Hoffman v. Transworld Systems Incorporated

Opinion

1 2 3

4 5 UNITED STATES DISTRICT COURT WESTERN DISTRICT OF WASHINGTON 6 AT SEATTLE 7 ESTHER HOFFMAN, et al., 8 Plaintiffs, 9 v. C18-1132 TSZ 10 TRANSWORLD SYSTEMS ORDER INCORPORATED, et al., 11 Defendants. 12

13 THIS MATTER comes before the Court on a motion to exclude the testimony of 14 Plaintiffs’ expert Michael Andrew, docket no. 329, filed jointly by defendants 15 Transworld Systems Incorporated (“TSI”), Patenaude and Felix, A.P.C. (“P&F”), 16 Matthew Cheung, and National Collegiate Student Loan Trusts 2004-2, 2005-2, 2005-3, 17 2006-1, 2006-3, and 2007-4 (collectively the “NCSLTs”). Also before the Court are 18 motions to exclude the testimony of TSI’s and the NCSLTs’ experts Sandy Goldstein, 19 docket no. 330, and Roger Saylor, docket no. 333, filed by plaintiffs Esther Hoffman, 20 Sarah Douglass, Anthony Kim, Il Kim, and Daria Kim. Having reviewed all papers filed 21 in support of, and in opposition to, the motions, the Court enters the following Order. 22 1 Background 2 Plaintiffs contend that Defendants filed false and misleading affidavits in the 3 collection or attempted collection of student loan debt. Second Amended Complaint

4 (“SAC”) (docket no. 61 at 2). Plaintiffs allege the affidavits are false and misleading 5 because Defendants do not possess the documents required to establish that the NCSLTs 6 own Plaintiffs’ student loans. Id. at ¶¶ 56–57, 80–81 & 109–10. Between 2015 and 7 2017, the NCSLTs filed individual lawsuits against Plaintiffs to collect unpaid student 8 loan debt. Id. at ¶¶ 42–47, 71–72 & 88–90. In support of these actions, the NCSLTs,

9 through their counsel, P&F, submitted “Affidavit and Verification of Account” 10 documents provided by TSI, the NCSLTs’ post-default servicer. Id. at ¶¶ 48–49, 73–74 11 & 98–99. Attached to these affidavits were “Pool Supplements” which allegedly 12 establish that the NCSLTs purchased Plaintiffs’ student loans from the originating 13 lenders. See id. at ¶¶ 51, 76 & 100–02. The Pool Supplements reference attached loan

14 schedules but none of the documents attached to the affidavits contained a list of loans, 15 individual borrowers, or account numbers. See id. at ¶¶ 52, 78 & 101–04. Plaintiffs 16 contend that the loan schedules referenced in the Pool Supplements are lost or never 17 existed. See id. at ¶¶ 56–57, 80 & 110. In contrast, TSI and the NCSLTs allege that the 18 loan schedules referenced in the Pool Supplements are comprised of six Excel

19 spreadsheets: (i) Bank of America Final Roster.xls; (ii) 2005-2 Lender Payout Summary 20 - FMC Master.xls; (iii) LENDER_ROSTERS_ALL IN.xls; (iv) BANK_ONE_RECON_ 21 SUMMARY - Revised.xls; (v) Bank_of_America_Post_Sale.xls; and (vi) 20074_Lender 22 Report - Post Sale-BANK OF AMERICA.xls. 1 Discussion 2 1. Legal Standard 3 The party offering an expert’s testimony bears the burden of establishing its

4 admissibility by a preponderance of the evidence. See Bldg. Indus. Ass’n of Wash. v. 5 Wash. State Bldg. Code Council, 683 F.3d 1144, 1154 (9th Cir. 2012). “A witness who is 6 qualified as an expert by knowledge, skill, experience, training, or education may testify 7 in the form of an opinion or otherwise if” (i) “the expert’s scientific, technical, or other 8 specialized knowledge will help the trier of fact to understand the evidence or to

9 determine a fact in issue,” (ii) “the testimony is based on sufficient facts or data,” 10 (iii) “the testimony is the product of reliable principles and methods,” and (iv) “the expert 11 has reliably applied the principles and methods to the facts of the case.” Fed. R. 12 Evid. 702(a)–(d). The trial judge is tasked with ensuring “that an expert’s testimony both 13 rests on a reliable foundation and is relevant to the task at hand.” Daubert v. Merrell

14 Dow Pharm., Inc., 509 U.S. 579, 597 (1993). In determining whether expert testimony is 15 reliable, the Court may consider certain factors, such as testing, peer review, error rates, 16 and acceptability in the relevant scientific community. See id. at 593–94. But “the test of 17 reliability is ‘flexible,’ and Daubert’s list of specific factors neither necessarily nor 18 exclusively applies to all experts or in every case.” Kumho Tire Co. v. Carmichael, 526

19 U.S. 137, 141 (1999). The Court, however, must take care “to assure that a proffered 20 witness truly qualifies as an expert, and that such testimony meets the requirements of 21 Rule 702.” See Jinro Am. Inc. v. Secure Invs., Inc., 266 F.3d 993, 1004 (9th Cir. 2001). 22 1 2. Michael Andrew 2 Plaintiffs retained Andrew to opine on the authenticity of the six Excel spreadsheet 3 files discussed above. Andrew is an expert in the field of computer forensics with over

4 eighteen years of experience in digital forensic examination. Andrew Decl. at ¶ 2 (docket 5 no. 237). Plaintiffs asked Andrew to analyze the metadata associated with the six 6 spreadsheet files. Id. at ¶ 3. According to Andrew, the metadata embedded in an Excel 7 spreadsheet records details such as the file’s “Author,” the date the file was “Created,” 8 and the “Last Modified” date. Id. at ¶ 4.

9 During his examination, Andrew found that the Excel spreadsheets had Last 10 Modified dates later than the effective dates documented in the corresponding Pool 11 Supplements. Id. at ¶ 7; Andrew Report, Ex. 3 to Defs.’ Mot. to Exclude (docket no. 12 329-3 at 6). Based on this finding, Andrew believes that “there is no proof for any direct 13 association” between the spreadsheets and the Pool Supplements. Andrew Decl. at ¶¶ 7–

14 14. Andrew opines that the spreadsheets are not the loan schedules referenced in the Pool 15 Supplements and “there is simply no way to trust that the provided Excel spreadsheets 16 represent the original files in anything other than filename.” Andrew Report (docket no. 17 329-3 at 6). 18 Andrew also compared the hash values of the six Excel spreadsheets he inspected

19 with the hash values from additional copies of the spreadsheets TSI and the NCSLTs 20 produced in discovery. Andrew Report (docket no. 329-3 at 2–3). Andrew explains that 21 “a hash value for a file or document is considered to be a reliable ‘digital fingerprint’ of 22 the file and if any question arises about the matter then all that is needed for verification 1 is to re-hash the data and compare the original and the copy; if the hash values are 2 identical then the two files are identical.” Id. at 5. Andrew found that some of the copies 3 of the spreadsheets produced in discovery had different hash values than the six

4 spreadsheets he inspected. Id. at 2–3. Finally, Andrew observed that there is no 5 “Created” date recorded in the metadata for the spreadsheet file named “Bank of America 6 Final Roster.xls.” Id. at 4. He believes that this “anomaly” is “unusual” and 7 demonstrates “attempts to hide or misrepresent the Created date on at least one 8 spreadsheet.” Id. at 4 & 6.

9 As an initial matter, Defendants do not contest that Andrew is qualified to extract 10 and analyze metadata. Instead, Defendants argue that some of his opinions are (i) not 11 relevant to the present action, (ii) beyond the scope of his qualifications, and (iii) not 12 based on reliable principles and/or methods. Defendants contend that Andrew’s 13 testimony regarding the purportedly mismatched hash values is not relevant because it is

14 of no consequence in this action. Defendants argue that this issue is a “red herring” 15 because the only spreadsheets relevant to the present matter are the six Andrew inspected, 16 not the additional copies of the spreadsheets produced during discovery. This argument, 17 however, ignores Plaintiffs’ theory that none of the Excel spreadsheets in TSI’s and the 18 NCSLTs’ possession are the original loan schedules referenced in the Pool Supplements.

19 See SAC at ¶¶ 56–57, 80–81 & 109–110 (docket no. 61). Andrew’s testimony that the 20 Defendants allegedly produced forensically different versions of the Excel spreadsheets 21 during the course of discovery is relevant because it raises questions about the 22 1 authenticity of the spreadsheets.1 See Primiano v. Cook, 598 F.3d 558, 565 (9th Cir. 2 2010) (“Expert opinion testimony is relevant if the knowledge underlying it has a valid 3 connection to the pertinent inquiry.”). Defendants’ motion to exclude, docket no. 329, is

4 DENIED as it relates to Andrew’s opinion regarding the mismatched hash values. 5 The Defendants also challenge Andrew’s opinion that “there is no proof for any 6 direct association” between the Excel spreadsheets and the corresponding Pool 7 Supplements. As discussed above, Andrew contends that the Excel spreadsheets cannot 8 be the original loan schedules referenced in the Pool Supplements because the

9 spreadsheets were purportedly modified (in an undisclosed way) after the Pool 10 Supplements became “effective.” Defendants contend that Andrew is not qualified to 11 offer this opinion because he has no specialized knowledge in assignments, 12 securitization, and the interpretation of legal documents. Andrew, for example, admits 13 that he does not know the legal significance of a Pool Supplement’s effective date or how

14 securitizations work. See Andrew Dep. at 30:18–23 & 31:12–16, Ex. 4 to Defs.’ Mot. to 15 16

17 1 Defendants also contend that Andrew’s testimony regarding the mismatched hash values is not based on 18 sufficient facts or data because Andrew did not examine the content of the spreadsheets. During his deposition, Andrew explained that he did not analyze the substantive content of any of the spreadsheets 19 and examined only the hash values. See Andrew Dep. at 33:14–19, Ex. 4 to Defs.’ Mot. to Exclude (docket no. 329-4). In contrast, TSI’s and the NCSLTs’ expert examined the substantive content of each of the spreadsheets with mismatched hash values and found them to contain identical content. Goldstein 20 Rebuttal Report, Ex. 3 to Leonard Decl. (docket no. 331-1 at 32) (“[T]he data in the rows and columns of all three spreadsheets in each set contain identical information.”). Goldstein attributes any mismatched 21 hash values to differences in data in the “Document Header” that is invisible to any user viewing the spreadsheets. Id. (“A Document Header is a tool used by computers to define the technical details of a forthcoming file.”). The Court concludes that Defendants’ criticism goes to the weight of Andrew’s 22 opinion, not its admissibility, and may be explored during cross-examination at trial. 1 Exclude (docket no. 329-4). Defendants also argue that Andrew’s opinion is not based 2 on sufficient facts or data or reliable principles and methods. 3 In performing its gatekeeping function, the Court must screen the jury from

4 unreliable opinions. See Alaska Rent-A-Car, Inc. v. Avis Budget Grp., Inc., 738 F.3d 960, 5 969 (9th Cir. 2013). In this case, the Court questions whether Andrew’s opinion is 6 reliable. Expert opinion testimony “is reliable if the knowledge underlying it has a 7 reliable basis in the knowledge and experience of the relevant discipline.” Id. at 969 8 (quoting Primiano, 598 F.3d at 565). There is no question that Andrew may testify about

9 the process he used to identify metadata within the spreadsheets such as their “Last 10 Modified” dates, or that he may opine on the hash values associated with the 11 spreadsheets. Andrew’s ultimate conclusion, however, appears to go beyond the scope of 12 his expertise and the data he reviewed. To reach his opinion that the six Excel 13 spreadsheets are not the actual loan schedules referenced in the Pool Supplements,

14 Andrew simply compared the effective dates listed in the Pool Supplements with the 15 “Last Modified” dates he identified.2 Andrew Report (docket no. 329-3 at 6). 16 Additionally, such a comparison is unlikely to assist a jury because comparing two dates 17 does not require any specialized skill or knowledge. See Alaska Rent-A-Car, 738 F.3d at 18 969–70 (“The district court is not tasked with deciding whether the expert is right or

19 wrong, just whether his testimony has substance such that it would be helpful to a jury.”). 20

21 2 Andrew did not analyze the metadata or their associated hash values within the Pool Supplements. Andrew Dep. at 26:9–16; Andrew Report (docket no. 329-3 at 5) (“The relevant dates referenced in the 22 language of the Pool Supplement documents was important as context.”). 1 Therefore, the Court DEFERS to trial a ruling on whether or to what extent Andrew may 2 testify about his opinion that the Excel spreadsheets are not the original loan schedules 3 referenced in the Pool Supplements.

4 3. Sandy Goldstein 5 TSI and the NCSLTs retained Goldstein to review and evaluate Andrew’s 6 processes, opinions, and conclusions. See Goldstein Report, Ex. 2 to Leonard Decl. 7 (docket no. 331-1); Goldstein Rebuttal Report, Ex. 3 to Leonard Decl. (docket no. 331-1). 8 Goldstein has more than thirty years of experience in technology-related matters such as

9 computer and document forensics. Goldstein Report (docket no. 331-1 at 15). Having 10 reviewed Andrew’s declaration and report, Goldstein opines that Andrew’s conclusions 11 are unsupported by the evidence. Id. at 16; Goldstein Rebuttal Report (docket no. 331-1 12 at 28). Goldstein believes that Andrew’s conclusion regarding the authenticity of the 13 loan schedules fails to consider how the Last Modified dates of the Excel spreadsheets

14 could have changed without affecting the files’ content. Goldstein Report (docket 15 no. 331-1 at 16–17). Goldstein explains that many activities can trigger updates to a 16 file’s Last Modified date field, such as upgrades to a computer’s operating system, 17 copying or moving a file from one location to another, or opening and saving a file 18 without changing any data.3 Id. at 18–19.

19 20

21 3 Goldstein also contends that Andrew failed to consider whether the mismatched hash values reflect substantive differences in the content of the relevant files. Goldstein Rebuttal Report (docket no. 331-1 at 22 30). Goldstein concludes that the files have the same substantive content. Id. at 32. 1 Plaintiffs argue that Goldstein’s testimony should be excluded because it is not 2 based on sufficient facts or data and is unreliable. Plaintiffs contend that Goldstein did 3 not check the hash values of the spreadsheets he reviewed and speculates about what

4 computer events could change a file’s Last Modified date. Plaintiffs fault Goldstein for 5 not completing the required testing to determine which activity “actually caused” the 6 spreadsheets’ Last Modified dates to change. But Goldstein’s testimony is not offered to 7 definitively establish what the spreadsheets’ metadata proves or disproves. Instead, 8 Goldstein’s testimony identifies potential deficiencies in Andrew’s methodology.

9 Plaintiffs’ challenges go to weight of Goldstein’s testimony, not its admissibility, and 10 their motion to exclude Goldstein’s testimony, docket no. 330, is DENIED. 11 4. Roger Saylor 12 TSI and the NCSLTs retained Saylor to rebut Andrew’s testimony and provide an 13 overview of the process by which the NCSLTs were structured and acquired student

14 loans. Saylor Report at ¶ 7, Ex. 1 to Henry Decl. (docket no. 334-1). From 2001 to 15 2009, Saylor served as the Managing Director of the Education Finance group within the 16 Citigroup Global Securitized Markets business. Id. at ¶ 2. Although he did not lead, 17 manage, or structure any of the NCSLTs involved in this action, he did serve as an 18 underwriter for several of the defendant NCSLTs. Id. at ¶ 3 (“[M]y group was

19 responsible for placing the Trust Certificates (the securities that provided holders with a 20 beneficial interest in Trust) with investors. This required me to review and become 21 familiar with all the relevant agreements, Prospectuses and Supplements governing the 22 securitization prior to the closing of the securitization.”). Saylor disagrees with Andrew’s 1 opinion that the Excel spreadsheets are not the original loan schedules referenced in the 2 Pool Supplements. He believes Andrew’s opinion is incorrect because “it is absolutely 3 appropriate and customary for an electronic schedule of loans pertaining to a particular

4 securitization to post-date the closing of the securitization.” Id. at ¶¶ 32–34. 5 Plaintiffs contend that Saylor’s testimony should be excluded because it is 6 unreliable and not based on sufficient facts or data. Plaintiffs argue that Saylor’s 7 opinions are based primarily on knowledge from his experience in the private student 8 loan industry and are unsupported because he did not review relevant documents for all

9 of the defendant NCSLTs and did not inspect the metadata associated with the Excel 10 spreadsheets.4 These arguments are unavailing. An expert can testify based on his or her 11 professional experience, see Woods v. Conagra Foods Lamb Weston, Inc., No. 15-cv- 12 05067, 2016 WL 4719886, at *2 (E.D. Wash. June 3, 2016), and Saylor explained why he 13 did not review all of the securitization documents for each of the six NCSLTs, see Saylor

14 Report at ¶ 17. Saylor reviewed the relevant documents pertaining to the 2005-2 and the 15 2007-4 NCSLTs. Id. Based on his professional experience, he believes “that the 16 securitizations for all of the NCSLTs worked in a virtually identical way to each other, 17 and that the agreements and related documents governing each NCSLT are virtually 18 identical.” Id. To the extent Plaintiffs contend that the securitizations and related

19 20 4 The Court rejects Plaintiffs’ assertion that Saylor’s opinions are unreliable because he has no experience with metadata. Despite Plaintiffs’ contention, Saylor does not opine on the metadata contained in the spreadsheets. The Court similarly rejects Plaintiffs’ additional argument that Saylor’s testimony should 21 be excluded because it is not relevant to the task at hand. Saylor’s opinions about the process and timing of securitizations is offered to challenge Plaintiffs’ theory that the loan schedules at issue in this action are 22 lost or never existed. 1 documents of the defendant NCSLTs are not virtually identical, they can address this 2 criticism in cross-examination. Plaintiffs’ motion to exclude Saylor’s testimony, docket 3 no. 333, is DENIED.

4 Conclusion 5 For the foregoing reasons, the Court ORDERS: 6 (1) Defendants’ motion, docket no. 329, to exclude the testimony of Michael 7 Andrew is DEFERRED in part and DENIED in part. The Court DEFERS until trial a 8 ruling on whether or to what extent Andrew may testify about his opinion that the Excel

9 spreadsheets are not the original loan schedules referenced in the Pool Supplements. All 10 other requested relief is DENIED; 11 (2) Plaintiffs’ motion, docket no. 330, to exclude the testimony of Sandy 12 Goldstein is DENIED; 13 (3) Plaintiffs’ motion, docket no. 333, to exclude the testimony of Roger Saylor

14 is DENIED; and 15 (4) The Clerk is directed to send a copy of this Order to all counsel of record. 16 IT IS SO ORDERED. 17 Dated this 26th day of September, 2022. 18 A

19 Thomas S. Zilly 20 United States District Judge