How are you monitoring non-HMDA-related loans for risk?

Many organizations might not have a repeatable or scalable process in place to consistently and accurately secure and maintain the necessary non-HMDA loan data elements needed to monitor fair lending risk.

Data extraction can often feel like a piecemeal process. For example, organizations might find that loan application fields are missing information, are filled in inconsistently, or have different naming conventions than the more uniform HMDA loan codes and fields. Organizations also might store their application data in different systems or include disparate pieces of information, such as the valuation of collateral for indirect or vehicle equity loans. Ensuring that consistent and correct data is available is paramount, and it highlights the need to address potential data governance process gaps.

One common proxy method used for meaningful fair lending analysis is Bayesian Improved Surname Geocoding (BISG), which relies on an applicant’s first and last names and street address to determine the probability of an applicant’s race, ethnicity, and gender. An organization can run a statistical analysis on known customer information, loan decisions, and pricing information to evaluate fair lending risk. But that still means organizations should have strong data governance to avoid incomplete or inaccurate results.

Fair lending risk is a challenge, and it can be hard to know where to start. Crowe fair lending risk specialists can help define the variables and loan attributes your organization needs to collect to determine the impact and potential scope of your data.

Crowe fair lending consultants combine data analytics with deep banking industry experience and regulatory knowledge to help organizations accurately assess fair lending risks and then develop the policies and procedures to address them. We can help you apply analytics, proactively assess risk, and prepare your program for what’s ahead.