Decompressing the Transition from Historical Data to Real-Time Data in Lending
Credit underwriting is shifting from using purely historical data to real-time data
[Editor’s note: This is the fourth article (see part one, part two and part three) in a special series we are publishing from Wharton Fintech ahead of LendIt Fintech USA. They are covering topics that will be addressed at the big annual event next week. This piece is by Kathleen Cordrey, Wharton MBA Class of ’22.]
It is becoming increasingly clear that COVID will be a catalyst for change in a number of industries. Within fintech, this couldn’t be truer – especially in the area of lending. Lenders play a crucial role in the US economy. Between mortgages, auto loans, small business loans, and personal loans, there are hundreds of billions of dollars in credit extended each year.
When evaluating borrowers, lenders have always looked at historical data – such as repayment history – to determine the risk that a particular borrower presents. Lenders will consider a borrower’s credit rating as a primary indicator of their reliability. A credit score is a unique number that, in theory, is the distillation of all historical data related to a borrower’s past credit and repayment patterns.
Historical data is, of course, useful in informing a lender whether a borrower can be trusted to pay off their debt in a timely manner. However, historical data does not always give the complete picture; it leaves room for information gaps that can harm the lender, the borrower, or both. The urgent need for credit that arose during the pandemic revealed some of the data gaps that exist in current underwriting processes. These shortcomings are largely the result of inadequate consideration of real-time and forward-looking data. For a long time, it was impossible to access this data and use it in a way that would be useful. However, today there are a handful of companies working to do just that – laying the groundwork for an exciting credit transformation in the years to come.
Consider two examples that can serve as potential use cases to better seize the opportunity:
Small businesses and PPP loans. Small businesses were among the hardest hit by the pandemic and many needed loans to stay open. Still, given a foreclosure environment, it’s worth considering whether credit scores are the best metric for assessing short-term credit risk. Yes, historical data is certainly an important factor to consider – but a credit score would not indicate the liquidity position, cash flow trends, payroll commitments, or how a business operates during the pandemic. . Many companies have not been able to operate at all during this time, so the future risk profile may not be better conjectured from historical data.
Mortgages and unemployment. Mortgage lenders have faced a increase in mortgage loan applications in 2020, given the attractive environment of low interest rates. As mortgage applications increased, unemployment unfortunately increased as well. Mortgage lenders may not have access to the most accurate up-to-date employment status of their applicants. Needless to say, extending credit to someone who may not have a job is not only a risk to the lender, but also detrimental to the borrower. To confirm employment status, lenders may request additional validation through laborious and inefficient processes (e.g., manual requests for real-time employment status updates, pay stubs, etc.).
In the above scenarios, the ability to have access to real-time predictive data would significantly improve the underwriting process.
Beyond COVID, there is an incredible opportunity to use real-time, forward-looking data to increase the lending and borrowing experience today. As another example, consider payroll related loans. In addition to up-to-date employment status checks, lenders could access real-time employment and payroll information enabling increasingly accurate underwriting. A direct connection to a payroll provider would also allow lenders to deduct repayments directly from the borrower’s paycheck. This type of information allows lenders to minimize their risk and lower costs for borrowers, making it a win-win solution.
It looks like it may be time to supplement (at least) traditional assessment metrics, like credit score, with the vast amount of data that exists today. There is a new wave of FinTech companies advancing this transformation. Of course, there are many challenges that will be faced by businesses disrupting these legacy systems. From regulatory hurdles and privacy concerns to data collection methods, there are a myriad of considerations within our extremely complex financial infrastructure that will require special attention. But it provides one more reason to be excited about the fintech companies that are propelling this transformation forward – and to hear from the executives of Argyle, Ocrolus, Spring labs and Capital One discuss some of these topics on day one of the Lendit Fintech conference.