Fair Lending: A Definition

A definition of fair lending is that it is about financial institutions treating similarly situated applicants similarly.  That is, for example, a prohibited basis applicant should be treated like a non-prohibited basis applicant if they have similar personal economics (DTI, LTV, credit score, etc.), are applying for loans that have similar risk characteristics such as loan term or fixed rate versus adjustable rate and have other similar features such as coming from the same MSA or were applied for in the same time period.  The fact that financial regulators (the Federal Reserve, the Office of the Comptroller of the Currency, the Federal Deposit Insurance Corporation and the Consumer Financial Protection Bureau) examine financial institutions to ensure that they are adhering to bank compliance regulations is what puts the topic at the top or near the top of most institutions’ priority list.  Failing to follow the fair lending compliance regulations can result in civil money penalties, restrictions on branching and significant reputational damage for an institution.

Avoiding Fair Lending Non-compliance

In order to avoid the above penalties of non-compliance with fair lending guidelines, financial institutions perform fair lending analyses.  Taken as a whole, banks establish fair lending programs to assess the fair lending risks each institution faces.  There are six fair lending risks banks and similar types of institutions face:  underwriting, pricing, steering, redlining, marketing and loan servicing.  For each of these risks, institutions may perform analyses to assess their risks.  The typical types of fair lending analyses include:  compliance management system evaluations, risk assessments and regression analyses.

Compliance management system evaluations generally are more qualitative in nature when compared to risk assessments and regression analyses and have five areas of review:  board oversight, the compliance program, compliance audit, complaint response and loan servicing.  The makeup of a particular institution’s compliance management system depends largely on the institution’s size, number of branches, organizational structure, business strategy, types of products, third party relationships, etc.  Nevertheless regulators want to see certain elements clearly demonstrated:

  • Board and management involvement in the institution’s compliance program as demonstrated, at a minimum, in Board and management meeting minutes.
  • Written bank compliance policies and procedures that are updated regularly and proactive monitoring systems.
  • Existence of a compliant response system even if there are no current complaints.
  • Compliance audit function to ensure the compliance management system is doing what it is designed to do.
  • Loan servicing system that ensures fair lending issues do not occur anywhere in the lending process from the intake of the application to the servicing of the loan.

Risk assessments are designed to tell an institution where its bank compliance issues or fair lending risks appear to be.  Often based on the institution’s HMDA data, the output of a risk assessment indicates where there may be fair lending issues by risk (underwriting, pricing, steering, redlining, marketing or loan servicing), focal point and prohibited basis group (American Indians, Asians, Blacks, Hawaiians-Pacific Islanders, Females, Hispanics, Seniors).  The results of a risk assessment are not definitive since they merely present the results of statistical calculations and give no indications as to why the results are what they are.  They merely indicate more investigation is needed.  For institutions with sufficient numbers of applications, regression analysis is typically the next step in investigating the possible fair lending risks indicated in a risk assessment.

Regression analysis uses advanced statistical techniques to determine why the results from the risk assessment are what they are.  The advantage of regression analysis is that it is conducted at the loan level rather than at the prohibited basis level.  In other words, regression analysis points out which applicants appear to have been treated differently (such an applicant is often referred to as an exception or outlier).  Furthermore, the analysis also matches similarly situated applicants to each exception applicant where matches exist.  Thus, a regression analysis allows an institution to do two things.  First, using the listing of exception applicants, the institution can go to the loan files of each exception applicant to be sure it can justify the credit or pricing decision that was made based on the documentation in the loan file.  Secondly, the institution can use the matching report to ensure it is being consistent in its decision making across the spectrum of applicants.  This type of review is often referred to as a matched pair analysis although there may be more than pairs to review.  It should be pointed out that this matched pair analysis is different from a second review process which generally looks to see if an individual loan can be made without considering the decisions that have been with respect to other similarly situated applicants.  In other words, second review processes look at the individual trees in the forest whereas the matched pair analysis looks at the forest itself.  A last comment with respect to regression analysis is that in many situations the regression results cast doubt on the risk assessment results and can show it is likely that no differential treatment is occurring.

How Do You Choose?

A final fair lending comment, the question often comes up should I buy fair lending software or should engage someone who does bank compliance consulting?  While there are many factors to consider in such a decision, a few of them are:  direct costs for software, training, hardware and personnel.  Additional decision factors may be:  support policy, updates, software ease of use and what is the experience of other banks my size using the software.  In general, it is a good policy to walk before you run.  That is, use an outside consultant for a period of time in order to get a sense what is involved in doing fair lending analysis, then, after you have gained a little experience with fair lending analysis .if you feel buying software is for you and your organization you should do it.