The coronavirus pandemic has disrupted an already overly complicated and legacy banking system. But it has also spotlighted major vulnerabilities and the need to accelerate digital transformation in credit scoring models to improve lending.
Despite a recession and global pandemic, the average FICO credit score hit an all-time high of 711 in 2020, according to Experian. Credit scores somehow managed to soar at the same time unemployment levels spiked to 14% in April 2020, surpassing the previous 10% peak during the financial crisis, according to the U.S Bureau of Labor Statistics.
Stimulus checks and other government relief programs used to pay down debt, coupled with an overall spending decrease as people remain in lockdown, may have been contributing factors to higher scores. But too many variables haven’t been taken into consideration, leaving financial institutions without the full picture about the consumer seeking credit.
FICO says its credit scores are used in more than 90% of lending decisions in the United States, so having a low or bad score seals a consumer’s fate with lenders. Banks typically rely on credit scores, data supplied by a credit bureau, recent customer activities with the bank and run that through credit risk management models periodically. Meanwhile, the credit score data may be subject to months of lags, even while some consumers may be trying to improve or correct their scores.
The standard models also fail to factor in how consumers respond to rapidly changing circumstances — such as the pandemic. Banks can make smarter decisions by supplementing credit scores with real-time data concerning the behaviors and needs of customers while using artificial-intelligence-based models to gain additional insights. Some lenders are beginning to use such an integrated and streamlined view of their customers for specific use cases in recent times.
For example, consider a real-case scenario in which a national private-label credit card issuer wanted to attract a specific segment of customers to its premium products. The PLCC tapped into a partnership between the outcome-based marketing provider Epsilon and TransUnion to create a scalable, accurate and transparent prescreen solution that combined real-time behavioral data with finance-related data and credit-active profiles from TransUnion. The PLCC could then define a creditworthy, credit-ready audience to target using direct mail and was also able to deliver appropriate, relevant offers based on each prospect’s current need.
Another use case could involve proactively managing and supporting customers who are still paying their bills on time but facing financial difficulty or uncertainty. Since they are not yet delinquent, their financial difficulty would not impact their FICO scores. During the pandemic, lenders have typically offered payment deferrals or other types of support to their customers who called in or emailed seeking help. This outreach program could have been made more targeted if the lenders had additional insights to identify the specific customers who were most challenged to pay.
In the case of customers who had turned delinquent and might be worried about escalating debt and high interest rates, banks would typically offer a hardship program. By leveraging data and insights, banks can potentially determine customers that are most likely to repay and proactively enroll them into targeted hardship programs. This way, banks can ensure a strong customer experience and appropriate payment structures while reducing the level of credit risks that they have to take on.
For example, American Express recently launched an Amex financial relief program to allow millions of its cardholders to have access to reduced annual percentage rates and payment amounts, and it waived late fees and annual fees. A key to the improved success of such financial relief programs is for lenders to draw upon real-time, in-depth and wide-reaching data to target the offers to the customers most in need instead of solely looking at credit scores and bureau data.
Through a broader data lens, banks can pick up on default signals earlier than they would have with simply a traditional credit score modeling. Lenders should look to leverage a broader set of data sources, apart from credit scores alone, to make better lending decisions that provide more credit access to consumers while preventing unforeseen credit risks to their portfolios.
Priya D. Bajoria is senior vice president, financial services and Steve Reiss is associate managing director at Publicis Sapient.
