摘要：This paper uses examples to demonstrate the generality of issues resulting from the heterogeneity of a population. Heterogeneity is fundamental to the marketing strategy of any firm. If one market segment is more expensive to service, then members of that segment might be charged higher prices or be left to one’s competitors. We pay particular attention to heterogeneity issues in financial services, specifically, insurance underwriting, lending, and investing. We examine techniques that practitioners use to assess heterogeneity in a population with respect to loss events. We employ the x2-goodness of fit test to assess whether observations in risk classes (sub partitions of population) are relatively homogenous. We use a sign-rank test to assess whether or not two loss distributions from different risk classes were drawn from the same probability distribution. We discuss how mutual ownership and price fixing have evolved as built-in safety nets to cope with modeling limitations.
Loading...
