Djeneba Sidibe v. Sutter Health

FOR PUBLICATION UNITED STATES COURT OF APPEALS FOR THE NINTH CIRCUIT DJENEBA SIDIBE; JERRY No. 22-15634 JANKOWSKI; SUSAN HANSEN; DAVID HERMAN; OPTIMUM D.C. No. GRAPHICS, INC.; JOHNSON POOL 3:12-cv-04854-LB & SPA, on Behalf of Themselves and All Others Similarly Situated, OPINION Plaintiffs-Appellants, v. SUTTER HEALTH, Defendant-Appellee, ______________________________ AETNA HEALTH OF CALIFORNIA, INC.; AETNA LIFE INSURANCE COMPANY; ANTHEM BLUE CROSS; BLUE SHIELD OF CALIFORNIA; UNITED HEALTHCARE SERVICES, INC.; KAISER FOUNDATION HEALTH PLAN INC., Intervenors. 2 SIDIBE V. SUTTER HEALTH Appeal from the United States District Court for the Northern District of California Laurel D. Beeler, Magistrate Judge, Presiding Argued and Submitted August 24, 2023 San Francisco, California Filed June 4, 2024 Before: Patrick J. Bumatay, Lucy H. Koh, and Roopali H. Desai, Circuit Judges. Opinion by Judge Koh; Dissent by Judge Bumatay SUMMARY* California’s Cartwright Act In an action brought by a certified class of individuals and businesses in Northern California who paid health care premiums to certain health plans (Plaintiffs), the panel reversed the district court’s judgment in favor of Sutter Health, the operator of a healthcare system, on Plaintiffs’ claims alleging that Sutter has abused its market power in the region to charge supracompetitive rates to these health plans, which were then passed on to the class in the form of higher premiums. * This summary constitutes no part of the opinion of the court. It has been prepared by court staff for the convenience of the reader.  SIDIBE V. SUTTER HEALTH 3 A jury returned a verdict in favor of Sutter following a trial on Plaintiffs’ claims under California’s Cartwright Act for tying and unreasonable course of conduct. The panel held that the district court contravened California law by removing “purpose” from the Judicial Council of California Civil Jury Instructions and thus failing to instruct the jury to consider Sutter’s anticompetitive purpose as to the unreasonable course of conduct claim, and that the legal error was not harmless. The panel also held that the district court abused its discretion under Fed. R. Evid. 403 in excluding as minimally relevant all evidence of Sutter’s conduct before 2006, which was five years before the specific contracts that Plaintiffs alleged caused them harm were negotiated and took effect. The excluded evidence concerned the inception, Sutter’s stated purpose, and effects of the conduct challenged during the trial. The panel held that Sutter failed to rebut the presumption that the error prejudiced Plaintiffs because, among other things, the excluded evidence would have rebutted Sutter’s testimony and arguments at trial. In a concurrently filed memorandum disposition, the panel affirmed the district court’s refusal to instruct the jury that the health plans were the relevant purchasers when defining the market and its denial of Plaintiffs’ motion for sanctions against Sutter for the destruction of evidence. Dissenting, Judge Bumatay would affirm the jury verdict. He wrote that this court should have left in the hands of the district court the decision as to what was a reasonable cutoff date for evidence; and that the majority crafts a new antitrust rule—justified by neither Supreme Court nor California precedent—under which anticompetitive purpose now becomes an element for every 4 SIDIBE V. SUTTER HEALTH antitrust case, regardless of the individualized circumstances of the case. COUNSEL Matthew L. Cantor (argued) and Jean Kim, Constantine Cannon LLP, New York, New York; James J. Kovacs and J. Wyatt Fore, Constantine Cannon LLP, Washington, D.C.; Azra Z. Mehdi, The Mehdi Firm PC, San Francisco, California; David C. Brownstein and David M. Goldstein, Farmer Brownstein Jaeger Goldstein Klein & Siegel LLP, San Francisco, California; Allan Steyer, D. Scott Macrae, and Suneel Jain, Steyer Lowenthal Boodrookas Alvarez & Smith LLP, San Francisco, California; Jill Manning, Pearson Simon & Warshaw LLP, San Francisco, California; for Plaintiffs-Appellants. Craig E. Stewart (argued), David C. Kiernan, and Matthew J. Silveira, Jones Day, San Francisco, California; Jeffrey A. LeVee, Jones Day, Los Angeles, California; Robert H. Bunzel, Oliver Q. Dunlap, and Patrick M. Ryan, Bartko LLP, San Francisco, California; for Defendant-Appellee. Raymond Wright (argued), Assistant Attorney General, California Department of Justice, Public Rights Division, Civil Rights Enforcement Section, Office of the California Attorney General, Los Angeles, California; for Amicus Curiae. Anna Tsiotsias, Williams & Connolly LLP, Washington, D.C., for Intervenors Aetna Health of California Inc. and Aetna Life Insurance Company.  SIDIBE V. SUTTER HEALTH 5 Michelle L. Cheng and Jarrad L. Wood, Reed Smith LLP, Los Angeles, California, for Intervenor Blue Cross of America dba Anthem Blue Cross. Elspeth V. Hansen and Christopher J. Kelly, Mayer Brown LLP, Palo Alto, California, for Intervenor Blue Shield of California. Maxwell V. Pritt, Boies Schiller Flexner LLP, San Francisco, California, for Intervenor United Healthcare Services Inc.. Mohammad Keshavarzi, Sheppard Mullin Richter & Hampton LLP, Los Angeles, California, for Intervenor Kaiser Foundation Health Plan Inc.. William A. Sokol, Weinber Roger & Rosenfeld, A Professional Corporation, Emeryville, California, for Amicus Curiae California Health Care Coalition. Elizabeth C. Pritzker, Pritzker Levine LLP, Emeryville, California; David A. Balto and Andre Barlow, Law Offices of David Balto, Chevy Chase, Maryland; for Amici Curiae Consumer Action and U.S. Public Interest Research Group. Dean M. Harvey and Brendan P. Glackin, Lieff Cabraser Heimann & Bernstein LLP, San Francisco, California, for Amici Curiae Professors of Law and Economics, Economists, and Health Policy Researchers. Randy M. Stutz, American Antitrust Institute, Washington, D.C.; Kristen G. Marttila and Joseph C. Bourne, Minneapolis, Minnesota; for Amici Curiae The Committee to Support the Antitrust Laws and American Antitrust Institute. Malinda Lee, Justin Lowe, and Raymond Wright, Deputy Attorneys General; Emilio Varanini, Supervising Deputy 6 SIDIBE V. SUTTER HEALTH Attorney General; Renuka R. George, Senior Assistant Attorney General; Rob Bonta, California Attorney General; Office of the California Attorney General, Los Angeles, California; Kwame Roul, Illinois Attorney General, Office of the Illinois Attorney General, Chicago, Illinois; Hector Balderas, New Mexico Attorney General, Office of the New Mexico Attorney General, Santa Fe, New Mexico; Josh Stein, North Carolina Attorney General, Office of the North Carolina Attorney General. Raleigh, North Carolina; Ellen F. Rosenblum, Oregon Attorney General, Office of the Oregon Attorney General, Salem, Oregon; Peter F. Neronha, Rhode Island Attorney General, Office of the Rhode Island Attorney General, Providence, Rhode Island; Karl A. Racine, District of Columbia Attorney General, Office of the District of Columbia Attorney General, Washington, D.C.; for Amici Curiae States of California, Illinois, New Mexico, North Carolina, Oregon, Rhode Island, and The District of Columbia. Andrew R. Dunlap, Lauren J. Zimerman, Anne L. Arcoleo, Selendy Gay Elsber PLLC, New York, New York, for Amici Curiae Scholars of Healthcare Economics. Anne L. Rauch and Trinette S. Sachrison, Berding & Weil LLP, San Diego, California; Daniel Rottinghaus, Berding & Well LLP, Walnut Creek, California; for Amicus Curiae Catalyst for Payment Reform. Jamie Crooks and Rucha Desai, Fairmark Partners LLP, Washington, D.C.; for Amicus Curiae Purchaser Business Group on Health. Boris Bershteyn, Skadden Arps Slate Meagher & Flom LLP, New York, New York, for Amicus Curiae American Hospital Association.  SIDIBE V. SUTTER HEALTH 7 Timothy M. Snyder and Jordan R. Goldberg, Latham & Watkins LLP, Washington, D.C.; Alfred C. Pfeiffer, Jr., Christopher S. Yates, and Sarah M. Ray, Latham & Watkins LLP, San Francisco, California; for Amici Curiae Economists and Antitrust Scholars. OPINION KOH, Circuit Judge: Djeneba Sidibe, Jerry Jankowski, Susan Hansen, David Herman, Optimum Graphics, Inc., and Johnson Pool & Spa (“Plaintiffs”) represent a certified class of individuals and businesses in Northern California who paid health insurance premiums to certain health plans run by Aetna, Anthem Blue Cross, Blue Shield of California, Health Net, and United Healthcare. Plaintiffs allege that Sutter Health (“Sutter”), which operates a healthcare system in Northern California, has abused its market power in the region to charge supracompetitive rates to these health plans, which were then passed on to the class in the form of higher premiums. Following a four week trial on Plaintiffs’ claims under California’s Cartwright Act for tying and unreasonable course of conduct, a jury returned a verdict in favor of Sutter. Plaintiffs now appeal the entry of final judgment in favor of Sutter. They contend that the district court impermissibly excluded relevant evidence, failed to instruct the jury to consider Sutter’s anticompetitive purpose, failed to instruct the jury that the relevant purchasers are the health plans, and wrongly denied Plaintiffs’ motion for sanctions against Sutter for the destruction of evidence. 8 SIDIBE V. SUTTER HEALTH We have jurisdiction under 28 U.S.C. § 1291. The district court erred by failing to instruct the jury to consider Sutter’s anticompetitive purpose and by excluding evidence of Sutter’s conduct before 2006. These errors were prejudicial, so we reverse.1 I. Plaintiffs represent a class of individuals and businesses who are or were insured by health plans that contract with Sutter, a healthcare system that spans 24 hospitals, five medical foundations, and 40 ambulatory surgery centers. Plaintiffs allege that Sutter charged supracompetitive rates to these health plans, which the health plans in turn passed on to Plaintiffs by charging higher premiums. Plaintiffs are therefore “indirect purchasers” of Sutter’s services, and their “theory of antitrust impact depends on two separate overcharges”: an overcharge by Sutter to the health plans, and an overcharge by the health plans to Plaintiffs. Olean Wholesale Grocery Coop., Inc. v. Bumble Bee Foods LLC, 31 F.4th 651, 684 (9th Cir. 2022) (en banc), cert. denied, 143 S. Ct. 424 (2022). In 2012, Plaintiffs filed this class action against Sutter, alleging violations of the Sherman Act, California’s Cartwright Act, and California’s Unfair Competition Law. Specifically, Plaintiffs alleged that Sutter engaged in unlawful tying, in violation of 15 U.S.C. § 1 and Cal. Bus. & Prof. Code § 16720; that Sutter engaged in an unreasonable course of conduct, in violation of the same; that Sutter engaged in monopolization and attempted 1 In a concurrently filed memorandum disposition, we affirm the district court’s refusal to instruct the jury that the health plans were the relevant purchasers when defining the market and its denial of sanctions.  SIDIBE V. SUTTER HEALTH 9 monopolization, in violation of 15 U.S.C. § 2; and that Sutter engaged in unlawful business acts or practices, in violation of Cal. Bus. & Prof. Code § 17200. This litigation has proceeded for over a decade, and so we discuss only a few relevant events from its history. First, in certifying a Rule 23(b)(2) and later a Rule 23(b)(3) class, the district court initially set a damages period beginning on September 28, 2008. This was four years before Plaintiffs sued in September 2012, reflecting the statute of limitations for all three statutes. See 15 U.S.C. § 15b; Cal. Bus. & Prof. Code § 16750.1; Cal. Bus. & Prof. Code § 17208. Second, the district court granted summary judgment to Sutter on Plaintiffs’ monopolization claims (section 2 of the Sherman Act) and for all claims between 2008 and 2010. Thus, the damages period began on January 1, 2011. Third, the district court granted several of Sutter’s motions in limine to exclude evidence. The court excluded: • Evidence relating to other litigation and investigations, including (1) a 1999 challenge to the merger of Sutter’s Alta Bates Medical Center in Berkeley and the Summit Medical Center in Oakland, and (2) two state court actions alleging similar anticompetitive conduct as this federal suit brought by UFCW & Employers Benefit Trust and the State of California (these latter two actions were later consolidated, and the parties settled in October 2019). • Evidence relating to Sutter’s practices before January 1, 2006 (that is, five years 10 SIDIBE V. SUTTER HEALTH before the damages period). However, the court left open the possibility that Plaintiffs might make an offer of proof as to a specific exhibit. • A 2006 memo by Strategy Advantage, which included statements about Sutter’s marketing position but “necessarily look[ed] back at” pre-2006 evidence. All three of these rulings were premised on Federal Rule of Evidence 403, which permits courts to exclude evidence that is nonetheless relevant if “its probative value is substantially outweighed by a danger of . . . unfair prejudice, confusing the issues, misleading the jury, undue delay, wasting time, or needlessly presenting cumulative evidence.” Fed. R. Evid. 403.2 Plaintiffs later made an offer of proof as to 23 pre-2006 documents they wished to introduce at trial. The district court denied the offer of proof, again citing Rule 403. The case proceeded to trial on Plaintiffs’ claims under the Cartwright Act only. The parties stipulated that because Plaintiffs sought damages only under the Cartwright Act, a jury would decide that issue. The parties also stipulated that, after the jury trial, the district court would decide (1) whether to award injunctive relief under the Cartwright Act, Sherman Act, and California’s Unfair Competition 2 Plaintiffs suggest that the district court relied on Rule 402 (the exclusion of irrelevant evidence) because the court stated “I don’t think it’s relevant” during a colloquy at trial. However, the district court’s written order ruling on the motions in limine discussed Rule 403, and the court orally clarified that it excluded the evidence because “their relevance was vastly outweighed by the danger, the sideshow, cumulative[,] confusing”—all hallmarks of a Rule 403 analysis.  SIDIBE V. SUTTER HEALTH 11 Law; and (2) whether Plaintiffs were entitled to restitution under the Unfair Competition Law. These determinations would likely concern the same evidence presented to the jury. Among other disputes over the jury instructions, the parties proposed differing instructions on how to prove an unreasonable course of conduct claim under the Cartwright Act. Both parties relied on sections 3405 and 3411 of the Judicial Council of California Civil Jury Instructions (“CACI”), which require a plaintiff to prove (among other things) that “the purpose or effect of [name of defendant]’s conduct was to restrain competition” and instruct the jury to weigh the “anticompetitive or beneficial purpose or effect” of a challenged restraint, respectively. Sutter, however, proposed instructions removing the word “purpose,” which it argued was necessary to conform with longstanding precedent holding that anticompetitive purpose alone does not offend the antitrust laws. Plaintiffs proposed leaving the word “purpose” in both instructions, citing the California Supreme Court’s decision in Corwin v. L.A. Newspaper Serv. Bureau, Inc., 583 P.2d 777 (Cal. 1978). The district court agreed with Sutter and excluded the word “purpose” from the proposed instructions. In fact, the district court was so thorough in erasing any mention of “purpose” from the jury instructions that the court replaced “[t]he reasonableness of the stated purpose for the restraint” with “[t]he reasonableness of the restraint” in CACI 3411, even though both Plaintiffs and Sutter had proposed keeping that language in that instruction. Plaintiffs objected to the removal of “purpose” from both instructions. The court overruled the objection without explanation. The jury was ultimately instructed that, to prove a claim for unreasonable course of conduct, Plaintiffs had to prove that “the effect of 12 SIDIBE V. SUTTER HEALTH Sutter’s conduct was to restrain competition,” and that the jury should consider whether “Sutter’s challenged restraint has an anticompetitive or beneficial effect on competition.” After a four week trial, the jury returned a verdict for Sutter on both the tying and unreasonable course of conduct claims. Plaintiffs then stipulated to entry of final judgment on their claims under section 1 of the Sherman Act and California’s Unfair Competition Law. Plaintiffs timely appealed. II. We first address Plaintiffs’ contention that the district court contravened California law when it omitted the word “purpose” from the jury instructions on Plaintiffs’ unreasonable course of conduct claim and that the legal error was not harmless. We agree. A. The Judicial Council of California is “the rule-making arm of the California court system.” NASD Disp. Resol., Inc. v. Jud. Council of State of Cal., 488 F.3d 1065, 1067 (9th Cir. 2007). The Council consists of the Chief Justice and one Associate Justice of the California Supreme Court, as well as three judges from the California courts of appeal, ten judges from the California superior courts, and various nonvoting members. CAL. CONST. ART. VI, § 6(a). The Council is empowered by the California Constitution to “adopt rules for court administration, practice and procedure, and perform other functions prescribed by statute,” so long as those rules are not “inconsistent with statute.” Id. § 6(d); In re Abbigail A., 375 P.3d 879, 883–84 (Cal. 2016); Cunningham v. California, 549 U.S. 270, 278 n.4 (2007).  SIDIBE V. SUTTER HEALTH 13 Among the California Rules of Court adopted by the Council is Rule 2.1050, the Judicial Council jury instructions. These model jury instructions “are the official instructions for use in the state of California” and are intended to “accurately state the law in a way that is understandable to the average juror.” Cal. R. Ct. 2.1050(a). Thus, use of the model jury instructions “is strongly encouraged,” unless the trial court judge “finds that a different instruction would more accurately state the law and be understood by jurors.” Cal. R. Ct. 2.1050(f). Accordingly, even though the “articulation and interpretation of California law [] remains within the purview of the Legislature and the courts of review,” Cal. R. Ct. 2.1050(b), the Ninth Circuit has generally treated the model jury instructions as a helpful, albeit not dispositive, interpretive tool. See, e.g., United States v. Ruiz-Apolonio, 657 F.3d 907, 913 & n.2 (9th Cir. 2011) (applying model criminal jury instructions to help interpret provision of California Penal Code); Hardeman v. Monsanto Co., 997 F.3d 941, 968 (9th Cir. 2021) (concluding that the district court’s jury instruction “was inconsistent with . . . CACI and California case law” (cleaned up)); Killgore v. SpecPro Pro. Servs., LLC, 51 F.4th 973, 984 (9th Cir. 2022) (relying on CACI as one of “[s]everal persuasive California sources” to support the court’s reading of the statute). B. This appeal concerns CACI 3405, which states that a plaintiff may prove the second element of an unreasonable course of conduct claim by proving either anticompetitive purpose or effect, and CACI 3411, which lists factors for a jury to consider in weighing the anticompetitive purposes or 14 SIDIBE V. SUTTER HEALTH effects of a defendant’s conduct. In its entirety, CACI 3405 reads: CACI 3405: HORIZONTAL AND VERTICAL RESTRAINTS (USE FOR DIRECT COMPETITORS OR SUPPLIER/RESELLER RELATIONS)—OTHER UNREASONABLE RESTRAINT OF TRADE—RULE OF REASON— ESSENTIAL FACTUAL ELEMENTS [Name of plaintiff] claims that [name of defendant] agreed to [insert unreasonable restraint of trade]. To establish this claim, [name of plaintiff] must prove all of the following: 1. That [name of defendant] [and [name of alleged coparticipants[s]]] agreed to [describe conduct constituting an unreasonable restraint of trade]; 2. That the purpose or effect of [name of defendant]’s conduct was to restrain competition; 3. That the anticompetitive effect of the restraint[s] outweighed any beneficial effect on competition; 4. That [name of plaintiff] was harmed; and 5. That [name of defendant]’s conduct was a substantial factor in causing [name of plaintiff]’s harm. Judicial Council of California Civil Jury Instruction 3405 (July 2023 Update). Although the second element may be proven by virtue of anticompetitive purpose or effect, the  SIDIBE V. SUTTER HEALTH 15 third element still requires a plaintiff to prove that anticompetitive effects outweigh any procompetitive effects. As a result, under CACI 3405, proof of anticompetitive purpose does not eliminate the need to provide proof of anticompetitive effect. CACI 3411, meanwhile, reads as follows: CACI 3411: RULE OF REASON — ANTICOMPETITIVE VERSUS BENEFICIAL EFFECTS In deciding whether [name of defendant]’s challenged restraint had an anticompetitive or beneficial purpose or effect on competition, you should consider the results the restraint was intended to achieve or actually did achieve. In balancing these purposes or effects, you also may consider, among other factors, the following: (a) The nature of the restraint; (b) The probable effect of the restraint on the business involved; (c) The history of the restraint; (d) The reasonableness of the stated purpose for the restraint; (e) The availability of less restrictive means to accomplish the stated purpose; (f) The portion of the market affected by the restraint; [and] (g) The extent of [name of defendant]’s market power; [and] 16 SIDIBE V. SUTTER HEALTH (h) [Insert other relevant consideration]. Judicial Council of California Civil Jury Instruction 3411 (July 2023 Update). Here, too, even though the introductory paragraph uses the phrase “purpose or effect,” the second factor emphasizes the relevance of the “probable effect[s] of the restraint,” and the sixth factor emphasizes the relevance of the “portion of the market affected by the restraint.” As we have discussed, the district court did not follow CACI. The district court adopted Sutter’s proposal to remove the word “purpose” from the second element of CACI 3405 and the introductory paragraph of CACI 3411, and it sua sponte changed “the reasonableness of the stated purpose for the restraint” in CACI 3411 to “the reasonableness of the restraint.” Plaintiffs objected, and the district court overruled the objection without explanation. The district court’s revised CACI 3405, which we call Instruction 3405, read as follows: UNREASONABLE COURSE-OF- CONDUCT CLAIM — RULE OF REASON — ESSENTIAL FACTUAL ELEMENTS In addition to their tying claim, the plaintiffs claim that Sutter entered contracts with insurance companies that unreasonably restrain competition for inpatient hospital services in the tied markets. The plaintiffs claim that the contracts contained terms that prevented the insurance companies from creating effective narrow network products or tiered products that would have allowed the insurance companies to steer patients to  SIDIBE V. SUTTER HEALTH 17 lower-cost non-Sutter hospitals within the health-plan network. To establish this claim, the plaintiffs must prove all of the following: 1. That Sutter and insurance companies entered into agreements that contain terms that prevented the insurance companies from steering patients to lower-cost non-Sutter hospitals within the health-plan network; 2. That the effect of Sutter’s conduct was to restrain competition; 3. That the anticompetitive effect of the restraint outweighed any beneficial effect of the restraint on competition; 4. That the plaintiffs were harmed; and 5. That Sutter’s conduct was a substantial factor in causing the plaintiffs’ harm. Thus, unlike CACI 3405, Instruction 3405 did not instruct the jury that Plaintiffs could prove the second element of their unreasonable course of conduct by proving anticompetitive purpose or effect. Instead, the jury was instructed that the second element required proof solely of anticompetitive effect. The district court’s revised CACI 3411, which we call Instruction 3411, read: UNREASONABLE COURSE-OF- CONDUCT CLAIM — RULE OF REASON — ANTICOMPETITIVE VERSUS BENEFICIAL EFFECTS In deciding whether Sutter’s challenged restraint has an anticompetitive or beneficial 18 SIDIBE V. SUTTER HEALTH effect on competition, you should consider the results that the restraint was intended to achieve or actually did achieve. In balancing these effects, you also may consider, among other factors, the following: (a) The nature of the restraint; (b) The probable effect of the restraint on the business involved; (c) The history of the restraint; (d) The reasonableness of the restraint; (e) The availability of less restrictive means to accomplish the stated reason for the restraint; (f) The portion of the market affected by the restraint; and (g) The extent of Sutter’s market power. The Instruction 3411 given to the jury, therefore, omitted “purpose” from the introductory paragraph (twice) and from the fourth factor, replacing “reasonableness of the stated purpose for the restraint” with “reasonableness of the restraint.” C. “In reviewing jury instructions, we do not employ a line- by-line examination. Instead, we use a practical approach, focusing on whether in the light of the issues and viewed as a whole, the instructions were complete, clear, correct, and adequate.” Ridgeway v. Walmart Inc., 946 F.3d 1066, 1081 (9th Cir. 2020) (internal quotation marks omitted). Although “we review for abuse of discretion a district court’s formulation of jury instructions,” we owe no deference in determining whether those instructions “accurately state the law.” Coston v. Nangalama, 13 F.4th 729, 732 (9th Cir.  SIDIBE V. SUTTER HEALTH 19 2021) (quoting Hung Lam v. City of San Jose, 869 F.3d 1077, 1085 (9th Cir. 2017)). If the jury was incorrectly instructed, then we will affirm only if the prevailing party below can show harmless error: that “it is more probable than not that the jury would have reached the same verdict had it been properly instructed.” Fierro v. Smith, 39 F.4th 640, 651 (9th Cir. 2022) (internal quotation marks omitted); see also BladeRoom Grp. Ltd. v. Emerson Elec. Co., 20 F.4th 1231, 1243 (9th Cir. 2021) (reviewing court will “presume prejudice” from the erroneous instruction, and the prevailing party has the burden to demonstrate otherwise). CACI’s use of the phrase “purpose or effect” stems directly from the text of the Cartwright Act, which outlaws “every trust,” Cal. Bus. & Prof. Code § 16726, defined as “a combination of capital, skill or acts by two or more persons for any of the following purposes: . . . .” Cal. Bus. & Prof. Code § 16720(a) (emphasis added). The Act likewise explains that an agreement or combination is not unlawful if its “purpose and effect” is “to promote, encourage or increase competition.” Cal. Bus. & Prof. Code § 16725 (emphasis added). Accordingly, the California Supreme Court has long defined the rule of reason analysis as whether a “contract, combination, or conspiracy . . . has as its Purpose or Effect an unreasonable restraint of trade.” Corwin, 583 P.2d at 784; see also Ixchel Pharma, LLC v. Biogen, Inc., 470 P.3d 571, 581 (Cal. 2020) (relevant factors include “the nature of the restraint and its effects, and the history of the restraint and the reasons for its adoption” (quoting In re Cipro Cases I & II, 348 P.3d 845, 861 (Cal. 2015))). The district court’s exclusion of “purpose” from both instructions was therefore at odds with both the text of the Cartwright Act and the California Supreme Court’s longstanding interpretation thereof. 20 SIDIBE V. SUTTER HEALTH This appeal solely concerns the Cartwright Act, which the California Supreme Court “no longer” treats as “coextensive with the Sherman Act.” Samsung Elec. Co., Ltd. v. Panasonic Corp., 747 F.3d 1199, 1205 n.4 (9th Cir. 2014). Nonetheless, we observe that federal law agrees with California law that anticompetitive purpose is a relevant factor, something Sutter also concedes. Because the Cartwright Act and Sherman Act both “carry forward the common law understanding that only unreasonable restraints of trade are prohibited,” Cipro, 348 P.3d at 861 (internal quotation marks omitted), decisions by the U.S. Supreme Court likewise emphasize the need to consider “the purpose or end sought to be attained” by a challenged restraint. Bd. of Trade of Chi. v. United States, 246 U.S. 231, 238 (1918); see also United States v. Topco Assocs., Inc., 405 U.S. 596, 607 (1972) (also listing as relevant factors “the history of the restraint and the reasons for its adoption”).3 It is therefore no surprise that federal model jury instructions, much like CACI, list “the history of the restraint” and “the reasons for adopting the particular practice that is alleged to be a restraint” as relevant factors in determining whether a course of conduct is reasonable or unreasonable. 3A Kevin F. 3 We do not suggest that every Supreme Court decision citing the rule of reason has directly invoked anticompetitive purpose, particularly where the rule’s specific factors were not the subject of the court’s analysis. See generally, e.g., Leegin Creative Leather Prods., Inc. v. PSKS, Inc., 551 U.S. 877 (2007) (holding that the rule of reason rather than per se illegality applies to vertical price restraints). Even in these cases, however, anticompetitive purpose remains an important factor: Leegin’s cited authorities for the factors to consider, 551 U.S. at 885, ultimately trace back to “the classic statement of the rule of reason in Chicago Bd. of Trade,” including consideration of “the purpose or end sought to be attained.” Arizona v. Maricopa Cnty. Med. Soc’y, 457 U.S. 332, 343 n.13 (1982) (internal quotation marks omitted).  SIDIBE V. SUTTER HEALTH 21 O’Malley, Jay E. Grenig, & Hon. William C. Lee, FED. JURY PRAC. & INSTR. § 150:21 (6th ed. Feb. 2024 update); see also ABA SECTION OF ANTITRUST LAW, MODEL JURY INSTRUCTIONS IN CIVIL ANTITRUST CASES, Ch. 1C, Instruction 3B (2016) (proof of harmful effect on competition may be shown by “the purpose and nature of the restraint”). Again, this case solely concerns CACI’s description of the Cartwright Act. It is a “cardinal principle of judicial restraint” that “if it is not necessary to decide more, it is necessary not to decide more,” and we express no opinion on other jurisdictions’ descriptions of the rule of reason. Ratha v. Phatthana Seafood Co., 35 F.4th 1159, 1168 (9th Cir. 2022) (quoting Midbrook Flowerbulbs Holland B.V. v. Holland Am. Bulb Farms, Inc., 874 F.3d 604, 617 n.13 (9th Cir. 2017)). We list these examples from Sherman Act decisions and federal jury instructions solely to demonstrate that our dissenting colleague is mistaken: our decision today announces no new legal rule but rather reflects the widespread consensus that consideration of anticompetitive purpose is an essential aspect of the rule of reason analysis under both the Cartwright Act and the Sherman Act. Accordingly, failing to instruct the jury to consider purpose misstates the law. To be sure, decisions interpreting the antitrust laws state that anticompetitive purpose is but one factor that a trier of fact may consider, not that it is required to do so. See, e.g., Cipro, 348 P.3d at 861 (stating that a court “may consider” factors including purpose). Our dissenting colleague latches on to these decisions to argue that proof of anticompetitive purpose is neither necessary nor sufficient to prove an antitrust claim under the rule of reason. This argument, however, fundamentally mischaracterizes the district court’s 22 SIDIBE V. SUTTER HEALTH error. By its very nature the rule of reason analysis is a flexible one that “requires courts to conduct a fact-specific assessment of market power and market structure.” Ohio v. Am. Express Co., 585 U.S. 529, 541 (2018) (internal quotation marks omitted). The ultimate goal is to determine the “restraint’s actual effect on competition.” Id. (cleaned up). As a means of determining anticompetitive effect, however, anticompetitive purpose is one of several relevant factors that a trier of fact may consider. The trier of fact is not required to rely on any one factor, but it must have the option of considering that factor, which is only possible if properly instructed that the factor exists. Here, the jury was not instructed that it could consider anticompetitive purpose. This was error. The district court’s change to the fourth factor listed in Instruction 3411 highlights the importance of instructing a jury that it may consider evidence of anticompetitive purpose. In CACI 3411, that factor reads, “[t]he reasonableness of the stated purpose for the restraint.” This factor appropriately instructs a jury that a defendant’s intent is relevant under the rule of reason. The district court, however, omitted this factor’s reference to Sutter’s intent, instead simply instructing the jury to consider “[t]he reasonableness of the restraint” itself. When combined with the district court’s other omissions in Instructions 3405 and 3411, this change failed to instruct the jury that it could consider evidence of Sutter’s anticompetitive purpose. Far from proving that the district court’s error was harmless, as the dissent suggests, Instruction 3411 makes clear that the district court’s instructions misstated the law. Because Sutter concedes that anticompetitive purpose is a relevant consideration under the Cartwright Act, Sutter’s only defense of the district court’s omissions is that no court  SIDIBE V. SUTTER HEALTH 23 has ever held that anticompetitive purpose alone can prove a claim of an unreasonable course of conduct. This is true. See, e.g., L.A. Mem’l Coliseum Comm’n v. Nat’l Football League, 726 F.2d 1381, 1395 (9th Cir. 1984) (“[A]nticompetitive purpose alone is not enough to condemn” challenged conduct (citing Bd. of Trade of Chi., 246 U.S. at 238)); Exxon Corp. v. Superior Ct., 60 Cal. Rptr. 2d 195, 200 (Cal. Ct. App. 1997) (emphasizing that a plaintiff “must prove that the restraint had an anticompetitive effect”). Were Sutter correct that CACI 3405 contained an inaccurate statement of law, then courts would be under no obligation to follow it. See Cal. R. Ct. 2.1050(b) (interpretation of California law is the purview of the legislature and courts, not the Judicial Council). However, Sutter misreads CACI 3405 by failing to consider the jury instructions “as a whole.” Ridgeway, 946 F.3d at 1081. CACI 3405 states that there are five elements of an unreasonable course of conduct claim and that the second element may be proven by anticompetitive “purpose or effect.” As we have already mentioned, however, the very next element—the third—requires a plaintiff to prove that “the anticompetitive effect of the restraint[s] outweighed any beneficial effect on competition.” CACI 3411, too, expressly instructs juries to consider “[t]he probable effect of the restraint” and “[t]he portion of the market affected by the restraint.” Sutter simply ignores these elements of the instructions, which make clear that a plaintiff cannot prevail without proof of anticompetitive effect. The problem is that the jury was not instructed that it could even consider anticompetitive purpose, as all parties agree California law requires. Thus, Sutter is incorrect that, had the district court followed CACI, the court would have misstated the law. To the contrary, not following CACI misstated the law. 24 SIDIBE V. SUTTER HEALTH When we view Instructions 3405 and 3411 in the context of the jury instructions as a whole—as Ridgeway requires— it is evident that the omission of “purpose” from both instructions contravened California law. We owe no deference to the district court’s decision. Coston, 13 F.4th at 732. As a result, the instructions cannot stand. D. Because legal error is established, we “presume prejudice” and Sutter has the burden to demonstrate that it is “more probable than not that the jury would have reached the same verdict had it been properly instructed.” BladeRoom Grp., 20 F.4th at 1243 (internal quotation marks omitted). Sutter has not met this burden. Sutter argues only that any error was harmless because the jury answered “no” to the fifth question on the jury form, corresponding to the first element of an unreasonable course of conduct claim, and so did not reach the elements affected by the erroneous instructions (questions 6 and 7). This argument lacks merit. Question 5 asked: “Did Sutter force the class health plans to agree to contracts that had terms that prevented the plans from steering plaintiffs to lower-cost non-Sutter hospitals within the plan network?” This question did not simply ask whether there was a restraint of trade but rather had two parts: (1) whether Sutter forced the health plans to accept the challenged contract terms, and (2) whether those terms prevented steering (i.e., were anticompetitive). Because this question asked the jury to consider whether Sutter “forced” the health plans to adopt the challenged contract terms, consideration of Sutter’s alleged anticompetitive purpose could have led the jury to answer Question 5 differently had it been properly instructed.  SIDIBE V. SUTTER HEALTH 25 Consideration of a party’s motives often shapes interpretations of that party’s actions, particularly under the rule of reason’s “fact-specific” analysis. Am. Express Co., 585 U.S. at 541. As Justice Brandeis famously articulated: “The history of the restraint, the evil believed to exist, the reason for adopting the particular remedy, the purpose or end sought to be attained, are all relevant facts. This is not because a good intention will save an otherwise objectionable regulation or the reverse; but because knowledge of intent may help the court to interpret facts and to predict consequences.” Bd. of Trade of Chi., 246 U.S. at 238. Sutter’s motives for adopting the challenged contract terms are therefore relevant to whether Sutter “forced” health plans to agree to terms that prevented the health plans from steering patients to lower-cost, non-Sutter hospitals. On harmless error review, Sutter has the burden to prove that it is “more probable than not” that the jury would have reached the same result if properly instructed. Fierro, 39 F.4th at 651 (internal quotation marks omitted). Sutter cannot do so.4 Finally, even though the erroneous instructions alone are sufficiently prejudicial to warrant a new trial, we also cannot view this error in isolation. The Ninth Circuit and many of our sister circuits have consistently held that errors in a civil 4 As mentioned, because the district court omitted “purpose” from Instruction 3411 as well as Instruction 3405, Instruction 3411 compounded rather than cured the prejudice caused to the Plaintiffs, notwithstanding Instruction 3411’s solitary reference to “the results that the restraint was intended to achieve or actually did achieve.” Even if this question were a close call, Sutter has failed to rebut the presumption of prejudice. See Obrey v. Johnson, 400 F.3d 691, 701 (9th Cir. 2005) (on harmless error review, in both civil and criminal cases, courts reverse even “in cases of equipoise” (cleaned up)). 26 SIDIBE V. SUTTER HEALTH trial must be considered “cumulatively” and that the combined effect of multiple errors “may suffice to warrant a new trial even if each error standing alone may not be prejudicial.” Jerden v. Amstutz, 430 F.3d 1231, 1240–41 (9th Cir. 2005). Thus, even if the jury understood that it could consider evidence of Sutter’s anticompetitive purpose, then it was all the more important that Plaintiffs be able to introduce evidence of such purpose. Here, however, Plaintiffs also contend that the district court improperly excluded evidence from before 2006, much of which was offered to show that Sutter’s conduct was motivated by anticompetitive purpose. That is, Plaintiffs argue that the district court not only failed to instruct the jury to consider Sutter’s anticompetitive purpose, as required by CACI and California law, but it also prevented the jury from hearing evidence that was highly probative of Sutter’s purpose in the first place. It is to this improper exclusion of evidence that we now turn. III. Plaintiffs’ second contention is that the district court abused its discretion in excluding pre-2006 evidence and that the error was prejudicial. Again, we agree. A. The probative value of the excluded evidence can only be understood in the context of Plaintiffs’ theory of antitrust injury, so before addressing the merits we explain the basics of Plaintiffs’ theory of the case. Like most healthcare providers, Sutter’s providers (hospitals, surgery centers, etc.) contract with health plans to be included in those plans’ “networks,” because insurers encourage patients to use in-network providers. When an  SIDIBE V. SUTTER HEALTH 27 insured individual or business uses an in-network provider, the insured generally pays less. In exchange for accepting this discounted rate for services, an in-network provider receives the benefit of higher patient volume. Prior to the late 1990s and early 2000s, each Sutter provider negotiated its own contracts with health plans. Consequently, which party had the upper hand in these negotiations—Sutter or the health plans—depended on local market conditions. If there were many providers in a market, then providers competed amongst themselves to offer lower prices to the health plans in order to be included in those plans’ networks or, if already included, to be treated as a “preferred” provider. If there were few providers in a market, or even only one provider, however, then that provider had market power and could charge higher prices to health plans. Because Sutter is a large healthcare system in Northern California, there are several markets, or geographic regions, in which there are few or no non-Sutter providers for inpatient hospital services. Plaintiffs allege that, accordingly, Sutter’s providers had market power in these regions and could charge higher prices. In other regions, however, Sutter’s providers competed with other providers, lacked market power, and were forced to charge lower prices. Plaintiffs’ primary allegation is that Sutter sought to use its market power in these uncompetitive regions to charge higher prices in other regions as well. To do so, around the turn of the millennium Sutter began contracting with health plans on a “systemwide” basis, meaning that one contract governs the relationship and imposes common terms between a health plan and all Sutter providers. Plaintiffs 28 SIDIBE V. SUTTER HEALTH allege that in 1997 and 1998, Sutter determined that systemwide contracting would give Sutter more leverage in negotiations with health plans and could drastically increase Sutter’s profits by nearly $200 million per year. This determination, Plaintiffs contend, is essential context to explain Sutter’s implementation of systemwide contracting. Plaintiffs further allege that Sutter effectively coerced the health plans into contracting on a systemwide basis by terminating all existing contracts and drafting a “model systemwide amendment” that Sutter would impose on all the health plans with which Sutter contracted. Plaintiffs allege that the health plans vigorously opposed the switch to systemwide contracting, indicating that Sutter had the market power to impose systemwide contracting over the health plans’ objections. Although systemwide contracting is itself lawful, Plaintiffs allege that systemwide contracting became the vehicle by which Sutter imposed anticompetitive contract terms on the health plans and charged supracompetitive prices. Sutter was able to accomplish this goal via a practice antitrust law calls “tying,” or conditioning the purchase of one product on the purchase of another. Eastman Kodak Co. v. Image Tech. Servs., Inc., 504 U.S. 451, 461 (1992). Sutter lacked market power in regions with many non-Sutter providers. Because the health plans wished to contract with Sutter in the regions with few or no non-Sutter providers, however, Sutter allegedly conditioned the product the health plans wanted (in-network participation of Sutter’s providers in uncompetitive or “tying” markets) on the health plans’ willingness to purchase a product they did not want or on terms they did not want (either in-network participation of or supracompetitive out-of-network rates for Sutter’s  SIDIBE V. SUTTER HEALTH 29 providers in more competitive or “tied” markets). See id. at 461–62. Plaintiffs further allege that Sutter employed three allegedly anticompetitive contract terms in its systemwide contracts with the health plans: “non-participating provider rates” (or “non-par rates”) clauses, “equal treatment” clauses, and “tiered products” clauses. Non-par rates are the rates health plans must pay to Sutter if a patient seeks care at an out-of-network Sutter hospital, as a percentage of the billed charge. In Sutter’s contracts, these non-par rates were often as high as 95% or 100% of the billed charges, whereas the in-network rates could be as low as 55%. Equal treatment clauses require health plans to treat Sutter’s providers the same as any other provider in the same benefit program or network. Accordingly, even if a Sutter hospital is much more expensive than a competing in- network hospital, health plans cannot implement financial incentives for patients to seek treatment at the competing hospital instead of the Sutter hospital. Tiered products clauses specify Sutter’s status in a health plan (in-network, out-of-network, preferred) in the systemwide contract. These clauses prevent health plans from changing Sutter’s providers’ tiers without first providing notice to Sutter so that the parties can negotiate new terms, thus preventing health plans from placing Sutter’s providers in a “non-preferred” tier that imposes higher copayments than at lower-priced, non-Sutter alternatives. All three of these provisions allegedly insulate Sutter from price competition by preventing health plans from 30 SIDIBE V. SUTTER HEALTH “steering” patients away from Sutter and toward lower- priced providers. High non-par rates make it impractical to exclude Sutter’s providers from a network, regardless of how many non-Sutter providers there are, because some patients will invariably seek emergency care at Sutter’s hospitals, at great expense to the health plan. Meanwhile, equal treatment clauses and tiered products clauses both prevent health plans from offering financial incentives to use non-Sutter providers. Plaintiffs allege that Sutter conceived of these contract terms as a means to charge higher prices and that Sutter began including these terms in its systemwide contracts between 2001 and 2005. Plaintiffs add that the health plans objected to these contract terms, too, further indicating that Sutter had the market power to impose the terms unilaterally. Finally, Plaintiffs allege that the results speak for themselves: after adopting systemwide contracts and imposing the challenged contract terms, Sutter began charging much higher rates to health plans, as much as 40 to 50 percent higher. By 2002, Sutter’s prices had “skyrocketed” relative to other Northern California providers. To sum up, Plaintiffs’ theory of the case was not simply that Sutter’s systemwide contracts contained anticompetitive terms. Rather, Plaintiffs contended that Sutter (1) had previously negotiated individual contracts; (2) then implemented systemwide contracting with the purpose of imposing anticompetitive terms to charge supracompetitive prices; and (3) ultimately imposed those terms over the health plans’ objections, resulting in exactly the supracompetitive pricing Sutter had predicted—and that all of this was implemented prior to January 1, 2006, the cutoff  SIDIBE V. SUTTER HEALTH 31 date that the district court imposed for the introduction of evidence. B. A district court’s evidentiary rulings are reviewed for an abuse of discretion and will not be reversed “unless the ruling is manifestly erroneous.” Gen. Elec. Co. v. Joiner, 522 U.S. 136, 141–42 (1997) (internal quotation marks omitted). Our review of the district court’s Rule 403 determination is therefore a delicate one. On the one hand, such determinations are “subject to great deference” because “the considerations arising under Rule 403 are susceptible only to case-by-case determinations” and require “examination of the surrounding facts, circumstances, and issues.” United States v. Cabrera, 83 F.4th 729, 736 (9th Cir. 2023) (quoting United States v. Hinkson, 585 F.3d 1247, 1267 (9th Cir. 2009) (en banc)), petition for cert. filed, No. 23-6976 (U.S. Mar. 5, 2024); see also Sprint/United Mgmt. Co. v. Mendelsohn, 552 U.S. 379, 384 (2008) (Rule 403 requires “on-the-spot balancing” (internal quotation marks omitted)). On the other hand, this deferential language is simply another way of saying that our review is for abuse of discretion rather than de novo. See id.; see also Henderson v. George Wash. Univ., 449 F.3d 127, 133, 141 (D.C. Cir. 2006) (rejecting view that trial judge retains “unfettered discretion in the application of Rule 403” and holding that exclusion of evidence was an abuse of discretion). Rule 403 itself, meanwhile, sets a high bar for exclusion: a court may exclude relevant evidence only if its probative value is substantially outweighed by one or more of the articulated dangers or considerations. As we have observed, “the application of Rule 403 must be cautious and sparing,” 32 SIDIBE V. SUTTER HEALTH because its “major function is limited to excluding matter of scant or cumulative probative force, dragged in by the heels for the sake of its prejudicial effect.” United States v. Hankey, 203 F.3d 1160, 1172 (9th Cir. 2000) (quoting United States v. Mills, 704 F.2d 1553, 1560 (11th Cir. 1983)); see also Henderson, 449 F.3d at 133 (“Rule 403 tilts, as do the rules as a whole, toward the admission of evidence in close cases.” (internal quotation marks omitted)). As a result, although decisions reversing Rule 403 exclusions are “relatively rare,” they are not “unknown.” United States v. Johnson, 89 F.4th 997, 1002–03, 1003 n.1 (7th Cir. 2024) (reversing and collecting similar cases from 12 of the 13 federal courts of appeals, including our circuit). Sutter argues that the requirement that Rule 403 be used “sparingly” is limited to the exclusion of evidence offered by the defendant in a criminal case. Nothing in the text of Rule 403, however, suggests that its standard differs from civil to criminal cases, and Sutter does not point to any decision by the Supreme Court or a federal court of appeals saying so. Indeed, our own brief investigation found decisions by the Second, Third, Fourth, Fifth, Seventh, Eighth, Tenth, Eleventh, and D.C. Circuits stating, in civil cases, that Rule 403 exclusions of evidence should be “spare.”5 5 See, e.g., George v. Celotex Corp., 914 F.2d 26, 30–31 (2d Cir. 1990); Blancha v. Raymark Indus., 972 F.2d 507, 516 (3d Cir. 1992); PBM Prods., LLC v. Mead Johnson & Co., 639 F.3d 111, 124–25 (4th Cir. 2011); Brady v. Fort Bend County, 145 F.3d 691, 715 (5th Cir. 1998); Cook v. Hoppin, 783 F.2d 684, 689 (7th Cir. 1986); Westcott v. Crinklaw, 68 F.3d 1073, 1077–78 (8th Cir. 1995); Eisenhour v. Weber County, 897 F.3d 1272, 1277 (10th Cir. 2018); Luxottica Grp., S.p.A. v. Airport Mini Mall, LLC, 932 F.3d 1303, 1318 (11th Cir. 2019); Joy v. Bell Helicopter Textron, Inc., 999 F.2d 549, 555 (D.C. Cir. 1993).  SIDIBE V. SUTTER HEALTH 33 Even if the district court abused its discretion by excluding evidence, its ruling will be reversed only if the error was prejudicial. However, “[w]hen error is established, we must presume prejudice unless it is more probable than not that the error did not materially affect the verdict.” Barranco v. 3D Sys. Corp., 952 F.3d 1122, 1127 (9th Cir. 2020) (quoting Boyd v. City & County of San Francisco, 576 F.3d 938, 949 (9th Cir. 2009)). Thus, as with their challenge to the jury instructions, Plaintiffs have the burden to show error, but if they succeed, the burden shifts to Sutter to demonstrate that the error was harmless. C. The district court’s justification for excluding evidence under Rule 403 was informed by its 2006 cutoff date for relevant evidence. Before discussing the excluded evidence in detail, therefore, we briefly discuss the choice of cutoff date. A trial court may “set a reasonable cut-off date, evidence before which point is to be considered too remote to have sufficient probative value.” Cont’l Ore Co. v. Union Carbide & Carbon Corp., 370 U.S. 690, 710 (1962). The district court here set the cutoff date at 2006 (five years before the start of the damages period in 2011), a date that the court acknowledged “is arbitrary.” The district court concluded that evidence predating 2006 was of “minimal relevance” because Sutter’s systemwide contracts were renegotiated regularly and the specific contracts that Plaintiffs alleged had caused them harm during the damages period were negotiated and took effect shortly before 2011. The district court’s reasoning conflated Plaintiffs’ theory of liability with any damages to which Plaintiffs would be entitled if they prevailed at trial. True, the specific contracts giving rise to Plaintiffs’ damages were negotiated shortly 34 SIDIBE V. SUTTER HEALTH before 2011. As with the jury instructions, however, the district court did not appreciate how the history and purpose of Sutter’s conduct was an essential aspect of Plaintiffs’ legal theory, not merely as context from before 2006 but as evidence of what Sutter did (and the effects thereof) during the class period. Indeed, pre-2006 evidence is highly relevant to both Plaintiffs’ tying claim and their unreasonable course of conduct claim. To prevail on their tying claim, Plaintiffs had to prove (1) the existence of a tying arrangement (that the “sale of the tying product” was “linked to the sale of the tied product”); (2) that Sutter had “sufficient economic power in the tying market to coerce the purchase of the tied product”; (3) that a substantial number of sales “was effected in the tied product”; and (4) that Plaintiffs “sustained pecuniary loss” as a result of the unlawful conduct. UAS Mgmt., Inc. v. Mater Misericordiae Hosp., 87 Cal. Rptr. 3d 81, 89 (Cal. Ct. App. 2008). As for Plaintiffs’ unreasonable course of conduct claim, Plaintiffs had to prove, under the “traditional rule of reason,” whether Sutter’s conduct “harms competition more than it helps.” Cipro, 348 P.3d at 861. Relevant considerations include “the facts peculiar to the business in which the restraint is applied, the nature of the restraint and its effects, and the history of the restraint and the reasons for its adoption.” Id. (quoting Topco Assocs., 405 U.S. at 607). A party’s intent to engage in anticompetitive conduct is probative of both claims. Consider an unreasonable course of conduct first. Again, “the history of the restraint and the reasons for its adoption” are crucial factors under the rule of reason. Id. (internal quotation marks omitted); see also Bd. of Trade of Chi., 246 U.S. at 238 (“The history of the restraint, the evil believed to exist, the reason for adopting  SIDIBE V. SUTTER HEALTH 35 the particular remedy, the purpose or end sought to be attained, are all relevant facts.”). Such evidence is not merely background context. Evaluating a party’s motives is particularly important when applying the rule of reason’s “fact-specific assessment.” Am. Express Co., 585 U.S. at 541. As the dissent concedes, the history of a party’s “conduct restraining trade” can explain, among other things, the “effects of the conduct.” That is, evidence from the past that a party engaged in certain conduct with the intent or belief that its conduct would have anticompetitive effects in the future is probative of whether that party’s conduct had anticompetitive effects. Moreover, in an unreasonable course of conduct claim, a defendant may rebut a prima facie case of anticompetitive effects with evidence of a “procompetitive rationale for the restraint.” Id. at 541–42; see also UAS Mgmt., 87 Cal. Rptr. 3d at 89 (rule of reasonableness, unlike per se violations, considers the “seller’s justifications”). Evidence— particularly contemporaneous evidence—that a defendant intended its conduct to have anticompetitive effects therefore also rebuts a later defense that the conduct was intended to benefit competition. As for tying, Sutter and our dissenting colleague are wrong to contend that, because tying arrangements are per se illegal, evidence of Sutter’s purpose is irrelevant. In cases alleging per se violations of the antitrust laws, the alleged conduct is “conclusively presumed to be unreasonable and therefore illegal without elaborate inquiry as to the precise harm they have caused or the business excuse for their use.” N. Pac. Ry. Co. v. United States, 356 U.S. 1, 5 (1958). In such cases, “it is often enough for the plaintiff to show” that the defendant “intentionally engaged in conduct which, if carried out as planned, would always or almost always 36 SIDIBE V. SUTTER HEALTH adversely affect competition.” Cascade Cabinet Co. v. W. Cabinet & Millwork Inc., 710 F.2d 1366, 1372 (9th Cir. 1983). In other words, without knowing what a defendant intended to do, it may be impossible to determine what a defendant actually did. In contending otherwise, the dissent conflates intent as a stated justification for otherwise anticompetitive conduct with intent as a form of evidence that a party engaged in anticompetitive conduct in the first place. Evidence of the former is indeed irrelevant to per se violations. See, e.g., UAS Mgmt., 87 Cal. Rptr. 3d at 89 (where a seller engages in a tying arrangement, “the seller’s justifications for the arrangement” are irrelevant); United States v. Joyce, 895 F.3d 673, 679 (9th Cir. 2018) (in case involving per se violation, defendant’s “attempt to persuade this court that his conduct was procompetitive” was legally irrelevant). As a more general matter, however, evidence of a defendant’s intent can often be important to prove that a defendant “intentionally engaged in conduct” constituting a per se violation. Cascade Cabinet Co., 710 F.2d at 1372; see also Cont’l Ore Co., 370 U.S. at 709–10, 710 n.15 (collecting cases and reaffirming that evidence of a defendant’s past acts can be important to “ascertain [the defendant’s] monopolistic intent or purpose”). Recall that the first two elements of a tying claim require a plaintiff to prove (1) that the defendant linked the sales of its tying and tied products and (2) that the defendant exercised market power in the tying market. UAS Mgmt., 87 Cal. Rptr. 3d at 89. Evidence that the defendant intended to link the sales of its tying and tied products or intended to exercise market power in the tying market, therefore, is probative of whether the defendant in fact engaged in either  SIDIBE V. SUTTER HEALTH 37 conduct—even if the specific evidence predates the class period. Cont’l Ore Co., 370 U.S. at 709–10. Pre-2006 evidence would have helped Plaintiffs prove both their tying and unreasonable course of conduct claims. To see why, consider the specific evidence Plaintiffs attempted to introduce but that the district court excluded. For convenience, we divide this evidence into five general (albeit somewhat overlapping) categories: (1) “admissions” by Sutter; (2) Sutter’s switch from individual to systemwide contracting and the health plans’ objections thereto; (3) Sutter’s imposition of the challenged contract terms during Sutter’s switch to systemwide contracting; (4) the health plans’ objections, before 2006, to the challenged contract terms; and (5) the 1999 Alta Bates-Summit merger litigation. All are highly probative of Plaintiffs’ claims. 1. Admissions by Sutter As evidence of Sutter’s admissions, the district court first excluded a 1997 Sutter memo explaining that Anthem would likely resist systemwide contracting “because of the increased leverage that twenty-one hospitals can achieve by working together” and estimating a potential gain to Sutter of $11 million. Second, the court excluded a 1998 memo in which Robert Reed, Sutter’s Chief Financial Officer (CFO), estimated that the “future benefit when all HMOs and PPOs contract on a system basis is” $198 million per year. Third, the court excluded Reed’s deposition testimony that the purpose of switching to systemwide contracting was to achieve “vastly better results,” by which Reed meant “better pricing.” Finally, the court excluded the 2006 Strategy Advantage memo, which included an interview in which future Sutter CEO Sarah Krevans (then the CEO of two individual Sutter hospitals) stated: “Related to the health 38 SIDIBE V. SUTTER HEALTH plans, we force them to pay us more. They do pay us more, and they don’t like us. In some cases, they have paid us more than the market. We’re working on it, though. There are lots of reasons why we pushed the health plans. Mainly, we pushed them because we could.” Because Sutter’s systemwide contracting was the mechanism by which Sutter allegedly imposed its anticompetitive contract terms, these admissions are highly relevant to Plaintiffs’ theory of the case. To be sure, as Sutter emphasizes, Sutter was not on trial for contracting on a systemwide basis. Plaintiffs’ theory, however, was that Sutter used its systemwide contracts to impose tying arrangements and other anticompetitive contract terms—that is, as Plaintiffs’ rebuttal to the jury put it, systemwide contracting becomes anticompetitive “when that systemwide contract is forced on you.” Evidence that Sutter’s intent in implementing systemwide contracting was to use those contracts as a vehicle to engage in tying or other anticompetitive behavior is therefore relevant to Plaintiffs’ legal theory, even if systemwide contracting itself is not unlawful. As a result, the admission by Sutter’s future CEO (Krevans) that Sutter not only intended to force health plans to pay above-market rates but actually did force the health plans to pay above-market rates, and furthermore did so “because we could,” is direct evidence of both anticompetitive effects (the central inquiry in an unreasonable course of conduct claim) and market power (the second element of a tying claim). Evidence that Sutter planned and began engaging in such conduct, even before the beginning of the class period, is probative of whether Sutter continued to do so. See, e.g., Cont’l Ore Co., 370 U.S. at 710 n.15 (defendant’s past acts used to “ascertain [the  SIDIBE V. SUTTER HEALTH 39 defendant’s] monopolistic intent or purpose”); U.S. Football League v. Nat’l Football League, 842 F.2d 1335, 1371 (2d Cir. 1988) (past anticompetitive conduct and the history of competition may “establish the intent, motive and method” of a restraint of trade). The fact that the district court excluded Krevans’s admission even though it was from 2006, simply because it “look[ed] back at” pre-2006 conduct, further highlights the district court’s error. The other excluded evidence was highly probative of Plaintiffs’ allegations as well. Consider Reed (Sutter’s CFO)’s deposition testimony that Sutter implemented systemwide contracting with the purpose of achieving higher pricing and Sutter’s contemporaneous memos quantifying the amount it stood to gain from switching to systemwide contracting. When combined with Krevans’s admission that, after implementing systemwide contracting, Sutter was able to force the health plans to pay above-market rates, this evidence would have bolstered Plaintiffs’ allegations that Sutter used systemwide contracting to link in-network participation in, or supracompetitive non-par rates at, its tying and tied hospitals (the first element of a tying claim). See Jordan v. Binns, 712 F.3d 1123, 1134 (7th Cir. 2013) (“People usually don’t make damaging admissions unless they are true.” (quoting Murrey v. United States, 73 F.3d 1448, 1455 (7th Cir. 1996))). Lastly, the excluded evidence of Sutter’s intent also could have undermined Sutter’s alternative explanations for its behavior. See Cerabio LLC v. Wright Med. Tech., Inc., 410 F.3d 981, 997 (7th Cir. 2005) (“blanket exclusion” of party’s evidence that had both affirmative and impeachment value caused “cumulative prejudice” to the party). For example, Sutter introduced evidence at trial that its purpose in adopting systemwide contracting was to make agreements 40 SIDIBE V. SUTTER HEALTH easier and more efficient to administer, testimony on which its expert witness later relied. In closing, Sutter repeated this contention, repeatedly emphasizing that systemwide contracting “provide[s] greater discounts. There are operational efficiencies. . . . [I]t would be very inefficient and costly to have [individual contracts] rather than one.” The district court’s exclusion of evidence denied Plaintiffs the ability to rebut this narrative and undermine the credibility of Sutter’s witnesses. 2. Imposing Systemwide Contracting and Objections Thereto The district court also excluded evidence explaining the timing and impact of Sutter’s implementation of systemwide contracting. This included: (1) a timeline of Sutter’s switch to systemwide contracting from 2001 to 2003; (2) acknowledgements by Sutter’s CFO Reed and Sutter’s Chief Contracting Officer (CCO) Melissa Brendt that Sutter contracted on an individual basis during the 1990s; (3) letters from Sutter to Health Net terminating individual hospital contracts; and (4) a graph prepared by Plaintiffs’ expert Dr. Tasneem Chipty showing that Sutter’s net patient revenue increased relative to other Northern California hospitals after 2002. Relatedly, the court excluded evidence that the health plans had objected to Sutter’s switch from individual to systemwide contracting. For example, Sutter’s 1997 memo acknowledged that Anthem did not want to negotiate on a systemwide basis. Anthem’s representative Steve Melody corroborated this acknowledgement, testifying in his deposition that Sutter abruptly terminated its existing individual hospital contracts with Anthem and insisted on systemwide contracting. Anthem protested, but it soon, in  SIDIBE V. SUTTER HEALTH 41 the district court’s words, “folded.” Melody added that after Sutter implemented systemwide contracting, rates subsequently increased by as much as 40 to 50 percent, an increase that was “far, far out of the ordinary.” Melody’s testimony and Dr. Chipty’s analysis corroborate Sutter’s own admission that it forced the health plans to pay above- market rates. Other health plan representatives likewise stated that Sutter had imposed systemwide contracting over their objections. Health Net’s representative Jenni Vargas recounted a similar experience to Anthem’s, testifying in her deposition that Sutter “told us that they were about to start negotiating as a system and they would be terminating all our Health Net contracts with all the Sutter affiliates” and that “[i]t was not a discussion . . . we were told that that’s what was going to happen.” Plaintiffs are correct that it is difficult to understand Sutter’s alleged tying and other anticompetitive conduct without understanding the history of its systemwide contracting practices. See, e.g., Reid Bros. Logging Co. v. Ketchikan Pulp Co., 699 F.2d 1292, 1305 (9th Cir. 1983) (holding, in a Sherman Act case, that “the background evidence is necessary to establish the defendants’ role in the limited market”). Additionally, contemporaneous evidence that Sutter was able to implement systemwide contracting over the health plans’ objections is essential evidence to prove that Sutter had market power, the second element of a tying claim under the Cartwright Act. UAS Mgmt., 87 Cal. Rptr. 3d at 89. The “extent of [Sutter]’s market power” is also relevant to an unreasonable course of conduct claim. Cipro, 348 P.3d at 861. Such evidence would have been particularly important given that Sutter, in closing, 42 SIDIBE V. SUTTER HEALTH repeatedly contended that Sutter “does not have sufficient market power to tie.” Our dissenting colleague observes that Plaintiffs were required to prove market power during the class period rather than from 2001 to 2003, but this only highlights why the excluded evidence was important. Because this evidence was excluded, in order to evaluate Sutter’s market power, conduct, and the effects of that conduct, the jury was only able to compare Sutter’s practices between 2006 and 2010 with Sutter’s practices during the damages period (2011 onward). This was particularly limiting because the parties appear to agree that Sutter’s conduct imposing the challenged contract terms was the same during these two periods. As a result, the jury could evaluate Sutter’s conduct after Sutter imposed the challenged contract terms, but the jury could not contrast that conduct with Sutter’s conduct before imposing the challenged contract terms, including Sutter’s previous practice of individual contracting, its motives for switching to systemwide contracting, its estimated resulting financial gain, and its ability to impose systemwide contracting over the objections of the health plans. 3. Imposing the Challenged Terms in Pre-2006 Contracts The third type of evidence excluded by the district court concerned anticompetitive terms in Sutter’s contracts between 2001 and 2005. The court excluded a timeline of Sutter’s first use of the challenged contract terms: equal treatment clauses in 2001, tiered products clauses in 2004, and high non-par rates clauses in 2005. The court also excluded examples of these contract terms in Sutter’s 2001 “Model Systemwide Amendment,” and agreements with  SIDIBE V. SUTTER HEALTH 43 Aetna, Anthem, Blue Shield, Health Net, and United Healthcare between 2001 and 2005. The fact that Sutter began including these challenged terms in its systemwide contracts shortly after implementing systemwide contracting from 2001 to 2003 supports Plaintiffs’ theory that Sutter’s purpose in switching to systemwide contracting was to engage in anticompetitive conduct and that systemwide contracting was the vehicle through which Sutter imposed anticompetitive contract terms. Again, even if these terms are identical to those challenged in 2011-onward contracts, the history and purpose of a course of conduct is indisputably relevant under the rule of reason. Cipro, 348 P.3d at 861. In the context of the other excluded evidence, evidence of these contract terms would have bolstered Plaintiffs’ theory that Sutter’s intent was to impose unfavorable contract terms that resulted in above-market pricing. Moreover, that Sutter was able to impose these contract terms despite the health plans’ objections (as discussed below) is indicative of Sutter’s market power.6 6 There is no basis for the contention, offhandedly raised by the dissent (but not by Sutter), that the excluded contracts were propensity evidence in violation of Federal Rule of Evidence 404(b). Evidence that a scheme to engage in anticompetitive conduct began in the past is not propensity evidence simply because the same scheme is alleged to have continued into the class period. That is why Rule 404(b) expressly permits the use of a party’s prior behavior to prove “motive,” “intent, preparation, plan,” or “knowledge.” Fed R. Evid. 404(b)(2); see also Cont’l Ore Co., 370 U.S. at 710 n.15 (past acts used to “ascertain [] monopolistic intent or purpose”); U.S. Football, 842 F.2d at 1371 (past conduct may “establish the intent, motive and method” of a restraint of trade). 44 SIDIBE V. SUTTER HEALTH 4. Health Plan Objections to the Challenged Contract Terms The fourth type of evidence excluded by the district court concerned the health plans’ objections, before 2006, to the challenged contract terms. This evidence included Anthem’s objections, in 2004, to high non-par rates, equal treatment clauses, and tiered products clauses. In 2003, meanwhile, Health Net objected to an equal treatment clause, specifically arguing that the clause “raises concerns about antitrust abuse.” The fact that health plans believed Sutter’s practices were anticompetitive before 2006 is highly probative of Plaintiffs’ claims. We have already discussed that Plaintiffs attempted to introduce contemporaneous evidence that Sutter both believed that imposing the challenged contract terms would have anticompetitive effects and conceded that it successfully achieved above-market pricing. The fact that the health plans also believed, again contemporaneously, that the contract terms would have anticompetitive effects is relevant corroboration. See, e.g., Browning v. Baker, 875 F.3d 444, 468 (9th Cir. 2017) (stressing the importance of contemporaneous evidence); Fontenot v. Crow, 4 F.4th 982, 1056 (10th Cir. 2021) (same). Also, that Sutter was able to impose the contract terms notwithstanding the health plans’ objections is evidence of both the existence and use of market power, which we repeat are relevant to both a tying and an unreasonable course of conduct claim. UAS Mgmt., 87 Cal. Rptr. 3d at 89; Cipro, 348 P.3d at 861. Just as importantly, Brendt (Sutter’s CCO) suggested in her testimony at trial that health plans did not begin complaining about the allegedly anticompetitive terms in  SIDIBE V. SUTTER HEALTH 45 Sutter’s systemwide contracts until 2012, after this suit was filed, thus implying that the complaints were insincere or opportunistic. The excluded evidence would have allowed Plaintiffs to rebut Brendt’s testimony. Cerabio LLC, 410 F.3d at 997. 5. The Alta Bates-Summit Merger Finally, the district court excluded two pieces of evidence from the State of California’s 1999 lawsuit to block the Alta Bates-Summit merger. Plaintiffs first attempted to introduce Sutter’s proposed findings of fact and conclusions of law from the merger case, in which Sutter argued that the merger would not lead to higher prices for patients because health plans could steer patients away from Sutter’s hospitals. Plaintiffs also attempted to introduce a 2011 study concluding that, following the merger with Sutter, Summit Medical Center raised its prices significantly more than other hospitals (28 to 44 percent larger than the control group) between 2002 and 2004. See generally Steven Tenn, The Price Effects of Hospital Mergers: A Case Study of the Sutter-Summit Transaction, 18 INT’L J. OF THE ECON. OF BUS. 65 (2011). Both pieces of evidence were highly relevant. Sutter’s opening statement and closing argument at trial both contended that tiered networks, one of the steering mechanisms used by health plans but prohibited in Sutter’s systemwide contracts, are harmful to and disliked by consumers. Such statements frame a case for a jury even though they are not evidence. Sutter’s submission in the Alta Bates-Summit merger trial could have rebutted the implication that Sutter’s tiered products clauses were implemented for the benefit of consumers and instead 46 SIDIBE V. SUTTER HEALTH indicated to the jury that Sutter’s purpose was to prevent steering and so raise costs for patients. As for Dr. Tenn’s 2011 study, part of Plaintiffs’ theory to the jury was that Sutter’s implementation of the challenged contract terms was responsible for a dramatic increase in healthcare prices in Northern California and that Sutter had implemented systemwide contracting in order to force the health plans to accept these contract terms. A study concluding that Sutter’s prices increased significantly more than other hospitals’ during the period in which Sutter began implementing systemwide contracts with the challenged terms (2002 to 2004) is therefore probative of both the purpose and effects of Sutter’s conduct. Moreover, Plaintiffs could have used the study’s conclusion that Sutter’s prices increased despite the nearby Kaiser Permanente Hospital to rebut Sutter’s contention that Kaiser and Sutter compete in the same market. * * * In short, Plaintiffs’ tying claim contended that Sutter linked in-network participation in (or supracompetitive non- par rates at) its tying and tied hospitals by forcing health plans to accept anticompetitive contract terms by negotiating on a systemwide basis rather than individually. Evidence that Sutter had previously employed individual contracts during the 1990s, switched to systemwide contracting in the early 2000s with the intent of imposing above-market prices, and then “forced” health plans to pay higher rates “because we could” is essential evidence both that Sutter did engage in tying and that Sutter amassed the market power to engage in tying—the first and second elements of a tying claim, respectively. Meanwhile, Sutter’s admission that it forced health plans to pay higher rates is direct evidence of  SIDIBE V. SUTTER HEALTH 47 anticompetitive effect, and evidence of Sutter’s switch from individual to systemwide contracting and Sutter’s belief that systemwide contracting would be more profitable is additionally probative of “the history of the restraint and the reasons for [the] adoption” of Sutter’s challenged contract terms, which are themselves means of proving anticompetitive effects. Cipro, 348 P.3d at 861 (internal quotation marks omitted). The district court failed to recognize that the excluded evidence was highly relevant. D. Exclusion of relevant evidence under Rule 403 is proper only if one or more of the rule’s articulated dangers or considerations “substantially outweigh[s]” the evidence’s probative value. Fed. R. Evid. 403. That all the evidence discussed above was highly relevant (as opposed to “minimally” so, as the district court concluded) does not, by itself, mean that the district court erred in excluding it under Rule 403. However, because the excluded evidence was highly relevant, any risk of prejudice or other dangers must be very high to justify exclusion. See, e.g., Johnson, 89 F.4th at 1009 (“The more probative the evidence, the more the court will tolerate some risk of prejudice.” (cleaned up)). The district court found that pre-2006 evidence would be cumulative of post-2006 evidence, and even if not cumulative, would confuse the issues, waste time, and delay the trial as the parties litigated “collateral issues.” The court added that the evidence from the 1999 Alta Bates-Summit merger litigation risked unfair prejudice, confusing the issues, and wasting time. Finally, the court concluded that the 2006 Strategy Advantage memo risked confusing the issues and was subject to the same concerns as the pre-2006 evidence. 48 SIDIBE V. SUTTER HEALTH We conclude that none of these dangers justified exclusion. 1. 2006 Cutoff Date As an initial matter, excluding all pre-2006 evidence was not justified by the district court’s discretion to set a “reasonable cut-off date” on the submission of evidence. Cont’l Ore Co., 370 U.S. at 710. Such cutoff dates are simply a means of ensuring that a trial does not become sidetracked by “collateral issues.” United States v. Socony- Vacuum Oil Co., 310 U.S. 150, 230 (1940). In this way they are much like other forms of balancing under Rule 403. Compare id. at 229–31 (discussing, in a case predating the modern Federal Rules of Evidence, evidence that is “merely cumulative,” could have “confused rather than enlightened the jury,” and would have “prolonged the inquiry and protracted the trial”), with Fed. R. Evid. 403 (relevant evidence may be excluded if “its probative value is substantially outweighed” by a risk of “confusing the issues, misleading the jury, undue delay, wasting time, or needlessly presenting cumulative evidence”). As we discuss below, none of Rule 403’s articulated dangers justified the wholesale exclusion of Plaintiffs’ pre-2006 evidence, so they cannot justify a blanket cutoff either. Thus, although district courts have the discretion to exclude evidence, United States v. Gonsalves, 675 F.2d 1050, 1054 (9th Cir. 1982), they must still carefully consider the proffered evidence to determine whether it is “collateral” or affects “matters of substance.” Socony-Vacuum, 310 U.S. at 230–31. Here, Plaintiffs attempted to introduce Sutter’s own admissions and other direct evidence of Sutter’s intended conduct and the effects thereof. These “matters of substance” are quite unlike the speculative and “collateral”  SIDIBE V. SUTTER HEALTH 49 evidence that trial courts may permissibly exclude. Id. at 228–31. Instead, the excluded evidence here is far more akin to the evidence that the Supreme Court has called “clearly material” to a party’s claims. Cont’l Ore Co., 370 U.S. at 709–10; see also Berkey Photo, Inc. v. Eastman Kodak Co., 603 F.2d 263, 296 (2d Cir. 1979) (evidence of anticompetitive actions before the statute of limitations period was relevant even though case depended on proof of overcharge during the limitations period). 2. Risk of Needlessly Presenting Cumulative Evidence We therefore turn to the primary reason the district court gave for excluding Plaintiffs’ pre-2006 evidence: that the evidence was needlessly cumulative. Here, the fact that the allegedly anticompetitive contracts were regularly renegotiated did not make the excluded evidence redundant, because Sutter’s intent and conduct prior to 2006 were probative of both whether Sutter’s conduct had anticompetitive effects, Cipro, 348 P.3d at 861, and whether any tying arrangement continued into the class period. Cont’l Ore Co., 370 U.S. at 710 n.15. We stress that the district court’s desire to avoid showing the jury every single systemwide contract that ever included an equal treatment or tiered products clause was prudent. However, this is not a case in which the plaintiffs attempted to introduce thousands of redundant documents over the course of decades, and as discussed, the history of Sutter’s course of conduct was a key element of Plaintiffs’ theory of the case. Evidence is not cumulative simply because it is probative of duration. United States v. Ives, 609 F.2d 930, 933 (9th Cir. 1979); see also generally Ben-E-Lect v. Anthem Blue Cross Life and Health Ins. Co., 265 Cal. Rptr. 3d 495 (Cal. Ct. App. 2020) (considering, in Cartwright Act claim 50 SIDIBE V. SUTTER HEALTH with damages period from 2011 to 2014, evidence of similar anticompetitive conduct from 2001 and 2006). Perhaps more importantly, when combined with the evidence that several of the health plans had objected to Sutter’s proposals to adopt the challenged contract terms, the fact that Sutter imposed the challenged terms regardless served as proof that Sutter had the market power to impose those terms unilaterally—something that Sutter repeatedly denied in closing. Our dissenting colleague identifies a few pieces of testimony at trial that were probative of Sutter’s pre-2006 conduct and intent, but these pieces of testimony were no substitute for the evidence that Plaintiffs were prevented from introducing. Recall that Melissa Brendt, Sutter’s CCO, contended at trial both that Sutter’s purpose in implementing systemwide contracting was to increase administrative efficiency and that the health plans did not object to the allegedly anticompetitive contract terms until 2012, after this suit was filed. Sutter’s expert witness relied on this testimony too, as did Sutter’s closing argument, which contended that systemwide contracting both has “operational efficiencies” and provides “greater discounts.” Likewise, Sutter repeatedly emphasized that it lacked the market power to force anything on the health plans, telling the jury in closing that Sutter “did not force these insurers to do anything. These insurance companies are some of the largest companies in the United States.” Sutter hammered the point home, adding that the health plans “want to control who makes the healthcare decisions,” and that if Sutter “really had that market power,” one would expect to see evidence that Sutter “could push around Aetna and the like”; instead, “Sutter needed the insurance companies with the memberships much more than the other way around.”  SIDIBE V. SUTTER HEALTH 51 Yet the district court excluded pre-2006 evidence that would have contradicted Sutter’s contentions and testimony. For example, Reed’s 1998 memo and deposition testimony stated that Sutter intended to implement systemwide contracting to achieve better pricing, which Reed estimated was worth nearly $200 million. The health plans’ written, contemporaneous objections to Sutter’s systemwide contracts and the allegedly anticompetitive terms therein— including Health Net’s specific contention that Sutter’s equal treatment clauses were anticompetitive—contradicted Brendt’s testimony that the health plans did not object before 2012. Meanwhile, Krevans’s admission that Sutter forced the health plans to pay above-market rates “because we could” suggested that Sutter did in fact have market power. The excluded evidence had three important advantages over the evidence Plaintiffs were able to introduce at trial. First, much of it was contemporaneous: Sutter’s intended course of conduct, its belief that it was charging above- market rates, and its ability to impose the new contract terms over the health plans’ objections. Contemporaneous evidence, particularly written evidence, is commonly understood to be more reliable than later recollections because it reduces the risks of “defective recollection or conscious fabrication”—hence why the hearsay rules make an exception for present-sense impressions. United States v. Green, 556 F.3d 151, 155–57 (3d Cir. 2009) (internal quotation marks omitted). Accordingly, courts have often emphasized the importance of contemporaneous evidence that either corroborates or contradicts later evidence. See, e.g., Fontenot, 4 F.4th at 1056 (corroborating); Browning, 875 F.3d at 468 (contradicting). Indeed, Sutter recognized the power of contemporaneous evidence when it contended 52 SIDIBE V. SUTTER HEALTH at trial that the health plans were only belatedly and opportunistically objecting to Sutter’s contracts. Second, Sutter’s own admissions would have had more persuasive force than other evidence. As with contemporaneous statements, the Federal Rules of Evidence recognize the value of a party’s own admissions. See, e.g., United States v. Mirabal, 98 F.4th 981, 987 (9th Cir. 2024) (exclusion of government’s prior statement, which should have been admitted as a party admission under Rule 801(d)(2), was not harmless because it could have persuaded the jury that defendant acted in self-defense); Jordan, 712 F.3d at 1134 (“People usually don’t make damaging admissions unless they are true.” (internal quotation marks omitted)). Third, the probative value of evidence is not simply a matter of the evidence’s substance but also its effect on a witness’s credibility. See, e.g., Henderson, 449 F.3d at 139 (Rule 403 did not justify exclusion of evidence that had probative value both as “affirmative evidence” and “for purposes of impeachment, rebuttal, and rehabilitation”); Cerabio LLC, 410 F.3d at 997 (reversing Rule 403 exclusion where party faced “cumulative prejudice” from “blanket exclusion” of evidence that had both affirmative and impeachment value). In sum, Sutter used the excluded evidence both “to shield [itself] from potentially damaging evidence” and as “a sword to slice through the foundation” of Plaintiffs’ case. Henderson, 449 F.3d at 140. These three points aside, we repeat that the pre-2006 evidence must be considered as a whole and in the context of Plaintiffs’ theory of this complex litigation. Relevance considers how each “brick” of evidence can, together, form a wall, not whether each “witness can make a home run.”  SIDIBE V. SUTTER HEALTH 53 Fed. R. Evid. 401 advisory committee’s note to 1972 proposed rules (internal quotation marks omitted). Plaintiffs’ contention was not just (as Sutter’s CFO Reed conceded at trial) that systemwide contracting was “better” for Sutter or (as Blue Shield’s witness Kristen Miranda testified) that systemwide contracting “would not have been [the health plans’] preference.” Rather, Plaintiffs’ contention was that Sutter (1) adopted systemwide contracting for the purposes of linking in-network participation in—or supracompetitive non-par rates at—its tying and tied hospitals, imposing unfavorable contract terms on the health plans, and ultimately achieving supracompetitive pricing; and (2) successfully implemented this plan and so achieved supracompetitive pricing. Far from having “scant or cumulative probative force,” Hankey, 203 F.3d at 1172 (quoting Mills, 704 F.2d at 1560), the excluded pre-2006 evidence was essential to prove these specific allegations and thus Plaintiffs’ theory of the case as a whole. 3. Risk of Unfair Prejudice Sutter’s contention of unfair prejudice is even weaker. That the evidence of Sutter’s admissions before 2006 and the timing and impact of Sutter’s implementation of systemwide contracting was damaging to Sutter does not mean that the admissions posed a risk of unfair prejudice, at least not one that substantially outweighed the evidence’s probative value. See, e.g., United States v. Thornhill, 940 F.3d 1114, 1123 (9th Cir. 2019) (emphasizing that even “highly prejudicial” evidence is “not necessarily unfairly prejudicial” (internal quotation marks omitted)); United States v. Portillo, 969 F.3d 144, 179 (5th Cir. 2020) (“Relevant evidence is inherently prejudicial.” (internal quotation marks omitted)); Cuff v. Trans States Holdings, Inc., 768 F.3d 605, 609 (7th Cir. 2014) (“The question under 54 SIDIBE V. SUTTER HEALTH Rule 403 is not whether evidence is ‘prejudicial’ . . . . It is inappropriate to exclude evidence under Rule 403 because it casts [the opposing party] in a really bad light.”). The history of a restraint and the reasons for adopting it are essential aspects of the rule of reason analysis under the Cartwright Act. Cipro, 348 P.3d at 861. There is nothing unfair about admitting evidence of precisely that. The only risk of unfair prejudice under Rule 403 is the possibility that the jury would have erroneously believed that Sutter’s use of systemwide contracting was itself a per se antitrust violation. However, Sutter’s use of systemwide contracting from 2006 onward was sufficiently discussed during the trial such that it is implausible that the marginal risk from admitting the pre-2006 evidence would substantially outweigh the evidence’s significant probative value. Moreover, because systemwide contracting and the allegedly anticompetitive contract terms were implemented around the same time, Plaintiffs would have been able to use the excluded evidence to argue that Sutter’s sizable rate increases during the early 2000s were attributable to the challenged contract terms, not systemwide contracting per se. In any event, a cautionary jury instruction could have cured any risk of unfair prejudice or confusion of the issues. See, e.g., Thornhill, 940 F.3d at 1123 (commending trial court for giving limiting instruction to mitigate risk of unfair prejudice); United States v. W.R. Grace, 504 F.3d 745, 765 (9th Cir. 2007) (emphasizing, as an alternative to excluding evidence, “the potential efficacy of a limiting instruction”); United States v. Boulware, 384 F.3d 794, 808 (9th Cir. 2004) (“Any danger” from admitting evidence “could have been dealt with by a cautionary instruction”). Nor was there any substantial danger in admitting evidence of the health plans’ pre-2006 objections to Sutter’s  SIDIBE V. SUTTER HEALTH 55 systemwide contracting and the challenged contract terms. Sutter, joined by our dissenting colleague, suggests that this evidence, which included Health Net’s letter indicating that Sutter’s practices “raise[] concerns about antitrust abuse,” would have invited the jury to hold Sutter liable for conduct from before the damages period began in 2011 or before 2008, when the four year statute of limitations period began. Sutter’s argument is meritless. The district court already permitted the jury to consider evidence from 2006 and 2007 because such evidence provided “context” to Plaintiffs’ allegations, so any additional risk from pre-2006 evidence would be minimal. As we have already discussed, evidence of Sutter’s acts prior to the damages period is relevant both to whether a tying arrangement continued into the class period and as direct evidence of whether Sutter’s conduct had anticompetitive effects. See Cont’l Ore Co., 370 U.S. at 709–10, 710 n.15 (collecting cases considering past conduct); Cipro, 348 P.3d at 861 (history and purpose are probative of effect). Particularly where the evidence of Sutter’s prior conduct concerned the inception of the same conduct alleged in this case, any risk of unfair prejudice did not outweigh the evidence’s probative value, let alone substantially so. See, e.g., U.S. Football League, 842 F.2d at 1371 (even evidence of prior antitrust judgments is not unfairly prejudicial where the “conduct underlying those prior judgments had a direct, logical relationship to the conduct at issue in this case”); Sullivan v. Nat’l Football League, 34 F.3d 1091, 1113 (1st Cir. 1994) (same). Regardless, again, a jury instruction would have mitigated this risk. Thornhill, 940 F.3d at 1123; W.R. Grace, 504 F.3d at 765; Boulware, 384 F.3d at 808. 56 SIDIBE V. SUTTER HEALTH 4. Other Rule 403 Dangers Finally, although we sympathize with the district court’s reluctance to relitigate the 1999 Alta Bates-Summit merger trial, the risk of confusing the issues or wasting time did not substantially outweigh the probative value of the two pieces of evidence Plaintiffs attempted to introduce. The jury instructions were already clear that Plaintiffs alleged that Sutter tied hospital services and imposed anticompetitive contract terms on health plans, not that its 1999 merger was illegal. Moreover, the district court had options less restrictive than exclusion, including admitting only the specific relevant portions of Sutter’s submission in that case and Dr. Tenn’s 2011 study (or, indeed, any other lengthy documents), and issuing a cautionary instruction. See, e.g., Obrey, 400 F.3d at 699 (concluding that, although the trial court had valid concerns with the risk of “mini-trials” stemming from some evidence, it “should have first addressed these concerns with the parties through other, less restrictive means”). * * * Viewing the evidence as a whole, it is apparent that the district court abused its discretion in its blanket exclusion of pre-2006 evidence. We recognize that, as Sutter emphasizes, the district court had “broad discretion” to conduct its Rule 403 analysis. Sprint, 552 U.S. at 384. Had the district court simply excluded a few specific pieces of evidence that we would not have excluded, then our difference of opinion would not amount to an abuse of discretion. Here, however, the district court excluded documents all based on the same fundamental error: failing to appreciate that pre-2006 evidence was highly relevant to both Plaintiffs’ tying and unreasonable course of conduct  SIDIBE V. SUTTER HEALTH 57 claims. Thus, the district court did not engage in the “on- the-spot balancing of probative value and prejudice” that normally warrants such broad discretion. Id. (internal quotation marks omitted). Instead, the 2006 cutoff date, which the district court conceded was arbitrary, in essence “applied a per se rule excluding” pre-2006 evidence. Id. at 385; see also id. at 387 (“Relevance and prejudice under Rules 401 and 403 . . . are generally not amenable to broad per se rules.”); Cerabio LLC, 410 F.3d at 994 (listing decisions that “have not hesitated to overturn blanket evidentiary rulings” excluding evidence from before a particular date). Our dissenting colleague again mischaracterizes our decision by accusing us of permitting parties to introduce evidence “from the inception of time,” “whenever ‘history’ is relevant,” and of barring courts from ever setting reasonable cutoff dates. Nothing in our decision today prevents district courts from setting reasonable limits on the introduction of evidence or from conducting the proper balancing test under Rule 403. All we hold is that, in this particular case, the district court used too blunt an instrument. The wholesale exclusion of evidence prior to 2006 was arbitrary because it deprived Plaintiffs of the evidence essential to proving their allegations. See, e.g., Henderson, 449 F.3d at 141 (district court abused its discretion by excluding evidence that, among other things, went “to the heart of a party’s case” and appeared “crucial to the outcome of the case”). Our review is deferential, but it is not a blank check, and we will not ignore evidentiary rulings that are “without support in inferences that may be drawn from the facts in the record.” Crawford v. City of Bakersfield, 944 F.3d 1070, 1077 (9th Cir. 2019) (quoting United States v. Espinoza, 880 F.3d 506, 511 (9th Cir. 58 SIDIBE V. SUTTER HEALTH 2018)). In short, “this is one of those rare cases where reversal is warranted.” Johnson, 89 F.4th at 999. The evidence at issue did not meet Rule 403’s high bar for exclusion. Accordingly, we conclude that the district court abused its discretion. E. Because the exclusion was error, we “must presume prejudice.” Barranco, 952 F.3d at 1127 (internal quotation marks omitted). Sutter is “[t]he party benefitting from the error” and so “has the burden of persuasion” to prove that the jury would have reached the same verdict even if the evidence had been admitted. Obrey, 400 F.3d at 701. Once again, Sutter has not met this burden. Sutter first argues that it presented sufficient evidence at trial that it did not link the in-network participation of its tying hospitals and the in-network participation of its tied hospitals. However, Plaintiffs alleged a second tying condition: not only the in-network participation of Sutter’s tied hospitals but also the payment of supracompetitive non- par rates at Sutter’s tied hospitals. The fact that Sutter presented evidence relevant to the network participation of its tied hospitals is not necessarily dispositive of whether Sutter engaged in tying at all. In any event, the jury might have come to a different conclusion about whether Sutter imposed a tie if the jurors had been permitted to consider the evidence about market conditions before Sutter began contracting on a systemwide basis and during Sutter’s implementation of systemwide contracting. Sutter likewise briefly alludes to the fact that the jury answered “no” to the first question relating to the tying claim on the verdict form. The verdict form, however, does not demonstrate that the erroneous exclusion of evidence was  SIDIBE V. SUTTER HEALTH 59 not prejudicial. The first question asked: “Did Sutter sell inpatient hospital services in one or more of the tying hospitals only if the buyer also purchased inpatient hospital services at one or more of the tied hospitals?” Sutter’s history of individual contracting, its motives for switching to systemwide contracting, the financial gain associated with that shift, and the health plans’ objections and ultimate acquiescence all could have impacted the jury’s answer to question 1. The excluded evidence could inform the jury’s interpretation of Sutter’s actions, Cascade Cabinet Co., 710 F.2d at 1372, and the credibility of Sutter’s testimony. Sutter has failed to prove that it is more probable than not that the erroneous exclusion of evidence “did not taint the jury’s verdict.” Obrey, 400 F.3d at 701–02. Sutter also contends that the evidence presented at trial demonstrated that Sutter’s course of conduct was not unreasonable. Once again, however, the history and purpose of a challenged restraint have long been considered relevant under a rule of reason analysis, under both the Cartwright Act and Sherman Act. Cipro, 348 P.3d at 861; Topco Assocs., 405 U.S. at 607. Moreover, as we have explained, the district court excluded direct evidence of Sutter’s admission that it was charging above-market rates and Sutter’s belief that its conduct would have anticompetitive effects, both of which are highly probative of whether Sutter’s conduct had anticompetitive effects. To the extent that Sutter repeats its argument that the verdict form proves harmlessness with respect to Plaintiffs’ unreasonable course of conduct claim, the argument has even less merit than with respect to the tying claim. Question 5 asked: “Did Sutter force the class health plans to agree to contracts that had terms that prevented the plans from steering patients to lower-cost non-Sutter hospitals 60 SIDIBE V. SUTTER HEALTH within the plan network?” Sutter contends that the trial evidence conclusively showed that the challenged contract terms did not prevent steering (that is, that there were no anticompetitive effects). Again, however, question 5 had two parts, asking both (1) whether Sutter forced the health plans to accept the challenged contract terms, and (2) whether those contract terms prevented steering. Sutter neglects the first half of the question. For example, the excluded statements by Krevans, Sutter’s future CEO, that “we force” the health plans “to pay us more,” that the health plans “do pay us more,” and that “they have paid us more than the market” certainly could have affected the jury’s deliberation about whether Sutter had market power and “forced” the health plans to accept the challenged terms. Proving both halves of question 5 depended on the excluded pre-2006 evidence. Sutter simply asserts, without evidence, that when the jury answered “no” to question 5, the jury’s answer must have been rooted in the second half of the question rather than the first. When pressed at oral argument, however, counsel for Sutter largely conceded that harmless error review does not permit such speculation. Sutter’s concession is correct: speculation about how the jury arrived at its “no” verdict to question 5 is insufficient for us to “say that the error was harmless; that is, we are unable to hold that ‘it is more probable than not that the error did not materially affect the verdict.’” United States v. Wiggan, 700 F.3d 1204, 1215 (9th Cir. 2012) (quoting Boyd, 576 F.3d at 949). In sum, the district court abused its discretion under Rule 403 by excluding the pre-2006 evidence, and Sutter has failed to rebut the presumption that the error prejudiced Plaintiffs.  SIDIBE V. SUTTER HEALTH 61 Finally, we return to the subject of cumulative error. Much of the excluded evidence was probative of whether Sutter’s purpose in implementing systemwide contracting and adopting the challenged contract terms was to engage in anticompetitive behavior. The damage from excluding this pre-2006 evidence was compounded by the district court’s failure to instruct the jury that Sutter’s purpose was relevant in the first place. See Jerden, 430 F.3d at 1240–41 (recognizing risks of cumulative error and collecting similar cases from other courts of appeals). The court’s improper exclusion of evidence and its erroneous omissions of “purpose” from the jury instructions therefore cannot be disentangled from one another: both errors stem from the district court’s belief that Sutter’s anticompetitive purpose is minimally relevant to this case. For all the reasons we have explained, this belief is both contrary to the Cartwright Act and a misunderstanding of Plaintiffs’ theory of the case. Either error would be sufficiently prejudicial to warrant a new trial. Even if neither error alone justified reversal, however, it is “not always” the case that “0 + 0 = 0,” because courts must “consider the whole picture” rather than make the “persistent error in legal analysis” of asking “whether a piece of evidence ‘by itself’ passes some threshold.” United States v. Vaughn, 62 F.4th 1071, 1072–73 (7th Cir. 2023) (cleaned up); see also United States v. Roper, 72 F.4th 1097, 1103 (9th Cir. 2023) (approvingly citing Vaughn). Here, particularly given the erroneous jury instructions, we cannot say that it is more probable than not that the jury would have reached the same result if the pre-2006 evidence had been admitted. Equally, particularly given the improperly excluded evidence, we cannot say that it is more probable than not that the jury would have reached the same result if properly instructed. See Obrey, 400 F.3d at 701–02 62 SIDIBE V. SUTTER HEALTH (improperly excluded evidence); Fierro, 39 F.4th at 651 (erroneous instruction). IV. For the foregoing reasons, the district court erred in failing to instruct the jury to consider Sutter’s anticompetitive purpose and in excluding highly relevant pre-2006 evidence. Accordingly, we reverse the district court’s entry of final judgment and remand for a new trial. REVERSED and REMANDED. Sutter shall bear costs on appeal. BUMATAY, Circuit Judge, dissenting: “As sands in the hourglass, so are the days of our lives.” In soap operas, it can be difficult to keep up with the latest twists and turns in their characters’ lives. Characters come and go (and sometimes come back again). Relationships begin and end (and often begin again). Now imagine trying to understand the latest plot development in a long-running soap opera you’ve never seen before. Of course, it would be highly entertaining and helpful to know what those characters were up to ten years ago. Certainly, to understand current relationships, it’d be useful to see who was in a relationship with whom back then. After all, each story arc builds on previous ones. And arcs often repeat. But to catch up on the lives of those characters today, it would be nearly impossible to watch every episode over those ten years. Indeed, given how frequently episodes are released, a span of ten years could mean almost 2,500 episodes. So to appreciate what’s happening in the soap opera now, it’s more practical to focus on the most recent episodes. And at some  SIDIBE V. SUTTER HEALTH 63 point, we must make choices on how far to go back. Deciding how many episodes to watch requires balancing competing concerns and making difficult decisions. In the end, while it would have been better to watch more episodes, sometimes we must live with what we can accomplish in the time we have. After all, just like the sand in that hourglass, the days of our lives are not endless. In complex litigation, district courts and litigants face the same pressures. The parties want to litigate disputes that sometimes have long histories but only involve more recent conduct. When that’s the case, we rely on the district court’s discretion to fashion reasonable limits for admissible historical evidence. After all, trials can’t last forever and jurors can’t always process 5, 10, or 20 years of evidence. The same is true in antitrust cases. The history of conduct restraining competition can often be long. Understanding that history, of course, can be helpful. It can help explain the nature of the industry, the market power of participants, the effects of the conduct, business purposes and justifications, and the like. So historical evidence may offer probative backdrops. But that doesn’t mean we can’t orient antitrust trials toward the most relevant time periods. That doesn’t mean courts shouldn’t set reasonable limits on admissible evidence. Under Rule 403 of the Federal Rules of Evidence, district courts have the authority to manage complex trials by setting reasonable cutoff dates for historical evidence. Such restrictions may focus trials on the most important times or issues. Otherwise, jurors will be forced to consider every tangential storyline a litigant wants to introduce. It should be obvious—the further back in time we look from the alleged wrongdoing, the more collateral the evidence 64 SIDIBE V. SUTTER HEALTH becomes. It can become stale, cumulative, confusing, and distracting for jurors. And the burdens on trial multiply— potentially adding days or weeks to already prolonged trials. So while historical evidence may be probative in many cases, perhaps especially in antitrust cases, that doesn’t mean a party may introduce every piece of evidence from the inception of time. Instead, we leave it to the district court’s discretion to place reasonable limits on admissible evidence. So broad is the district court’s discretion in this context that, to my knowledge, no federal circuit court has ordered a re- trial based on the setting of a reasonable evidence cutoff date. We are now the first. In this case, Djeneba Sidibe and a class of businesses and individuals claimed they paid inflated health insurance premiums to health plan providers because of anticompetitive conduct by Sutter Health. They assert that Sutter, which owns a chain of hospitals in Northern California, forced the health insurers into contracts with anticompetitive terms, causing them to pay more expensive premiums. The class alleges that Sutter’s antitrust conduct began in the late 1990s. But because of the lack of cognizable damages from before 2011, this suit only involves contracts which allegedly caused damages from 2011 to 2020. So while Sutter’s conduct may have stretched back 20 years, only more recent conduct is relevant to the case. To manage the trial, the district court limited the historical evidence that the parties could introduce. The district court, which had presided over the litigation for almost a decade, ruled that five years prior to the damages period, meaning 2006, was a reasonable cutoff for evidence. We should have left that decision in the hands of the district court.  SIDIBE V. SUTTER HEALTH 65 Today, we limit the discretion that district courts enjoy in managing trials and second guess their ability to set reasonable evidentiary limits. We reverse the jury verdict because we don’t like the district court’s choice of five years. But we offer no other guiding principles. Instead, according to the majority, district courts may no longer exclude historical evidence from trial whenever “history” is relevant. This applies not only in every antitrust case—but potentially countless others. Thus, after our ruling, litigants—not judges—get to choose how far back in time they want to present evidence. This is a mistake. *** If this wasn’t enough, there’s something even more damaging here. In its rush to overturn the jury verdict, the majority also crafts a new antitrust rule. The majority states that rule-of-reason cases require consideration of the defendant’s anticompetitive purpose and any jury instruction that does not directly mandate its consideration is reversible error. Anticompetitive purpose now becomes an element for every antitrust case, regardless of the individualized circumstances of the case. Worse yet, the majority makes this rule the law for every California and federal antitrust case going forward. Never mind that neither Supreme Court nor California court precedent justifies this novel rule. Indeed, it flies in the face of decades of precedent establishing that rule-of-reason claims require case-by-case determination. *** Because the majority gets both Rule 403 law and antitrust law wrong, I respectfully dissent. 66 SIDIBE V. SUTTER HEALTH I. Background Before jumping into the facts here, it’s worth starting with a brief overview of antitrust law. I then dive into the most important facts of the case. A. Antitrust Law The Cartwright Act is California’s “principal antitrust law.” In re Cipro Cases I & II, 61 Cal. 4th 116, 136 (2015). By its terms, the Act declares that “every trust is unlawful, against public policy and void.” Cal. Bus. & Prof. Code § 16726. Prohibited “trust[s]” include any “combination . . . by two or more persons” to “create or carry out restrictions in trade or commerce” or to “prevent competition in manufacturing, making, transportation, sale or purchase of merchandise, produce or any commodity.” Id. § 16720(a), (c). But much like the Sherman Act, its federal counterpart, the Cartwright Act doesn’t reach “every agreement within the four corners of its prohibitions.” In re Cipro, 61 Cal. 4th at 136 (simplified). It “prohibits not all agreements restraining trade, but rather agreements that unreasonably restrain trade.” Flagship Theatres of Palm Desert, LLC v. Century Theatres, Inc., 55 Cal. App. 5th 381, 399 (2020) (simplified). So the focus of the antitrust inquiry is on the anticompetitive effects of any agreement. After all, the law is designed “to prevent undue restraints upon trade which have a significant effect on competition.” Corwin v. L.A. Newspaper Serv. Bureau, Inc., 22 Cal. 3d 302, 314 (1978) (simplified).  SIDIBE V. SUTTER HEALTH 67 Again, following federal law, Cartwright Act antitrust violations can be broken into two categories: (1) “per se” and (2) “rule of reason” violations. Flagship Theatres, 55 Cal. App. 5th at 399–400. Whether conduct is a “per se” or “rule of reason” violation depends on the nature of its anticompetitive effects. Some agreements have such a “pernicious effect on competition and lack of any redeeming virtue” that they “are conclusively presumed to be unreasonable and therefore illegal without elaborate inquiry as to the precise harm they have caused or the business excuse for their use.” Id. at 399 (quoting N. Pac. R. Co. v. United States, 356 U.S. 1, 5 (1958)). These “manifestly anticompetitive” practices are unreasonable “under any circumstances”—so much so that anticompetitive effects are presumed. Id. (simplified). And similarly, there’s no need to prove a precise anticompetitive intent or balance any procompetitive justification. See Cascade Cabinet Co. v. W. Cabinet & Millwork Inc., 710 F.2d 1366, 1372 (9th Cir. 1983) (discussing how, in per se violations, “anti-competitive purpose” only requires a showing that “the defendants intentionally engaged in conduct which, if carried out as planned, would always or almost always adversely affect competition”); United States v. Hilton Hotels Corp., 467 F.2d 1000, 1002 (9th Cir. 1972) (explaining that “[t]he ultimate objective defendants sought to achieve is immaterial” for a per se violation of the Sherman Act); United States v. Container Corp. of America, 393 U.S. 333, 340 (1969) (Marshall, J., dissenting) (“Under the antitrust laws, numerous practices have been held to be illegal per se without regard to their precise purpose or harm.”). So if conduct falls within this category, it’s an antitrust violation—no balancing is required. See Chavez v. 68 SIDIBE V. SUTTER HEALTH Whirlpool Corp., 93 Cal. App. 4th 363, 369 (2001). These are “per se” violations. All other cases fall within the “rule of reason,” which— as the name implies—requires balancing of the conduct’s effects with any procompetitive justifications. These cases require determining whether the “practice unreasonably restrains trade by assessing its actual competitive effects under the rule of reason.” Flagship Theatres, 55 Cal. App. 5th at 400. Put simply, “[t]he rule of reason weighs the anticompetitive effects of the conduct in the relevant market against its procompetitive effects, and determines whether, on balance, the practice harms competition.” Id. Two theories of antitrust conduct demonstrate the difference between “per se” and “rule of reason” violations—tying and anti-steering. Tying is a quintessential “per se” violation. Tying occurs when a seller “use[s] its market power in one market to force or coerce a buyer to purchase its product or service in a distinct market in which the seller does not have such market power or to refrain from buying from the seller’s competitor.” UAS Mgmt., Inc. v. Mater Misericordiae Hosp., 169 Cal. App. 4th 357, 368–69 (2008). So a seller is prohibited from conditioning the sale of one product on “the buyer also purchas[ing] a different (or tied) product.” Suburban Mobile Homes, Inc. v. Amfac Cmtys., Inc., 101 Cal. App. 3d 532, 542 (1980) (simplified). It is a “per se” violation because it serves no purpose other than to “deny competitors free access to the market for the tied product.” Id. In contrast, anti-steering violations fall within the “rule of reason.” See Ohio v. American Express, 585 U.S. 529, 541 (2018). In a competitive market, the ability of buyers to  SIDIBE V. SUTTER HEALTH 69 steer away from more expensive sellers keeps prices low. That’s because those sellers must compete with cheaper ones or lose business. But anti-steering provisions defeat this market force. Anti-steering allows the seller to prevent buyers from shifting to less expensive alternatives. In other words, anti-steering means that a buyer is stuck with the seller no matter the price. As a “rule of reason” violation, anti-steering claims must balance any “anticompetitive effect[s]” of the restraint with the seller’s “procompetitive rationale” for steering. Id. B. The Facts Sutter Health operates a network of 24 hospitals in Northern California. As part of its business, Sutter negotiates contracts with various health insurance providers. The big-name insurers include Aetna, Anthem Blue Cross, Blue Shield of California, Health Net, and United Healthcare. The contracts set the rates that the insurers will pay Sutter for the medical services provided to their customers. Djeneba Sidibe represents a class of employers and individuals who paid health insurance premiums to insurers who contracted with Sutter Health (collectively, “Sidibe”). Sidibe alleges that she and her co-plaintiffs paid inflated insurance premiums based on Sutter’s anticompetitive conduct. In 2012, Sidibe sued Sutter under California’s Cartwright Act, alleging two claims—one for tying and another for its anti-steering contract provisions. Sidibe sought damages for Sutter’s conduct from 2008 onwards. 70 SIDIBE V. SUTTER HEALTH First, according to Sidibe’s complaint, Sutter refused to allow its seven most desirable hospitals to participate in insurers’ “network” plans (i.e., the “tying” hospitals) unless insurers also included four of its less desirable hospitals within the plans (i.e., the “tied” hospitals). Second, Sidibe claims that Sutter included anti-steering provisions in its systemwide contracts designed to make it impossible for the insurers to push its patients to lower-cost hospitals. Sidibe argues that three particular provisions— “non-par rate,” “equal treatment,” and “tiered products” clauses—prevented insurers from including Sutter hospitals in lower premium tiers or shifting patients to lower cost competitors. Each clause does something different: The “non-par rate” term encourages plans to keep Sutter hospitals in network; the “equal treatment” clause prevents health plans from making Sutter hospitals less available to patients; and the “tiered products” clause prevents health plans from including Sutter hospitals in cheaper, tiered networks. After almost a decade of litigation, the case was ready for trial. But in the run up to trial, the district court narrowed the case. It granted summary judgment for Sutter on all conduct from before 2011 because Sidibe couldn’t establish any injury from that period. So the only conduct left for trial was Sutter’s contracts with insurance companies impacting service from 2011 to 2020. Before trial, Sutter moved in limine to exclude evidence from before 2006—five years prior to the damages period— under Federal Rule of Evidence 403, as too remote in time and too disconnected from the conduct at issue at trial. The district court granted the motion, but left the door open for Sidibe to introduce particular pre-2006 evidence at trial.  SIDIBE V. SUTTER HEALTH 71 Rather than wait for trial, Sidibe moved shortly before trial to admit 23 pre-2006 documents. These 23 documents generally fall into four buckets: Sutter’s Transition to Systemwide Contracting in Late 1990s/Early 2000s. The first bucket covers evidence from Sutter’s transition from individualized contracting to systemwide contracting in the late 1990s to early 2000s. Prior to systemwide contracting, Sutter negotiated contracts with insurers hospital-by-hospital. The bucket includes internal Sutter memoranda from 1998 on how moving to systemwide negotiations could lead to “better pricing” and “better results.” It also includes a 1999 litigation filing showing that Sutter knew that health insurers’ ability to “redirect” patients would constrain prices. Insurers’ Objections to Systemwide Contracting from 1999–2004. The second bucket deals with evidence of the health plans’ objections to Sutter’s shift to systemwide contracting from 1999 to 2004. This bucket includes concerns from insurers that the systemwide contracting would give Sutter “significant market power” and result in insurers paying more. Sutter Contract Amendments from 2002–2005. The third bucket contains contract amendments from 2002 to 2005 with insurers like Blue Shield, Blue Cross, Aetna, and Health Net. According to Sidibe, they contained provisions that tied or prevented steering, like “non-par rate,” “equal treatment,” and “tiered products” provisions. Evidence of Price Increases After Systemwide Contracting. The fourth bucket includes evidence allegedly showing that Sutter’s prices increased after the move to systemwide contracting, such as a graph from Sidibe’s 72 SIDIBE V. SUTTER HEALTH expert depicting a boost in Sutter’s net patient revenue in 2002. Sidibe asserts that this historical evidence would show that Sutter shifted to systemwide contracting in the 1990s to leverage market power and raise prices; that insurers objected the shift but were forced to accept it; and that the shift allowed Sutter to contract for anticompetitive terms with insurers. After a hearing, the district court denied Sidibe’s motion to admit these 23 documents. The case then went to trial. At the end of the month-long trial, the jury returned a complete verdict for Sutter. Sidibe now appeals. In this opinion and dissent, we address Sidibe’s challenges to the district court’s Rule 403 ruling and jury instructions. II. Federal Rule of Evidence 403 Federal Rule of Evidence 403 authorizes district courts to “exclude relevant evidence if its probative value is substantially outweighed by a danger of . . . unfair prejudice, confusing the issues, misleading the jury, undue delay, wasting time, or needlessly presenting cumulative evidence.” Rule 403 “favors admissibility, while concomitantly providing the means of keeping distracting evidence out of the trial.” United States v. Fleming, 215 F.3d 930, 939 (9th Cir. 2000) (simplified). District courts receive “great deference” in applying Rule 403. United States v. Cabrera, 83 F.4th 729, 736 (9th Cir. 2023) (simplified). After all, it’s the district court that lives with a case, understands the realities of the litigation, and manages the trial. That’s why  SIDIBE V. SUTTER HEALTH 73 we affirm Rule 403 rulings unless they’re illogical, implausible, or unsupported by the record. See United States v. Hinkson, 585 F.3d 1247, 1262 (9th Cir. 2009) (en banc). At issue here is whether the district court abused its discretion in setting a cutoff for admissible trial evidence under Rule 403. The district court excluded any evidence predating 2006—five years before the 2011 damages period—as being only “marginally relevant,” “cumulative,” and “confusing.” Given the latitude in placing evidence cutoffs, the reasonable assessment of the historical evidence’s relative relevance, and the Rule 403 concerns, the district court conclusion was no abuse of discretion. And if it was, any error was harmless. A. Evidentiary Cutoffs When excluding evidence, relevance is not the end of the story. Long ago, the Supreme Court acknowledged that district courts enjoy great discretion to fashion a reasonable cutoff date for admitting even relevant evidence. See United States v. Socony-Vacuum Oil Co., 310 U.S. 150, 230 (1940). That’s because “[t]erminal points are necessary . . . [in] trial[s] involving intricate business facts and legal issues.” Id. In Socony-Vacuum Oil, the government charged several oil companies and individuals with conspiring to fix oil markets, which led to a surge in gas prices. Id. at 167. At trial, the defendants sought to admit evidence that other factors caused the rise in gas prices. Id. at 229. This evidence was fresh data, starting from 1932—just a few years before the alleged 1935 conspiracy. Id. Even so, the 74 SIDIBE V. SUTTER HEALTH district court excluded the evidence from the three-and-half- month trial. The Supreme Court affirmed the exclusion of evidence. Though the evidence “was not wholly irrelevant,” the Court held it “was clearly collateral.” Id. at 230. The Court first noted that “the record is replete with evidence showing the condition of the oil industry at the time” and that “ample testimony” was already introduced on the “other causal factors” which the defendants claimed were responsible for the price increase. Id. Thus, the Court ruled that “[m]uch of the refused testimony was merely cumulative in nature.” Id. The Court then recognized the district court’s “wide range for discretion in the exclusion” of evidence of collateral matters. Id. Admitting such evidence “might have confused rather than enlightened the jury” and “would have prolonged the inquiry and protracted the trial.” Id. Practical considerations governed the matter, the Court said. As Justice Holmes colorfully put it, excluding evidence on collateral issues is “a concession to the shortness of life.” Id. (quoting Reeve v. Dennett, 145 Mass. 23, 28 (1887)). So, as the Court ruled, “a new trial will not be ordered for alleged errors in exclusion of evidence where matters of substance are not affected” and in which a “great mass of evidence was received [and] the range of inquiry was wide.” Id. at 231. And all this was in the context of a criminal trial, when a defendant’s right to present a defense is at its highest. The principles of Socony-Vacuum Oil were reaffirmed two decades later. In another antitrust case—this time involving the monopolization of metal ore—the Supreme Court reiterated that district courts may place “reasonable limits upon [clearly material] evidence.” Cont’l Ore Co. v. Union Carbide & Carbon Corp., 370 U.S. 690, 710 (1962).  SIDIBE V. SUTTER HEALTH 75 The Court then endorsed district courts “set[ting] a reasonable cut-off date, evidence before which point is to be considered too remote to have sufficient probative value to justify burdening the record with it.” Id. This is also the law of our circuit. See United States v. Gonsalves, 675 F.2d 1050, 1054 (9th Cir. 1982) (“The trial court has the discretion to exclude stale evidence.” (simplified)). Other circuit courts have also applied this rule to affirm reasonable evidence cutoffs. See Int’l Shoe Mach. Corp. v. United Shoe Mach. Corp., 315 F.2d 449, 460 (1st Cir. 1963) (affirming the district court’s decision to exclude evidence predating the limitations period because of the district court’s authority to “set a reasonable cut-off date” (simplified)); Berkey Photo, Inc. v. Eastman Kodak Co., 603 F.2d 263, 296 (2d Cir. 1979) (allowing evidence from before the limitations period, but warning it wasn’t a license for “antitrust plaintiffs . . . to embark on a . . . time-warped fishing expedition” and emphasizing that “a trial court in its discretion may always set a reasonable cut-off date” (simplified)). All of this follows the district court’s duty under Rule 403 to exclude even relevant evidence when its probative value is substantially outweighed. With this background, we should have easily affirmed the district court’s evidentiary cutoff. The district court understood that Sidibe sought to introduce the evidence to show Sutter’s market power and intent and recognized that the “history of Sutter’s practices is relevant within some reasonable time period” predating 2011. But it questioned “the point of putting in evidence from a time that precedes the relevant class period by not just five or six years, but by more than a decade in some instances.” Indeed, no one disputes that Sutter’s contracts were renegotiated “regularly” and the relevant contracts took effect shortly 76 SIDIBE V. SUTTER HEALTH before 2011. Having evidence that “stretches back too far,” the district court then concluded, may confuse the jury on “[w]hat’s at issue: [t]he contracts during the period [with] challenged . . . anticompetitive [terms].” The district court thus ruled pre-2006 evidence would be cumulative, “too attenuated,” “confuse[] the issues,” “waste[] time,” and “add[] delay.” Given its “marginal relevance,” the district court determined that the pre-2006 evidence did not “merit the mini-trial” that would result from its admission. So the district court understood that the historical evidence was relevant to market power, history, and intent, but that the trial needed some limits. After all, we can’t always admit evidence having any semblance of relevance— no matter how remote, stale, or collateral. On balance, placing a five-year cutoff on admissibility reasonably achieves Sidibe’s objective to show intent and history while mitigating the concerns for jury confusion. And given that the district court had lived with this case for almost a decade, it wasn’t an abuse of discretion in making this call. Contrary to the majority’s contention, the district court’s choice of a five-year cutoff date was not “arbitrary.” As it was explained, Sutter advocated for the 2006 cutoff because that was the year that Sutter negotiated the contracts that set the prices for 2009, the first year in Sidibe’s expert report. So it’s wrong to suggest that the district court simply picked the cutoff date out of thin air. And the majority borders on disrespect to claim that “the district court did not appreciate how the history and purpose of Sutter’s conduct was an essential aspect of [Sidibe’s] legal theory.” Maj. Op. 34. After presiding over the case for almost ten years, it’s safe to assume that the district court understood Sidibe’s legal theories. Indeed, the majority describes the evidence and  SIDIBE V. SUTTER HEALTH 77 legal theories in the same terms as the district court did—so this criticism rings hollow. Of course, choosing a cutoff date will always require some discretionary calls—that’s the nature of the thing. And as with many discretionary decisions, not everyone would make the same choice. The majority clearly wouldn’t. But virtually every deadline or cutoff date is somewhat arbitrary as a necessary part of discretion. See, e.g., Fed. R. Evid. 803(16) advisory committee’s note to 2017 amendment (“The Committee understands that the choice of a cut-off date has a degree of arbitrariness.”); Fed. R. Evid. 901(b)(8) advisory committee’s note to 1972 amendment (“Any time period selected is bound to be arbitrary.”). But that doesn’t mean that district courts can’t draw a line somewhere or that the district court abused its discretion by doing so here. Indeed, the majority offers no insight on how district courts should structure a cutoff date aside from unhelpful guidance that it shouldn’t be “too blunt an instrument.” Maj. Op. 57. But through its analysis, the majority seems to suggest that no cutoff date is acceptable whenever “history” is relevant to the claims at trial. That simply can’t be the rule and the majority provides little justification for this view. And the majority makes no attempt to reconcile its decision with the Supreme Court’s clear guidance that district court’s may “set a reasonable cut-off date” for even relevant evidence. Cont’l Ore Corp., 370 U.S. at 710. So ordering a new trial based on this evidentiary ruling is both unprecedented and wholly unnecessary. And as the following shows, the evidentiary cutoff was entirely appropriate given the pre-2006 evidence’s minimal relevance and its cumulative, confusing, and prejudicial nature. 78 SIDIBE V. SUTTER HEALTH B. Degree of Relevance Relevance is often in the eye of the beholder. At base, the majority disagrees with the district court on the historical evidence’s degree of relevance. The district court viewed the pre-2006 evidence as “minimally relevant” while the majority asserts it’s “highly relevant.” This discretionary call is one we normally leave to district courts. And it’s baffling to understand why we choose this case to impose our views on the degree of relevance. While the majority may have decided differently, the district court’s ruling doesn’t even come close to being “beyond the pale of reasonable justification under the circumstances.” Boyd v. City & Cnty. of San Francisco, 576 F.3d 938, 943 (9th Cir. 2009) (simplified). And the district court wasn’t illogical in drawing the line on the historical evidence. According to Sidibe, the pre- 2006 evidence can be used to establish two things: (1) Sutter’s intent to move to systemwide contracting to achieve better prices and force insurers to pay supracompetitive rates, and (2) the history of the restraint, including Sutter’s historical market power, the inclusion of anticompetitive contract terms, and increased prices after systemwide contracting. For the tying claim, a per se violation, which cares little about past intent or past market conditions, the historical evidence has almost no relevance. And while the evidence is more relevant for the rule-of- reason claim, it doesn’t take the “essential” role that the majority ascribes to it. See Maj. Op. 34–35. Indeed, it was unnecessary for Sidibe to establish Sutter’s late-1990s/early- 2000s purpose or market power to prove its post-2011 anti- steering claim.  SIDIBE V. SUTTER HEALTH 79 1. Relevance to Tying Start with Sidibe’s tying claim. Recall that tying arrangements are illegal per se under California law. UAS Mgmt., Inc., 169 Cal. App. 4th at 368–69. So pernicious is such conduct that they “serve hardly any purpose beyond the suppression of competition.” Suburban Mobile Homes, 101 Cal. App. 3d at 542 (simplified). Thus, we don’t measure the “seller’s justifications for the arrangement,” UAS Mgmt., Inc., 169 Cal. App. 4th at 369 (simplified), or care about the seller’s “precise purpose,” Container Corp. of America, 393 U.S. at 340 (Marshall, J., dissenting). And a seller’s historical purpose is even further afield. The history of the restraint is also irrelevant. Indeed, the very point of the per se rule is to “avoid[] the necessity for an incredibly complicated and prolonged economic investigation into the entire history of the industry involved.” Suburban Mobile Homes, 101 Cal. App. 3d at 541; see also United States v. Joyce, 895 F.3d 673, 677–78 (9th Cir. 2018) (describing the same under federal law). Consistent with this, neither anticompetitive purpose nor history of the restraint is “essential” to establishing tying under California law. See UAS Mgmt., Inc., 169 Cal. App. 4th at 368–69. Instead, assuming a tying arrangement, the focus is on whether Sutter “had sufficient economic power in the tying market to coerce the purchase of the tied product.” Id. (simplified). As was explained, “where the seller has no control or dominance over the tying product so that it does not represent an effectual weapon to pressure buyers into taking the tied item any restraint of trade attributable to such tying arrangements would obviously be 80 SIDIBE V. SUTTER HEALTH insignificant at most.” N. Pac. Ry. Co. v. United States, 356 U.S. 1, 6 (1958). So history of the restraint has nothing to offer for the tying claim here. While Sidibe argues that pre-2006 history can show Sutter’s historical market power, such ability to coerce insurers in the distant past is irrelevant. “[T]he inquiry is whether the defendant plays enough of a role in the relevant market to significantly impair competition.” Ben-E-Lect v. Anthem Blue Cross Life & Health Ins. Co., 51 Cal. App. 5th 867, 875–76 (2020) (emphasis added) (simplified). So only evidence of market power shortly before 2011—not in the late 1990s/early 2000s—could help establish the relevant market. Indeed, if Sidibe could establish Sutter’s market power in the timeframe before 2011 as required under California law, why would market conditions from the early 2000s be probative? So it’s a mistake to think, as the majority does, that decade’s old market power (from as far back as 1997) can prove Sutter’s market power in the runup to 2011. See Maj. Op. 41–42. That’s like looking to the market for dialup modems in 2000 to prove today’s WiFi market. Sure, it’s interesting history, but it doesn’t tell you that much. The same goes for pre-2006 evidence of Sutter’s anticompetitive purpose. Establishing an anticompetitive purpose is not “essential” if Sutter sought a tying arrangement or condition during the relevant period. See UAS Mgmt., Inc., 169 Cal. App. 4th at 368–69; Cascade Cabinet Co., 710 F.2d at 1372. If Sidibe could prove that Sutter intentionally sought to make a tying arrangement in the runup to the 2011 period, it doesn’t matter whether Sutter also did it for supracompetitive gain. It’s a per se violation. Recall, for such violations, “anti-competitive purpose” only requires proof that “the defendants intentionally engaged in  SIDIBE V. SUTTER HEALTH 81 conduct which, if carried out as planned, would always or almost always adversely affect competition.” Cascade Cabinet Co., 710 F.2d at 1372. In other words, Sidibe doesn’t have to prove any separate anticompetitive intent apart from the purpose of arranging a tying scheme. So whether Sutter intended to extract better prices from insurers through systemwide contracting back in the late 1990s is not essential to the per se violation. In deeming this historical evidence “essential” for proving purpose, the majority makes two errors here. First, it conflates systemwide contracting with tying. The majority seems to think that having the purpose to engage in systemwide contracting is the same thing as having the purpose to tie hospitals. But negotiating for hospitals as a group is very different than coercing insurers to buy tied medical services. So, contrary to the majority’s belief, knowing what Sutter “intended to do” back when it switched to systemwide contracting doesn’t help “determine what [Sutter] actually did” during the relevant period for Sidibe’s tying allegation. Maj. Op. 36. Second, the majority mixes historical purpose with contemporary purpose. Just because a seller had a purpose in the past doesn’t mean a seller had the same purpose 20 years later. This is especially true when the conduct we are comparing (systemwide contracting v. tying) is so different. This all shows how tenuous this supposed purpose evidence is. 2. Relevance for Steering That leaves the anti-steering claim. Anti-steering provisions aren’t per se antitrust violations and so fall under the rule of reason. See Chavez, 93 Cal. App. 4th at 369. The rule of reason requires balancing the restraint’s economic 82 SIDIBE V. SUTTER HEALTH effects with its possible procompetitive justifications. See Marin Cnty. Bd. of Realtors, Inc. v. Palsson, 16 Cal. 3d 920, 934–35 (1976). We consider several factors here including the “restraint’s history, nature, and effect.” Leegin Creative Leather Prods., Inc. v. PSKS, Inc., 551 U.S. 877, 885 (2007) (simplified). Sidibe argues that the historical evidence would show how Sutter’s shift to systemwide contracting was the lynchpin to forcing insurers to accept the three “anti- steering” clauses. So the evidence, Sidibe says, was relevant for two questions: (1) whether Sutter “forced” the health plans to sign contracts, and (2) whether those contracts contained terms that prevented insurers from steering patients to lower-cost, non-Sutter hospitals. While Sidibe is correct this evidence is relevant, the district court wasn’t wrong to say the pre-2006 evidence was only marginally so. Take each question separately. Forcing. No one disputes that the pre-2006 evidence shows that Sutter coerced insurers into systemwide contracting over their objection. So it reflects Sutter’s ability to flex its market power back in the late 1990s and early 2000s. The question is whether this evidence is “marginally relevant,” as the district court found, or “essential,” as the majority believes. It’s no abuse of discretion to group this evidence in the former category. Why? Because it’s also uncontested that the contracts were renegotiated frequently and that the contracts for the relevant damages period were negotiated shortly before 2011. So Sidibe had to show Sutter’s ability to force health plans almost a decade after much of the pre-2006 evidence. While giving some context, it’s really more propensity evidence—if Sutter forced contracts in the early 2000s, it must be forcing contracts  SIDIBE V. SUTTER HEALTH 83 today. But that’s not the proper use of historical evidence. See Fed. R. Evid. 404(b). Anticompetitive Terms. The question about anticompetitive terms is even easier. Here, Sidibe had to prove Sutter included the anticompetitive terms—“non-par rate,” “equal treatment,” and “tiered products” clauses—in contracts that caused damages after 2011. Once again, the relevant contracts were signed shortly before 2011. So the historical evidence, which includes contracts from the early 2000s, simply doesn’t speak to the contract terms in effect during the damages period. If Sidibe could prove the three anticompetitive terms existed in the contracts impacting 2011 and after, why would their existence in 2002 or 2003 be probative at all—let alone highly probative? Either the circa-2011 contracts contained the terms or not. Circa-2000 contracts won’t move the needle on that question. *** So even under the lax standard for relevance, the historical evidence isn’t “essential” to Sidibe’s claims, as the majority contends. And it wasn’t an abuse of discretion to find the pre-2006 evidence only marginally relevant. C. Rule 403 Balancing As stated above, relevance isn’t the end of the evidentiary inquiry. Even with highly relevant evidence, the district court must balance its probative value against the other Rule 403 factors—“unfair prejudice, confusing the issues, misleading the jury, undue delay, wasting time, or needlessly presenting cumulative evidence.” Fed. R. Evid. 403. Here, the district court didn’t abuse its discretion in finding the probative value of the historical evidence 84 SIDIBE V. SUTTER HEALTH substantially outweighed by the Rule 403 concerns. In this case, the historical evidence was cumulative, risked jury confusion, and could lead to unfair prejudice. 1. Cumulative Evidence Rule 403 allows for the exclusion of evidence that is needlessly cumulative of other evidence available to the jury. That’s the case here. To illustrate why, let’s consider the four basic points Sidibe offered the evidence to prove: (1) Sutter shifted from hospital-by-hospital contracting to systemwide contracting; (2) the insurers opposed the shift to systemwide contracting but Sutter forced insurers to accept it; (3) Sutter was motivated by profit to shift to systemwide contracting; and (4) the insurers agreed to systemwide contracts that contained anticompetitive terms, like anti- steering provisions. The trial contained evidence for each of these points. Shift to Systemwide Contracting. Sidibe thought it important to show the jury the “before and after” of Sutter’s contracting practices. But the jury heard such evidence. Take the testimony from a former Blue Shield employee, Kristen Miranda. Miranda testified at trial that “Sutter had many hospitals” and “up until a certain point in time, those contracts were really negotiated sort of separately.” This is the “before”—when there were individualized negotiations. Miranda didn’t stop there. She went on to describe how, “[a]t some point -- I think it may have predated the 2007 systemwide agreement--. . . Sutter began to require that the entire system would be negotiated as part of a single -- a single contract and a single negotiation.” This is the “after”—when Sutter shifted to systemwide contacting.  SIDIBE V. SUTTER HEALTH 85 Thus, systemwide contracting was extensively covered at trial. At closing, even Sidibe’s counsel admitted that jurors “heard an exhaustive amount” about the “systemwide contracts that Sutter has imposed on the health plans.” Forcing. Next, Sidibe wanted to show that insurers objected to the shift to systemwide contracting, which demonstrates Sutter’s ability to force insurers to accept contracts. Again, that evidence already came out at trial. For example, Miranda was asked, “did Blue Shield want to contract with Sutter on a systemwide basis?” Her answer, “That would not have been our preference.” Sidibe’s counsel followed up, “Did Sutter . . . insist that the contract be on a systemwide basis?” Miranda responded, “Yes.” At this point, Sutter’s counsel objected and tried to exclude the evidence as “pre-2006.” The district court overruled the objection. Miranda then said that Sutter refused to negotiate on an individual basis and “made it very clear during this time that the only option for negotiating a contract with Sutter was on a systemwide basis.” When asked if Blue Shield would have wanted a contract term requiring the insurer to “take all of [Sutter’s] hospitals or none,” Miranda responded, “Goodness, no.” But that’s not all. Aetna’s former contract negotiator, Chandra Welsh, told the jury something similar. Sidibe’s counsel asked her why Aetna agreed to the systemwide contract even though the contract contained terms Aetna didn’t like. Welsh explained that “in order to get a -- a contract done and keep the providers in our network, Sutter insisted. They refused to take it out. So we ultimately had to sign it.” Sidibe also extracted the same evidence from the horse’s mouth—Robert Reed, Sutter’s former CFO. Reed conceded 86 SIDIBE V. SUTTER HEALTH at trial that multiple health insurers objected to the systemwide contracting. Reed testified that Blue Shield, Aetna, and Health Net all objected to Sutter’s systemwide contracting. So Blue Shield and Aetna’s former employees informed the jury that the health insurers felt compelled to negotiate on a systemwide basis. And Sutter confirmed it. Any other evidence of forcing would simply be cumulative. Profit Motive. Moving on to Sutter’s alleged purpose to exact supracompetitive prices through systemwide contracting. Trial evidence readily fleshed this out. Once again, Blue Shield’s Miranda’s testimony is one data point. When asked “why Sutter insisted on systemwide contracting,” Miranda said “that it was to put them in the best possible position to negotiate the most favorable terms for Sutter.” Miranda added that Sutter’s prices were “materially higher-cost relative to other providers in northern California.” Even more probative, Reed, Sutter’s former CFO, disclosed the same thing. When asked whether “it [is] true that you believe that getting all the Sutter hospitals to act in a cohesive fashion would result in a better outcome for the group than for any of the individual hospital affiliates,” Reed admitted, “I think that’s true.” Sidibe’s counsel didn’t stop there. He then asked, “[I]n your view, systemwide contracting resulted in better results for the group hospitals than any individual hospitals between 2006 and the present; correct?” Once again, Reed admitted “that’s true.” So it’s clear that the jury heard from witnesses who explained Sutter’s profit motivations and the higher prices associated with systemwide contracting.  SIDIBE V. SUTTER HEALTH 87 Anticompetitive Terms. That leaves showing anti- steering provisions. Sidibe contends that the pre-2006 contracts contained “equal treatment,” “tiered products,” and “non-par rate” clauses. But the post-2006 contracts contain these terms as well. Indeed, Sidibe needed to show that these terms existed for the contracting impacting the post-2011 damages period. Each of these terms was extensively discussed at trial. Take former Aetna employee Welsh’s testimony. She testified about all three. She explained that Sutter drafted the “non-par” clause, that Aetna did not want the provision, and that Aetna objected to it. Same with the “equal treatment” and “tiered products” clauses. Welsh said that Aetna would have preferred to remove the two clauses from their Sutter contracts, but Sutter refused to do so. Considering these examples, admitting pre- 2006 contracts with similar terms would’ve been purely cumulative. 2. Jury Confusion Excluding the historical evidence also wasn’t an abuse of discretion given the risk of confusing or misleading the jury. Let’s focus on three points here. First, the historical evidence may have confused the jury with its targeting of systemwide contracting. Sidibe wanted to introduce voluminous evidence of Sutter’s shift to systemwide contracting, which Sidibe portrayed as the precursor to charging supracompetitive rates. So a great deal of the pre-2006 evidence is designed to give the jury the impression that there was something inherently nefarious about systemwide contracting. But systemwide contracting isn’t illegal, and Sutter’s use of the practice doesn’t amount to an antitrust violation. Put another way, the jury didn’t 88 SIDIBE V. SUTTER HEALTH have to answer whether Sutter’s systemwide contracting practices had anticompetitive effects on the market. It had to answer whether Sutter’s supposed tying or anti-steering in systemwide contracts had those effects. Focusing so much on systemwide contracting by itself may lead the jury into confusing those questions. For example, it may confuse the jury into thinking that systemwide contracting is the same thing as tying—the same mistake the majority seemingly makes. Second, the historical evidence may have confused the jury on the relevant timeframe. Why have so much evidence of contracts from outside the damages period? The trial evidence already included several post-2006 contracts— contracts that actually affected the damages period. The pre- 2006 contracts Sidibe wanted to admit would only confuse things. Take one Sutter/Blue Shield contract Sidibe wanted to introduce. It was entered into and became effective on January 1, 2002. Almost ten years before the damages period! Or a 2004 email containing redlines to the contract that became effective on January 1, 2005. It’s 80 pages long! Why make the jury sift through that? Subjecting the jury to voluminous evidence on similar contracts may confuse jurors about which contracts mattered for the damages period. Finally, the historical evidence may have confused the jury by requiring a minitrial. Much of the historical evidence would have taken significant resources to admit. Look at the evidence from the 1999 litigation over the merger of Alta Bates Medical Center and Sutter’s Summit Hospital. Sidibe sought to introduce Sutter’s proposed findings of fact from the litigation. Sidibe claims this evidence shows that Sutter understood that insurers’ ability to steer patients away from its hospitals was a means of checking price increases. But  SIDIBE V. SUTTER HEALTH 89 Sidibe needed two witnesses just to admit the proposed findings—one witness to lay the foundation for the document and another witness to presumably testify to what it says. And Sutter surely would have challenged Sidibe’s version of events, requiring even more witnesses. So the likelihood was high that the jury would have been bogged down in a minitrial on collateral issues. That’s a core 403 concern. 3. Unfair Prejudice Lastly, consider the issue of unfair prejudice. Of course, evidence is, more often than not, prejudicial to some degree. Prejudice becomes unfair, however, when it “appeals to the jury’s sympathies, arouses its sense of horror, provokes its instinct to punish, or otherwise may cause a jury to base its decision on something other than the established propositions in the case.” United States v. Skillman, 922 F.2d 1370, 1374 (9th Cir. 1990) (simplified). So the question is whether the pre-2006 evidence might have caused the jury to reach a conclusion for an improper reason. All of Sutter’s actions from the pre-2006 evidence predate the damages period—meaning they’re nonactionable. But some of the evidence from this period is the most direct and damning. Such as Health Net’s objection to Sutter’s “equal treatment” clause from 2003—almost a decade before the start of the damages period. Health Net complained that the clause could “suppress competition and [the] exercise of informed choice on the cost and quality of services in the marketplace.” It also “raises concerns about antitrust abuse to not only insist on high levels of reimbursement, but to then also seek to frustrate the marketplace’s judgement on those prices, as reflected in 90 SIDIBE V. SUTTER HEALTH physician and patient choices, not imposed by Health Net.” Finally, Health Net asserted, “[i]t is not only impractical for Health Net to agree to language containing even a suggestion of such an obligation, but language that would risk achieving such an effect would pose an unreasonable constraint on the operations of the market itself.” In essence, Health Net directly accused Sutter of violating antitrust laws back in 2003. It’s easy to see how admitting such inflammatory accusations could encourage the jury “to base its decision on something other than the established propositions in the case,” and punish Sutter for events not at issue in the trial. Skillman, 922 F.2d at 1374 (simplified). A letter like Health Net’s may serve to elicit an instinctual response from the jury based on the use of buzzwords like “antitrust abuse” and “constrain the marketplace.” But even if Sutter committed antitrust violations in 2003, admitting such evidence risks punishing Sutter for nonactionable conduct. That’s unfair prejudice. So making the trial about events so far in the past would’ve invited the jury to base its decision on events that were distinct from the conduct at issue at trial. *** Given the cumulative nature of the evidence, risks of jury confusion, and unfair prejudice, the district court’s ruling does not even approach an abuse of discretion. D. Harmless Error But even if the district court got the Rule 403 balance wrong, we must ask—where’s the prejudice? See City of Long Beach v. Standard Oil Co. of Cal., 46 F.3d 929, 936 (9th Cir. 1995) (“Reversal will not be granted unless  SIDIBE V. SUTTER HEALTH 91 prejudice is shown.” (simplified)). To establish prejudice, the district court’s error must have “more probably than not tainted the verdict.” Crawford v. City of Bakersfield, 944 F.3d 1070, 1077 (9th Cir. 2019) (simplified). In other words, Sutter—as the benefitted party—needs to show that “it is more probable than not that the jury would have reached the same verdict even if the evidence had been admitted.” Obrey v. Johnson, 400 F.3d 691, 701 (9th Cir. 2005). There’s simply no prejudice for the tying claim. The verdict form shows that the jury found that Sidibe failed to prove the threshold question—that Sutter tied its hospitals during the damages period. The jury stopped at that point and didn’t reach any other question on the tying claim. Because pre-2006 evidence couldn’t prove any tying for the post-2011 period, there’s no chance it tainted the jury verdict for the tying claim. The exclusion of the historical evidence was also likely harmless for the anti-steering claim. Once again, the verdict form shows that the jury found that Sidibe failed to prove the threshold question—that Sutter forced the class health plans to agree to contracts with terms preventing the plans from steering patients to lower-cost, non-Sutter hospitals during the damages period. So the jury had to consider two things here: (1) whether Sutter forced the health plans to sign agreements, and (2) whether those agreements prevented health plans from steering patients to lower-cost options. The pre-2006 evidence would have no impact on the jury’s consideration of the steering question. Given that the jury had to find anti-steering provisions in contracts impacting the post-2011 world (which would be a matter of reviewing those contracts), any pre-2006 contracts would be irrelevant. 92 SIDIBE V. SUTTER HEALTH While a closer call, it’s more probable than not that the exclusion of the historical evidence didn’t taint the jury verdict on the forcing question. Sidibe argues that the pre- 2006 evidence could have made a difference in showing Sutter’s ability to force the insurers to enter agreements. That’s true, but, as described above, the cumulative nature of the forcing evidence minimizes any chance of prejudice from exclusion here. See City of Long Beach, 46 F.3d at 937 (holding no prejudicial error when excluded evidence was “at best, cumulative since the record reveals that the plaintiffs presented numerous other pieces of evidence designed to show” defendant’s knowledge). It’s hard to imagine the jury changing its mind simply by hearing more of the same evidence—but this time from an older, more remote time. *** Given everything above, there’s no basis to overturn this four-week trial and the jury’s verdict based on an evidentiary ruling left to the discretion of the district court. III. Jury Instructions As problematic as the majority’s Rule 403 ruling is, its jury instructions ruling may be even more so. Here, the majority reverses the jury’s verdict on the rule-of-reason claim based on its view of a faulty jury instruction. But there’s no cause to do so. First, the majority adopts a questionable new legal rule for antitrust law to reach this result. In doing so, the majority imposes its novel rule on both California and federal law. Second, any instructional error was harmless. See BladeRoom Grp. Ltd. v. Emerson Elec. Co., 20 F.4th 1231, 1243 (9th Cir. 2021) (explaining  SIDIBE V. SUTTER HEALTH 93 that an instructional error is harmless if it’s “more probable than not that the jury would have reached the same verdict had it been properly instructed” (simplified)). Thus, we should have affirmed this ground as well. A. Rule-of-Reason Instruction The parties dispute the jury instruction used for the anti- steering claim. They disagree on whether juries must be instructed to consider the defendant’s anticompetitive “purpose or effect” as an element of the claim or whether consideration of anticompetitive “effect” is enough. Under the California pattern jury instructions, the second element of a rule-of-reason claim provides that the plaintiff must prove: “That the purpose or effect of [name of defendant]’s conduct was to restrain competition.” Judicial Council of Cal., Civil Jury Instructions No. 3405 (2023 ed.). Sidibe advocated for the pattern instruction. Sutter instead asked that the instruction be modified so that the plaintiff must prove only: “That the effect of Sutter’s conduct was to restrain competition.” So the major difference between the parties was the elimination of anticompetitive “purpose” as satisfying the second element of the claim. The district court sided with Sutter and required a finding of an anticompetitive “effect” as an element. But the district court left the instruction that the jury “should consider the results that [Sutter’s] restraint was intended to achieve” as part of its balancing analysis. The majority reverses. In doing so, the majority breaks new legal ground and announces a new rule for all antitrust law—not for just California, but for the Nation. According to the majority, “consideration of anticompetitive purpose is 94 SIDIBE V. SUTTER HEALTH an essential aspect of the rule of reason analysis under both the Cartwright Act and the Sherman Act” and “failing to instruct the jury to consider purpose misstates the law.” Maj. Op. 21. Although unclear what is meant by the term “essential aspect” here, if the majority is saying that consideration of “anticompetitive purpose” is now an element of all rule-of-reason claims and it’s reversible error to omit it from jury instructions, that is simply wrong. Neither California law nor federal law supports this. The California Supreme Court has never described anticompetitive purpose as an element that must be considered when evaluating a rule-of-reason claim. In fact, the California Supreme Court has only ever described anticompetitive purpose as one of several factors a jury may consider when assessing such a claim. See, e.g., In re Cipro, 61 Cal. 4th at 146 (“[A] court may consider the facts peculiar to the business in which the restraint is applied, the nature of the restraint and its effects, and the history of the restraint and the reasons for its adoption.” (emphasis added) (simplified)). This isn’t surprising because rule-of-reason claims are inherently case or market specific. See Flagship Theatres, 55 Cal. App. 5th at 400. And the majority’s support for its broad view of federal law falls short. The majority relies on two Supreme Court cases for its proposition under the Sherman Act: Chicago Board of Trade v. United States, 246 U.S. 231 (1918) and United States v. Topco Associates, Inc., 405 U.S. 596 (1972). Contrary to the majority’s contention, neither Supreme Court case said that consideration of anticompetitive purpose is essential to a rule-of-reason claim, requiring inclusion in its elements or jury instructions. Instead, in Chicago Board of Trade, the Court listed several considerations including “purpose or end sought to be  SIDIBE V. SUTTER HEALTH 95 attained,” viewing them “all [as] relevant facts.” Chi. Bd. of Trade, 246 U.S. at 238. No relevant fact, like “the evil believed to exist” or the “reason for adopting the particular remedy,” id., was given primacy, and the majority here doesn’t claim that these considerations must also be included in jury instructions as essential to antitrust claims. Consider as well that the Chicago Board of Trade Court reasoned that “intention” alone is not dispositive, but “knowledge of intent may help the court to interpret facts and to predict consequences.” Id. Topco Associates holds the same. There, the Court said that purpose is “include[d]” as a consideration, but it didn’t make it mandatory. See Topco Assocs., 405 U.S. at 607 (“An analysis of the reasonableness of particular restraints includes consideration of the facts peculiar to the business in which the restraint is applied, the nature of the restraint and its effects, and the history of the restraint and the reasons for its adoption.”). The majority’s mandatory rule also conflicts with other Supreme Court precedent. Rule-of-reason claims require “weigh[ing] all of the circumstances of a case.” Leegin Creative Leather, 551 U.S. at 885 (simplified). It has never been said that one factor is dispositive or required. Indeed, Leegin lists several important factors to consider— “information about the relevant business,” “market power,” and “the restraint’s history, nature, and effect.” Id. (simplified). Notably, it doesn’t include “anticompetitive purpose,” undermining the majority’s claims that its consideration is “essential.” In short, the majority can point to no Supreme Court case holding that anticompetitive purpose must be considered in every rule-of-reason case. The reason for that is simple. No such rule exists and adopting it contravenes a century’s worth of cases saying otherwise. 96 SIDIBE V. SUTTER HEALTH So based on the shakiest of foundations, the majority crafts a new rule that alters the direction of antitrust law. B. Harmless Error But even if the majority were correct, any instructional error was harmless. First, while an anticompetitive purpose is not a required element, an anticompetitive effect is. So the jury needed to find an anticompetitive effect no matter what. See, e.g., Marsh v. Anesthesia Servs. Med. Grp., Inc., 200 Cal. App. 4th 480, 495 (2011) (explaining that “where an antitrust plaintiff alleges vertical restraints, facts must be pled showing some anticompetitive effect in the larger, interbrand market” (simplified)). So anticompetitive purpose alone cannot sustain a Cartwright Act claim. Even if the instructions failed to mention purpose, it would make no difference—the jury still had to find an anticompetitive effect regardless of any anticompetitive purpose. Second, the jury instructions still directed the jury to consider anticompetitive purpose. Nothing in the jury instructions precluded the jury from considering anticompetitive purpose. In fact, on the very same page of the jury instructions, the district court told the jury that it “should” consider what Sutter’s conduct was “intended to achieve”—which is the same thing as purpose. So if it were essential for the jury to consider purpose, it was instructed to do so. Yet, the majority ignores this fact and proclaims that “the jury was not instructed that it could consider anticompetitive purpose.” Maj. Op. 22. Finally, and most importantly, the jury never even got close to analyzing whether Sutter’s conduct had an  SIDIBE V. SUTTER HEALTH 97 anticompetitive effect or purpose because the jury found that Sidibe failed to establish that Sutter prevented insurers from steering its patients. Consider the verdict form: Based on the jury’s finding that Sutter didn’t stop insurers from steering their patients, that was the end of its factfinding mission and it never went on to the other elements. No consideration of effects or purpose. In short, regardless of any error in Question 6, it’s harmless because Sutter couldn’t be held liable without a “yes” on Question 5. And the jury said “no.” This is thus the classic case of harmless error. 98 SIDIBE V. SUTTER HEALTH Perhaps recognizing the simple reality that any instructional error was harmless, the majority retreats to claiming that the jury instructions and the pre-2006 evidence together amount to “cumulative error.” Maj. Op. 25–26. But as shown above, no error occurred here. “Zero” plus “zero” equals “zero.” So cumulative error has no place here. IV. For these reasons, we should have affirmed the jury verdict here. I respectfully dissent.