Last week, the annual Community Reinvestment Act & Fair Lending Colloquium took place in Austin, Texas. Two officials from the U.S. Department of Justice (DOJ) discussed in detail the “Combatting Redlining Initiative” led by the DOJ using a “whole of government” approach, the current state of redlining investigations, and the future direction of enforcement. In prepared remarks Assistant Attorney General Kristen Clarke stated, “we are proud of the work we have been able to accomplish in these past two years through the Combatting Redlining Initiative. But we are by no means done. We are also focusing on unlawful practices such as reverse redlining, and steering.”

Since its October 2021 launch, the DOJ’s initiative has resolved 10 redlining cases, securing over $107 million in relief for communities of color that have experienced lending discrimination. On a “Red Alert on Redlining” panel, Lucy Carlson, Deputy Chief of the Housing and Civil Enforcement Section, confirmed recent public statements by Attorney General Merrick Garland that approximately two dozen redlining investigations are underway. She further highlighted two common trends the DOJ is seeing in redlining enforcement actions: lenders’ awareness of redlining risks, sometimes for years, without taking corrective action; and evidence of discrimination in employee or manager emails, i.e., disparaging descriptions of certain neighborhoods or overt animus towards protected groups.

Assistant Attorney General Clarke offered guidance to lenders on how to avoid redlining practices, based on the DOJ’s past two years of experience:

• Engage with regulators and promptly implement any recommendations following a fair lending examination.
  • Lenders that ignored regulators’ recommendations that could have made a difference in identifying and addressing redlining risk are more often subject to investigation by the DOJ.
• Understand the full scope of the communities where the institution lends.
  • Some lenders could have proactively identified risk if they had stepped back to determine whether there were communities of color adjacent to their lending areas that they could have reasonably served.
• Engage with community groups which have insight about and experience with the credit needs of local communities of color.
  • Community groups are critical partners in connecting lenders to potential borrowers, as some of these groups provide services that are creating pathways to home ownership, such as credit counseling and matched savings programs.
• Ensure that the institution’s compliance management system is accurately measuring redlining risk.
  • The DOJ frequently sees institutions relying on insufficient metrics that give them a false sense of security.
• Information from community credit needs assessments and discussions with community organizations should be used to refine the institution’s internal fair lending monitoring program.
  • In February 2022, the DOJ joined the Consumer Financial Protection Bureau (CFPB), the Department of Housing and Urban Development, and prudential regulators in an interagency statement encouraging creditors to explore opportunities to develop special purpose credit programs (SPCPs) to better expand access to credit. SPCPs can be an effective way to help mitigate redlining risk.
  • Over the last 24 months, the DOJ has closed investigations without further enforcement action when it is clear that the lender promptly addressed redlining risk that it or its regulator identified — not as a reaction to an imminent referral to the DOJ — but because the institution was proactive in its efforts to do better.

Clarke also identified two other priority areas for the DOJ moving forward. The first is addressing appraisal discrimination. The DOJ is one of 13 federal agencies on the Interagency Task Force on Property Appraisal and Valuation Equity (PAVE) collaborating to develop a proactive, federal response to appraisal discrimination. In 2022, PAVE issued a report detailing specific agency commitments to address bias, including the DOJ’s commitment to expand fair lending enforcement and drive accountability in the industry. The second is combatting algorithmic bias and discrimination. “The use of algorithms to underwrite loans – from mortgages to student loans to credit cards — has created a new playing field for obtaining credit. Lending institutions are responsible for ensuring that the playing field is level for applicants regardless of their race, sex, familial status or national origin. Lenders should proactively and consistently review and test their underwriting process – including their automated steps — to ensure that credit decisions do not turbocharge discrimination by disproportionately rejecting loans to applicants of color for reasons unrelated to creditworthiness.” Clarke noted that monitoring the use of algorithms is a priority given President Biden’s recent Executive Order on artificial intelligence, discussed here, which assigns federal agencies the task of assessing the civil rights implications of artificial intelligence and other emerging technologies.

Our Take

As noted by Assistant Attorney General Clarke in her remarks, the DOJ and the other federal agencies are clearly moving “full speed ahead” in pursuing redlining investigations and enforcement under the Combatting Redlining Initiative, which remains one of the signature fair lending efforts of the Biden administration and the federal government. In preparation for future redlining examinations, lenders should ensure that their fair lending compliance risk management programs incorporate appropriate redlining analyses, including root cause analysis to identify which factors are driving risk and whether corrective action is needed.

Lenders should also note that the DOJ and federal agencies will continue to prioritize their efforts in addressing appraisal bias and algorithmic discrimination in underwriting, and take steps to mitigate those risks. The CFPB in particular has been highly focused on potential discrimination in credit underwriting models that use artificial intelligence or machine learning.