The Use of Cash-Flow Data in Underwriting Credit

Overview

Even before the Covid-19 pandemic massively disrupted the nation’s economy, millions of consumers and small businesses struggled to access affordable credit because of gaps and weaknesses in traditional financial information systems. Approximately 20 percent of U.S. consumers lack sufficient credit history to predict their repayment risk using conventional scoring models. Small businesses also face particular information barriers in accessing credit, especially those businesses that have not yet built financial track records.

To fill these gaps, both traditional incumbents and new entrants have been experimenting with various sources of “alternative” or “non-traditional” data. One of the most promising of these alternatives is cash-flow data — such as records from consumers’ deposit and card accounts and from small businesses’ accounting software — because it provides a relatively detailed and comprehensive picture of how applicants manage their finances on an ongoing basis. Yet while recent technological and market developments are making it easier for lenders to access cash-flow information electronically, the adoption of underwriting models that rely on detailed analyses of such information is uneven in the United States.

FinRegLab set out to investigate the use of cash-flow data in credit underwriting by conducting an empirical assessment of its benefits and risks, as well as market and policy analyses of the challenges to its wider adoption. We view the project as a useful case study at the intersection of two broader financial innovation trends: (1) the transformation of automated credit underwriting as firms experiment with new data and analytical techniques; and (2) efforts to structure the new data transfer system to enhance customer control and spur greater competition and innovation in financial services markets.

Our groundbreaking empirical research used data from six non-bank financial services providers—Accion, Brigit, Kabbage, LendUp, Oportun, and Petal—that are using cash-flow data in an effort to provide unsecured, relatively short-term credit to consumers and businesses who may have difficulty obtaining loans from traditional sources. We retained Charles River Associates to help us design and conduct an independent analysis of the participants’ cash-flow variables and scores relative to actual loan performance for purposes of assessing general predictiveness, inclusion, and fair lending risk in connection with credit eligibility determinations.