Even as other parts of this research series address questions of the appropriateness of age discrimination in our legislation and public polices, a separate question is the use of age by private entities. This report considers how age is used in the selling of life, individual health, and automobile insurance to Canadians. The first finding is that age is a statistically reliable classification variable for all three lines of personal insurance considered and its use can be practically implemented. Therefore, the primary concerns about the use of age as a classification and/or underwriting variable centre around the issue of socially acceptability. Specifically, is there an intuitive and causal link between the classification variable and the risk that it is measuring. While age carries a strong intuitive causal link for life and health insurance, that is not the case with driving ability. For that reason this report focuses on issues of age and driving which immediately become the issues of age and auto insurance. Age is associated with conditions, such as inexperience and immaturity for young drivers and restricted mobility and decreased cognitive ability for older drivers, which affects driving ability. In Canada, the frequency and severity of auto accidents are highly correlated with age. Young drivers have the greatest frequency of accidents, but when adjusted for distance traveled, seniors over the age seventy-five have the worst accident rate per kilometer driven. The result is a distinctive U-shape when accident history is graphed against age. However, heterogeneity in driving abilities for both younger and older ages emphasizes that this relationship is one of correlation not causation. But could insurers be using a variable that captures a stronger (causal) relationship between the classification variable and driver risk? Or that captures the same level of correlation as age? To answer these questions, we explored a range of possible alternatives to age. The variables examined are number of years licensed, driving record, safe driving discounts for youth, insurance credit scores, and distance driven. The performance of these variables and any disproportionate impact they might have on young and elderly drivers are discussed. None of these variables, either individually or in combination, captures a driver‘s risk with the same degree of accuracy as can be achieved using age. The insurance score performs the best, but it too is controversial and may ultimately be found to be discriminatory by regulators and consumer groups. For these reasons, though reliance upon chronological age as may be reduced somewhat if provincial licensing bodies can better test fitness to drive and restrict those drivers that are less fit, sufficient justification exists for its the continued use in the rating and underwriting of insurance. (Author/publisher)
Samenvatting