REU 2010 – Minimum Bias Approach to Determining Automobile Insurance Premiums

The goal of the project was to assist Hanover’s development of updated insurance premiums for various automobile insurance coverages within Massachusetts. Historical premiums and paid losses were trended to current levels, so that an apples-to-apples experience study could be performed. The experience was sorted by numerous driver/vehicle/coverage attributes in order to determine the best predictors of risk. A minimum bias approach was employed to achieve a “best fit” to historical data. Regulatory constraints were overlaid on these results, and from this, a set of “indicated premiums” was presented to Hanover’s management for their consideration.