The regulatory world in which financial companies operate is a vast sea of uncertainty that is seemingly in a constant state of churn. Although the laws and regulations themselves are not necessarily constantly changing, regulatory interpretation and shifting priorities, coupled with bifurcated and shifting agendas, create a sort of organized chaos.
This makes compliance difficult, because what “compliance” looks like may differ by examination, examiner, and agency. This applies to not only what may constitute an issue but also what the focal points and areas of emphasis are. This also includes where the bar is set in terms of compliance, which can often be a moving target.
This makes for a very unstable foundation for compliance risk managers who are charged with managing and reducing regulatory risk for their institutions. At its core, risk-reduction only occurs to the extent that uncertainty is reduced. To reduce risk is to reduce uncertainty; if uncertainty cannot be effectively reduced, there is no tangible corresponding reduction in risk.
Facing Fair Lending Uncertainty
Most fair lending practitioners would agree that “uncertainty” is perhaps the best word to describe the fair lending regulatory environment today.
How is one able to reduce fair lending risk in such an environment when the ground beneath you seems to be constantly shifting? This is a complex question, and one that could differ considerably by institution and circumstances.
In general, however, an effective fair lending risk mitigation framework must be aggressive, proactive, forward-looking, adaptable, and agile. It is also critical for the compliance management function of the institution to have executive management’s full attention when necessary.
Likewise, management must be committed to fair lending compliance and have complete respect and trust in their leadership from their respective teams. A customer-focused service culture is also more conducive to reducing fair lending risk. Institutions that rely solely on profit-directed priorities will likely encounter difficulty in the current environment.
There are a number of landmines and pitfalls that can hinder institutions along the way. Despite the best efforts and intentions, and because regulatory priorities and the standards do shift and change, even the best institutions, with otherwise stellar compliance management systems, may find themselves in the fair lending crosshairs.
When Reducing Risk Isn’t Enough
This may be in the form of an intensive examination that may potentially result in a fair lending enforcement action, or worse, an enforcement action itself. This is where the role of regulatory forensics and the effective application thereof may mean success or failure.
The term forensics, and more specifically as it applies to statistics and mathematics, refers to the use of quantitative methods applied to scientific evidence and data. Today, statistics and quantitative methods are the currency of fair lending regulation and enforcement. These are the building blocks upon which any significant regulatory action is based. It may, therefore, be necessary at some point for any institution to counter alleged discrimination by similar means.
There are (3) types of potential discrimination recognized by the regulatory and enforcement agencies. These are (1) overt discrimination, (2) disparate treatment, and (3) disparate impact.
Overt discrimination is the most obvious form and what comes to my mind when most people think about discriminatory practices. This form is rare, however, because it is blatant and obvious, easily detectable and, thus, largely avoidable.
The latter two forms, disparate treatment and disparate impact, and primarily disparate treatment, are where the vast majority of fair lending issues occur and constitute the majority of the risk fair lending financial institutions face. These types of fair lending violations are largely established by data, statistical patterns, and relationships within these data that suggest discriminatory practices.
These types of discrimination are largely the focus of the agencies because the regulations refer to “patterns or practices” of discrimination, and these forms may stem from unintended systemic issues or hidden biases within an institution that, nevertheless, contain fair lending risk.
The Role of Statistical Analysis
Although there may be other supporting factors surrounding an alleged fair lending violation, the primary evidence presented are typically statistical patterns present in the data.
In other words, the data indicates there are differences in treatment between protected and non-protected groups; and these differences are unexplained by legitimate, non-discriminatory factors. Accordingly, these form the basis for the alleged discrimination. This evidence is then further subjected to more rigorous statistical and econometric methods to add the “scientific” layer described earlier, and that is determining if the differences noted in treatment between protected and non-protected groups could have occurred by chance, or if they are “statistically significant”.
Such is often sufficient to form the basis for a pattern or practice of discrimination. A statistically significant finding indicates that there is:
- Correlation between treatment of applicants with respect to loan outcomes based on protected and non-protected class status, such as higher rates charged to female borrowers when compared to male borrowers, and
- That this correlation is unlikely due to chance or random variation.
Although correlation does not equal causation, there is an implicit causal relationship assumed when a statistical finding is used as a basis for determining that discrimination has occurred.
The methodological framework applied is a target – control group structure in order to test differences between these groups. Think about an experiment where the effectiveness of a medication is being tested, where there would be a test group (target group) that received the medication and a placebo (control group) that did not receive the medication. The principle is the same in that in fair lending we are comparing outcome differences between groups.
In the former, however, using more of an experimental methodology where the researcher has control over the environment, it would be more appropriate to interpret results at least somewhat causally. As opposed to the typical fair lending analysis in which the opportunity for this type of control over the environmental factors is much weaker, it is much more difficult to establish a causal relationship, scientifically speaking.
That said, however, often such a statistical relationship is a sufficient basis for the regulatory and enforcement agencies to infer a pattern or practice of discrimination has occurred and, therefore, a violation of law, with the data and statistical findings constituting the building blocks. This is where the role of regulatory forensics becomes critical.
How Regulatory Forensics Can Help Fight the Battle
There are a number of different reasons why the data on the surface may sometimes suggest bad acting on the part of the institution that simply does not hold with further investigation.
The multitude of potential reasons are well beyond the scope of this post, but there are typically alternative explanations that may have been overlooked, methodological problems with the analysis, erroneous assumptions, or issues in the data. These types of issues can stem from miscommunication or misunderstanding of relevant facts and an almost endless range of possibilities.
During the typical inquiry, examination teams must review a large volume of information and documentation, gather information from discussions with management, and juggle a number of different tasks. There are different layers in the process with significant opportunity for critical pieces of information to get lost in the translation.
Whatever the cause, these are the types of things that could be uncovered in a forensic statistical analysis. The “building blocks” upon which the findings were based would be dissected and examined to determine their validity.
Fair lending is one of the dominant risks facing financial institutions of all sizes today. The best solution for fair lending compliance is to have effective management and aggressive and focused programs to, as feasible, reduce risk. With the uncertainty of the environment today, however, this is becoming more and more of a challenge.
The blunt reality is that any institution may find itself past the routine examination and into a more extensive inquiry; and the possibility of it taking a negative turn, or at least getting worse before it gets better, is a distinct possibility. This applies to ALL institutions, and it is a contingency that must be considered and planned for in order to have comprehensive management.
If your institution is in need of guidance or a forensic analysis, our team is experienced and ready to work with you. Reach out to us below to discuss your bank’s needs today.
