Concurring opinion filed by Circuit Judge MILLETT.
Dissenting opinion filed by Circuit Judge KAVANAUGH.
ROGERS, Circuit Judge:This expedited appeal arises from the government’s successful challenge to “the largest proposed merger in the history of the health insurance industry, between two of the four national carriers,” Anthem, Inc. and Cigna Corporation. Appellees Br. 1. In July 2015, Anthem, which is licensed to operate under the Blue Cross Blue Shield brand in fourteen states, reached an agreement to merge with Cigna, with which Anthem competes largely in those fourteen states. The U.S. Department of Justice, along with eleven States and the District of Columbia (together, the “government”), filed suit to permanently enjoin the merger on the ground it was likely to substantially lessen competition in at least two markets in violation of Section 7 of the Clayton Act. Following a bench trial, the district court enjoined the merger, rejecting the factual basis of the centerpiece of Anthem’s defense, and focus of its current appeal, that the merger’s anticompetitive effects would be outweighed by its efficiencies because the merger would yield a superior Cigna product at Anthem’s lower rates. The district court found that Anthem had failed to demonstrate that its plan is achievable and that the merger will benefit consumers as claimed in the market for the sale of medical health insurance to national accounts in the fourteen Anthem states, as well as to large group employers in Richmond, Virginia.
Anthem and Cigna (hereinafter, Anthem) challenge the district court’s decision and order permanently enjoining the merger on the principal ground that the court improperly declined to consider the claimed billions of dollars in medical savings. See Appellant Br. 10.1 Specifically, Anthem maintains the district court improperly rejected a consumer welfare standard — what it calls “the benchmark of modern antitrust law,” id. — and generally abdicated its responsibility to balance likely benefits against any potential harm. Ac*349cording to Anthem, the merger’s efficiencies would benefit customers directly by reducing the costs of customer medical claims through lower provider rates, without harm to the providers. The government has not challenged Anthem’s reliance on an efficiencies defense per se. Rather, it points out that Anthem neither disputes that the merger would be anticompetitive but for the claimed medical cost savings, nor challenges the district court’s findings on the relevant market definition, ease of entry, the effect of sophisticated buyers, or innovation. Instead, Anthem’s appeal focuses principally on factual disputes concerning the claimed medical cost savings, which the government maintains were not verified, not specific to the merger, and not even real efficiencies.
For the following reasons, we hold that the district court did not abuse its discretion in enjoining the merger based on Anthem’s failure to show the kind of extraordinary efficiencies necessary to offset the conceded anticompetitive effect of the merger in the fourteen Anthem states: the loss of Cigna, an innovative competitor in a highly concentrated market. Additionally, we hold that the district court did not abuse its discretion in enjoining the merger based on its separate and independent determination that the merger would have a substantial anticompetitive effect in the Richmond, Virginia large group employer market. Accordingly, we affirm the issuance of the permanent injunction on alternative and independent grounds.
I.
Under Section 7 of the Clayton Act, a merger between two companies may not proceed if “in any line of commerce or in any activity affecting commerce in any section of the country, the effect of such [merger] may be substantially to lessen competition.” 15 U.S.C. § 18.
A burden-shifting analysis applies to consider the merger’s effect on competition. United States v. Baker Hughes Inc., 908 F.2d 981, 982 (D.C. Cir. 1990). First, the plaintiff must establish a presumption of anticompetitive effect by showing that the “transaction will lead to undue concentration in the market for a particular product in a particular geographic area.” Id. The most common way to make this showing is through a formula called the Herfin-dahl-Hirschman Index (“HHI”), which compares a market’s concentration before and after the proposed merger. See id. at 983 n.3. By squaring the market share percentage of each market participant and adding them together, a market’s HHI can range from >0 to 10,000 (i.e., a pure monopoly, or 1002). Dept. of Justice & Fed. Trade Comm’n, Horizontal Merger Guidelines § 5.3 & n.9 (Aug. 19, 2010) (the “Guidelines”). Under the Guidelines, a market will be considered highly concentrated if it has an HHI above 2500, and if the merger increases HHI by more than 200 points and results in a highly concentrated market, it “will be presumed to be likely to enhance market power.” Id. § 5.3. Although, as the Justice Department acknowledges, the court is not bound by, and owes no particular deference to, the Guidelines, this court considers them a helpful tool, in view of the many years of thoughtful analysis they represent, for analyzing proposed mergers. See Baker Hughes, 908 F.2d at 985-86.
The burden shifts, once the prima facie case is made, to the defendant to rebut the presumption. Id. at 982. To do so, it must provide sufficient evidence that the prima facie case “inaccurately predicts the relevant transaction’s probable effect on future competition,” or it must sufficiently discredit the evidence underlying the initial presumption. Id. at 991. “The more compelling the prima facie case, the more evi*350dence the defendant must present to rebut it successfully,” but because the burden of persuasion ultimately lies with the plaintiff, the burden to rebut must not be “unduly onerous.” Id.
Upon rebuttal by the defendant, “the burden of producing additional evidence of anticompetitive effect shifts to the [plaintiff], and merges with the ultimate burden of persuasion, which remains with the [plaintiff] at all times.” Id. at 983.
II.
Anthem is the second-largest seller of medical health insurance to large companies in the United States, and it serves approximately 38.6 million medical members. It is a member of the Blue Cross Blue Shield Association, a group of thirty-six health insurance companies licensed to do business under the Blue Cross and/or Blue Shield brands. Anthem holds an exclusive license to the Blue brands in all or part of fourteen states (the “Anthem states”), and it may also compete for business outside those states if it receives permission from the Blue licensee in the relevant area. Anthem also owns non-Blue subsidiaries through which it may operate both in and outside of the Anthem states, subject to Anthem’s “Best Efforts” obligations in its licensing agreement with the Blue Cross Association. Under these “Best Efforts” provisions, at least 80% of Anthem’s revenue within the Anthem states must come from Blue-branded products, as must at least 66.67% of its revenue nationwide. Failure to comply could result in termination of Anthem’s license, which would trigger a $2.9 billion fee to the Association.
Cigna, the third-largest seller of health insurance to large companies in the United States, serves approximately 13 million medical members nationwide and in more than 30 countries, in addition to offering other specialty products such as dental and vision insurance. Unlike Anthem, which has historically been able to leverage its size to negotiate steep discounts from providers, Cigna’s provider discounts have generally not been as good, so Cigna has developed a different and innovative value proposition in order to compete for customers. Under its more collaborative arrangements with providers, and through the integrated, customized wellness programs it offers its customers’ employees, Cigna’s focus is on reducing employees’ utilization of expensive medical procedures and promoting wellness through behavioral supports and lifestyle changes. This offers customers a different means of lowering health care costs than the traditional model relying heavily on provider discounts.
On July 23, 2015, Anthem reached an agreement to merge with Cigna. The merger would leave Anthem as the surviving company, with a controlling share of the merged company’s stock and a majority of seats on the merged company’s board of directors. Within the Anthem states, Cigna customers would be permitted to remain with Cigna, at least for the time being, but Anthem and Cigna would otherwise no longer compete with one another in those states. Outside the Anthem states, Cigna’s existing business would allow Anthem a bigger foothold to compete, subject to Anthem’s ' “Best Efforts” obligations. The merger agreement extends until April 30, 2017.
On July 21, 2016, the United States, along with California, Colorado, Connecticut, Georgia, Iowa, Maine, Maryland, New Hampshire, New York, Tennessee, Virginia, and the District of Columbia, sued to enjoin the merger. Relying on Section 7 of the Clayton Act, 15 U.S.C. § 18, plaintiffs alleged that the merger would substantially lessen competition in the market for the *351sale of health insurance to national accounts in both the Anthem states and the United States as a whole, as well as in the market for the sale of health insurance to large group employers in 35 local markets. Plaintiffs also alleged that the merger would substantially lessen competition for the purchase of services from healthcare providers in the 35 local markets by giving the combined company anticompetitive buying power.
Following a six-week bench trial, the district court permanently enjoined the merger on the basis of its likely substantial anticompetitive effect in the market for the sale of health insurance to national accounts in the Anthem states, as well as in the market for the sale of health insurance to large group employers m Richmond, Virginia. United States v. Anthem, Inc., No. CV 16-1493 (ABJ), — F.Supp.3d —, —, 2017 WL 685563, at *68 (D.D.C. Feb. 21, 2017). It first defined the relevant national accounts market, accepting the government’s proposed definition of “national account” as an employer purchasing health insurance for more than 5,000 employees across more than one state. It also found that the market properly included both fully insured and “administrative services only” (“ASO”) plans. Under a fully-insured plan, the employer pays for claims adjudication, access to the insurer’s provider network (including whatever discounted rates the insurer has negotiated), and coverage of the employees’ medical costs. Under an ASO plan, the employer pays for claims adjudication and network access, but the employer self-insures and thus takes on the risk of its employees’ medical costs. Finally, the district court found that the relevant geographic market for national accounts was the fourteen Anthem states, because that is where Anthem and Cigna currently compete most prominently, given the geographical restrictions imposed on Anthem under its Blue Cross license.
With the national accounts market so defined, the district court then found a presumption of anticompetitive effect based on the combined company’s market share. It determined that the merger would increase HHI by 537 to 3000, while the Guidelines threshold is an increase of 200 to 2500, resulting in a highly concentrated market. Guidelines § 10. It also noted that under any variation performed by plaintiffs’ expert the resulting numbers were still well over the presumptive Guidelines limits: considering only national accounts where 5% of employees reside in another state, HHI would increase 641 to 3124; considering only ASO customers with 5% out-of-state employees, HHI would increase 880 to 3675; and considering all ASO national accounts, HHI would increase 771 to 3663. Anthem objected that these calculations overstated Anthem’s market share by including all Blue customers even if they were not Anthem’s, but the district court found that this was appropriate. Anthem’s own internal calculations include these customers, and a key part of Anthem’s value proposition to customers is that they can access all non-Anthem Blue networks nationwide.
Next, the district court found that Anthem had provided sufficient evidence to rebut the government’s prima facie case. It relied on evidence that Anthem’s primary competitor for national accounts is United Healthcare, not Cigna; that national accounts tend to be sophisticated, well-informed customers and thus better able to thwart an attempted price increase; that new entrants to the market will constrain pricing; and that the combined company would have incentives to innovate in its collaborative care arrangements with healthcare providers.
*352Finally, the district court found that the merger’s overall effect in the Anthem states would be anticompetitive by reducing the number of national health insurance carriers from four to three. It rejected Anthem’s efficiencies defense, which posited the combined company would realize $2.4 billion in medical cost savings through its ability to (1) “rebrand” Cigna customers as Anthem in order for them to access Anthem’s existing lower rates; (2) exercise an affiliate clause in some of its provider agreements to allow Cigna customers access to Anthem rates; and (3) renegotiate lower rates with providers. First, it found that the claimed savings were not merger-specific because they were based on the application of rates that either company was already able to attain, and thus presumably each company could attain the other’s superior rates on its own. It also found that for Cigna customers that would be rebranded to Anthem, any related savings would not be merger-specific because Cigna customers could simply purchase the Anthem product today. It rejected the notion that the merger was necessary to allow Anthem customers access to Cigna’s popular product offerings because Anthem had failed to show that it could not develop and offer these products on its own. Second, the district court found that the claimed savings also failed because they were not sufficiently verifiable. It found that Anthem’s plan to exercise the affiliate clause in its provider contracts was unlikely to work as Anthem suggested. That is, exercise of the affiliate clause would likely give rise to provider resistance because the providers were unlikely to accept lower rates and provide more services without getting anything in return. The district court also found, as a matter of fact, that attempts to achieve the claimed savings through renegotiation of provider contracts would run into similar problems. It found that any savings would take- time to be realized, and that Anthem’s expert failed to account for utilization, i.e., the amount of medical services that would be consumed by a given customer. In sum, it found the claimed savings were aspirational inasmuch as every proffered strategy either floundered in the face of business reality or was achievable without the merger, or both. The district court also expressed doubt as to whether the type of efficiencies claimed by Anthem, which merely redistribute wealth from providers to Anthem and its customers rather than creating new value, are even cognizable under Section 7.
Additionally, with regard to the Richmond market for large group employers, the district court found a presumption of anticompetitive effect based, on the fact that Anthem and Cigna were the city’s first- and second-largest competitors, with a combined market share of between 64% and 78%. It found that Anthem rebutted the presumption by challenging the government’s calculations, pointing to additional competitors outside the Richmond area and claiming that Anthem customers in the Federal Employee Program skewed its Richmond market share. Overall, however, the district court credited the testimony of the government’s expert that even accepting all of Anthem’s claimed efficiencies, the merger would still have a net anticompetitive effect. Because Anthem had not shown that the remaining competition (or potential market entrants) could likely constrain a price increase by the combined company, it found that the merger should be enjoined on that additional basis as well.
III.
Our review of the district court’s decision whether to issue a permanent injunction under the Clayton Act is limited to determining whether there was an *353abuse of discretion. United States v. Borden Co., 347 U.S. 514, 518, 74 S.Ct. 703, 98 L.Ed. 903 (1954); see FTC v. H.J. Heinz Co., 246 F.3d 708, 713 (D.C. Cir. 2001) (“Heinz”). The district court’s conclusions of law are reviewed de novo, and its findings of fact must be affirmed unless clearly erroneous. Heinz, 246 F.3d at 713. If a finding of fact rests on an erroneous legal premise, then the court “must examine the decision in light of the legal principles [it] believe[s] proper and sound.” Id. (quoting Ambach v. Bell, 686 F.2d 974, 979 (D.C. Cir. 1982)).
A.
It is undisputed that the government met its burden to demonstrate a highly concentrated post-merger market, which would be reduced from four to just three competing companies. Anthem also does not dispute the definition of the national accounts market, nor that such a market will be even more highly concentrated post-merger. Anthem’s appeal instead hinges on the district court’s treatment of its efficiencies defense. The premise of its defense was explained by its expert, Mark Israel, Ph.D. According to Anthem, Dr. Israel quantified the medical cost savings that the combined company could achieve post-merger using a “best of best” methodology, based on the economic theory that the combined company, with its greater volume, would be able to obtain discount rates that are no worse than either of the companies could achieve separately. Using claims data from Anthem and Cigna, he calculated that the merger would generate $2.4 billion in medical cost savings through improved discount rates, 98% of which he predicted would be passed through to customers, the large national employers with which Anthem and Cigna contract. Of the $2.4 billion in claimed savings, Dr. Israel projected that $1,517 billion would result from Cigna customers accessing Anthem’s lower rates, while $874.6 million would result from Anthem customers accessing Cigna’s lower rates; when viewed in terms of self-insured versus fully-insured customers, the former would purportedly see $1,772 billion of the claimed $2.4 billion, while the latter would see $619.8 million. Using merger simulation models, he balanced these projected savings against potential anticompetitive effects from the loss of the rivalry between the two companies and found the savings easily outweighed any potential harm. See Appellant Br. 5-6. But, as Anthem tends to ignore, the government offered its own evidence and experts to challenge these conclusions, as we discuss below.
Despite, however, widespread acceptance of the potential benefit of efficiencies as an economic matter, see, e.g., Guidelines § 10, it is not at all clear that they offer a viable legal defense to illegality under Section 7. In FTC v. Procter & Gamble Co., 386 U.S. 568, 87 S.Ct. 1224, 18 L.Ed.2d 303 (1967), the Supreme Court enjoined a merger without any consideration of evidence that the combined company could purchase advertising at a lower rate. It held that “[possible economies cannot be used as a defense to illegality. Congress was aware that some mergers which lessen competition may also result in economies but it struck the balance in favor of protecting competition.” Id. at 580, 87 S.Ct. 1224. In his concurrence, Justice Harlan criticized this attempt to “brush the question aside,” and he “accepted] the idea that economies could be used to defend a merger.” Id. at 597, 603, 87 S.Ct. 1224 (Harlan, J., concurring). No matter that Justice Harlan’s view may be the more accepted today, the Supreme Court held otherwise, id. at 580, 87 S.Ct. 1224, and no party points to any subsequent step back by the Court.
*354Nor does our dissenting colleague, despite his wishful assertion that Procter & Gamble can be disregarded by this court because it preceded the “modern approach” adopted in cases like United States v. General Dynamics Corp., 415 U.S. 486, 94 S.Ct. 1186, 39 L.Ed.2d 530 (1974), and Continental T. V., Inc. v. GTE Sylvania Inc., 433 U.S. 36, 97 S.Ct. 2549, 53 L.Ed.2d 568 (1977). See Dis. Op. 375-77, 378-79. The Supreme Court made no mention of Procter & Gamble in General Dynamics, 415 U.S. 486, 94 S.Ct. 1186, and it cannot be read to have implicitly overruled the earlier decision because it did not involve efficiencies. See id. at 494-504, 94 S.Ct. 1186; see also 4A Phillip E. Areeda & Herbert Hovenkamp, Antitrust Law ¶ 976c2, at 115 (2016) (“Areeda & Ho-venkamp”) (distinguishing between an efficiencies defense and General Dynamics’ “competitive significance” defense). And whatever significance Continental T. V. may have in the area of vertical restraints on trade, 433 U.S. at 54-59, 97 S.Ct. 2549, it did not do the yeoman’s work that the dissent apparently ascribes to it here, for it did not involve efficiencies, mergers, or Section 7 of the Clayton Act. Even stranger is the dissent’s suggestion that our decision in Baker Hughes, 908 F.2d at 986, blessed an efficiencies defense, see Dis. Op. 376-77, because Baker Hughes did not concern efficiencies and, like Heinz, 246 F.3d at 720, it could not overrule Supreme Court precedent. Nor has this court even hinted, as the dissent proclaims, that General Dynamics overruled Procter & Gamble's efficiencies holding. See Baker Hughes, 908 F.2d at 988 (citing Procter & Gamble favorably); Heinz, 246 F.3d at 720 & n.18 (interpreting Procter & Gamble’s efficiencies holding). Put differently, our dissenting colleague applies the law as he wishes it were, not as it currently is. Even if “the Supreme Court has not decided a case assessing the lawfulness of a horizontal merger under Section 7 of the Clayton Act” since 1975, Dis. Op. 376, it still is not a lower court’s role to ignore on-point precedent so as to adhere to what might someday become Supreme Court precedent.
Despite the clear holding of Procter & Gamble, 386 U.S. at 580, 87 S.Ct. 1224, two circuit courts, and our own, have subsequently recognized the use of efficiencies evidence in rebutting a prima facie case. Heinz, 246 F.3d at 720 (citing, inter alia, FTC v. Tenet Health Care Corp., 186 F.3d 1045 (8th Cir. 1999); FTC v. Univ. Health, Inc., 938 F.2d 1206 (11th Cir. 1991)); see also ProMedica Health Sys., Inc. v. FTC, 749 F.3d 559, 571 (6th Cir. 2014). The Eighth Circuit, in holding that the government had produced insufficient evidence of a well-defined market, acknowledged that the district court may have properly rejected the efficiencies defense, while observing evidence of enhanced efficiencies should be considered in the context of the competitive effects of the merger. Tenet Health Care Corp., 186 F.3d at 1053-55. The Eleventh Circuit similarly concluded that whether an acquisition would yield significant efficiencies in the relevant market is “an important consideration in predicting whether the acquisition would substantially lessen competition,” University Health, Inc., 938 F.2d at 1222, while noting both that “[i]t is unnecessary ... to define the parameters of this defense now,” and that “it may further the goals of antitrust law to limit the availability of an efficiency defense,” id. at 1222 n.30. Other circuits have remained skeptical and simply assumed efficiencies can rebut a prima facie case, before finding that the merging parties had not clearly shown the merger would enhance rather than hinder competition. See, e.g., FTC v. Penn State Hershey Med. Ctr., 838 F.3d 327, 348 (3d Cir. 2016); *355Saint Alphonsus Med. Ctr.-Nampa, Inc. v. St. Luke’s Health Sys., Ltd., 778 F.3d 775, 790 (9th Cir. 2015). These very recent decisions put to rest the dissent’s notion that “no modern court” recognizes the continued viability of Procter & Gamble, see Penn State Hershey Med. Ctr., 838 F.3d at 348; Saint Alphonsus Med. Ctr., 778 F.3d at 789, while even á cursory reading of the court’s opinion today puts to rest any suggestion that it “espouses the old ... position that efficiencies might be reason to condemn a merger.” Dis. Op. 379 (emphasis added) (quoting Ernest Gellhorn et al., Antitrust Law and Economics in a Nutshell 463 (5th ed. 2004)).
“Of course, once it is determined that a merger would substantially lessen competition, expected economies, however great, will not insulate the merger from a [Ejection 7 challenge.” Univ. Health, 938 F.2d at 1222 n.29. Notably, Professors Ar-eeda and Hovenkamp have observed that “Congress may not have wanted anything to do with an efficiencies defense asserted by a firm that was already large or low cost within the market and to whom the efficiencies would give an even greater advantage over rivals.” Areeda & Hovenkamp, supra, ¶ 950f, at 42; id. ¶ 970c, at 31. As our subsequent analysis shows, this court, like our sister circuits, can simply assume the availability of an efficiencies defense to Section 7 illegality because Anthem fails to show that the district court clearly erred in rejecting Anthem’s efficiencies defense.
This court was satisfied in Heinz, in view of the trend among lower courts and secondary authority, that the Supreme Court can be understood only to have rejected “possible” efficiencies, while efficiencies that are verifiable can be credited. 246 F.3d at 720 & n.18 (discussing 4 Phillip Areeda & Donald Turner, Antitrust Law ¶ 941b, at 154 (1980)). The issue in Heinz was whether under Section 13(b) of the Federal Trade Commission Act, 15 U.S.C. § 53(b), preliminary injunctive relief would be in the public interest. 246 F.3d at 727. The court held that the district court “failed to make the kind of factual findings required to render that defense sufficiently concrete to rebut the government’s prima facie showing,” id. at 725, and, upon weighing the equities, remanded for entry of a preliminary injunction. Id. at 726-27. The court expressly stated however: “It is important to emphasize the posture of this case. We do not decide whether ... the defendants’ claimed efficiencies will carry the day.” Id. at 727. These are not the issues in Anthem’s appeal from the grant of a permanent injunction. See LaShawn A. v. Barry, 87 F.3d 1389, 1393 (D.C. Cir. 1996) (en banc).
Consequently, the circuit precedent that binds us allowed that evidence of efficiencies could rebut a prima facie showing, Heinz, 246 F.3d at 720-22, which is not invariably the same as an ultimate defense to Section 7 illegality. Cf. generally Saint Alphonsus Med. Ctr., 778 F.3d at 789-90 (and authorities cited therein). In this expedited appeal, prudence counsels that the court should leave for another day whether efficiencies can be an ultimate defense to Section 7 illegality. We will proceed on the assumption that efficiencies as presented by Anthem could be such a defense under a totality of the circumstances approach, see Baker Hughes, 908 F.2d at 984-85 (citing General Dynamics, 415 U.S. at 498, 94 S.Ct. 1186), because Anthem has failed to show the district court clearly erred in rejecting Anthem’s purported medical cost savings as an offsetting efficiency. Guidelines § 10; cf. Heinz, 246 F.3d at 720-22. Additionally, because the district court could permissibly conclude that the efficiencies defense failed, because the amount *356of cost saving that is both merger-specific and verifiable would be insufficient to offset the likely harm to competition, this court has no occasion to decide whether the type of redistributional savings claimed here are cognizable at all under Section 7. It bears noting, though, that all of those other issues pose potentially substantial additional obstacles to this merger.
One further preliminary analytical point. Amici supporting Anthem invite the court to disregard the merger-specificity and verifiability requirements on the ground they place an asymmetric burden on merging parties that could doom beneficial mergers. See Br. for Antitrust Economists and Business Professors as Amicus Curiae in Support of Appellant and Reversal (“Amici Economists”) at 5-7. Anthem itself has not adopted this argument. See Burwell v. Hobby Lobby Stores, Inc., — U.S. —, 134 S.Ct. 2751, 2776, 189 L.Ed.2d 675 (2014); Eldred v. Reno, 239 F.3d 372, 378 (D.C. Cir. 2001). We note, however, that Amici Economists misapprehend the nature of Anthem’s claimed efficiencies as “direct price reductions,” id. at 6-7, rather than as potential price reductions subject to a number of uncertainties. For customers to realize any price reduction, Anthem would first have, to succeed in reducing providers’ rates, and to that extent the purported reductions would not be a direct effect of the merger. By contrast, the merger would immediately give rise to upward pricing pressure by eliminating a competitor, see, e.g., Tr. 960:12-18, and Anthem could unilaterally raise its prices in response. Further, Amici Economists ignore that fully-insured customers, and potentially self-insured customers depending on the terms of their contracts with Anthem, will not see any savings until Anthem takes a second action, renegotiating the customers’ contracts to pass through the savings. This illustrates the reason for the verifiability requirement: Perhaps Anthem is certain to take those actions, and there will be no impediments to the savings’ realization, but that showing is still necessary for a court to conclude that the merger’s direct effect (upward pricing pressure) is likely to be offset by an indirect effect (potential downward pricing pressure). See Guidelines § 10. As for merger-specificity, Amici Economists point to no logical flaw in the policy that consumers should not bear the loss of a competitor if the offsetting benefit could be achieved without a merger. See Heinz, 246 F.3d at 722.
B.
Any claimed efficiency must be shown to be merger-specific, meaning that it “cannot be achieved by either company alone because, if [it] can, the merger’s asserted benefits can be achieved without the concomitant loss of a competitor.” Heinz, 246 F.3d at 722. The Guidelines frame the issue slightly differently: an efficiency is said to be merger-specific if it is “likely to be accomplished with the proposed merger and unlikely to be accomplished in the absence of either the proposed merger or another means having comparable anticompetitive effects.” Guidelines § 10. Anthem faults the district court for considering whether the efficiencies “could” be achieved absent the merger, without regard to likelihood, Appellant Br. 24, even though in Heinz, 246 F.3d at 722, this court spoke repeatedly in terms of possibility (“can” or “could”).
Heinz, 246 F.3d at 721-22, cited the Guidelines with approval in describing the standard for merger-specificity. Both the current and then-current Guidelines refer to “practical” alternatives to achieving the efficiency short of merger, alternatives that are more than “merely theoretical.” Guidelines § 10 (2010); Guidelines § 4 *357(1997). Similarly, in Heinz, 246 F.3d at 722, the court considered whether it was practical for the company to obtain better baby food recipes by investing more money in product development, or whether that would cost more money than the merger itself. The real question is whether the alternatives to merger are practical and more than merely theoretical, see id.; Guidelines § 10. Even assuming there is any difference between the two standards, it would not affect the outcome here on this factual record. Viewed under either articulation, certain of Anthem’s claimed efficiencies fall away.
The crux of Anthem’s argument regarding merger-specificity is the theory that the combined company will allow Anthem to create a “new product” that is “unavailable on the market today”: a product that features both “Cigna’s customer-facing programs” and Anthem’s “generally lower ... rates.” Appellant Br. 26. One way Anthem maintains the merger will result in this new product is through reb-randing. According to Anthem, “rebrand-ing means [the combined company] retain[s] the Cigna product but brand[s] it under the Anthem name with Anthem’s negotiated provider rates.” Appellant Br. 34. The record, however, refutes rather than substantiates Anthem’s proposed reb-randing approach. In fact, the record evidence Anthem cites for its rebranding plan is the testimony of Anthem Senior Vice President Dennis Matheis. But in that testimony, Matheis confirmed that, at least “[i]n the short term,” rebranding would simply involve Anthem “offer[ing] Cigna customers Anthem products,” in a manner that is “no different” than Anthem “selling new business in the market.” Tr. 1599:20-25. In other words, when a Cigna customer rebrands, the immediate effect is that the customer gives up a Cigna contract and Cigna product in favor of an Anthem contract and Anthem product. Indeed, it is only “[o]ver the long haul,” Matheis testified, that Anthem could actualize its “vision ... [to] combine Cigna features ... with Anthem features,” Tr. 1606:17-21, and then rebranding might result in a former Cigna customer obtaining some semblance of the former Cigna product at the new Anthem rate. But rebranding in the immediate aftermath of the merger would involve a Cigna customer switching to the extant Anthem product, and that is not a merger-specific outcome; that is just more successful marketing of the existing Anthem product. And Anthem expressly “does not contend that ... a customer simply switching from a Cigna product to an existing Anthem product[] results in merger-specific efficiencies.” Appellant Reply Br. 21.
Instead, Anthem claims only that reb-randing over the long haul, once it has successfully rolled out an improved, Cigna-like product, will result in a merger-specific benefit, and maintains that the district court clearly erred in finding Anthem could simply develop and offer an improved product on its own. Just as in Heinz, 246 F.3d at 722, the evidence offered by Anthem is woefully insufficient to show that it cannot develop better customer-facing programs. Anthem points to testimony from two witnesses that Anthem has failed to replicate Cigna’s products, for reasons unknown. In particular, Anthem’s President of Specialty Business Pam Ke-haly testified that Cigna offers a “packaged integrated wellness approach where [Anthem offers] disparate pieces that employers kind of have to piece together on their own.” Kehaly Depo. Tr. 87:12-15 (Apr. 28, 2016). According to Kehaly, Anthem has been trying to solve the problem for “probably a decade” but for whatever reason it just has “not been able to crack this nut.” Id. at 88:3-13. She did not indicate how intensive the effort has been, how *358many hours were devoted to it, or how much money Anthem has allocated toward it. Anthem’s Regional Vice President of Sales Brian Fetherston also testified that Cigna has “done a really good job of building wellness programs” and that Anthem has tried but failed to catch up. Fetherston Depo. Tr. 170:14-19 (May 6, 2016). The district court could properly find that failure likely results more from Anthem’s own no-frills culture or flawed marketing strategies than from any inherent difficulties in pulling together an integrated wellness program. For instance, Fetherston testified that Cigna is “significantly better at marketing” its wellness program, while by contrast Anthem “just [was not] actively promoting” its own, and indeed, Anthem recently decided to “dial back some of [its] disease management programs,” which Fetherston believed was a mistake. Id. 169:1-170:6, 323:1-23. To the extent Anthem has failed to devote the resources needed to improve its product, it is in no position to claim that consumers will benefit from it swallowing up Cigna’s superior product.
Put differently, rebranding does not create a merger-specific benefit in either the short- or long-term. Perhaps Anthem could create some brief, interim benefit in the mid-term by integrating Cigna’s product faster than it could develop a comparable product of its own. Guidelines § 10 n.13 (“If a merger affects not whether but only when an efficiency would be achieved, only the timing advantage is a merger-specific efficiency.”). But Anthem made no sufficient factual showing in the district court on this point. It has offered no evidence to show how long it would take, once the necessary resources were allocated, to develop an improved product. Nor has it shown how long it would take to roll out a hybrid Anthem-Cigna product. At oral argument, Anthem claimed that it could do so in six months, but at trial, Anthem’s Senior Vice President Matheis allowed that it might not be able to do so within two-and-a-half years.
To the • extent Anthem also maintains that none of Dr. Israel’s claimed savings are dependent on rebranding, it ignores the reality of his economic model. Without one of its mechanisms to get current Cigna customers access to Anthem rates, none of the $1,517 billion in claimed Cigna savings could be realized. Although Dr. Israel may have been agnostic as to which mechanism is used to achieve those savings, he acknowledged that rebranding would achieve a portion of them: “If there was rebrand-ing as a way to get the discounts ... that would just be another way to get them faster.” Tr. 2108:9-11. Given that rebrand-ing is the linchpin of Anthem’s post-merger strategy, because it is the only option that helps Anthem comply with its “Best Efforts” obligations, the inability to credit rebranding savings seriously undermines Anthem’s efficiencies defense.
The district court further found that none of the medical cost savings are merger-specific because they are based on an application of rates that each of the companies has already achieved on its own. Anthem, quite reasonably, challenges this finding with regard to Cigna’s rates on the ground that “there is no dispute that [Cig-na] has generally secured less favorable provider rates than Anthem for years and has been unable to close that gap despite its best efforts.” Appellant Br. 27. The record shows that, by its own account, Cigna has been unable to match Anthem’s volume-based discounts and instead has had to compete on quality and innovation. Even the government’s expert, Dr. David Dranove, agreed that a true Cigna product at Anthem rates would not be achievable absent the merger. That the district court clearly erred in finding that the application of Anthem rates to customers that choose *359to remain with Cigna is not merger-specific, however, is immaterial to the district court’s ultimate conclusion that the merger would be unlawful because these claimed efficiencies are not sufficiently verifiable.
C.
Under the Guidelines, projected efficiencies will not be credited “if they are vague, speculative, or otherwise cannot be verified by reasonable means.” Guidelines § 10. Anthem maintains that the district court clearly erred because the $2.4 billion in projected post-merger savings was verified by two independent sources (Dr. Israel and an integration planning team from McKinsey & Company, which had access to each company’s internal files). In Anthem’s view, the district court also erred as a matter of law by imposing a “virtually insurmountable burden” of persuasion, Appellant Br. 38, when all that is required is to show “probabilities, not certainties,” Baker Hughes, 908 F.2d at 984.
As discussed, Anthem plans to achieve the claimed savings through a combination of three mechanisms: rebranding, renegotiating provider contracts, and exercising Anthem’s affiliate clause. The district court found that practical business realities would undermine the execution of that plan, making achievement of the savings speculative, and therefore unverifiable. With regard to the affiliate clause, the district court focused on evidence of the potential for provider discontent if the lower Anthem rates are forced on providers that must expend extra effort and resources to deliver the Cigna product, without any corresponding increase in value for providers. This evidence included testimony by both Anthem and Cigna witnesses as well as documents from Anthem and Cigna that acknowledged the likely “abrasifon].” E.g., Pls.’ Ex. 89. The record indicates that physician contracts can be terminated by either party with only 90 days’ notice, so the affiliate clause would accomplish little if the contract is terminated or renegotiated soon after the clause is exercised. Hospital contracts tend to involve three-year commitments, so the affiliate clause may bind them to offer lower rates for a longer period. Still, when those hospital contracts expire, large delivery systems with greater leverage “could push back hard” in renegotiation. Pls.’ Ex. 717. In either event, it is probable, as Cigna CEO David Cordani testified, that some providers will eventually “react [by] renegotiating ... and putting upward pressure on rates, which has been a market force to date.” Tr. 443:17-23. That “very few” Anthem providers have preemptively sought to renegotiate proves little, see Tr. 1686:15-25, because the feared abrasion would not occur until Anthem invokes the affiliate clause, assuming it ever does so.
This raises another practical difficulty with the affiliate clause: although it is theoretically useful to Anthem, in reality it is unlikely to be widely exercised because it works counter to Anthem’s contractual obligations. Under the “Best Efforts” clause in Anthem’s licensing agreement with the Blue Cross Blue Shield Association, 80% of Anthem’s revenue within the Anthem states must be Blue-branded, as must 66.67% of its revenue nationwide. The merger would immediately throw Anthem out of compliance and so Anthem intends to rebrand a “lion’s share” of current Cigna customers in order to count that revenue as Blue-branded. Tr. 1600:17-21 (Anthem Sr. VP Matheis). By contrast, widespread exercise of the affiliate clause would remove any incentive for Cigna customers to convert to Anthem because those customers would then be receiving the Cigna product at Anthem prices, Dr. Israel’s much-touted “best of both worlds” scenario. Anthem fails to ad*360dress this reality when it maintains that 80% of the savings to Cigna customers could be achieved rapidly using the affiliate clause. See Appellant Br. 40. Because doing so would work contrary to Anthem’s own contractual obligations, its witnesses conceded that it will instead rely heavily on rebranding, which, as discussed, gives rise to no merger-specific benefits.
As for renegotiation, the short answer is that if Anthem cannot persuade providers to extend lower rates to Cigna under its affiliate clauses — where it has apparent contractual recourse to do so — then it is speculative that Anthem could get them to agree to do the same thing through negotiations absent compulsion. Anthem assumes, as did Dr. Israel’s model, that in all instances renegotiation would result in providers accepting the lower Anthem rates. That assumption appears questionable in the case of a provider that has just terminated a contract because Anthem mandated, through an affiliate clause, the acceptance of those very rates. Instead, Cigna’s CEO Cordani predicted such renegotiation would put upward pressure on the Anthem rate, and to the extent Anthem were to adopt a take-it-or-leave-it approach, the provider could simply choose to walk away. See Br. for Amici Am. Med. Ass’n. & Med. Soc’y of D.C. (“AMA Br.”) 11-12. This is especially true for large hospital networks with significant bargaining power.
To the extent that some medical savings would be achieved for Cigna customers at the bargaining table due to the combined company’s volume, the district court expressed concern over how long such savings would take to be realized. Anthem’s CEO Swedish testified that capturing medical savings requires a “long gestation period,” in part because existing hospital contracts span three to five years and would not be subject to renegotiation “for a considerable period of time.” Tr. 337:21-338:16. He also rejected the idea that Anthem would simply “drop[] the hammer” on providers by insisting on maximum discounts across-the-board because Anthem instead relies upon “customized relationship-driven eontract[s]” that seek to optimize performance on a case-by-case basis, rather than focusing solely on discounts. Tr. 294:20-295:11. Anthem’s expert agreed that renegotiations in the ordinary course of business will take place over time. The longer it takes for an efficiency to materialize, the more speculative it can be, see Guidelines § 10 & n.15, so the district court was on solid ground to give less weight to the claimed renegotiation savings.
In sum, although renegotiation will lead to a decrease in Cigna’s rates, the assumption that it will in every instance lead to the Anthem rate is farfetched. See Tenet Health Care Corp., 186 F.3d at 1054. Indeed, as the district court observed, “the Department of Justice is not the only party raising questions about Anthem’s characterization of the outcome of the merger” because Cigna itself had “provided compelling testimony undermining the projections of future savings.” Anthem, Inc., — F.Supp.3d at —, 2017 WL 685563, at *4; see also Pls.’ Ex. 722.
Whatever mechanism is employed to achieve the savings, the district court had “reason to question ... whether the quality of the Cigna offering will in fact degrade” as a result of the merger, Anthem, — F.Supp.3d at —, 2017 WL 685563, at *61, which further undermines the purported efficiency claims. Guidelines § 10. For those that choose to stay with Cigna post-merger — and thus would access lower rates through renegotiation or exercise of the affiliate clause — the abrasion problem arises because providers would be asked to *361continue offering the high-touch, collaborative Cigna service, with its added behavioral, wellness, and lifestyle programs, for less money. See AMA Br. 10-11. It was perfectly reasonable for the district court to find that some providers, even if they are willing to accept less money, will simply respond by offering customers less in the way of Cigna high-touch service. Furthermore, according to Cigna’s CEO Cor-dani, the value of the Cigna offering will be diminished because Anthem’s rebrand-ing strategy will siphon business away from Cigna, leaving behind an atrophied Cigna customer base that is less attractive to providers. This will in turn diminish Cigna’s capacity for further innovation with its collaborative model. And for Cigna customers that agree to migrate to Anthem (or are pushed into doing so because the company refuses to extend their expiring Cigna contracts), provider abrasion again rears its head, this time with providers being asked to offer Anthem customers better, and more resource-intensive, collaborative service for the same rates they have historically received. Anthem does not respond meaningfully to these concerns, simply labeling them “speculation.” Appellant Reply Br. 10. In light of the numerous Anthem witnesses who acknowledged the abrasion problem, the district court did not err in finding it “dubious” that Anthem would be able to offer a true Cigna-like product, or that legacy Cigna would be able to maintain the quality of its own product. Anthem, — F.Supp.3d at —, 2017 WL 685563, at *59, *61.
The fact is, it is widely accepted that customers value the existing Cigna product, and that Cigna is a leading innovator in collaborative patient care. That threat to innovation is anticompetitive in its own right. Cf. United States v. Cont’l Can Co., 378 U.S. 441, 465, 84 S.Ct. 1738, 12 L.Ed.2d 953 (1964). And the problem is neither answered by Anthem’s evidence nor offset by its purported efficiency of offering a degraded Cigna product at a lower rate.
In addition to claimed savings to current Cigna customers, Dr. Israel also projected that $874.6 million in savings would be realized if Anthem’s customers were able to access superior rates that Cigna has already negotiated. In focusing almost entirely on the other side of the ledger, Anthem offers little reason to think that the district court clearly erred in rejecting the claimed savings to existing Anthem customers. See Appellant Br. 41-42. To the extent Anthem argues on appeal that Anthem customers could access Cigna’s superior rates through rebranding or exercise of an affiliate clause, the only witness it cites was actually discussing the affiliate clause in Anthem’s contracts that would apply to Cigna’s customers. And even assuming that Cigna’s contracts contain an affiliate clause, Blue Cross Association rules would prohibit Anthem from exercising it. Further, rebranding Anthem customers to Cigna would only exacerbate Anthem’s “Best Efforts”' problem, which indicates why Anthem Senior VP Matheis testified that Anthem would rebrand a lion’s share of Cigna customers to Anthem, not the other way around. Renegotiation would be the only viable option for realizing the projected savings to Anthem customers.
• Moreover, Anthem has not explained why these projected savings would even exist. The record is clear that Anthem, unlike Cigna, has already achieved whatever economies of scale are available. According to Anthem’s expert Dr. Robert Willig, in the 35 local markets identified in the government’s complaint, the data did not show that Anthem’s size correlated with its provider discounts. To the contrary, Dr. Willig testified that Anthem is “already past the threshold of having *362enough size to do what it needs to do in terms of offering volume to providers.” Tr. 2231:9-12. Similarly, Anthem’s CEO Joseph Swedish denied that Anthem would seek to negotiate even greater volume-based discounts after the merger because post-merger Anthem would “certainly not [pay] less than what [it is] now paying as Anthem.” Tr. 294:10-19. The evidence indicates that where Cigna has better discounts than Anthem, that is a result of factors other than volume, and the district court reasonably questioned whether those atypical discounts would remain available post-merger. In the absence of an additional volume-based discount, then, Anthem makes no effort to show how its current customers would see lower prices as a result of the merger, and certainly not to the tune of $874.6 million. Consequently, the district court did not clearly err in rejecting these alleged medical cost savings as unverifiable.
Next, the claimed medical cost savings only improve consumer welfare to the extent that they are actually passed through to consumers, rather than simply bolstering Anthem’s profit margin. See Univ. Health, Inc., 938 F.2d at 1223. After all, the merger potentially harms consumers by creating upward pricing pressure due to the loss of a competitor, and so only efficiencies that create an equivalent downward pricing pressure can be viewed as “sufficient to reverse the merger’s potential to harm consumers ..., e.g., by preventing price increases.” Guidelines § 10; see also Areeda & Hovenkamp, supra, ¶ 971a, at 48 (“[A] sufficient amount of any efficiencies [must] be passed on that the post-merger price is no higher than the pre-merger price.”). Dr. Israel testified that absent monopsony (ie., the exercise of market power to gain subcompetitive prices from providers), any cost savings will create downward pricing pressure, and while this is unobjectionable, the amount passed through to consumers indicates the strength of that pressure. See Br. of Professors as Amici Curiae in Support of Ap-pellee and for Affirmance 7-8 (“Amici Professors”). The district court rightly cast doubt on Dr. Israel’s estimated pass-through rate of 98%, which was unsupported by the evidence and treated self-insured and fully-insured customers identically.
Because ASO customers pay their employees’ medical costs directly, any reduction in medical rates would result in savings that automatically pass through to the customer, absent some corresponding ASO price increase by Anthem. This would improve the quality of one aspect of the ASO product (ie., access to more deeply discounted network rates), and it could thus be procompetitive even if it did not immediately result in an ASO price decrease. See Guidelines § 10. Dr. Israel’s analysis rested on the assumption that rather than raising ASO prices to capture the medical cost savings, Anthem would attempt to increase its market share by providing a much superior product at only a slightly higher price, thereby maximizing its profits through increased sales. The district court highlighted internal Anthem documents that discussed ways to keep those savings for itself, in particular where Anthem listed seven alternatives with 100% pass-through to ASO customers considered last. Contrary to Dr. Israel’s assumption, then, Anthem apparently concluded that total pass-through was not the profit-maximizing, “optimal solution to' capture the most value from [the] deal,” and that it could actually lose business if customers initially saw savings that were not sustained over the long term. Pis.’ Ex. 727. Amici Professors offer another reason why Anthem might have come to this conclusion: in highly concentrated markets, already-large insurers are less constrained by competition and thus tend to find it *363more profitable to capture medical savings and increase premiums. Amici Professors Br. 6-9; see also Arbeda & Hovenkamp, supra, ¶ 971f, at 56 (in highly concentrated markets “there is less competition present to ensure that the benefit of efficiencies will flow to consumers”). That corroborates rather than remediates anticompetitive concerns.
As for fully insured customers, which comprise $619.8 million of the projected savings, the estimated pass-through is even less likely given that the savings would automatically inure to Anthem’s benefit absent some corresponding price decrease to its customers. Dr. Israel recognized this dynamic at trial, and yet his model takes no account of it, applying a pass-through rate of 98% to both ASO and fully insured accounts. The record indicates that ASO customers, which pay medical costs directly to providers, are keenly attuned to fee transparency, but it is unclear fully-insured customers are afforded the same transparency. That is, if Anthem negotiates provider rates and pays providers directly, how would a customer be aware that Anthem had achieved medical savings, in order to be able to seek a pass-through in the form of a lower negotiated price? Further, when would fully-insured customers realize that renegotiated price, given that their existing contracts would not pass though any savings? See Tr. 2107:17-21 (Dr. Israel: ordinary-course renegotiation of employers’ contracts “will take place over time”). Neither Anthem nor Dr. Israel answers (or addresses) these problems.
Finally, the district court did not clearly err when it criticized Anthem’s failure to account in its projected savings for utilization, which is a signature aspect of the Cigna product. Dr. Israel’s model was based on discounts that either company was able to achieve on its own multiplied by the total claims value, but as Anthem’s CEO Swedish testified, “We don’t live in a discount world any longer.” Tr. 295:11. Cigna’s CEO Cordani agreed: “If you’re looking [only at] a discount calculation, if [Anthem] has a 2 percent lower discount for the emergency room service, you would assume that that’s a savings,” unless Cig-na’s wellness program helps the patient avoid that emergency room visit altogether. Tr. 443:10-16. Anthem maintains that Dr. Israel and the McKinsey & Co. team did account for utilization, because Dr. Israel testified that lower utilization would result in a lower total claims value, a value that factored into both his and the McKin-sey & Co. models. But this ignores that on cross-examination, Dr. Israel conceded that he did not control for the different risks and features of each company’s population at a particular provider, which would be necessary to compare utilization, and so his model did not account for whether one company’s utilization w;as better than the other’s. And although Anthem nevertheless maintains that no evidence shows accounting for utilization would materially reduce the claimed savings, Dr. Dranove testified that any error or incorrect assumption would have a significant effect on the overall projected savings. See Tr. 2327:15-2329:11. Thus, the problem is less that the failure to account for utilization would necessarily reduce the projected savings, and more that it undermined the district court’s confidence in the reliability and factual credibility of those savings calculations.
Both sets of projections suffered from additional, basic analytic flaws. For instance, Dr. Israel did not agree with the district court’s national accounts market definition (employers with 5,000 + employees), so his savings projection was based on the broader market definition that he believed appropriate (large-group employers with either 50+ or 100+ employees). *364In other words, Dr. Israel’s claimed $2.4 billion in savings is unmoored from the actual market at issue, and there is no indication of what portion is properly derived from the national accounts market. Similarly, the McKinsey & Co. analysts based some of their savings on a comparison of Cigna and Anthem rates where only one of the companies had negotiated a discount with that particular provider. This apples-to-oranges comparison of in-network versus out-of-network rates overstated the true disparity between the companies’ existing discounts and thus necessarily inflated McKinsey & Co.’s projected savings. Even Dr. Israel acknowledged as much: his model only compared in-network rates because he concluded that “that’s what the economics tells you you need to do to get the answer right.” Tr. 1855:2-22. This could help to explain why Dr. Israel’s otherwise similar methodology resulted in a projection that was almost a billion dollars less than McKinsey & Co.’s.
The savings projected by McKinsey & Co. and Dr. Israel — uncritically relied on by the dissent, e.g., Dis. Op. 373-75, 380— were without a doubt enormous. The problem is, those projections fall to pieces in a stiff breeze. If merging companies could defeat a Clayton Act challenge merely by offering expert testimony of fantastical cost savings, Section 7 would be dead letter.
D.
Having considered the totality of circumstances, see Baker Hughes, 908 F.2d at 984, we hold that the district court reasonably determined Anthem failed to show the kind of “extraordinary efficiencies” that would be needed to constrain likely price increases in this highly concentrated market, and to mitigate the threatened loss of innovation. Cf. Heinz, 246 F.3d at 720. Given the record evidence, Anthem’s objection that the district court “abdicated its responsibility” to balance the merger’s likely benefits against its potential harm, Appellant Br. 46, rings hollow. Anthem seems to insist upon a dollar-for-dollar comparison after discounting whatever claimed efficiencies were properly rejected, in responding that “so long as at least one-third of the $2.4 billion of savings are likely to be achieved, the merger is pro-competitive.” Appellant Reply Br. 10. This would apparently require the court to calculate, for instance, a more realistic pass-through rate than the rejected 98% figure, or to estimate what percentage of the claimed $2.4 billion was attributable to customers with fewer than 5,000 employees and thus outside of the relevant market. Anthem has pointed to no relevant expert economic evidence on which to base such an imprecise calculation, and Anthem, not the district court, has the burden of showing what portion of the claimed efficiencies will result from the merger itself. Even assuming it were possible on this record, see University Health, 938 F.2d at 1223, “[ejconomies cannot be premised solely on dollar figures, lest accounting controversies dominate § 7 proceedings,” Procter & Gamble, 386 U.S. at 604, 87 S.Ct. 1224 (Harlan, J., concurring). Because “the state of the science does not permit such refined showings,” commentators have recommended simply giving the government the benefit of the doubt in a close case. See Areeda & Hovenkamp, supra, ¶ 971f, at 56. In any event, this is not a close case.
The dissent’s critique of the court’s opinion is not well founded. Its fundamental flaw is the failure to engage with the facts shown in the record as they pertain to merger-specificity and verifiability. Repeated references to unspecified evidence, see, e.g., Dis. Op. 373, 377, 378, 380, on which the dissent bases sweeping conclusions, speak volumes. Rather than engage *365with the record, much less adhere to our standard for reviewing findings of fact by the district court, the dissent offers a series of bald conclusions and mischaracter-izes the court’s opinion. For instance, the dissent repeatedly claims that the court “does not fully accept the fact ... that providers rates would actually be lower,” Dis. Op. 380; id. at 379, when in fact the court accepts that rates would be lower for some existing Cigna (but not Anthem) customers post-merger. Those who rebrand with Anthem, however, will see no merger-specific savings, Op. 357-58, and the few that Anthem fails to rebrand will see far fewer savings than Anthem claims, due in large part to provider abrasion and big hospital systems that will stand their ground in renegotiation, Op. 359-61. The dissent baldly asserts that the efficiencies “are merger-specific by definition,” Dis. Op. 374-75, without addressing the paucity of evidence that Anthem would be unable to develop a Cigna-like product without merging. So too, it baldly asserts that the savings were “sufficiently verified,” while admitting that it is not clear “just how much the employers would benefit from this merger.” Id. at 375 (emphasis omitted). In other words, Anthem estimated an astronomical amount of savings, so even if that amount were wildly overstated, the dissent expects the court to trust that, as an unknown fraction of a large number, the result “would be large.” Id.
To the extent the dissent notes any of the major factual problems with Anthem’s depiction of the merger, it brushes them aside. It dismisses as “highly speculative” the provider abrasion problem that was conceded by both Anthem and Cigna witnesses, Dis. Op. 380, a problem that undermines Anthem’s plans for realizing the savings through the affiliate clause and renegotiation. It characterizes record evidence that squarely contradicts Anthem’s pass-through estimates — Anthem’s own internal PowerPoint presentation — as “secret Anthem plans to dramatically raise [its] fees,” id., which is precisely what the evidence reflects. It attacks straw men like supposed reliance on “friction between the Anthem and Cigna CEOs,” id., when the court does not so rely, noting indeed the limited probative value of that evidence. Op. 348 n.1. It fails entirely to address Anthem’s “Best Efforts” obligations, which make it likely that Anthem will rely predominantly on rebranding, a strategy that gives rise to no merger-specific benefit. Op. 359-60. The “Best Efforts” clause also creates a verifiability problem with regard to Anthem’s other savings strategies, for instance undercutting Anthem’s assertion that 80% of the savings to Cigna customers “could be achieved simply [and rapidly] by invoking the affiliate clause.” Appellant Br. 40. Again, pulling at any one loose thread quickly unravels Anthem’s narrative, but the dissent is simply unwilling to do so.
Ultimately, the dissent concludes that “[o]n this record, there is little basis to doubt that the cost savings for employers as a result of the merger would be large,” without evincing any real awareness of the record beyond the testimony of Anthem’s expert and consultants. See Dis. Op. 375. To wit, the dissent suggests that Anthem’s savings estimates went unrebutted at trial, Dis. Op. 374, when the record shows Dr. Dranove not only offered his own estimate of $100 million to $500 million but explained why those savings were unlikely to be realized, essentially for the reasons discussed in this opinion. See, e.g., Tr. 3802:25-3803:11. Similarly, it recognizes no distinction between savings to existing Cigna customers (some of which the government concedes will materialize) and savings to existing Anthem customers (the existence of which even Anthem cannot *366explain in light of the testimony of both its CEO and expert, see supra Part III.C). Likewise, in concluding that the quality of the Cigna product (its wellness programs and the high-touch service that providers offer in support of the programs) will not degrade post-merger, the dissent does not go so far as to say there is no evidence to support the district court’s contrary finding, Anthem, — F.Supp.3d at — - —, 2017 WL 685563, at *59, *61; rather, it asserts it does not consider this evidence “persuasive” or “convincing.” Dis. Op. 380. Such de novo analysis throughout the dissent betrays no meaningful effort to engage with the district court’s factual findings, which are subject only to review for clear error.
Furthermore, the dissent’s assumption that the prices paid by consumers (regardless of the quality of the resulting product) are the sole focus of antitrust law is flawed. “The principal objective of antitrust policy is to maximize consumer welfare by encouraging firms to behave competitively.” Kirtsaeng v. John Wiley & Sons, Inc., 568 U.S. 519, 133 S.Ct. 1351, 1363, 185 L.Ed.2d 392 (2013) (emphasis added) (quoting 1 P. Areeda & H. Hovenkamp, Antitrust Law ¶ 100, at 4 (2006)). This single-minded focus on price ignores that in highly concentrated markets like this one, lower prices, if they occur at all, may be transitory. Owing to the lower level of competition in highly concentrated markets, when presented with lower supply input prices, companies have a greater ability to retain for themselves the input savings rather than pass them on to consumers. The Clayton Act, as the Supreme Court “ha[s] observed many times, [is] a prophylactic measure, intended primarily to arrest apprehended consequences of in-tercorporate relationships before those relationships could work their evil.” Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477, 485, 97 S.Ct. 690, 50 L.Ed.2d 701 (1977) (internal quotation marks and citation omitted). The ability of a firm to obtain lower prices for inputs for its product (here, provider services) should, especially in light of the prophylactic nature of the Clayton Act, be viewed skeptically when high market concentrations may have the future effect of permitting capture of those savings. The dissent uncritically accepts Dr. Israel’s rosy testimony to the effect that ASO savings “will be passed through to employers,” Dis. Op. 379, but fails to address contrary Anthem documents and the historical tendencies of large companies in highly concentrated markets to capture savings. E.g., Areeda & Hovenkamp, supra, ¶ 971f, at 56. The dissent also ignores the district court’s numerous and not clearly erroneous findings, as previously discussed, that total or nearly total pass-through is unlikely. See, e.g., Anthem, — F.Supp.3d at —,- —, 2017 WL 685563, at *4, *62. Even if ASO savings would pass through in the short term, that does not “practically guarantee[ ]” that Anthem would not then raise its prices correspondingly. But see Dis. Op. 380.
Additionally, the dissent fails to recognize that lower prices may arise due to, or ultimately lead to, a decrease in product quality. Everyone would agree that rock-bottom provider rates seem beneficial to consumers, but when those rock-bottom prices lead to inferior medical services, any benefit to the consumers’ wallets is diminished by the harm to their health. As the Guidelines state, if merging firms “would withdraw a product that a significant number of customers strongly prefer to those products that would remain available, this can constitute a harm to customers over and above any effects on the price or quality of any given product.” Guidelines § 6.4; see also id. § 10 (“[PJurported efficiency claims based on lower prices can be *367undermined if they rest on reductions in product quality or variety that customers value.”). And a decrease in product quality is not merely speculative here — every dollar of medical cost savings realized by consumers will come at the expense of providers. It thus is quite plausible that paid less, the medical providers will provide less. These inconvenient facts do not jibe with the dissent’s superficial, thirty-thousand-foot view of this case, and it is thus unsurprising that they are addressed in passing, if at all.
IV.
Anthem fares no better in its challenge to the district court’s independent and alternative determination that the merger should be enjoined on the basis of its anticompetitive effect in the Richmond, Virginia market for the sale of health insurance to “large group” employers with more than fifty employees. There, the government’s prima facie case was even stronger than in the market for national accounts in the fourteen Anthem states. Depending on how market share was calculated {i.e., including all Blue customers as Anthem or not, including fully insured customers or just ASO), the companies’ combined market share ranged from 64% to 78%. Even under the calculation most favorable to Anthem (ASO-only, disregarding non-Anthem Blue customers), the merger would raise an overwhelming presumption of anticompetitive effect: HHI would rise 1511 to a post-merger total of 4350, where the Guidelines presumption threshold is an increase of 200 to a post-merger total of 2500. As the President of Anthem Virginia acknowledged, Anthem has the biggest share of the large group employer market across all of Virginia, and in Richmond, Cigna is its strongest competitor.
Anthem principally challenges the district court’s reliance on a chart prepared by Dr. Dranove, the government’s expert witness, showing the merger would have an anticompetitive effect in Richmond even crediting all of Dr. Israel’s' claimed efficiencies. The chart included an asterisk next to the Richmond entry signifying that “no amount of cost savings could offset employer harm due to decreased competition.” Pis.’ Ex. 760. On cross examination, Dr. Dranove was asked whether that meant even a savings of $10 billion or $20 billion would not offset the merger’s harm, and he acknowledged that he could not recall the foundation for his statement. Given this inability to address that extreme hypothetical, Anthem maintains that the district court should not have relied on the statement or even on the chart as a whole.
The record shows that the district court did not rely on the “asterisk” statement and explained at trial that it would not do so because it was unnecessary to finding a substantial anticompetitive effect. As to the broader question whether Dr. Dranove’s inability to explain the asterisk meant that the district court should have disregarded his chart (and related testimony) altogether, the district court did not abuse its discretion. See Heller v. District of Columbia, 801 F.3d 264, 272 (D.C. Cir. 2015). Leaving the asterisk statement aside, Anthem raises no real objection to the substance of the chart, only urging that Dr. Dranove’s inability to explain the asterisk was so damaging that it called into doubt the reliability of his overall analysis. The district court, which heard extensive testimony from Dr. Dranove about the anticompetitive effects revealed by his economic models and relied on it heavily throughout its opinion, clearly concluded otherwise. The district court had “considerable leeway” to do so in deter*368mining that all other aspects of the chart and his testimony were reliable. Kumho Tire Co. v. Carmichael, 526 U.S. 137, 152, 119 S.Ct. 1167, 143 L.Ed.2d 238 (1999); cf. Snyder v. Louisiana, 552 U.S. 472, 477, 128 S.Ct. 1203, 170 L.Ed.2d 175 (2008).
Anthem’s remaining challenges amount to an ineffectual attack on the district court’s weighing of rebuttal evidence. It incorrectly states that the district court relied solely on Dr. Dranove’s chart to find anticompetitive harm while ignoring evidence of “enormous” medical savings, Appellant Br. 53, when in fact Dr. Dranove testified that his chart credited 100% of Anthem’s claimed savings and still found a net anticompetitive effect. Anthem posits that there would still be five or more competitive insurers in Richmond post-merger, but even assuming that is true (one of the two witnesses it cites identified only four, including the combined company), the mere existence of competitors may not be sufficient to constrain a larger Anthem that would control 64% to 78% of the market. See Guidelines § 5. Indeed, one of those competitors, Optima, was said to have struggled in the Richmond market, and Anthem shows no clear error in the district court’s finding that Optima “does not appear able to compete on the same field as the merged company.” Anthem, — F.Supp.3d at —, 2017 WL 685563, at *68. Nor does Anthem show clear error in the finding that other companies do not appear interested in entering the Richmond market, or that even if they did, their entry would be insufficient to constrain the combined company. The evidence cited by Anthem shows only that other companies may intend to enter Richmond (Innovation), or may have the ability to enter Richmond (Piedmont, VCU), or may have a marginal or embryonic presence in Richmond (Kaiser, Bon Secours), not that entry by these companies would offset the merger’s anticompetitive potential.
Tellingly, our dissenting colleague offers a single mention of the district court’s Richmond holding (in a parenthetical, no less), which itself is a sufficient basis for enjoining the merger. Any suggestion that the claimed savings would make the merger procompetitive in Richmond ignores the record evidence, namely that even crediting all of the claimed savings, the merger of Richmond’s two biggest large-group insurers would give the combined company such a vast market share that the overall effect of the merger would still be anticom-petitive. As Dr. Dranove testified at trial, his analysis “still predicts a price increase” in the Richmond market “even [after] crediting every penny of th[e] efficiencies” estimated by Dr. Israel. That is, even ignoring Anthem’s failure to show that the savings were merger-specific and sufficiently verifiable, see supra Part III.B-C, the proposed merger would cause an already highly concentrated market to become overwhelmingly so, with Anthem controlling as much as 78% of the market and two or three other companies fighting to maintain relevance. Although the dissent recognizes this appeal raises “fact-intensive question[s],” Dis. Op. 378, it has persistently failed to engage with the facts.
In conclusion, the district court did not clearly err in its factual findings that the merger would have anticompetitive effects in the Richmond market, and importantly, Anthem does not allege any error of law with respect to that determination. Thus, the district court did not abuse its discretion in enjoining the merger on the basis of the merger’s anticompetitive effects in the Richmond market. And, as previously noted, this holding provides an independent basis for the injunction, even absent a finding of anticompetitive harm in the fourteen-state national accounts market.
*369Accordingly, we affirm the issuance of the permanent injunction on alternative and independent grounds.
. Cigna has become a reluctant supporter of the merger, stating in its appellate brief that "[i]n accordance with the merger agreement, Cigna has appealed and defers to Anthem.” Cigna Br. 3. Indeed, the district court noted the "elephant in the courtroom,” for at trial Cigna executives dismissed various of Anthem's claims of savings, cross-examined the merging parties' expert witness, and refused to sign Anthem’s proposed findings of fact and conclusions of law. United States v. Anthem, Inc., No. CV 16-1493 (ABJ), - F.Supp.3d -, -, 2017 WL 685563, at *4 (D.D.C.2017). Anthem suggested this is a " 'side issue,' a mere 'rift between CEOs.’ ” Id. That their relationship may have deteriorated has little to do with the anticompetitive effects of the proposed merger.