ship between pollution potential and all measures of toxic releases. This finding, again, is expected given the fact that we are examining toxic releases, and the construction of this measure equates with existing and operating manufacturing capacity. Environmental interests are also related to all three measures of toxic releases in the hypothesized direction. State-level environmental interests are concerned with maximizing environmental quality, and the size of these interests throughout the state appears to lower the level of toxic releases in the substate regions. The state business climate, in this phase of the analysis, increases the level of toxic releases. Most probably, state-level actions to encourage and support economic development are having an effect at the substate level, and this effect is to minimize concerns about environmental protection--leading to higher levels of toxic releases. Apparently, the substate regions have little control over either the benefits the state awards to business or to the consequences of those benefits--increased pollution.
Although legislative professionalism exerts the hypothesized effect on the level of global toxic releases--that is, a professional legislature decreases this level of environmental harm--its effect disappears when dealing with the specific air-pollution measures of toxic releases. Partisanship also evidences a marginal effect on two measures of toxic releases. Although Republican partisanship is marginally related to decreased levels of state air toxic releases, it is positively related to fugitive air toxic releases. Finally, large counties appear to have higher global TRI releases and stack air toxic releases.
This chapter has assessed the environmental racism hypothesis with regard to three measures of toxic pollutants within roughly 70 percent of U.S. counties as of 1995. In both instances, our simple tripartite model produced misleading results when compared to our more complex analysis. Indeed, in all instances, the results from our more complex model provided evidence of complicated relationships between our measures of race and toxic releases. To aid the reader, our findings from the reduced complex test are summarized in Figure 6.2.
Overall, our findings do tend to confirm that our two minority classifications were exposed to higher levels of toxic releases; and insofar as other research has reported similar, but less complicated, results, our findings are compatible ( Burke, 1993; Downey, 1998; Gould, 1986; Lester and Allen, 1996; Lester, Allen, and Lauer, 1994; Polloch and Vittas, 1995; Ringquist, 1997; Szasz, 1994). However, our results also go beyond simple formulations of the race-risk nexus. Instead of simple linear relationships between the percent black population and toxic releases, we found that this relationship was considerably stronger within the Sunbelt. Furthermore, the rela