- CBA on
- CBA Media
Joint HMDA Comment Letter to CFPB
Re: Request for Comment on Proposed Amendments to Regulation C to Implement Amendments to Home Mortgage Disclosure Act, Docket No. CFPB-2014-0019
Dear Ms. Jackson:
The Consumer Bankers Association (“CBA”), American Bankers Association (“ABA”), Financial Services Roundtable (“FSR”), Housing Policy Council (“HPC”), and Mortgage Bankers Association (“MBA”), Consumer Mortgage Coalition (together “the Associations”) appreciate the opportunity to comment on the subject proposed rule (“the Proposal”) issued by the Consumer Financial Protection Bureau (“CFPB” or “Bureau”) amending Regulation C, which implements the Home Mortgage Disclosure Act (“HMDA”) amended by the Dodd-Frank Wall Street Reform and Consumer Protection Act (“the DoddFrank Act”).
Our members are committed to ensuring that all customers receive fair treatment and we oppose any form of illegal credit discrimination. While we support the purpose of HMDA--to provide information on the availability of credit in the home mortgage market--we are concerned that the Proposal to markedly increase HMDA data reporting and coverage goes beyond the law’s purposes in some areas and will unduly harm competition and increase costs in others. At the same time, we do not believe that the Proposal adequately ensures consumers’ privacy and data protection in an age where unwelcome breaches must be anticipated. Consequently, if the Proposal is finalized as proposed, we are concerned that it will result in significant adverse and unintended consequences that harm consumers and unduly increase costs. For the reasons detailed in this letter, we urge the CFPB to:
• Weigh the consequences and value before adding fields;
• Limit Regulation C’s coverage to only home mortgage loans;
• Protect consumers from invasions of privacy and address data security concerns through rulemaking before the contents of the HMDA fields added by the Proposal are released to the public;
• Establish a reasonable implementation schedule for these changes given that technology resources will be consumed by other demands, especially the TILA RESPA Integration through August 1, 2015 and beyond;
• Establish workable data integrity standards for HMDA reporting considering the many new data points proposed to be collected and reported;
• Minimize unnecessary reporting requirements to preserve credit options and avoid unnecessary costs;
• Conform Regulation C to other rules and to MISMO as proposed; and
• Coordinate with other regulators on the Community Reinvestment Act (“CRA”) implications of these changes.
Our nation’s financial system remains unique due to the diverse array of mortgage lenders and products to serve consumers’ home finance needs. These lenders include a wide range of institutions including the very largest banks to the smallest community banks, credit unions, and independent mortgage bankers of all sizes. While our members remain supportive of HMDA’s purpose, they are also concerned that the new rules and their costs are making loans less affordable and less available to the very consumers the market should serve.
Further, institutions of all sizes have had to pay much greater costs to comply with the myriad of new regulations, and to manage risks associated with regulatory uncertainty, have made it difficult to innovate and provide consumers with products and services that meet their needs. The increased cost of mortgage lending is particularly stark. MBA data shows the cost of originating a mortgage loan has increased from approximately $5,000 to as much as $8,000 in a few short years, with a considerable amount of these costs attributable to regulatory compliance. 7 Moreover, the HMDA data indicates that regulatory costs are taking a toll on smaller lenders. This year’s HMDA data shows a continuing reduction in the number of reporting lenders from 8,900 HMDA filers in 2006 to 7,190 in 2012, a downward trend that shows no sign of abating. These factors combine to constrain the availability of credit to consumers. Through this lens, we consider and comment on this Proposal that would greatly increase Regulation C’s data reporting requirements and the number of institutions required to report.
II. WE RESPECTFULLY ASK THE BUREAU TO WEIGH THE CONSEQUENCES AND VALUE BEFORE ADDING INDIVIDUAL FIELDS
The Associations have heard criticisms from many members that echo points raised by small entity representatives (“SERs”) during the Small Business Regulatory Enforcement Act (“SBREFA”) process as well as others that have identified issues concerning implementation of the Proposal. These criticisms range from disputing the analytical value of certain proposed fields, pointing out problems in reporting values of certain fields consistently, and raising compliance questions about how to report fields under varied loan circumstances. The more our members imagine living under the complex reporting regime to be imposed on them under the Bureau's Proposal, the more concerns are being articulated. This reaction will not end with the conclusion of the comment deadline-any more than they have with respect to the remittance rule or the qualified mortgage (“QM”) / ability to repay (“ATR”) rules.
No data set can fully explain all underwriting or pricing decisions; there are often legitimate, necessary, and non-discriminatory factors influencing underwriting or pricing decisions, and such information, when reduced to individual data points, will be of limited use in understanding credit decisions. Consequently, we question the proposed inclusion of many data points, including automated underwriting recommendations; borrower paid origination charges; total points and fees; total discount points; risk-adjusted pre-discounted interest rate; interest rate; prepayment penalty; QM status; and HELOC first draw amount. Each of these data points is either not used in underwriting decisions, subject to variation between lenders that is not relevant to HMDA, affected by borrower choice, or otherwise not an appropriate metric by which a lender's origination activity should be evaluated.
The Associations will continue to assemble these points to share with the Bureau and with the prudential regulators that have supervisory and enforcement authority over HMDA reporting. We believe that continuing to engage on these issues can improve HMDA rulemaking and HMDA compliance performance. As the Bureau has experienced, continuing dialogue produces better regulatory results. In addition, as the Bureau's creation of a permanent office of rule implementation evidences, final rules never resolve all the issues inherent in their implementation.
Accordingly, the Associations seek offer to work with the Bureau and the interested prudential regulators to make a better regulation with appropriate scope, to develop an updated and reliable HMDA: Getting it Right guide to create workable interagency examination procedures that support material reporting compliance.
III. LIMIT REGULATION C’S COVERAGE TO HOME MORTGAGES
Congress enacted HMDA in 1975 after concluding “some depository institutions have sometimes contributed to the decline of certain geographic areas by their failure pursuant to their chartering responsibilities to provide adequate home financing to qualified applicants on reasonable terms and conditions.” The statute declares that its purpose “is to provide the citizens and public officials of the United States with sufficient information to enable them to determine whether depository institutions are filling their obligations to serve the housing needs of the communities and neighborhoods in which they are located and to assist public officials in their determination of the distribution of public sector investments in a manner designed to improve the private investment environment.”
HMDA was intended to provide insight into the home mortgage market and gauge the availability of credit to consumers. The legislative history does not indicate that loans for other than home financing were to be addressed and the Dodd-Frank Act provisions amending HMDA make no mention of such loans.
Based on the clear reading of the statute as amended and its legislative history, the Associations urge the CFPB not to extend the scope of Regulation C to require reporting of data on loans for purposes other than home mortgage financing, including commercial loans, repurchased loans, home improvement loans secured by a dwelling, or reverse mortgages. Similarly, we urge the Bureau not to require the reporting of data relating to home equity loans or home equity lines of credit (“HELOCs”). As explained below, the expansion of HMDA’s coverage to loans for purposes other than housing finance would distort the data and undermine HMDA’s important purpose while unduly increasing costs to consumers.
A. Regulation C Should Not Apply to Commercial Purpose Loans
Put simply, commercial purpose loans work differently from mortgage loans. In the commercial context, lenders frequently add different kinds of collateral at various points in the loan process. Occasionally, a commercial lender may require a residential property as additional collateral out of an “abundance of caution” notwithstanding that the loan has a business purpose unrelated to the property. In home mortgage lending, the home itself always secures and is the objective of the loan.
Notwithstanding the clear differences between commercial and residential lending, the Proposal would apply Regulation C whenever a residence is taken as collateral, even in the case of a commercial loan where the underlying transaction may have nothing to do with housing. Reporting non-housing loan data will distort the usefulness of HMDA data for analyzing home financing activity.
In addition, in commercial real estate mortgage lending, including loans secured by multifamily property, the purpose of the loan is the purchase or refinancing of an incomeproducing business — a multifamily property or other commercial real estate. Such commercial lending programs are proprietary, terms are negotiated and the loan is structured to mitigate risks associated with the property, borrower, market and loan terms, as well as the needs of the investor. MBA is also submitting a separate letter explaining its position that commercial real estate loans secured by multifamily properties should not be covered by HMDA.
Another challenge associated with reporting data about commercial loans is that the commercial loan process is very different from the residential mortgage loan process. Unlike residential mortgage lending, commercial loans frequently do not involve a formal application process. Moreover, the software systems handling these loans do not use data points typical to HMDA data collection.
Many of the data points that are proposed to be collected such as QM status, credit score and debt to income (“DTI”) ratio are irrelevant to commercial purpose loans. As a result, many data fields would have “N/A” entries. HMDA was not designed, nor will it work for commercial purpose loans. If Regulation C is extended to commercial lenders, these lenders – including many smaller lenders – will face significant costs and compliance challenges.
Congress is clearly concerned about unnecessarily costly regulation and its impact on both consumers and the industry. Hence, the Dodd-Frank Act provides a mandate that “outdated, unnecessary, or unduly burdensome regulations are regularly identified and addressed [by the CFPB] in order to reduce unwarranted regulatory burdens.” 14 If the CFPB adopts a regulation requiring lenders to implement costly systems and compliance measures to collect data on loans that are not related to housing needs, it would go against Congressional concerns.
B. Reporting of HELOCs Should Remain Optional
For similar reasons, we urge the CFPB to maintain the existing provision in Regulation C making reporting of HELOCs optional. While some percentage of HELOCs are used for home improvements, many of these loans are used to finance educational needs, to purchase vehicles, in case of emergency, to consolidate outstanding debt, and for other purposes unrelated to housing.
Requiring lenders to report data about HELOCs will not enhance the data available on home financing or for fair lending analysis, but will conflate HMDA numbers and impose new and costly compliance requirements. When the Federal Reserve last amended Regulation C, it specifically found that data about HELOCs would not be useful in promoting the purposes of HMDA and for that reason, decided to make HELOC reporting optional.
HELOC data will provide little useful information for housing purposes. Since homeowners can use the equity in their homes for many reasons that have nothing to do with housing, adding the data into the mix will produce misleading and unhelpful information for housing purposes. Moreover, the inclusion of many borrowers who secure a HELOC solely for emergency purposes and never draw down the line will further distort the data and its utility.
Also, the process of making small dollar equity loans does not require the same level of data from consumers as traditional mortgages and lenders consequently do not collect the same level of data. Thus, the added compliance from the proposal will be costly because systems and procedures to collect that data will need to be created. Considering the compliance costs, members have indicated they may no longer offer smaller HELOCs because these transactions may no longer be economically viable.
It also is important to note that HELOCs are frequently originated and managed through different systems and in different parts of companies than those that originate mortgage loans. Mortgage lending divisions originate mortgages and consumer lending divisions frequently originate HELOCs. These systems and other differences likely explain why few lenders report HELOC HMDA data on a voluntary basis.
C. Regulation C Should Not Apply to Repurchased Loans
We also urge the Bureau not to require lenders to report data relating to repurchased loans because these transactions are not “consumer financial products” and because the loans already have been reported. The decision to repurchase loans is based on operational considerations and is not reflective of home mortgage lending decision making. Like other loans discussed above, the inclusion of these loans in the data will not shed light on the housing market and reporting them would be contrary to HMDA’s purpose.
Repurchased loans which were originated prior to the effective date of the revised TILA RESPA regulation will lack information that will be difficult – if not impossible – to collect. More important, though, is that loans may be repurchased for any number of reasons which have nothing to do with housing. Since the information on the loan already will have been reported at the time of origination and the data needed for housing identified at that time, the information on repurchased loans would be redundant. Although the data may be interesting and significant for other reasons, HMDA is not the vehicle to collect and report data on repurchases.
D. Regulation C Should Not Apply to Reverse Mortgages
We also urge the Bureau not to require lenders to report reverse mortgages pursuant to Regulation C. Reverse mortgage loans are generally used for purposes that are unrelated to housing finance. Moreover, the data collected on these loans will not provide insight into the objectives that HMDA serves. Finally, lenders are already exiting the reverse mortgage market due to regulatory demands and uncertainties with these products.
While we understand that the Bureau has been reviewing the reverse mortgage market, we do not believe that requiring reporting for these loans is consistent with HMDA’s purpose. We also are particularly concerned about the unintended consequences of coverage for this market and the consumers who may be served by these products. Good lenders report that the constraints and regulatory burdens have driven them from or caused them to forgo this market, and added data collection burdens and complex reporting regimes will do nothing to reverse the trend; if anything, it will help continue it.
E. Regulation C Should Not Apply to Home Improvement Loans
Finally, we support the Proposal’s exclusion of home improvement loans from reporting. Reporting these loans would distort the HMDA data because Regulation C requires loans to be categorized as home improvement even if only a small portion of the loan is used for that purpose. For example, on a $500,000 refinancing, if as little as $2,000 is used toward home improvement costs, that small amount is enough to cause the entire loan to be classified as a home improvement loan. It also may not be apparent on the application that a loan is intended for home improvement, which will be confusing and burdensome for consumers and lenders.
IV. HMDA DATA SHOULD NOT BE MADE PUBLICLY AVAILABLE UNLESS CONSUMER PRIVACY IS PROTECTED AND DATA SECURITY CONCERNS ARE ADDRESSED
A. HMDA Data Presents Particular Privacy and Data Security Concerns
Massive breaches of consumers’ private information collected and maintained by companies and government--affecting millions or even tens of millions of consumers--have become commonplace, making information protection and data security a matter of highest priority to our members.
Although HMDA data does not presently include personal identifiers, loan-level HMDA data collected and made available to the public in combination with other publicly available data sources, if provided for all data fields, could easily enable data prospectors, bad actors and others to “reidentify” individual borrowers’ exceedingly confidential data and exploit it for their own purposes.
A recent White House report analyzing Big Data’s benefits and challenges (“Big Data Report”) defined re-identification as the process where previously de-identified information is re-connected to reveal the identity of the person.15 It noted a “mosaic effect” is used to infer a person’s identity from datasets that do not include personal identifiers.
The Big Data Report cautions that, even if information does not include personal identifiers, “it is difficult to predict how technologies to re-identify seemingly anonymized data may evolve. This creates substantial uncertainty about how an individual controls his or her own information and identity, and how he or she disputes decision-making based on data derived from multiple datasets.”
If the CFPB adopts the Proposal in its present form, the number of required fields for each record will increase from 36 to 72. These new fields include confidential financial data such as credit score, DTI, and combined loan to value (“CLTV”) ratio. Consequently, if this data are inadvertently or knowingly released to the public, the harm associated with re-identification would be even greater.
As discussed below, the Associations strongly support the CFPB’s proposal to redact the new data including confidential financial data such as credit score, DTI, and CLTV as well as other confidential data including age and personal identifiers.
Nevertheless, while the Proposal indicates that the CFPB plans at a later date to seek public input on the application of a balancing test to release data that lenders currently make available to FFIEC, we urge that additional data should not be made publicly available until the CFPB has adopted detailed rules for collecting and releasing data and only after public notice and comment. These rules should specifically address the treatment of each HMDA data field subject to release as well as conditions on release outside government for research or other purposes.
Similarly, to protect against data breach, we urge the CFPB to detail the types of data security safeguards it will undertake and publish them for public comment. The possibility of a breach of confidential financial data is even more troubling when consumers cannot control distribution of data concerning them, as is the case with HMDA. If a consumer wants to buy a home, he or she has no choice but to provide confidential financial data that in turn must be reported for HMDA purposes and potentially could be released.
B. To Protect Consumers New Data Fields Must be Kept Confidential and Not Released
As previously indicated, the Associations strongly support the CFPB’s proposed decision to redact confidential HMDA data points before publication. By doing so, we believe the CFPB is taking an essential step to protect confidential privacy data on consumers that is so sensitive that it should not be publicly released.
Due to the sensitivity of the data, we believe new data collected under the rules must only be available to government authorities and not made publicly available for the near term. Before data is provided to other parties, including researchers, standards should be established in the form of rules.
According to former Federal Reserve Board Senior Advisor Glenn Canner, approximately 95 percent of loan records are “unique,” meaning loan amount and census tract can be attributed to a single person. With a cross match to private lien transfer records, one can identify these individuals in 95 percent of the cases.
Accordingly, we strongly urge the Bureau to keep all data collected under the Proposal private pending further rulemaking as discussed in the following sections.
C. The Bureau Should Promulgate Privacy and Data Security Rules and Protect the Confidentiality of HMDA Data
The Associations believe that to fully comply with the Dodd-Frank Act and other governing privacy laws,21 the Bureau should issue privacy and data security rules for public comment.
1. Privacy Rules
The Proposal indicates that the CFPB plans at a later date to seek public input on the application of the balancing test to the public release of the data that lenders currently make available to FFIEC. The Associations strongly urge the CFPB to go through a rulemaking process or other procedure that provides an opportunity for comment.
Under the Dodd-Frank Act, Congress tasked the Bureau with protecting privacy interests including requiring lenders to delete sensitive information before releasing the loan application register (“LAR”) to the public. Since certain information is so sensitive that a lender may face litigation if it releases the information to the public, the Bureau may require the redaction of certain information to protect lenders from violating federal or state privacy laws.
Additional safeguards should address this issue as well as the release of data to third parties for various purposes including research and in response to Freedom of Information Act (“FOIA”) requests.
If the Bureau provides HMDA data to researchers, the Associations believe the Bureau should establish fair criteria for selection of researchers that ensure research from all perspectives, detailed procedures including written agreements that require conformity with the research plans and make clear the liability for violating confidentiality.
Notably, under the Dodd-Frank Act and the Privacy Act of 197424 any proposed disclosure of personal information is required to be addressed through “rules” and “notice” to the public, subject to all of the rulemaking requirements in the Administrative Procedures Act25 (“APA”) and other statutes, such as the Regulatory Flexibility Act26 and the Paperwork Reduction Act.27 The Dodd-Frank Act requires the CFPB to “promulgate regulations providing for the confidentiality of certain types of information and protecting such information from public disclosure.”
The final rule the CFPB previously issued “pertain[ing] to the protection and disclosure of confidential information”28 would not authorize the release of confidential information in the HMDA data through any process short of notice and comment rulemaking. We also do not believe mention of the balancing test to be applied later is sufficient for this purpose.
The APA only exempts “interpretative rules, general statements of policy, or rules of agency organization, procedure, or practice” from formal notice and comment rulemaking requirements where “notice or hearing” is not required by statute. This would not apply to the future release of confidential HMDA data because the Dodd-Frank Act specifically requires “rules” regarding confidential treatment of consumer information29 and the Privacy Act contemplates providing “notice” to the public of rules regarding the disclosure of personal information.30 Accordingly, we ask the Bureau to provide for a process where stakeholders have the opportunity to provide comment on the proposed information disclosure whether through formal rulemaking or other procedure.
2. Data Security Rules
The collection of additional data including credit scores, DTI, CLTV, age and postal address makes HMDA data an attractive target for hackers whether it is held by companies or by the Government. In addition, the annual bulk submission of the data makes it vulnerable to attack.
While we appreciate that the CFPB has indicated that it will take steps necessary to ensure data security, we urge the CFPB to detail these steps to the extent feasible considering the danger of hacking, and make the rules available for public comment before data is collected. As an alternative, we would suggest that the Bureau consider collecting some of this data in aggregate form while it decides upon data security procedures. As the Bureau is aware, a recent Government Accountability Office (“GAO”) Report found the CFPB (1) lacks written privacy procedures; and (2) has not created a comprehensive privacy plan that merges policies and guidance.
Considering the confidential nature of the data to be collected pursuant to the Proposal, the Associations believe it is crucial for the Bureau to remedy the gaps identified by the GAO before gathering data. The Associations share GAO’s view that the CFPB’s implementation of proper written procedures would help assure the public that the Bureau is meeting statutory privacy requirements and not unduly placing consumers’ privacy – and their financial well-being – at risk.
In its report, GAO noted the CFPB lacks several key elements necessary for implementing National Institute in Standards and Technology (“NIST”) guidance on risk assessments and remedial action plans. To help protect consumer data, GAO recommends CFPB establish written procedures for: the data intake process; anonymizing data; assessing and managing privacy risks; monitoring and auditing privacy controls; and documenting NIST implementation.
The Federal Reserve Board’s Office of Inspector General (“OIG”) has issued a report raising similar concerns about CFPB data collection and security mechanisms. The OIG report, Security Control Review of the CFPB’s Cloud Computing-Based General Support System, made recommendations for the CFPB to enhance security with regards to system and information integrity, configuration management, contingency planning, and incident response.
Considering the highly confidential nature of the new data fields and the risks to consumers of data breach, we believe it is imperative that the CFPB address these concerns, as well, before collecting additional loan level HMDA data.
3. Losses from Data Breaches will be Borne by Banks and Ultimately Consumers
Our members report that if data are required to be collected on the scale proposed – including both data points that could lead to identification of an individual borrower (personally identifiable information or “PII”) and financially sensitive information – the costs of data security will rise significantly. Inevitably, costs will escalate in two areas: (1) implementing and maintaining more robust data security systems, and (2) compensating customers for losses in the event of a breach.
While our members go to great lengths to ensure data is protected, the collection and transfer of greatly increased and more sensitive data will necessitate even more robust and costlier systems. Further, it will heighten the potential for fraud and identity theft. While our members help customers victimized by security breaches, invariably these expenses must be addressed by increasing charges or eliminating products. Data breaches harm consumers as well as lenders.
V. IMPLEMENTATION SCHEDULE SHOULD BE REASONABLE
Over the last year, the mortgage industry has implemented an unprecedented number of rules that pervade the mortgage origination, underwriting, and servicing process. Because of the magnitude of that task, implementation continued well beyond the effective date for these rules.
Over the next eight months, the industry must retool its systems and revise its business processes to implement the TILA RESPA integration involving a complete overhaul of disclosures to consumers at the time of application and closing. Virtually all technological resources are engaged in this process and, based on experience from implementing mortgage rules earlier this year, implementation efforts can be expected to continue throughout 2014 and throughout 2015.
The complexity and breadth of this overhaul to the mortgage process has also affected the Bureau’s resources. Questions that arise must be addressed and consistent guidance presented to avoid confusion, errors, and undue burden. The Bureau must ensure that it has sufficient capability – particularly the needed expertise – to respond to the inevitable questions that will arise. This Proposal should be factored into that process.
Accordingly, we urge the Bureau to require reporting of HMDA data under this rule no earlier than 24 months after the January 1st following issuance of a final rule. Under HMDA as amended by the Dodd Frank Act, institutions are not required to report new data before the first January 1st that occurs nine months after the Bureau issues the final rule. Thus, we are only asking the Bureau to delay implementation for a year after the statutorily mandated timeframe. Assuming the final HMDA rule is issued in the spring of 2015, and TILA RESPA integration will continue throughout 2015, developing and implementing system changes for new HMDA requirements cannot begin until the beginning of 2016.
Changes to reporting requirements – which will remain unknown until this Proposal is finalized – can be expected to require significant system and process changes for lenders. Moreover, as discussed, while we support the adoption of MISMO standards for HMDA reporting, necessary training and system changes from MISMO should be factored into any final time frame. Considering all of these factors, we do not believe that lenders will be ready for data collection until January 2017. Accordingly, we urge the Bureau not to require reporting of such data until 2018 at the earliest. Moreover, if quarterly reporting is required of some lenders, additional implementation time should be provided for those changes.
We also urge the Bureau to provide transition rules for applications that are taken before the effective date, but reach final disposition after the effective date.
VI. ESTABLISH WORKABLE DATA INTEGRITY STANDARDS
Our members are committed to reporting accurate data and strive to do so but the current supervisory expectation of a near-zero error rate is virtually impossible to achieve. As community banks and other small lenders pointed out to the Bureau during the SBREFA panel, the doubling of the number of reported fields can be expected to cause the error rate to increase exponentially.
Several small business participants expressed concern about the lack of tolerances for errors during supervisory examinations and emphasized that the new data points will greatly increase the potential for examiner criticism. In turn, this leads to higher costs for reviewing data prior to reporting to go with the significant costs of training, systems and business process changes.
Compounding the problem, no data integrity standards currently are codified in Regulation C or the CFPB’s examination manual. Moreover, Regulation C does not distinguish between “key” fields that are essential to determining compliance with applicable law and “nonkey” fields that are useful but not critical for HMDA analysis. The absence of clear – and reasonable – standards produces different outcomes from one examination team to the next and ultimately results in increased costs to consumers.
We urge the CFPB to codify data integrity standards with reasonable tolerances either in Regulation C or in authoritative guidance. In the process, we urge that particular attention be given to establishing reasonable tolerances in fields where there is no value threshold. Specifically, we ask the CFPB to allow for reasonable tolerances in at least the following fields: rate spread, property value, DTI, total origination charges, application date, loan amount, credit score, any date field, and any loan amount field. In these fields, the absence of exact figures does not materially impact analysis.
A. Distinguish “key” and “non-key” fields and provide appropriate tolerances for each
The Associations urge the CFPB to distinguish “key” and “non-key” fields and tailor the regulatory accuracy standards to reflect the fact that certain fields are more critical than others in determining compliance with applicable law. We recommend a 5% tolerance for key fields and 10% tolerance for non-key fields.
In addition, institutions should not be required to re-file due to inaccuracies in non-key fields unless the error rates affect a material percentage of loans of the total loans reported in a Metropolitan Statistical Area (“MSA”).
We suggest that the data integrity standards distinguish among the following types of fields and provide for appropriate standards for each type of field:
• Key Field - Borrower Characteristics – These fields include Race, Ethnicity, Sex, and Age and fields that identify low-to-moderate income borrowers and geographies. We believe these are key fields and lower error rates are necessary to have a reliable unbiased analysis of the protected class or LMI impact.
• Key Fields - Loan Attributes – These fields include fields that relate to pricing, loan terms and the action taken and reflect the underwriting decision. An analysis of HMDA data will generally assess how outcomes differ for different protected classes or LMI borrowers or households. Random errors in outcome fields will not be biased because the distribution of the errors would also be random. We suggest that a 5% error rate for outcome fields would be sufficient for valid analytical results.
• Define All Other Fields as Non-Key Fields – subject to a 10% tolerance
B. Eliminate the File Error Rate
We believe the file error rate should be eliminated as it does not have a material impact on fair lending analysis as long as the field error rates are met. Not only is there no benefit, the file level standard is impractical and virtually impossible to meet. Even if the error rates are kept below 1% per field, the whole application can still “fail” if more than 10% of the categories contain more than one error. The Proposal will make the per file rate even more burdensome because it increases the number of fields from 36 to 72. Increasing the number of reportable fields, as proposed, and maintaining file error rate standards would greatly increase costs which will be passed onto consumers with no appreciable benefit to either lenders or borrowers.
VII. AVOID UNDUE COVERAGE OF INSTITUTIONS AND REPORTING REQUIREMENTS
A. The Reporting Threshold Should be Increased The Proposal would lower the threshold requiring reporting under HMDA for nondepository institutions so that a lender with an office in a MSA making 25 or more loans a year or just over two loans per month would be required to report. The Proposal would also adopt a new threshold for depository institutions so that an institution making 25 or more loans per year would be required to report if it also met the relatively small asset size requirements for banks and had an office in an MSA.
The Associations urge CFPB to reconsider the proposed change that would reduce the threshold for non-depositories from 100 loans and establish a new 25 loan threshold for depository institutions. While adopting a consistent threshold makes sense, the Associations believe this new 25 loan threshold will increase the burden on non-depository lenders unnecessarily and risk depriving consumers of credit and increasing their costs. At the same time, while a new threshold for depository institutions is welcome, adopting a 25-loan threshold will do little to alleviate the burden. We urge that the Bureau adopt a higher threshold of at least 250 home mortgage financing transactions each year. This is particularly important for smaller depository and non-depository lenders that operate in rural areas.
Increasing the threshold would not compromise the integrity of the information reported. Increasing the threshold would also be greatly beneficial to lenders that operate in rural agricultural communities. It is estimated that the exclusion of lenders with fewer than 250 transactions will still capture 95% of the loans made. Certain lenders report so few mortgage loans that their involvement in the housing market simply is not meaningful in assessing market activities. Requiring these entities to report will force them to make expensive technology upgrades as well as business process and personnel changes.
Most importantly, if this new threshold is established, it is likely that the costs and risks that attach to HMDA reporting will cause lenders that make a small number of loans to exit the mortgage lending market, in turn reducing credit access and increasing consumer costs.
We also support the recommendation made during the SBREFA process that the Bureau adopt a two-year look back period when determining whether a lender must report under a threshold.41 This approach has been effective in the CRA context as prudential regulators categorize institutions by asset size looking at two years of their assets.42 By using a two-year period, the CFPB will get a more consistent pool of reporters and alleviate the reporting burden for lenders that simply had a spike in lending in the previous year. Accordingly, we ask that the Bureau establish that institutions must report only if they met the loan threshold in both of the two preceding years.
1. Expanding HMDA Raises the Barrier to Entry for Smaller Institutions and Decreases Access to Credit
The presence of various types of lenders in the mortgage market – including large banks, smaller banks, credit unions, and independent mortgage bankers – promotes competition that benefits consumers’ access to credit and lowers mortgage costs. To ensure a continuing variety of lenders to serve all aspects of the market is another reason the Associations urge the Bureau to raise the reporting threshold to 250 loans. Any lower level will further consolidate the mortgage market and negatively affect consumer access to credit. While the Bureau stated it does not anticipate “any material adverse effect on credit access,” requiring institutions that originate 25 loans will drive smaller lenders out of the market or incentivize lenders to offer fewer loans. This is even more likely given the weight of regulatory changes already straining resources and adding to the sluggish economic recovery.
The Associations believe that the CFPB’s cost benefit analysis of the Proposal woefully understates the costs of compliance. The impact on lenders and ultimately consumers is likely to be much greater than estimated.
The burden of these proposed changes on smaller lenders was stressed repeatedly by participants in the SBREFA panel. These changes will in many cases either price smaller lenders out of the market or force them to abandon or greatly restrict their lending portfolios limiting credit for consumers. Some of our smaller members indicate they will be inclined to constrain their lending below 25 loans per year or leave the market because being forced to comply with HMDA could potentially drive them out of business.
2. Higher Operational Costs Limits Innovation and Increases Costs to Consumers
The latest Banking Compliance Index (“BCI”) underscores the regulatory costs faced by smaller institutions that are ultimately borne by consumers. According to the BCI, community financial institutions saw a 26 percent increase in the number of hours and employees required to meet regulatory compliance demands between the second and third quarter of 2014. 43 Additionally, the BCI “found that the average community bank needed to devote 653 additional hours, or the equivalent of 1.86 full-time employees, to manage the 82 new regulatory changes added in the third quarter. To meet those needs, the average institution had to spend an additional $45,264 on compliance last quarter.”44 We urge regulators to consider the consequences of these increased compliance expenses on innovation, service to consumers, and costs to consumers for obtaining a mortgage.
Additionally, a recent Goldman Sachs report (“the Goldman Report”) concluded, with regard to both credit cards and mortgages, higher regulatory costs have disproportionately affected low income consumers and effectively priced many out of the market. In the mortgage market, the Goldman Report found the differential between the rates paid by borrowers with lower and higher credit scores has grown post-crisis. Specifically, pre-2008, a borrower with a FICO score of 620 paid approximately 3.5% more than a borrower with a score of 800, while the same borrower today would pay 8.7% more.
Further, due to heightened regulatory scrutiny, mortgage credit to sub-prime borrowers has practically dried up, originations in the jumbo mortgage market are half of the 2000-2007 annual average, and the home equity market has experienced a pricing surge and dramatic decrease in originations. The only segment of the market that has expanded post-crisis is mortgages guaranteed by the Federal Housing Administration or the Veterans Administration. While these loans only represent 20% of the market, this further underscores how policy interventions shift the allocation of credit.
B. HMDA Data Should Be Reported Annually
The Proposal directed larger institutions to report quarterly, which is defined by lenders that processed 75,000 applications in the previous year. However, we do not see a material benefit in switching from annual to quarterly reporting and offer several reasons why such a change is inadvisable.
It is not evident that quarterly reporting will improve the supervisory process or public review since a meaningful analysis of fair lending cannot be based on one quarter of reporting, and perhaps not even on one year, and data will still only be released to the public annually. Moreover, as discussed below, quarterly data is likely to be less accurate and thus less useful for stakeholders.
Quarterly reporting will result in significant data integrity issues, beginning with the timing of reporting loans that are in process at the end of a quarter. While lenders may be able to adjust any data discrepancies at year-end, quarterly reporting will raise the possibility for errors.
In addition, each quarterly report will require an extensive quality control review before the report is submitted. The process includes manual review, corrections, systems reviews and involves considerable personnel time and costs. Quarterly reporting necessitates a quadrupling of the quality control process and a quadrupling of associated personnel time and costs.
The current annual reporting cycle enables financial institutions to perform rigorous quality control and deliver to regulators and all stakeholders a much more accurate picture of their mortgage lending operations than quarterly reporting will allow.
It is also important to understand that systemic errors can take months to resolve, which would be unachievable in a condensed three month reporting timeframe. As a result, the data integrity of each quarter prior to corrections will suffer and be less reliable.
Further, we note the confusion created by transactions that extend beyond the quarter. For example, if a loan is originated in one quarter and sold in a subsequent quarter of the same year, the earlier filing will need to be updated to reflect the sale. If only loans that reach final disposition in the quarter are to be reported, it is unclear how to correct errors in submissions. Alternatively, if each quarterly submission includes all of the loans that have reached final disposition so far that year, the recording would overwrite the earlier submissions, which would then be of little use.
It is also worth noting that the Bureau underestimates the impact quarterly reporting would have. If the Proposal were adopted without change, the number of loans that would be reportable would significantly expand and the amount of data to be collected and reported would grow exponentially. Our members indicate that adding HELOCs alone will double – or even triple – the number of loans to be reported. Therefore, the number of institutions required to report quarterly would be far greater than what the Bureau estimates.
VIII. CONFORM REGULATION C TO RELATED MORTGAGE REGULATIONS AND INDUSTRY STANDARDS
The Associations urge the CFPB to align HMDA with established mortgage industry standards to minimize confusion and increase data integrity. To further increase clarity, we also urge the CFPB to conform HMDA requirements to the items outlined below.
A. TILA/ RESPA
The Associations support the CFPB’s proposals that would align HMDA with TILA/ RESPA standards and urge the Bureau to adopt consistent standards whenever possible. Specifically, we ask the CFPB to conform the following Regulation C fields and principles to the TILA/ RESPA Rule:
• Loan or Application Type
• Purpose of Loan or Application should be defined as purchase, refinance, construction, or home equity
• Pricing fields including using the same definition of Rate Set or Lock Date when computing the rate spread
• Definition of Prepayment Penalty (excluding recapture of third party closing costs)
• Loan Term
• Introductory Rate Period
• Non-Fully Amortizing Features
• Entity-created loan identifier ID
• Individual loan originator ID#
Further, the definitions of “home improvement loan,” “home purchase loan,” and “refinancing” in the Proposal continue to include a loan secured by one property where the loan proceeds are used to improve, purchase or refinance a lien on another property. This is inconsistent with Regulation Z because under the TILA/ RESPA rule “purchase” and “refinance” loans are loans where the proceeds are used to purchase or refinance the property securing the loan.
Where loan proceeds are used to purchase, refinance or improve a property that is not security for the loan, that loan should be reported in the “home equity” category. This would be consistent with how such loans are underwritten and priced and with how consumers and the public view them. Data integrity for both Regulations C and Z would be improved by eliminating unnecessary inconsistencies.
Finally, where a loan has more than one purpose, we ask that the hierarchy to determine the purpose reported be the same as TILA/ RESPA (Purchase, Refinance, Construction or Home Equity) rather than the current hierarchy (Purchase, Home Improvement, Refinancing).
The Associations appreciate the Bureau’s efforts to use Mortgage Industry Standards Maintenance Organization (“MISMO”) standards in HMDA reporting. These standards are developed using a voluntary consensus process in accordance with OMB Circular A119 that calls for government to leverage voluntary consensus industry standards rather than proprietary formats. Today, most participants in the mortgage industry use MISMO directly or indirectly and the Associations welcome the CFPB’s efforts to align HMDA reporting with MISMO standards to provide a common language for exchanging data across the mortgage industry.
For those industry entities that are not familiar with MISMO, we believe that training should be provided and the Associations are willing to help. We nevertheless believe that this is a factor, which should be considered when establishing a reasonable implementation schedule.
We also ask the Bureau to consider using, or at least explicitly state that lenders may use, the MERS Mortgage Identification Number (MIN) as the core of a universal loan identifier (“ULI”) for the HMDA program and other areas where a ULI is required.
IX. COORDINATE WITH REGULATORS ON COMMUNITY REINVESTMENT ACT EFFECTS OF DATA CHANGES
We urge the Bureau to coordinate with the prudential regulators before making amendments to HMDA that may affect CRA reporting, ultimately conveying results that are not representative of CRA progress. Our members are committed to serving low- and moderateincome (“LMI”) communities and take great pride in their community reinvestment initiatives. However, they are concerned that including commercial and small business loans on the HMDA LAR will confuse the public and understate the institution’s CRA performance. If the CFPB requires commercial loans, including Small Business Administration (“SBA”) loans, secured by residential real estate to be reported for HMDA purposes, it will add a significant loan volume to the HMDA LAR, even though many commercial loans are not for the purposes of home purchase, refinance, or home improvement. Additionally, due to differences between the mortgage and equity markets, the addition of HELOCs will conflate CRA numbers, but will not shed light on consumer access to credit because homeowners can use the equity in their homes for many reasons that have nothing to do with housing. These revisions have the potential to distort the data and alter public perception.
For these reasons, the Associations urge that the Bureau coordinate with regulators on the effects of HMDA changes on the Community Reinvestment Act.
The Associations appreciate the opportunity to comment on this Proposal and the Bureau’s work to implement Dodd-Frank’s requirements to provide additional insight into the mortgage markets. However, given the statutory purpose of HMDA, we respectfully urge the CFPB to confine the data requirements to loan data relevant to home mortgages and to consider additional concerns about particular data fields expressed by our members. We also urge the Bureau to move forward deliberately to protect and secure consumers’ confidential information and provide an opportunity for feedback. We also recommend that new data not be released to the public until the issues we raised have been appropriately resolved and certain data only be collected and reported in aggregate form. Finally, we urge consideration of our request for sufficient time to implement these rule changes and our other comments.
If you have any questions, please contact any of the undersigned Associations. We would welcome an opportunity to meet with Bureau representatives to discuss these comments and hope to continue to engage in a beneficial dialog as the rulemaking is finalized.
American Bankers Association
Consumer Bankers Association
Consumer Mortgage Coalition
Financial Services Roundtable
Housing Policy Council
Mortgage Bankers Association