During the last 15 years there have been notable changes in regulatory enforcement of potential redlining situations. This article addresses those changes.
Originally, redlining prosecutions were limited to situations in which the lender’s intent to avoid or otherwise restrict access to mortgage credit was demonstrable by specific cases. However, beginning with the Obama Administration in 2009, a statistics-driven approach began to be implemented by regulators using statistics and “statistical significance” as evidence of “disparate impact”, or the unintentional discriminatory effect on protected class groups. This coincided with the advent of the CFPB which was authorized, unlike the other prudential regulators, to file its own fair-lending actions independent of the DOJ.
To implement this statistics-based theory it was necessary to determine what the “market” was that would be the basis of statistical analysis. The concept of a “REMA” or “reasonably expected market area” was the answer to that question. A REMA, conceptually, is an attempt by regulators to determine what is the “real” market for a bank and which could be the basis for establishing a statistical reference point upon which disparate impact could be inferred.
Originally, examiners were directed (there is no statutory or regulatory definition of a REMA) in examination manuals to consider various factors that would indicate if a bank’s true market deviated from its CRA Assessment Area. But the methodology used by regulators to determine a REMA for a financial institution has been changing. Not only that, but there is an indication that the regulators may be also taking a different approach when determining what lenders qualify as peers for redlining purposes.
From 2009 to 2021 examiner manuals instructed examiners to review certain factors when deciding how to delineate a REMA. For example, the geographic distribution of mortgage applications was identified as a specific factor. Examiner manuals also cited other considerations, such as loan production offices and the “coverage” areas for marketing campaigns.
Then in 2022, the agencies announced they would no longer consider a REMA to be anything less than an entire MSA, MD, or statewide non-MSA. This approach is a radical and critical departure from the historic approach to determining the configuration of a REMA. No longer are reasonable factors such as the actual geographic dispersion of a bank’s mortgage applications a consideration.
The adoption of the new approach already has had a very negative impact on the enforcement of potential redlining allegations by distorting the market in which the alleged redlining has occurred. For redlining analysis itself to be reasonable, analysis must be based on the market a bank can “reasonably” be expected to serve. If a market is not realistic the comparisons to the market are not realistic and this in turn can lead to meaningless and even misleading conclusions.
In previous articles I have used the Los Angeles MSA as an example of how distorted redlining analysis can be when using an unrealistic market as the reference point for benchmarks and comparisons on which statistical significance is applied.
In our practice we are now seeing what appears to be a new development: regulators seem to be evolving a new approach to defining competition and peers for redlining purposes. The standard practice has been to identify “peers” in a REMA based on application volumes in the REMA for all lenders who process from 50% to 200% of the volume of the subject bank. Lately, we have observed a deviation from that approach. We are now seeing situations in which regulators are now defining peers as banks and credit unions that have processed 50% to 200% of the application volume of the subject bank and that maintain a deposit-taking facility within the REMA.
This refinement may be an implicit acknowledgement of the criticism expressed by many that the inclusion of mortgage companies as peers creates an unfair benchmark derived from the lending activities of entities that have far different business models from banks and that tend to capture a much larger share of the subprime market.
Not surprisingly, in the situations we have evaluated so far using the new approach, the comparisons have yielded more favorable results for banks under examination for potential redlining. This suggests that when banks do self-analysis for potential redlining, they include this new approach in addition to the standard approach to determine if the results affect statistical significance outcomes.
This will require the identification of banks and credit unions that maintain depository facilities with a bank’s potential REMA. That information can be accessed through the FDIC and the NCUA. Once those institutions are identified they can be used as filters in the HMDA A&D data released annually by the CFPB and the FFIEC.
A statistical analysis based on comparisons to “peer” lenders as determined in the foregoing manner can yield more realistic statistical analysis. Banks conducting internal analysis of potential redlining problems can use this new approach to defining peers to compile an alternative redlining analysis to complement their standard approach. It may be especially impactful if a bank determines it does have statistically significant results using the heretofore standard redlining approach.
