UNITED STATES DISTRICT COURT FOR THE DISTRICT OF COLUMBIA
July 29, 2009
NATIONAL ASSOCIATION OF MORTGAGE BROKERS, INC., PLAINTIFF,
SHAUN DONOVAN, SECRETARY, U.S. DEPARTMENT OF HOUSING AND URBAN DEVELOPMENT, DEFENDANT.
The opinion of the court was delivered by: James Robertson United States District Judge
The National Association of Mortgage Brokers, Inc. (NAMB) sues Shaun Donovan, the Secretary of the United States Department of Housing and Urban Development (HUD), asserting arbitrary and capricious action in his promulgation of a "Rule to Simplify and Improve the Process of Obtaining Mortgages and Reduce Customer Settlement Costs," 73 Fed. Reg. 68,204 (Nov. 17, 2008) ("Final Rule"). The parties have cross-moved for summary judgment. The defendant's motion will be granted.
Among its many provisions, the Final Rule creates a new standardized form that originators of federally related mortgages*fn1 must provide to prospective borrowers. HUD asserts that the new form provides borrowers with a clearer picture of loan terms and conditions, allowing them to compare loan offers more easily, and promoting competition between loan providers. NAMB argues that the Final Rule is procedurally and substantively deficient: procedurally, because HUD developed the new form with the help of data that it did not disclose during the notice-and-comment period; and substantively, because HUD did not adequately explain why it imposed disclosure requirements on mortgage brokers different from those required of direct lenders.
A few words about terminology: Mortgage loan originators are either direct lenders or brokers -- direct lenders originate loans and provide funds, while brokers originate loans and arrange for third parties to fund them.*fn2 See 24 C.F.R. § 3500.2. The various tasks that go into processing an application and originating a loan -- title searches, title examinations, credit examinations, real estate appraisals, and so on -- are known as "settlement services," for which the loan originator charges and the borrower pays a settlement fee. See 12 U.S.C. § 2602(3). Under long-standing federal law, loan originators must inform prospective borrowers of loan terms and settlement fees on a "good-faith estimate" form, or "GFE." See 24 C.F.R. Pt. 3500 App. A & B. The Final Rule at issue in this case creates a new GFE.
Broadly speaking, NAMB opposes the new GFE because of the way it discloses "yield spread premiums" (YSPs). A YSP is a payment by a lender to a broker that compensates the broker for originating a loan with an "above-par" interest rate. The "par rate" is the interest rate at which the lender will fund 100% of the loan with no premiums or discounts. Administrative Record [hereinafter, "A.R."] 241, 1636. If the par rate for a certain $100,000 mortgage is, say, 5%, and the broker originates that mortgage at a 5.5% interest rate, then the lender might deliver $100,500 at closing -- $100,000 that will be disbursed to the borrower, plus a $500 YSP for the broker.
YSPs can benefit certain borrowers. Consider a borrower who wants a mortgage but is unable to pay the entire settlement fee at closing.*fn3 If she is working with a direct lender, she might agree to pay a higher rate of interest on her loan in order to reduce her settlement fee. In such a case, the lender trades off the smaller up front fee for larger monthly interest payments, or, as we have seen in recent years, for the ability to package and sell a more valuable loan on the secondary market. The broker, unlike the direct lender, gets no benefit from the higher-interest loan. It is the YSP that gives him the incentive to accept a lower settlement fee from the borrower.
Borrowers only benefit from YSPs, however, if they can understand and make intelligent choices about the tradeoffs between short-term settlement costs and long-term interest payments. That was the idea behind the Real Estate Settlement Procedures Act(RESPA), 12 U.S.C. §§ 2601-2617, which provides mortgage borrowers "with greater and more timely information on the nature and costs of the settlement process." Id. § 2601(a). Section 4 of RESPA requires the HUD Secretary to "develop and prescribe a standard form for the statement of settlement costs which shall be used . . . as the standard real estate settlement form in all transactions in the United States which involve federally related mortgage loans." Id. § 2603(a). "Such form shall conspicuously and clearly itemize all charges imposed upon the borrower and all charges imposed upon the seller in connection with the settlement." Id.
The Final Rule is the result of a very long administrative process motivated in part by HUD's dissatisfaction with the way the existing GFE disclosed the impact of YSPs on loan terms. In HUD's view, "YSP payments [were] not required to be included in the calculation of the broker's total charge for the transaction, nor [were] they clearly listed as an expense to the borrower, even though the borrower promise[d] to pay the YSP through interest payments." Dkt. 15, at 10.
Thus, in 1995, HUD began the process of designing a new GFE that, among other things, would more clearly demonstrate the inverse relationship between settlement fees and interest rates, and would allow borrowers to compare competing loan offers quickly and accurately. HUD issued a proposed rule and sought public comment on alternative approaches to disclosing various indirect fees, including YSPs. A.R. 1783-90. The HUD Secretary convened a negotiated rulemaking advisory committee consisting of industry groups (like NAMB), consumer groups, state organizations, and government-sponsored enterprises. Id. 1777. Though this initial process reached a dead end, it laid the foundation for a proposed rule issued in 2002. That rule would have revised the existing GFE, simplified and standardized the disclosure of settlement costs, and modified other disclosure requirements. Id. 91. After receiving over 40,000 comments --mostly supportive of HUD's goals -- HUD decided to withdraw the rule in 2004 and gather additional information about settlement costs and the options for improving disclosure. Id.
HUD's first step was to assess the strengths and weaknesses of the existing GFE. It hired the Urban Institute to conduct a study of 7,560 home loans. Id. 3119. The study found that borrowers could not describe the basic aspects of their mortgages, particularly in more complex transactions involving YSPs. Id. 223-24. Unwittingly, borrowers tended to accept small reductions in their settlement fee in exchange for larger increases in their interest payments, resulting in significantly higher costs over the life of the loan. Id. 275, 316. In HUD's view, the Urban Institute study, combined with much of the research in the field, demonstrated that the existing GFE was too opaque for most borrowers to understand. Id. 246-67.
HUD then tried to design a more effective GFE. Between 2002 and 2008, HUD had the Kleimann Communications Group (KCG) conduct seven rounds of consumer testing. In the first three rounds, KCG tested HUD's draft GFE by asking current and potential homeowners for their thoughts on the GFE's clarity.
Id. 360, 2092. HUD discovered that customers better understood the tradeoff between settlement costs and interest rates if YSPs were disclosed more prominently. Id. 23, 364.
But highlighting YSPs had the potential to give direct lenders a competitive advantage over brokers. HUD was concerned that borrowers might mistakenly believe that loans originated by brokers were more expensive than comparable loans originated by direct lenders, because only brokers would have to disclose the YSP payment.
Accordingly, in rounds four and five of the consumer testing, HUD focused on creating a form that would disclose YSP payments while avoiding anti-broker bias. Round four found that the proposed GFE failed to meet that standard: borrowers saw the YSP as a cost unique to broker-originated loans, and tended to choose more costly lender-originated loans. Id. 2065-66. In response, HUD made several revisions to the GFE, including requiring lenders to check a box stating, "The credit or charge for the interest rate you have chosen is included in 'Our service charge' (See Item 1 above)." Id.
Those revisions appeared to work. In round five, KCG found no evidence of bias against brokers when borrowers were asked to choose between a selection of loans with different terms. Id. 2094. Prospective borrowers identified the lowest cost loan at least 90% of the time, and chose that loan 88% of the time, regardless of whether the loan was offered by a lender or a broker. Id. 2067, 2218-23.
In round six, KCG did more qualitative testing, prompting HUD to make some minor modifications to the GFE. Id. 370, 2070-71, 2094. HUD concluded that round six affirmed that the proposed GFE produced no anti-broker bias: participants were still able to correctly identify the lowest-cost loan 90% of the time. Id. 2034.
Satisfied, the HUD Secretary published a Notice of Proposed Rulemaking on March 14, 2008, which solicited public comments on all aspects of the Rule, including the proposed GFE. Id. 113. HUD received approximately 12,000 comments from interested parties. Id. 9096-104,866. HUD revised the proposed GFE in response to some of these comments, and subjected the new form to an additional round of testing -- round seven. In round seven, KCG found that prospective borrowers could identify the least expensive loan only 80% of the time -- less reliably than in rounds five and six. Id. 2039. Although KCG concluded that the decline between rounds five and six and round seven was not statistically significant, HUD chose to eliminate many of the changes it had made to the GFE in response to public comments, and return the form to the format that had tested well before. Id. 2039. HUD did not publish the methodology and results of round seven for public comment.
The Secretary issued the Final Rule on November 17, 2008. The final GFE is three pages long. The first page provides a summary of the loan terms and settlement costs, and makes no mention of YSPs or other specific settlement charges. Id. 54. The second page contains a chart titled "Understanding your estimated settlement charges," which is divided into two parts: one breaking down "adjusted origination charges," and the other enumerating the charges for "all other settlement services." Id. 55. A broker is required to include the YSP payment it will receive from a lender as part of its origination charge for the loan. For two identical loans, this requirement means that the broker-originated loan will show a higher adjusted origination charge on page two than will a loan made by a direct lender. But because the YSP should offset the charges for the other settlement services the broker offers, a broker in a competitive market should disclose a commensurately smaller charge in the "all other settlement services" section of the GFE. The result should be that the broker and the direct lender disclose the same amount of overall settlement charges on the first page of the form.
The third and final page of the GFE features a "tradeoff table" and a "shopping chart." Id. 56. The tradeoff table is optional for both brokers and lenders. It presents the prospective borrower with information about the settlement costs and monthly interest payments for three potential loans from the same originator: (1) the loan spelled out in detail in the GFE, (2) that same loan with lower settlement charges and a higher interest rate, and (3) that same loan with higher settlement charges and a lower interest rate. The shopping chart is to be filled out by the prospective borrower. It provides a template for the borrower to compare loan offers across different originators.
The Final Rule requires loan originators to begin using the new GFE by January 1, 2010. Id. 2.
Under the APA, a reviewing court will declare agency action unlawful if it is "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law." 5 U.S.C. § 706(2)(A). NAMB argues that the Final Rule is arbitrary and capricious for three reasons: first, because HUD failed to supply a reasoned explanation for the Rule; second, because HUD failed to consider reasonable alternatives to the Rule and explain why it rejected them; and third, because HUD relied substantially on data that it never produced for public comment.
"The scope of review under the 'arbitrary and capricious' standard is narrow and a court is not to substitute its judgment for that of the agency." Motor Vehicle Mfrs. Ass'n of U.S., Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983). The party challenging the agency action must show that there was no "rational connection between the facts found and the choice made." Id. (citation omitted). NAMB has not done so.
A. Failure to Provide a Reasoned Explanation
NAMB contends that HUD does not offer a reasoned explanation for its asymmetrical disclosure requirements. NAMB argues that if brokers are required to disclose YSP payments, then direct lenders should have to disclose any premiums they earn from selling a higher-interest loan on the secondary market. In NAMB's view, "[i]n both cases, the compensation is derived in the same way using the same information, and can be calculated at the time that a mortgage is originated." Dkt. 16, at 6 (emphasis in original).
But HUD has offered a reasoned explanation for its decision -- and a compelling one. As an initial matter, HUD submits that disclosing YSPs is necessary to achieve its ultimate goal of creating a more competitive mortgage market:
The third round of testing did not include the YSP disclosure, and the important finding was that, without the YSP disclosure, consumers did not understand the existence of the tradeoff between interest rates and origination charges as well as when the YSP was disclosed. Helping consumers understand this tradeoff is a fundamental goal of HUD's RESPA reform effort and of the design of the GFE form. The third round of testing confirmed that inclusion of the YSP disclosure helped consumers understand the tradeoff, and that if they take a loan with a relatively high interest rate, they should pay lower settlement charges.
HUD then offers two explanations for its decision not to compel direct lenders to disclose the secondary market premiums they receive. The first is that RESPA only requires the disclosure of all charges imposed "in connection with the settlement." 12 U.S.C. § 2603(a). It similarly requires that the GFE include an estimate of charges "for specific settlement services the borrower is likely to incur in connection with the settlement." Id. § 2604(c). HUD argues that the premiums a direct lender may receive on the secondary market are separate from any settlement-related costs, and that it cannot lawfully mandate their disclosure under RESPA. A.R. 106.
The second explanation is that direct lenders cannot disclose the premiums they may receive on the secondary market because those premiums are unknown at settlement. Unlike YSPs, which are known and definite at settlement, the value of any premium a direct lender may earn depends on when the lender sells the loan, what the market conditions are like, and other factors that are not established with certainty at settlement. Id. 106, 271, 385-86. Direct lenders cannot be expected to disclose what they do not know.
While HUD makes a persuasive case for clear YSP disclosure by brokers, and for exempting direct lenders from disclosing premiums they have not yet realized, its most difficult task is to show that the new GFE can do both of those things without producing anti-broker bias. As HUD itself admits, the ultimate goal of the new GFE is to promote competition in the mortgage industry. Id. 23. If the new GFE distorts the marketplace by providing an artificial advantage to direct lenders, it would scuttle HUD's reform effort, and run afoul of the APA's requirements.
There is certainly evidence in the administrative record that raises doubts about the desirability of asymmetric disclosures. A 2004 study conducted by the Federal Trade Commission (FTC) found that an "asymmetric disclosure policy will cause a large proportion of consumers to view identically priced loans from brokers and lenders very differently." James M. Lacko & Jamis K. Pappalardo, Fed. Trade Comm'n, The Effect of Mortgage Broker Compensation Disclosures on Consumers and Competition: A Controlled Experiment ES-7 (2004) (A.R. 5268). The study warned that the resulting market distortion could cause "mortgage consumers to incur additional costs of hundreds of millions of dollars per year." Id. 5286. Both NAMB and the FTC cited the 2004 FTC study, and echoed its concerns, in their comments opposing HUD's proposed revisions to the existing GFE. See FTC Comment, at 24 (A.R. 24207); NAMB Comment, at 32-33 (A.R. 15923-24).
HUD adequately addresses these concerns, however. It notes first that the FTC study tested only excerpted portions of a draft GFE found in the 2002 proposed rule. A.R. 365. By contrast, in round five of consumer testing -- which found no statistically significant anti-broker bias -- HUD evaluated a different version of the GFE, and gave the study participants the entire GFE, not just excerpts. Id. 366. Thus, the FTC study does not directly undermine the study on which HUD bases the new GFE.
The FTC does take issue with HUD's claim that round five uncovered no anti-broker bias. In its Comment, the FTC notes that while anti-broker bias "does not appear in the results . . . in which the broker loan . . . is cheaper than the lender loan, it does still appear in the tests in which the broker and lender loans cost the same." Id. 24203. The FTC observes that round five showed a "17-18 percentage point bias in the proportion of customers preferring the lender loan even though the two loans had the same interest rates, closing costs, and payments." Id. 24204.
Once more, HUD offers a credible rebuttal. It points out that the 17-18 percentage point bias was only found in the first iteration of round five testing, when the study participants were given a draft GFE that did not have a shopping chart. When the shopping chart was included in subsequent iterations of round five testing, the results actually suggested a slight pro-broker bias. Id. 2068-69, 2466, 2502.
And HUD adequately addresses the finding that borrowers tended to select the lender loan over the broker loan when the loam terms were the same. The Urban Institute study found that, in the market, origination charges vary considerably even among similar loans with identical interest rates, id. 3150-51, so borrowers are rarely offered a choice among identical loans. Id. 2067, 2466. HUD notes that, under more realistic conditions, when prospective borrowers are asked to distinguish between disparate loan offers, consumers were able to identify the least expensive loan 92% of the time, and to select that loan at nearly identical rates whether it was offered by a lender or a broker. Id. 2464, 2466, 2494-95.
HUD passes APA review "[s]o long as [it] has examined the relevant data and provided a reasoned explanation supported by a stated connection between the facts found and the choices made." North Carolina v. FERC, 112 F.3d 1175, 1189 (D.C. Cir. 1997). Though the FTC, NAMB, and other interested parties expressed their doubts about asymmetric disclosure requirements, HUD provided ample evidence to support the efficacy of its chosen GFE. It is not my role to second-guess that well-supported finding.*fn4
B. Failure to Consider Reasonable Alternatives
"It is well settled that an agency has a 'duty to consider responsible alternatives to its chosen policy and to give a reasoned explanation for its rejection of such alternatives.'" City of Brookings Mun. Tel. Co. v. FCC, 822 F.2d 1153, 1169 (D.C. Cir. 1987) (citations omitted). NAMB contends that HUD refused to consider two reasonable alternatives: requiring symmetrical disclosures for brokers and lenders or postponing promulgation of the Rule "to conduct further research into the operation of mortgage markets and the efficacy of the proposed disclosures." Dkt. 16, at 33.
I have already catalogued HUD's stated reasons for rejecting symmetrical disclosures. As for the virtues of conducting further research, the mass of academic studies and consumer testing that HUD has already compiled -- six years of study and seven rounds of testing -- is enough.
C. Failure to Produce Data on Which it Relied
NAMB argues that HUD violated the APA by not publishing the results of round seven of consumer testing before issuing the Final Rule. "An agency commits serious procedural error when it fails to reveal portions of the technical basis for a proposed rule in time to allow meaningful commentary." Owner-Operator Indep. Drivers Ass'n, Inc. v. Fed. Motor Carrier Safety Admin., 494 F.3d 188, 199 (D.C. Cir. 2007). But the APA permits an agency to "use 'supplementary' data" during and after the notice and comment period as long as that data "expands on and confirms information contained in the proposed rulemaking and addresses 'alleged deficiencies' in the pre-existing data, [and] so long as no prejudice is shown." Chamber of Commerce v. SEC, 443 F.3d 890, 900 (D.C. Cir. 2006) (citations omitted).
HUD conducted six rounds of consumer testing before publishing its Notice of Proposed Rulemaking. In response to some of the comments it received, it revised the GFE and subjected it to a seventh round of testing. Round seven found that the proposed GFE was more effective than the revised one: prospective borrowers were able to identify the lower-cost loan 90% of the time using the proposed GFE, and only 80% of the time using the revised version. A.R. 2039. Accordingly, HUD eliminated many of the revisions so that the final GFE closely resembled the version that had tested well in rounds five and six. Id.
The very purpose of the notice-and-comment process is to encourage the agency to consider revisions to its proposed rule before issuing the final version. HUD did just that, and took the extra precaution of testing those revisions before promulgating the Final Rule. When those revisions proved ineffective, HUD scrapped them. The process worked as intended.
Indeed, to accept NAMB's argument would place agencies in an untenable position. They could either: (1) refuse to make changes to the proposed rule in response to public comments, forcing them to explain why they rejected suggestions they may have actually found helpful; (2) make changes in response to public comments, but not subject those changes to any testing, forcing them to justify the changes without any empirical data; or (3) open a new comment period after any additional testing they conduct, leading to an endless cycle of revisions and comments, see Cmty. Nutrition Inst. v. Block, 749 F.2d 50, 58 (D.C. Cir. 1984) ("Rulemaking proceedings would never end if an agency's response to comments must always be made the subject of additional comments."). That is why the Court of Appeals has approved the use of "supplementary data" -- such as round seven here -- during the comment period.
NAMB has failed to meet its burden of showing that HUD acted arbitrarily or capriciously in promulgating the Final Rule. Accordingly, the defendant's motion for summary judgment [#15] will be granted by the accompanying order.