Throughout the history of consumer lending of all types, a major component of credit underwriting has been credit history. National credit bureaus use a person’s financial history, such as on-time payments, capacity of credit used and length of credit history, to create what is called a credit score. The credit score is intended to predict the likelihood that a borrower will repay a loan. The vast majority of consumer loans in the U.S. are underwritten using the FICO score, a proprietary score developed by FICO.
However, imagine a market where credit scores are unavailable or where some argue traditional credit scoring models are outdated. Imagine you wanted to make consumer loans to individuals in developing countries who do not have traditional credit records. Or perhaps you want to make loans in the U.S. to consumers who have poor credit scores due to previous market distortions, which resulted in numerous poor loan product matches with borrowers. Yet given improvements in the economy and new realities in financial markets, many of these people actually are now likely to repay appropriately matched loan products.
To determine creditworthiness, you would need to rely on factors besides a traditional credit score or credit history. You might instead seek to learn about the personal characteristics of your potential borrower and use their habits, personality or alternative aspects of their credit history (when available) to determine whether they are likely to repay a loan. Or perhaps you might focus on a combination of factors that relies on the current picture of a borrower and not prior behavior that in many cases, even if available, may not be an accurate predictor of repayment now.
In fact, this is exactly what has begun happening around the world. As a substitute for traditional credit history and scoring, lenders have been underwriting loans based on a borrower’s personal characteristics, like honesty, ethics, drive, motivation, optimism, intelligence and business skills. This type of lending, termed psychometric lending, currently plays a major role in lending in China, Peru, India, South Africa and other countries in Asia, Africa and Latin America, and has recently spread to the U.S. as lenders seek alternative credit scoring for consumers who once had a solid repayment history for smaller balance loans but who may have defaulted on improperly matched larger balance loans or loans with cost increases during the repayment period that exceeded borrower income growth. Silicon Valley startups have begun using psychometric values to assess creditworthiness of tens of thousands of U.S. consumers who don’t have traditional credit histories, and some larger companies have plans in the works for credit scoring based on historical retail transaction data.
Psychometric data about borrowers in developing countries was initially collected through questionnaires, but now data is increasingly being gathered from borrowers’ smartphones. Borrowers give lenders access to a smartphone’s stored data by approving and installing a data collection app, and then predictive algorithms go to work analyzing the collected data to find subtle behavior patterns that can be used to assess creditworthiness. For example, those who add last names to their contacts or those who are known gamblers (those who receive texts from a gambling company) are likely to be more creditworthy. On the other hand, those who drain their battery quickly or travel infrequently are likely to be less creditworthy. Often algorithms analyze 10,000 so-called signals per customer, resulting in a more accurate creditworthiness determination than is provided by a traditional credit score.
The advantages of psychometric lending are significant: it opens previously untapped markets to lenders and drastically reduces underwriting costs by allowing lenders to use inexpensive technology to evaluate borrowers. Additionally, as predictive algorithms improve, this type of underwriting is likely to get more and more accurate.
The Government Takes Note
In the U.S., although not necessarily psychometric data focused, signs of a movement away from traditional credit scoring models are emerging. For example, in December 2015, Rep. Ed Royce (R-CA) introduced H.R. 4211, the Credit Score Competition Act of 2015, cosponsored by Reps. Terri Sewell (D-AL), James Himes (D-CT), Krysten Sinema (D-AZ), Robert Pittenger (R-NC) and Gregory Meeks (D-NY). The bill would allow Freddie Mac & Fannie Mae to use other credit scoring models besides the FICO score. The legislation is intended in part to address problems faced by would-be homebuyers who do not meet traditional FICO score standards but otherwise have a high likelihood of repayment.
Other policymakers in Washington, D.C. have also exhibited an interest in other means of assessing a borrower’s likelihood of repaying a loan. For example, the Consumer Financial Protection Bureau’s (CFPB) 2013 Ability to Repay Rule established a framework to mitigate the harmful effects of market distortions by requiring creditors to use specific minimum standards during underwriting in the hopes of restricting the availability of certain mortgage loan products to borrowers who cannot repay them. Conspicuously absent from the Rule, however, were credit scores as a mandatory standard for determining a borrower’s ability to repay. In fact, the Rule specifically states in official commentary that creditors are not required to consider a consolidated credit score or prescribe a minimum credit score that must be applied. The commentary further states that creditors may look to nontraditional credit references such as rental payment history or public utility payments. In the issuing release of the Rule, the CFPB also indicated an interest in and the possibility of further research into using a borrower’s residual income as a reliable predictor of repayment.
Regardless of whether psychometric data is used or some other real-time metric similar to residual income at the time the loan is made, there are compliance and privacy issues to be considered.
Charter, usury and secondary market financing
Many FinTech entrants thus far (and probably many new entrants on the psychometric lending scene) are not banks. These nondepositories, however, are often subject to the laws of the states they would like to operate in, such as state licensing, usury, advertising and disclosure laws. To start generating revenues as quickly as possible, state licensed entities can follow the tried and true method other nondepositories have used over the years by obtaining licenses and opening for business first in the states with larger economies and more target borrowers such as California, Texas and Florida.
Banks, however, are generally not subject to state usury and lender licensing laws and therefore have a leg up in some ways on nondepositories hoping to enter the space. For a variety of reasons, however, banks may be less willing to enter this business from scratch. Banks on the other hand may still be able to participate in the burgeoning space by engaging with nondepository psychometric lenders to offer loans in ways that would not subject the psychometric lender to state usury and licensing laws while allowing the bank to share in new opportunities and meet prudential mandates to offer credit to underserved markets.
These arrangements often entail the bank acting as the true lender and selling the loans soon after closing to minimize bank portfolio risk. There are a variety of ways this can be accomplished. In either situation, whether operating through a state license or bank charter, the loans are often pooled and securitized into various loan balance and credit risk tranches. These securities can be appealing to investors because they potentially offer higher returns with shorter durations and more accurate real time risk assessments of the borrowers of the collateral in the pool. The securities can be an exempt offering to certain accredited and institutional investors or registered with the SEC and sold to a wider array of buyers.
Regardless of charter approach and strategy to finance and sell the loans, regulators are quickly moving to catch up or take action, which could have significant consequences. Many of the prudential regulators such as the Office of the Comptroller of the Currency (OCC) and Federal Reserve Board, state regulators in California and New York and the SEC have discussed plans to focus on the FinTech space in 2016, including the new psychometric lenders. The California Department of Business Oversight has even sent requests for information to FinTech companies operating in the state. In other cases, regulators are actively engaged in rulemaking for existing businesses that could have consequences for FinTech lenders. The CFPB, for example, will be issuing new payday loan and debt collection rules that may impact all short-term, small dollar closed loan markets in the U.S.
Another area of great concern for psychometric lenders is consumer privacy. The behavioral data collected by psychometric lenders that is the backbone of the business model would likely be a rich trove of non-public personal information for hackers and other wrongdoers. Lenders that gather such information without authorization or inappropriately share or safeguard information may violate Gramm-Leach-Bliley requirements and also engage in unfair, deceptive or abusive acts and practices (UDAAP).
Marketing and UDAAP
Psychometric lenders must also be clear in how they offer credit. Depending on the methods psychometric lenders collect and use consumer data to make a credit decision, they could be engaging in unfair, deceptive or abusive acts and practices that result in significant consumer harms. Psychometric lenders should inform consumers how their data will be collected, compiled, analyzed and used. In many cases, data will be collected via the borrower’s smartphone based on use of various apps, activities and geographic locations. Consumers should reasonably understand the credit product, what information will be collected and how their information will be used. Psychometric lenders should not sell data to others without receiving authorization from the borrower and ensuring that the buyer appropriately safeguards and does not share the information without proper authorization. Sales of data could also result in the lender being deemed a credit reporting agency under the Fair Credit Reporting Act, which brings additional regulatory and data integrity requirements.
How psychometric lenders use collected data could also create fair lending issues if use of the data results in credit discrimination against protected classes. The Equal Credit Opportunity Act, for example, prohibits discrimination in the extension of credit to otherwise creditworthy persons based on age, sex, marital status, race, country of origin, color, religion or the fact that all or part of the applicant’s income derives from a public assistance program. Unlike many traditional underwriting factors that have been used for years and even decades, many psychometric factors have likely not been tested from a fair lending perspective. Factors relied upon for underwriting should be monitored to ensure that there are not significant correlations between their use and disparate impacts such as higher pricing and credit denials involving protected classes.
Psychometric lending offers exciting opportunities to reach more consumers, provide them with better suited products and improve credit risk. A potential winner for individuals and the broader economy. However, even though real-time data gathering and predictive analytics are continually being applied in new ways, many of the same old rules and regulations applicable to traditional consumer lending and the structured financing of these activities are still in place. For sure, some of these rules may be outdated, but regulators will likely not sit back and watch a new market form that could be subject to abuse and distortions that have harmed and even tainted prior innovative cycles (think “.com” and subprime mortgages). If used properly, these new underwriting methods could actually result in real-life improvements for millions of consumers.
Alston & Bird is uniquely situated to assist psychometric lenders and other participants along the psychometric continuum at each step with established practice teams focused on corporate law, M&A, federal and state consumer finance regulation, bank regulation, payments systems, securitization, privacy and data security, investment advisers and securities broker-dealers.
This advisory is published by Alston & Bird LLP’s Financial Services & Products practice area to provide a summary of significant developments to our clients and friends. It is intended to be informational and does not constitute legal advice regarding any specific situation. This material may also be considered attorney advertising under court rules of certain jurisdictions.