not allege a per se violation of the Sherman Act. That’s because (1) it alleged a price-fixing restraint with both horizontal and vertical aspects and (2) caselaw and economic analysis shows that category of restraint may have procompetitive effects.

11

a. Per se rule versus the rule of reason Despite its broad language, § 1 of the Sherman Act only prohibits unreasonable restraints of trade. Leegin, 551 U.S. at 885. There are two dominant ways to determine whether a restraint is unreasonable: the rule of reason and the per se rule. 5 The rule of reason is the default. Leegin, 551 U.S. at 885. It requires that “the factfinder weigh[] all of the circumstances of a case in deciding whether a restrictive practice should be prohibited as imposing an unreasonable restraint on competition.” Cont’l T.V., Inc. v. GTE Sylvania Inc., 433 U.S. 36, 49 (1977). This involves inquiries into the specific business and market, along with the restraint’s history, nature, and effect, to identify the specific restraint’s actual competitive impact. Leegin, 551 U.S. at 885–86.

Certain “categories of restraints,” however, have been held per se unreasonable.

Leegin, 551 U.S. at 886. In other words, the nature of the restraint makes it “necessarily illegal” without inquiry into—or evidence about—the particular restraint’s anticompetitive impact. Id. But this blanket illegality isn’t applied loosely; it is “confined to restraints . . . ‘that would always or almost always tend to restrict competition and decrease output.’” Id. (quoting Bus. Electrs. Corp. v. Sharp Electrs. Corp., 485 U.S. 717, 723 (1988)). So the per se rule can be applied to a category of restraints only after economic evidence shows the restraint has “manifestly anticompetitive effects” and “lack[s] . . . any redeeming

5

Note that we say these two rules are the dominant rules, not the only rules. See Cal. Dental Ass’n v. F.T.C., 526 U.S. 756, 779–80 (1999). In civil cases, the district court often must decide which mode of analysis will be used at trial. But when a district court is asked whether a criminal indictment properly states a per se offense, it must decide only whether the per se rule applies to the allegations.

12

virtue.” Id. at 886–87 (quotations and citations omitted). Put differently, the per se rule only applies automatically to restraints that already have been held to be devoid of procompetitive effects; it cannot be extended to new categories of restraints except through economic analysis that shows the new type of restraint is always anticompetitive. Id. at 887 (noting that the per se rule applies “only after courts have considerable experience with the type of restraint at issue” and have “confidence it would be invalidated in all or almost all instances under the rule of reason” (citations omitted)).

Whether the per se rule applies didn’t always turn on the restraint’s economic effects. Previously, the per se rule was extended to new categories of restraints largely because they resembled other per se restraints. For example, some vertical restraints were declared per se unreasonable largely because similar horizontal restraints were already subject to the per se rule. See, e.g., United States v. Arnold, Schwinn & Co., 388 U.S. 365, 379 (1967), overruled by GTE Sylvania, 433 U.S. 36; Albrecht v. Herald Co., 390 U.S. 145, 152–54 (1968), overruled by State Oil v. Khan, 522 U.S. 3 (1997). Economic evidence that the vertical restraint increased competition, or that the restraint “may have different consequences” in different contexts, was ignored. See Albrecht, 390 U.S. at 152–53; Schwinn, 388 U.S. at 384 (Stewart, J. concurring).

But the Supreme Court has since cautioned courts against over-analogizing in the antitrust context, recognizing that the classes of restraints subject to per se condemnation should be narrowly construed. Leegin, 551 U.S. at 888. Rather than look only at the label attached to the restraint, such as “price fixing” or “market allocation,” courts must consider the restraint in context—including how the parties are related—before applying the per se

13

rule. Id. And when a case involves a category of restraint not yet classified under either the rule of reason or the per se rule, “departure from the rule-of-reason standard must be based upon demonstrable economic effect rather than . . . upon formalistic line drawing.” GTE Sylvania, 433 U.S. at 58.

In determining whether a category of restraint’s “demonstrable economic effect”

warrants per se treatment, the Court has repeatedly told us that antitrust’s North Star is the restraint’s impact on interbrand competition. See, e.g., id. at 52 n.19. Rather than intrabrand competition (i.e., competition among the retailers who sell one manufacturer’s product), interbrand competition (i.e., competition between manufacturers of similar products) “is the primary concern of antitrust law.” Id. And the Court now routinely accepts that restraints boosting interbrand competition at the expense of intrabrand competition do not warrant per se treatment.

For this reason, the Court in GTE Sylvania overruled Arnold, Schwinn & Co., which had held that vertical territorial restraints were per se unreasonable. Reviewing the economics of such restraints, the Court determined that the vertical restraints at issue there were not devoid of “any redeeming virtue,” even though they reduced intrabrand competition, because they increased interbrand competition via distributional efficiencies. Id. at 54–58. Similarly, in Khan and Leegin, the Court overruled cases holding that vertical price-fixing restraints warrant the per se rule. A central reason for both reversals was that

14

economics showed that vertical restraints can promote interbrand competition. Khan, 522 U.S. at 15–20; Leegin, 551 U.S. at 889–92. 6 The takeaway is that, before the per se rule applies outside of its predetermined, narrow confines, economic evidence must make certain that a restraint has solely anticompetitive effects. And after GTE Sylvania, Khan, and Leegin rejected applying the per se rule to vertical restraints, the per se rule’s predetermined, narrow confines only extend to certain categories of horizontal restraints, including agreements between competitors to fix prices or divide markets. Leegin, 551 U.S. at 886.

This has led courts to determine whether the per se rule applies by first asking whether the alleged restraint is horizontal or vertical. The district court here did just that. And that is the right threshold question—in a broad sense. For if the restraint is horizonal,

6

To illustrate why vertical restraints may increase interbrand competition, pretend you are a shoe store selling Nikes. If you know that every other store in town is selling Nikes, you may conclude that you’re better off relying on those retailers’ Nike marketing efforts rather than spending your own resources. In other words, you decide to free-ride. But if you decide to free-ride, it’s likely the other retailers do, too. That means less Nikes are sold, and there’s less competition between Nike and other brands like Adidas. However, let’s say that Nike places a territorial restriction, limiting the sale of Nikes to a single retailer in a given city. If you’re that retailer, you no longer can free-ride on others. You want to get the Nikes out the door, and so you spend your own resources on marketing and value-adding services to entice customers. Thus you sell more Nikes. And that increases Nike’s interbrand competition with Adidas. The same logic applies when it comes to prices; if Nike sets a minimum resale price on its shoes, its retailers cannot undercut each other and cause a race-to-the-bottom that sinks Nike’s profits. Instead, they are incentivized to draw people into their stores with marketing and services that lead to increased sales at increased prices.

15

then the per se rule will generally apply. 7 And if the restraint is vertical, then the rule of reason will apply.

The Supreme Court has explained that whether a restraint is horizontal or vertical depends on the relationship between the parties to the agreement that imposes the restraint. Horizontal restraints are “restraints ‘imposed by agreement between competitors.’” Am. Express, 138 S. Ct. at 2283 (quoting Bus. Electrs., 485 U.S. at 730). And vertical restraints are “restraints ‘imposed by agreement between firms at different levels of distribution.’” Id. (quoting Bus. Electrs., 485 U.S. at 730). 8 So we ask here if the agreement between Contech and Pomona was made by competitors or by firms at different levels of distribution. See Bus. Electrs., 485 U.S. at 730 n.4 (“[A] restraint is horizontal not because it has horizontal effects, but because it is the product of a horizontal agreement.”).

b. The indictment alleged a hybrid restraint that hasn’t been held to be per se unlawful.

Going off the Supreme Court’s definitions, the indictment here alleged a restraint that was both horizontal and vertical. On the one hand, it alleged Pomona and Contech both submitted bids for NCDOT aluminum projects. That would make them competitors

7

Some horizontal price-fixing restraints have been held to require the rule-of-reason analysis. See, e.g., Nat’l Collegiate Athletic Ass’n v. Bd. of Regents, 468 U.S. 85, 104 (1984); Broad. Music, Inc. v. Columbia Broad. Sys., Inc., 441 U.S. 1 (1979). But these are exceptions to the rule that purely horizontal price-fixing restraints are per se unreasonable. See Leegin, 551 U.S. at 886.

8

In an earlier decision, we suggested that the horizontal-vertical distinction turned not on the relationship but on the “purpose” of the agreement. Donald B. Rice Tire Co. v. Michelin Tire Corp., 638 F.2d 15, 16–17 (4th Cir. 1981). That dicta cannot survive the Supreme Court’s later pronouncements.

16

within the aluminum-project market, suggesting their agreement was a horizontal restraint. On the other hand, the indictment alleged that Pomona “served as a dealer” for Contech, with Contech supplying Pomona aluminum that Pomona then used to compete against “others” in NCDOT aluminum-structure projects. J.A. 46. And the indictment alleged that the benefit Contech got from the agreement was that “it supplied aluminum pieces to [Pomona] for use in” the projects Pomona won. J.A. 56. That means Contech was a manufacturer and Pomona its dealer, placing them at different levels of distribution, and indicating that their agreement was vertical. In short, the restraint alleged in the indictment doesn’t fit neatly into either the horizontal or vertical definition—it fits into both.

But does the per se rule apply to such a hybrid restraint? The Supreme Court has not yet told us. 9 Still, we are not without guidance. We must begin with a “presumption in favor of a rule-of-reason standard.” Bus. Electrs., 485 U.S. at 730. And we know that “problems in differentiating vertical restrictions from horizontal restrictions” do not alone “justify a per se rule.” GTE Sylvania, 433 U.S. at 58 n.28. Displacing the presumptive rule-of-reason analysis is possible only when demonstrable economic evidence shows that

9

The only restraints that the Supreme Court has held to be per se unreasonable are purely horizontal, or, in other words, are agreements between entities who are only related as competitors. See, e.g., Arizona v. Maricopa Med. Soc’y, 457 U.S. 332 (1982) (holding that price fixing between medical organizations is per se unreasonable); Catalano, Inc. v. Target Sales, Inc., 446 U.S. 364 (1980) (same for beer wholesalers); United States v. Topco Assocs., 405 U.S. 596, 608–10 (1972) (same for supermarket chains); United States v. Socony-Vacuum Oil Co., 310 U.S. 150, 166 (1940) (same for oil companies); Palmer v. BRG of Ga., Inc., 498 U.S. 46 (1990) (same for market division between bar-review companies). These include restraints between competitors and nominally vertically related entities that are, in reality, instrumentalities the competitors use to facilitate the restraint among them. See Topco, 405 U.S. at 602–05, 608–09; United States v. Sealy, 388 U.S. 350, 352–56 (1967).

17

the type of restraint at hand “always or almost always” has “manifestly anticompetitive effects” and “lack[s] . . . any redeeming virtue.” Leegin, 551 U.S. at 886–87.

Before turning to these economic effects, however, we address the government’s three arguments for why we should apply the per se rule without considering economics.

The government first argues that the restraint alleged in the indictment isn’t a hybrid restraint at all. Rather, it contends that the restraint is simply a horizontal restraint. That’s because, from the government’s view, whether a restraint is horizontal or vertical depends not on the relationship of the parties to the agreement en total; it only depends on which part of their relationship is restrained by the agreement. And the government asserts the restraint alleged in the indictment only limited how Contech and Pomona could act when bidding on aluminum-structure projects—i.e., in their relationship as competitors. It did not, in the government’s perspective, limit how they could act when Contech sold its aluminum pieces to Pomona. So the government urges us to ignore the vertical aspect of Contech and Pomona’s relationship and see the restraint as straightforward horizontal bid rigging between Contech-as-bidder and Pomona-as-bidder.

We decline to do so, however, because we cannot disregard the parties’ broader relationships when classifying a restraint. As explained above, when determining whether a restraint is horizontal or vertical, we are instructed to look at the relationship of the parties, not just the nature of the limitation imposed. See Bus. Electrs., 485 U.S. at 730 & n.4. This is because agreements that otherwise look identical in form produce different economic effects based on how the parties relate to one another. Cf. Leegin, 551 U.S. at 888 (warning against analogizing between similar restraints in different contexts). A price-

18

fixing agreement between two competing parties involves different dynamics and produces different effects on competition than one between parties who simultaneously compete and collaborate. We cannot simply ignore this relational difference because of the specific form that the agreement itself takes.

Moreover, the government’s approach would force us to engage in arbitrary and likely impossible line-drawing. We do not normally artificially split a business entity into pieces in order to conclude that only one part of the entity—for example, the part that acted as the other party’s competitor—was the actual “party” to the agreement. The Sherman Act doesn’t ignore reality; it treats the entire business entity as the single party it is. See Copperweld Corp. v. Indep. Tube Corp., 467 U.S. 752, 771–74 (1984); Texaco, 547 U.S. at 5–6. But this artificial division is what the government would have us do here. Contech is acting as both manufacturer and co-distributor in its arrangement with Pomona. The government would ask us to parse the form of agreement to see which part of Contech is affected. 10 Antitrust law does not turn on such artificial mental gymnastics.

10

For instance, imagine if Contech required Pomona to bid its aluminum at a certain price, and then Pomona required Contech to bid its services at a certain price. Or imagine that Contech required Pomona to bid at a certain price, and then set its own bid price to be slightly higher. Would we have a horizontal restraint between competitors to fix the aluminum-structure bid? Or would we have vertical, minimum-price restraint? The Supreme Court tells us that the agreements like the former are per se illegal, while those like the latter must be assessed under the rule of reason. Yet the government offers no principled way to distinguish between the two. This only goes to show that focusing on the form of the restraint, instead of the relationship between the parties, is a fool’s errand. See Mark A. Lemley & Christopher R. Lesli, Categorical Analysis in Antitrust Jurisprudence, 93 Iowa L. Rev. 1208, 1238–40 (2008) (noting that classifying a restraint between parties related both vertically and horizontally “as either horizontal or vertical” by looking to the restraint’s purpose, effect, or source is a “laborious process” that requires “significant resources” and is “exactly backwards”).

19

The government next attempts to classify the restraint as per se unlawful by relying on cases in which the per se rule has been applied to conspiracies involving competing companies and their vertical supplier or distributor. See, e.g., Klor’s Inc. v. Broadway- Hale Stores, Inc., 359 U.S. 207 (1959); Kotteakos v. United States, 328 U.S. 750, 755 (1946); United States v. Apple, Inc., 791 F.3d 290, 322–25 (2d Cir. 2015). In these so- called hub-spoke-and-rim conspiracies, competitors agree to restrain trade and are encouraged to do so by a shared vertical entity. But—crucially—the vertical entity is not a party to the competitors’ agreement; it is merely an encourager of it, for example, through separate vertical agreements. In other words, the parties to the competitors’ agreement are related only horizontally. The vertical entity’s relationship to the competitors is separate from the agreement. It is to this purely horizontal restraint that the per se rule applies. See Apple, 791 F.3d at 323 (noting that “the relevant ‘agreement in restraint of trade’” determined to be per se unlawful was “not Apple’s vertical Contracts with the Publisher Defendants” but “the horizontal agreement that Apple organized among the Publisher Defendants” (emphasis added)); Phillip E. Areeda & Herbert Hovenkamp, Antitrust Law ¶ 1402c (4th ed. 2013) (stating that hub-and-spoke conspiracies have only been assessed as per se Sherman Act violations where there is a “traditional horizontal conspiracy” with a “vertically related facilitator”). The separate vertical agreements between the vertical entity and each competitor, however, “if challenged, [would] have to be evaluated under the rule of reason.” Apple, 791 F.3d at 323.

So the restraints in hub-spoke-and-rim conspiracies to which the per se rule applies are purely horizontal restraints that a vertical entity encouraged. The restraint alleged in

20

this indictment is of a different kind. It alleges a single agreement between two parties related both vertically and horizontally. The government’s attempted analogy thus fails.

Lastly, the government argues that the indictment alleged a per se unlawful restraint because it stated that Brewbaker and Contech “rigged bids” and we have held that bid rigging is per se unlawful. See Portsmouth Paving, 694 F.2d at 325 & n.18. But, in so holding, we defined per se unlawful bid rigging as an “agreement between competitors.” Id. That is precisely how the Supreme Court defines a horizontal restraint. Am. Express, 138 S. Ct. at 2283. So, for the same reasons the indictment doesn’t simply allege a horizontal restraint, it doesn’t allege what we have held to be per se unlawful bid rigging. This is reinforced because the restraint in Portsmouth Paving was between parties with a purely horizontal relationship. The parties to the agreement were all paving companies that rigged bids for certain paving projects. See 694 F.2d at 315–16. None had a vertical relationship with another party. See id.

More pointedly, Portsmouth Paving predates Leegin. That is, it was decided when both horizontal and vertical price fixing were per se unlawful. Thus, even if it could be read to prohibit price fixing beyond that between purely horizontal parties, Leegin mandates it be assessed anew.

In sum, the Supreme Court has instructed that vertical price restraints are subject to the rule of reason and that horizontal price restraints are per se illegal. To determine which applies, we must look to the relationship of the parties to the agreement. Doing so here shows that Contech and Pomona had a hybrid relationship with both vertical and horizontal components. And the Supreme Court has not told us how to analyze an agreement between

21

two parties with that type of relationship. 11 But it has instructed that, before we apply the per se rule to a new category of restraint, we must apply a presumption in favor of the rule of reason that may be overcome only with demonstrable economic effect.

c. Dr. Elzinga’s affidavit addressing the economic effect of this category of restraint should be considered.

Before addressing the demonstrable economic effect of the category of restraint the indictment alleged (i.e., a hybrid price-fixing agreement), we pause to ask whether the district court erred by categorically refusing to consider Dr. Elzinga’s affidavit in addressing Brewbaker’s motion to dismiss the Sherman Act count.

The district court was right so far as it held that it was prohibited from considering any extrinsic factual evidence—including any portions of Dr. Elzinga’s affidavit that outlined Contech and Brewbaker’s version of the facts. See United States v. Engle, 676

11

We note that United States v. McKesson & Robbins, Inc., 351 U.S. 305 (1956), does not respond to our question. There, the Supreme Court interpreted an exception to the Miller-Tydings Act. Id. at 311. That Act exempted certain retail-price-maintenance contracts from the per se rule—because, at that time, such manufacturer-mandated price restraints would have been per se unreasonable. See id. The exception that the Court interpreted stated that the Act “shall not make lawful any contract or agreement, providing for the establishment or maintenance of minimum resale prices . . . between persons, firms, or corporations in competition with each other.” Id. According to the Court, the last five words of the exception meant a restraint between a manufacturer who acted as a wholesaler and its other wholesalers still faced the per se rule, as the agreeing parties were, in part, competitors. Id. at 312–15. But this, at most, shows that the Court read the Miller-Tydings Act to exempt only purely vertical restraints from the per se rule at a time when all price- fixing restraints were per se illegal. Thus, per the Act, any price-fixing restraints between parties with any other type of relationship remained per se unreasonable. However, the Miller-Tydings Act was repealed in 1976, and post-Leegin, vertical restraints are subject to the rule of reason without a statutory exemption. So McKesson doesn’t illuminate whether hybrid restraints are subject to the per se rule or the rule of reason. See also Areeda & Hovenkamp, supra, at ¶ 1605a (noting that McKesson “has little bearing” on the problem of deciding whether restraints are subject to the per se rule).

22

F.3d 405, 415 (4th Cir. 2012). A district court is limited to considering the factual allegations in the indictment and must accept them as true in ruling on a motion to dismiss. Id. That is because a district court “lack[s] authority to review the sufficiency of the evidence supporting the indictment.” United States v. Wills, 346 F.3d 476, 488 (4th Cir. 2003); cf. Kaley v. United States, 571 U.S. 320, 333 (2014). Just as we do on appeal, the district court had to accept the facts as found by the grand jury in the indictment.

But the district court, and not the grand jury, must decide questions of law. Cf.

Papasan v. Allain, 478 U.S. 265, 286 (1986) (explaining that, on a motion to dismiss, courts “are not bound to accept as true a legal conclusion couched as a factual allegation”). So, while an indictment may not be dismissed simply because the district court doesn’t think the government can prove what it has alleged, an indictment may be dismissed if the district court concludes that the allegations in the indictment—even if proven—would not satisfy the elements of the charged offense. Engle, 676 F.3d at 415 (recognizing that a “district court may dismiss an indictment under Rule 12 where there is an infirmity of law in the prosecution” (cleaned up)). Here, we must determine whether the indictment’s factual allegations, if true, stated the charged per se violation of the Sherman Act. And that is a legal question—both when asking whether the alleged agreement falls in a category of restraint that has already been held to be per se unlawful, and when asking whether the per se rule should be extended to a new category of restraint in which the alleged agreement falls. See Areeda & Hovenkamp, supra, at ¶ 1909b.

When making that second legal determination—whether the per se rule should be applied to a new category of restraint—the Supreme Court instructs that courts must

23

consult economic evidence to determine whether the category of restraint has plausible procompetitive effects. See Leegin, 551 U.S. at 885–87. 12 And to properly evaluate economic effects, courts may look to economic analysis by economists. We, of course, may look at caselaw or academic literature when making any other legal determination. So too may caselaw and academic literature be consulted when making the legal determination of whether the per se rule should be expanded to a new category of restraint. And an economist’s analysis of the competitive effects of the category of restraint is equally relevant, given how the Supreme Court has defined the question. See Leegin, 551 U.S. at 896–98. So the district court should not have categorically excluded Dr. Elzinga’s academic analysis drawing on supporting literature about the competitive effects of the category of restraint alleged in the indictment.

d. Economic evidence shows the category of restraint alleged in the indictment wouldn’t have solely anticompetitive effects.

So we turn to asking whether demonstrable economic evidence showed that the category of restraint alleged in the indictment “always or almost always” has “manifestly anticompetitive effects” and “lack[s] . . . any redeeming virtue.” Id. at 886–87 (quotations

12

To be clear, a court may not consider economic evidence when ruling on a motion to dismiss an indictment if the restraint alleged in the indictment falls into a category that has already been held to be per se illegal, such as purely horizontal price-fixing restraints. See United States v. Aiyer, 33 F.4th 97, 118–19 (2d Cir. 2022). Based on binding legal precedent, that restraint would be properly charged as a per se criminal violation without more. So consideration of that particular restraint’s economic effect or reasonableness would be prohibited. Id. Here, however, the indictment alleges a category of restraint that hasn’t been held to be per se illegal—hybrid price-fixing restraints. And the answer to the question that Leegin thus requires us to ask—whether to extend the per se rule to this new category of restraint—depends on the competitive effects of that category of restraint.

24

and citations omitted). To the contrary, economic analysis—including that provided by Dr. Elzinga, academic literature, and Supreme Court opinions—shows that this type of restraint has possible procompetitive effects. 13 The vertical-horizontal setup alleged in the indictment is known in antitrust law and economics as a “dual distribution” arrangement: Contech was both supplying Pomona with aluminum and competing against Pomona for aluminum projects. See Areeda & Hovenkamp, supra, at ¶ 1600c2 (“[A] manufacturer practices ‘dual distribution’ by selling its product directly to some consumers in competition with the independent dealers handling its product.”).

Although couched in economic terminology, dual distribution is something we’re all familiar with. To illustrate, let’s say you want some new Nikes. There’s more than one way you could buy them. You can order them online at Nike.com. Or you can drive to Foot Locker and buy them. If you go with the first option, you are buying directly from the manufacturer. If you go with the second, you are buying from the manufacturer’s dealer. So Nike is both supplying Foot Locker with shoes to sell (a vertical relationship)

13

We need not, and do not, decide whether the particular restraint here was actually procompetitive. That determination could be made only after applying the rule of reason to this particular restraint. See Robert Zwirb, Dual Distribution and Antitrust Law, 21 Loy. L.A. L. Rev. 1273, 1342 n.219 (1988) (“The first step is concerned mainly with an arrangement’s ‘potential,’ while the second step (the rule of reason balancing inquiry) is concerned with resolving whether the potential is ‘actual.’ The analysis in the first step, therefore, is not and need not be as comprehensive as that required in the second step of the rule of reason inquiry.”). And the indictment did not allege an antitrust offense based on the rule of reason.

25

and is competing with Foot Locker when selling the shoes directly to consumers (a horizontal relationship).

The indictment here alleged nothing different. It stated that Contech was both supplying Pomona with aluminum to sell to NCDOT and competing with Pomona to sell that aluminum to NCDOT. Despite this straightforward analogy, the district court concluded (and the government argues on appeal) that the indictment doesn’t allege a Nike- Foot-Locker-type relationship because the bids Contech and Pomona were competing for were not for the exact product Contech supplied to Pomona—aluminum pipe—but for completed aluminum-pipe projects. That may be true. Yet, as an economic matter, it is beside the point. Part of the reason for dual-distribution arrangements is that a retailer can offer services alongside the manufacturer’s products that the manufacturer cannot (or doesn’t want to) offer itself. With Nike and Foot Locker, Foot Locker offers the added service of in-person customer assistance. And you need only think of the car you’d need to drive to Foot Locker to have another example: You could buy a tire online directly from a manufacturer, but unless you’re particularly handy, you are more likely to buy it from the auto shop that installs it for you.

Of course, the alleged Sherman Act violation here is not the mere fact that Contech and Pomona had a dual-distribution arrangement. It’s that they imposed a price-fixing restraint within it. The government doesn’t dispute that, post-Leegin, if a manufacturer like Contech wasn’t also selling directly to a customer (here, NCDOT), any price restraint it imposed on its distributors would be adjudged under the rule of reason. The inquiry is thus whether the fact that a manufacturer is also selling directly to consumers eliminates

26

the potential interbrand procompetitive effects that supported the Supreme Court’s holding in Leegin.

It does not. Start with the general economic reason for dual distribution: It increases distributive efficiency. See J.A. 110. More sellers means it’s easier to find a product. If it is easier to find a product, then it is easier to buy. And if it’s easier to buy, then, presumably, sales will increase. See J.A 117; Gregory T. Gundlack & Alex G. Loff, Dual Distribution Restraints: Insights from Business Research and Practice, 58 Antitrust Bull. 69, 79 (2013). Further, as mentioned, retailers can provide services the manufacturer cannot or will not, further increasing consumer reach. J.A. 114–15; see Leegin, 551 U.S. at 891. Plus, if distributors fail to make their sales (or, as relevant here, fail to place a bid), the manufacturer’s sales serve as a stopgap to ensure the manufacturer still makes money. J.A. 118, 138. This can be good for intrabrand competition, as more outlets are selling the same good. And it can be good for interbrand competition, because the greater reach of a certain brand means greater competition between that brand and other, competing brands.

But a manufacturer selling alongside a distributor may cause issues that undermine the economic efficiencies of the vertical relationship between them, harming manufacturers and interbrand competition alike. Economists call these “channel conflicts.” J.A. 119; Andy A. Tsay & Narendra Agrawal, Channel Conflict and Coordination in the E- Commerce Age, 13 Prod. & Op. Mgmt. Soc. 93 (2004). For example, if a manufacturer cuts its own prices, the independent distributor may lose the incentive to provide valuable additional services or to market—and thus sell—the product itself. J.A. 118; Malcolm B. Coate & Mark R. Fratrik, Dual Distribution as a Vertical Control Device 14 (Fed. Trade

27

Comm’n, Working Paper No. 143, 1986). More than that, a distributor may become so upset with the manufacturer for undercutting it that it decides to stop distributing the manufacturer’s product completely. See J.A. 115–16 (“To undercut one’s distributor . . . would be the business equivalent of shooting oneself in the foot.”). And this would be especially detrimental in a market where the number of potential distributors is limited. Coate & Fratrick, supra, at 15. In both scenarios, consumers and competition lose out. When fewer distributors sell one manufacturer’s goods, other manufacturers’ goods face less interbrand competition. J.A. 119; Tsay & Agrawal, supra, at 94 (“Elimination of intermediaries may cause an erosion of profits, market share, or both.”).

So manufacturers have to find ways to mitigate these conflicts. One way is by ensuring their direct-sale prices are equal to or higher than their distributors’ prices by fixing the distributors’ resale prices or the manufacturer’s own. J.A. 119; see Reuben Arnold, Neill Norman & Daniel Schmierer, Resale Price Maintenance and Dual Distribution, Distrib. and Franchising Comm.: ABA Section of Antitrust L. 12 (2016). As stated, outside of a dual-distributor setup, this type of vertical price fixing would not be subject to the per se rule after Leegin. Yet the same potential boons to interbrand competition don’t disappear just because a manufacturer also acts as a distributor. J.A. 119–20; cf. Leegin, 551 U.S. at 890–91. The price restraints still incentivize distributors to continue to vigorously sell the manufacturer’s product and to offer additional services, therefore increasing interbrand competition. Arnold, Norman & Schmierer, supra, at 12 (explaining that a dual-distribution manufacturer that sets its direct sale price equal to its distributors “may strengthen the competitiveness of [its] brand and thereby enhance inter-

28

brand competition”). In fact, on remand, the Fifth Circuit recognized that the restraint in Leegin was a dual-distribution restraint but noted that the manufacturer’s position in the retail market made it “no different from a manufacturer that does not have retail stores.” See PSKS, Inc. v. Leegin Creative Leather Prods., Inc., 615 F.3d 412, 421 (5th Cir. 2010). So the per se rule was inapplicable. Id.

The same logic applies to the restraint alleged in this indictment. The alleged bid rigging (a type of price fixing) could allow Contech to maintain its relationship with Pomona by making sure it never undercut, and thus upset, its distributor. J.A. 115. So— just like in GTE Sylvania, Leegin, and Khan—while the bid rigging had the effect of eliminating intrabrand competition between Contech and Pomona, it also could benefit interbrand competition. By increasing Pomona’s sales of Contech’s aluminum, the restraint could lead to greater competition between Contech and other aluminum manufacturers. J.A. 115–16.

The potential interbrand procompetitive effects show that the category of restraint alleged in the indictment would not invariably lead to anticompetitive effects. Yes, it may lead to some. See Leegin, 551 U.S at 892–94. But it is exactly that economic uncertainty that shows the indictment did not allege a per se violation. For we cannot “predict with confidence that” the dual-distribution bid rigging alleged in the indictment “would be

29

invalidated in all or almost all instances under the rule of reason.” Id. at 886–87; see also id. at 894. 14 When the government indicted Brewbaker, it decided to include detailed factual allegations. It wasn’t required to. See United States v. Quinn, 359 F.3d 666, 673 (4th Cir. 2004). But the government did. It alleged that that Pomona wasn’t only Contech’s co- bidder on NCDOT projects; it was also its distributor, and the restraint between them benefited Contech because of the vertical nature of its relationship with Pomona. Supplied with these allegations, the district court had the responsibility to ensure that the indictment stated a per se violation. Yet the indictment alleged neither a restraint previously held subject to the per se rule nor one that economics showed would invariably lead to anticompetitive effects. So Count One of the indictment should have been dismissed for failing to state an offense, and we reverse Brewbaker’s Sherman Act conviction.

14

While sometimes applying different analyses, this Court and nearly all other lower courts have adjudged hybrid restraints with vertical and horizontal aspects under the rule of reason. See, e.g., Donald B. Rice, 638 F.2d at 16; Hampton Audio Electrs., Inc. v. Contel Cellular, Inc., 966 F.2d 1442 (4th Cir. 1992) (unpublished); Electr. Com. Corp. v. Toshiba Am. Consumer Prods., Inc., 129 F.3d 240, 243–44 (2d Cir. 1997); Krehl v. Baskin-Robbins Ice Cream Co., 664 F.2d 1348, 1357 (9th Cir. 1982); Dimidowich v. Bell & Howell, 803 F.2d 1473, 1481 (9th Cir. 1986); AT&T Corp. v. JMC Telecom, LLC, 470 F.3d 525, 531 (3d Cir. 2006); Abadir & Co. v. First Miss. Corp., 651 F.2d 422, 428 (5th Cir. 1981); Davis- Wakins Co. v. Serv. Merch., 686 F.2d 1190, 1197–1202 (6th Cir. 1982); see also Areeda & Hovenkamp, supra, ¶ 1600c2 (noting that dual-distribution “restraints are generally tested by the rules governing ordinary vertical restraints”).

30

B. Brewbaker’s fraud convictions stand.

Brewbaker makes two arguments for why reversing his Sherman Act conviction also requires reversing his wire- and mail-fraud convictions. We disagree with both arguments and affirm the fraud convictions.

Brewbaker was convicted of conspiracy to commit mail and wire fraud, as well as four counts of mail and wire fraud relating to specific misleading submissions. As the jury was instructed here, mail fraud (18 U.S.C. § 1341) and wire fraud (18 U.S.C. § 1343) require the defendant to have (1) knowingly devised or participated in a scheme or artifice to obtain money or property by means of false or fraudulent pretenses, representations, or promises that were material; (2) acted with the intent to defraud; and (3) used the mails or wire communication in furtherance of the scheme. J.A. 1674, 1681.

Brewbaker does not assert that there was insufficient evidence for the jury to convict him of each fraud count. Rather, he argues that the jury instructions on the Sherman Act count “infected” the jury’s consideration of the fraud counts. Appellant’s Br. at 63.

The trouble with Brewbaker’s argument is that we operate under the “crucial assumption that jurors carefully follow instructions.” United States v. Rafiekian, 991 F.3d 529, 550 (4th Cir. 2021); see also Francis v. Franklin, 471 U.S. 307, 324 n.9 (1985) (“The Court presumes that jurors, conscious of the gravity of their task, attend closely the particular language of the trial court’s instructions . . . and follow the instructions given them.”). And here, the fraud instructions did not incorporate or reference the Sherman Act instructions. Nor did the fraud counts depend on finding Brewbaker guilty under the Sherman Act. Plus the court specifically instructed the jury that they “must consider each

31

count separately” and that guilt on one count “shouldn’t control your verdict as to the other counts.” J.A. 2588.

We see nothing that sufficiently undercuts our assumption that the jury followed these instructions. As the indictment alleged, the fraud counts turned on the false certification that the bids were “submitted competitively and without collusion.” J.A. 53, 55. That certification was materially false, the government argued, because Brewbaker and Contech colluded with Pomona to obtain their total bid price and submit a non- competitive, intentionally higher bid. See J.A. 55–56. The falsity of the certifications thus turned on whether Brewbaker submitted competitive and non-collusive bids—not on whether doing so was a per se Sherman Act violation. Brewbaker doesn’t contest that, at trial, the government proved he obtained Pomona’s bid prices and used them to submit Contech’s higher bids. See, e.g., J.A. 1834–35, 2320. So the jury had good reason— independent of any Sherman Act instruction or violation—to believe the certifications were materially false. As a matter of common parlance, it’d be hard to say a bid was submitted “competitively” when Contech’s bid was intentionally higher, or “without collusion” when it was previously agreed-upon. Therefore, we refuse to find that the jury disregarded the court’s instructions to consider these charges separately.

Making one last-ditch effort, Brewbaker points to the jury’s request during deliberation for an explanation of “collusion” regarding the NCDOT certification. J.A. 2641. With Brewbaker’s assent, the district court told the jury that collusion was mentioned with regard to the nature of the crime charged in Count 2 (mail- and wire-fraud conspiracy). It then explained: “There isn’t a legally defined explanation of collusion . . . .

32

I remind you to consider all the facts and circumstances in evidence in reaching your understanding of the crime charged, and consider all of the Court’s instructions as a whole.” J.A. 2645 (emphasis added). According to Brewbaker, the reference to the instructions as a whole directed the jury to consider the Sherman Act instructions and conviction. 15 Yet, as we have explained, the jury was instructed to consider each count separately. In the face of our assumption that juries follow instructions, we will not presume that the jury understood “consider all of the Court’s instructions as a whole” to mean “abandon the Court’s instruction to consider the counts separately.”

We can only overcome the presumption that a jury follows instructions in “extraordinary situations.” Francis, 471 U.S. at 324 n.9. This is no such situation. See Bruton v. United States, 391 U.S. 123, 136–37 (1968) (jury instruction to disregard co- defendant’s confession that inculpated the defendant as hearsay was insufficient); Jackson v. Denno, 378 U.S. 368, 377–78 (1964) (improper to have jury decide, simultaneously with the defendant’s guilt, whether defendant’s confession was voluntary); United States v. Lindberg, 39 F.4th 151, 164–65 (4th Cir. 2022) (erroneous jury instruction on one count was repeated during the instructions on another count, and therefore “infected” the latter). So we hold fast to our trust in the jury and conclude that the Sherman Act instructions did not bear on Brewbaker’s mail- and wire-fraud convictions.

15

Notably the Sherman Act instructions directed that “the exchange of information about bid prices is not, by itself, illegal. The fact that defendant and alleged co-conspirators exchanged such information does not establish an agreement to rig bids. There may be other legitimate reasons that would lead competitors to exchange information about bid prices.” J.A. 1658.

33

* * *

Whether an indictment states an offense “is a question of law, to be decided by the court, not the prosecutor.” United States v. Cruikshank, 92 U.S. 542, 559 (1875). This indictment did not state a per se antitrust violation under the Sherman Act. So that count should have been dismissed. But the fraud convictions stand, and we remand to the district court for resentencing on those counts alone. The district court’s judgment is

REVERSED IN PART,

AFFIRMED IN PART,

AND REMANDED.