Small businesses are the backbone of the U.S. economy. They recovered steadily after the pandemic and have shown resilience in recent years amid rising rates and high inflation. According to the U.S. Chamber of Commerce, 33 million small businesses employ nearly half the entire American workforce and represent 43.5% of America’s GDP.

The key factors that determine Small Business success are access to capital and the ability to obtain lending as business needs evolve.

The above chart from the Federal Reserve Bank of St. Louis indicates new business applications have nearly doubled after the pandemic and there is pressure on financial institutions to serve the rising demand of higher lending appetite of these businesses in a timely manner.

According to the findings from the 2023 Small Business Credit Survey(SBCS)about half of applicants (51%)were fully approved for the financing for which they applied, resulting in financial gap challenges for the remaining 49% of applicants. Moreover, small businesses owned by people of color or immigrants led to even lower approval rates due to poor/thin business credit files which is a significant headwind in getting loans based on the traditional commercial credit score.

According to SBCS, business loan applicants most often cited an existing relationship with their lender as a factor influencing where they applied. Therefore, banks can leverage the transactional data of existing clients who apply for business loans in addition to the traditional commercial scores provided by the credit bureaus’ financial institutions.

Moreover, the advent of open banking and CFPB’s guidance around the Personal Financial Data Rights Rule will further enable banks to access their client’s cash flow data from every financial relationship.

This treasure of data can be valuable to building holistic measures around the financial health of clients and developing internal credit scores. Financial institutions can capitalize on these data assets and achieve business lending growth for their existing clients while maintaining the necessary exposure to credit risk.

Machine Learning (ML) models and generative AI (GenAI) can be leveraged to decipher valuable financial insights from non-traditional data sources to gain a comprehensive understanding of the small business owner’s finances. Key financial ratios around a business’s balance sheet can be built based on the monthly outflows (Payments to vendors/utility, etc.) and monthly revenue generated that can be critical inputs to predict future cash flows.

Large Language Models (LLMs) can help mine through the contracts being signed by small businesses to determine the future flow of funds. These models can also incorporate macro industry performance as well as the activity in the geographical footprint of the business.

Layering in the local government growth projects and estimated economic activity can provide an overall understanding of the macro variables to append in credit decisioning. Based on all these attributes one can build an internal scoring algorithm and use it in context with the traditional credit score to expand the credit population as well as swap previously approved high-risk populations with lower risk based on a combination of scores.

As organizations explore revamping underwriting models, they can perform below steps:

1. Identify data assets and derived financial ratios: Understand what attributes related to the financial health of a customer exist in their database. Partner with the business team to define base attributes and derived financial ratios. Explore the external data already purchased related to the client’s cash flows.

2. Analysis: Leverage ML techniques to quantify the lift in the lendable population while maintaining credit risk, which is significant based on statistical measures.

3. Deployment: Document and work with model risk teams to socialize the model and get necessary approvals due to the highly regulated nature of the work.

4. Monitoring: Measure the outcomes as we get the actual performance of booked loans and iteratively fine-tune the model to deliver consistent results.

Conclusion

The small business financing space is growing and traditional banks need to evolve their approach to provide time-sensitive tailored solutions in the highly competitive space of fintechs/neo banks. Lending relationships are long-term and bring along the stickiness of clients that can in the future provide opportunities to deepen with higher revenue-generating merchant/treasury solutions as small businesses grow in scale.

*Disclaimer: The views expressed in this article are solely those of the author and do not represent the opinions or policies of his employer.

About the Author:

Jasmeet Bhatia is VP, Decision Science Team Lead at Truist. He has 15 years of experience in the Financial Services industry and has led data science teams’ efforts to leverage advanced analytical techniques to understand customer journeys and build personalized solutions based on unique client needs.

Bhatia has three patents for ML algorithms to enhance credit underwriting using alternate data. He has been a regular speaker at multiple Data Analytics conferences and the Symposium of Data Science and Analytics organized by the American Statistical Association.