Squires and Velez 65 The substantive focus of most research has been on the outcome of insurance industry practices and the range of social costs associated with insurance redlining. Biases in the industry's perception and treatment of urban communities, particularly those with large minority populations, have been widely documented.13 The high price or unavailability of adequate insurance has been linked with greater difficulties in purchasing and maintaining homes or businesses, exacerbating disinvestment and contributing to the decline of urban communities.' And insurance redlining has often been found to constitute a violation of federal and state fair housing laws and other civil rights and unfair trade practices acts. 15 Much has been learned, though much debate remains, regarding the conditions that give rise to redlining and the effects of such practices. The focus of this research is on the immediate practice of redlining; the communication that occurs when residents call agents and ask for an insurance policy. This study examines the process of discrimination on the part of property/casualty insurance companies. TESTING FOR RACIAL DISCRIMINATION-DATA AND A recent study of the distribution of homeowners' policies written by major insurers in Milwaukee County found a strong and statistically significant bias against inner city minority communities and for white suburban areas. In that study the number of homeowners' policies written in Milwaukee neighborhoods was associated with the racial composition of those communities even after the effects of family income, poverty level, age of housing, and rate of residential turnover were taken into consideration. In recent years the Metropolitan Milwaukee Fair Housing Council has received several complaints from residents charging insurance companies with racial discrimination when their policies were not renewed, despite good claims records." In light of these developments the fair housing council conducted a test of three large insurance companies between the months of March and May, 1986.18 16 Eight teams of testers, individuals who would pose as homeowners interested in purchasing homeowners' insurance, were trained. Four teams had two members and four teams had three members. Within each team one member represented a predominantly white neighborhhood and one or two people represented an integrated or predominantly minority area. (Responses on virtually all items to callers from integrated or predominantly minority areas were not significantly different. Therefore, 66 The Review of Black Political Economy / Winter 1988 they will be treated as one category.) Within each team all members were matched in terms of income, occupational status, construction and value of home, claims experience, and other personal, property, and neighborhood characteristics that affect an individual's insurability. The only difference among team members was the racial composition of the neighborhood in which they purported to reside. Obviously, identical matches in terms of the insurability of testers were virtually impossible to secure. Where differences did occur, however, the member from the predominantly white neighborhood would have the slightly less attractive objective characteristics. For example, where differences in the age of the home occurred, it would be the caller from the predominantly white neighborhood who lived in the relatively older home. Or if incomes differed, the tester from the white neighborhood would report the lower income. Such a procedure, of course, provides a built-in bias against a finding of discrimination on the basis of the racial composition of the neighborhood. Any finding of such discrimination, therefore, would constitute an underestimate of the actual extent of discrimination that may prevail. Each team member contacted the three insurance companies to request information on how to obtain a homeowners insurance policy. A total of 60 tests, 24 from white communities and 36 from nonwhite or integrated areas, were completed. During each test call, testers took detailed notes and then completed a questionnaire on the contents of the phone call. The objective was to determine whether or not insurers treat individuals differently on the basis of the racial composition of the neighborhood in which they reside. FINDINGS The major conclusion to be drawn from these tests is that agents exhibited more eagerness to sell insurance policies to testers representing white neighborhoods and placed more hurdles in the paths of callers from integrated or minority areas. Equally significant, however, was the absence of blatant discrimination or outright refusal to offer policies in neighborhoods of any particular racial composition. The most consistent finding of these tests was the agents' interest in geographic location. Each agent obtained the street location of testers and in 93.3 percent of the tests the first, or one of the first, questions asked by the agent was the specific address or zip code of the caller. Neighborhood clearly remains a central consideration in the sale of property insurance. Agents frequently asked testers for information on the age, value, con struction, size, and other characteristics of the home. They also asked some questions about the employment and income of family members and previous claims history of the homeowner. The solicitation of such information, however, was not associated with the racial composition of the neighborhoods represented by the callers. But there were significant differences in the frequency with which certain questions were asked among callers from white and nonwhite areas that indicated a preference to pursue the former as customers and to place additional barriers before the latter. First, agents requested the names of the callers from white areas in 91.6 percent of the cases but in only 77.8 percent of the cases of callers from nonwhite areas (see Table 1). The simple act of asking for the potential customer's name has been found to be a significant factor in sales agents' efforts to create a warm sales climate, particularly by real estate agents, in fair housing research.1 Soliciting a person's name is frequently an indication of a desire to develop a friendly, trusting relationship for the purpose of ultimately closing the sale. Not asking a name is certainly no confirmation that the agent would not sell the product. But the association between racial composition of neighborhood and the frequency with which the name was requested reveals a preference for where this group of insurance agents would like to do business. This preference is revealed further by the frequency with which agents asked callers for information about their current policies. Callers from 68 The Review of Black Political Economy / Winter 1988 TABLE 2 Agent Request of Current Policy Information By Area Racial Composition white areas were asked detailed information about their current homeowners' policies in 91.7 percent of their conversations with agents compared to 66.7 percent for testers in nonwhite neighborhoods (see Table 2). Knowledge of the current policy (e.g. its price, coverage, terms) helps agents know what they have to offer to increase their chances of getting the business. Again, failure to solicit such information does not mean the agent would refuse to sell a policy. But the racial disparity on this item reinforces the notion that agents more actively pursued and were more eager to sell policies in white areas. Another indication of the racial preferences of these agents is their utilization of inspections. A majority (58.3 percent) of testers representing white neighborhoods were offered policies without any inspection. Only 47.2 percent of testers from nonwhite communities were offered a similar option. While the relationship between area racial composition and inspection requirement was in the anticipated direction, it was not statistically significant. What proved to be significant, however, was the difference in the type of inspection required of testers in white and nonwhite areas when the agent indicated that an inspection would be necessary. The most frequent type of inspection involves a drive by the house and a photograph of the exterior, often referred to in the industry as a "drive by and Polaroid." The basic objectives of such an inspection are to verify that the house exists on the property identified by the applicant during the initial phone conversation and to confirm its value. According to William Sirola, Chicago regional manager of the Insurance Information Institute, such an inspection might cause the agent to change the coverage, price, or terms of a policy but would not result in the denial of a policy previously offered over the phone.20 In some cases, however, a more detailed inspection of the exterior and occasionally even the interior of a home was required of the Milwaukee testers. That examination as well as a photograph would become part of the inspection report. Among those testers for whom an inspection was required by the agents, 44.4 percent of those representing white communities received the "drive by and Polaroid." Only 15.8 percent of the properties that needed to be inspected in the nonwhite areas received such a light inspection. Conversely, 84.2 percent of the testers in the nonwhite areas requiring an inspection were told that application for insurance of their properties would have to include the more rigorous inspection along with the photograph. Just 55.6 percent of the testers in white areas whose properties needed to be inspected were required to undergo the heavier inspection (see Table 3). Once again, requiring an inspection, or a more rigorous inspection, is not tantamount to a refusal to insure. But given the association between inspection policies and neighborhood racial composition, it is clear that testers from nonwhite communities had at least some additional hurdles to overcome in order to be insured. These inspections were not completed as part of this test. Therefore it cannot be determined whether such in |