Event Summary: Lending and Credit Risk in the Digital Age

The financial services industry is undergoing a digital transformation with advanced computing and increased amounts of data available that can be used to assess credit risk in lending. How is the digitization of lending impacting small businesses and consumers alike? On October 26, the Bipartisan Policy Center held a panel discussion to better understand the impact of technological change on credit underwriting and the economy of lending.

Congressional Insights

Rep. Zach Nunn (R-IA-3) kicked off the conversation by explaining the vital role he sees digitization playing in the financial services sector in his home district in Iowa. Nunn highlighted both the heightened risks and benefits that digitization can bring to our economy. He noted how digitization can provide benefits, such as increased access to credit and a fairer financial system, for the underbanked. However, Rep. Nunn cautioned that novel digitization is not without novel risks, such newly digitized services being used for fraud and other illicit activity. Nunn cited the work he has done in promoting the Financial Protection Act of 2023 to counter illicit activity and bad actors. Nunn also discussed artificial intelligence (AI) and its potential effects on the financial system. According to Rep. Nunn, Congress is up to the task of regulating AI. He added that Congress wants community banks to be adequately prepared to invest in AI, while ensuring privacy and discrimination laws are rightfully enforced.

After hearing from Rep. Nunn, we were joined by FinRegLab’s Kelly Thompson Cochran, Enova International’s Joe DeCosmo, and Brookings Institution’s Aaron Klein for a panel moderated by BPC’s John Soroushian. The discussion began with panelists sharing their individual perspectives on how digitization will affect the credit economy in the near future.

Digitization and Finance

Cochran opened the discussion by exploring the significant impact of digitization and novel forms of data analysis on credit and lending in the U.S. She elaborated on the benefits that small businesses and underbanked individuals would reap through the application of innovative and novel data techniques in creditworthiness assessment. Klein echoed similar benefits but voiced concerns about the potential homogenization of credit factors raising the risk of excluding some individuals from the credit system. DeCosmo emphasized the importance of having humans “in the loop” when using AI and highlighting risks related to fraud and cybersecurity.

Credit Analysis Basics and Risk-Based Pricing

Klein explained how lenders use credit scores, which are derived from credit reports assessing individuals’ past credit usage, to make lending decisions. He went on to discuss how the Fair Credit Reporting Act governs credit data reporting, and the Equal Credit Opportunity Act regulates how lenders allocate credit, requiring explanations for credit denials.

Cochran stated that risk-based pricing enables lenders to charge higher interest rates on borrowers with a higher likelihood of default. She noted the debate between proponents, who argue it facilitates credit for riskier applicants, and opponents, who contend it may increase the likelihood of default for such applicants.

DeCosmo supported risk-based pricing, stating it allows lenders, such as his firm Enova, to better match and provide credit to consumers and small businesses. He emphasized that, for him, risk-based pricing involves not just the interest rate but also factors like loan amount and term length.

Klein pointed out that some products, like overdraft fees, don’t depend on risk-based pricing, highlighting the tension between risk-based pricing and cross-subsidization.

New Datasets and Alternative Data

Cochran explained how incorporating new datasets can help evaluate the credit risk of individuals who cannot be scored under traditional credit systems, thereby promoting financial inclusion. She pointed out challenges with new datasets, including flaws in the data, bias against protected classes, and relationships may not hold over time.

Klein noted that the existing system faces similar issues, with stricter rules on alternative data. Further, Klein discussed the potential risks and benefits associated with the use of alternative data sources in credit assessment. While acknowledging that alternative data can increase access to credit for the traditionally underbanked, he cautioned that, without appropriate guardrails, alternative data may serve as a proxy for protected classes, potentially leading to more bias and discrimination in lending. DeCosmo emphasized how some alternative datasets are already helping his firm make better lending decisions and broaden access to credit, adding that cash flow-based models are an increasingly important source of information.

Artificial Intelligence

The panel also explored artificial intelligence and emerging trends. Klein discussed the opacity of models commonly used for credit decisions and emphasized the importance of explainability in utilizing AI tools for credit analysis. DeCosmo emphasized how his firm uses machine learning, not generative AI, for underwriting. While generative AI helps boost Enova’s internal productivity, its credit decision-making does not employ the technology. Cochran added that while automated decision-making has existed for some time, machine learning models are now considerably more complex, necessitating robust and curated datasets to ensure fair and explainable credit decisions.

Final Thoughts

Panelists underscored the importance of leveraging innovative financial data for assessing creditworthiness while addressing the relevant challenges. They expressed a shared vision for the positive impact of digitization on the American economy — provided appropriate policies and safeguards are in place. Policymakers in Washington need to begin laying groundwork for an effective system to ensure data efficacy in the future. Preparing our regulatory system now will allow us to leverage digitization for a fairer, stronger, and more inclusive economy tomorrow. BPC will continue to engage experts and policymakers on digitization’s impact on lending for small businesses and consumers. You can read our past work on small business lending and financing here.