During the Biden presidency the prudential bank regulators and the DOJ began to prosecute cases of alleged redlining much more aggressively than had been historical practice. A key concept in almost all cases was the idea of a “REMA”, or “Reasonably Expected Market Area.” A REMA is not defined in statutes nor in regulations, nevertheless it played an extremely important if not critical role in redlining cases in recent years.
REMA background and history
A REMA is a concept developed and applied in the field beginning around 2009 when it appeared in regulatory compliance examination manuals. Shortly thereafter between 2010 and 2020, bank regulators broadcast webinars explaining the concept and how examiners were to apply it in the field. Notwithstanding explanations in the examiner compliance examination manuals and the years of repeated webinars explaining the concept, the absence of an official definition of a REMA left room for ambiguity and potential confusion in the banking community as legal cases applying the concept wound their way through the courts.
The murkiness surrounding the definition of a REMA became even more murky when, under the Biden Administration, the bank regulators and the Department of Justice abruptly adopted a radically different approach to how a REMA was defined. The DOJ announced the “Combatting Redlining Initiative” in October 2021 and shortly thereafter bank regulators proclaimed that a REMA would be nothing less than a MSA or MD (metropolitan Division) or multiple counties in non-MSA markets. This new definition is wildly impractical for many community banks. Nevertheless, it was used in some redlining cases prosecuted by the DOJ.
This history exposes 3 problems with the REMA concept as applied by regulators. First, the concept was never clearly defined in regulations (although an informal explanation in the form of webinars was broadcast by regulators from 2009 to 2021). Second, the concept as applied by examiners abruptly changed in 2022. Third, the new definition was not realistic for many community banks.
Why the REMA concept is so important and needs clarity
The REMA concept is incredibly important to redlining cases because it’s necessary to define the market(s) in which redlining is alleged to have been practiced. The population demographics of any market reveal (1) where majority-minority neighborhoods exist and (2) determine the significance of the minority population within the entire community. Minority populations may be concentrated in “majority-minority” tracts or may be disbursed more evenly throughout a market and manifested in census tracts with “relatively concentrated minority population tracts” (in practice in the field, tracts with a minority population of 20% to 50%).
Even more important, a REMA affects the apparent geographic dispersion of a bank’s lending to and within minority neighborhoods (a bank’s “penetration rate”) relative to its lending in the REMA.
Expanding or contracting a market where redlining is alleged to have happened has a direct impact on the market statistics to which a bank is being compared as well as the bank’s apparent penetration rate lending in majority-minority tracts. These appearances can be distorted and therefore misleading.
A good example of the foregoing problems is demonstrated by the following table which depicts results in a theoretical market consisting of 2 counties. One county has no majority-minority census tracts while the second county has MMCTs.
In County 1 all banks have no mortgages in MMCTs because there are no MMCTs in that county. But in County 2 where there are MMCTs, Bank 1 significantly outperforms the other banks lending in the MMCTs in that county with a 60.0% penetration rate while Bank 2’s MMCT penetration rate is 12.0% and Bank 3’s is 11.6%. When the 2 counties are combined into a REMA Bank 1 has a statistically significant low MMCT penetration rate of 6.0% while Bank 2 has a MMCT penetration rate of 11.8% and Bank 3 has a penetration rate of 10.2%.
The outcome in this situation shows that the bank (Bank 1) with the highest MMCT penetration rate in the only county with MMCTs (and therefore is the only county where redlining could have occurred) would have a statistically significant low MMCT penetration rate for the REMA (combined counties) and would be a candidate for prosecution of alleged redlining. This is an example of Simpson’s Paradox in which statistical analysis of disaggregated data contradicts statistical results from aggregated data. The validity of a statistically significant result is suspect in such situations and the conflicting results are evidence of an unidentified “confounding” variable or hidden factor.
What should regulators do to fix the REMA problem?
First, there should be a regulatory definition of a REMA to remove any ambiguity about what it is and how it is applied by examiners. By adopting a regulatory definition, regulators will remove the ambiguity of an undefined term and make it more difficult to abruptly and arbitrarily change that definition as was done under the Biden Administration.
Second, regulators should recognize that statistical analysis, while very useful as a screening tool, can be very misleading in some cases. In practice, statistical analysis should incorporate a two-prong approach which uses both aggregated data and disaggregated data to determine if a result is consistently statistically significant.Statistical analysis is not infallible. It is a useful tool to identify potential discrimination, but it is not conclusive. As the Supreme Court said in Inclusive Communities statistically significant results alone are not sufficient to prove discrimination under the legal standard.
[View source.]
