Every banker should know by now that the Department of Justice announced an “Anti-Redlining” crusade in October 2021. In his October 2021, press release, Attorney General Merrick Garland, proclaimed that redlining is “pervasive” today and the DOJ will be working with the prudential bank regulators and state attorneys general as “force multipliers” in a new initiative to combat redlining. Since that October 2021 press conference, the DOJ has publicized numerous redlining complaints it filed and related settlements. The DOJ has also said it has received a record number of redlining referrals from the prudential bank regulators.
The question discussed in this article is, “Is anti-redlining enforcement fair as practiced in today’s regulatory environment?”
A review of one of the fundamental tools used by examiners, statistical significance models, suggests that the assumptions underlying statistical significance analysis often are flawed and can result in inaccurate and misleading conclusions. In the DOJ’s jihad against redlining, it’s imperative that every banker knows when these facts apply and be sure to call them to the attention of examiners.
As explained in the Comptroller’s Handbook for Consumer Compliance, “Redlining refers to the illegal practice of refusing to make residential loans or imposing more onerous terms on any loans made because of the predominant race, national origin, etc., of the residents of the neighborhood in which the property is located. Redlining violates both the FH Act and the ECOA.”
Redlining typically is interpreted in terms of a bank’s low number of mortgages in minority census tracts. Consequently, examiners focus on the counts of mortgage applications and originations processed by a bank in the minority neighborhoods within a bank’s defined community.
The major tool used to determine if a bank is redlining measures “statistical significance” which compares the number of mortgage applications and originations as they are distributed in a bank’s reasonably expected market area (“REMA”) among tracts classified as “Majority-Minority” tracts and various minority submarkets (Majority-Minority tracts, Majority-Black tracts, Majority-Hispanic tracts, etc.). The parameters are the number of a bank’s applications and originations in minority tracts relative to a bank’s overall mortgage lending in the market. Sometimes this is referred to as the “penetration rate” in the minority tracts. That result is compared to other lenders’ penetration rates in the same market. Typically, regulators will compare a bank’s penetration rate to the penetration rates of a group of “peer” lenders, defined as any lender (including non-banks) that processed residential mortgage applications ranging from 50% to 200% of the mortgage applications processed by the subject bank.
If a bank’s minority tract penetration rate falls below the average minority tract penetration rate of its peers, how far below the market benchmark can it go before it is considered to be redlining? Statistical significance models measure that issue.
A fundamental assumption underlying statistical analysis as applied by regulators for redlining analysis is that “a mortgage is a mortgage is a mortgage”. For redlining analysis there is no distinction between or among mortgages. Essentially, it’s a simplistic numbers game that equates all mortgages as the same. The underlying weakness of this method is that it implicitly assumes and counts all mortgages as the same value for analytical purposes. In other words, a $100,000 mortgage for a single-family house in a minority census tract is given the same weight as a $100,000,000 mortgage that finances 500 housing units including 200 affordable housing units in the same tract.
In real life the distinction explained above happens. Right now, we are advising a bank in a situation like the above. The bank originated only 10 mortgages in majority-minority tracts within its REMA. The bank would need to extend about 30 or more mortgages in the MM Tracts to fall outside the statistically significant threshold. But because the bank originations fall short by about 20 mortgages in the Majority-Minority tracts examiners are considering referring the bank to the DOJ even though the bank financed hundreds of housing units more than the count of their mortgages in the MMCTs and even though hundreds of those units were affordable.
Adding to the irony is that the bank not only financed many more housing units than mortgage counts mandated by the statistical significance model it is by far the #1 lender in terms of the value of its mortgages in the Majority–Minority tracts compared to all lenders, not just the bank’s peers. Is that the picture of a lender that is redlining majority-minority neighborhoods? Yet the bank is being threatened with a referral to the DOJ. This makes no sense. In fact, redlining should not even be an issue, but in today’s zealous regulatory enforcement climate the bank is now in the formal process of responding to regulator questions.
To be sure, the Comptroller’s Handbook for Regulatory Compliance does instruct examiners to use “any other data useful for conducting a redlining and marketing analysis” but it would not be prudent to leave that task to examiners in today’s adversarial regulatory climate. The lesson to be learned is that bankers must anticipate how their mortgage volume in the minority tracts in their markets looks from a statistical significance point-of-view and to determine other useful information that may paint a more complete and accurate picture of their performance. In fact, that is the kind of risk management component that is essential to an effective fair lending risk management monitoring program.
