Analysis of Effects of Race on Insurance Companies' Operations in Chicago, December 1977 - November 1978

摘要

Dataset Chicago is about the residential fire insurance policies issued in Chicago over December 1977 through February 1978; categorized as either voluntary, or involuntary, which includes state-offered Fair Access to Insurance Requirement (FAIR) plans (Wriggins 2010) for people who would be denied insurance because of high risk classification, after Fair Housing Act (FHA) has been in act since 1968 (Calmore 1997). Data size is relatively small as n = 47 Zip codes. Theft, fire and the age of the house are also provided, along with the median income on the size of the expected loss and insolvency. Main purpose of this article is to explore the relationship between insurance activity and the variable race. S-Plus, rather than numerical methods such as risk theory models, was employed by using Data Plots to exam data itself, Regression Fitting to get possible candidates of linear expressions, and Model Selection to pick the best model under different scenarios. The final result proves insurance companies were using race as a determinative factor of underwriting insurance contacts, but not the only determinative factor.