U.S. Bank National Association v. Goldman Sachs Mortgage Company
- Paul Gardephe
- 1:19-cv-02305
- U.S. District Court · Southern District of New York
- 18
In U.S. Bank v. Goldman Sachs, Judge Nathan denied U.S. Bank’s motion to use loan sampling to prove contract liability and damages.
U.S. Bank National Association, as trustee for the two identified trusts, and Goldman Sachs Mortgage Company and the related defendants. The ruling determines that U.S. Bank may not obtain summary judgment establishing that the contracts clearly permit its proposed statistical-sampling method; it does not finally resolve the underlying breach-of-contract claims.
What happened
U.S. Bank National Association, as trustee for two mortgage-securities trusts, sued Goldman Sachs Mortgage Company and related defendants over allegedly defective mortgage loans and alleged failures to cure or repurchase them. In U.S. Bank National Association v. Goldman Sachs Mortgage Company, U.S. Bank asked to use statistical sampling rather than examine every loan individually.
U.S. Bank proposed examining samples of liquidated loans, calculating the percentage with warranty breaches, and using that percentage to estimate liability and damages. Goldman argued that the contracts required loan-by-loan proof and did not clearly allow sampling. The opinion did not decide whether the proposed sampling would be reliable or admissible under the evidence rules.
Judge Alison J. Nathan denied U.S. Bank’s motion for partial summary judgment and denied both parties’ requests for oral argument. The court held that the contracts’ loan-specific language did not clearly authorize sampling at this stage, and ordered the parties to file a revised case-management plan within two weeks. The court did not resolve Goldman’s alternative argument about judicial estoppel or the parties’ dispute about the contract’s notice requirement.
The detailed version
- U.S. Bank National Association v. Goldman Sachs Mortgage Company · No. 1:19-cv-02305
- Paul Gardephe
- Mar. 10, 2022
Background
These two related cases concern two trusts that pooled mortgage loans to back residential mortgage-backed securities. U.S. Bank served as trustee. Goldman Sachs Mortgage Company sponsored the trusts and underwrote the loans, while GS Mortgage Securities Corp. deposited the loans into the trusts. The trusts contained approximately 8,500 loans worth approximately $1.7 billion.
The agreements included representations and warranties about the loans and required Goldman Sachs Mortgage Company to cure or repurchase loans when a breach materially and adversely affected the value of a loan or the trust’s interest in it. The agreements described repurchase as the sole remedy for certain breaches. The repurchase price depended on individual-loan characteristics, including unpaid principal balance, mortgage interest, servicing advances, and related expenses.
U.S. Bank received investor notifications stating that at least 617 loans in one trust and 1,041 loans in the other materially breached one or more representations or warranties. U.S. Bank notified Goldman and demanded that it cure or repurchase the identified loans and other defective loans. The opinion states that Goldman had not cured any breaches or repurchased any loans when the motion was litigated.
U.S. Bank alleged breach of contract and sought specific performance of the repurchase obligation and damages. The court had previously denied Goldman’s motion to dismiss and allowed U.S. Bank’s claims to proceed.
The Proposed Sampling Method
U.S. Bank sought partial summary judgment on whether it could rely on statistical sampling to prove liability and damages at trial. Its proposed method involved randomly selecting 400 loans from each loan group in each trust, for 1,600 liquidated loans total. The sample would be drawn from liquidated loans with at least $100 in losses. U.S. Bank would use loan-by-loan proof for active loans it alleged were defective.
U.S. Bank’s experts would reunderwrite the selected loans by comparing each loan with the contractual warranties and examining information about the borrower, loan, and mortgaged property. They would determine the rate of defective loans in the sample, extrapolate that rate to the relevant loan groups, and calculate damages. U.S. Bank asserted that this process would require reunderwriting approximately 2,700 fewer loans than a loan-by-loan approach.
The parties agreed that whether the contracts permitted this method was a contract-law question. The court expressly noted that the reliability or admissibility of the experts’ methodology under Federal Rule of Evidence 702 was a separate issue that could be raised later.
Legal Standard
Summary judgment may be granted only when there is no genuine dispute about any material fact and the moving party is entitled to judgment as a matter of law. For a contract claim, summary judgment is appropriate on the contract’s meaning only when the relevant language is wholly unambiguous and has a definite meaning. If the contract is ambiguous and the parties’ intent must be examined, the issue generally presents a question of fact.
New York law governed the contract dispute. Under that law, a contract is ambiguous if a reasonably intelligent person could objectively interpret its language in more than one way. Ambiguity is determined from the contract itself rather than outside sources.
Court’s Analysis
The court concluded that the agreements did not speak directly to whether statistical sampling was permitted. But the court rejected U.S. Bank’s argument that the absence of an express prohibition unambiguously authorized sampling. In the court’s view, the agreements addressed how breach, material and adverse effect, and the repurchase price were to be established, and those provisions consistently used loan-specific language.
The representations and warranties agreement required Goldman to cure or repurchase “such Mortgage Loan” after a qualifying breach. The court reasoned that determining whether a breach materially and adversely affected value required examining information specific to the particular loan, including information about its originator, borrower, and property.
The court also found that the remedy was loan-specific. The agreement referred to repurchasing or substituting for “a Mortgage Loan,” and the repurchase-price formula depended on individual-loan facts. The court therefore concluded that the contractual structure generally called for proof of breach, materiality, and damages on a loan-by-loan basis.
U.S. Bank argued that its proposal was limited to liquidated loans and that no liquidated loan would actually be transferred back to Goldman. The court held that this distinction did not resolve the problem because identifying a breach, determining material harm, and calculating the repurchase price remained loan-specific even for liquidated loans.
The court considered conflicting decisions from other courts. Some decisions had allowed sampling in similar mortgage-securities cases, while others had required loan-by-loan proof. The court was not persuaded that the sampling methodology’s possible reliability or admissibility answered the contract-interpretation question. It also declined to follow the reasoning of decisions that treated sampling as broadly accepted evidence without first determining whether the specific contracts authorized it.
The court acknowledged that denying sampling would increase the number of loans requiring reunderwriting from 1,600 to 4,305. It found that burden significant but not prohibitive and held that cost or inconvenience could not change the meaning of the parties’ contracts.
Unresolved Issues and Disposition
The court did not resolve Goldman’s argument that U.S. Bank should be barred from arguing for sampling because of positions U.S. Bank allegedly took in earlier cases. The court also did not resolve whether U.S. Bank’s notice to Goldman satisfied the agreements’ notice requirement or whether compliance with that requirement was a prerequisite to filing suit.
The court DENIED U.S. Bank’s motion for partial summary judgment. It also DENIED the parties’ requests for oral argument and ordered the parties to file a revised case-management plan within two weeks. The order resolved specified docket entries in both related cases.
Read the full 18-page opinion on CourtListener, the free public archive maintained by the Free Law Project.