The first and only tri-bureau credit scoring models for greater consistency, predictability and accuracy.
The Model for Credit Scoring Innovation
VantageScore models are developed by using anonymized data from each credit reporting company (CRC); providing greater consistency, predictability, stability as well as new customer opportunities with each model release.
Features and Benefits
VantageScore 4.0 is the fourth-generation generic, tri-bureau credit scoring model from VantageScore Solutions. Building on the advancements of VantageScore 3.0, this model sets a new standard for predictive performance and modeling innovation; pioneering several industry firsts that benefit lenders and consumers alike.
+ 40 Million More Consumers Scored 1 -Including those who cannot obtain credit scores when other commercially available scoring models are used; enabling lenders to increase their universe of potential applicants and extending mainstream financial services to historically underserved consumers.
+ First and Only Tri-Bureau Model to Incorporate Trended Credit Data – Newly available from all three national credit reporting companies (CRCs), Equifax, Experian and TransUnion, trended credit data reflects changes in credit behaviors over time, in contrast to the “snapshot in time” approach used in other models. VantageScore 4.0 uses trended credit data to gain deeper insight into borrowing and payment patterns; providing a 20 percent improvement in predictive performance when used in combination with traditional static attributes.
+ Built Using Machine Learning Technology – To further enhance the predictiveness of scores of applicable to the population of conventionally unscoreable consumers, VantageScore 4.0 model developers leveraged machine learning techniques in the development of scorecards for those consumers who have limited credit histories. This has led to a performance lift in the scoring of consumers with dormant credit files (i.e., consumers who have scoreable trades but have had no update to their credit file in the last six months); specifically, 16.6 percent for bankcard originations and 12.5 percent for auto originations.
+ Built In Anticipation of the National Consumer Assistance Plan – VantageScore 4.0 is more predictive than prior VantageScore models while relying less on certain public record information, for example, tax liens, civil judgments and certain medical collections information, which were called out by the National Consumer Assistance Plan (NCAP). 2
+ Refreshed Data Set – Development was based on anonymized consumer credit files from all three CRCs for the years 2014-2016. VantageScore 4.0 also takes into account the latest credit products and trends in consumer behavior for a more accurate assessment of risk.
Who Can Use This Model?
VantageScore 4.0 offers predictive performance lift across credit industries, including auto, single-family residential mortgage, credit card, installment loans and other consumer credit products; outperforming VantageScore 3.0 in all major credit industries, with a significant overall predictive lift. They model may be used in all functions that require a credit score, from prospecting to origination, account management, loss prevention and recovery as well as in computations for loan loss reserves and Current Expected Credit Loses (CECL) reporting.
VantageScore 4.0 Performance
The latest performance assessment study highlights VantageScore 4.0’s superior performance on many levels. On a U.S. population representative of mainstream consumers (consumers conventionally scored by generic scoring models):
+ VantageScore 4.0 outperforms prior versions of VantageScore’s models, as well as CRC proprietary models, across all product types for both new originations and existing accounts.
+ VantageScore 4.0 continues to significantly outperform VantageScore 3.0 in bankcard, auto, and mortgage originations.
The inclusion of trended attributes provides significant improvement in predictive performance, particularly in higher credit score ranges where trended attributes contribute more heavily to the consumers’ credit score.
On a U.S. population representative of newly scorable consumers (consumers who cannot be scored by other commercially available scoring models), VantageScore 4.0 maintains its predictive power. This predictiveness is provided by the machine learning techniques that were used during the model’s development to address this population.
Aided by the use of levelled credit attributes (including trended attributes), a high degree of score consistency is maintained across CRCs, with VantageScore 4.0 representing the most consistent VantageScore model yet.
There is no statistical bias in credit scores when VantageScore 4.0’s performance is compared across ethnic groups or between mainstream and newly scorable consumers.
Features and Benefits
Introduced in 2013, VantageScore 3.0 model pairs industry-leading analytics with more granular data from consumers’ credit files to create a highly predictive and consistent credit scoring model.
+ Generates Scores for 40 Million Previously “Unscoreable” Consumers – Similar to earlier models, VantageScore 3.0 model scores consumers whose oldest trade is less than six months old, providing a highly predictive score after the first tradeline update. The model also takes into account rent, utility and telecom data when such data is present in consumer credit files. All model attributes were designed to be incorporated in a regulatory compliant generic risk model. When used in consumer credit decisions, the model provides lenders with all the required reason codes.
+ Score Consistency Across All Three CRCs – VantageScore models are the first and only “tri-bureau,” generic credit scoring models available at all three CRCs. The same VantageScore model, or algorithm, is deployed at all three CRCs. The ability to use the same algorithm is made possible because all model attributes undergo a unique, patent-protected characteristic leveling process.
+ Data Strength – In the development of VantageScore 3.0, data was blended from two different timeframes, 2009–2011 and 2010–2012, to capture a broad development sample of recent consumer behaviors, including activity occurring after the 2008 economic crisis.
VantageScore 3.0 Performance
Use of the same model reduces score variance across the CRCs and provides more consistent risk assessment. Testing demonstrates that when consumer scores are simultaneously pulled at all three CRCs, approximately 90 percent of those scores fall within a 40-point range. Any difference in the credit scores pulled simultaneously from the three national CRCs is attributable solely to data differences in consumers’ credit files at the respective CRCs.
Use of a common model across the three CRCs also enables a consistent understanding of risk regardless of which bureau a lender partners with; helping to reduce overall portfolio risk, operational costs and credit losses.
More consistent credit scores may also enhance consumers’ understanding of their scores and the credit scoring process.
Features And Benefits
VantageScore 2.0 was launched in October 2010 in response to the substantial shifts in credit markets, most notably observed in single-family mortgage lending from 2007 to 2009 as a result of the Great Recession. While the underlying architecture of the model - the leveled characteristic design and segmentation structure - remained robust, opportunities were identified to improve model performance by updating the underlying consumer credit data set.
VantageScore 2.0 reflects consumer behavior from multiple outcome periods over an extended timeframe, 2006 to 2009, in order to capture both non-recession and recession-period credit patterns. Additionally, 45 million credit files (nearly twice as many as those used for VantageScore 1.0) were used to create the dataset for greater predictability.
VantageScore 2.0 Performance
+ VantageScore 2.0 delivers predictive performance lift over the benchmark models and VantageScore 1.0 for all primary credit products, both in originations and in portfolio management.
+ Built on a dataset covering activity both before and during the 2008 recession.
+ VantageScore 2.0 provides lenders with highly consistent risk assessments across all three CRCs as a result of the characteristic leveling process.
Features And Benefits
Built in in 2006, VantageScore 1.0 introduced competition in credit scoring that has fueled innovation and benefitted both lenders and consumers.
This model provides consistent credit scores across the three CRCs through VantageScore’s introduction of patent-protected characteristic leveling and the use of a synchronized development window. The model takes into account rent, telecommunications and utility data when such data is present in consumers’ credit files. The model scores 15 million more consumers than other commercially available credit scoring models.
VantageScore 1.0 Performance
Our annual performance assessments continue to show:
+ Score consistency between the three national CRCs is maintained.
+ Continued ability to rank order over time.
- Based on a 2018 study performed on a nationally representative random sample of 15 million consumers.
- The National Consumer Assistance Plan is an initiative launched by the three nationwide consumer credit reporting companies – Equifax, Experian and TransUnion – to make credit reports more accurate and make it easier for consumers to correct any errors on their credit reports.