The VantageScore 3.0 model by the numbers

Date: June 22, 2020

Credit scoring accuracy depends on highly complex math, but you don’t have to be an analytics whiz to run these numbers on the competitive advantages of using the VantageScore 3.0 model:

  • The VantageScore 3.0 model provides as much as 25% performance improvement, or increase in predictive accuracy, with the prime and near-prime credit tiers.
  • In 99% of tests, the VantageScore 3.0 model outperforms the already strong VantageScore 2.0 model and all other benchmark models provided by the three credit reporting companies (CRCs)—Equifax, Experian and TransUnion.
  • The model generates scores for 30–35 million American consumers who can’t receive a score from other traditional models. (Click here for more on this topic.)

The VantageScore 3.0 model sets new standards for:

  1. Consistency
  2. Performance on both originations and account management
  3. Lender risk alignment, which is nearly identical across all three CRCs
  4. Real estate performance

More granular data, better accuracy
The key driver of the VantageScore 3.0 model’s improved performance is its use of more granular data from the three national credit reporting companies. The deeper and broader data offers significant new insight into consumer behaviors. In fact, while developing the model, the VantageScore analytics team was able to define and consider 900 credit characteristics before selecting the most predictive 150 characteristics available, 40% of which are new compared to previous VantageScore models.

This data’s enhanced predictive attributes include:

  • The ability to distinguish first mortgages, home equity loans and home equity lines of credit (HELOCs), greatly enhancing predictiveness related to a borrower’s real estate debts. In previous scoring models, all real estate loans were considered together in a single category.
  • Installment loans are identified separately as auto loans, student loans, personal loans and installment loans.
  • Secured revolving loans are distinct from unsecured revolving loans.

An additional design innovation enhances VantageScore 3.0 predictiveness for both loan originations and management of existing accounts. By extracting greater predictive insight from available data, VantageScore 3.0 provides performance on par with models optimized solely for originations.

These advances in scoring predictiveness are crucial in a credit world that is vastly different from any we’ve seen before: “Exotic mortgages” are no more. Default rates in the student loan market are reaching astronomical levels. Some consumer segments are reducing debt dramatically, while others are sinking under their debt loads. Mortgage modifications, short sales and foreclosures are at historic highs, while auto and credit card delinquencies are at historic lows. Many moving parts are in play.

Credit cycles change. Consumer behaviors change. And this most recent credit cycle has had a profound impact on most consumers. These are the reasons we saw a need to create a credit scoring model with the stability, inclusiveness, and predictive power of VantageScore 3.0.