By Stephanie Kalahurka and Mat Petersen
On
March 30, 2023, the Consumer Financial Protection Bureau ("CFPB") issued its
final rule ("Final Rule") implementing the changes to the Equal Credit
Opportunity Act ("ECOA") made by Section 1071 of the Dodd-Frank Act. The Final Rule requires financial
institutions to collect and report to the CFPB data on applications for credit
from "small businesses," which includes any business that had $5.0 million or
less in gross annual revenue for its preceding fiscal year. When implemented, the Final Rule will require
banks and other lenders to collect 20 separate data points on each covered
small business loan application, including data regarding the minority-owned
status of the business.
Much
of the initial attention given to the Final Rule has focused on implementation,
and concern with the administrative burden of data collection and reporting
requirements. While those aspects will
be time-consuming and onerous, financial institutions are accustomed to shouldering
new regulatory-mandated workloads (e.g., the 2013 "ability to repay" mortgage
rules and TILA-RESPA Integrated Disclosure rules). Rather than being overly concerned about
short-term implementation, community banks should be far more
concerned about how the Final Rule's reported data points will be used
in future fair lending examinations. If
the data is used in the way that the CFPB has stated that it will be—to
identify unexplained discrepancies and to facilitate fair lending enforcement—the
Final Rule has potentially devastating implications for the relationship small
business lending models employed by most community banks. Under the bank regulators' existing
enforcement methodology, the Final Rule is likely to remove almost all lender
flexibility from small business loan pricing and underwriting, effectively
eliminating the strategic advantage of community banks to customize loans to
the individual need of a particular small business.
Understanding the
Regulatory Process for Fair Lending Examination and Enforcement.
The full implications of the Final Rule must be
considered in the context of the current, and largely statistical, regulatory
process for fair lending examination and enforcement. Fair lending exams are rigorous, and the regulatory
process for proving discrimination under fair lending examination standards are
largely unpredictable. Within the
current framework, a pattern or practice of discrimination does not
require any evidence of discriminatory intent. While discrimination may be proven with overt
evidence, more often, violations are supported by statistical evidence of
"disparate treatment" or "disparate impact."
Data collected under the Final Rule will be used primarily to support
findings of ECOA violations based upon evidence of "disparate treatment."
Disparate
treatment occurs when a lender treats a credit applicant differently because of
a prohibited basis, which generally include the applicant's: (i) race or color;
(ii) religion; (iii) national origin; (iv) sex; (v) marital status; (vi) age
(provided the applicant has the capacity to contract); (vii) receipt of income
derived from any public assistance program; or (viii) exercise, in good faith,
of any right under the Consumer Credit Protection Act. Importantly, a violation based upon evidence
of disparate treatment does not require any showing that the treatment was
motivated by prejudice or a conscious intention to discriminate against a
person beyond the difference in treatment itself. In other words, the currently regulatory
standard requires only a showing of difference in treatment, and the regulatory
agency may then presume that difference is due to protected class
status, resulting in a fair lending violation, unless the lender can prove
otherwise.
As
an example, regulators often employ a statistical regression or similar
analysis when testing for disparate treatment as evidence of pricing or
underwriting discrimination in consumer lending. As part of that process, examiners will typically
identify a sample (sometimes, a small sample) of a specific type of loan, such as
mortgage loans or unsecured consumer loans.
Examiners gather data relating the pricing or underwriting of those
loans and information on the associated applicants from a combination of file
review and interviews with bank officers.
This data is sent to the regulators' statisticians, who perform a
complex regression analysis to determine whether there are statistically
significant differences in rates charged, or approvals/denials with respect to
protected class versus non-protected class borrowers. If statistically significant differences are
identified, then based solely upon the statistical modeling, the
regulators may conclude that an ECOA violation has occurred. This process of collection and analysis takes
time and may cause a compliance examination to be held open for a year or more. When the analysis is complete, and if a
statistical disparity has been detected, the institution is typically given a
mere fifteen days to respond to the preliminary findings. If the bank is unable to disprove the purported
violation, the regulators are permitted under applicable examination procedures
to presume that the disparity is due to applicants' protected class status and
proceed with enforcement.
Applying
current regulatory examination procedures, the standard puts a significant
burden on community banks to prove the absence of discrimination, where the
regulators benefit from statistical techniques that are subject to sample size
and omitted variable bias and the right of examiners to second guess an
institution's policy with 20/20 hindsight.
Historically, application of this type of statistically driven
enforcement has been limited to consumer loans, which are fairly uniform with
respect to both credit terms and underwriting criteria; and mortgage loans, for
which uniformly reported data (e.g., HMDA data), including the applicant's
minority status, are available.
Importantly, this statistical approach has not previously focused on
small business loans, largely because the testing data was not as uniform or
readily available. With the CFPB's
issuance of the Final Rule, that is all about to change.
Implications for
the Final Rule under Statistical Fair Lending Enforcement Processes.
After
the Final Rule has been implemented, regulatory agencies will have quantitative
data available to: (i) identify the protected class status of small business
loan applicants; and (ii) determine whether there are unexplained
disparities in the underwriting or pricing of small business loans made to
similarly situated borrowers. Under
currently regulatory guidelines, the determination of whether borrowers are
"similarly situated" may, in the discretion of the regulator, depend solely
on reported data points.
The
regulators' data-based approach to fair lending testing has reshaped how banks are
required to price and underwrite consumer loans, and banks have adapted. It will not be so easy to adapt in the
commercial loan context. Due to the larger and more complex nature of
commercial loans, a statistical approach to fair lending enforcement will not
be consistent with prudent credit practices, will not benefit small business
applicants and will ultimately strip community banks of their primary strategic
advantage. Community banks' primary
strategic advantage in small business lending has been their ability to lend
based upon intangible, "relationship" factors, for the benefit
of their small business customers. Larger
banks tend to have strict, uniform underwriting criteria and more centralized decision-making
that can leave small businesses without access to credit. Community banks have been able to fill the
gap in small business lending by taking a more qualitative approach that recognizes
the individualized characteristics of a particular business that makes it
creditworthy.
The
regulators' statistically-driven fair lending enforcement methodology has
driven most banks to utilize stringent underwriting metrics and pricing
guidelines for consumer loans. Risk of
unexplained discrepancies have eliminated almost all loan officer discretion
from the consumer lending process. To
mitigate fair lending risk, banks of all sizes now employ consistent
underwriting and pricing criteria for consumer loan products. Those criteria are typically uniform and
quantifiable, including factors such as (among others) the size of the loan,
the age and type of collateral, the borrower's income, credit score, and the
amount of the borrower's deposit relationship.
For consumer loans, the risk-mitigating benefits of these types of
quantitative controls can be achieved without materially hampering a bank's
ability to serve its customers or to underwrite credit in a safe and sound
manner.
The
same is not true for commercial lending.
The inherent complexity and higher level of credit risk associated with
commercial loan products mandates a dramatically different approach to underwriting
and pricing. Commercial loans are
typically larger, more complex, and accordingly, pose a higher risk to the
lender than consumer loans. As such, a
safe and sound approach to underwriting small business loans requires
consideration of, and the interplay between, numerous and varied
non-discriminatory factors that are not uniformly present in consumer loan
applications. Those factors also extend well
beyond the 20 data points to be collected under the Final Rule.
For
small business loans, lenders cannot simply "run the numbers." The loan officer must understand the
business. To that end, he or she
must consider all relevant information contained in the borrower's loan
application, in the customer's file or on the bank's data system if the
borrower is an existing customer, and information from third parties and other
data sources. This information generally
includes factors that can be quantified, such as debt to income and loan to
value. This information may also include
business revenues, type of credit, time in business and other of the data
points to be collected under the Final Rule.
Importantly, however, this information must also include factors
that are not quantifiable, such as the borrower's business acumen and
management experience, market conditions, industry-specific risks and the
borrower's reputation among customers and suppliers.
For
commercial loans, the bank's own experience may even factor in. For example, heightened underwriting
standards might be appropriate for a small business loan where the
institution's credit personnel have less experience and knowledge of the
industry, or where they have less practical ability to monitor the ongoing
financial health or business prospects.
For small business loans, it would not be practical, or even possible,
to translate all these legitimate, non-discriminatory and necessary
considerations into quantitative policy matrices or limits. Ultimately, the exceptions would negate the
policy, or alternatively, the inflexibility of the policy would undermine the
bank's ability to effectively serve its customers.
Notably,
the bank regulatory agencies acknowledged that community banks' more flexible,
and non-standardized approach to underwriting small business loans benefits
small businesses. The following is from
the FDIC's Small Business Survey:
"An important
potential difference between small and large banks stemming from their distinct
approaches is the way they manage loan requests from small businesses, and
specifically whether they apply preset criteria. Two common preset criteria are a required
minimum loan amount and the use of standardized loan products. Small banks' use of nonstandard
information may lead to greater flexibility and willingness to customize loans
according to the individual needs of a small business… The advantages to small
businesses from using standardized loan products are that the underwriting can
be quicker to process and the products can offer more competitive interest
rates. The disadvantages are that only a
subset of small businesses may be able to satisfy the standard criteria
required to qualify for them, and the small businesses seeking the loans may
prefer loan terms different from the ones offered…
The
survey supports the understanding that small banks are relationship lenders and
approach small business lending in a more flexible and customized, case-by-case
way compared with large banks; and as a result of this approach may be that
less established firms are more likely to receive credit. Small banks are found less likely than large
banks to use minimum loan amounts on their top products or to rely on
standardized loan products. And small
banks are more likely than large banks to accept real estate collateral, a
practice that is consistent with small banks' having a more intimate knowledge
of their local communities. Further,
small banks … often evaluate a wide set of additional information, including
relationship-based soft information such as owner's experience or the
management team's skills."
Community
banks engage in traditional relationship lending that is based, in many cases,
upon numerous, non-discriminatory, but often intangible factors. They can do this because they have personal
knowledge of both their customers and their community. These factors cannot accurately be captured
in any number of data points under the Final Rule. They cannot be quantified or controlled for
in the statistical models currently used for fair lending testing. Accordingly, after implementation of the
Final Rule, community banks that continue to act as relationship lenders to
their small business customers will have significantly heighted levels of fair
lending risk to due to the inevitably of having to answer for "unexplained
disparities" in their small business loan data.
Conclusion.
From
a legislative and regulatory policy standpoint, we should remember and
continually emphasize that bankers are in the business of managing risks. As part of that process, bankers may decide
that they simply cannot manage the heightened regulatory risks posed by the current,
purely statistical, "no-intent-required" approach to fair lending
enforcement. Rather than tiptoe this
perilous regulatory tightrope, many bankers have simply decided to cease
offering certain types of loans at all.
Smaller unsecured consumer and mortgage loans, which are more often
subject to heightened fair lending scrutiny, have been widely de-emphasized or even
abandoned by many community banks in favor of larger, commercial loan products. As regulated financial institutions exit this
space, we have seen non-bank and payday lenders step in, eager to satisfy the
resulting demand, but at higher rates and without the same level of
oversight.
The CFPB has openly stated that data collected under the Final Rule is intended to be used for fair lending enforcement in small business lending. For the reasons discussed in this article, the more flexible, qualitative small business lending model used by most community banks, which ultimately benefits many small business borrowers, will not fare well under this type of statistical examination and enforcement. At the end of the day, the current statistical approach to fair lending enforcement will likely have the unintended consequence of "protecting" many small businesses out of broad access to lower-cost, more individualized commercial credit. For these reasons, community bankers should continue to support their associations' and others' efforts to oppose or seek modification of the Final Rule. Lawmakers and the federal banking agencies should consider the consequences of the Final Rule, and at a minimum, adjust examination and enforcement procedures to accommodate and permit the continued tradition of relationship lending by smaller community banks.
Originally posted May 4, 2023