Back to Blog

How to Underwrite Loans When Everyone is a Higher Risk

COVID-19 has rewritten so many rules about the economy. It is now more difficult than ever to underwrite risk and ultimately understand if a consumer will pay back their loan.

The Wall Street Journal reported late last month that many lenders have implemented stricter lending requirements because of this challenge. In some ways, this is necessary for banks to protect themselves. However, the more stringent standards also create hardships for consumers who could really use some extra cash right now.

Policymakers have intervened to encourage banks to loosen their lending standards to meet consumer needs during this time. Banks are being told not to pay attention to credit as much as they used to and to not collect more than a year’s worth of data for underwriting.

“There is significant pressure by the Small Business Administration to make unsupported loans,” said career banker and author Richard Lawless. “Banks are being told, ‘don’t pay attention to bad credit.’ This will result in loan losses of 10%, or more. All of which amounts to the new CDC guidance for banks, ‘don’t wear you mask, don’t wash your hands, touch everything, and gather in large groups. It’s okay, the government has got your back.'”

Fortunately, non-traditional underwriting models have been gaining popularity in the fintech space. Many of these models don’t rely on a borrower’s financial standing, but instead pull data from alternative sources such as social media. Two things fueling this recent explosion include the availability of more data and the advanced expertise of AI.

California-based Neener Analytics relies on both of these aspects– the abundance of data as well as AI– for its risk outcome predictor. The company offers businesses an “automated psychologist” that tells companies the likelihood that a prospective borrower will pay back a loan. Unlike the way many companies analyze creditworthiness, Neener Analytics doesn’t look at whether or not the consumer’s financial situation is in good shape. “The question isn’t can they pay us back– that’s easy to figure out,” said CEO Jeff LoCastro during his demo at FinovateSpring 2019. “The question is will they.”

The company places a lot of weight on what it considers small data and human data. Regarding the impact of the COVID-19 crisis on consumer credit scores, LoCastro said, “The market is going to be hit with a tidal wave of newly undecisionable consumers: consumers who on a Monday were a good bet, but by Friday will suddenly be unacceptable. They missed payments because of a global COVID shut-down…not because they are a bad risk; this is a health crisis, not a financial one. But the big data algorithms can’t account for that… Only small data can see beyond COVID; only through small data is the consumer still a distinctive individual human being endowed by a unique matrix of conditions and domains that manifests in binary outcomes.”

To help businesses underwrite risk in this new environment, Neener Analytics’ tool turns to social media. With over 70% accuracy, this “automated psychologist” tool can be summoned via a one-click decisioning tool or a chatbot dubbed ARIA. Both methods eliminate the need for lenders to ask more questions on loan applications, which often leads to abandonment.

“We all know sometimes bad things happen to good people,” added LoCastro. “The only way to bridge this is through human data . . . not through more underwhelming historical, transactional, or relational approaches. With Neener Analytics, consumers who were a good bet on Monday . . . will still be a good bet on Friday.”