Validation results for vantagescore 4 0

Validation results for VantageScore 4.0

Date: June 22, 2020

Validation Study for VantageScore 4.0 Credit Scoring Model Shows Increased Predictive Performance and Alignment of Credit Scores

VantageScore 4.0 outperforms across all major lending industries in existing account management and new account originations;

Results Posted Publicly to Aid in Model Governance

VantageScore Solutions announces its inaugural validation study for the VantageScore 4.0 model, which showcases how the latest model outperforms pre-existing versions of VantageScore and credit reporting company (CRC) scoring models. With the inclusion of trended credit data and the application of machine learning techniques, the VantageScore 4.0 model shows predictive lift that can significantly improve a lender’s business.

Highlights of the 2018 validation study results include:

  • Overall, VantageScore 4.0 provides a predictive lift as compared with earlier VantageScore models and the CRC benchmark model for mainstream consumers in all credit categories.
  • Consistency – 93% of consumers receive credit scores from the three CRSs, which fall within a 40-point span (95% for the mortgage, 94.8% for bankcard).
  • Universe Expansion – 4.2% to 9.9% performance lift as compared with VantageScore 3.0 in the categories of account management and originations (respectively) among newly scored consumers.

As part of its commitment to transparency, VantageScore Solutions annually validates all VantageScore models and shares the results publicly. VantageScore Solutions considers this a best practice, which facilitates compliance of users of the model.

This is the first validation study of VantageScore 4.0 since its launch last year. A standard two-year timeframe was used for the analysis (2015-2017).

Trended Credit Data

The VantageScore 4.0 model incorporates first-to-market innovations such as the use of trended credit data, which captures the trajectory of borrower behaviors over time.

In the validation study, trended attributes’ contribution to the score more than doubles in predictiveness for low risk (Prime and Superprime) credit tiers compared with high risk (Subprime, Near-Prime, and Thin & Young) credit tiers. This provides a better separation in lower risk segments, a highly desired population for many lenders.

Machine Learning

VantageScore 4.0 is also the first and only tri-bureau credit scoring model to leverage machine learning techniques in their development of scorecards for those with dormant credit histories (i.e., those with scoreable trades but with no update to their credit file in past six months).

In the validation study, VantageScore 4.0 outperforms VantageScore 3.0 by nearly 10% for new account originations and 4.2% for existing account management trades.

“Our validation studies are a hallmark of what sets VantageScore apart from other credit scoring models. Every year, we rigorously test our credit scoring models and share the results because that’s how much confidence we have in our models,” said Barrett Burns, CEO, and president, VantageScore Solutions. “This year is no different. Through the latest innovations we developed for VantageScore 4.0, we unequivocally proved a more predictive and consistent model can be accomplished without lowering credit risk standards.”

Validation results for VantageScore models are always shared publicly online. This year’s study can be found at