Opinion
14 C 1437
06-22-2021
GREGORY GREENE, individually and on behalf of all others similarly situated, Plaintiff, v. MARK KARPELES, Defendant.
MEMORANDUM OPINION AND ORDER
The court's prior opinions, familiarity with which is assumed, summarize the history of this suit, which concerns the collapse of the Mt. Gox bitcoin exchange. Docs. 199-200 (reported at 169 F. Supp. 3d 855 (N.D. Ill. 2016)); Docs. 229-230 (reported at 206 F. Supp. 3d 1362 (N.D. Ill. 2016)); Docs. 311-312 (reported at 289 F. Supp. 3d 870 (N.D. Ill. 2017)); Docs. 373-374 (reported at 327 F.R.D. 190 (N.D. Ill. 2018)); Docs. 409-411 (reported at 2019 WL 1125796 (N.D. Ill. Mar. 12, 2019)); Docs. 489-490 (reported at 2020 WL 3250715 (N.D. Ill. June 16, 2020). As matters now stand, Gregory Greene, the sole remaining plaintiff, seeks on behalf of a putative class to hold Mark Karpeles, Mt. Gox's principal and the sole remaining defendant, liable under a common law fraud theory for financial losses arising from Mt. Gox's collapse. Docs. 245, 311, 428, 464. The court has denied Karpeles's motion for summary judgment on that claim. 2020 WL 3250715, at *5. Before the court is Greene's motion to certify the class, Doc. 508, and Karpeles's motions to exclude the opinions of Greene's expert, Professor Carol Goforth, Doc. 501, and to strike two exhibits that Greene attached to his class certification motion, Doc. 517. Greene's class certification motion is denied, and Karpeles's motions are denied without prejudice as moot.
Background
"Unlike a motion under Federal Rule of Civil Procedure 12(b)(6), a motion to certify a class under Rule 23(c) is not one for which the court may simply assume the truth of the matters as asserted by the plaintiff. Instead, if there are material factual disputes, the court must receive evidence and resolve the disputes before deciding whether to certify the class." Priddy v. Health Care Serv. Corp., 870 F.3d 657, 660 (7th Cir. 2017) (citation, alterations, and internal quotation marks omitted). Still, "[i]n conducting this analysis, the court should not turn the class certification proceedings into a dress rehearsal for the trial on the merits." Messner v. Northshore Univ. HealthSystem, 669 F.3d 802, 811 (7th Cir. 2012).
Karpeles was the CEO and primary owner of Mt. Gox from early 2011 until the exchange went offline in 2014. Doc. 466 at pp. 2-3, ¶¶ 4-7; id. at p. 8, ¶ 19. Users could create a Mt. Gox account and use it to buy, sell, and store bitcoins and fiat currency. Id. at p. 3, ¶ 8. Greene was one such user, purchasing bitcoin on the exchange on January 5, 2012 and March 4, 2013. Docs. 526-7, 526-6.
On or about January 20, 2012, Mt. Gox for the first time posted Terms of Use on its website. Doc. 510-1 at 14 (94:15-20); Doc. 467-14 at ¶ 2.c. Karpeles directed the drafting and dissemination of the Terms. Doc. 510-1 at 14-15 (96:14-97:18). Karpeles admits that any user who created a Mt. Gox account after the Terms were posted was required to agree to them. Doc. 510-2 at 5. He disputes, however, Greene's assertion that users who created an account before January 20, 2012 also had to accept the Terms as a condition of their continued use of the exchange. Doc. 509 at 9; Doc. 526 at 13. The evidence supports Greene's view. First, the Terms themselves stated: "Please read these Terms carefully and do not use the Site or the Platform unless you accept them." Doc. 510-3 at 2. That language at least purported to bind all existing users who continued to use Mt. Gox after January 20, 2012. Second, Karpeles himself confirmed that the Terms were "not optional for exchange users" and "applied to all users who agree[d] to use the website." Doc. 510-1 at 15 (99:10-17).
That is not to say that any user had to read the Terms of Use. True enough, Greene asserts that users were required not only to abide by the Terms, but also to read them. Doc. 509 at 9. Yet the materials Greene cites—Karpeles's deposition testimony and one of Karpeles's interrogatory responses—do not support that assertion. Karpeles testified only that users were "bound" by the Terms, Doc. 510-1 at 14 (95:9), and his interrogatory response states only that users were "required to ... agree to" the Terms, Doc. 510-2 at 5. A contract can bind a person whether or not the person reads it. See Hill v. Gateway 2000, Inc., 105 F.3d 1147, 1148 (7th Cir. 1997) ("A contract need not be read to be effective; people who accept take the risk that the unread terms may in retrospect prove unwelcome."). So the fact that all Mt. Gox users had to accept the Terms does not imply that they read them or knew their contents.
Greene did testify that he himself read the Terms of Use, Doc. 510-11 at 8 (17:18-22), but there is no evidence that all or most—or even the majority of—Mt. Gox users did so. To the contrary, both Greene's expert, Goforth, and Karpeles's expert, Professor David Pelleg, testified that many users likely did not read the Terms. When asked at her deposition whether "everybody who signed up to use the Mt. Gox exchange after January of 2012 read all of the Terms of Use," Goforth replied, "[a]bsolutely not." Doc. 526-8 at 3 (138:14-17). Pelleg agreed, testifying that "the vast majority [of Mt. Gox users] would not have read the Terms of Use." Doc. 526-9 at 5-6 (173:24-174:5).
The Terms of Use stated that Mt. Gox could be used to "buy and sell" bitcoins and that the exchange "allow[ed] all registered users of the Platform ('Members') to transfer funds and Bitcoins to other Members and to use Bitcoins for purchasing specific items." Doc. 510-3 at 2. The Terms defined "Members" as "Buyers and Sellers as well as any holder of an Account." Ibid. "Buyers" and "Sellers" were defined—in somewhat circular fashion—to mean "Member(s) that are submitting an offer" to buy or sell bitcoins on Mt. Gox. Ibid.
Under the heading "Mt. Gox's Obligations," the Terms of Use stated that members "agree that, when completing Transactions, they are trading with other Members, and Members accept that MtGox acts only as an intermediary in such Transactions." Id. at 4. Mt. Gox promised in the Terms to "hold all monetary sums and all Bitcoins deposited by each Member in its Account, in that Member's name as registered in their Account details, and on such Member's behalf." Ibid. In a September 2012 forum post, apparently referring to that provision of the Terms, Karpeles stated that, "[a]s described in our Terms of Service, customer funds are kept in full, and none are loaned." Doc. 510-4 at 2. Mt. Gox also promised in the Terms to "use all reasonable care and skill in facilitating" bitcoin trades. Doc. 510-3 at 4.
Under the heading "Members' Obligations," the Terms of Use required members to warrant that fiat currencies deposited to buy bitcoins "are actual currencies corresponding to actual assets in [the member's] bank account and coming from legal sources" and that "Bitcoins offered to sell or to transfer correspond to actual Bitcoins." Ibid. Accounts that violated the Terms were subject to suspension. Ibid.
Mt. Gox lost large sums of bitcoins and fiat currency in a series of hacks in 2011. In January, $50,000 was stolen from a Mt. Gox account. Doc. 476 at ¶ 10. In March, an individual or individuals gained unauthorized access to Mt. Gox and stole 80,000 to 90,000 bitcoins. Id. at ¶ 13. In June, there was another hack of Mt. Gox, which caused the price of bitcoin to plummet and additional bitcoins to be lost. Id. at ¶ 20; Doc. 476-3 at 2-3; Doc. 510-7. After the March 2011 hack, Mt. Gox no longer had enough bitcoins to cover all users' deposits, and Karpeles was not sure that Mt. Gox would be able to satisfy users' withdrawal requests. Doc. 476 at ¶¶ 14, 27; Doc. 510-1 at 6-7 (52:8-54:8), 9-10 (72:7-73:22), 16 (105:5-15).
Karpeles used a script called the "Gox Bot" to automatically conduct trades on the exchange. Doc. 476 at ¶ 26. At its inception, the Gox Bot purchased bitcoins using the commissions that Mt. Gox collected on users' trades, but in 2011 Karpeles started having the Gox Bot execute trades using "debt for Mt. Gox." Id. at ¶¶ 26, 28-30. The parties dispute whether it is fair to say that those purchases were made with "fake funds," as Greene calls the alleged "debt." Id. at ¶ 29. It is undisputed, however, that the Gox Bot bought and sold bitcoins using funds that Karpeles manually credited to the account upon which the Gox Bot drew to conduct purchases. Doc. 510-1 at 11-12 (77:20-82:3). Karpeles was convicted in Japan of criminal data manipulation due to his conduct relating to the Gox Bot. Doc. 476 at ¶ 35.
On February 7, 2014, Mt. Gox suspended all bitcoin withdrawals. Doc. 466 at p. 8, ¶ 18. On February 24, the Mt. Gox website went offline. Id. at p. 8, ¶ 19. At that time, Greene had 42.97104 bitcoins and $2.60 in his Mt. Gox account. Id. at p. 14, ¶ 39.
Discussion
The court's analysis of class certification "is not free-form, but rather has been carefully scripted by the Federal Rules of Civil Procedure." Chi. Teachers Union, Local No. 1 v. Bd. of Educ., 797 F.3d 426, 433 (7th Cir. 2015). To be certified, a proposed class must satisfy the four requirements of Rule 23(a): "(1) the class is so numerous that joinder of all members is impracticable; (2) there are questions of law or fact common to the class; (3) the claims or defenses of the representative parties are typical of the claims or defenses of the class; and (4) the representative parties will fairly and adequately protect the interests of the class." Fed. R. Civ. P. 23(a); see Bell v. PNC Bank, N.A., 800 F.3d 360, 373 (7th Cir. 2015). The proposed class also must fall within one of the three categories in Rule 23(b), which the Seventh Circuit has described as: "(1) a mandatory class action (either because of the risk of incompatible standards for the party opposing the class or because of the risk that the class adjudication would, as a practical matter, either dispose of the claims of non-parties or substantially impair their interests), (2) an action seeking final injunctive or declaratory relief, or (3) a case in which the common questions predominate and class treatment is superior." Spano v. Boeing Co., 633 F.3d 574, 583 (7th Cir. 2011); see also Bell, 800 F.3d at 373. Finally, the class must be "identifiable as a class," meaning that the "class definition[] must be definite enough that the class can be ascertained." Oshana v. Coca-Cola Co., 472 F.3d 506, 513 (7th Cir. 2006); see also Mullins v. Direct Digit., LLC, 795 F.3d 654, 659-61 (7th Cir. 2015).
Greene bears the burden of showing that each requirement is satisfied. See Priddy, 870 F.3d at 660; Chi. Teachers Union, 797 F.3d at 433; Messner, 669 F.3d at 811. As the Seventh Circuit has explained, "a district court must make whatever factual and legal inquiries are necessary to ensure that requirements for class certification are satisfied before deciding whether a class should be certified, even if those considerations overlap the merits of the case." Am. Honda Motor Co. v. Allen, 600 F.3d 813, 815 (7th Cir. 2010); see also Kartman v. State Farm Mut. Auto. Ins. Co., 634 F.3d 883, 889-90 & n.6 (7th Cir. 2011). The Seventh Circuit has instructed district courts to exercise "caution" before certifying a class. Thorogood v. Sears, Roebuck & Co., 547 F.3d 742, 746 (7th Cir. 2008).
Greene seeks to certify under Rule 23(b)(3) a class of "[a]ll persons in the United States who had bitcoins or money stored with Mt. Gox on February 24, 2014." Doc. 508 at 1; Doc. 509 at 9, 13. Karpeles contends that the proposed class fails to satisfy Rule 23(b)(3)'s predominance requirement. Doc. 526 at 20-30. He is correct.
A proposed class satisfies predominance if "the questions of law or fact common to class members predominate over any questions affecting only individual members ... ." Fed. R. Civ. P. 23(b)(3). "Mere assertion by class counsel that common issues predominate is not enough." Parko v. Shell Oil Co., 739 F.3d 1083, 1085 (7th Cir. 2014) (emphasis omitted). Although similar to commonality, predominance is "far more demanding." Amchem Prods., Inc. v. Windsor, 521 U.S. 591, 624 (1997). "The predominance inquiry 'asks whether the common, aggregation-enabling, issues in the case are more prevalent or important than the non-common, aggregation-defeating, individual issues.'" Tyson Foods, Inc. v. Bouaphakeo, 577 U.S. 442, 453 (2016) (quoting 2 William B. Rubenstein et al., Newberg on Class Actions § 4:49, at 195-96 (5th ed. 2012)). As such, predominance "is not determined simply by counting noses: that is, determining whether there are more common issues or more individual issues, regardless of relative importance," Parko, 739 F.3d at 1085, and it cannot be satisfied where liability determinations are individual and fact-intensive, see Kartman, 634 F.3d at 891. "Ultimately, the [district] court must decide whether classwide resolution would substantially advance the case." Suchanek v. Sturm Foods, Inc., 764 F.3d 750, 761 (7th Cir. 2014).
"Analysis of predominance under Rule 23(b)(3) 'begins ... with the elements of the underlying cause of action.'" Messner, 669 F.3d at 815 (quoting Erica P. John Fund, Inc. v. Halliburton Co., 563 U.S. 804, 809 (2011)). The elements of common law fraud under Illinois law are: "(1) a false statement of material fact; (2) known or believed to be false by the person making it; (3) an intent to induce the plaintiff to act; (4) action by the plaintiff in justifiable reliance on the truth of the statement; and (5) damage to the plaintiff resulting from such reliance." Newman v. Metro. Life Ins. Co., 885 F.3d 992, 1003 (7th Cir. 2018) (quoting Doe v. Dilling, 888 N.E.2d 24, 35-36 (Ill. 2008)); see also Toulon v. Cont'l Cas. Co., 877 F.3d 725, 734 (7th Cir. 2017) (same). Greene argues that all 50 States and the District of Columbia define common law fraud in the same manner. Doc. 509 at 15-16; Doc. 510-13. Karpeles does not contest the point, Doc. 526 at 21-25, so the court will assume that variation in state law is not itself a barrier to certification of a nationwide class here.
As discussed at greater length in the court's summary judgment opinion, 2020 WL 3250715, at *3, Greene's theory of fraud is that the Mt. Gox Terms of Use falsely represented that Mt. Gox held all assets on its users' behalf and that trades involved actual assets, that Karpeles knew those representations were false and intended to deceive Mt. Gox users, and that the users "kept assets on Mt. Gox in reliance on Karpeles's misrepresentations." Doc. 509 at 17-19. As Greene correctly acknowledges, his fraud theory "turns on Karpeles's drafting and dissemination of the Mt. Gox Terms of Use." Id. at 17. Greene thus premises his fraud claim on Mt. Gox's alleged misrepresentations in its contracts with its users.
A fraud claim can be based on representations made in a contract, provided that the elements of fraud are satisfied. See United States ex rel. Main v. Oakland City Univ., 426 F.3d 914, 917 (7th Cir. 2005) ("[F]ailure to honor one's promise is (just) breach of contract, but making a promise that one intends not to keep is fraud."); Restatement (Second) of Torts § 530, cmt. c (ALI 1977) ("[I]t is immaterial to the tort liability that the damages recoverable are identical with, or substantially the same as, those which could have been recovered in an action of contract if the promise were enforceable."). A key distinction between contract claims and fraud claims is that fraud, unlike a breach of contract, requires "actual reliance on a fraudulent misrepresentation." Greenberger v. GEICO Gen. Ins. Co., 631 F.3d 392, 401 (7th Cir. 2011). A contracting party suing for breach does not need to prove his subjective reliance on the contract's terms because "the objective manifestations of the parties, including the language they used in the contract[,]" control. Gore v. Alltel Commc'ns, LLC, 666 F.3d 1027, 1033 (7th Cir. 2012) (quoting Carey v. Richards Bldg. Supply Co., 856 N.E.2d 24, 27 (Ill. App. 2006)); see also Gupta v. Morgan Stanley Smith Barney, LLC, 934 F.3d 705, 711 (7th Cir. 2019) ("A contract has ... nothing to do with the personal, or individual, intent of the parties. A contract is an obligation attached by the mere force of law to certain acts of the parties ... which ordinarily accompany and represent a known intent.") (alterations in original) (quoting Hotchkiss v. Nat'l City Bank of N.Y., 200 F. 287, 293 (S.D.N.Y. 1911) (Hand, J.)). In a fraud suit, by contrast, reliance is necessary to show a "fraudulent act distinct from the alleged breach of contract." Greenberger, 631 F.3d at 401. The success of a fraud claim based on alleged misrepresentations articulated in a contract thus depends on how the plaintiff did understand the contract, not on how a reasonable person would understand it. See Field v. Mans, 516 U.S. 59, 70-71 (1995) (holding that the "justifiable reliance" element of common law fraud imposes a subjective, not an objective, standard).
Karpeles suggests that a fraud claim's reliance element categorically prevents class treatment of fraud cases. Doc. 526 at 21-22. That goes too far. The Advisory Committee Notes to Rule 23 approve the use of class actions in fraud cases so long as common issues predominate:
[A] fraud perpetrated on numerous persons by the use of similar misrepresentations may be an appealing situation for a class action, and it may remain so despite the need, if liability is found, for separate determination of the damages suffered by individuals within the class. On the other hand, although having some common core, a fraud case may be unsuited for treatment as a class action if there was material variation in the representations made or in the kinds or degrees of reliance by the persons to whom they were addressed.Fed. R. Civ. P. 23(b)(3) advisory committee's note to 1966 amendment (reported at 39 F.R.D. 69, 103 (1966)). Consistent with this guidance, the Seventh Circuit has held that "fraud claims ... do not automatically fail the predominance test." Beaton v. SpeedyPC Software, 907 F.3d 1018, 1029 (7th Cir. 2018); see also Pella Corp. v. Saltzman, 606 F.3d 391, 393 (7th Cir. 2010) ("While consumer fraud class actions present problems that courts must carefully consider before granting certification, there is not and should not be a rule that they never can be certified."). Fraud claims thus have no uniquely disfavored status under Rule 23. As with any predominance inquiry, the court must weigh "the common, aggregation-enabling, issues" against "the non-common, aggregation-defeating, individual issues." Tyson Foods, 577 U.S. at 453.
Some elements of Greene's fraud claim present issues common to the proposed class. Whether the Terms of Use contained materially false statements, whether Karpeles knew the statements to be false, and whether Karpeles intended to induce Mt. Gox's users to act based on the statements are all common questions. Resolving those questions would not depend on the experience of any one user, but rather on the text of the Terms, the operations of Mt. Gox, and Karpeles's state of mind. See Mullins, 795 F.3d at 673 (holding that whether the defendant's statements were "false or misleading" was "a 'common contention' that is 'capable of classwide resolution'") (quoting Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 350 (2011)); Schleicher v. Wendt, 618 F.3d 679, 687 (7th Cir. 2010) (holding, in a securities fraud case, that "whether a statement is materially false is a question common to all class members and therefore may be resolved on a class-wide basis after certification").
Reliance, however, is particular to each class member under the fraud theory advanced by Greene. To have relied on statements in the Terms of Use, a Mt. Gox user must—at a bare minimum—have been aware of what the Terms said. See Vill. of Bensenville v. City of Chicago, 906 N.E.2d 556, 591 (Ill. App. 2009) (affirming dismissal of a fraud claim where "there [was] no allegation that these misrepresentations flowed to plaintiffs and influenced their actions") (emphasis added); Premier Elec. Const. Co. v. Morse/Diesel, Inc., 628 N.E.2d 1090, 1099 (Ill. App. 1993) ("[A] common law fraud claim requires actual reliance which means that the misrepresentations must reach the plaintiff who must reasonably rely on them.") (emphasis added). As the Seventh Circuit has explained, reliance in a fraud case requires that the misrepresentation be "both believed ... and acted on" by its alleged victims. AMPAT/Midwest, Inc. v. Ill. Tool Works Inc., 896 F.2d 1035, 1041 (7th Cir. 1990). It is impossible to "believe," let alone "act[] upon," a statement of which one is unaware.
The question whether Mt. Gox users read or otherwise learned of the contents of the Terms of Use cannot yield a common answer across all or even most of the class. Greene estimates that the putative class comprises over 30,000 users. Doc. 509 at 14. No reasonable factfinder could simply assume that all or most of those users read or otherwise learned of the Terms. In fact, as noted, the available evidence points in the other direction. When asked at her deposition whether "everybody who signed up to use the Mt. Gox exchange after January of 2012 read all of the Terms of Use," Greene's expert Goforth replied, "[a]bsolutely not." Doc. 526-8 at 3 (138:14-17). Karpeles's expert expressed a similar opinion, testifying that Mt. Gox users "had to click 'okay' on the terms and conditions and they probably didn't even see them" and that "the vast majority [of users] would not have read the Terms of Use." Doc. 526-9 at 5-6 (173:21-174:5).
Goforth also acknowledged that awareness of the Terms of Use was not uniform across the class because many users, in deciding whether and how to use the Mt. Gox exchange, relied not on the Terms, but instead on their pre-existing views of how cryptocurrency exchanges normally work. Specifically, Goforth opined that "[e]ven members who did not actually read the Terms of Use would have anticipated that whoever they were dealing with was a legitimate user, with accounts backed by real [bitcoin] or fiat currency[,]" because that is "the way that crypto exchanges operate." Doc. 510-8 at ¶ 51; see also id. at ¶ 54 ("[E]ven for Mt. Gox users who did not actually read such terms, this would have been the customary way in which crypto exchanges were set up to operate."). Greene cites this opinion in contending that a reasonable factfinder could infer reliance on the Terms across the entire putative class. Doc. 509 at 27-28. But Goforth's opinion in fact conveys the contrary view. While granting that many users did not read the Terms, she opines that such users would nevertheless have held certain views about how Mt. Gox functioned, not because of anything stated in the Terms, but because of "the customary way in which crypto exchanges were set up." Doc. 510-8 at ¶ 54. Perhaps Karpeles could somehow be held liable for upsetting users' expectations even if they lacked awareness of the Terms. But that is not the question Greene proposes to resolve through his proposed class action, as it does not match the theory of fraud underlying his claim.
In any event, even if all or most Mt. Gox users were at least aware of the Terms of Use, reliance still would not present a common issue. The key question in showing reliance is not whether class members simply knew about the Terms, but how they understood the Terms and how that understanding "influenced their actions," Bensenville, 906 N.E.2d at 591, or, put another way, how they "acted on" that understanding, AMPAT/Midwest, 896 F.2d at 1041. Greene understood the Terms to convey "that users' assets would be held on the Exchange in full, could be recovered on demand, and that all trades on the Exchange would be made with other users and always involve real fiat currency and real bitcoin," Doc. 509 at 9-10, and submits that he acted upon that understanding "when deciding to move bitcoins onto, and keep them on, Mt. Gox," id. at 13. He highlights four provisions of the Terms that, in his view, convey those promises on Mt. Gox's behalf. Doc. 466 at p. 18, ¶¶ 5-8. To find reliance on a classwide basis, a factfinder would need to determine, in one fell swoop, that all or most class members shared Greene's interpretation of those four provisions and acted on it. But examining those provisions shows that a reasonable factfinder could not possibly determine that all or most putative class members would have arrived at Greene's interpretation—that is, a reasonable factfinder could not possibly answer the question, "how did the class members understand those provisions," in a uniform manner across the class.
The first provision cited by Greene reads: "Members acknowledge and agree that, when completing Transactions, they are trading with other Members, and Members accept that MtGox acts only as an intermediary in such Transactions and not as a counterparty to any trade." Doc. 510-3 at 4. That provision did not clearly prohibit Mt. Gox itself from buying or selling bitcoin on the platform. The Terms of Use defined "Member" simply to mean "Buyers and Sellers," so when Mt. Gox bought or sold bitcoin, it could have been considered a "Member." Id. at 2. Or at least that is a plausible reading of the Terms, different from Greene's professed understanding that Mt. Gox itself did not buy or sell bitcoins on the exchange.
The second provision reads: "[Mt. Gox] will use all reasonable care and skill in facilitating the matching of offers of the Members on the Site via the Platform to conclude Transactions." Id. at 4. That provision addressed only the processing of individual transactions, which is not the subject of Greene's fraud claim. Moreover, "reasonable care and skill" is highly generic and thus unlikely to support even an individual fraud claim, let alone to be understood uniformly across the class. See Corley v. Rosewood Care Ctr., Inc. of Peoria, 388 F.3d 990, 1010 (7th Cir. 2004) ("An indefinite statement like this is not the stuff of which fraud claims are made."); In re Marriage of Bower, 409 N.E.2d 75, 77 (Ill. App. 1980) ("An indefinite statement will not qualify to support an action for fraud.").
The third provision reads: "[Mt. Gox] will hold all monetary sums and all Bitcoins deposited by each Member in its Account, in that Member's name as registered in their Account details, and on such Member's behalf." Ibid. That provision certainly promised that Mt. Gox would hold users' deposits. But the provision did not explicitly say, as Greene professes to have understood it to say, that deposits would be "held ... in full" or "recover[able] on demand." Doc. 509 at 9-10. Some or most users might instead have thought that, like a bank, Mt. Gox held only a portion of users' deposits at any one time. In fact, the forum discussion attached to Greene's motion shows a user asking, in September 2012, whether "[Mt. Gox] maintain[s] full reserve," so apparently the answer was unclear to users despite what Greene professed to glean from the Terms. Doc. 510-4 at 2. True, Karpeles responded to the user's question by saying that Mt. Gox retained funds "in full," ibid., but Greene grounds his fraud claim not on that response, but on the Terms themselves. Greene must show that the Terms—whose representations are the grist for his fraud claim—were subject to a uniform interpretation, and they plainly were not.
Finally, the fourth provision reads: "Seller warrants that the Bitcoins offered to sell or to transfer correspond to actual Bitcoins." Doc. 510-3 at 4. That represents a warranty from sellers on the Mt. Gox exchange to buyers, not a promise from Mt. Gox to its users.
In sum, while Greene's reading of the Terms of Use may not be implausible, his is far from the only plausible reading, or even the best reading, and there is no basis to suppose that all or most of the 30,000-member putative class shared his reading. In fact, Goforth testified that it is improbable that "all of the people who signed up for the Mt. Gox exchange would have cared about ... the Terms of Use," Doc. 526-8 at 3 (138:18-22), thus confirming that Greene's theory of fraud would depend on individualized showings of actual reliance.
Resisting this conclusion, Greene points to several cases holding that reliance can be resolved on a classwide basis. Doc. 509 at 23-27. For instance, CGC Holding Co. v. Broad & Cassel, 773 F.3d 1076 (10th Cir. 2014), held that "reliance can be disposed of on a classwide basis without individualized attention at trial ... where circumstantial evidence of reliance can be found through generalized, classwide proof." Id. at 1089. CGC Holding involved a fake lender that demanded large upfront fees from loan applicants, denied the applications, and pocketed the fees. Id. at 1082. In such a scheme, "evidence of payment" alone could allow a factfinder to infer reliance across the class without any "individualized consideration." Id. at 1091. In so holding, however, the Tenth Circuit cautioned that "the inference of reliance here is limited to transactional situations—almost always financial transactions—where it is sensible to assume that rational economic actors would not make a payment unless they assumed that they were receiving some form of the promised benefit in return." Id. at 1091 n.9.
As CGC Holdings teaches, where every class member received the same clear promise and paid for the promised service, it can be reasonable to infer a common understanding of the promise and reliance thereon across the class. Other cases, including those cited by Greene, make the same point. See, e.g., In re U.S. Foodservice Inc. Pricing Litig., 729 F.3d 108, 120 (2d Cir. 2013) ("In cases involving fraudulent overbilling, payment may constitute circumstantial proof of reliance based on the reasonable inference that customers who pay the amount specified in an inflated invoice would not have done so absent reliance upon the invoice's implicit representation that the invoiced amount was honestly owed."); In re First All. Mortg. Co., 471 F.3d 977, 992 (9th Cir. 2006) (holding that reliance could be inferred uniformly across the class where borrowers signed mortgage agreements based on similar oral promises); Klay v. Humana, Inc., 382 F.3d 1241, 1259 (11th Cir. 2004) ("It does not strain credulity to conclude that each plaintiff, in entering into contracts with the defendants, relied upon the defendants' representations and assumed they would be paid the amounts they were due."); Vasquez v. Superior Ct., 484 P.2d 964, 972-73 (Cal. 1971) (holding that reliance could be inferred on a classwide basis where all plaintiffs were induced by specific, uniform misrepresentations to purchase freezers and frozen food); Seekamp v. It's Huge, Inc., 2012 WL 860364, at *10 (N.D.N.Y. Mar. 13, 2012) (holding that reliance could be inferred where it was "clear that every plaintiff would have relied on the implicit representation of the [product's] legality and beneficialness in deciding whether to purchase it").
But the general proposition that reliance can sometimes pose a common question does not help Greene. His theory of fraud does not rest upon a simple transaction where the mere act of payment could imply a common understanding of a misrepresentation tied to the transaction and reliance thereon. To the contrary, Greene points to representations in a contract (the Terms of Use) that many class members almost certainly did not read and, for those class members who did read it, infers an interpretation that many almost certainly did not share. In addition, Greene posits not a discrete purchase by each class member following the alleged misrepresentations, but rather decisions by class members over the course of weeks, months, or years to continue leaving money with Mt. Gox. Doc. 509 at 19, 27. These facts distinguish CGC Holding, U.S. Foodservice, Klay, and other cases with a closer and easily inferred link between a clear misrepresentation and a purchase. No reasonable factfinder could infer classwide reliance based on the fraud theory that Greene presses here.
With reliance individualized in this case, the final question under Rule 23(b)(3) is whether the common questions of falsity, knowledge, and intent nevertheless predominate over the non-common question of reliance. A class can be certified to resolve predominating common questions, even if additional process will be necessary to determine liability as to each class member. See Beaton, 907 F.3d at 1030 ("[I]ndividualized inquiries [can] be handled through 'streamlined mechanisms' such as affidavits and proper auditing procedures."); Pella Corp., 606 F.3d at 394 ("A district court has the discretion to split a case by certifying a class for some issues, but not others, or by certifying a class for liability alone where damages or causation may require individualized assessments."); Arreola v. Godinez, 546 F.3d 788, 800 (7th Cir. 2008) ("Rule 23(c)(1)(B) specifically recognizes the possibility of certifying not just 'class claims,' but also class 'issues.'"). Greene urges the court to do so here, arguing that if the common issues are resolved in his favor, then it will be "a 'straightforward matter' to resolve any remaining individual issues." Doc. 509 at 29 (quoting Suchanek, 764 F.3d at 760).
That approach would be inappropriate because the common issues do not predominate. The issue of reliance lies at the heart of the fraud claim and would be hotly disputed as to each of the 30,000 members of the putative class. Resolving that issue would require receiving evidence as to whether each user read or was made aware of the pertinent provisions in the Terms of Use, how each user subjectively understood those provisions, and whether that understanding influenced each user's decision to leave fiat currency or bitcoins deposited with Mt. Gox through February 24, 2014. It is nearly certain that only testimony from each user could resolve such questions. One matter pertinent to the question of whether to certify a Rule 23(b)(3) class is "the likely difficulties in managing a class action." Fed. R. Civ. P. 23(b)(3)(D). Holding over thirty thousand mini-trials to determine how each class member understood and whether each class member relied upon a contract they accepted nearly a decade ago would present insurmountable difficulties. See Kartman, 634 F.3d at 891 ("[T]he class-action device is not appropriate for resolving ... highly individualized questions of fact."); Thorogood, 547 F.3d at 746-48 (holding that common issues did not predominate where each class member's claim turned on the extent to which she relied on or was damaged by the defendant's allegedly deceptive advertising). Class treatment is not appropriate.
Conclusion
Greene's class certification motion is denied. Because that motion is denied, there is no need to decide whether Goforth's testimony and the exhibits challenged by Karpeles should have been excluded in deciding the motion. Karpeles's motions are accordingly denied as moot, without prejudice to renewal if Greene's individual case is tried. June 22, 2021
/s/_________
United States District Judge