By Melissa Koide, Larry Rosenberg
America’s credit system is under serious pressure as it faces the most sudden and severe downturn since the Great Depression. In the months and years ahead, access to safe and prudent credit will make a critical difference to families and businesses across the country — as well as to the performance of economic aid initiatives.
So, as we seek to build a fairer and more inclusive financial system to support recovery from the COVID-19 pandemic, we need a credit system to match. That requires moving beyond traditional data and models to incorporate new transaction and payment information that is more timely, more insightful, and more inclusive.
Bank account, payment and other cash-flow information is one of the most promising sources of underwriting data relative to traditional credit history because it reflects both income and expenses and is available for more consumers and small businesses.
While it is not likely to close all gaps between demographic groups, FinRegLab’s research finds this data may help to expand credit to millions of underserved consumers and small businesses. The pace at which the use of more nuanced financial data gets scaled up will depend on how quickly lenders and investors decide that it makes sense to incorporate it into their business processes — as well as how quickly they can secure access to the data. Interest has increased substantially in the past six months, but many companies have meaningful work to do in order to validate existing products, adjust their models and secure increased access to data.
Federal financial regulators have signaled increasing openness to incorporating cash-flow data into credit underwriting both before and since the onset of COVID-19. For instance, in December 2019 federal banking regulators issued a joint statement highlighting the role of transaction data in promoting more inclusive credit. As FinRegLab’s report shows, though, more concrete engagement from both regulators and Congress will be necessary to making this data accessible and making sure consumers and their data are protected.
These efforts have become even more urgent in light of the ways that the pandemic and recession have exacerbated weaknesses in our existing financial and economic systems. The credit system plays a foundational role in helping millions of families and small-business owners build better lives for themselves, strengthening local communities and the national economy in the process. Yet millions of individuals and small businesses are unable to benefit from the credit system because many lenders continue to rely only on traditional credit bureau data. This information is insufficient for assessing the credit risk of millions of individuals and small businesses who lack a credit history or have a limited one.
Today’s system depends heavily on analyzing applicants’ past credit history to determine whether and on what terms to extend new loans. But that creates a Catch-22 for people who haven’t yet been able to access the kinds of loans that are typically reported to the credit bureaus unless additional data is used. The system often lumps together borrowers who mismanage their finances with those who run into financial difficulties through no fault of their own. And reporting lags are especially problematic when economic conditions are evolving, such as when the economy is heading into or out of a recession.
Even before COVID-19, an estimated 50 million U.S. adults lacked sufficient credit history to be scored reliably by traditional models based solely on credit bureau data. Another 80 million often paid higher rates or were rejected outright because they had “nonprime” scores. Together, that’s more than half of the U.S. adult population. Similar data gaps and hurdles affect small businesses as well, particularly since lenders often check owners’ personal credit histories when underwriting commercial loans.
Blacks and Latinos have substantially lower average credit scores than non-Hispanic whites and are nearly twice as likely to lack sufficient history to be scored at all. Much of this is driven by disparities in income, wealth and economic opportunities, disparities that have been generated in part by historical discrimination.
The economic downturn caused by the COVID-19 pandemic threatens to exacerbate these barriers to credit because of widespread small business closures and unemployment. And Blacks and Latinos are both substantially more likely to have suffered financial shocks due to illness and economic disruption and more likely to have encountered delays and other difficulties in accessing pandemic relief programs than white and Asian peers.
Existing credit models are not designed for these conditions, and traditional credit records are lagging indicators of both shocks and recovery. In the face of macro-economic uncertainties, simply tightening existing credit risk standards across the board is a blunt instrument that negatively impacts many responsible households, especially Blacks, Latinos and low-income households.
The financial sector recognizes these limitations for some consumer segments and is seeking a way forward. Lenders are beginning to explore new scoring products, sources of data and analytical methods in order to try to separate meaningful signals from the noise. But identifying which changes reduce existing barriers —
rather than raise them — will require coordinated effort from businesses, advocates and government.
Greater investment in the human side of the credit system is also critical to achieving fairer and more inclusive outcomes over time. For example, lenders can use diverse project teams to develop and vet new models. They can also increase their outreach to populations who have limited or negative historical experiences with the credit system by designing better product fits.
Much like vaccine development, this process will need to strike the right balance between speed and safeguards. There is no time to waste, though. After the Great Recession of 2008-09, major credit scoring vendors did not roll out updated models until 2013-14. Even then, their uptake was limited — the most widely used models in the market today were actually developed using pre-2008 data. If we want to credit to help deliver a more rapid, inclusive and fair recovery this time, it’s clear we need to adopt more insightful financial data in lending.