There are many success stories featuring predictive models, but what does not get as widely reported are the failures: mistakes that range from subtle misinterpretations and minor miscues to unvarnished disasters.
Executive SummaryIn the third in a series of articles about the pitfalls of predictive modeling, Ira Robbin presents six hypothetical scenarios of models that fail and explains the reasons why. For the complete report, download the entire whitepaper, "Predictive Modeling Pitfalls Whitepaper: When to Be Cautious."
This three-part series focuses on the use of predictive models in property/casualty insurance and illustrates several pitfalls. Many of the pitfalls have nothing to do with technical aspects of model construction but rather with false assumptions about the data or misapplications of model results.
In Part 3, we present six hypothetical pitfall scenarios.
Pitfall 1: Thinking a Predictive Model Predicts the Future
Joe, the chief pricing actuary of a medium-sized personal lines writer, laid off a few of his conventional actuaries and hired some statistical modelers. They developed an excellent predictive model that was used to derive new relativities for private passenger auto customers.