Minority Move-in, and Environmental Justice

Previous research suggests that minority residential areas have a disproportionate likelihood of hosting various envirotnmental hazards. Some critics have responded that the contemporary correlation of race and hazards may reflect post-siting minority movein, perhaps because of a risk effect on housing costs, rather than discrimination in siting. This article examines the disproportionate siting and minority move-in hypotheses in Los Angeles County by reconc iling tract geography and data over three decades with firm-level intbrmation on the initial siting dates for toxic storage and disposalfacilities. Using simple t-tests, logit analysis, and a novel simultaneous model, we find that disproportionate siting matters more than disproportionate minoritv move-in in the sample area. Racial transition is also an important predictor of siting, suggesting a role for multiracial organizing in resisting new facilities. In recent years, policy makers have become increasingly responsive to the perception of racially inequitable exposure to various environmental hazards concerns. As early as 1994, a Presidential Executive Order directed all federal agencies to take into account the potentially disproportionate burdens of pollution or hazards existing in US minority communities. In 1998, the Southern California Air Quality Management District-charged with cleaning up the country's dirtiest air-decided, under pressure from grassroots activists, to create its own task force on environmental justice. One year later, the California legislature adopted a law directing the state's Office of Policy Research to develop environmental justice guidelines for Califomia's *Direct correspondence to: Manuel Pastor, Latin American and Latino Studies, University of Califbrnia, Santa Ctruz, Casa l atina, Merrill College, Santa Cruz, CA 95064. E-mail: [email protected] JOURNAL OF URBAN AFFAIRS, Volume 23, Number 1, pages 1-21. Copyright ©D 2001 Urban Affairs Association All rights of reproduction in any form reserved. ISSN: 0735-2166. 2 1 JOURNAL OF URBAN AFFAIRS I Vol. 23/No. 1/2001 various agencies, forcing a scramble among policy makers to better define both the problem and appropriate remedies. Despite the ongoing response at the policy level, the research on disproportionate exposure by race has yielded mixed results. Making use of simple cross-tabulations, basic correlation analysis, and case studies, the earliest work in this field found that minority neighborhoods hosted a disproportionate share of the environmental hazards and toxins produced by an industrialized society (Bullard, 1990; UCC, 1987). Subsequently, some researchers found that race was not a significant factor when controlling for income, employee proximity, and other reasonable variables (Anderton, Anderson, Oakes, & Fraser, 1994: Anderton, Anderson, Rossi, et al., 1994). However, a more recent wave of research, also controlling for other explanatory factors, has tended to confirm the racial disproportionality hypothesis (see, for example, Been, 1995). Virtually all of this research has amounted to a "snapshot in time" of the distribution of environmental hazards. Recognizing where hazards are and whom they might affect is of immediate utility to those public officials calculating health risks, planning emergency measures, or seeking to redevelop contaminated land. But such a cross-section analysis does not fully address a question of central concern to policy makers: Were the hazards disproportionately sited in minority communities or did minority residents move in after hazards were sited? The debate between the "disproportionate siting" and "minority move-in" hypotheses matters greatly for urban and environmental policy. If the problem of disproportionate exposure by race is due to siting, then it would be appropriate for policy makers to revise zoning and permitting procedures to eliminate any elements of discrimination. But suppose the pattern emerges because the siting of hazards detracts from neighborhood livability and thereby diminishes land values, inducing an exodus of middle-class (often Anglo) homeowners and an influx of lower-class (often minority) residents. While health precautions would still call for buffers between industrial and residential uses as well as other safeguards, the notion that the process is market-driven may lead some to suggest that individuals are simply choosing to trade increased neighborhood health risks for slightly larger or better (in other ways) housing. The role for policy in this view might be confined to: (1) ensuring access to data about neighborhood health risks so that individuals who choose to trade risk for affordable housing are not acting on incomplete information (see Burby & Strong, 1997), and (2) continuing the enforcement of existing statutes that limit the steering of minority house-seekers to particular neighborhoods. Indeed, if information is complete and housing discrimination is limited, then some might argue that there is little reason to be concerned about a contemporary pattern of disproportionate exposure: after all, market dynamics suggest that those neighborhoods with hazards will eventually become predominantly minority anyway. Is the current pattern of environmental inequity a field of bad dreams: Build it and minorities will come? This article contributes to disentangling the role of disproportionate siting and minority move-in with a study of the temporal dynamics in Los Angeles County, a region where there is clear evidence of disproportionate contemporary exposure to toxic storage and disposal facilities (TSDFs), toxic air releases, and other environmental negatives (Boer, Pastor, Sadd, & Snyder, 1997; Burke, 1993; Sadd, Pastor, Boer, & Snyder 1999; Szasz, Meuser, Aronson, & Fukurai, 1993). We focus on TSDFs, facilities that operate under a U.S. EPA permit to store hazardous wastes (any non-petroleum substance which is ignitable, corrosive, reactive, or toxic) as defined in the Resource Conservation and Recovery Act of 1976. Most TSDFs are private, for-profit businesses that accept waste from other generating facilities. We link the siting dates and addresses of Los Angeles' high-capacity TSDFs (those which handle more than 50 tons a year) to a database that tracks changes in selected socio-economic variables through the period 1970 to 1990, with all data geographically indexed to the 1990 . Which Came First? 1 3 census tract shapes. We subject the resulting data to a variety of tests. including logistic regressions to predict future siting and a simultaneous model that accounts for both minority move-in and disproportionate siting. The results indicate that disproportionate siting matters more than minority move-in within the sample area. The results also suggest that areas undergoing ethnic transition may be as vulnerable to siting as areas with older or more established minority populations. This finding reinforces the activist argument that residents should organize on a multiracial basis to resist increased exposure to environmental hazards. The article proceeds as follows. The first section reviews previous studies and outlines our approach. The second section discusses basic trends in the data. The third section offers logitstyle regressions that attempt to predict the likelihood that a hazardous site will be located in a particular area. The fourth section tests for the possibility of minority-move-in, both after the siting and during the period of siting. The final section concludes with possible lessons for both policy makers and activists. LITERATURE REVIEW AND METHODOLOGY There is now a burgeoning literature examining the pattern of contemporary location of environmental hazards (see the extensive review in Szasz & Meuser, 1997). While the evidence is often more mixed than many activists have believed, the bulk of the research does seem to point to disproportionate exposure to hazards in minority communities. The most recent work about California is strongly supportive of disproportionality in the Golden State. MorelloFrosch (1997). for example, focuses on hazardous air pollutants (HlAP) at the county and census tract level and finds a consistent association between the percentage of minorities and both HAP concentrations and estimated likelihoods of pollutant-related cancer risk. Our own previous work (Boer et al., 1997; Sadd et al., 1999) explores the distribution of hazardous waste storage and disposal facilities (TSDF) and toxic air releases in Southern California and finds strong evidence of a racial pattern, even when controlling for reasonable variables such as land use, manufacturing employment, and income. Many have assumed that contemporary inequity is the result of discriminatory siting practices. The general argument is that low levels of political power in minority communities may induce polluters to locate hazards in these areas (Hamilton, 1995). Such a political argument is often implicitly based on notions of social capital and community efficacy: Where residents have more ability to organize and affect policy, perhaps because of their income or racial status in a stratified society, they will be more able to resist the placement of a hazardous facility. Of course, social capital may in fact be affected by other factors, such as the level of education of residents or the ability to bridge differences between minority groups, a topic we explore below (see also Briggs, 1998; Temkin & Rohe, 1998). An alternative argument suggests that disproportionate exposure simply reflects the market: Both minorities and undesirable land uses will be attracted to areas with lower housing values, and in fact, minorities may move in after the arrival of a new locally undesirable land use (Been & Gupta, 1997). In our view, this market-based account of minority move-in is unlikely, at least in Southern California; after all, if race still matters when income is held constant in a cross-section regression, then any disproportionate move-in of ethnic residents would seem to reflect different consumer tastes for exposure to this type of risk. In fact, however, one survey suggests that minority residents may be even more concerned about environmental risk, particularly in the contemporary period in which environmental justice has become a key organizing buzzword in selected communities (Burby & Strong, 1997). Still, the minority move-in argument persists and a more sophisticated version can incorporate the potential role of housing discrimination in limiting the locational opportunities for minorities. 4 1 JOURNAL OF URBAN AFFAIRS I Vol. 23/No. 1/2001 Despite the importance of the issue, there is very little solid research on the dynamics of disproportionate siting versus minority move-in. Yandle and Burton (1996) provided an early longitudinal look at hazardous landfills in metropolitan areas in Texas but their work has been sharply cnrticized on methodological grounds (see, for example. Anderton, 1996; Mohai, 1996). More recently, Shaikh and Loomis (1999) looked at the decadal percentage change in minorities in Denver neighborhoods after the siting of a stationary source of air pollution. Not only did they find no evidence of minority move-in but they also found some evidence suggesting that communities without polluting sites experienced larger increases in the percentage of minorities; however, the areal units in their study are zip codes, a less uniform geography which has largely been eschewed in favor of census tracts in most recent research efforts. Thus, the two most significant and reputable longitudinal studies are Oakes, Anderton, & Anderson (1996) and Been & Gupta (1997). The Oakes, et al. (1996) study uses the 1992 Environmental Services Directory to determine beginning dates for commercial TSDFs nationwide. Comparing tracts that received TSDFs over the 1970 to 1990 period to the rest of the county, the authors found no evidence of either disproportionate siting by race-or of a subsequent mnove-in of minorities that exceeds the pattern for areas with similar industrial characteristics. They then conducted a more formal multivariate regression analysis on the TSDF tracts and a stratified sample of non-TSDF tracts: neither race nor poverty was significant and the only variable with real predictive power was the percentage of local residents involved in industrial employment. As for post-siting changes, a TSDF tends to have a negative (not positive) impact on African American or Latino in-migration, albeit at an insignificant level. Thus, in both the simple comparisons and the multivariate setting, neither disproportionate siting nor minority move-in are shown to exist. In a similar nationwide longitudinal study, Been and Gupta (1997) arrived at slightly different results. Like Oakes et al. (1996). they used a national sample and conducted multivariate regressions on tracts that received TSDFs and a stratified sample of tracts that did not. They also found no evidence for a market dynamics story of minority move-in subsequent to the siting of a TSDF. However, they did find that the percentage of Latinos had a significant impact on the likelihood of receiving a TSDF (as did the percentage of local industrial employees and population density). While this overall pattern of results tends to offer some weak support to the usual claims of environmental justice proponents, there is no evidence that the percentage of African American residents has an impact on siting and the percentage of residents in poverty is actually found to have a negative impact on 1 980s sitings. Our own approach involves several modifications from the previous studies. First, we look only at one region, Los Angeles County. This limited geographic scope is partly due to our view that the nature of hazards is related to the industrial clusters of a region-Los Angeles's furniture making and metal plating industries are not likely to drift north to Seattle, and Microsoft is not likely to move south to Los Angeles-so it is the distribution of hazards within a region that matters. Logistically, this regional focus also allowed us to obtain siting information from original business records and permit applications, as well as to accurately locate and verify each TSDF by conducting visits to actual facility locations. Focusing on one region also allowed us to employ a Califomia Department of Finance (DOF) database that allocates certain variables from the 1970 and 1980 censuses, including ethnicity, to the 1990s tract boundaries. Therefore, we could consider all host and non-host tracts rather than a stratified sample as in earlier work. A second difference is our use of geocoded site location and GIS procedures to determine affected tracts. Both Been and Gupta (1997) and Oakes, et al. (1996) focused on the demographic characteristics of tracts that contained TSDFs. Yet as Anderton, Anderson, Rossi. et al. (1994) and Anderton, Anderson, Oakes, et al. (1994) pointed out, TSDFs are often located J Which Came First? 1 5 near a tract boundary and a simple tagging of only the host tract will ignore the impact on immediately adjacent neighborhoods. Therefore we pinpointed the actual facility and used a circular buffer distance of one-quarter mile and one mile to define the potentially affected tracts and residential population. As a result of this procedure, there are slightly more affected tracts than there are TSDFs at the one-quarter mile level and, of course, even more affected tracts when we extend out to the one-mile circle. Third, we go beyond previous work in considering the post-siting dynamics. With little theoretical justification, other researchers have tended to employ the same variable set to predict move-in as they did to predict siting; we instead nest our analysis of post-siting effects in a simple model of neighborhood demographic change. We also advance the field by constructing and testing a simultaneous (or two-stage least squares) model. After all, disproportionate siting and minority move-in often occur at the same time and a regression strategy that accounts for this may be the best way to estimate the separate effects. Fourth, we focus on the effects of a new dimension of ethnic change. Previous work has stressed the percentage of minorities. But while a 40% increase in Latinos that is matched by a corresponding 40% decrease in African Americans may leave the percentage of minorities unchanged, the neighborhood will in fact be transformed. Such ethnic transitions may weaken the usual social bonds constituted by race and make an area more susceptible to siting. We investigate this "social capital" effect below, finding that it does indeed have an effect on the likelihood of receiving a TSDF. Before presenting the methods and results, we should acknowledge several clear limits to our work and that of others. One is that we are testing for effects at the neighborhood level. Such a focus on the social ecology of an area does not mean that particular subpopulations or individuals are necessarily exposed in the same rate as their census tract; actual exposure can vary depending on a variety of factors. This neighborhood effects approach. however, is characteristic of almost all environmental justice studies, primarily for reasons of data collection (for an exception based on an original survey, see Burby, 1999; Burby & Strong, 1997). In addition, some epidemiological studies have demonstrated a significant relationship between residential proximity to hazardous waste storage facilities and increased health risk and disease, especially among pregnant mothers and infants (Berry & Bove 1997; Croen, Shaw, Sanbonmatsu, Selvin, & Buffler, 1997; Goldman, Paigen, Magnant, & Highland, 1985: Knox & Gilman 1997; see also Shaw, Schulman, Frisch, Cummins, & Harris, 1992). A second related limit is that we do not really establish the actual risk associated with living near a TSDF. Once again, there are few efforts in the environmental justice literature that tackle actual risk; an exception is Morello-Frosch's effort to use public health methodologies to transform cumulative exposure to hazardous air pollutants into estimated cancer risk. However, few people believe that living near a TSDF enhances their quality of life and, as Burby and Strong (1997) argue, proximity to environmental negatives does seem to have a significant impact on perceptions of neighborhood quality. People are more likely to be alarmed about hazards when their sense is that they are being exposed involuntarily or that exposure is unfair. In short, the distribution of perceived risk and perceived fairness also matters. Finally, while this study was conducted with as much rigor as possible, it still requires all the qualifications necessarily associated with the statistical work in this field. Most specifically, associating race with siting decisions, even in multivariate exercises, may establish pattern but it does not establish intent. The real rationales for location will need to be uncovered by specific case studies, of which there are a few excellent examples (see Boone & Modarres, 1999; Pulido, 1996). This study simply offers a framework of plausibility for the more detailed and qualitative work ahead for other researchers. 6 I JOURNAL OF URBAN AFFAIRS T Vol. 23/No. 1/2001 DATA SET AND BASIC TRENDS The data set we use merges selected variables from the 1970, 1980, and 1990 censuses, all recalculated to the 1990 tract shapes, with information on the location and siting dates of TSDFs in Los Angeles County. Because many facilities obtained permits long after siting (partly because they were sited prior to current regulations or operated with interim permits), the recorded permit dates used in many studies are often inaccurate. To correct this, we filed a series of public records act requests to obtain the original forms identifying when any particular facility began operation. We focused on the high-capacity TSDFs-those that process or store at least 50 tons of hazardous substances annually. Although slightly less than half (39 of 83) of the TSDFs in the study area are classified as high capacity, these facilities handle nearly all of the hazardous waste among TSDFs in the region (644,136 of 644,511 total tons). Of these 39, our records search proved unable to identify the siting date of one of these facilities. Given the circular buffers, we ultimately examined 55 tracts in the quarter mile radius (for which there was at least 50 tons allocated to the tract when casting a circle of effect around the facility), and 245 tracts in the one-mile radius as of 1990, all within a county with 1,652 tracts. Figure 1 shows all TSDFs in the county; Figures 2 and 3 plot two possible date-location combinations for the high capacity TSDFs (existed prior to 1970. or was placed 1970 to 1990) against median household income and percentage of minorities in 1990. There is a definite visual correlation between these socio-economic variables and the contemporary location of high-capacity TSDFs. This association is confirmed in the t-tests shown in Table 1. The focus in this article is on the high-capacity TSDFs so all references in the statistical tables and the following discussion are to the high-capacity variety only. TSDF locations . Low Capacity (<50 tons/yr) A High Capacity (> 50 tonslyr) Los Angeles freeway system 1990 Census tract boundaries