Fintech partnerships with conventional lenders are providing banks with the technology they need to both access and analyze such data . Siless says that , even a few years ago , only those with highly technical skills could develop and train credit models to incorporate alternative data sources . Now , however , the accessibility of cloudbased software means that more organizations can incorporate ML and alternative data into their own credit decisions . “ It ’ s much easier ,” she says .
ML is also helping to identify new predictors of creditworthiness and evaluate their impact on consumer credit scores faster , according to 2022 Oliver Wyman / Experian research : “ Analytical techniques turn data on prospective borrowers into information about their likelihood to default ,” the report says . “ In recent years , the application of [ ML ] techniques to credit data has enabled several opportunities for enhancement .”
Lastly , consumer literacy regarding alternative credit and even digital payments has increased exponentially . Alexandra Rizzi , senior director for responsible data practices at the Center for Financial Inclusion ( CFI ) says the pandemic played a big role , with many government assistance payments relying on digital pathways . In Paraguay , for example , government-to-person payments during the crisis yielded 1.5 million new mobile wallets , according to a CFI report .
“ Better identifying and serving the creditworthy consumers [ lenders ] miss today is an opportunity to grow their business while doing good .” — Oliver Wyman / Experian report
The alternative data promise The growing accessibility of advanced technology that utilizes alternative data may open up entire new credit markets — and improve borrowers ’ lives . Dimuthu Ratnadiwakara , assistant professor of finance at Louisiana State University , says that while traditional models lock out people with low credit scores , “ some of them are more creditworthy than their credit score suggests .”
In a recent Harvard / LSU study , Ratnadiwakara analyzed the outcomes of three million borrowers who received personal loans from Upstart , one of the first online lenders to expand credit access using alternative data and ML models . The analysis showed Upstart provided loans to consumers who wouldn ’ t usually be able to access credit . Even better , those who secured loans improved their credit scores and gained access to additional credit opportunities . For example , the borrowers who got a loan from Upstart were more likely to obtain a first mortgage and receive better rates on future credit products .
The Oliver Wyman / Experian report estimates that , by using currently available alternative data and advanced analytics , lenders could score 21 million people who are unscorable because they have thin credit files and score a majority of 28 million applicants who have no credit history . “ The endeavor will require some investment in new analytical techniques , new data sources , new products and potentially new ways of engaging with communities ,” the report says . For lenders , “ better identifying and serving the creditworthy consumers they miss today is an opportunity to grow their business while doing good .”
New frontier , new challenges As alternative data use tips into the mainstream , regulators and nonprofits are exploring best practices to help ensure that data sharing doesn ’ t harm consumers . For example , Rizzi says fintechs could provide more transparency
PHOTO COURTESY OF QUIPU
38