Sanchez et al v. Wells Fargo & Company et al

Filing 245

ORDER ON MOTION CHALLENGING DAMAGE STUDY AND SEEKING CLASS DECERTIFICATION by Judge William Alsup [granting in part and denying in part (200) Motion in case 3:07-cv-05923-WHA]. (whasec, COURT STAFF) (Filed on 5/5/2009)

1 2 3 4 5 6 7 8 9 10 VERONICA GUTIERREZ, et al., Plaintiffs, v. WELLS FARGO & COMPANY, et al., Defendants. / INTRODUCTION In this certified consumer class action, defendant Wells Fargo Bank, N.A., challenges plaintiffs' damage study and moves to decertify the two classes previously allowed. The main holding of this order is that the record does not justify resorting to an aggregate fluid-recovery method of proving injury and damages. Decertification is the proper remedy for one class but not yet for the other. ANALYSIS This consumer action challenges certain overdraft-charge practices by Wells Fargo Bank. Both practices allegedly gouge depositors and one supposedly even lures them into writing overdrafts by overstating the "available balance." The full background of this case is ORDER ON MOTION CHALLENGING DAMAGE STUDY AND SEEKING CLASS DECERTIFICATION No. C 07-05923 WHA IN THE UNITED STATES DISTRICT COURT FOR THE NORTHERN DISTRICT OF CALIFORNIA

United States District Court

11
For the Northern District of California

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

1 2 3 4 5 6 7 8 9 10

set forth in companion orders filed today. Suffice it here to say that two classes were certified earlier on as follows (Dkt. 98): the "re-sequencing" class was defined as "all Wells Fargo customers from November 15, 2004, to June 30, 2008, who incurred overdraft fees on debit card transactions as a result of the bank's practice of sequencing transactions from highest to lowest." the "including and deleting" class was defined as "all Wells Fargo California customers with consumer checking accounts from November 15, 2004, to June 30, 2008, who incurred overdraft fees on debit card transactions after dissemination by Wells Fargo of available-balance information that once reflected and later deleted a debit card transaction." Now that expert reports have been exchanged and fact discovery has closed, Wells Fargo challenges plaintiffs' proof regarding injury and damages. Although the class period covers a 43-month period, the expert damages study covered only one month. From this one month, plaintiffs' counsel proposes to extrapolate to the full 43-month period and seek hundreds of millions in alleged damages. More specifically, during fact discovery, plaintiffs requested and Wells Fargo provided transaction data for 10,000 customers within the one-month period between May 14 and June 13, 2008. The data included the number of overdrafts and the amount of overdraft fees for each customer during the period. (Those actually penalized by overdraft charges was a much smaller group than the 10,000 customers.) Plaintiffs did not request information outside the one-month period and no claim is made that the bank would have refused to supply it. 1. THE RE-SEQUENCING CLASS.

United States District Court

11
For the Northern District of California

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

With respect to the "re-sequencing claim" class, Computer Expert Art Olsen analyzed the data to calculate the amount of overdraft fees for actual versus different hypothetical scenarios. Olsen wrote software code to put the transactions in the order as sequenced by Wells Fargo for each customer each day (i.e., the challenged re-sequencing practice of highest to lowest). He also wrote a code to alternatively sequence the transactions in chronological order in addition to other possible sequencing orders. When a customer received less overdraft charges as a result of the alternate chronological sequencing, then the code determined the difference in the amount of overdraft fees. Finally, the overdraft fees for all 2

1 2 3 4 5 6 7 8 9 10

customers who received less overdrafts under the alternate chronological sequencing were totaled. According to plaintiffs, Olsen's analysis showed that the amount of overdraft fees for the thirty-day period was much larger using the bank's re-sequencing order than when the chronological order was used. He also compared it against yet another more benign sequencing policy, namely, the bank's pre-2001 sequencing debit card transactions before checks and ACH and then sequencing each group from lowest to highest. That analysis produced an even larger disparity in overdraft fees. Olsen provided his calculations to Expert Statistian Charles Cowan who then extrapolated from the thirty-day to the 43-month period. Using Olsen's calculations, Cowan determined that the bank would have assessed 13.97 percent less in overdraft fees if it would have sequenced transactions in chronological order and the bank would have assessed 15.95 percent less in overdraft fees if it would have sequenced transactions from lowest to highest. The total amount of overdraft fees assessed to California consumers during the class period was $1.7 billion. Cowan then applied the 13.97 percent increase to the total amount of overdraft fees that Wells Fargo assessed -- $1.7 billion dollars -- to calculate the total additional overdraft fees assessed by Wells Fargo due to re-sequencing. He then reduced that figure by twenty percent to adjust globally for overdraft fees that were waived or not collected by Wells Fargo. He calculated the total damages for the re-sequencing class to be $198.1 million. Earlier, Wells Fargo served interrogatory requests on plaintiffs seeking identification of the algorithm or methodology that identified individual members of the certified classes (Jolley Exh. 8 at 2­3). In addition to referring defendant to the code used by Olsen, plaintiffs responded: "The script utilized by Mr. Olsen does not output a listing of customers. However, it would be a simple matter of added script code to the existing script code to output a listing of customers for each scenario. . . The methodology utilized for the sample would be the same as utilized for the universe of class members" (id. at 3).

United States District Court

11
For the Northern District of California

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

3

1 2 3 4 5 6 7 8 9 10

When certifying the "re-sequencing" class, the class order relied on testimony from Expert Lewis Mandell, who stated that software could be written and applied against the bank's database to determine precisely which customers had been charged overdraft fees for transactions that occurred on days when the customers had positive account balances (Dkt. 98 at 25). For these customers, the program could compare the number and amount of overdraft charges they received with the number and amount they would have received had the sequence been more benign. The class order concluded: "Given the bank's large wealth of computerized account information, this order is confident and so finds that software will be designable to extract the vital elements needed to assess damages and causation" (id. at 26). Despite his earlier representations, Expert Mandell did not conduct any such study. Rather than use the bank's computer database to reconstruct the actual alleged overcharges paid by the actual alleged victims of the challenged practice, counsel has opted for a short-cut method of proof via different experts, namely to estimate globally all overcharges for the 43-month period based solely on a single-month database. Counsel desires that a gross judgment be entered for this large amount to be followed by an administrative claims process. The excess left (inasmuch as not all customers would submit claims) would be paid as a bonus to those who submitted claims or would be given away under a cy pres theory. This damage approach is known as fluid recovery. The main question presented by this motion is whether the record presented justifies resorting to such a short-cut method of proof. * * *

United States District Court

11
For the Northern District of California

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

The extent to which fluid recovery can be obtained in a litigated class action on an aggregate basis rather than by tallying up individual claims (even if on a formulaic basis using a computerized database) has been a point of contention for decades. The Supreme Court has not spoken on the issue. In the Second Circuit, such aggregation of damages is not allowed, period. See Eisen v. Carlisle & Jacquelin, 479 F.2d 1005, 1018 (2d Cir. 1973) (rejecting fluid recovery as a "solution of the manageability problems of class actions"). The main reason is that actual injuries and damages for each absent class member are required elements of substantive proof and either a class-wide method of proving those elements must be supplied 4

1 2 3 4 5 6 7 8 9 10

or, once liability is otherwise determined, individual injury/damages proceedings must be held. Otherwise, Rule 23 would be used to truncate the required substantive elements of proof by each claimant in violation of the Rules Enabling Act, 28 U.S.C. 2071­77. The main argument the other way is that it is sometimes hard to prove injury and damages on a member-by-member basis yet easy to reliably estimate class-wide damages in the aggregate and it is better that the wrongdoer disgorge all ill-gotten gains even though windfalls will result. In the Ninth Circuit, the best reading of the law is that fluid recovery is allowed only where conventional methods of proof are demonstrably unavailable and, even then, only in an extraordinary circumstance such as the Marcos case described below. The seminal decision in our circuit is In re Hotel Telephone Charges, 500 F.2d 86, 89­90 (9th Cir. 1974). Citing Eisen, Judge Ely wrote for the circuit: "allowing gross damages by treating unsubstantiated claims of class members collectively significantly alters substantive rights under the antitrust statutes" and "is clearly prohibited by the Enabling Act." Id. at 90; see also Six Mexican Workers v. Arizona Citrus Growers, 904 F.2d 1301, 1305­06 (9th Cir. 1990); Molski v. Gleich, 318 F.3d 937, 954 (9th Cir. 2003). One circumstance where it appears the circuit would relax this requirement is where a defendant has not preserved its records so as to allow a more precise calculation of damages. Such a defendant cannot complain of an inability of plaintiffs to do a more accurate analysis. Cf. Bigelow v. RKO Radio Pictures, 327 U.S. 251 (1946). With the foregoing in mind, the crux of this order is that plaintiffs have not demonstrated that the bank's records cannot be used to calculate the actual damages suffered by each class member. Instead, plaintiffs have simply resorted to a short-cut method of aggregation and attempted to extrapolate from one month to 43 months without actually doing the hard work. All 43 months of data were available but only one month was requested. The bank has not been shown to have stonewalled or destroyed evidence. Although they have not yet done so, plaintiffs' counsel assert the same process used for the one-month period can be used for the class as a whole. On this record, plaintiffs have not shown that they could not use

United States District Court

11
For the Northern District of California

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

5

1 2 3 4 5 6 7 8 9 10

the bank's records to calculate damages on a member-by-member basis and, indeed, concede that it could have been done.1 * * *

The Court has carefully considered all of the federal authorities cited by plaintiffs to justify use of a short-cut method of aggregation. None contradicts what the undersigned believes the law to be in the Ninth Circuit -- fluid recovery or aggregate damages cannot be used to circumvent individual proof requirements with the limited exceptions noted above. The undersigned recognizes that fluid recovery is sometimes used in settling class actions but the extent to which fluid recovery may be used in litigated class actions is subject to the rule stated. Most authorities cited by plaintiffs did not even involve class actions but addressed statistical evidence in some other context. For example, Salas by Salas v. Wang, 846 F.2d 897 (3rd Cir. 1988), did not involve a class action.2 While Dukes v. Wal-Mart, Inc., 509 F.3d 1168 (9th Cir. 2007), was a class action, statistical projection was used in that case to demonstrate the disparity between the treatment of employees in order to raise an inference of discrimination. Significantly, it was not used as a method of aggregating class-wide damages. (Also, of course, the entire decision is being reheard en banc.) To be sure, in Schwab v. Philip Morris USA, Inc., 449 F. Supp. 2d 992 (E.D.N.Y. 2006), Judge Weinstein allowed a sampling technique in a class action to calculate damages. But this very point was reversed on appeal, an important item plaintiffs' counsel neglected to add, it is sad to have to say. McLaughlin v. American Tobacco Co., 522 F.3d 215, 231 (2d Cir. 2008).

United States District Court

11
For the Northern District of California

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

1 Plaintiffs' Expert Olsen submitted a declaration stating among other things that the same code used for the sample of Wells Fargo customers can be applied to the entire class to provide a list of affected customers and the amount for each customer. 2

As to the state court cases cited, state courts are not bound by the Rules Enabling Act.

6

1 2 3 4 5 6 7 8 9 10

In In re Static Random Access (SRAM) Antitrust Litigation, 2008 WL 4447592, *6 (N.D. Cal. Sept. 29, 2008), Judge Wilken certified a class of direct purchasers in an antitrust case and stated in part: Plaintiff has proffered three methodologies for calculating damages on a class-wide basis: the first compares SRAM prices before and after the period of the price-fixing conspiracy; the second compares SRAM prices for comparable products; and the third uses Defendants' cost data to estimate what competitive prices for SRAM should have been. Plaintiff has proffered methods for calculating aggregate damages for overcharges paid by class members, based on average market prices. The validity of those methods will be adjudicated at trial based upon economic theory, data sources, and statistical techniques that are entirely common to the class. This passage sheds little light on the issue at hand because the statistical techniques alluded to were not described in Judge Wilken's opinion. In addition, the issue of whether or not the statistical techniques were valid was left for another day. In In re Sugar Industry Antitrust Litigation, 1976 WL 1374, 27 (N.D. Cal. 1976), Judge Boldt said "[w]ith respect to damages determined on a class-wide basis, decisional law, including Ninth Circuit decisions, approves proof of injury by just and reasonable inferences where the facts and circumstances of a case necessitate such a method." But the decisions then cited by Judge Boldt did not involve class actions; at most, they stood for the familiar proposition that where a defendant has done wrong, the defendant, as the wrongdoer, cannot object to a damage study that is the best possible with the data available. See, e.g., Bigelow v. RKO Radio Pictures, 327 U.S. 251 (1946). The best federal authority cited by plaintiffs is Hilao v. Estate of Marcos, 103 F.3d 767, 782 (9th Cir. 1996). That decision allowed statistical sampling for class damage claims in a narrow and extraordinary context. After a trial against the estate of Philippines dictator Ferdinand Marcos, 10,059 claims were received from class members of which 518 were deemed facially invalid, leaving 9,541claims. From these, 137 claims were randomly selected based on expert testimony that there would be a 95 percent statistical probability that the same percentage determined to be valid among the examined claims would be applicable to the totality of claims filed. The district court then appointed a special master who oversaw 7

United States District Court

11
For the Northern District of California

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

1 2 3 4 5 6 7 8 9 10

depositions taken in the Philippines of the 137 claimants. The special master recommended damages for 131 valid claims in the category of "torture" claims, a separate amount for "summary-execution" claims, and yet a third amount for "disappearance" claims. The Ninth Circuit stated: "While the district court's methodology in determining valid claims is unorthodox, it can be justified by the extraordinarily unusual nature of this case." Id. at 786. Given the extraordinary nature of the case in that it involved extreme violations of human rights against many victims, the Ninth Circuit held that the methodology did not violate due process. The present case, however, does not involve extraordinary circumstances such as human rights violations. This order holds that plaintiffs have not justified the use of an aggregate class-wide method of proving fluid-recovery damages. The statistical methodology in question might be admissible for other purposes, such as to prove the bank's intention to maximize profits rather than to act in good faith vis-a-vis its customers. But it does not come even close to a proper damages study. * * *

United States District Court

11
For the Northern District of California

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

On the other hand, two bank arguments have no merit. First, with respect to the issue of how to account for after-the-fact waivers by the bank of overdraft charges, the answer is simple. They could reasonably be accounted for on a last-in-first-out basis, meaning that waivers should be applied to erase the most recent penalty charges first, and then working backwards in time. This presumption is entirely reasonable given that normally it would be the most recent spate of charges that would bring a customer into the bank to seek a waiver. If the bank wishes to prove some other specific individualized agreement by a specific customer on how to allocate a waiver across a history of charges, the bank would be permitted to try to rebut the presumption on an individualized basis, provided it has timely disclosed the evidence under Rule 26. Second, contrary to the bank, customers who continued to stay with Wells Fargo after being forced to pay overdraft penalties did not consent to a "voluntary" payment. A depositor retains his or her right under the contract, i.e., his or her right to good faith and fair dealing, 8

1 2 3 4 5 6 7 8 9 10

even if the bank violates that right and the depositor fails to switch banks. The bank's argument to the contrary is like saying an employee subjected to unlawful working conditions voluntarily consents thereto by not quitting his or her job. * * *

A common misconception about class actions is that all counsel must do at trial is to prove the case for the class representative and then the case will be automatically proven for each class member, at least as to liability. This is not so. While the fact pattern of the class representative's story may serve to illustrate the problem, proof must be furnished to cover each element of the class for each class member. Fortunately, this can often be done via class-wide avenues of proof. For example, the same standardized consumer contract (with the same covenant of good faith and fair dealing) could extend to all class members, although actual evidence of the class-wide usage of the contract must be presented at trial. Thus, while class-wide proof can prove elements of even the class representative's own claim, the reverse is not true, i.e., facts bearing on the representative's individual case will not automatically go class-wide. Another misconception is that in a class action an individualized issue is a showstopper under Rule 23. The rule tolerates some individualized issues so long as they do not predominate over the common issues. For example, damages are almost always individualized issues and, at worst, they can be handled via individualized trials after a claims process. Other elements of proof may occasionally have to be individualized or at least be proven by chapters using subclasses. How many localized fact patterns to accept in the mix just depends on how problematic the unique questions will be in the context of the overall case, always keeping in mind that we must ever be fair to the absent class members whose rights will be extinguished by any judgment under res judicata. * * *

United States District Court

11
For the Northern District of California

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

9

1 2 3 4 5 6 7 8 9 10

As to the re-sequencing class, decertification seems too severe a remedy on the present record, at least without further considering less drastic alternatives. Two less drastic remedies will be first considered, reserving on whether decertification will ultimately occur: · Alternative One: Proceed to trial and if the class prevails on liability, invite class members to submit individual claims, allowing trials as necessary as to disputed claims, both as to injury and as to amount. · Alternative Two: Postpone the trial and allow plaintiffs' counsel to mine the bank data for all of the damage analysis practicable, submit new reports, and entertain a new round of motions in limine. Were this allowed, counsel would be well advised to address all of the bank's myriad other objections to the study, at least as an alternative branch of analysis. This order addresses only the main flaws and, those being sufficient, finds it unnecessary to reach the other attacks. Given that we are within a month of the trial date, within SEVEN CALENDAR DAYS, counsel shall please submit separate memoranda limited to ten double-spaced pages advising on which alternative (or some other alternative) is preferred. Please also advise whether plaintiffs' counsel should be required to pay the bank's incremental costs and fees associated with Alternative Two if that course is adopted, meaning the costs and fees that could have been saved had counsel done it right to begin with. Within THREE CALENDAR DAYS thereafter, each side may then respond, limited to five double-spaced pages. 2. THE INCLUDING-AND-DELETING CLASS.

United States District Court

11
For the Northern District of California

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

The including-and-deleting claim is predicated on a multi-step fact pattern that allegedly lures customers into making overdrafts. Customers go online and consult their available balance shortly after a purchase and see that very purchase listed among the items reflected in the available balance. Usually, the item stays deducted and is paid and posted when the item is presented, so that no problem usually arises. If, however, the item is not presented for payment within three days (an occasional scenario), then the bank has an undisclosed policy of reversing the item out of the available balance. If the same customer then consults and relies on the 10

1 2 3 4 5 6 7 8 9 10

available balance, he is apt to make an overdraft, not realizing that the bank has reversed out the item and not realizing that the available balance shows higher than it really is or so it is alleged. The following diagram illustrates the chronological steps for this claim:

United States District Court

11
For the Northern District of California

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 With respect to the "including-and-deleting claim" class, the same thirty-day sample was used to project damages. Olsen wrote a different software code to identify the customers in the sample that were affected by the bank's including-and-deleting practice. Significantly, he had to make some assumptions. Here is a critical one: Olsen assumed that every time a customer accessed his or her account online, at an automatic teller machine, or in a branch store, the customer learned the then-current available balance. Then, Expert Cowan determined that 8.8 percent of all overdraft fees were incurred because of the including-and-deleting practice. Cowan applied 8.8 percent to $1.7 billion -- the aggregate overdraft fees charged by Wells Fargo -- and then reduced that number by 8.75 percent to account for fees that the sample showed were waived. Cowan calculated the total 11

1 2 3 4 5 6 7 8 9 10

damages for the including-and-deleting class to be $141.7 million. According to plaintiffs, this number requires further deduction by twelve to twenty percent to reflect overdraft fees that were waived or not collected by Wells Fargo. Significantly, Cowan included all cases wherein the customer consulted the online available balance between steps 2 and 3 even if the customer had not consulted the available balance between steps 1 and 2. As with the companion class, the including-and-deleting class was certified on the representation that plaintiffs could utilize a computer program to ascertain precisely who belongs in the class and the extent of any injury: "It seems reasonable that a computer program can be fashioned and applied against the bank's database to track when customers accessed their available-balance information, determine the available balance displayed, and verify whether an overdraft fee was charged. Presumably, the program could also determine whether a transaction was included in the available balance and later deleted" (Dkt. 98 at 24). * * *

United States District Court

11
For the Northern District of California

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

For the including-and-deleting class, the threshold deficiency is the same as above -- counsel cannot use a one-month sample in place of reconstructing the alleged wrongs done to each class member, account-by-account, for all 43 months involved. This deficiency might be curable but other flaws are not curable. One concerns proof of the reliance element, this being a deception claim. In a class action, as stated, it is still necessary to prove each class member's full claim, including the elements of proof needed for the claim. Some individualized proofs can, of course, be furnished in a class action but common proofs must predominate. Otherwise, the proceeding will be unmanageable. Where fraud is alleged, it is necessary to prove reliance by each class member. In an earlier era, defendants regularly argued that reliance was inherently an individual, subjective issue and was not amenable to any class-wide method of proof. For some species of fraud cases, however, the courts have developed presumptions of reliance to serve as class-wide methods of proof. One example is fraud-on-the-market theory of reliance in securities class actions. Another is the standardized-sales-pitch theory in real-estate fraud cases (wherein the wrongdoer makes the same false sales pitch directly to all purchasers with the manifest intent 12

1 2 3 4 5 6 7 8 9 10

that they rely thereon). Except where a class-wide method of proof is furnished, however, the federal courts have regularly disfavored fraud class actions in litigated cases, for the individual issues usually will predominate over the common issues. Here, plaintiffs have supplied no class-wide method of proving reliance as to the "including and deleting" class. Even if plaintiffs' streamlined and revised claim were allowed to go forward (a point addressed below), the essence of the claim would still be that customers were led to believe that an item was reflected in the available balance when, in fact, it had been backed out without notice so as to make the available balance appear higher than it really was, leading the customer to incur overdrafts. As stated, plaintiffs' expert assumed that every time any customer accessed his or her account online, or via an ATM, or in person (by visiting a teller window), he or she would have then learned and relied on the then available balance. This categorical assumption is mere ipse dixit and defies common experience. It is easy to use an ATM without obtaining one's balance. It is easy to make a teller transaction without obtaining the balance. It is possible to consult an account online for reasons other than obtaining one's balance, such as to remind oneself of the amount of a specific charge so as to apply for a reimbursement. The record includes no foundation whatsoever to support plaintiffs' blanket and universal assumption that such events communicate the available balance to the depositor. To be sure, many customers specifically consult the available-balance information -- and surely rely on it to write checks and make debit purchases. But it is too far a leap to presume that all customers who access their accounts in the ways described invariably seek or even learn their account balances, there being too many other plausible purposes in the customer contacts described. If we could isolate those instances in which the available balance was consulted from those where it was not, the next question would be whether those who did consult the available balance actually relied on it to make an innocent overdraft, putting aside the further question whether any such reliance was reasonable. Here a presumption might be available. If it could be proven that the bank intentionally misled customers into believing that they could rely on the "available balance" without the customer reviewing what items were or were not reflected 13

United States District Court

11
For the Northern District of California

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

1 2 3 4 5 6 7 8 9 10

therein, then it might be proper to presume that any new checks or debits incurred in the immediate aftermath of such a consultation would have been in actual reliance on the misinformation. This would be akin to presuming reliance when a wrongdoer intends that a victim rely on misinformation, as in the standardized sales-pitch cases. The reason for the "immediate aftermath" qualifier is that most checking accounts are in perpetual motion. Most customers would not rely on stale balance information. If, for example, a bank gave out a misleading available balance on a Monday, it would be unlikely that customers would still rely on it by the end of the week. Too many things could have changed. Any presumption, were one to be used at all, ought to extend only to fresh misinformation, perhaps one-day old or less. (The seven-day period used by plaintiffs would be far too long.) Even then the records might show that the customer would have overdrawn irrespective of the misinformation, in which case it would be unreasonable to make any presumption of reliance. To repeat, however, the present record does not allow us to even reach this step in the analysis, for plaintiffs have advanced no plausible class-wide method for sorting those who actually learned any misinformation from those who did not. This leads to the final flaw evaluated by this order. This claim was originally sold to the Court as one in which the bank customer consulted the available balance between steps 1 and 2, noticed that a specific item had been subtracted, and then later consulted the available balance between steps 2 and 3 and relied on it without realizing that it looked larger than it really was due to the backed-out items. Now, however, plaintiffs seek to skip the first part and to broaden the claim to depositors who merely consulted the online balance between steps 2 and 3, whether or not they consulted it between steps 1 and 2. The damage study thus wholly ignores the very scenario that formed the basis for class certification. The bank says this expansion has been done to triple the amounts involved in the damage study. At the time of class certification, this very attempt was expressly rejected by the certification order (Dkt. 98 at 13­14): To be clear, the only viable claim concerning the available-balance information concerns Wells Fargo's practice of including but then 14

United States District Court

11
For the Northern District of California

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

1 2 3 4 5 6 7 8 9 10

deleting transactions without adequate notice. The mere fact that the available balance appearing online or at an ATM is sometimes not fully accurate is not, by itself, enough to state a claim. One reason is that the bank has no way of knowing the extent of unpresented checks and the extent of debit transactions that will turn out to be more than what was initially authorized. The definition published online for "available balance" itself states: (Please note that some transaction activity may not be immediately recorded to your account and will then not be reflected in the available balance . . . .) Therefore, with respect to the available-balance information, only plaintiffs' claims as related to Wells Fargo's practice of including items and then deleting them without adequate notice, thereby deceiving customers into thinking they have a larger balance and inducing them to go into overdraft situations, will move forward. In disregard of this limitation, plaintiffs' damage theory severely overreaches and tries to include the very category of claims expressly rejected. Significantly, plaintiffs did not even calculate a number based on the very clear definition supplied in the certification order. Under the certified issue, the reliance problem is even more exacerbated. A class-wide method of proving reliance should have been found for both times the customer consulted the available balance. And, as to the first (between steps 1 and 2), we would need a class-wide presumption that the customer not only consulted the "available balance" but specifically noted that a particular item was listed as a deduction. This is a hopelessly individualized question, not amenable to any plausible method of common proof (and may well be the reason that counsel sought to switch theories). This order need not reach all of the other challenges presented, for the foregoing is sufficient. * * *

United States District Court

11
For the Northern District of California

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

For all of the foregoing reasons, the including-and-deleting class is hereby DECERTIFIED and shall not proceed as a class claim. This order finds that, contrary to the views that led to the original certification order, individual issues will plainly predominate over common issues with respect to the including-and-deleting class. At their expense, plaintiffs' counsel shall promptly cause notice to be given to class members advising them of this development, and the reasons, so that any class members who have been relying on the pendency of this action can timely bring 15

1 2 3 4 5 6 7 8 9 10

their own claims. Within SEVEN CALENDAR DAYS, please submit a form of notice and a plan of notice distribution. CONCLUSION For the foregoing reasons, the motion to decertify is GRANTED IN PART AND DENIED
IN PART.

IT IS SO ORDERED.

Dated: May 5, 2009.

WILLIAM ALSUP UNITED STATES DISTRICT JUDGE

United States District Court

11
For the Northern District of California

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 16