Social Problems

, Vol. 54, Issue 3, pp. 343–370, ISSN 0037-7791, electronic ISSN 1533-8533.
© 2007 by Society for the Study of Social Problems, Inc. All rights reserved. Please direct all requests for permission to photo-
copy or reproduce article content through the University of California Press’s Rights and Permissions website at www.ucpress-
journals.com/reprintinfo/asp. DOI: 10.1525/sp.2007.54.3.343.
Racial Inequality in the Distribution
of Hazardous Waste: A National-Level
Reassessment

PAUL MOHAI,

University of Michigan

ROBIN SAHA,

University of Montana

National-level studies examining racial disparities around hazardous waste treatment, storage, and disposal
facilities have been very influential in defining the academic and political debates about the existence and impor-
tance of “environmental injustice.” However, these studies tend to employ methods that fail to adequately control
for proximity between environmentally hazardous sites and nearby residential populations. By using GIS and
applying methods increasingly used in environmental inequality research that better control for proximity, we con-
duct a comprehensive reassessment of racial inequality in the distribution of the nation’s hazardous waste facilities.
We compare the magnitude of racial disparities found with those of prior studies and test competing racial, eco-
nomic, and sociopolitical explanations for why such disparities exist. We find that the magnitude of racial dispari-
ties around hazardous waste facilities is much greater than what previous national studies have reported. We also
find these disparities persist even when controlling for economic and sociopolitical variables, suggesting that factors
uniquely associated with race, such as racial targeting, housing discrimination, or other race-related factors are
associated with the location of the nation’s hazardous waste facilities. We further conclude that the more recent
methods for controlling for proximity yield more consistent and definitive results than those used previously, and
therefore argue for their wider utilization in environmental inequality research. Keywords: environmental justice,
environmental inequality, environmental racism, racial inequality, hazardous waste, GIS.

Racial inequalities in life circumstances and outcomes have long been studied, including
inequalities in education, employment, income, housing, life satisfaction, poverty, health sta-
tus, and mortality (Beggs 1995; Eggebeen and Lichter 1991; Hayward et al. 2000; Hughes and
Thomas 1998; James and McCammon 1997; Jargowsky 1996; McCall 2001). Since the mid-
1980s, there has been increasing attention to racial inequalities in the distribution of environ-
mental quality. Attention to this form of racial inequality began as an “environmental justice”
movement emerged in the 1980s and 1990s to protest the placement of waste sites and pol-
luting industrial facilities in predominately African American and Latino communities (Bryant
and Mohai 1992; Bullard 1990; Cable and Benson 1993; Schlosberg 1999; Szasz 1995). The
impacts of this movement have been significant, spurring much public and academic dis-
course. Indeed, interest in examining the extent of social inequalities in the distribution of

This research was supported by grants from the Sociology Program and Geography and Regional Science Program of
the National Science Foundation (#0099123) and the University of Michigan State Policy Research Fund. The authors are
indebted to Professor Vicki Been of the New York University School of Law and her co-author, Francis Gupta, for providing
us the names and addresses of the hazardous waste treatment, storage, and disposal facilities used in their national envi-
ronmental justice studies. The authors also are grateful to Ken Guire of the Center for Statistical Consultation and Research
at the University of Michigan for his helpful suggestions regarding the statistical analyses for this article, Shannon Brines of
the School of Natural Resources and Environment’s GIS Lab for his assistance with performing spatial regression analyses,
and graduate research assistants Betsy Boatner Marsh, Sara Cohen, and Theresa Weber for their assistance compiling Cen-
sus data, verifying facility locations, and conducting GIS analyses. Direct correspondence to: Paul Mohai, School of Natural
Resources and Environment, University of Michigan, Ann Arbor, MI 48109-1041. E-mail: pmohai@umich.edu.

344

MOHAI/SAHA

environmental quality, their causes and consequences, and potential remedies has spread
rapidly in the past decade, not only in sociology but across a multitude of disciplines (Brown
1997; Freudenburg 1997; Pellow 2001; Taylor 2000).
Public policy activities around this issue have also been significant, as evidenced by the
U.S. Environmental Protection Agency’s (EPA) attempts to formulate policies to remedy envi-
ronmental injustices, including its creation of an Office of Environmental Justice, by the issu-
ance in 1994 of Presidential Executive Order 12898 calling upon all federal agencies, not just
the EPA, to take into account the environmental justice consequences of their actions, and by
the introduction of numerous environmental justice bills in the U.S. Congress and many state
legislatures across the country (Rechtschaffen and Gauna 2002; Ringquist 2003). At the same
time, the environmental justice movement has continued to grow. Given the extraordinarily
rapid rise in prominence of environmental justice as an important social issue in public and
academic discourse, as well as its implications in current debates about whether race as a fac-
tor affecting life outcomes is declining in significance (see, e.g., Cancio, Evans, and Maume
1996; Hughes and Thomas 1998; Wilson 1987), the attention given by sociologists to this
form of racial inequality appears warranted.
Many quantitative studies examining racial and socioeconomic disparities in the distribu-
tion of environmentally hazardous sites have been conducted over the past decade. Various
approaches have been applied to assess such disparities. These approaches have tended to be
of two types: (1) pollution dispersion assessments and (2) site proximity assessments. Pollu-
tion dispersion assessment studies involve collecting data about the volumes and toxicities of
various air and water emissions, timing of emission releases, stack heights, wind directions
and speeds, and other factors (Ash and Fetter 2004; Chakraborty and Armstrong 1997; Glickman,
Golding, and Hersh 1995). From these data, estimates are made about the geographic disper-
sion and deposition of the toxic emissions. Census data are then employed to determine the
demographic characteristics of those most likely to live where pollution and toxicity levels are
concentrated. Some pollution dispersion studies have gone as far as attempting to conduct
risk assessments in which human exposure and expected lifetime cancer risks are estimated
(for an example, see Hamilton 1999). Obtaining complete and accurate information for mod-
eling pollution dispersions and toxicity levels has been difficult; however, this has been espe-
cially so for risk assessments. As a result, relatively few environmental inequality studies
employing pollution dispersion or risk assessment methods have been conducted.
By far, the most frequently employed approach for conducting quantitative environmental
inequality analyses has been to assess the proximity of hazardous sites to nearby populations.
While nearly all national-level environmental inequality studies have involved proximity
assessments, all national-level studies of the distribution of hazardous waste treatment, storage,
and disposal facilities (TSDFs) have done so. These studies have been very influential in spur-
ring policy development and further research in the area of environmental justice. Although
most have found these disparities to be statistically significant (Lester, Allen, and Hill 2001;
Ringquist 2005; Saha and Mohai 2005), there has been considerable variation in the magnitude
of racial and socioeconomic disparities found. Some studies have found no race and income dis-
parities associated with the presence of environmentally hazardous sites and locally unwanted
land uses (Anderton et al. 1994; Davidson and Anderton 2000).
In another paper, the authors (Mohai and Saha 2006) hypothesized that a likely source
of these uncertainties has been wide reliance in environmental inequality research on what
has been termed “unit-hazard coincidence” methodology. This approach involves selecting a
pre-defined geographic unit (such as zip code areas or census tracts), determining which sub-
set of the units is coincident with the hazard and which not, and then comparing the demo-
graphic characteristics of the two sets. Implicit in this approach are two assumptions: (1) that
adverse impacts tend to be concentrated within close proximity of the hazards, and (2) that
populations living within the host units are located closer to the hazard under investigation
than populations living in the non-host units. However, we demonstrated that this latter