Artificial Intelligence

As Pandemic Reshapes Consumer Behavior And Credit Risk: In AI We Trust?

The hoarding might be done. Toilet paper is no longer a precious commodity. Grocery shelves are (relatively) well-stocked.

As the pandemic continues to take root in the U.S., there are some shifts in consumer behavior that will be long-lived, perhaps permanent. We’re all living life online and having the essentials of daily life delivered.

Unemployment, of course, is skyrocketing. More Americans are living paycheck to paycheck. Banks are reserving tens of billions of dollars against potential credit card and loan defaults. They’re eyeing risk exposure while at the same time trying to help consumers get back on their feet.

In an interview with Karen Webster, Mastercard Senior Vice President and Head of Brighterion Sudhir Jha said that advanced technologies – specifically artificial intelligence (AI) – can help FIs weather the storm. Used correctly and ethically, AI can even help us, as a society, re-emerge from the shadows of the coronavirus by aiding public health efforts and getting consumers back into the physical world and spending.

Shifting Consumer Sentiment

And, of course, in those places where they are spending, especially in (limited) brick-and-mortar settings, contactless rules the day, as we shun handling paper bills and coins, or even handing over plastic cards.

“People are afraid to use certain payment instruments,” remarked Jha. The transition to digital payments, long on Mastercard (and other firms’) roadmaps, has been hastened by a public health crisis.

There’s also been a marked shift in consumer sentiment about their personal finances. Amid job losses and reduced wages, they’re pulling back on credit card spending and embracing debit transactions.

Mastercard, for example, said in its latest earnings call that debit/prepaid transactions were up 9.5 percent year over year. Elsewhere, merchants increasingly have been offering buy now, pay later (BNPL) options, recognizing that consumers extend payments far out into the future, when they hope to have firmer financial footing after things normalize.

Touching on how consumer behavior may evolve longer term, Jha echoed comments from management on the latest Mastercard earnings call, which laid out a roadmap across containment, stabilization, normalization and growth.

The first stage, containment, was where the hoarding was happening. Jha noted that the frenzied buying behavior “is going away as we get into the normalization phase.”

With a nod toward Mastercard’s own capabilities through Brighterion, Jha said that “we’ve already had different models for all these different types of transactions. And the good news is that over the last two months of these [coronavirus-related] changes, we have done well in terms of scoring the new change in dynamics.” 

As more spending has come online, and as people buy more items all at once (to be shipped to the house), he said that Mastercard has not been declining those transactions.

Recalibrating Risk

The prospect of extending credit, of course, brings key issues into focus, primarily those of risk and fraud. The global macro environment can be charitably termed rocky. Beyond bracing for credit defaults, banks have been monitoring credit lines and credit limits. They also have been sensitive when it comes to collecting past-due payments as customers ask for forgiveness, forbearance or deferrals.

Such activity, however, can short-circuit credit scoring efforts, especially moving forward. That’s because, as Jha explained, traditional risk scoring models have been predicated on payment history, and how payments may be spread across multiple cards and loans.

“How up to date you are on the payments plays a very big role in the models … and that data may not be very useful,” he said.

Jha noted that Brighterion has created a range of “sub-models” that can be sensitive to such changes in activity and focus more on transactional data – viewed across multiple merchants and multiple cards – than past payment history.

This way, he said, issuers can focus on the 90 percent of their customer base that will prove to be a good credit risk, and who can become even stronger customers in the future.

The goal of an AI-driven scoring model, Jha said, should be twofold: not just to find and pinpoint the risk of default, but also to illuminate customers of good standing who are in need of additional credit to get them through the next several months of the pandemic.

There’s a proactive element, too, where AI-powered analysis can help Mastercard reach out to customers well before delinquencies happen to help negotiate terms that can keep the lender/borrower relationship intact.

“Some issuers are already seeing this as an opportunity, because there's going to be a lot of demand for credit,” Jha pointed out.

“And if you can actually have a really good AI-enabled model to distinguish good customers versus bad customers, you can take advantage of this – and grow. And the second part is that advanced marketing can give you much more flexibility to deal with customers and help them repay loans or credit,” he added.

The key to any good model, of course, is flexibility: As Jha noted, the expectations may be that in three, four or five months, commerce will return to some semblance of normalcy.

“And if it doesn't, again, your model has to predict a ‘new normal,’” he said.

Beyond Risk Scoring

AI’s utility is not confined simply to financial services. As Jha and Webster discussed, advanced technology also plays a role in public health, if used judiciously and ethically.

As the world rushes to find a coronavirus vaccine, consumers need to feel more comfortable re-engaging in the physical world in the ways they once did. Contact tracing can help re-establish that comfort level, tracking symptoms, possible exposure and, ultimately, the virus’ progress.

But in order to leverage technology in the service of public health, there must be a framework in place.

Yet it will be nearly impossible to do contact tracing manually. To gain that requisite level of trust, he said, scientists – who do have the public trust – can advocate for the use of advanced tech in combatting the virus through testing and tracing.

“Some of the trust that technology has lost is hurting us,” Jha told Webster. “This might be an opportunity for us to earn that trust back.”



About: Accelerating The Real-Time Payments Demand Curve:What Banks Need To Know About What Consumers Want And Need, PYMNTS  examines consumers’ understanding of real-time payments and the methods they use for different types of payments. The report explores consumers’ interest in real-time payments and their willingness to switch to financial institutions that offer such capabilities.