I have a question regarding the relationship between policy or other types of rules used in the origination process, versus what elements that may also be used (or were considered) in the development of the scoring model in production. I am concerned that there may be some double counting of the same information or worse conflicting impacts of the same data elements. How can I connect the pieces so I can feel confident in how they work together?
Dear Double Duty,
You raise an issue that has been debated since the mass introduction of scoring into the credit process over 30 years ago! Lenders tend to think in terms of the individual case, and of course the scoring model is based upon the thousands of cases it considered in its sample database.
If a particular data element was considered and rejected because it did not have sufficient information value by itself, or it was highly correlated with other elements in the model, than using it de facto, adds no value. If you put weight on that same data element in your origination process you are placing it in a value system arbitrary. The impact of this will reduce your acceptance rates (these rules are almost always used to decline an applicant that passes the score cutoff) and past studies have shown that they rarely add value.
At the same time it is hard to accept an applicant who has a debt ratio approaching 100%. In a recent post, we looked at debt ratio in some detail and it is hard to convince management not to consider it. Fair enough, but take a hard look at the level the rule is set at. We have seen rules based upon a 40% mark. That may have been fine 5-10 years ago but with increasing housing and transportation costs it is hard to find applicants who can meet the test. It might be time to do a short study and see what your debt ratio distribution looks like and adjust your rule based upon those results. It is hard to defend a rule that is catching over half of your applicants.
Other rules may be divided into hard and soft sets. You must be 21 to legally sign a contract but any minimum income requirement is open to debate. Don’t forget to keep an eye on the number of cases you are turning down based upon any arbitrary test and how your overall override rates are trending. If they exceed 3-5% you are most likely turning down good business. So, in the case of double duty, study your distribution, take a close look at the rules and speak up if something isn’t working!
Ask Gini Terms
Content provided in this blog is for entertainment purposes only. Ask Gini blogs do not reflect the opinion of BankersLab. BankersLab makes no representations as to the accuracy or completeness of information in this blog. BankersLab is not liable for any errors or omissions in this information nor for the availability of this information. These terms and conditions of use are subject to change at anytime and without notice.