August 30, 2022 | Ali Abbas
Estimated reading time: 9 minutes
Compliance Confessionals: the enduring issue of fair lending failures
I am sure everyone in the risk world has read the recent consent order issued by the Consumer Financial Protection Bureau (CFPB) the Department of Justice (DOJ) against Trident Mortgage LLC (now defunct), a Berkshire Hathaway Holding. For those of you not familiar here is a brief recap.
“..From at least 2015 through 2019, Trident engaged in a pattern or practice of unlawful discrimination against applicants and prospective applicants, on the basis of race, color, or national origin, including by redlining majority-minority neighborhoods in the Philadelphia-Camden-Wilmington, PA-NJ-DE-MD Metropolitan Statistical Area (Philadelphia MSA), and engaging in acts and practices directed at prospective applicants that would discourage prospective applicants from applying for credit. For the purposes of this Complaint, “minority” refers to people who identified as Black, Hispanic, Asian, Native American, Native Hawaiian, or Pacific Islander in the U.S. Census…” (CFPB/DOJ Consent Order 27 Jul 22)
The consent order struck me. The specific areas identified as problematic pretty much covered every good compliance officer’s Fair Lending nightmare:
- Office locations were concentrated in majority-white neighborhoods
- Mortgage loan officers served majority-white neighborhoods, but not majority-minority neighborhoods
- Marketing targeted majority-white neighborhoods and avoided majority-minority neighborhoods
- Racist and/or discriminatory emails and photos were exchanged by lending staff
- There were disproportionately low numbers of home loan applications from majority-minority and high-minority neighborhoods
- Disproportionately low numbers of home loans were offered in majority-minority and high-minority neighborhoods
- There were intentional failures to improve lending to black and Hispanic applicants and prospective applicants in majority-minority communities
The consent order resulted in a $24.4 million fine. I won’t sensationalize it by leaving that big number out there. In reality the break down tells us a lot more. The civil monetary penalty was $4 million, $18.4 million for loan subsidy program, and $1.6 million set aside for various marketing/advertising, training, and community outreach programs. Trident closed its doors in April 2021 and Prosperity Mortgage (another Berkshire Hathaway Holding) took over its loans, employees, and the carrying out of the Order.
I am going to digress briefly to bring my old finance experience into my assessment of this BIG fine…
The last reported annual revenues I could find on Trident was $103 million. In essence, the fine was about 24% of that last reported revenue. Given the last few years of hay-day in mortgage lending, this was – in reality – a drop in the ocean. Their behavior spanned 5 years as cited by the CFPB (2015 – 2019). In addition, the reality is that the actual cash out the door, in the form of monetary penalty, was only $4 million so that’s about 1% of the last reported revenue. The $18.4 million put into a reserve for a loan subsidy program will still generate loans that go into Prosperity Mortgage portfolio. Yes, these will be at a higher cost and slightly lower than the prevailing interest rates (but in a rising rate environment). But the bottom line is that these are loans that should have been made anyway and will most likely still net a profit to Prosperity Mortgage.
The rest of the fine amounts to re-allocating dollars for specific efforts that will leave Prosperity Mortgage (from a reputation perspective) looking like they truly care about these under-represented communities. People and the media have short attention spans and over the course of time some will forget why Prosperity Mortgage is carrying this out — to fulfil Trident’s Consent Order. The end result is Berkshire Hathaway hasn’t lost much and should, in fact, come out much stronger in the fair lending arena and the public image if the contents of the Consent Order are truly integrated in all of its lending and real estate subsidiaries. At a minimum, fair lending issues should not be a problem again… or Warren Buffett didn’t have a serious enough conversation with the heads and board members of his many subsidiaries.
Now back to compliance — one thing I learned a long time ago is not to just read the summaries or business synopsis of Consent Orders or other published findings. As a compliance professional, I read every line so I can ensure that I understand the implications and mentally (if not actually) perform an assessment of the impacted area within my organization against the finding. This way I know if there is any risk that needs to be addressed.
The Consent Order against Trident Mortgage followed close on the heels of another recent HUD/DOJ settlement against Meta “Facebook” for FHA violations in the amount of $115,054 (the fine is limited by FHA rule cited penalties). The settlement was based on “…alleging that Meta’s housing advertising system discriminates against Facebook users based on their race, color, religion, sex, disability, familial status, and national origin…” (DOJ 21 Jun 22).
I have been reflecting on why it is that we are still having fair lending regulatory issues in the 21st century…
Organizations in the lending business – banks, mortgage lenders, FinTechs – are in the business of making profits. You would think that the “almighty dollar” would be the one driving force behind any lending program and far outweigh the supposed bygone era of discriminatory practices – apparently not.
Let’s, for now, put aside the individual behavior of lending officers/sales associates’ interactions with the consumers identified in Trident’s Consent Order. This comes as a result of poor hiring decisions, training, culture, and top-down leadership problems that the Human Resource and Executive Management in every company should be taking a serious look at if they actually want to compete in the lending business.
In the 21st century – the digital era – a consumer’s credit history, income, and ability to repay should be the driving focus behind loan programs and all of the underpinnings, such as marketing/advertising and credit decisioning.
These programs are supported by a lot of automated AI technology that is constantly evolving to help companies capture more market share. In both Trident’s and Meta’s case, the one overlapping issue was marketing and advertising efforts. In Trident’s case, it seems that it was overt and intentional while in Meta’s case it seemed unintentional. In either case the outcome still leads to discriminatory practices and “redlining”.
Both of these companies had bias built into their decision-making. Trident consciously determined that their efforts should be focused on “whites” based on the assumption that marketing to “whites” was the best way to get qualified candidates and good properties in their portfolio. Meta, on the other hand, used algorithms built by humans (who also have predisposed bias) marketed certain real estate to minorities while marketing other types of real estates to “whites” because of a bias (unconscious or not) around where minorities may want to live or price points they could afford.
Either way, I kept wanting to scream “where are their compliance staff during all of this decision making?!”
Over the years, the compliance teams focus on marketing and advertising has improved. However, truth be told, even in my last few years working in compliance with direct oversight of Fair Lending Analytics and routine interaction with the Compliance Fair Lending Leader, there was not dedicated fair lending resources involved in marketing/advertising programs. That is likely the case for most organizations who depend on the compliance generalist to cover all aspects of marketing/advertising compliance. Looking at every new campaign through the lens of fair lending is not always there. Secondly, having the time to understand complicated algorithms of 3rd party marketing vendors, as well as the potential risks associated with those algorithms, is not realistic given all of a compliance officer’s day to day responsibilities.
Most fair lending teams are well positioned to perform after-the-fact analysis on applications, approvals, and declines based on fair lending criteria. It’s easy to identify disparate impact this way. My analytical team was cracker jack at producing this information. We had routine schedules for each portfolio and the MSA peer analysis with the annual release of the HMDA data. Over time we started developing trending and baseline benchmark reports for MSA , based on where the business wanted to improve their statistics.
However, like most organizations, areas that needed more care included spending time with marketing/advertising and looking at branch closure (a topic for another day). Now, don’t get me wrong the Fair Lending Compliance team did a good job using the analytics and reviewing results with the risk and business partners. But with limited resource there was only so much time that could be spent. As is often the way, since we weren’t having any problems either demonstrated by the Fair Lending Analytics or identified by the regulators – everything was being covered using a risk-based approach.
Looking now into my rear-view mirror, I really do believe that marketing/advertising programs were taking leaps and bounds leveraging analytics and AI, while the compliance team was being left on the side lines observing.
If I could do it over, with proper resources, I would have staffed up the fair lending team (data scientists and Fair Lending Compliance Specialists) dedicated to understanding marketing and advertising programs, growing their knowledge as technology morphed.
Imagine building the understanding of how to analyze the potential impact of marketing/advertising decisions from a fair lending lens, layered on top of these sophisticated marketing AI technologies. It would benefit any organization from a risk management perspective. More importantly, the organization would reap the rewards by having a larger group of consumers to market lending products if we could proactively weed out these biases upfront, and the fair lending implications as a result.
For FinTechs, mortgage lenders, and other small bank or non-bank entities, don’t let the lesson of Trident escape you. The CFPB and other regulators are now making it clear that these entities will not continue to fly under the radar. Fair lending is not only a mortgage related topic, so non-mortgage lenders may want to start beefing up reviews of marketing/advertising programs, as well as the tools that third parties leverage. Like Meta, what may look like a targeted marketing effort can quickly become unintentional discrimination. Compliance Fair Lending teams need to expand and deepen their analysis and engagement in decisions around:
- Marketing/Advertising Campaigns
- New Product Offerings
- Physical brick and mortar office openings and closures
- Virtual sites leveraged in promotions
- Exceptions made for interest rates, credit scores, etc.
- Employee Fair Lending Training
In the 21st century technology can make work easier, but it may also open an entire new avenue of things that we must think about to manage risk. That is indeed the case with Fair Lending. For organizations that have embedded overt discrimination practices into their culture – whether intentionally or otherwise – they will soon be defunct. There is no longer any room in a global economy for this behavior.