The Top Five Predictors of Subprime Risk

Better assess subprime opportunities by looking at data and attributes that are often more predictive than traditional credit scores

Most of the conversation around automotive finance is currently focused on the growth of originations to consumers with subprime credit scores, but there is more remarkable growth in orginations made to consumers who do not have a credit score at all.

As seen in Chart 1, subprime originations (designated by credit scores between 550 and 619) increased 2.16 percent from 2013 to 2014. Growth was even higher in the deep subprime segment (designated by credit scores below 550), with originations increasing 2.9 percent from 2013 to 2014. Yet the group that grew the most were consumers with no score at all, with originations growing 7.89 percent from 2013 to 2014.

Chart 1

These originations are not only growing in number, but also performing quite well. Chart 2 displays subprime auto delinquency rates from 2006 to 2015. In the last five years, both the number of delinquent subprime accounts and the amount of balances owed have decreased overall, with that trend looking to continue in 2015.

Chart 2

It would seem to be common sense that lenders would avoid consumers who do not have a traditional credit score. After all, it should be difficult to accurately gauge an individual’s financial situation without that three-digit number. So how can subprime originations be growing and performing well, especially among consumers with no credit scores?

The answer is that lenders are starting to leverage non-traditional financial attributes that are often more predictive for the subprime segment as well as consumers without a traditional credit history. In the past, these attributes were used anecdotally and  reliant on information that consumers  shared willingly with lenders. Moreover, it took time for consumers to hunt for their latest pay stub to prove they currently had a job and  stable income. This ultimately led to delayed or  derailed sales opportunities, which are lose-lose situations for all parties involved. Now lenders have access to alternative risk scores and databases of comprehensive financial information.

Many of these emerging databases are more than a simple pooling of data sourced from different companies and public records, with data providers and consumer reporting agencies going a step further to generate state-of-the-art risk models to analyze information about subprime borrowers. These models are the result of analyzing financial characteristics that have been prioritized by statistical algorithms. Using these databases and algorithms can reveal that different individuals who have the same subprime credit scores may actually have entirely different financial situations.

For example, two borrowers applying for an automotive loan could have the same subprime credit score despite a glaring difference – one of them has recently filed for bankruptcy, while the other has no reported bankruptcies. Additionally, the second individual with no bankrupticies has also established a steady track record of paying off cell phone and utility bills every month. Some lenders would deny both of these applicants based on their subprime credit scores alone. However, a closer look at that second individual  reveals a person who may be  more likely to stay current on an auto loan.

A growing number of lenders are looking at these alternative attributes to find  subprime borrowers similar to the second individual in the example above – individuals who are rebuilding their credit history after hard times to demonstrate they are more likely to remain current on an auto loan. These alternative databases can be a goldmine of information, and lenders may be surprised at which financial attributes are the most predictive at assessing the risk of a potential borrower. Some of the most important financial attributes identified by these databases and algorithms include:

• Size of Delinquent Telco and Utility Balances: Individuals having larger telecommunications or utility balances tend to be a greater risk for auto lenders. This is particularly true for Thin File individuals or those with a bankruptcy on file
• Presence of an Involuntary Disconnection: Individuals who have had their utilities, cell phones, cable service or other telco or utility service disconnected due to nonpayment represent greater risk for lenders.
• Number of Address Changes: Individuals who have changed their physical address multiple times represent greater risk for lenders.

These attributes are only the tip of the iceberg – there is a wealth of alternative data that can provide lenders with the insight they need to formulate a more comprehensive evaluation of consumers in the subprime market. With automotive sales remaining strong, lenders can use these resources to quickly and efficiently assess applicants, communicate with their partners and help close more deals. A version of this article ran in the July-August edition of Non-Prime Times.