Any Room Left for Skimming the Cream? [Discussion]

Mr. Bailey is to be commended for the excellent work that he is doing, as revealed by this and other papers of his, in bringing mathematical analysis to bear on the problems of rating systems. These problems are extremely difficult, and the final analytical solutions are still to be made, but every contribution, such as Mr. Bailey’s, is another step along the way. In this paper Mr. Bailey considers the problem of measuring the amount of skimmable cream to be found in the classification system for automobile liability insurance. He draws several conclusions from his analysis, his final conclusion being that the present rating system is not perfect and still has skimmable cream in it. No one will disagree about there still being cream in the rating system; Mr. Bailey, himself points out that perfection can only be achieved if there is a separate rate for each risk or, more precisely for each group of risks with the same accident-potential. As soon as you combine into one rate-class a group of risks with differing accident-potential, no matter how slight the difference, there will of necessity be some risks that are better than average: thus there will always be cream.