I was shocked to read this report provided by the CFPB’s Consumer Credit Panel. There is a lot of information but one caught me by surprise. Did you know that our scores could be pre-determined by creditor scoring models.
Here’s an example: You are a single mom, that works 1 low wage job with 2 children. Did you know that your credit score could be based on the average of outcome of a single mom with 2 children and 1 low wage job is. If the computer model predicts that other single mom’s that fit this category usually fall behind on their credit card payments, car payments, utility bills, etc. Creditors will develop a “Class” for that profile and they will score your credit using that model, NOT your actual true credit scores.
Below is the complete study is you’d like to read it.
Data Sources The data used in this study come from three sources.
The first is the CFPB’s Consumer Credit Panel (CCP), a longitudinal sample of approximately 5 million de-identified credit records that is nationally representative of the credit records maintained by one of the NCRAs. This study primarily uses data from December 2010; however, as described below, we also use information for these same consumers from December 2014 in cleaning the data. For each time period, the entire credit record is supplied in the CCP, excluding any direct identifying personal information (such as name, address, or Social Security Number). In addition to the credit records, the CCP includes a commercially-available credit score, which we use to indicate which records were scored and which were not.
For each unscored record, an “exclusion code” is provided indicating why the record could not be scored using the model for the commercially-available credit score. Like most credit scoring models, the model that generated the scores in the CCP was built to predict future credit performance (that is, the likelihood, relative to other borrowers, that a consumer will become 90 or more days past due on a credit obligation in the following two years).4 In some cases, the model builders will determine that a credit record does not contain enough information to make a suitably reliable prediction. During score development, these records are excluded and are unscored by the model going forward.
CLICK HERE TO SEE YOUR CREDIT REPORTS AND SCORES.