National Association of Mortgage Brokers, Inc. v. Preston

UNITED STATES DISTRICT COURT
FOR THE DISTRICT OF COLUMBIA

NATIONAL ASSOCIATION OF MORTGAGE :
BROKERS, INC., :
:
Plaintiff, :
:
v. : Civil Action No. 08-2208 (JR)
:
SHAUN DONOVAN, Secretary, U.S. :
Department of Housing and Urban :
Development, :
:
Defendant. :

MEMORANDUM

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.

Background

Among its many provisions, the Final Rule creates a new

standardized form that originators of federally related

mortgages1 must provide to prospective borrowers. HUD asserts

that the new form provides borrowers with a clearer picture of

1
A “federally related mortgage” is essentially any loan for
residential property that is insured by the federal government,
or that it is originated by an entity that is regulated by the
federal government. See 12 U.S.C. § 2602(b).
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.2 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

2
This distinction is a transactional one: an entity may
call itself a mortgage broker but provide funding for a loan that
it originates. Such an entity is considered a direct lender,
while an entity that holds itself out as a direct lender is
considered a broker when it does not fund the loan it originates.

- 2 -
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.3 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

3
This is a common occurrence. As HUD had observed, “[o]ne
of the primary barriers to home ownership and homeowners’ ability
to refinance and lower their housing costs is the up front cash
needed to obtain a mortgage,” which goes to pay for “closing
costs and origination fees association with a mortgage loan.”
2001 Statement of Policy, 66 Fed. Reg. at 53,053-054 (A.R. 1505-
06).

- 3 -
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

- 4 -
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

- 5 -
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.

- 6 -
 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.

- 7 -
 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

- 8 -
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.

- 9 -
 The Final Rule requires loan originators to begin using

the new GFE by January 1, 2010. Id. 2.

Analysis

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

- 10 -
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.

A.R. 23.

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

- 11 -
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.

- 12 -
 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

- 13 -
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

- 14 -
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.4

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

4
NAMB’s claim that HUD failed to explain why it made the
tradeoff table optional is also unavailing. HUD cannot legally
compel lenders to provide information on other loan options that
might be available to the borrower. A.R. 2041. And, in any
event, consumer testing revealed that there would be market
pressure on lenders to fill out the tradeoff table, and that even
a blank tradeoff table helped borrowers understand the
relationship between settlement costs and interest rates. Id.
2041-42.

- 15 -
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

- 16 -
‘alleged deficiencies’ in the preexisting 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

- 17 -
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.

Conclusion

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.

JAMES ROBERTSON
United States District Judge

- 18 -