Opinion
11 Civ. 5201 (DLC)
03-26-2013
FEDERAL HOUSING FINANCE AGENCY, etc., Plaintiff, v. UBS AMERICAS, INC., et al., Defendants.
For the plaintiff: Philippe Z. Selendy Kathleen M. Sullivan Adam M. Abensohn Manisha M. Sheth Jordan A. Goldstein Quinn Emanuel Urquhart & Sullivan, LLP For defendants: Jay B. Kasner Scott D. Musoff Robert A. Fumerton Skadden, Arps, Slate, Meagher & Flom LLP Four Times Square
OPINION & ORDER
APPEARANCES: For the plaintiff:
Philippe Z. Selendy
Kathleen M. Sullivan
Adam M. Abensohn
Manisha M. Sheth
Jordan A. Goldstein
Quinn Emanuel Urquhart & Sullivan, LLP
For defendants:
Jay B. Kasner
Scott D. Musoff
Robert A. Fumerton
Skadden, Arps, Slate, Meagher & Flom LLP
Four Times Square
DENISE COTE, District Judge:
The UBS Defendants in this action have moved to require the Federal Housing Finance Agency ("FHFA") to pay either $250,000 or at least $85,000 of the cost incurred in identifying certain files produced by third parties. For the following reasons, the application is denied.
The defendants are UBS Americas, Inc., UBS Real Estate Securities, Inc., UBS Securities, LLC, Mortgage Asset Securitization Transactions, Inc., David Martin, Per Dyrvik, Hugh Corcoran, and Peter Slagowitz.
BACKGROUND
This action (the "UBS Action") is one of sixteen actions filed by FHFA in this district. The litigation of these sixteen actions is being coordinated and supervised by this Court. In these actions FHFA, serving as conservator for the Federal National Mortgage Association ("Fannie Mae") and the Federal Home Loan Mortgage Corporation ("Freddie Mac") (collectively, the "GSEs"), has sued many banks and related entities and individuals in connection with the packaging, marketing and sale of residential mortgage-backed securities that the GSEs purchased in the period from 2005 to 2007. FHFA has identified essentially three misrepresentations that it alleges the defendants made about the securities. See FHFA v. UBS Americas Inc., 858 F. Supp. 2d 306 (S.D.N.Y. 2012). One of those three misrepresentations is at issue here: the defendants' representations that the underlying mortgage loans complied with certain underwriting guidelines and standards. Id. at 330.
One of the sixteen actions, FHFA v. General Electric Co., et al., 11 Civ. 7048 (DLC), has been resolved.
All parties agree that the trials of this component of FHFA's case will require a comparison of the loan files and the underwriting guidelines for loans associated with the securities that FHFA purchased. The supporting loan groups for the relevant securities in these sixteen actions include over 1,100,000 individual home mortgages.
At initial conferences, the Court urged the parties to litigate these actions on the basis of an agreed upon sample from each securitization. FHFA agreed and made a proposal for statistical sampling of the loans. For example, there are approximately 44,000 loans in the supporting loan groups in the UBS Action, and FHFA has selected a sample of 2,400 loans.Although at one point it appeared that the defendants might agree to that approach, ultimately the defendants refused. Instead, they insisted that each of the loan files for the home mortgages ("Loan Files") and its associated underwriting guidelines ("Guidelines") had to be obtained. In the face of the defendants' insistence that all Loan Files and Guidelines be obtained, FHFA reserved its right to offer evidence from Loan Files outside its sample. Reluctantly, the Court ruled on June 13, 2012 that, given the parties' failure to reach agreement, it would not attempt to require these cases to be litigated from a sample of loans in each securitization.
Originally, FHFA suggested a sample of only 1,060 loans for the UBS Action.
As a consequence, the defendants undertook a massive and expensive effort to locate all of the over 1,100,000 Loan Files and associated Guidelines for the loans that constitute the supporting loan groups for each of the securities purchased from the defendants by the GSEs. The documents have come from defendants' files, from FHFA, and from scores of third parties. FHFA has assisted in the effort, but the brunt of the work has been borne by the defendants. In the UBS Action alone, Loan Files have been sought from sixty entities.
In some instances, defendants have located Loan Files and Guidelines within their own files that must be produced in actions in which they are not named as defendants.
The sixteen actions have been divided into four tranches for trial and organized as a pyramid. The first tranche has one member: the UBS Action. It will proceed to trial in January 2014. The second tranche has two members, who will proceed to trial separately in June 2014. These two actions represent the largest losses experienced by the GSEs. The third tranche has three members, who will each proceed to trial in September 2014. The remaining cases, which are the lawsuits in which the smallest damages are sought, are in the fourth tranche and will be tried in January 2015.
The defendants in the three cases that comprise the first and second tranches requested that FHFA be required to disclose the results of its factual re-underwriting review evaluating compliance with underwriting guidelines for each of its sample loans far in advance of the scheduled exchange of expert reports. Over the objection of FHFA, the Court agreed. As a result, FHFA and the defendants in those three actions agreed to give priority to the production and identification of the Loan Files and Guidelines for FHFA's sample loans, to try to reach agreement that the assembled documents were the best representation of those Loan Files and Guidelines that the parties have been able to recreate, and to negotiate a worksheet that FHFA would use to identify the underwriting deficiencies. The parties also agreed to a schedule by which FHFA would identify its perceived deficiencies in this aspect of the underwriting process and by which these defendants would reply. UBS was required to complete, or at least substantially complete, its production of the Loan Files for FHFA's sample loans by December 17, 2012.
The defendants in tranches three and four have also been instructed to give priority to production of the loans from FHFA's samples and to attempt to reach agreement that the produced Loan Files and Guidelines are "the best representation of the Loan File and Guidelines existing at the time of the loan's origination that the parties have been able to recreate as of the time of such agreement."
On November 6, 2012, the UBS Defendants informed the Court that some of the Loan Files produced by third parties would require "cracking." While it is often possible to identify which Loan Files in the hands of a third party are associated with a particular securitization and constitute one of FHFA's sample loans by using identifying file numbers and other information, that is not always the case. For example, a loan in the FHFA sample may be identified only by a loan originator's loan identification number, but the Loan File may be in the possession of the loan servicer and only be associated with the servicer's loan identification number. When identifying numbers cannot confirm that the correct Loan File has been produced, Loan Files must be "cracked" open to obtain identifying information, such as a borrower's address. With that information, the Loan File can be identified as a member of FHFA's sample.
On November 15, the UBS Defendants advised the Court that approximately 10,000 third party Loan Files would need to be cracked. UBS also asked that FHFA be required to bear half the cost of the cracking process. FHFA was skeptical that cracking was necessary and offered to work with UBS and the third parties to avoid it if possible. On November 19, the Court denied UBS's request that FHFA be required to share the cost of any cracking.
On December 3, FHFA admitted that cracking would be necessary since some third parties produced documents "in a manner that makes it impossible . . . to match loan files in those productions to our samples without opening the files." The difficulty involved 7,400 Loan Files from various third parties and 8,400 from Ocwen Loan Servicing, LLP ("Ocwen"). On December 14, the Court ordered UBS to bear the costs of cracking third party Loan Files to determine if those files are part of FHFA's sample.
On December 17, UBS submitted an affidavit from a Senior Managing Director within FTI Consulting indicating that it would take the firm eight weeks beginning January 7 -- or until March 4 -- to identify the 15,800 Loan Files by cracking them open. It estimated the cost as between $940,000 and $1.8 million, depending on the whether an optical character recognition search or manual review process were used. UBS renewed its request that FHFA be required to bear half the cost. At a conference that day, the Court expressed skepticism that the cracking would be so time consuming and costly. FHFA represented that it could do the cracking far faster and less expensively, but that it was still talking with Ocwen to see if the cracking process could be avoided for its 8,400 files. The Court granted UBS's request that it be allowed to brief the issue of sharing with FHFA the costs of the cracking process, and ordered the parties to confer so that the sample's Loan Files could be identified as quickly as possible. As the Court noted, however, UBS was to bear the burden of any cracking of Loan Files pending the Court's decision on the motion.
By December 18, FHFA was able to tell UBS that its vendor could crack the group of 7,400 unidentified Loan Files by December 31 at a cost of just $210,000. UBS promptly agreed to that process being undertaken. On December 27, FHFA advised UBS that "[d]espite both parties' efforts, we have not received information that allows us to identify the Loan Files produced by Ocwen without a manual examination." FHFA's vendor estimated that the examination of the Ocwen files could be completed by January 7 at a cost of $220,000. The next day, UBS provided FHFA with a spreadsheet containing information that UBS claimed would allow FHFA to identify a portion of the Loan Files without cracking, and asked for more time for Ocwen to provide information that could identify the remainder of the Loan Files. FHFA responded that in its view the spreadsheet did not contain information sufficient to render cracking unnecessary for any portion of the Loan Files, and asked UBS whether it had a different understanding. FHFA received no response, and two days later, on December 31, it began cracking the Ocwen files itself, without waiting for an agreement from UBS.
On January 7, 2013, FHFA's vendor finished cracking the Ocwen files, at a final cost of approximately $85,000. Cracking was thus completed on both sets of files on the date on which UBS's vendor proposed to begin cracking. The total cost of cracking both sets of files ended up being approximately $250,000, far less than the $940,000 to $1.8 million UBS's vendor estimated the task would cost.
FHFA provided UBS with spreadsheets generated by the cracking process on January 5 and 7. From that data, FHFA advised UBS on January 11 that of the 15,500 Loan Files that had been cracked, it was able to identify 275 Loan Files as matching loans in FHFA's sample.
UBS filed the instant motion on January 8. FHFA filed its opposition on January 23, and UBS filed a reply on January 30.
DISCUSSION
UBS contends that FHFA should pay the $250,000 associated with cracking the two sets of Loan Files to identify which of those files are associated with FHFA's sample loans. UBS makes essentially two arguments for this cost shifting. First, it characterizes the work as "litigation costs" and the creation of "work product" rather than costs associated with the production of discovery materials, and argues that each party must in the ordinary course bear its own litigation costs. Second, it argues that FHFA unilaterally decided to crack the 8,400 Ocwen files, and must as a result bear at least that cost.
UBS is of course correct that the "American Rule" normally requires each party to bear its own litigation costs. Alyeska Pipeline Serv. Co. v. Wilderness Soc'y, 421 U.S. 240, 247 (1975). On the other hand, the traditional "presumption is that the responding party must bear the expense of complying with discovery requests." Oppenheimer Fund, Inc. v. Sanders, 437 U.S. 340, 358 (1978). This distinction is also at the heart of the amount of discretion granted to district courts to allocate such costs. While the American Rule forbids a district court from shifting the parties' litigation costs absent express statutory authority, W. Va. Univ. Hosps., Inc. v. Casey, 499 U.S. 83, 86-87 (1991), Congress has provided that authority in several provisions of the Federal Rules. See 28 U.S.C. §§ 2072-74 (authorizing promulgation of Rules); Oliveri v. Thompson, 803 F.2d 1265, 1271, 1274 (2d Cir. 1986) (Rule 11 among exceptions to American Rule). Rule 26(c), in particular, authorizes federal courts to shift the costs of discovery "for good cause." See Oppenheimer Fund, 437 U.S. at 358. The propriety of placing the cost of the cracking exercise on UBS therefore hinges, in large part, on whether cracking the Loan Files is properly characterized as part of document production or a litigation expense like attorney work product.
Attorney work product is generally not discoverable, and the scope of what constitutes work product is closely tied to policy considerations surrounding discoverability. See United States v. Adlman, 134 F.3d 1194, 1196 (2d Cir. 1998) (protection of work product "is intended to preserve a zone of privacy in which a lawyer can prepare and develop legal theories and strategy"); In re Steinhardt Partners, L.P., 9 F.3d 230, 234 (2d Cir. 1993) ("The logic behind the work product doctrine is that opposing counsel should not enjoy free access to an attorney's thought processes."). Thus work product is usually defined by reference to the "thoughts" and judgment of an attorney. See United States v. Nobles, 422 U.S. 225, 238 (1975) ("At its core, the work product doctrine shelters the mental processes of the attorney, providing a privileged area within which he can analyze and prepare his client's case".). Work product has been described as including "preparing legal theories, planning litigation strategies and trial tactics, and sifting through information." Steinhardt Partners, 9 F.3d at 234 (citing Hickman v. Taylor, 329 U.S. 495, 510-11 (1947)). Work product protection is codified in the federal rules, which do not allow discovery of "documents and tangible things that are prepared in anticipation of litigation or for trial by or for another party or its representative." Fed. R. Civ. P. 26(b)(3)(A).
Cracking does not qualify as work product. Cracking is simply opening a Loan File and retrieving information that allows the Loan File to be identified. The end result of the cracking process may be a spreadsheet containing the resulting information, which typically includes such uncontroversial data as the loan number, the borrower's name and address, and the amount of the loan. Cracking is therefore more like a kind of transcription, albeit time-consuming and costly transcription, than the thought processes or analysis of an attorney. See Steinhardt Partners, 9 F.3d at 234; Riddell Sports Inc. v. Brooks, 158 F.R.D. 555, 559 (S.D.N.Y. 1994) ("[T]he collection of evidence, without any creative or analytic input by an attorney or his agent, does not qualify as work product.").
Nor does cracking qualify as work prepared by or for FHFA or its agent, such that it falls within the definition of work product in Rule 26. The cracking at issue here was conducted by FHFA's outside vendor for the benefit of both parties. As a result, FHFA provided the spreadsheets to UBS as soon as the cracking task was completed. For this very reason, there is also no suggestion that the fact that the task was performed by an agent for the plaintiff as opposed to an agent for the defendant should bear on the analysis. Cracking is a simple, ministerial act. In this case, because it was performed by the plaintiff's agent, it was done more cheaply and quickly, and UBS does not question the reliability of the results.
Cracking is more properly seen as part of the cost of document production. In responding to a typical document request, a party must sift through documents so that it can identify and produce those documents being requested. See Rothman v. Emory Univ., 123 F.3d 446, 455 (7th Cir. 1997) (producing party ordinarily has an "obligation to sort through the documents and produce only those responsive to [the opposing party's] request"). In this case, FHFA elected to try its case based on a sample set of loans, a restricted subset of the large number of loans contained in the supporting loan groups for each securitization. To determine which Loan Files in fact corresponded to the sample loans, it proved necessary in this instance to crack open some of the Loan Files themselves and extract certain information from them. Indeed, both parties must be able to identify the Loan Files in FHFA's sample in order to begin the reunderwriting process, which is where the true analysis of the Loan Files begins. Because in this instance it proved a necessary step in identifying what was being produced, cracking is part of the cost of production, not work product.
Of course, the set of Loan Files being produced is not limited to FHFA's sample, because defendants refused to consent to try this case on the basis of a sample. This decision provided at least some of the impetus for requiring defendants to bear the cost of this production and cracking exercise. Plaintiffs initially sought production of only the Loan Files corresponding to loans in their sample, a set of some 2,400 loans out of a total of roughly 44,000 loans in the securitizations at issue in the UBS Action. It was defendants' refusal to agree to try the case on the basis of a sample that necessitated production of all Loan Files, and vastly increased the burden on the parties and third parties. Furthermore, the extent to which UBS will decide to rely on loans outside the sample in rebutting FHFA's proof, or in presenting its affirmative defenses, remains to be seen. Placing the cost of any future cracking that may be necessary on UBS therefore serves to help keep costs in check for all parties.
Although UBS complains that defendants are "bearing virtually all of the costs associated with the production of third party loan files", it does not ask in this motion that FHFA be required to share those costs generally. As the Court has explained, "we are in this very expensive, burdensome document production, which has enormous ramifications for your clients, the defendants, the plaintiffs, and now third parties. And the defendants will bear the cost of that."
--------
The cases cited by UBS do indicate that courts should be hesitant to force a party to create and produce new evidence for the benefit of its adversary. See, e.g., Paramount Pictures Corp. v. Replay TV, CV 01-9358 (FMC), 2002 WL 32151632, at *2-3 (C.D. Cal May 30, 2002) (company not required to produce customer-use data that had never been collected); Alexander v. FBI, 194 F.R.D. 305, 310 (D.D.C. 2000) (FBI not required to produce list that did not exist). The mere transcription involved in cracking, however, is not closely analogous. See Riddell Sports, 158 F.R.D. at 559 ("[S]ince the mere assembly of evidence by one party is not work product, its adversary will be able to obtain that evidence . . . [and] is not thereby liable for the opposing counsel's fees for time expended in collecting the evidence."). Cracking is necessary to identify the Loan Files that are being produced, based on criteria over which there is no dispute and on information that is in existence and is in fact contained in the Loan File itself. Cracking is also both beneficial and necessary for both FHFA and UBS. As the parties have made clear, the defendants have an equal (if not greater) interest in ensuring the completeness and accuracy of the Loan Files. Cracking is part of that process, and its benefits accrue to the defendants as well as FHFA.
UBS argues that even if it is required to bear the cost of cracking generally, it should not be required to bear the cost of cracking the Ocwen files, because FHFA "unilaterally" decided to ask its vendor to crack those files without waiting for permission from UBS. UBS suggests that, at the time FHFA asked its vendor to crack the Ocwen files, UBS was still waiting for additional information from Ocwen that might render cracking unnecessary. This information was to arrive via a spreadsheet on January 2; it never arrived because UBS instructed Ocwen to abandon its efforts when it learned that FHFA's vendor had already started cracking. FHFA, on the other hand, argues that even if the information had been produced, it would not have been sufficient to render cracking unnecessary.
It should be remembered that cracking the Ocwen files was completed by FHFA's vendor at a considerable savings in time and money to UBS. It should also be remembered that UBS had been on notice of the need for cracking long before December 31, 2012, when FHFA began cracking the Ocwen files. On November 15, the Court ordered FHFA and UBS to work together to see whether providing additional information to the servicers would render cracking unnecessary, and on November 19, the Court ordered UBS to bear the cost of cracking if this effort did not succeed. On December 17, UBS was again ordered to undertake the cracking process, with the issue of which party would ultimately pay being reserved for this briefing. In short, UBS had ample time to confer with Ocwen to determine whether it could provide information that would render cracking unnecessary, and no reason to expect that it could delay cracking indefinitely by withholding its consent. There is thus no reason to treat the cracking of the Ocwen files differently from the non-Ocwen files.
Finally, it should be noted that UBS does not contend that the cost associated with the cracking of the 15,800 Loan Files is unreasonable. The entire exercise was completed on the day UBS's vendor was prepared to begin, having taken less than three weeks instead of the eight weeks UBS estimated it would take. And it was done for a fraction of the cost: $250,000 instead of $940,000 to $1.8 million. Had UBS been required to shoulder just half of that cost, it would have paid between $470,000 and $900,000, and its work with the sample loans would have been delayed by two months.
CONCLUSION
The UBS Defendants' January 8, 2013, motion to require FHFA to bear the financial burden associated with cracking is therefore denied.
SO ORDERED: Dated: New York, New York
March 26, 2013
____________
DENISE COTE
United States District Judge