Article Details


Understanding the Total Exceptions Picture - Carl G. Pry

ABA Bank Compliance
March-April 2015
Carl G. Pry

One of the trickiest issues for compliance officers to confront is that of lending exceptions (whether underwriting or pricing). This is an important issue, as any time a deviation from credit policy is permitted it presents fair lending risk that must be recognized and monitored. Are exceptions granted more often to non-minority applicants or borrowers? Even if exceptions are distributed equally, are those that are permitted less beneficial to minorities?
A common theme within banks with fair lending problems is lack of awareness that exceptions are happening at all. This is especially true for lenders following secondary market (investor) guidelines; the assumption is that since only saleable loans will be made, exceptions are impossible. While this certainly could be true, it should never be assumed.
Even if exceptions are permitted, many banks are unaware of the extent to which they occur. But to clearly grasp and quantify fair lending risk, banks must clearly understand the degree to which exceptions are permitted and granted.
Defining exceptions: by any other name…

Ask the following simple question to a lender: “Do you make exceptions?” The answer may be no, never, or maybe once in a while. But is that really the case? Recognize that the problem may be with the question rather than the answer. Lenders may understand the term “exception” to mean something entirely different than the compliance officer does.
What exactly is an “exception” then? The easy definition is any time there is a policy deviation for whatever reason. For example, an approval is granted even though the applicant wouldn’t normally qualify, or a loan’s rate is lower than the pricing matrix would dictate. Whether labeled a variance, concession, allowance, adjustment, discount, exclusion, override, modification, deviation, or something else, the concept is the same. Don’t let terminology get in the way. The proper question should have been, “Are there any situations where credit policy is not followed?”

Step one therefore is to get past what exceptions are called and understand whether and how often they happen. But even then it is not that simple; all exceptions are not created equal. In fact, some situations where policy is not followed don’t create fair lending risk and need not be considered in a fair lending analysis.
Take the situation where, due to an overload of applications in the pipeline, an applicant’s rate-lock agreement expires. It’s not the customer’s fault that the bank took too long to close, but to make things right the applicant is given the locked-in lower price even though rates had risen in the meantime. Normally this would appear as an exception, as the given rate is lower than it should have been. But this “exception” doesn’t present any fair lending risk since it had nothing to do with the customer or his/her qualifications. These types of “customer service adjustments” should not be considered exceptions for fair lending purposes.

Reason codes
Identifying exceptions is only the beginning. Why was each one granted? Too often banks can identify the number of exceptions but are unable to explain why they were granted. Often coding is built into loan originations systems or an exception form completed (with approvals) to identify compensating factors or other reason(s) that led to an exception being granted.

While certainly a good practice, two commonly-seen reasons can be problematic. One is a rate exception provided to “match the competition.” This is certainly a valid reason to offer a lower rate (especially if the bank is not the low-price leader), but can it be proven? Many banks shop the competition, but often this is causal or anecdotal, and not documented. Rarer still is a bank requiring a customer to provide a competitor’s GFE. With no proof requirement, the reason becomes a catch-all for anything and everything.

A similar situation results when an exception is granted because of the customer’s “relationship” with the bank. This may be easier to document, but what constitutes a relationship significant enough for the exception to be made is often ill-defined and subjective. Criteria that changes from applicant to applicant in and of itself presents fair lending risk.

The result is these codes are frequently overused. With no set standard or proof requirement, they can be used in situations where neither is applicable. They then become meaningless, which does not help mitigate any fair lending risk.

What are the limits?

Many banks place limits on exceptions, which is a good idea. Caps can be placed on total number of exceptions allowed or amount of leeway allowed (such as how low a credit score may be below the stated minimum or amount of price reduction or fee waivers allowed, for example).
Arbitrary limits within a set time can backfire, however. For example, if only 10 exceptions are allowed per month, what happens to the 11th applicant who otherwise should have received the same exception as one or more of the first 10? And if that 11th applicant happens to be a minority, that arbitrary limit could set up a fair lending issue – it’s an easily identified matched pair.

The most carefully considered limits don’t serve much of a purpose if they are not properly monitored. Both low-side (decisions to approve or price lower when standards are below the minimum) and high-side (deny or price higher even though standards are met) exceptions should be separately tracked. And they don’t net against each other.

If limits are set but then routinely ignored or exceeded, it’s almost worse than having no limits at all. It points to a lack of control over the process, which puts the credibility of the bank’s compliance management system in doubt.

How many, or what percentage, of exceptions is acceptable? This is an impossible question to answer since exception policies run the gamut from very strict to anything goes. There is no arbitrary number or percentage that is either good or bad, but two things are clear: 1) at some point when too many are permitted, exceptions become the rule (which defeats having credit policies in the first place); and 2) the more discretion allowed, the higher the fair lending risk; therefore the more critical monitoring becomes.
Analyze for fair lending

It should go without saying that exceptions should be analyzed for fair lending issues. Patterns should be identified and comparisons made. Who got an exception and who didn’t, all other things being equal? And why was this so? With increasing attention being placed on fair lending in general, accurate recognition of exceptions granted, as well as why they were, is more important than ever.