![Sample our Politics & International Relations journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5947/TOC-Subject-Sample-2021-Politics-International-Relations.jpg"})
Abstract
Congress passed the Real Estate Settlement Procedures Act (RESPA) in 1974 based upon documented instances of kickbacks between settlement service providers, unearned fees, and expensive and unnecessary closing costs paid by buyers and sellers of residential real estate. It opted for a disclosure strategy accompanied by few substantive prohibitions. Over the last 37 years, only a handful of studies attempted to measure the success of the mortgage loan disclosures. This article uses a uniquely rich database to examine this question. We find evidence that closing costs increased since 1972 and fee types proliferated. The early cost estimate underestimated the final closing costs and projected cash to borrowers in a majority of cases, lending credence to complaints of baiting and switching. These observations call into question the efficacy of the RESPA disclosure scheme. Further, they point to the need for detailed data collection, routine monitoring of whether RESPA is meeting its legislative intent, and rigorous debate about whether RESPA's goals can be achieved more effectively by another strategy. This article is particularly timely because Congress instructed the new Bureau of Consumer Financial Protection to combine RESPA and related Truth in Lending Act disclosures into a single, integrated form by the summer of 2012.
Acknowledgments
We wish to thank Patricia A. McCoy, the George J. & Helen M. England Professor of Law at the University of Connecticut School of Law and the Director of its Insurance Law Center, and Stephen L. Ross, Professor of Economics at the University of Connecticut, for their critical roles in the creation and design of the NMDR. The NMDR project would not have been possible without the generous support of the Ford Foundation and George McCarthy. The authors also thank Emily Douglas, Jessica Goldberg, Deanne Loonin, Patricia McCoy, Diane Thompson, presenters and commentators attending the Symposium at Valparaiso School of Law on March 26, 2010, and colleagues at Albany Law School, especially Vincent Bonventre, James Gathii, and Timothy Lytton, for providing insightful and invaluable feedback. John Forbush, Reannon Froehlich, Tamar Malloy, and Jason Stone assisted with legal research. We especially appreciate the diligence and commitment of Lisa Chin, Warren Montgomery, Shirley Poon-Cho, and Trevor Tattan who entered the data into the NMDR database.
Notes
1The history of this transformation reveals that federal deregulation and pre-emption of state law combined to produce a system of dual regulation of home mortgages that precipitated a race to the bottom in mortgage-lending standards. Prior to the legislative and regulatory responses to the recent financial crisis, this left borrowers with disclosure as their primary form of protection against abusive lending. (McCoy and Renuart 2008, 110–18).
2Congress created this Bureau and transferred jurisdiction over many of the federal consumer protections laws to it in the Dodd-Frank Wall Street Reform and Consumer Protection Act of 2010. Pub. L. No. 111–203, § 1032(f), 124 Stat. 1376 (July 21, 2010).
3US Senate Banking and Currency Committee Report No. 91–761, 1970, reprinted in 1970 USCCAN 3488, 3489–90 (noting that interest rates on Federal Housing Administration (FHA) and Veteran Administration (VA) insured loans reached 8¼ percent and carried as many as six to ten points; and noting that conventional rates rose to 9 ¼ percent).
4 Idem. at 3506.
5Emergency Home Finance Act of 1970, Pub. L. No. 91–351, § 701, 84 Stat. 450, 523, 537. The Act also encouraged and expedited the construction and financing of a substantial number of new and existing homes by expanding existing mortgage credit facilities and creating new secondary market facilities to broaden the availability of mortgage credit. US Senate Banking and Currency Committee Report No. 91–761, 1970, reprinted in 1970 USCCAN 3489–90.
6Department of Housing and Urban Development and Veterans' Administration (1972) (hereafter the “HUD/VA Report”). The insurance or guaranty encourages lenders to originate mortgage loans through these programs because these agencies cover a high percentage of losses lenders may incur in the event of default and foreclosure.
7Data gathering accompanied interviews with lenders, title personnel, and others. Local HUD and VA offices collected settlement statements from lenders, recorded the data, and submitted a 10 percent sample of the total for study.
8HUD/VA Report p. 35. The “settlement costs” used to create these percentages included real estate brokers commissions and discount fees. “Closing costs,” were a portion of “settlement costs,” as discussed later in more detail. HUD and the VA also calculated the average “closing costs” for the borrower alone to be $243, and $315 for the seller alone; whereas total settlement costs were $454 for the buyer and $1,483 for the seller. Idem.
9US Senate Banking, Housing and Urban Affairs Committee Report No. 93–866, 1974, reprinted in 1974 USCCAN 6546, 6564.
10In particular, a series of newspaper articles published in early 1972 outlined several types of abuses, including: kickbacks paid by attorneys and title companies to developers, lenders, real estate brokers, and builders; large variations in closing costs across different cities and regions; hidden fees that boost cost; mystifying charges; and, conflicts of interest. Ronald Kessler, “The Settlement Squeeze,” Washington Post January 9–12, 1972.
11US Senate Banking, Housing and Urban Affairs Committee No. 93–866, 1974, reprinted in 1974 USCCAN 6546, 6547.
12Indeed, HUD proposed regulations along these lines in 1972. HUD, 37 Fed. Reg. 13186 (July 4, 1972). However, Congressional action in 1974 trumped the finalization of these rules.
13The concepts of capping fees or utilizing disclosures were embodied in two rival bills introduced by Senators Proxmire and Brock. US Senate Banking, Housing and Urban Affairs Committee No. 93–866, 1974, reprinted in 1974 USCCAN 6546, 6547–8. “Lender pay” refers to a third approach suggested by Senator Proxmire as early as 1971. A mortgage lender would pay all of the settlement charges it required as a condition of extending credit. This formula assumed that: “(i) lenders will initially increase their interest rates to cover the costs of settlement charges; (ii) over time, the superior economic bargaining power of lenders will force a reduction in excessive or unnecessary settlement charges; and (iii) competition between lenders will result in the savings being passed on to the general public.” Idem. at 6560.
14US Senate Banking, Housing and Urban Affairs Committee No. 93–866, 1974, reprinted in 1974 USCCAN 6546, 6548. The House Committee on Banking and Currency also linked the provision of advance disclosures with increased market efficiencies in the form of reduced costs. US House of Representatives Banking and Currency Committee Report No. 93–117, 1974, 4.
15Pub. L. No. 93–533, § 3(1), 4, 88 Stat. 1724, codified at 12 USC § 2601 et seq. While the Congressional definition of a federally related mortgage loan covered the entire home loan market, HUD limited the Act's application to home purchase loans. HUD, 40 Fed. Reg. 22448, 22450 (May 22, 1975), codified at 24 CFR § 82.2(e).
16Pub. L. No. 93–533, §§ 4, 5, 6.
17Pub. L. No. 93–533, §§ 8, 9, 12.
18US House of Representative Banking, Currency and Housing Committee Report No. 94–667, 1975, reprinted in 1975 USCCAN 2448, 2449.
19 Idem. at 2449–2450. Representative Lenore Sullivan dissected the validity of this testimony in her dissent accompanying this U.S. House of Representatives Report. She concluded that: “It took five years of hard work and some bitter battles … to enact a law last year to protect home buyers against predatory abuses, unconscionable overcharges and flagrant ‘featherbedding’ practices in the transfer of residential real estate; it is now taking only five months of real estate lobbying pressure to convert that law into a hollow shell which would permit elements of the industry to resume doing many of the very things which made the original law necessary.” Idem. at 2457.
20Pub. L. No. 94–205, § 4, 89 Stat. 1157.
21Pub. L. No. 94–205, §§ 6, 19, 89 Stat. 1157. Congress has never fixed these enforcement glitches.
22HUD, 41 Fed. Reg. 1672, 1674 (Jan. 9, 1976), codified at 24 CFR § 82.8. Lenders could add charges on blank lines, in addition to filling in charges specific to the loan on lines with preprinted fee names.
23HUD, 57 Fed. Reg. 49600 (November 2, 1992).
24Page two of the form itemizes the settlement costs. Real estate broker commissions appear in the 700 series; loan origination fees appear in the 800 series; escrow and advances appear in the 900 series; reserves deposited with the lender appear in the 1000 series; title charges appear in the 1100 series; government recording and transfer charges appear in the 1200 series; and miscellaneous settlement fees appear in the 1300 series. Line 1400 contains the total of all of the settlement charges. Page one contains summaries of the buyer's and seller's transactions related to the property sale, seller contribution to the costs, the bottom line amount due from borrower, and how much cash, if any, the borrower must bring to the closing.
25There are minor but notable differences between the HUD-1's and HUD-1A's pre-printed fee types. Line 807 on the HUD-1 is for “Assumption fee,” while line 807 on the HUD-1A is for “Mortgage broker fee” (perhaps reflecting HUD's apparent determination that purchase loans will not require a broker fee and refinance loans will not require an assumption fee). In the 1300 series, the HUD-1A has two additional pre-printed lines for items not listed on the HUD-1: “Architectural/engineering services” (line 1303) and “Building permit” (line 1304).
26An amendment to RESPA in 1992 clarified that the Act covered refinances of existing mortgage loans and subordinate lien loans. Pub. L. No. 102–550 § 908, 106 Stat. 3874. HUD responded by drafting the HUD-1A.
27HUD, 24 CFR § 3500.8. A sample HUD-1A appeared at the end of the regulations accompanied by instructions.
28HUD, 73 Fed. Reg. 14,030 (March 14, 2008). As a part of this process, HUD hired a communications firm to vet several versions of GFEs and HUD-1s with focus groups formed in 15 cities. Six rounds of testing occurred over a five-year period (US Department of Housing and Urban Development 2008b). The agency also evaluated the potential costs and savings of the amended templates (US Department of Housing and Urban Development, 2008a). This 2008 effort followed an aborted 2002 rulemaking process. HUD, 67 Fed. Reg. 49134 (July 29, 2002).
29HUD, 73 Fed. Reg. 68204 (November 17, 2008).
30The agency inserted a box to list credits and debits against the “origination charge” due to the payment of yield spread premiums and discount points. In an attempt to make the tradeoffs in lender-paid broker compensation more apparent to borrowers, HUD added a table on the final page of the GFE that lists four pricing metrics – the interest rate, initial loan amount, initial monthly payment, and total settlement costs – for the loan applied for and two other loans, with lower and higher lender-paid broker compensation. However, lenders may opt to leave the comparison information blank. HUD, 73 Fed. Reg. 68204, 68258 (November 17, 2008).
31 Idem. at 68239–68241, codified at 24 CFR §§ 3500.2(b), 3500.7(a)(1), 3500.7(a)(4), 3500.7(c), 3500.7(f). The GFE remains binding if the borrower moves forward with the application and decides to accept the loan. Lenders can issue a revised GFE whenever “changed circumstances” intervene. “Changed circumstances” consist of a revision to the loan amount, new information affecting the borrower's credit quality, or the borrower choosing a new product but cannot include increases in the price of third-party settlement service providers. If the lender discovers an “inadvertent” misstatement, the lender can “cure” the faulty disclosure up to 30 days after closing by issuing a new disclosure and offering the borrower the amount by which the costs were under-disclosed. Over-disclosures need not be cured. The loan terms are not locked unless the lender explicitly agrees. 25 CFR § 3500.7(g).
32Pub. L. No. 93–533, § 14, 88 Stat. 1724.
33The study's definition of closing costs appears to coincide with that used in the HUD/VA Report. HUD (1981, App.A, II-33).
34In 2010, HUD announced a plan to revisit concerns related to controlled business relationships. HUD, 75 Fed. Reg. 31334 (June 3, 2010). However, the agency has not publicly followed up on this project as of April 2011.
35Congress did not act upon HUD's suggestions. Much later, HUD and the Board of Governors of the Federal Reserve System proposed to Congress that lenders give borrowers more reliable closing cost information earlier in the process and choose between guaranteeing settlement costs and providing a GFE that would be accurate within a specific tolerance. HUD also requested Congress to grant it and relevant state agencies the authority to enforce any violation of RESPA and to create a private right of action for borrowers, as in the original Act – Board of Governors of the Federal Reserve System and the U.S. Department of Housing and Urban Development, 1998, 84 (hereafter “FRB/HUD Report”). Again, Congress did not act.
36There also were no large-scale independent studies of RESPA's effectiveness, likely due to the absence of publicly-available data.
37“Loan-related service charge” equaled the sum of origination-related fees (excluding yield spread premiums), third party charges, and title fees (Woodward 2008, 40–41, III-3–5). It is unclear whether she included discount points in this total. Unfortunately, this definition appears to overlap with but not match the measure of “closing costs” used in the HUD/VA Report. One major reason was that the VA/FHA Settlement Costs Code Sheet did not list yield spread premiums or mortgage broker fees, presumably because these were not common fees charged in the market at the time (Engel and McCoy 2011, 15) (describing the home mortgage market in the 1970s in which consumers went directly to banks and savings and loan associations to obtain mortgage credit).
38Their “closing costs” included yield spread premiums.
39Shroder defined closing costs to be those fees that belong in the 800, 1100, and 1300 series, excluding discount points.
40These fees represented charges for: loan origination, appraisal, credit report, mortgage broker, tax related service, processing, underwriting, wire transfer, title search, document preparation, attorney, title insurance, recording, survey, and flood survey. He excluded loan discount points, closing or escrow fee, city/county tax/stamps, notary, and other charges (Collins 2010, 23).
4112 USC §§ 2807–2810 and FRB Regulation C, 12 CFR. §§ 203.1–203.6.
42The NMDR collected the following document types: loan application, GFE, Truth in Lending Act disclosure, HUD-1 or HUD-1A, loan note and riders, mortgage riders that pertained to loan price, arbitration agreement, and the notice of right to cancel.
43The NMDR collected information about borrower costs only (although seller closing costs are part of the HUD-1/1A document, they were not part of that research project).
44Although many documents in the NMDR were acquired from attorneys, not every loan had been the subject of litigation or even negotiation. In many cases, the NMDR received mortgage loan paperwork that clients had provided to legal services attorneys as part of an agency's intake process. These offices accepted some of these clients for representation and rejected others. Materials acquired from state agencies – including offices of attorneys general and banking regulators – were obtained through Freedom of Information Act requests. These agencies obtained these documents through investigations of one or more of the participants in the lending process. However, not all of the investigations were concerned with the data that were the subject of the NMDR. For example, in one case, the NCLC received a large number of loans from an agency that had investigated a broker for licensure and other violations unrelated to loan pricing and other terms.
45The number of documents is greater than the number of loans because some loan sets contained multiple iterations of a document type. We sometimes received partial documents (i.e., a page was missing) or copies of documents on which a portion was illegible. In total, 1,373 (74 percent) of our HUD-1/1As and 827 (74 percent) of our GFEs represent complete documents with no illegible portions.
46A list of the lenders that made five or more loans in our dataset appears in Appendix B, Table B1. Table B2 states the geographic distribution.
47To arrive at these numbers, we matched the list of lenders whose HUD-1/1As appeared in this dataset to HUD's Subprime Lender List, available at http://www.huduser.org/portal/datasets/manu.html. We then counted the number of loans represented by the subprime lenders in our database whose HUD-1/1As and GFE's we used in our analyses that follow.
48HUD and the VA classified “settlement costs” as “closing costs” plus loan discount fees, prepaid items, and real estate commissions. This accounts for the 10 percent ratio between 1972 average borrower and seller settlement costs and average sale price. Members of Congress focused on this number when debating the need for RESPA. This study focused only upon the borrower costs.
49Three states represent a sizeable proportion of the whole: Massachusetts (20.3 percent); North Carolina (23.5 percent); and New York (20.9 percent).
50For example, from 1971 through 1979, HUD regulations capped interest rates but permitted an origination fee of up to 2.5 percent of the principal on construction loans or 1 percent on all other loans plus reasonable and customary closing related fees and a discount charge in the case of loan constructions and refinances – 24 CFR §§ 203.20, 203.27 (1971) and (1979). We reviewed the regulations for those years because HUD and the VA collected its data in 1971 and HUD's research firm collected its data in 1979 for those two studies against which we compare our results. By 1994, the first year in which loans contained in the NMDR were made, HUD had eliminated the interest rate cap but the rules related to charging other fees remained essentially the same except that the agency allowed a discount fee in any type of loan – 24 CFR §§ 203.20, 203.27 (1994). These regulations did not change in any material way through 2007 – 24 CFR §§ 203.20, 203.27 (2007). For the same periods, VA regulations capped the allowable interest rate and prohibited brokerage fees but permitted reasonable and customary charges for a list of specified services, 1 percent of the loan amount for all other charges, and an additional 2 percent for construction, alteration, improvement or repair loans – 38 CFR §§ 36.4311, 36.4312 (1971) and (1979). Other than the elimination of the usury cap in 1996, the VA regulations remained essentially unchanged though 2007 – 38 CFR §§ 36.4311, 36.4312 (1996) and (2007).
51Using the same methodology described in Note 48 above, we counted the number of loans represented by the subprime lenders in our database whose HUD-1/1As we used in this closing cost analysis.
52In total, the NMDR used 116 codes to capture fees disclosed on the HUD-1/1A and GFE, but for this analysis, we grouped similar codes together according to the HUD/VA's coding instructions to produce an apples-to-apples comparison.
53For this analysis, we used HUD's definitions of “origination service” and “title-related service.” HUD, 73 Fed. Reg. 68204, 68 239 (November 17, 2008), codified at 24 C.F.R. § 3500.2.
54“Final charges at settlement often include ‘junk fees,’ which increase the original estimates … ‘Junk fee’ is a term used throughout this document to mean any fee charged for a service to a borrower that has little or no value in relation to the charge, and/or may be duplicative, to increase a loan originator's profit … The current requirements allow … a loan originator to charge several fees for origination, document preparation, and document review … Excessive itemization thus enables originators to charge more than if borrowers could review and shop the total origination charges … The types of fees charged by loan originators, title agents and other service providers have multiplied in recent years making it steadily more difficult for borrowers to compare settlement costs – HUD, 67 Fed. Reg. 49134–49135 and n. 6 (July 29, 2002).
5524 C.F.R. § 3500.7(c)(2), appearing in HUD, 61 Fed. Reg. 13232, 13236 (March 26, 1996).
56He compared estimated and final costs using fees that should be placed in the 800, 1100, and 1300 series, regardless of their actual location on the real documents. This definition of “closing costs” differs from that used by the HUD/VA Report in one way: the HUD/VA included recording and transfer fees, whereas Shroder excluded them. To make our results comparable to his, we replicated his definition.
57We did not examine whether particular types of charges were predominantly responsible for discrepancies, largely because of the dilemmas imposed by fee unbundling. For example, some lenders put all origination charges on line 801, for “Origination fee,” while others impose a series of origination-related fees, each separately itemized.
58We selected $100 because there is regulatory support for this number. The Truth In Lending Act (TILA) allows lenders to err up to $100 when they disclose the finance charge to consumers – 15 USC § 1605(f)(1)(A).
59For the purposes of this analysis, we excluded as outliers those observations that had difference values (i.e., final closing costs minus estimated closing costs) greater than three standard deviations from the mean.
60As outlined above, lenders complained of increased effort and costs after RESPA's birth in 1974 to such a degree that Congress amended it significantly less than a year later. More recently, in its 2008 Report regarding the regulatory impact of the proposed disclosure changes, HUD summarized literature describing “perverse” competition and excessive fees (HUD 2008a, 2–60–2–93).
61Many of the additional fee types we created based upon our initial review of sample documents had been used fewer than 20 times, showing that they represented uncommon charges.
62We excluded fee types intended to capture items that one might expect to be described with varied language and also combined similar fee types. For example, we excluded: items coded “escrowed funds,” which were listed in many different ways because funds were placed in escrow for many genuinely different purposes; and items coded “insurance-other,” which captured numerous kinds of insurances that borrowers purchased. Examples of fee types we combined include: “document preparation” and “document fee-other” (counted as one fee type); and “settlement or closing cost” and “settlement or closing cost-other” (also counted as one fee type).
63Unfortunately, whether the origination charge refers to just the fee listed on line 801 or all origination-related fees that belong in the 800 series is unclear.
64HUD, 73 Fed. Reg. 68240–68241, codified at 24 CFR §§ 3500.7(e)(1), 3500.7(f)(2), § 3500.7(f)(3)(2010).
65Whether or not it is transparent to borrowers that these eight fee types are origination-related is an open question.
66HUD, 73 Fed. Reg. at 68241, 68244–68245, codified at 24 CFR § 3500.2(b)(2010) and Appendix A to Part 3500 – Instructions for Completing HUD-1 and HUD-1a Settlement Statements. “Origination service means any service involved in the creation of a mortgage loan, including but not limited to the taking of the loan application, loan processing, and the underwriting and funding of the loan, and the processing and administrative services required to perform these services.”
67Bundling should also reduce differing interpretations of what charges go into a specific line and lender confusion about where a specific charge should be placed. Lenders and settlement service providers raised these complaints when surveyed in 1979–1980. HUD 1981, App. A IV-13–14.
68HUD, 73 Fed. Reg. at 68240, codified at 24 CFR § 3500.2(b); Appx. A to Part 3500 “Instructions for Completing the HUD-1 and HUD-1A Settlement Statements.” “Title services” refers to any service involved in the provision of title insurance and for conducting the settlement. The regulation provides a non-exhaustive list of charges that represent title services: title examination and evaluation; preparation and issuance of title commitment; clearance of underwriting objections; preparation and issuance of a title insurance policy or policies; and the processing and administrative services required to perform these function.
69There were an additional 258 HUD-1/1As that disclosed positive cash-to-borrower values, but they were accompanied by GFEs that failed to disclose any information about estimated cash-to-borrower values.
70See our explanation for selecting $100 in Note 59 above.
7124 CFR § 3500.8(a), appearing in HUD, 61 Fed. Reg. 13232, 13237 (March 26, 1996).
7224 C.F.R. Appx. A to Part 3500 “Line Item Instructions for Completing the HUD-1 and HUD-1A Settlement Statements,” appearing in HUD, 59 Fed. Reg. 6506, 6515–16 (Feb. 10, 1994). The Instructions to settlement agents contain no guidance regarding where to list disbursements when the HUD-1 is used for a refinance loan, nor whether the reimbursements should be summed with new charges into the total settlement charges line at the bottom of the page. Compare 24 CFR Appx. A to Part 3500 “Instructions for Completing the HUD-1 and HUD-1A Settlement Statements,” appearing in HUD, 73 Fed. Reg. 68204, 68243–68247 (Nov. 17, 2008).
73Even in the case of complete documents – those we received with no pages missing and no illegible portions – there were instances in which the total settlement charges line was simply left blank.
74It was not possible to make a determination for the remaining 20.8 percent of cases.
75Disclosures do not operate alone in the RESPA scheme. The Act's prohibitions against kickbacks and steering between service providers may have failed as well. Unfortunately, the NMDR does not permit analysis of the influence that substantive regulations play either separately or in combination with disclosures.