Case Study 24.1 — The Marketing Services Agreement: A Decade of Section 8 Enforcement

A real, public enforcement arc. Every matter named below is a public action with a public order. This case study describes theory and pattern only. It prints no settlement amounts, no penalty figures, and no consent-order terms — partly because this book does not fabricate them, and partly because the figures are the least useful thing in the record. What you need is the shape of the conduct the Bureau found, because you will be offered that shape.

By 2010 the residential mortgage industry had run a thirty-six-year experiment in Section 8 compliance and arrived at a stable answer.

Cash for referrals: unlawful, obviously, since 1974. Buying the referral source: lawful, if you survive the affiliated business arrangement's three conditions — disclosure, no required use, return on ownership only — and if the entity is not a sham. That left one door, and it was the widest one in the statute.

Section 8(c) permits payment for goods actually furnished or services actually performed. Advertising is a service. Therefore: enter a written agreement under which a real estate brokerage performs marketing services for a lender — signage in the office, a banner on the brokerage's website, an insert in the client newsletter, materials distributed at open houses — and pay a monthly fee.

The structure is genuinely lawful. A brokerage with a real website and real traffic is selling real advertising, and a lender that buys it at what such advertising costs has bought a service. Nothing in RESPA prohibits that, and nothing in this case study suggests otherwise.

The problem is that the identical document can describe an arrangement in which no marketing happens at all. From outside, a lawful MSA and an unlawful one look the same: a contract, an invoice, a monthly payment. The difference lives in facts the document does not contain — whether the services were performed, whether the price came from the market or from the referral count, and whether the payment survives when the referrals stop.

Between roughly 2013 and 2017 the Consumer Financial Protection Bureau went looking for those facts, and what it found became the enforcement record that still governs how your compliance department thinks.

The issue: five patterns, found repeatedly

Read across the public actions and the same five features recur. They are not a legal test — the legal test is the statute — but they are the pattern that draws attention, and any one of them is enough to lose an argument you thought you were winning.

1. The payment tracked referrals rather than services. The clearest version: a fee renegotiated when volume changed, or quietly reduced in a quarter when the referrals slowed, or increased after a brokerage's production rose. The service scope did not change. Only the referral count did. Regulation X § 1024.14(e) treats a thing of value received repeatedly and connected in any way to the volume or value of referrals as evidence of an agreement or understanding — which is precisely the element Section 8(a) requires and the one everyone assumed could not be proven without a smoking gun.

2. The services were never performed, or were performed nominally. Agreements listed deliverables. No one produced them. No one asked for proof. In matter after matter, the party paying could not show what it had bought, because nobody had ever created a record — and a party that cannot demonstrate the service was performed has, functionally, made an unearned payment. That is Section 8(b) territory as well as 8(a): a charge for which no or only nominal services are performed is an unearned fee under § 1024.14(b).

3. Fair market value was established after the price. The recurring artifact is a valuation opinion dated after the agreement's effective date, or built on assumptions supplied by the party who wanted a particular answer. The point of an independent valuation is to prove the price came from the market. A valuation obtained afterward proves the opposite: that the number came first and the justification was procured.

4. The MSA was the price of admission. Where a brokerage's business was, in substance, available to whoever signed, the agreement was not buying advertising. It was buying position — and position purchased from a person who controls referrals is the thing the statute names.

5. The arrangements were stacked. MSAs bundled with desk license agreements, with lead agreements, with sponsorships and event funding, so that no single item looked large while the total flowing to one referral source was substantial. Compliance reviews that examined one agreement at a time saw nothing. Examinations that summed all payments to a single counterparty saw something else.

The record: what actually happened, in order

2014 — a title agency's marketing services agreements. The Bureau entered a consent order with Lighthouse Title, Inc., a Michigan title insurance agency, concerning marketing services agreements it had entered with companies that referred business to it. The Bureau's theory was that the payments were connected to referrals and exceeded the value of any marketing services actually performed. This was the first widely-read signal that "we have an MSA" was not a conclusion.

2015 — Genuine Title, and the part every loan officer should read twice. The Bureau, together with the Maryland Attorney General, brought actions arising from the practices of Genuine Title, a Maryland title company. The alleged pattern: Genuine Title provided cash payments and free marketing materials and services — including data, letter-generation, and mailing services — to loan officers in exchange for referrals of settlement business. Actions were brought against large lenders.

And individual loan officers were named. That is the fact to carry out of this case study. Section 8 prohibits any person from giving and any person from accepting. It does not require the government to proceed against your employer first, and it does not treat "my manager set it up" as a defense. Careers ended. Licenses were affected. The arrangement had been characterized, inside the business, as ordinary marketing support.

October 2015 — Bulletin 2015-05. The Bureau issued Compliance Bulletin 2015-05, RESPA Compliance and Marketing Services Agreements, describing the risks it had observed. It was widely read as hostile to MSAs generally, and over the following two years a large part of the industry exited them entirely — not because the bulletin banned them, but because institutions concluded the litigation risk exceeded the marketing value.

2016–2018 — PHH, and a court's reading of Section 8(c). In the PHH Corporation matter, involving captive mortgage reinsurance, mortgage insurers ceded reinsurance premiums to a lender-affiliated reinsurer; the Bureau's theory was that the payments were consideration for referrals of mortgage insurance business and exceeded the value of the risk actually transferred. In 2016 a panel of the D.C. Circuit rejected the Bureau's reading of Section 8(c), holding that Section 8(c)(2) permits bona fide payments for services actually performed at reasonable market value, and further holding that applying a changed interpretation retroactively raised due process concerns. The 2018 en banc decision addressed the separate constitutional question of the Bureau's structure and left the panel's statutory RESPA holdings undisturbed. (Read the opinions. Do not rely on summaries, including this one.)

January 2017 — Prospect Mortgage. The Bureau took action against Prospect Mortgage, LLC, alleging a network of arrangements with real estate brokers and others that functioned as payments for referrals — including marketing services agreements, lead agreements, and desk license agreements. The Bureau also acted against parties on the receiving end, including real estate brokerages and a mortgage servicer, on the theory that accepting the payments violated Section 8 just as giving them did. (The parties and terms are in the public orders; consult them directly.)

Two features of that matter deserve emphasis because they generalize:

  • The desk license agreement was treated as part of the same picture, not as a separate innocent lease. Where what is purchased is position inside a referral source's office rather than square footage, the label on the agreement does not control.
  • The Bureau pursued the recipients. Both sides of a Section 8 arrangement are exposed. An originator who has been told "we're the ones paying, so you're fine" has been told something the statute does not say.

October 2020 — the reversal in guidance. The Bureau rescinded Bulletin 2015-05 and replaced it with RESPA Section 8 Frequently Asked Questions covering Sections 8(a), 8(b), and 8(c), gifts and promotional activity, and MSAs. In substance, the FAQs state that MSAs are not per se illegal and that the analysis returns to the statutory standard: services actually performed, reasonable market value, not tied to referrals.

2023 — the same problem, online. The Bureau issued an advisory opinion on digital mortgage comparison-shopping platforms and payments to their operators, addressing platforms that present themselves as neutral marketplaces while ranking or steering based on what providers pay. The structure is new. The theory is 1974's.

What it shows

First: the statute never moved. Section 8 has read the same way since 1974. What changed across this decade was the Bureau's chosen method of communicating about it — a bulletin, then no bulletin, then FAQs, then an advisory opinion. A practitioner who tracked the guidance experienced a decade of whiplash. A practitioner who tracked the statute experienced no change at all, because the question was always the same three-part question: did anything of value move, was there an agreement or understanding, and was the payment for something of genuine market-value substance.

Second: the evidence is ordinary business records. No matter in this arc turned on a hidden document. It turned on emails, invoices, valuation opinions with revealing dates, and spreadsheets that put referral counts next to marketing spend. All of it was produced in the normal course, by people who did not think they were writing evidence.

Third: rescinding a bulletin is not a safe harbor. The most dangerous misreading of the 2020 rescission is "MSAs are fine again." What the FAQs restored was the statutory standard — which is demanding, fact-specific, and exactly what the enforcement actions were about. A restatement of the test is not a relaxation of it.

Fourth: individual liability is real and it is not theoretical. This is the difference between Section 8 and most of the compliance material in this book. A TRID timing error is your employer's problem, cured with a corrected disclosure. A Section 8 arrangement you personally accepted is your problem, and it follows your NMLS record.

Outcome

The industry's practical settlement, arrived at the hard way, looks roughly like this — and it is what your compliance department is likely enforcing:

  • Many lenders exited MSAs entirely and have not returned, treating the marketing value as not worth the exposure.
  • Those that maintain them do so with independent pre-pricing valuations, narrowly specified deliverables, monthly documented proof of performance, fixed fees insulated from volume, and periodic re-valuation.
  • Desk license agreements are prohibited outright at many institutions, again as policy rather than as law.
  • Aggregate-exposure review — summing every payment to a single referral counterparty across all agreements — became standard, because stacking was how the arrangements grew without anyone noticing.
  • Individual originators are trained on Section 8 annually, and the training exists because individuals were named.

The lesson

The MSA is not a loophole and it never was. It is Section 8(c) applied to advertising, and it works exactly as well as your documentation of three things: that the service was specified, that it was performed, and that the price came from the market rather than from the referral count.

Which means the compliance test and the business test are the same test, and it fits in one sentence:

If the referrals stopped tomorrow, would you keep paying?

If yes, you bought marketing and you can prove it. If no, you bought referrals, and the title on the agreement is not going to help you.

Discussion questions

  1. Regulation X § 1024.14(e) says an agreement or understanding may be established by a practice, pattern, or course of conduct. Identify three artifacts in an ordinary loan officer's own records that could establish such a pattern without anyone ever having discussed one.

  2. The D.C. Circuit panel in PHH held that Section 8(c)(2) permits bona fide payments for services actually performed at reasonable market value. Does that holding make an MSA safer, or does it simply relocate the entire argument to the words "bona fide" and "reasonable market value"? Which party bears the practical burden of proof, and why does that matter more than the legal burden?

  3. Bulletin 2015-05 was rescinded and replaced with FAQs saying MSAs are not per se illegal. A branch manager cites the rescission as authority to enter one. Write the three questions you ask before agreeing, and explain what each question is actually testing.

  4. The Bureau pursued both givers and recipients. Construct the argument a loan officer might make that they were merely following a company-approved program — then explain why that argument fails under Section 8's text, and what the originator should have done at the moment they were handed the agreement.

  5. Prospect Mortgage's alleged arrangements included MSAs, lead agreements, and desk license agreements with the same counterparties. Design the review procedure you would want your compliance department to run monthly. What would it measure, and at what unit of aggregation?

  6. The 2023 advisory opinion applies a 1974 statute to online comparison-shopping platforms. Name a marketing channel that exists today and did not exist when the statute was written, and work the three elements against it. Where does the analysis get genuinely hard?