DID YOU KNOW: Why the VantageScore model has been rebuilt three times in 10 years?

Date: June 24, 2020

Every few years credit score model developers, like other software developers, rebuild and relaunch their models. VantageScore has three models: 1.0, 2.0 and, most recently, VantageScore 3.0.

Obviously, these developments lead to a situation where consumers potentially have a large number of different credit scores. While the data overwhelmingly demonstrates consumers know that they have many scores and believe it is a net positive, it is still important to explain why models must be rebuilt.

The main reasons why models must be rebuilt generally fall into two categories.

First off, innovation is now an important part of the credit scoring market, and new innovations must be built into models. Model developers are continually testing new approaches, taking advantage of more innovative model architectures. As these approaches get refined, they are incorporated into new models. Examples of this include the exclusion of paid collection accounts and the addition of specific scorecards for scoring consumers with sparse credit histories.

Secondly, the credit environment changes. Shifts in the economy and interrelated changes in consumer credit behaviors contribute to a breakdown in the predictiveness of any credit scoring model over time. To develop its model, VantageScore uses two blended timeframes to capture two different, two-year periods representative of current market conditions. As those timeframes become less current, it becomes more likely that there will be a breakdown in the predictiveness of the model. When this occurs, model developers may use more current blended timeframes in order to build a new model that is more reflective of current conditions.

This is why VantageScore has launched three different models in its ten-year history. The development of new models increases innovation and predictiveness, which in turn benefits lenders, consumers and the economy.