In the
United States Court of Appeals For the Seventh Circuit ____________________ No. 22-1553 UNITED STATES OF AMERICA ex rel. MICHELLE CALDERON, Plaintiff-Appellant,
v.
CARRINGTON MORTGAGE SERVICES, LLC, Defendant-Appellee. ____________________
Appeal from the United States District Court for the Southern District of Indiana, Indianapolis Division. No. 1:16-cv-00920-RLY-MJD — Richard L. Young, Judge. ____________________
ARGUED NOVEMBER 30, 2022 — DECIDED JUNE 14, 2023 ____________________
Before WOOD, JACKSON-AKIWUMI, and LEE, Circuit Judges. WOOD, Circuit Judge. Michelle Calderon sued Carrington Mortgage Services on behalf of the United States for alleged violations of the False Claims Act. Calderon is a former em- ployee of Carrington. She alleges that Carrington made false representations to the U.S. Department of Housing and Urban Development (HUD) in the course of certifying residential 2 No. 22-1553
mortgage loans for insurance coverage from the Federal Housing Administration (FHA). Carrington moved for summary judgment on the basis that Calderon did not meet her evidentiary burden on two el- ements of False Claims Act liability. First, it asserted that she could not show that the allegedly false representations were material to HUD’s decisions to pay out various claims under the federal mortgage insurance program. Second, it con- tended that she could not show that the false representations caused HUD to suffer a monetary loss. The district court sided with Carrington on both elements and granted summary judgment, disposing of Calderon’s lawsuit. Though we conclude that Calderon does have suffi- cient proof of materiality, we agree that she has not met her burden of proof on the element of causation. We therefore af- firm the district court’s decision. I A Federal mortgage insurance is designed to create a path to homeownership for borrowers who might be considered too risky to qualify for a traditional mortgage because of their lack of savings, poor credit history, or low income. The Direct Endorsement Lender program is one through which HUD provides mortgage insurance to approved, private lenders. HUD covers the losses of private lenders in the event of a loan default to encourage the issuance of these higher risk mort- gages. Carrington has been a Direct Endorsement Lender for many years. If a potential Direct Endorsement Lender such as Carring- ton wishes to cover a loan with federal mortgage insurance, it No. 22-1553 3
must first submit to an underwriting process during which it assesses the prospective borrower’s eligibility for federal in- surance. HUD publishes handbooks that provide the under- writing guidelines for lenders and promulgates regulations that govern Direct Endorsement lending. 1 Carrington hires its own Direct Endorsement Underwriters and operates its own quality control system. Its goal is to ensure that it properly evaluates a borrower’s financial information, determines the degree of risk involved in issuing the loan, and complies with all federal requirements. After the lender approves the loan, the lender submits the loan to HUD for review and endorse- ment. Through this submission, the lender certifies to HUD that the borrower meets the minimum standards of HUD’s underwriting guidelines. HUD relies on these certifications to issue the necessary insurance coverage. Next, all loans submitted for federal insurance are subject to a pre-endorsement review by HUD. The parties dispute the scope of that review, but HUD’s own regulations indicate that the agency is focused on verifying that all necessary docu- ments are present, rather than on assuring the accuracy of the information it finds in the loan file. See HUD, 4155-2, Lender’s Guide to the Single Family Mortgage Insurance Process 8.C.1.b (2010). Once a loan passes pre-endorsement review, HUD issues federal insurance to the lender for that loan. If a loan file is missing some of the required documentation, the lender instead receives a notice of return that specifies the
1 See, e.g., HUD, 4155-1, Mortgage Credit Analysis for Mortgage In- surance, at https://www.hud.gov/sites/documents/41551HSGH.PDF (“HUD 4155-1”); HUD, 4155-2, Lender’s Guide to the Single Family Mort- gage Insurance Process, at https://www.hud.gov/sites/docu- ments/41552HSGH.PDF (“HUD 4155-2”). 4 No. 22-1553
deficiencies and corrective action needed before the loan can be federally insured. HUD may conduct further examination, even after it is- sues the insurance for the lender. It subjects approximately 5 percent of loans to a post-endorsement technical review in which it evaluates the loan using the federal underwriting re- quirements and confirms the accuracy of the information in the loan file. When a post-endorsement review reveals mate- rial noncompliance with HUD’s underwriting guidelines, HUD will require the lender to agree to an indemnification agreement, under which the lender must abstain from filing an insurance claim in the case of default or reimburse HUD if HUD makes a payment on an insurance claim for that mort- gage. Federal regulations define which violations may qualify as “serious and material.” 24 C.F.R. § 203.255(g)(3). 2
2 Because materiality is central to this appeal, we furnish the full text of section 203.255(g)(3) here: (3) Serious and material violation. The mortgagee shall indemnify HUD for an FHA insurance claim paid within 5 years of mortgage in- surance endorsement, if the mortgagee knew or should have known of a serious and material violation of FHA origination requirements, such that the mortgage loan should not have been approved and en- dorsed by the mortgagee and irrespective of whether the violation caused the mortgage default. Such a serious and material violation of FHA requirements in the origination of the mortgage may occur if the mortgagee failed to, among other actions: (i) Verify the creditworthiness, income, and/or employment of the mortgagor in accordance with FHA requirements; (ii) Verify the assets brought by the mortgagor for payment of the required down payment and/or closing costs in accordance with FHA requirements; or No. 22-1553 5
B Calderon worked at Carrington from March 2013 to March 2015 as a Direct Endorsement Underwriter. During that time, HUD classified loans in one of four ways following a post-endorsement review: conforming, deficient, unac- ceptable, or mitigated. Deficient loans were those with docu- mentation deficiencies or processing errors that presented only low-risk issues. Unacceptable loans contained a high- risk error or omission and did not meet the basic eligibility requirements for federal mortgage insurance. Once a loan was deemed unacceptable, lenders had an opportunity to ex- plain or correct the identified deficiencies. If they were able to rectify the problem, the loan was reclassified as mitigated. If not, HUD issued an indemnification agreement. Concerned about what she was seeing on the job, includ- ing allegedly reckless and inappropriate underwriting prac- tices, Calderon brought this lawsuit against Carrington under the False Claims Act. See 31 U.S.C. § 3729. The Act allows a private party to sue for violations on behalf of the govern- ment; successful suits result in a payment to the initiator. Id. § 3730. A plaintiff such as Calderon must plead and ulti- mately prove four elements: 1) the defendant made a false
(iii) Address property deficiencies identified in the appraisal af- fecting the health and safety of the occupants or the structural integ- rity of the property in accordance with FHA requirements, or (iv) Ensure that the appraisal of the property serving as security for the mortgage loan satisfies FHA appraisal requirements, in accord- ance with § 203.5(e) [the regulation governing appraisals in the Direct Endorsement process]. 6 No. 22-1553
statement (falsity); 2) the defendant knew the statement was false (knowledge); 3) the false statement was material to the government’s payment decision (materiality); and 4) the false statement caused the government’s loss (causation). United States v. Molina Healthcare of Ill., Inc., 17 F.4th 732, 739–40 (7th Cir. 2021). Understanding the process by which federal mortgage in- surance is issued clarifies how a company such as Carrington might be held liable under the False Claims Act. The alleged false statements occur if Carrington wrongly certifies that a loan meets federal underwriting requirements. If HUD would have withheld federal insurance or issued an indemnification agreement had it known of the noncompliance, then the false certification of compliance is material to HUD’s payment de- cision. And if the loan defaults and HUD covers the cost of the default, then the false certification of compliance causes the government’s loss when the loan’s noncompliance is the fore- seeable cause of the default. C Calderon asserts that during her time at Carrington she observed “reckless and inappropriate underwriting practices at Carrington,” including the false certification of several loans as meeting HUD’s minimum underwriting guidelines. Essentially, she alleges, if HUD had known of the errors in Carrington’s loan files, it would not have endorsed those loans for federal insurance or, in the alternative, if all Carring- ton’s loans had been subjected to a full post-endorsement re- view, many of them would have been characterized as unac- ceptable and HUD would have issued indemnification agree- ments. Further, if Carrington had complied with HUD’s un- derwriting guidelines, fewer of its federally insured loans No. 22-1553 7
would have defaulted because only borrowers with appropri- ate risk levels would have received loans. To meet her evidentiary burden, Calderon provided a sample “re-underwrite” of 349 federally insured loans that Carrington issued between 2013 and 2015; all of those loans later defaulted. Calderon’s analysis identifies several alleged deficiencies in these loan files, to which we refer as the “re- viewed loans.” The flaws included instances where Carring- ton overstated the borrower’s income or provided insufficient documentation. Calderon also gives examples where the orig- inal underwriter improperly omitted borrower debt, failed to assess creditworthiness, permitted excessive debt-to-income ratios, or improperly assessed compensating factors—that is, borrower characteristics that can offset a bad credit score, such as documented cash reserves, no discretionary debt, sig- nificant additional income, or residual income. Calderon also offered herself as an expert and intended to testify about the shortcomings in Carrington’s quality control department, as well as about the elements of materiality and causation. But the district court excluded the bulk of her ex- pert opinion. After Calderon was precluded from testifying about certain aspects of materiality and causation, Carrington moved for summary judgment on just those elements, argu- ing that Calderon could not meet her evidentiary burden on the available record. The district court granted the motion, agreeing with Carrington that as a matter of law Calderon could not prove either materiality or causation. Calderon has appealed. 8 No. 22-1553
II Calderon first challenges the district court’s decision un- der Federal Rule of Evidence 702 to exclude portions of her own proffered expert testimony. Calderon also disputes the district court’s decision to admit the testimony of Carring- ton’s expert, Kori Keith. We decide independently whether the district court followed Rule 702’s framework. United States v. Brown, 973 F.3d 667, 703 (7th Cir. 2020). We review the court’s decision to admit or exclude expert testimony for abuse of discretion. Id. Federal Rule of Evidence 702 governs the admissibility of expert testimony and requires that the district court deter- mine that a witness is qualified as an expert by her “knowledge, skill, experience, training, or education.” Rule 702 then sets out four factors to determine admissibility: (a) the expert’s scientific, technical, or other specialized knowledge will help the trier of fact to understand the ev- idence or to determine a fact in issue; (b) the testimony is based on sufficient facts or data; (c) the testimony is the product of reliable principles and methods; and (d) the expert has reliably applied the principles and methods to the facts of the case. The district court held that Calderon could not offer her opinion about the effectiveness of Carrington’s quality control program because she never worked in the quality control de- partment, did not review any manuals or procedures govern- ing that department, and did not compare Carrington’s qual- ity control processes to those required by HUD. In so ruling, the court probably overdid it by insisting that Calderon No. 22-1553 9
needed to have worked in Carrington’s own quality control department to develop the necessary expertise. Nonetheless, its overall criticisms of Calderon’s methods are sound. Given that Calderon did not read any of the materials that detailed Carrington’s policies and practices, she is not qualified to opine on their effectiveness. The court was well within its rights to exclude this portion of her proffered expert testi- mony. The district court also excluded the bulk of Calderon’s opinions on materiality and causation, finding them “too speculative.” It reasoned that because Calderon never worked for HUD or in loan servicing, she could not opine about “what is material to HUD’s decision to pay out claims,” nor could she suggest that all of Carrington’s false statements caused a loss to HUD. But the district court did permit Calderon to tes- tify from her own experience and explain what happened when she saw HUD reject loans or require indemnification. Again, the district court placed too much weight on where Calderon worked. Rule 702 does not suggest that specialized knowledge can be developed only in certain ways, and the Supreme Court’s decision in Kumho Tire Co. v. Carmichael, 526 U.S. 137, 150 (1999), points in the opposite direction—experts and expertise come in many different forms. For present pur- poses, working for HUD is not the sole path for a proffered expert to become sufficiently knowledgeable about HUD’s decision-making processes or the causes of its payments on federal insurance claims. Many people with long careers in the residential mortgage industry can and do develop this ex- pertise. Nevertheless, the district court did not abuse its dis- cretion in its ultimate ruling. Both materiality and causation are determined through nuanced, multi-factored analyses. 10 No. 22-1553
The district court correctly recognized that Calderon cannot testify about the existence of a conclusive list of what is mate- rial to HUD’s insurance-endorsement decisions or requests for indemnification. Nor could she opine that certain defects always cause losses to HUD. On appeal, Calderon claims that the scope of her expert testimony was narrower than the district court intimated. For example, as to materiality, she says that her opinion would be “that a particular loan was insured with a material defect that she has seen in her experience to result in rejection when HUD is aware of the defect.” In our view, the district court’s order contemplated the admission of such testimony. The district court said that Calderon “can discuss her experience and why her loans have been rejected in the past” and that she could “give her opinions on [the reviewed] loans so long as they stem from her experience as an underwriter” and do not “in- vade[] the province of the jury.” Otherwise, the court was well within its discretion to confine Calderon’s expert opinions to the bounds of her actual experience. The court reasonably prevented her from offering more sweeping conclusions about materiality and causation. Finally, the district court admitted Carrington’s rebuttal expert, Kori Keith, even though Keith lacked some of the un- derwriting qualifications that Calderon has. Calderon argues that if Calderon’s testimony was excluded, Keith’s testimony must be excluded too. But the scope of Keith’s expert testi- mony was different from the scope of Calderon’s. Keith was called to testify about underwriting practices generally and the purpose of the federal mortgage insurance program. She also planned to rebut the process Calderon followed when she evaluated the reviewed loans. The district court did not No. 22-1553 11
abuse its discretion when it determined that Keith’s twenty- plus years of experience in the residential mortgage industry qualified her as an expert, and that her review of the com- plaint, the reviewed loans, and the applicable HUD regula- tions were sufficient to support her proffered opinion. III This brings us to the heart of the appeal: Calderon’s chal- lenge to the grant of summary judgment in favor of Carring- ton. Our examination of that decision is de novo. Dunlevy v. Langfelder, 52 F.4th 349, 353 (7th Cir. 2022). Summary judg- ment is appropriate only when “there is no genuine dispute as to any material fact and the movant is entitled to judgment as a matter of law.” Fed. R. Civ. P. 56(a). “We draw all justifi- able inferences in favor of the nonmoving party” and view the facts in the light most favorable to Calderon, the nonmovant. Scaife v. U.S. Dep’t of Veterans Affairs, 49 F.4th 1109, 1115 (7th Cir. 2022). Calderon’s primary evidence of both materiality and cau- sation comes from her review of the 349 loans. The district court concluded that her evidence failed to support both these elements. We look first at materiality, and then at causation. A Under the False Claims Act, a false claim is “material” if it has a “natural tendency to influence, or be capable of influ- encing, the payment or receipt of money or property.” 31 U.S.C. § 3729(b)(4). In Universal Health Services, Inc. v. United States ex rel. Escobar, the Supreme Court clarified that a false claim would be material in “two circumstances”: “(1) ‘[if] a reasonable man would attach importance to [it] in determin- ing his choice of action in the transaction’; or (2) if the 12 No. 22-1553
defendant knew or had reason to know that the recipient of the representation attaches importance to the specific matter ‘in determining his choice of action,’ even though a reasona- ble person would not.” 579 U.S. 176, 193 (2016) (alterations in original) (quoting the Restatement (Second) of Torts § 538 (1976)). Escobar did not end with the identification of those two cir- cumstances; it also provided additional guidance on how to approach materiality in these cases. For example, it confirmed that mere regulatory violations, even if the regulation is la- beled by the government as “material” to governmental deci- sion-making, are not automatically material for the purposes of the Act without additional evidence that the government “consistently refuses to pay claims in the mine run of cases based on noncompliance with the particular statutory, regu- latory, or contractual requirement.” Id. at 195. Additionally, “if the Government pays a particular claim in full despite its actual knowledge that certain requirements were violated, that is very strong evidence that those requirements are not material.” Id. The district court held that Calderon had failed to estab- lish materiality because she did not present evidence that would permit a reasonable factfinder to conclude that HUD viewed the alleged underwriting deficiencies as important. It also criticized her for failing to rebut the possibility that HUD was aware of Carrington’s potential violations but took no ac- tion. Although the question is close, we conclude that when all permissible inferences are properly drawn in Calderon’s favor, there is enough evidence of materiality to clear the summary judgment hurdle. No. 22-1553 13
Calderon proffered evidence that (if believed) would show that the deficiencies she identified in the reviewed loans are material under 24 C.F.R. § 203.255(g). As we noted earlier, that regulation explains that “serious and material viola- tion[s]” will require the lender to indemnify HUD because, but for those violations, “the mortgage loan should not have been approved and endorsed.” Id. § 203.255(g)(3). Relevant material violations include the failure to: (i) verify the creditworthiness, income, and/or employment of the mortgagor in accordance with FHA requirements; (ii) verify the assets brought by the mortgagor for payment of the re- quired down payment and/or closing costs in accordance with FHA requirements…. Id. § 203.255(g)(3)(i)–(ii). In the reviewed loans, Calderon found several instances of overstated income, improperly omitted debt, and insufficient documentation of assets, in- come, and debt—problems that mirror the material violations identified in the regulation. And even though Escobar in- structs that simply labeling these violations as “material” in the regulation does not establish materiality without further proof, that does not mean that the regulatory evidence is be- side the point. The regulations provide some guidance, in HUD’s own voice, about the false certifications that would improperly induce the issuance of federal insurance, and those are precisely the false certifications present here. Calderon also proffered evidence that HUD issued indem- nification agreements to Carrington when faced with similar loan deficiencies in the past. She did so with her own expert testimony. As the district court ruled, Calderon may testify that she has seen certain types of underwriting deficiencies in 14 No. 22-1553
her experience, and that those deficiencies resulted in re- quests for indemnification from HUD. On top of that, Calde- ron proffered a series of letters that HUD sent to Carrington, detailing material loan deficiencies that HUD identified dur- ing its post-endorsement technical review of nine Carrington loans. HUD explained that it would issue an indemnification agreement if the deficiencies were not addressed. The types of material deficiencies that HUD identified in those letters— improper sourcing of closing funds, failure to explain large deposits, inadequate closing funds, failure to analyze debts correctly, overstated income, and inadequate documentation for credit analysis—are akin to the deficiencies that Calderon identified in the reviewed loans. These letters therefore con- firm that HUD “refuses to pay claims … based on noncompli- ance” with these kinds of underwriting requirements. See Es- cobar, 579 U.S. at 195. The letters also show that Carrington was on notice about which false certifications were serious in HUD’s eyes and “had reason to know that the recipient of the representation attaches importance to the specific matter.” Id. at 193. Carrington responds that “HUD’s conclusion that a spe- cific defect is material when reviewing one loan cannot sup- port a reasonable inference that similar defects will always be material for all loans.” But this line of argument asks us to weigh the evidence, a task that is best reserved for a fact- finder. For example, Calderon says that overstated income in a particular borrower’s loan file is a serious deficiency; Car- rington responds that overstated income occurs in varying degrees and does not, in every instance, constitute major non- compliance. Carrington’s rebuttal does not entitle it to sum- mary judgment, because the degree of the deficiency is genu- inely disputed. No. 22-1553 15
Similarly, Carrington points out that it was able to miti- gate several of the deficiencies identified in the letters and was forced to indemnify only one loan in the nine that were reviewed. It argues that the relative ease with which it could have mitigated any alleged deficiency is significant because, once mitigated, HUD will make payments on any insurance claims; HUD’s willingness to issue payment undermines the materiality of the violation. But again, the fact-intensive job of evaluating the identified deficiencies and deciding which ones could have been mitigated is a job best left to the jury. At summary judgment, the question is only whether Cal- deron has proffered sufficient evidence to warrant a reasona- ble inference in her favor. We do not doubt that Carrington might be able to convince a rational trier of fact that the de- fects in the reviewed loans are factually distinguishable from the material defects identified in HUD’s letters or are capable of correction. But that does not warrant summary judgment. Calderon has identified several false certifications in 349 loans, she has shown that those false certifications are mate- rial according to federal regulations and her own underwrit- ing experience, and she has proffered evidence that HUD has in the past deemed similar violations material. A factfinder would be permitted to infer materiality from this evidence. We are also concerned that the district court placed too much weight on its belief that HUD knew about Carrington’s false certifications. If HUD had knowledge of the false certifi- cations and nonetheless issued the insurance (or refrained from demanding indemnification), Escobar instructs that this would be strong evidence that the certifications were not ma- terial. The district court was impressed by HUD’s pre-en- dorsement review of all Carrington’s loan files, the selective 16 No. 22-1553
post-endorsement technical reviews of a sizable sample, and Calderon’s own allegations, which she reported to HUD when she initiated this lawsuit. But the extent of HUD’s knowledge is contested. The rec- ord does not establish that HUD’s pre-endorsement review would have revealed widespread underwriting violations. HUD’s own regulations suggest that the pre-endorsement re- view looks only to see if the required documents are all in the file; the content of those documents is not examined. For that matter, HUD’s post-endorsement reviews provide infor- mation on only a select sample of Carrington’s loans and do not show that HUD possessed a broad awareness of all al- leged deficiencies in Carrington’s federally insured loans. Fi- nally, the record is unclear as to how much HUD learned from Calderon; in her deposition, she could not remember what she had told government representatives in those initial con- versations. In United States ex rel. Yannacopoulos v. General Dynamics, we rested our materiality ruling on the fact that the agency had “actually learned of the supposed misrepresentation.” 652 F.3d 818, 831 (7th Cir. 2011). Such certainty of knowledge is a far cry from the district court’s conclusion here that “HUD was likely aware of any violation.” To infer HUD’s knowledge is to draw an inference in the moving party’s favor from a heavily disputed set of facts. A juror could reasonably conclude that HUD was unaware of Carrington’s alleged cul- ture of reckless underwriting. The extent of HUD’s knowledge is thus still up for grabs; that issue was not suita- ble for summary judgment. No. 22-1553 17
B Turning now to causation, we must briefly address the False Claims Act’s damages provisions. Some circuits have in- terpreted the Act as creating “two sorts of liability.” See United States ex rel. Schwedt v. Planning Research Corp., 59 F.3d 196, 199 (D.C. Cir. 1995). As they see it, the first form of liabil- ity can be found in the Act’s civil penalty, which, according to Schwedt, may be imposed “regardless of whether the submis- sion of the claim actually causes the government any dam- ages; even if the claim is rejected, its very submission is a basis for liability.” Id. The second form of liability they identify is the Act’s provision of treble damages, which is triggered only “for damages that the government sustains because of the submission of the false claim.” Id. Under this understanding, Calderon’s claims could move forward as claims for civil pen- alties, based on her satisfaction of the other three elements of a claim under the Act. Though Calderon suggested to the district court that she should proceed to trial even without sufficient evidence of causation (i.e., by using the first approach mentioned above), she has not renewed that argument on appeal. Because she has not asked us to do so, we do not consider the possibility of a claim limited to civil penalties. Whether we agree with the Schwedt approach is a question we leave for another day. In order to avoid summary judgment, therefore, Calderon had to proffer evidence covering all four elements of the claim, including causation. Indeed, since United States v. Luce, we have required a plaintiff to establish both actual and prox- imate cause to recover under the Act. 873 F.3d 999, 1014 (7th Cir. 2017). Carrington’s false certifications to HUD must not only be material, they also must cause a foreseeable harm: “a 18 No. 22-1553
type that a reasonable person would see as a likely result of his or her conduct.” Id. at 1012 (emphasis in original) (quoting Blood v. VH-1 Music First, 668 F.3d 543, 546 (7th Cir. 2012)). Simple but-for causation is not enough. To show proximate causation, Calderon had to put for- ward evidence indicating that the false certifications in the re- viewed loans were the foreseeable cause of the later defaults. We focus on the defaults because that is what triggers HUD’s payment obligations under the federal insurance program. We recognize that when we adopted the proximate-cause standard in Luce, we did not explicitly state that proving prox- imate cause in cases about federal mortgage insurance re- quires proving the causes of defaults. We did, however, rely heavily on the Fifth Circuit’s reasoning in United States v. Mil- ler, 645 F.2d 473 (5th Cir. 1981), and the Third Circuit’s rea- soning in United States v. Hibbs, 568 F.2d 347, 351 (3d Cir. 1977). In both of those cases, the courts made explicit state- ments about the need to prove what caused the defaults. In Miller, the Fifth Circuit held that “[i]n the context of a federal housing case, the United States must show that the false statements in the application were the cause of subse- quent defaults.” 645 F.2d at 476. And in Hibbs, the Third Cir- cuit said that because it is the default that causes the loss to the United States, a plaintiff must show some connection be- tween the false certifications and the default. 568 F.2d at 351. Where a default is caused “by a flood or some other unin- sured catastrophe,” a defendant’s false certifications cannot be said to have caused the government’s loss. Id. This focus seems sound to us. To ensure that the false certifications were a substantial factor in bringing about HUD’s losses and that the losses were foreseeable to the defendant, the plaintiff must No. 22-1553 19
show that the false certifications played some role in causing or increasing the risk of a subsequent default. The parties dispute how Calderon might meet that bur- den. Calderon argues that she should be allowed to extrapo- late causation from a generalized statistical analysis of Car- rington’s federally insured loans. Carrington responds that Calderon had to proceed loan-by-loan through the 349 loans and show how each allegedly false statement caused each loan’s default. But because we find that Calderon did not meet her burden under either method, we do not resolve which method was appropriate. We consider the statistical method first. Calderon claims to have offered proof of causation in the form approved in United States v. Hodge, 933 F.3d 468 (5th Cir. 2019). There the Fifth Circuit found sufficient evidence to support a jury ver- dict against the defendant where the government showed that the defendant maintained a culture of “reckless under- writing” and that this culture resulted in elevated default rates for its federally insured loans when contrasted with the national default rate for such loans. Id. at 475–76. Calderon and Carrington dispute whether Calderon has provided suf- ficient proof of reckless underwriting practices at Carrington. But Calderon can avail herself of this method of proof only if she proffers evidence on which a trier of fact could rely to find that Carrington had an elevated default rate. To support this fact, Calderon points us to Carrington’s 2014 annual financial report, where it disclosed a total fore- closure rate of 10.59 percent. That rate is 500 percent higher than the national foreclosure rate of 2.15 percent, and so it might seem to be a strong support for her position. But the problem is that the report provides little information about 20 No. 22-1553
the performance of Carrington’s federally insured loans. Cal- deron needed to identify the total number of federally insured loans that Carrington serviced, determine Carrington’s de- fault rate for those loans, and then compare that to the na- tional default rate of federally insured loans. Having failed to do that, Calderon’s attempted statistical analysis falls short. Without some evidence that Carrington had a higher-than- average default rate for its federally insured loans, a jury could not find that Carrington’s underwriting practices, reck- less or not, had any effect on subsequent loan performance. Turning to the loan-by-loan method, Calderon proffers her analysis of the 349 reviewed loans, in combination with Carrington’s documented reason for each loan default. The reason for default is recorded in code form. Calderon asserts that the codes, which identified problems such as “Excessive Obligations,” were plain-English descriptions that were ac- cessible to any jury. Perhaps so, but that was not the problem with them. The sticking point is vagueness: the codes do not contain any information that would permit a reasonable fact- finder to determine the cause of default. For example, as Car- rington argues, a code such as “Curtailment of Income” would not explain whether the default was caused by a false statement regarding income in the borrower’s file or by recent unemployment. Calderon proffered no expert who could interpret the codes, explain common causes of defaults, or discuss the de- faults in the reviewed loans. Nor has Calderon herself under- taken the kind of analysis of the loan servicing records that would permit her to opine to the jurors about the events that caused a “Curtailment of Income” and whether those events were related to the initial false statement. While the district No. 22-1553 21
court agreed to let Calderon testify about the underwriting errors that led in her experience to defaults, she has not shown how that experience-based testimony would help a jury. For example, Calderon’s re-underwrite of Borrower 14’s loan file revealed an inflated income and a default code of “Curtail- ment of Income.” But a lay factfinder reviewing the loan file would find no evidence that would allow her to identify whether the reported income was spot on, too high, or too low. Pay stubs in the file revealed what a reasonable person might calculate as $7,801.40 in monthly income; the total in- come reported by Carrington was $7,659. Calderon’s re-un- derwrite provides scant guidance on how to assess or recreate her findings. Further, that same lay factfinder would have no information about what “Curtailment of Income” meant for Borrower 14 and why that borrower stopped being able to make their monthly payments. The inference that Calderon is asking the jury to draw as to causation in loan files like this one is speculative and impermissible. Other courts have noted that misrepresentations about a borrower’s income, debt, and assets foreseeably increase the risk of default by their very nature. See, e.g., United States v. Spicer, 57 F.3d 1152, 1159 (D.C. Cir. 1995) (holding that the de- fendant’s intentional misrepresentations about the size of bor- rower down payments foreseeably caused HUD’s losses when the loans defaulted, even where there was limited evi- dence to explain each default and even though other factors might have played a role). But we are reluctant to extend that reasoning to this case. On the present record, it is not clear how a factfinder would even spot the alleged false statement in each loan file, let alone evaluate its seriousness and scope. And though Calderon asserts that the misrepresentations in this case are of the type identified in Spicer, we do not see 22 No. 22-1553
much in the record to support that point other than Calde- ron’s assertions. Without more evidence from which a jury could conclude that Carrington’s alleged misrepresentations in each loan caused the subsequent defaults, the nature of those misrepresentations is not enough to get past summary judgment. IV Because Calderon did not proffer evidence that would permit a reasonable trier of fact to find that Carrington’s vio- lations of the False Claims Act caused any harm to HUD, we AFFIRM the judgment of the district court.