Plaid introduces LendScore for real-time credit risk assessment

Plaid on Wednesday introduced LendScore, a new credit risk model that uses real-time cash flow data and account connection insights to give mortgage lenders a more complete view of a borrower’s financial situation.

The launch comes amid significant shifts in consumer borrowing behavior and credit scoring. Plaid pointed to a recent TransUnion study, which found that Americans currently hold about $257 billion in unsecured personal loans, and more than half of borrowers who took out a personal loan in the past five years are still repaying it.

And as borrowers increasingly use alternative products like buy now, pay later (BNPL) services while relying more on debit cards than credit cards, credit scores are proving to be less reflective of consumers’ financial lives.

Meanwhile, the credit system has seen more change in the past year than in the previous decade, with FICO planning to start selling scores directly to mortgage lenders and the adoption of VantageScore 4.0 by the government-sponsored enterprises (GSEs).

With lenders ready to explore alternative risk data — such as cash flow underwriting — Plaid’s LendScore aims to build on that shift by leveraging the company’s network, which connects 150 million people to more than 7,000 financial apps and services.

LendScore uses real-time cash flow data along with other digital financial activity from Plaid’s vast network. After a borrower gives permission, Plaid analyzes the data and generates a score between 1 and 99, plus “reason codes” that explain the score.

The model is delivered through Plaid Check, the company’s consumer reporting agency, ensuring compliance with the Fair Credit Reporting Act and fair lending rules.

According to Plaid, early testing showed the score reduced relative credit risk by up to 20% for subprime and near-prime borrowers. This potentially saves lenders more than $1 billion in losses and cuts borrowers’ annual interest costs by an estimated $8 billion.

“Traditional scores still matter, but they miss what’s happening right now,” said Michelle Young, Plaid’s credit product lead. “LendScore helps lenders see the full financial story.

“We built LendScore in close partnership with unsecured lenders, which helped shape the attributes we prioritized, the structure of our reason codes, and how we measured model performance,” she added. “In testing, we generated millions of scores and delivered a 25% lift in predictive performance for lenders, compared to traditional credit data alone.”