Fair lending keeps morphing

When Jim Bedsole compares fair-lending laws and the Community Reinvestment Act to Dr. Seuss’ “Thing 1” and “Thing 2,” it’s not because the veteran compliance officer wants to lighten the mood. He’s trying to portray the confusing and uncomfortable—and at times surreal—position that banks increasingly appear to find themselves in when examined for both types of compliance.

In Seuss’ The Cat In The Hat, the two Things are impish twin creatures in red jumpsuits who create pandemonium in the home of two children, in the course of helping the uninvited Cat to “entertain” the kids. In compliance, the two bodies of law, although connected, are supposed to be distinct. Simultaneously, the enforcement of both has grown further apart and yet closer.

In some ways, Bedsole and other speakers at a panel on the interplay of the two laws at the recent American Bankers Association Regulatory Compliance Conference portrayed a situation where these two regulatory “things” wreak their own havoc.

Bedsole, senior vice-president and chief compliance and risk officer at BankSouth, pointed out that CRA’s intent is to be sure banks are meeting the credit needs of their assessment areas. And most institutions obtain either a “Satisfactory” or “Outstanding” rating. Continuing, he pointed out that fair-lending performance is one of the components of a CRA exam. And there is evidence that fair-lending concerns can lower preliminary CRA ratings.

Something odd going on?

Yet in the six most-recent redlining cases announced by the federal government at the time of the June conference, each bank had obtained a “Satisfactory” rating, pointed out fellow panelist Brian McCormally, partner specializing in fair-lending cases at Arnold & Porter LLP. McCormally said that all federal regulatory definitions of redlining define the term as a form of disparate treatment, implying discriminatory behavior, rather than a form of disparate impact, where the end result of behavior is perceived as discriminatory even if there was no overt act of discrimination.

The situation grows stranger. Increasingly, purchased loans do not bring a bank any credit toward fair-lending performance, leaving only loans originated and held by that institution as qualifying, for example. This is so, McCormally said, even though for some organizations loan purchase is the only effective means of penetrating a market in order to bring credit to a neighborhood.

McCormally reviewed several recent settlements, and pointed out where things didn’t appear to line up. For example, he represented $35 billion-assets Hudson City Savings Bank, which settled with the Department of Justice and the Consumer Financial Protection Bureau for $32.8 million in September 2015.

The attorney, who held senior legal and compliance posts at banking agencies prior to entering private practice, said the government presented no evidence of actual disparate treatment. Instead, the government case hinged on such factors as alleged redlining through loan officer placement and branching decisions—McCormally says his client actually maintained 14 branches in majority minority neighborhoods but received no recognition of that.

However, the core of the case rested on statistical analysis of the bank’s application data in affected areas compared to its peers.

“We no longer rely on traditional treatment analysis,” said McCormally. “The agencies no longer feel the need to show that someone was done actual harm.” He maintained that now simply being below the peer average in a minority census tract creates risk that the bank will be accused of redlining.

How does CRA fit today?

“I’m currently dealing with a bank that has an ‘Outstanding’ CRA rating,” said McCormally, “where the government says there may be a redlining issue.”

McCormally added that in addition to CRA becoming irrelevant in recent fair-lending discussions, the bureau, DOJ, and HUD have at times forced changes to banks’ ostensibly self-established assessment areas. In addition, the government has moved from a focus on low-to-moderate income census tracts to a newer focus on “high-black and Hispanic” census tracts, which he defined at containing 80% or more of those minorities.

Actual discrimination has always been difficult to prove, McCormally said, so he thinks the government has moved from a focus on disparate treatment and disparate impact towards more of a focus on redlining as a policy tool to address a perceived shortfall in credit availability. He says that of 25 open fair lending investigations at yearend 2014, only a third have been resolved. The remainder are somewhere in the legal pipeline, and many of those cases involve redlining.

“The sands have shifted, and the government is evaluating these cases from a different perspective,” said McCormally. He says he has repeatedly asked about the apparent shift and been told nothing has changed. Given his experience in the field, he does not believe this, he said.

“It’s not to say that the government is wrong,” said McCormally. “It’s to say that we have to understand how it has shifted, why it has shifted, and to be able to be prepared to address the issues going forward.”

Ultimately, without more clarity, banks have less than ever to go on.

Furthermore, speaking with what appeared to be irony, McCormally said he didn’t want to “accuse the government of backing into a conclusion.” However, he said that the issue of what will constitute a bank’s peers for the sake of fair-lending analysis has come to be the central negotiating point of every fair-lending case in which he participates.

[After the conference, in late June, a $10.6 million settlement involving $13.9 billion-assets BancorpSouth Bank was announced. The bank was accused of redlining, and two years earlier had received a Satisfactory rating. Among the additional allegations in the case was that the bank maintained “an explicitly discriminatory denial policy” and that it engaged in both discriminatory underwriting and discriminatory pricing. A recording of a 2012 internal meeting at the bank demonstrated discriminatory attitude, according to the CFPB, and “clearly articulates the Bank’s policy or practice to reject minority applicants more quickly than White applicants, as well the Bank’s perception of African Americans. The meeting participants were all White …” As recounted in the complaint, the recorded conversation included a racial epithet. CFPB also conducted a matched-pair test that put the bank in a bad light. This was the bureau’s first use of the testing technique in a discrimination investigation.]

By the numbers, but which ones?

 “We’re in an environment where there don’t appear to be any written rules,” observed McCormally. Or, if there are written rules, the government is changing when it wants to apply them and when it doesn’t.

“It is increasingly difficult for institutions to know how to address this,” McCormally added.

But if there is a ground zero for confusion, it would appear to be in the analysis that banks voluntarily do from their Home Mortgage Disclosure Act data versus the analysis that the government does in screening institutions for potential investigation and then for potential enforcement or settlement.

“Flabbergasted” is one word that speaker Marsha Courchane used when describing her reactions to the statistical cases that federal agencies has been bringing against lenders in recent enforcement actions. Courchane is vice-president and practice leader, financial economics, for Charles River Associates, which assists banks with such issues as fair-lending data analysis.

A key element in federal cases now is how an institution’s lending record compares to peers. Courchane explained that while firms like hers could assist a bank in performing peer analysis of its record, there is no assurance that the peer organizations selected by the bank will be the same group that the government selects for its case.

In the post-conference BancorpSouth Bank complaint, the government frequently compares the bank’s behavior unfavorably to that of peer institutions. For example, the complaint states that the bank drew only 3.2% of its applications from high-minority neighborhoods from 2011-2013, versus 17.6% by peers. No mention is made in the complaint nor in any associated maps or links regarding who the peer group consisted of. The complaint also notes that the bank’s own internal analysis indicated that black applicants were denied for first-mortgages in 2011-2013 at a rate 9.5 percentage points higher than whites. (At the time of the settlement the bank stated that it had been working on improvements to its fair-lending and other compliance systems since 2012, and that it had settled to conserve resources.)

As integral as statistical analysis has become in fair lending, Courchane portrayed the landscape as fraught with uncertainty.

“I can’t predict who they are going to compare you to,” said Courchane. “And you probably can’t either. So don’t waste time doing that. Figure out who you consider your peers to be.” She recommended writing up the analysis into a document that can be presented to examiners and investigators defending the bank’s fair-lending record.

Peer selection may vary based on whether an analysis is being performed for CRA or fair-lending review. Courchane said it is unclear what the CFPB’s standard is for selecting peers for redlining analysis—it may vary from case to case, apropos of McCormally’s remark. The Federal Reserve’s rule used to be pretty clear to people in her business, she said.

Picking up on McCormally’s point about assessment areas, Courchane warned that federal authorities increasingly seem to resist banks’ desire to set up assessment areas based on portions of counties or Metropolitan Statistical Areas. They tend to prefer whole counties, for example. This may not correlate well with a bank’s historical markets.

Post-recession, a wrinkle is seen in the authorities’ concept that originations in a majority minority market represent an equal penetration of that area compared to the penetration by the bank of predominately white markets. As lenders have tightened up in recent years on credit standards, this has become harder for lenders to do, and therefore more of them could potentially be accused of redlining, as the term is currently used. Courchane said that none of the recent investigations take differences in credit quality in different tracts into account. This can leave lenders exposed.

Courchane said banks could begin asking regulators for some clarity on peer selection. If enough bankers pushed the issue, some guidance might come back, but Courchane was skeptical. She said she didn’t think it was clear that regulators had firm answers to give.

Thus, when an audience member asked what pre-packaged software her bank might purchase to help perform analysis, Courchane doubted that much automated help was out there that was equal to the task.

“I don’t know how you would select canned software,” said Courchane, “because the rules are different in every case.”

Dealing with the “things”

Looking at data, banker Jim Bedsole indicated that lenders must work hard on HMDA number integrity. He said this was critical so banks could “say with confidence that their data is accurate.

“If you can count on your data, then the only bad thing that can happen to you is having someone reach bad conclusions from your data,” said Bedsole. More proactively, he said, once a bank knows that its data is sound, it can tap outside expertise to demonstrate that it has a good story to tell on access to credit.

“We’ve got to understand our own data as well as or better than our examiners,” said Bedsole. “We have to understand what the data means, and then how to tell that story.”

Given that a disconnect appears to exist between fair lending and CRA, Bedsole suggested steps that banks can take to reconnect the related compliance tasks internally:

• Bring staff responsibilities closer. Often fair-lending and CRA have been seen as separate roles with differing focuses and priorities. Bedsole suggested building stronger links.

• Bring in Marketing. Even today, Marketing and Compliance can have a variable relationship. However, Bedsole sees Marketing’s understanding of demographics and its experience in crafting advertising and other promotional campaigns as critical to success in CRA.

“CRA, at its core, is a marketing regulation,” said Bedsole.

He believes Marketing can be of assistance, equally, in helping a bank achieve the market penetration necessary to demonstrating fair-lending compliance.

• Make sure systems can talk to each other. Bedsole said that in some institutions the tracking systems for the two areas may not even be compatible. If the bank’s own analysis is to be performed consistently, he said, this is essential.

• Create a linkage in oversight. At his own bank, Bedsole said, board-level review of CRA and fair-lending issues and performance have historically not been conducted at the same time. He said he has been working to synch the reporting and oversight regimes for the two disciplines.

For community banks, the fair-lending challenge, like many others, comes down to resources, according to Bedsole.

“How many of you have all the resources you want, all the resources you have to have to earn a ‘Satisfactory’?” asked Bedsole. “None of us do.”

The solution lies in using risk assessment techniques to determine where to apply scarce resources, Bedsole explained. From there, a community bank can tap external resources for many tasks, saving internal resources for what can’t be obtained through public sources, trade associations, networking, and even the local university.