Management is considering increasing our bad debt reserves because of the growth in our unsecured credit portfolio, with our credit card portfolio being the most significant portion. They’re questioning the age of our models and validation results that drive our bad debt projections, based in part on our Basel II calculations, since our probability of default is driven by score. What would be your recommendation on model redevelopment timeframes?
What a great question! Credit scoring model lifecycles have been the subject of considerable research. Kind of like 4 out of 5 dentists recommend a specific brand of gum, most experts agree that models have a two-to-five year lifespan. That can vary based on the volatility of the market.
At the time of the research study, we found the models of stable markets like Australia (g’day mate) validated extremely well even after seven years. “How?” you ask. The market didn’t change. Credit products remained similar. Various economic factors, such as income, were not in the models. On the other hand, the models of rapidly changing markets like China (nǐ hǎo péng yǒu) must be replaced in less than two years. By the way, “g’day mate” and “nǐ hǎo péngyǒu” both mean “hello, friend.”
Nowadays, many companies have adopted an automatic bi-annual replacement process to maintain a score-to-risk relationship with which they feel comfortable. “Oh wise one, how do they do this?” you ask. Simple – today’s technology allows them to implement moving sample databases and gives them the ability to modify existing and implement new scoring models very quickly. Their score-to-risk relationships developed bi-annually drive both their Basel calculations as well as their credit strategies. Now, model development timeframes are less of an issue because of the ability to maintain score-to-odds relationships across models and time periods.
Here’s the “fly in the ointment” as they say. Building and installing new models doesn’t come without costs. Those costs can be significant, especially if you contract with outside resources. If that is true in your company’s case, you will first need to look closely at your validation process. Then you will need to determine the best tradeoff between model deterioration and the cost of rebuilding and reinstalling your model.
Making this determination for your company may be a fairly simple process. Yes, Gini likes to look on the bright side of things. Given your management’s initiative to increase your reserve for bad debt, which has an immediate bottom line effect, you can easily determine the benefits of bringing your models back to 100%, then fund the development rather than increase bad debt expense.
Have a question for Gini?
Please send your question by using the form below:
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.