By Loretta Kirkwood
Managing Director
and
Perry Wilson, Manager
CrossCheck Compliance
The Home Mortgage Disclosure Act (HMDA) continues to generate challenges for financial institutions and mortgage lenders year-after-year. The accuracy and completeness of HMDA data is critical to fair and responsible lending risk management, as well as lending distribution analysis for the Community Reinvestment Act (CRA).
Risks occur in the normal course of business – new systems, new loan programs, new reporting software, staff changes, and process changes. Add to that, the significant increase in the number and amount of Civil Money Penalties (CMPs) and the proposed laundry list of additional information to be collected and reported for HMDA … and, data quality processes have to be constantly adjusted to address the increased risk of inaccurate data.
The simple truth is … HMDA data quality is at the top of a very long list of regulatory risks. “Getting it right” is no longer just a guide, it is the destination!
Regulatory Focus
HMDA data is used in fair and responsible lending examinations, investigations, and enforcement actions, as well as in CRA performance evaluations. Inaccurate data leads to flawed results, delays in the examination process, and often mandated HMDA scrubs – which can be very costly and time consuming. It has become clear over the past several years that regulators have shifted attention to more sophisticated data analysis and testing throughout the examination process.
Inaccurate submission of HMDA data can additionally result in the assessment of CMPs, which have increased in number and amount in recent years. Although the number of HMDA CMPs in 2013 was less than2012, this downward trend is not expected to continue. It is important to note that the total amount of CMPs is similar at $486,000 for 2013 and $446,700 for 2012. CMPs during this period ranged in size from $1,500 to $60,000, with the exception of the CFPB’s $425,000 assessment against Mortgage Master. Asset sizes of institutions receiving HMDA CMPs range from $52 million to $13 billion and number of LAR entries from two (2) to 15,000. The take away from this analysis is – no lender is exempt from enforcement action relating to HMDA data quality.
Regulators aren’t simply focused on enforcement – it is about providing the public with information on financial institutions and access to regulators to communicate complaints against financial institutions. The CFPB has launched several data driven initiatives which are distributed on-line and immediately available to the public. The complaints database and enhanced HMDA analysis tools are providing access to information and data not previously available. This generates a different kind of risk for lenders … attention from a much broader group including special interest groups, community groups, media, and the public in general.
Internal Controls
Important things to consider in managing HMDA data quality include not only data capture, collection, testing and submission, but also developing an effective HMDA compliance management system. Board and management oversight, training and education, policies and procedures, and independent reviews are key components of HMDA data integrity risk management.
Training and education must be reviewed and adapted to ensure that the content, materials and presentation provide guidance not only on the regulatory requirements, but also address internal policies, systems, procedures and accountability. Everyone involved in the lending process generally understands the regulation and their respective duties with regard to ensuring the accurate and timely collection of data. But, do they truly understand the higher risk areas that can’t be “reviewed” in file reviews – such as, customer contact either in person, via email or by telephone. It is a good idea to include customer service level training programs and develop standard scripts for explaining the regulatory requirement to prospective borrowers.
Policies and procedures should clearly define regulatory interpretations specifically with regard to a lender’s business practices, system controls, and processes. Data management is an on-going process, starting with data capture at the point of application, through the credit management process and ultimately submitted to the lender’s primary federal regulator.
Procedures should address specific areas of risk in order to identify gaps in data capture, processing, and reporting:
Regulatory interpretations
o All HMDA fields – action, purpose, etc.
o Identifying reportable and excluded transactions
o Define the differences in prequalification, preapproval, and application
Data Capture
o List of source systems / extracts used
o Map of system extract data to HMDA fields
o Description of system controls to ensure data is captured at application
Errors and Omissions Testing
o Definition of manual and systemic processes
o Source of initial geocode and validation
o Method for validation of rate spread calculations
Corrective action plan for identified deficiencies
Monitoring and testing programs, internal audits, quality control functions and ultimately HMDA “scrubs” play a role in ensuring the accuracy of HMDA LAR data. It’s important to develop uniform testing methodologies with consistent knowledge and training for individuals involved in these processes to ensure interpretations and conclusions are without conflict. A collaborative approach across lines of business and support functions such as compliance and audit ensure an effective approach to managing data quality.
Technology
As society continues to move, not away, but toward new technology, new communication methods, social media, and even a virtual currency – it’s time to consider a more sophisticated approach to managing HMDA data integrity.
Random sampling methodologies are still a component of the regulatory exam process … but, depending on the number of applications, decision centers, lines of business, and distribution channels – it may not be the best solution for internal data integrity testing.
Data quality review programs should include both errors and omissions testing. Historically, these functions have been manual in nature – using system reports for reconciliation and physical files to ensure the accuracy of HMDA LAR data. Using technology is the most effective approach for two reasons – 1) most applications are “keyed” into an automated underwriting system or uploaded from a point-of-sale system with paper applications and other documents printed from the system data, and 2) data elements contained in application and loan systems are much easier and more reliable than a review of paper documents.
Systemic data integrity testing can be simple or complex, depending on the number of application processing systems, lines of business, and loan servicing systems. The process involves reconciling HMDA data to application and loan system data using information and data not commonly considered “HMDA” fields. Some examples of systemic validation options include:
Using the check box on the 1003 that asks if the borrower(s) will occupy the property as a primary residence to validate occupancy.
Purpose of refinance narrative fields can be used to check for home improvement purpose applications.
Call and collateral codes on loan systems can be used to identify residential secured and multifamily loans for omissions testing.
Consideration should also be given to the accuracy of HMDA aggregate data. HMDA data is used for more than assessing individual lenders; it contributes to trend analysis of housing needs at a community level. It is also used by lenders and regulators for peer reviews – what percentage of errors would be required to result in inaccurate conclusions? Remember the “garbage in – garbage out” theory of information technology … clearly described as “unintended or inaccurate data IN, produces undesired or negative results OUT”.
New Data / New Processes
Although the timing is uncertain, the reality is that new data requirements are coming and it is important that lenders prepare for the change now. The CFPB recently announced that they are considering proposing rules that change not only what data is reported for HMDA, but also how data is reported.
We are all too familiar with the data requirements defined in the Dodd-Frank Act. The CFPB is proposing to add to the original list as shown in the following chart:
Category
Field
Dodd Frank
CFPB
Borrower Information
Age
x
Credit score
x
Lender Information
Application channel (broker,etc.)
x
Loan originator identifier
x
Loan Information
Universal loan identifier
x
Property parcel number
x
Property value
x
Loan term
x
Rate spread (all loans)
x
Total origination points and fees
x
Negative amortization
x
Prepayment penalty term
x
Term of introductory rate period
x
Interest rate
x
Debt-to-income (DTI)
x
Combined loan-to-value (CLTV
x
Automated underwriting system (AUS) results
x
Qualified mortgage (QM) status
x
Total origination charges
x
Total discount points
x
The CFPB is also proposing changes that will:
Align the HMDA data set with Mortgage Information Standards Maintenance Organization (MISMO) data standards, and
Automate delivery using Application Programming Interface (API) technology to connect lenders software to CFPB back-end systems.
Wrap up
The simple truth is that HMDA compliance management systems and data integrity testing programs are more important than ever before. Staying ahead of impending changes is critical to managing the impact on the mortgage lending process. Lenders must develop and implement action plans and internal reviews to identify current and future risk associated with HMDA data quality.
Fair and responsible lending risk management requires reviewing the past, keeping up with current events, and trying to forecast what is yet to come. HMDA data integrity is an integral part of identifying and managing fair and responsible lending risk associated with housing related programs.
Loretta Kirkwood
Perry Wilson
Loretta is a regulatory compliance and risk management executive at Cross Check Compliance. Loretta has over 30 years of experience in the financial services industry. Her most significant area of focus is fair and responsible lending risk management, including compliance management system and program development, qualitative and quantitative risk assessments, data quality management, process mapping, and staff training.
Perry Wilson is a manager at Cross-Check Compliance LLC with nearly 30 years of experience as a mortgage banking professional. She has extensive knowledge of both banking and mortgage related regulations and assists clients with integration of policies, practices, and internal controls. Perry can be reached at: pwilson@crosscheckcompliance.com
The post HMDA … Hmm, is My Data Accurate? appeared first on Mortgage Compliance Magazine.