Embracing Technology to Assess Risk and Improve Your Competitive Edge

Have you ever bought a new car and, like me, just hopped in and started driving without reading the owner’s manual to discover what features and options the vehicle had?  Months, and maybe years, go by without knowing you have some really cool features.  Or maybe, you have a feature you know about, but you don’t use it because you don’t fully understand the benefit it provides.

Back in the mid-eighties, financial institutions began to price loans based on risk, but regulations related to equal credit opportunity required lenders to assess risk using an empirically derived model.  Well, 25-30 years ago, most of us had did not have computers on our desk to allow us to access data and run regression analysis on large data sets to create these models.  So, in order to have an empirical method for establishing risk, lenders used risk scoring systems offered by the credit bureaus.  These models have existed for a number of years, and have been improved over time, but sole reliance on these models for credit and pricing decisions is an archaic portfolio management method in today’s marketplace.

The good news is, most lenders, regardless of size, now have access to more data and technology that will enable them to complement credit scores in deriving a more precise scoring method for making credit decisions and pricing loans.  It may be time to discover what you are capable of doing with information and data you already own.  In fact, your own application and loan performance data can tell you more about how future loans will perform than even the credit score.

Recently, CU Direct’s Advisory Services conducted a study on a single credit union, and found that there are six independent variables, besides the credit score, that can be used to predict future loan performance. These variables include loan-to-values, credit score migration, term and whether or not the loan had a co-borrower.  Now, look what wasn’t identified as a predictor.  Debt-to-income, whether a vehicle was new or used at origination, and other common factors used in credit decision and pricing were not found to be significant predictors.

Be careful, though; do not interpret these results to be common to all lenders.  Using generic decision and pricing models, like FICO scores alone, is no longer effective for properly assessing risk and can leave a lender at a competitive disadvantage, as other lenders move to more sophisticated and precise models that allow them to offer credit to more members, and to a broader sector of the population.  Pause here for just a moment and look at the rate sheet used by your organization for pricing loans.  Most likely, you have a rate sheet that uses credit score (risk tier) and term as the primary risk factors for pricing a loan.  However, term is not as significant in predicting default as loan-to-values in most cases.  One might be able to say that, historically,  there is an interest rate risk associated with longer term loans, but can one say that in the current rate environment where rates are not rising?

Every lender should look at their decision and pricing models and ask a simple question: Why?  Why are the predictors that are used, being used, and what is their level of significance?  If you can’t answer these questions confidently, I’d encourage you to get the answers you need to ensure your credit union is maximizing its portfolio performance.  CU Direct’s Advisory Services can analyze your data and provide you with insights into how you can improve your competitive edge.  Not using the technology that is available today is like insisting on manually locking your car doors, even though there are power door lock buttons in every car.