Another advanced feature of the VantageScore 3.0 model is that it was developed using an exceptionally wide breadth of consumer credit-file data. This allows the model to retain its predictive power for longer periods of time and through economic volatility.
The years following the last recession brought dramatic changes in U.S. consumers’ credit behavior, and VantageScore analysts noted significant shifts in the types and combinations of behaviors that were most predictive of defaulting on loans. This inspired an approach to developing the VantageScore 3.0 model that would account for consumer behaviors both in times of economic upheaval and of stability.
To that end, VantageScore secured 15 million anonymized consumer credit files from each of the national credit reporting companies (CRCs—Equifax, Experian and TransUnion), yielding a development sample of 45 million credit files. And instead of taking the traditional approach of basing the model on borrower behavior during a single two-year period, the team based the VantageScore 3.0 model on blended behavioral data from two time frames: 2009-2011, the height of the economic crisis; and 2010-2012 the more stable period that followed. This approach reduced the model’s sensitivity to volatile behavior that can be found in a single timeframe, and extended the model’s performance stability.
Visit our Video Hub to see both recent and classic releases from VantageScore Solutions.