Driver Negligence vs. Odometer Miles: Rival Theories to Explain 12 Predictors of Auto Insurance Claims

This paper challenges accepted explanations for well-established correlations. The correlations allow insurers to assign cars to groups that predict accident and claim frequencies per 100 insured car years. Correlations with driver age and regional geography can be explained by variations in accident involvements per million miles established by engineering studies and are not considered in this research. Instead this paper examines a dozen predictor variables that have no obvious connection to physical risk. Previous studies focus on one or two predictors. For example, driver sex is used by Brockett, Golden, and Dunn (2005) as a benchmark predictor in seeking to explain why lower credit scores predict higher accident frequencies. My research includes these two predictors—driver sex and credit score—with 10 others to test rival theories for their ability to provide causal explanations and avoid conflict with established economic theory.