Subconscious Bias: The Silent Exposure to Non-Compliance

Irresistible website designs and kickass marketing programs that deliver results

Learn More
hero-jobbies-7

Could your lending practices be biased, even if you take steps to ensure that your inputs and systems are not using any prohibited factors? 

The answer is yes. According to the Consumer Financial Protection Bureau and Regulation B, even if you completely avoid all prohibited factors, your lending practice could be non-compliant if its outcomes are inadvertently discriminatory under the “effects test” legal doctrine. Carrying a maximum penalty of $500,000 in punitive damages in addition to compensatory damages, it is critical that your lending practice be cognizant of bias on factors such as race, color, sex, religion, national origin, marital status, or age. 

Unfortunately, even the major credit reporting agencies can fall short of this standard. Due to legacies of discriminatory practices, such as redlining, certain demographics continue to face disproportionately low credit scores when calculated on decades-old data and criteria practices. Thankfully, Trust Science® can help: its Credit Bureau 2.0 ® uses proprietary AI-powered models that harness thousands of data sources to provide up-to-date, insightful and fully compliant risk scores capable of accurately assessing creditworthiness for everyone, even those left behind by the traditional bureaus.

Legal Background

One of the many statutes governing lending practices, the Equal Credit Opportunity Act (ECOA) ensures that creditors do not discriminate against applicants on the basis of a number of prohibited factors, including race. Supported by Regulation B, enforced by the Consumer Financial Protection Bureau (CFPB), creditors must ensure that they are not biased on these prohibited factors to remain compliant, facing substantial financial penalties for violations. Lenders, with extremely limited exceptions, are fully cognizant of and take actions to eliminate consciously discriminatory practices by excluding factors from which bias can be construed. Where some may fall short is a far more difficult to identify form of discrimination: namely, subconscious biases which produce disparate results.

Under § 1002.6(a) of Regulation B, the CFPB states that the standard for discriminatory non-compliance is the “effects test”, originally intended for the employment field and later applied to the credit industry by Congressional intent. Under the regulation, even if a creditor has no intent to discriminate and has neutral-appearing practices, that creditor may still be considered discriminatory if that practice produces results with a disproportionately negative impact on a prohibited basis. Furthermore, if there is evidence that a discriminatory action was, consciously or inadvertently, regular practice, then the Department of Justice may file a lawsuit under ECOA for a pattern or practice of discrimination, causing even well-meaning lenders to face crippling legal and financial sanctions.

Why Traditional Credit Fails

Subconscious lending biases, especially in a virtual setting of increased applicant anonymity, are often a product of poor data or biased criteria which inadvertently exclude certain demographics. For marginalized demographics, conventional credit factors often seem impossible to achieve. The on-ramp to the credit market is broken and caught in a catch-22: in order to build credit, you need to have been approved for a credit product, and in order to be approved for a credit product, you need to have built credit. Due to past discrimination, many marginalized demographics were unable to develop the financial and credit history that others may take for granted, leaving them vulnerable and outside of the credit market. With the market’s reliance on old-school scoring mechanisms, many of these people are scored disproportionately low or cannot be scored at all, leaving a vast wake of credit invisibles that are a product of subconscious lending biases and a remarkable market opportunity for lenders, if they are able to find a reliable and unbiased way of assessment.

The traditional credit scoring models are based on limited and overlapping criteria, almost all centering around your credit and repayment history. As a result, the credit score systematically excludes those without a credit background. In addition to the problems with creating the credit catch-22 and excluding roughly one in four American households aside, racialized (especially Black) communities are disproportionately overrepresented in the financially underserved market segment, resulting in a subconscious negative bias that subjects millions of potentially creditworthy individuals to inaccurately low credit scores. 

Of the many factors which contribute to this inequality of outcome, home ownership and mortgages, with their substantial weighting in traditional credit scoring algorithms, are particularly impactful. Despite overt redlining having been prohibited since 1968, a 2018 study by the National Community Reinvestment Coalition found that, as a result of the prior inability to secure mortgages, BIPOC Americans lack the robust mortgage and credit history to develop strong credit scores, instead resorting to rental payments which fall outside of the scope of consideration for traditional credit scoring agencies. Essentially, the traditional credit scoring methodologies use an excessively limited view of the personal financial picture and are unable to process various financial transactions and factors that could break down the subconsciously exclusionary boundaries of traditional credit scoring.

The Trust Science® Solution - Credit Bureau 2.0 ®

While conventional approaches rely on a limited number of restrictive data points, Trust Science® leverages tens of thousands of data sources using its proprietary AI and machine learning technology to integrate seamlessly with a variety of loan origination and management systems and provide lenders with a highly predictive, fully explainable and fully compliant Six°Score™. Trust Science® uses information that will include as many people as possible, including those financially marginalized for any reason. By using consented banking, rental, telephone, and utility billing data, in addition to data used in present credit scoring, Trust Science® helps deserving people get the loans they deserve, breaking free of the constraints of old data to eliminate much of the traditional subconscious lending biases that may have left lenders vulnerable to costly litigation while bringing substantial returns on investment.
Trust Science® also takes part in and consistently passes regular and voluntary compliance checks which screen extensively for biased factors and results. This rigorous commitment to eliminating bias ensures that Trust Science® scores and practices are fully compliant with even the most stringent regulations. For Trust Science® clients, this means being able to assess credit invisibles with confidence, leading to increased quality loan originations, improved rates of default and delinquency, and total compliance in your regulating environment. Trust Science® is offering lending leaders a compliant, fair and ethical way to score underbanked and financially stressed applicants, right now. Learn more about what Trust Science® can do to help your lending company be at its most compliant and most effective by visiting our website, or contact us to schedule a demo.

Contact Us