Racial residential segregation has become an increasingly important factor in research on variation in levels of formal social control. This is particularly true of research from the conflict perspective, and especially in tests of the effect of perceived racial threat on city-level variation in police size, police expenditures, and arrest rates. The theoretical expectation is that higher levels of segregation will reduce levels of racial threat among whites, which subsequently reduces the need for crime control. Empirical findings are mixed, but these inconsistent results may be partly due to an improper conceptualization of segregation vis-a-vis the racial threat perspective. Despite theoretical emphasis on the concept of exposure, all prior studies have employed a measure of segregation that, instead, reflects the extent to which groups are equally dispersed across neighborhoods. This measure is not necessarily related to exposure either conceptually or empirically. The present study utilizes a novel measure of racial residential exposure, developed with spatial analytic methods, to provide a more conceptually true test of this aspect of the racial threat hypothesis. Specifically, we examine the effect of whites' aggregate levels of residential exposure to blacks and Hispanics on variation in police expenditures across U.S. cities. ..PAT.-Unpublished Manuscript [ABSTRACT FROM AUTHOR]