potential removal of tax lien and civil judgment data from credit files would not significantly impair vantagescore model performance

Potential Removal of Tax-Lien and Civil-Judgment Data from Credit Files Would Not Significantly Impair VantageScore Model Performance

Date: June 26, 2020

The predictive performance of the VantageScore credit scoring models would not be impaired significantly if the three national credit reporting companies (CRCs—Equifax, Experian and TransUnion) were to institute several changes under consideration in their treatment of public-records information.

The CRCs are considering various changes as part of the National Consumer Assistance Plan (NCAP), an initiative launched by the CRCs in March 2015. The NCAP is designed to improve the accuracy of credit reports and make it easier and more transparent for consumers dealing with their personal credit information.

Developed in conjunction with the trade group Consumer Data Industry Association (CDIA), the NCAP includes a review of all aspects of consumer-credit data reporting and management. Its final determinations have not been disclosed yet, but options under consideration include the CRCs’ removal of some or all of the tax-lien or civil-judgment records from consumer credit files.

In light of that possibility, VantageScore Solutions analyzed what impact the removal of those entries could have on consumer credit scores calculated using the VantageScore 3.0 credit scoring model.

Tax liens and civil judgments are among the many potential credit-file entries that scoring models evaluate when predicting the likelihood that any given consumer will default on his or her loan payments—i.e., by going 90 days or more past-due on a payment. Tax liens and judgments are known to be valid predictors of future default. The loss of those entries or any other predictive data could hinder the effectiveness of a credit scoring model—both in its ability to accurately assign scores to individual consumers and to accurately rank order consumers according to their likelihood of default.

Methodology and results

VantageScore’s team of decision-analytics scientists obtained credit files for four million consumers from one CRC and compared the scoring results before and after the removal of all tax liens and civil judgments from those files—a scenario representative of the “maximum-impact” treatment of tax-lien and judgment records.

The analysis, now available as a whitepaper entitled “Impact to VantageScore 3.0 Credit Score Model From Revisions to Public Record Reporting,” reveals a number of important findings, including:

  • Removal of tax-lien and civil-judgment records led to changes in VantageScore 3.0 scores for slightly more than 8 percent of the scoreable U.S. population.

  • For that fraction of the population whose scores changed, the average difference was a score increase of 10 points.

  • The VantageScore 3.0 model’s predictive performance decreased only slightly. There was minimal impact because each scorecard within the model considers multiple derogatory factors, so that other, overlapping negative information in the credit files that were affected largely compensated for the removal of the tax liens and civil judgments.

As details of the NCAP emerge, it is possible that the decision will be made to retain some of these data in the credit files, and the negative impact on model performance could be even less than that observed under the “maximum impact” scenario tested by VantageScore.