Over the past year, financial institutions have paid closer attention to fair-lending issues. Even institutions regulated by the Department of Housing and Urban Development and the National Credit Union Administration — two agencies that have not been as visible to lenders — are under pressure.
Why this sudden change in attitude over fair lending?
There is no question that the crisis surrounding the current mortgage industry, specifically regarding subprime lending, option adjustable-rate mortgage products, rampant delinquencies, defaults, foreclosures and repurchases, has caused regulators to scrutinize lenders' activities. The result has been several high-profile penalties at state and federal levels.
In 2010 there were two fair-lending settlements that shook the industry: In March, two subsidiaries of American International Group settled with the Department of Justice and agreed to pay $6.1 million to African-American borrowers for alleged discriminatory pricing practices; and in September, Golden Empire Mortgage settled with the Federal Trade Commission and agreed to pay $5.5 million to Hispanic borrowers for allegedly giving higher-priced loans based on discretionary pricing policies.
The size of the settlements made national headlines, but it was the size of the institutions that made them eye-openers to many financial executives and risk managers.
In particular, Golden Empire Mortgage caught everyone's attention because, according to public Home Mortgage Disclosure Act data, it only submitted an average of 7,200 applications per year from 2004 to 2009. If an institution with relatively low volume like Golden Empire can be singled out, then every institution is at risk.
Many mortgage companies have serious concerns about their fair-lending risk because they do not have fair-lending compliance programs in place, and truth be told, most of them allow the same discretionary pricing as Golden Empire. Without a solid program and software analysis of the specified risk factors, these institutions are facing the same risks as AIG and Golden Empire.
But as the cliche goes, the best defense is a good offense. There are basic steps an institution should perform to help eliminate fair-lending risk.
First, update fair-lending policies and procedures. This can be done with internal personnel if your institution has the knowledge and resources. It can be costly and time-consuming, however, so evaluate your resources and determine the best approach for your institution.
At the same time, start analyzing data. Every lender's data tells a story, and it is much better if the institution knows what the data is saying before the examiners get involved.
There are two basic methodologies for analyzing your data: risk-factor scorecards and regression analysis.
Risk-factor scorecards analyze absolute disparities in pricing, denial rates and other sources of potentially inconsistent treatment, and are useful in situations where a statistical model is not available.
Regression analysis takes variables into consideration to model lending decisions, and can be used to refine an examination by eliminating false positives.
Lending volume, the range of products and services you offer, the general demographic makeup of your lending areas and many other factors will determine which method or combination of methods you should use.
Once you understand the picture, you'll be able to determine if there is a problem.
If you detect disparate treatment or disparate impact from your data, you will need to consider appropriate corrective action. Your policies and procedures will need to be adjusted to address these problems, and you will need to review your training program to help prevent problems from recurring.
Finally, continued monitoring of data will be necessary to make sure these adjustments are working as intended.
Regulations and their enforcement are only getting tougher with the formation of the Consumer Financial Protection Bureau and the passing of the Dodd-Frank Act.
Fair-lending analysis, however, can be a complex task, and depending on what you find during your analyses, it can become even more complex.
Cleaning up and normalizing your data, and setting up accurate regression models are not jobs for a novice. Institutions should form a dedicated and knowledgeable team.
Remember, it can be expensive, but the alternative could potentially be a multimillion-dollar settlement, as AIG and Golden Empire can attest.
John A. Woloshen is executive vice president and chief operating officer of RATA Associates, a provider of HMDA/CRA data compliance software and services.
