Examining the Use of Alternative Data in Underwriting and Credit Scoring to Expand Access to Credit

  • Endnotes
Share

Dear Chairman Lynch and Ranking Member Hill:

I would like to thank the Financial Services Committee for convening its Task Force on Financial Technology and organizing its hearing titled “Examining the Use of Alternative Data in Underwriting and Credit Scoring to Expand Access to Credit” on July 25, 2019. I am writing to express my views regarding the topic of that hearing.

My name is Dan Quan and I am an adjunct scholar at the Cato Institute’s Center for Monetary and Financial Alternatives. I also advise high-growth, disruptive fintech companies. Previously, I led the Consumer Financial Protection Bureau’s (CFPB) fintech office, Project Catalyst.

The United States has the most developed and competitive consumer credit market in the world. For example, the credit card industry has seen robust growth since the financial crisis, with $1.07 trillion in outstanding balances as of May 2019.1 Fintech lending has seen even greater growth, accounting for 38% of the $138 billion unsecured personal loan market in 2018.2 However, the inconvenient truth buried in those rosy numbers is that we still have a huge financial inclusion challenge. 45 million Americans, or 19.3% of the adult population, don’t have access to credit.3 These consumers tend to be disproportionately African American, Hispanic, young, and lowincome.4 Additionally, 53% of small businesses surveyed by the Federal Reserve could not obtain the full financing they sought.5

Now, however, the use of alternative data in credit scoring and underwriting holds great promise for bringing greater access to credit and capital to struggling consumers and small businesses. Underwriting models that include alternative data can increase lending volume, lower interest rates for borrowers, and improve the accuracy of default predictions. In short, alternative data can make lending more plentiful, more affordable, and sounder — with historically underserved borrowers and communities benefitting most.6

Alternative data can include anything that is not currently part of consumers’ traditional credit reports. It can range from account transaction history (known as cash-flow underwriting), to educational and occupational information, to social media use and other online or mobile activities. The topic of alternative data sometimes raises concerns about whether it could increase discrimination against protected classes or intrude on borrowers’ privacy. While these are legitimate concerns, it is counterproductive to prevent or tightly constrain the use of alternative data in lending. Restrictions would likely hurt, not help, marginalized borrowers who are overwhelmingly low-income and minorities. In fact, a recent study on the racial and ethnic disparities in credit access finds that the lending gap attributed to credit discrimination is much smaller for fintech firms than it is for traditional lenders.8 Discouraging innovation out of a concern for potential discrimination is therefore likely to undermine financial inclusion.

In my experience, not many firms use alternative data in credit scoring and underwriting today. Those that do mostly use cash flow data. For example, fintech lenders such as Oportun use cash flow data to provide credit to "unscorable" or "credit invisible" consumers. Other lenders, like payment processors Square and PayPal, use transaction histories to help them lend to their merchants effectively and efficiently. According to Square, its average loan size is $6,000,8 an amount that most traditional financial institutions find unprofitable to finance. Individual consumers can opt to include data like their on-time bill payment history to boost their FICO scores.9 In my conversations with fintech lenders, none report any convincing evidence that social media data can predict consumers’ repayment behavior. No lender is using it for underwriting purposes in the U.S. A recent paper from the Federal Reserve Bank of Philadelphia agrees that underwriting with alternative data can better predict loan outcomes, resulting in improved terms for borrowers who, under traditional credit criteria, would receive higher-priced loans.10

There is no need for new laws or rules just for the use of alternative data. The same laws and rules that ensure fair credit access, privacy protections, and transparency in underwriting decisions equally apply to lenders who use alternative data. On the other hand, regulatory agencies should encourage the responsible use of alternative data by providing greater clarity to lenders in two key areas.

First, a transparent, secure, and frictionless data sharing ecosystem is necessary for increasing the use and reliability of alternative data. Regulators should support the Treasury Department’s position that consumers have a right to permission their own financial data for third-party use.11

Second, the CFPB needs to issue clear guidance on what responsibilities under the Fair Credit Reporting Act that third-party data aggregators and financial institutions have when consumers request to share their financial data through a secure mechanism such as an API. Uncertainty about these responsibilities creates confusion, friction, and can potentially result in consumer harm.

Additionally, the private marketplace, through industry and consumer group collaboration, can work together to ensure consumers fully understand their data rights, including how lenders will use their data, before they consent to sharing them.

The use of alternative data will enhance the affordability of credit and make our credit system more inclusive. While policymakers should remain vigilant, they must also stay open-minded about how to encourage further developments in market innovation that will benefit consumers, small businesses, and the economy at large.

I appreciate the opportunity to comment on this important hearing.

Dan Quan

Task Force on Financial Technology
Financial Services Committee
United States House of Representatives

Notes