Copyright (c) 2003 North Carolina Law Review

North Carolina Law Review

March, 2003

81 N.C.L. Rev. 1031

LENGTH: 47308 words

ARTICLE: DISTRIBUTIVE JUSTICE AND THE ENVIRONMENT

Alice Kaswan*

* Professor of Law, University of San Francisco School of Law. B.S., University of California, Berkeley, 1984; J.D., Harvard Law School, 1991. I would like to thank Barbara Kautz for her invaluable assistance, as well as Shanee Michaelson for her earlier assistance. I thank John Applegate, Craig Anthony Arnold, Henry Brown, Josh Davis, Rick Diamond, Sheila Foster, Susan Freiwald, Gerald Torres, and Tim Iglesias for their insightful comments. I also thank my colleagues at the University of San Francisco for their helpful direction during a Faculty Scholarship Lunch devoted to this Article.

SUMMARY:

... The Waterfront South neighborhood is already a popular location for the siting of industrial facilities. ... In fact, the land use siting process and the dynamics of the housing market likely skew undesirable land uses toward poor and minority communities regardless of those communities' preferences. ... Considering the community preference model of distributive justice, we come to the not surprising conclusion that the objective criteria provide no assurance that siting decisions will gravitate to the locations where residents desire them. ... In sum, minority and poor communities are less likely to have effective voting power and less able to participate effectively in land use planning and zoning decisions than other communities, thus lessening the likelihood that their preferences are equally met. ... As with zoning ordinances, the existing distribution is a consequence not only of siting decisions being made now, but of past siting decisions. ... Demographic stability following a land use siting decision, like that found in Professor Been's national study of housing dynamics following the siting of hazardous waste facilities, does not prove that, on balance, the residents desired the facilities in question. ...  

TEXT-1:

[*1033]

Introduction

Waterfront South, a neighborhood in Camden, New Jersey, is not lined with marinas, yachting clubs, or fancy estates.

As one court has described:

The population of Waterfront South is 2,132, forty-one percent of whom are children. Ninety-one percent of the residents ... are persons of color ... . The residents of Waterfront South suffer from a disproportionately high rate of asthma and other respiratory ailments. n1

[*1034]

In 1990, the median household income of residents of Waterfront South was $ 15,082, and the per capita income was $ 4,709. Over 50% of the residents of Waterfront South live at or below the federal poverty level. n2

The Waterfront South neighborhood is already a popular location for the siting of industrial facilities. It contains the Camden County Municipal Utilities Authority, a sewage treatment plant, the Camden County Resource Recovery facility, a trash-to-steam plant, the Camden Cogen Power Plant, a co-generation plant, and two United States Environmental Protection Agency ("EPA") designated Superfund sites. Four sites within one-half mile of [a] proposed facility are currently being investigated by the EPA for the possible release of hazardous substances. The [New Jersey Department of Environmental Protection] has identified fifteen known contaminated sites in the Waterfront South neighborhood. n3

In 1999, the St. Lawrence Cement Company selected Waterfront South as the venue for a new cement manufacturing facility. n4 According to a court reviewing the siting decision, the facility will emit "particulate matter (dust), mercury, lead, manganese, nitrogen oxides, carbon monoxide, sulphur oxides and volatile organic compounds." n5 Each year, 35,000 trucks will cross the neighborhood to make deliveries to the facility and 42,000 trucks will depart from the neighborhood. n6 Since deliveries will come from a barge, the deliveries will be concentrated, with 500 truck deliveries per day on the eighty days a year scheduled for deliveries. n7

The story in Waterfront South is not unique. This pattern of disproportionately siting locally undesirable land uses, or "LULUs," in poor and minority n8 neighborhoods is common throughout industrial America. n9 Similar patterns emerge when one considers [*1035] other types of "undesirable" land uses, such as halfway houses for former prison inmates, homeless shelters, and the like. n10

Beginning in the early 1980s, these conditions prompted the emergence of the "environmental justice" movement. In particular, minority communities and organizations began to question whether they were being exposed to more than their fair share's worth of environmental harms and whether decisions distributing environmental harms were being made fairly. n11 Local grassroots organizations around the country began to explore the connection between race and exposure to undesirable environmental problems. n12 Over the last decade, these grassroots efforts have coalesced into a broadly-based movement for environmental justice n13 that has challenged environmental organizations and government agencies to address the fairness and implications of their actions. n14

Some commentators have suggested that "distributive injustice" - that is, evidence of disproportionate land use patterns - is not of regulatory concern unless it can be shown that the unequal patterns were caused by identifiably discriminatory or biased processes. n15 Under this view, distributional disparities not caused by [*1036] tainted processes might be regrettable, but are not the appropriate focus of regulatory action. For example, a disparity would be worth redressing if it were caused by an intentionally discriminatory siting decision. The disparity would not be worthy of regulatory concern, however, if it was not caused by a tainted decision-making process, but was instead traceable to the operation of "market forces," such as differences in land costs or the relative efficiency of one location over another. Under this view, the Waterfront South siting decision described above would be worth redressing only if caused by intentional discrimination.

I first argue that distributive justice is a critical issue even in contexts devoid of intentional discrimination or other process failures. n16 Without demeaning the significance of the environmental justice movement's attention to decision-making processes and the structural causes of disadvantage, I argue that distributive justice is an important goal in its own right because it responds to the conditions that people actually experience n17 and because it may be more amenable to remedy than many forms of discriminatory treatment. n18

The issue is not simply abstract: the importance one attaches to distributive justice has a significant impact on the willingness to develop public policies to address it. Consider, for example, a government policy prohibiting government agencies from siting facilities where there is already a high concentration of polluting facilities. If an agency nonetheless located a facility in such an area, it could be held in violation based solely on the outcome of its decision, even if the basis for its decision was not discriminatory and there was no bias in the decision-making process. Such a policy might prohibit siting an industrial facility in Waterfront South, the community described above. Proposals of this nature have been highly controversial, with the controversy turning to a considerable extent on attitudes about the relative importance of purely distributive justice. n19

[*1037] Assuming the importance of distributive justice, the second issue I address is whether the current distribution of undesirable land uses is distributively just. The answer to this question is contested, and turns upon how we define, and therefore measure, distributive justice. This Article identifies and clarifies the two most important theories at issue, and argues that we have failed to achieve distributive justice regardless of which theory of justice we adopt.

The most prevalent theory of distributive justice, what I call the "equal division" model, measures justice by the degree to which undesirable land use distributions are physically equal. n20 Because studies indicate that many poor and minority communities experience a greater concentration of LULUs than other communities, n21 we have a distributive justice problem under this model.

Professor Lynn Blais, in her article Environmental Racism Reconsidered, has, however, challenged the assertion that there is a distributive justice problem through proposing what is, implicitly, [*1038] another way of conceptualizing distributive justice. n22 Under her approach, which I characterize as the "community preferences" model, disparities in the siting of LULUs are not necessarily unjust. n23 This model defines distributive justice according to the degree to which the location of LULUs meets community preferences. Because what might be undesirable to one community might be desirable to another, a physically unequal distribution could be "just" if the distribution satisfied host communities equally. Professor Blais argues that the market in preferences works well enough to conclude that, overall, disparities are generally justified by differing preferences. n24

I argue that the market in community preferences does not work, and that the community preferences model therefore fails to provide an "out" from the problem of distributive justice. Economic and political markets do not function to meet community preferences equally. In fact, the land use siting process and the dynamics of the housing market likely skew undesirable land uses toward poor and minority communities regardless of those communities' preferences. n25 The distributive justice problem does not evaporate through applying a community preferences model; whether we measure justice by it or by the equal division model, we encounter significant inequities deserving serious attention.

The choice of model - equal division versus community preferences - and the assessment of whether justice is achieved under each model, have real-world implications for public policy. Under the equal division model, evidence of disparities provides a strong basis for seriously considering government policies to address unfair concentrations of LULUs. But if one instead adopts the community preferences model and assumes that it provides a satisfactory explanation for existing disparities, then government efforts to equalize the siting of LULUs become unnecessary: there is no distributive justice problem to remedy. Moreover, under this theory, government efforts imposing across-the-board distributional edicts could be counterproductive. If one assumes that the market in [*1039] preferences works, government efforts to equalize distributions would interfere with the market's ability to match differing community preferences, leaving communities equally burdened, but unequally satisfied. Under this model, public policies to equalize siting would be an unnecessary and counterproductive intrusion into neutral agency actions.

Since the community preferences model threatens to erase the distributive justice problem, it is important to determine whether its underlying assumptions are accurate. Otherwise, it risks lulling us into complacency about distributive injustice and thwarting efforts to address it.

In sum, the analysis in this Article has four primary goals: (1) to argue for the importance of addressing distributive justice in its own right; (2) to illuminate the relevant theories of distributive justice applicable in the land use context; (3) to demonstrate that we have failed to achieve distributive justice, whether one adopts the equal division or the community preferences model; and (4) to demonstrate more broadly the fallacy of assuming that "the market" has the capacity to resolve significant public policy issues, such as the inequitable siting of locally undesirable land uses. This Article does not attempt to propose specific public policy initiatives. Instead, the Article addresses the fundamental questions about the importance and nature of distributive justice that must undergird future policy development.

Part I provides a brief introduction to the environmental justice movement and then identifies its primary claims: "distributive," "political," and "social" justice. Part I then explores the various causes of distributive injustice, ranging from the invidious to the innocent. It concludes by observing why the issue of distributive justice necessarily implicates racial and economic justice. Part II addresses the importance of distributive justice as a focal point for policy attention, even in the absence of intentional discrimination or identified process failures. Part III explores general theories of distributive justice and analyzes how they lead to the particular models that are relevant in the land use context.

Part IV introduces the first model of distributive justice, the equal division model, and then concludes that, notwithstanding certain methodological controversies, we have a distributive justice problem under this model.

Part V introduces the community preferences model of distributive justice. This model has been used to challenge the assumption that an unequal division of LULUs is unjust, and posits [*1040] that unequal distributions may be justified by differences in community preferences. The model itself is controversial. However, this Article accepts the model on its own terms. Part V then explains the methodology for determining whether preferences are equally met. Since it is impossible to determine the extent to which community preferences are met through direct empirical analysis, it is necessary to analyze the likelihood that they are met by exploring the factors most likely to determine existing land use distributions: land use siting processes and post-siting housing market dynamics. Hence the apparent irony: an article on distribution that thoroughly explores land use processes. The irony is resolved, however, by observing that I generally evaluate processes not for their own sake, n26 but to determine the likelihood that the resulting land use distributions are likely to satisfy preferences equally.

Part VI evaluates land use siting processes to determine whether they are likely to have led to distributions of land uses that satisfy community preferences equally. It analyzes objective factors (such as market forces and regulatory requirements), political processes expressed through general zoning and individualized siting decisions, and the special role of public participation provisions. It concludes that siting decisions often fail to reflect equally the preferences of facilities' neighbors, and instead are often skewed against the preferences of minority and low-income residents.

Part VII addresses the contention that, even if land use siting processes do not meet community preferences equally, post-siting dynamics in the housing market could rectify such failures. Residents who did not like the LULUs could move away; residents who wanted to live near them could move closer. Part VII concludes that post-siting housing market dynamics are not likely to result in distributions that match neighborhood preferences equally. Differences in income and housing discrimination mean that some will be better able to move toward or away from their preferred land uses than others, thereby deepening distributional disparities. The "market" does not lead to equity.

The Article concludes that advocates of the community preferences model have incorrectly assumed a functioning private preferences market in land use siting and in the post-siting housing [*1041] market. The community preferences model's misplaced faith in the market should not lead us to dismiss important governmental efforts to address distributive justice.

This Article is a starting point. Important questions remain, such as the appropriate model for distributive justice, the weight to be given distributive justice when it competes with other social policies, what government efforts would be likely to succeed given the complex causes of distributive inequities, and the specific goals and structure of government initiatives to achieve distributive justice. These issues are, however, subjects for future scholarship.

I. The Environmental Justice Movement and Its Claims for Justice

A. The Environmental Justice Movement's Impact

Since the 1980s, the rise of the environmental justice movement has begun to impact environmental and land use policy. As the movement challenged the fairness of environmental impacts and decision-making, academics, government agencies, and others undertook systematic studies that suggested that LULU distributions were correlated with race and income. n27 These studies and the movement's gathering strength resulted in various environmental justice initiatives in the 1990s. The Environmental Protection Agency ("EPA") established the Environmental Equity Workgroup, which analyzed existing evidence on the distribution of environmental risks and in 1992 published a study called "Environmental Equity: Reducing Risk for All Communities." n28 In February 1994, President Clinton issued Executive Order 12,898, which requires federal agencies to consider the implications of federal decisions on the poor and on minorities and which urges federal agencies to develop mechanisms for ensuring the participation of all groups in decision-making processes affecting them. n29 The EPA, in conformance with the Executive Order, has issued several guidance documents indicating how it will take environmental justice issues into consideration. n30 Pursuant to the Executive Order, other federal [*1042] agencies have, to varying degrees, integrated environmental justice concerns into their programs. n31

The federal government has not acted alone. Various states have enacted or are considering legislative proposals to control the siting of undesirable environmental land uses. n32 In addition to government actions, a number of mainstream environmental groups have responded to the movement with greater attention to environmental justice issues. n33 Although opinions differ as to the sincerity and depth [*1043] of these efforts, n34 environmental justice issues play a much greater role in debates about environmental policy than was evident just a decade ago. Critical issues remain, however, as to the appropriate nature and scope of environmental justice initiatives.

B. The Environmental Justice Movement's Claims for "Justice"

At its core, the environmental justice movement raises two distinct types of claims for justice: distributive justice and political justice. n35 To provide context for this Article's subsequent focus on distributive justice, this Section sketches the basic parameters of each form and introduces one broader category: "social justice." n36 It also analyzes the degree to which distributive injustices are, and are not, likely to be caused by political or social injustices. The primary focus is on claims for justice raised in the context of land use siting decisions.

1. Distributive Justice

Claims for "distributive" justice focus on whether communities bear more than their fair share of LULUs. n37 While all communities [*1044] must bear some of the burdens of industrial society, the critical question is whether these burdens are distributed equitably. The environmental justice movement raises both narrow and broad claims of distributive justice. When individual local communities argue that they are subject to more than their respective fair shares of LULUs, they are raising "narrow" claims regarding their particular circumstances. n38 For example, if one community is subject to all of the waste disposal facilities for an entire metropolitan area, then that community has a claim of distributive injustice in the narrow sense. But many in the environmental justice movement make the broader claim that, overall, more LULUs are located in poor and minority neighborhoods than in other neighborhoods. n39 Although both narrow and broad claims are critical parts of the environmental justice movement, the broader claim is the primary focus of this Article.

It is important to note that distributive justice is concerned with outcomes, not with the causes of those outcomes. n40 Thus, if the distribution of waste disposal facilities in an area is highly skewed, the community has a claim of distributive injustice regardless of whether the disparity was caused by intentional discrimination or, instead, by "objective" siting criteria bearing no relation to the community's demographics. Similarly, looking at the distribution of LULUs more broadly, unequal distributions are of concern regardless of whether they were determined by discriminatory processes or ostensibly neutral market factors.

[*1045]

2. Political Justice

The other central claim the environmental justice movement makes is one I have characterized as a claim for "political justice." n41 A claim for political justice looks at the fairness of decision-making processes rather than the discrete distributional outcome of those processes. n42 As H.L.A. Hart has stated, "a choice, made without prior consideration of the interests of all sections of the community would be open to criticism as merely partisan and unjust." n43 If the interests of some are given more weight than the interests of others, then we have an instance of political injustice. n44 For example, if a community were selected as a site for a waste disposal facility because its interests were not treated with the same respect as those of others, then that community has a claim of political injustice. More pointedly, if the community were selected as a site because of intentional racial discrimination, that claim would implicate political injustice.

Although the distributional outcome of a decision-making process may be relevant to assessing its fairness, n45 the central inquiry under political justice concerns the process rather than the outcome. When evaluating discrete decisions, the standard for political justice is fair treatment, without any guarantee of a fair outcome. n46 Here, too, [*1046] the environmental justice movement makes narrow and broad claims. In a particular context, it might claim that a particular siting decision was motivated by invidious factors. n47 More broadly, those in the movement argue that there is systemic discrimination in decision-making institutions and structures. n48

Many have used the term "procedural justice" to describe what I term "political justice." n49 This Article uses the term "political justice" rather than "procedural justice" because the issue goes beyond the question of procedure to the substance of the deliberative process. n50 If the issue is framed as one of "procedural justice," then decision-makers might argue that they have solved the "fair treatment" problem through the creation of procedures that ensure participation of all groups in decision-making processes. It is not clear, however, that procedural requirements enhancing public participation will necessarily lead to substantive decisions that are more responsive to public opinion. n51 While enhancing participation procedures to equalize opportunities is an important step in creating the preconditions for political justice, it provides no guarantee that the [*1047] substantive decision will embody political justice. One must, of course, be mindful of the risk of creating a false dichotomy between "substance" and "procedure." Clearly, the procedures agencies use may affect the substance of their ultimate decisions, and this hope is one of the motivating forces behind the creation of improved procedural requirements. n52 Nonetheless, a focus on procedure is one step removed from establishing explicit expectations for the substantive respect that various groups' interests should be accorded in decision-making processes.

3. Social Justice

Distributive and political justice mark the basic categories of justice at issue in the environmental justice movement. Under the rubric of "social justice," environmental justice claims look beyond immediate siting processes and outcomes to consider the wide web of political, economic, and social forces that influence the distribution of LULUs. n53 This inquiry focuses on such issues as inequities in wealth, historic and present zoning practices, implicitly or explicitly discriminatory decision-making structures that indirectly affect siting decisions, housing discrimination, and employment discrimination. The "social justice" rubric is useful because it indicates a broadly-focused inquiry. At base, however, it combines elements of distributive and political justice and does not present a fundamentally different category of justice.

***

None of these claims for justice - distributive, political, or social - are raised in a vacuum. Many communities raise all of them. n54 Speaking narrowly, siting decisions are challenged because they add to the disproportionate burden a community must bear and because the bases for the siting decisions are believed to be discriminatory. n55 [*1048] Speaking broadly, those in the movement decry inequalities in national distributions and the widespread racism, power disparities, and institutional biases that are believed to cause them. n56

C. The Causes of Distributive Injustice

The extent to which distributional disparities are caused by political and social injustice is highly contested. n57 Although I argue that the presence of distributional disparities is sufficient to trigger regulatory concern regardless of cause, the causes are not irrelevant. Where distributive injustice is linked to political or more general social injustice, it presents multiple bases for concern. It is thus worthwhile to at least sketch the possible relationships. n58

In some circumstances, distributive injustice could be caused by political injustice in discrete siting decisions. Historic and current discrimination in siting LULUs would likely lead to a concentration of LULUs in disfavored communities. n59 Theoretically, these distributional disparities could be erased by post-siting housing dynamics: those who did not like certain land uses could move away; those who liked them could move closer. n60 Even if this were true, however, there would be some period of time during which disparities would exist. As elaborated below, moreover, such post-siting housing market dynamics are not, in fact, likely to equalize the demographics surrounding LULUs. n61 Thus, past and present discrimination in siting processes are likely to lead to distributional disparities.

In other circumstances, siting disparities might not be caused by political injustice in discrete decisions, but might nonetheless be a product of the broader form of social injustice discussed above. Historic or present zoning laws might concentrate undesirable land [*1049] uses in certain neighborhoods. n62 Alternatively, social injustice may have led to disparities after an initially neutral siting decision. If at the time of an initial decision, an area had been unpopulated, or was white and affluent, then the siting decision itself was probably not politically unjust. n63 But broader social injustices, like poverty and housing discrimination, could have led to a subsequent concentration of poor and minority residents, since they would be less able to flee undesired LULUs, or might be attracted to lower-valued housing in areas with LULUs. n64

Social injustice is also a factor in disparities that arise from certain market forces, like land values. Proponents of undesirable land uses are drawn to cheaper land, land that is often located in poor and minority neighborhoods. n65 In such cases, there is no political injustice in the discrete siting decision. But the problem of social injustice is evident. Land values in minority neighborhoods are often cheaper than in comparable non-minority neighborhoods, suggesting that the difference in valuation is a product of lingering discrimination. n66 Thus, even if certain distributional disparities are not attributable to political injustice in discrete siting decisions, they may nonetheless be attributable to broader forms of social injustice. n67

Finally, it is possible for distributive disparities to arise without implicating either political injustice in a siting decision or broader issues of social justice. A community could be treated with the same concern and respect as all other communities, but nonetheless be selected for a disproportionate share of LULUs due to some [*1050] objective factor that does not implicate political or social justice. n68 For example, geological features might explain a concentration of waste disposal facilities. In such an instance, a distributional disparity in disposal facilities would have been caused by the objective factor of stable geological formations, not by the siting agency's failure to treat the waste-site beleaguered community with less fairness and respect than others. In this situation, we could identify a distributional injustice, but it would not be linked to a political or social injustice. (Arguably, however, if decision-makers know their decision will result in an unintended disparate impact, but fail to take that impact into consideration, their decision demonstrates a lack of respect for the affected community that does suggest political injustice. n69)

The core of my thesis is that distributional injustice is a matter of concern regardless of its cause: it is a concern in all of the situations described above, whether the product of outright discrimination or utterly neutral factors. Nonetheless, this sketch of possible causes of distributive justice suggests that, in many instances, distributive injustice presents a concern not just in its own right, but because it implicates other forms of injustice as well.

D. Distributive Justice and Its Relation to Racial and Economic Justice

Theoretically, a problem of distributive justice exists whenever any individual or group is subjected to an "unfair" proportion of LULUs. On its own terms, the concept of distributive justice does not implicate race or class. If the population were completely integrated by race, class, and other demographic features, then the question would simply be whether Neighborhood A had more and/or worse LULUs than Neighborhood B. Under these circumstances, the question of distributive justice might be important, but it would not implicate the questions of racial and economic justice that are critical to the environmental justice movement.

[*1051] But patterns of residential living in the United States are highly segregated by race and class. n70 Neighborhood A is likely to have a completely different make-up from Neighborhood B. Economic segregation is widespread, and particularly evident in the sharp lines between wealthy suburbs and poor inner cities. n71 Although levels of racial segregation for African-Americans have been slowly decreasing since 1970, n72 a recent Brookings Institute study suggests, based on 2000 census figures, that "the large number of American metropolitan areas with extremely high levels of segregation remains quite striking." n73 For African Americans, the most segregated group in the United States, n74 the national average of a key segregation index is in the "hypersegregated" range, n75 and segregation is significantly above [*1052] this national average in the nations' most populous areas. n76 The slow decreases in segregation over the last three decades n77 have been achieved through the integration of formerly all-white census tracts, not through the integration of heavily African-American census tracts, n78 and the decreases in segregation have been smallest in the areas with the greatest African-American populations n79 and the greatest amount of historic segregation. n80 Almost a third of the nation's African Americans currently live in neighborhoods that are 80% African-American or more. n81

Levels of segregation for Latinos are significant, but slightly lower than those for African Americans. n82 In the Northeast, Latino segregation has reached "hypersegregation" levels. n83 As with African Americans, the highest levels of segregation exist in those regions with the highest percentages of Latinos. n84 In addition, while the levels of African-American segregation have been slowly decreasing over the last thirty years, n85 key aspects of Latino segregation have been [*1053] increasing, not decreasing. n86 The reality, therefore, is that neighborhoods differ greatly in their demographic make-up.

To the extent that disparities in distribution are correlated with differences in race or income, the issue of distributive justice implicates fundamental notions about equity in our society that would not arise in a more integrated world. Although any inequity deserves concern, inequities that stem from and in turn create the pernicious effects of segregation change what could be an abstract question of distributive justice into one that is necessarily related to broader issues of racial and economic justice.

II. On the Importance of Distributive Justice

Distributional inequities are frequently discounted unless and until one demonstrates that they were caused by discrimination - by political injustice. n87 As one author has noted, in light of the difficulty [*1054] of proving that a given distributional injustice was caused by direct political injustice by siting authorities, "critics have easily scrutinized the distributive paradigm of environmental injustice with causation objections." n88 Some of these critics thus imply that, if the disparity was not caused by a racist or otherwise discriminatory siting decision, then the disparity is not problematic. n89 For example, if an area is subject to a disproportionate burden due not to discrimination, but to lower property values that attract industry, then "such is life;" we have not implicated "justice." n90 The argument echoes constitutional equal protection jurisprudence, under which the disparate impact of a government action is not relevant unless it can be linked to an intent to discriminate. n91 In the philosophical tradition, the view against addressing distributive justice is expressed most forcefully by Robert Nozick, who argued that efforts to achieve justice should not attempt to achieve particular distributional goals. n92 He argued that "whether [*1055] a distribution is just depends upon how it came about." n93 Process, not distributional outcome, should be the touchstone.

I argue that distributive justice is of critical independent significance. By focusing on distributive justice, one concentrates on what is experienced, regardless of cause. If inner-city African-American children are exposed to far more air pollutants than children in other neighborhoods, then that problem is one worth considering. The absence of a racist motive does not change the troubling distributional disparity. As Alan Freeman has stated in the Equal Protection Clause context, an exclusive focus on affixing blame directs attention to the "perpetrator" without sufficiently considering the conditions experienced by the "victim." n94 For the communities experiencing harmful and debilitating disparate impacts, it is problematic to argue that those impacts are not worthy of attention without an identified discriminatory cause, n95 or that the primary significance of the impacts is in what they tell us about their cause. A preoccupation with assessing why an inequity has arisen should not impede the effort to remove the disparity.

Distribution has also been recognized as an important aspect of justice by political philosophers. John Rawls's A Theory of Justice is a pivotal work articulating a conception of justice that includes a [*1056] strong distributive component. n96 He argues that "the justice of a social scheme depends essentially on how fundamental rights and duties are assigned and on the economic opportunities and social conditions in various sectors of society." n97 His reference to the distribution of "social conditions" clearly implicates the type of concerns raised by the environmental justice movement. Other formulations of the importance of distribution to justice abound. Professor Nicholas Rescher states that "justice consists in realizing to the greatest possible extent a distribution that renders to each a "fair share' of the good (or evil) at issue." n98 Professors John Arthur and William Shaw have stated that the "issue which lies at the heart of all philosophizing about society [is]: What constitutes a just distribution of the benefits and burdens of economic life?" n99 A focus on distribution thus resonates with significant philosophical theories on the nature of justice.

Moreover, at least in the short-term, the distributional dimension of an environmental justice conflict may be more amenable to improvement than the political dimension. In some instances, distributional injustice may arise without political injustice. Recall the example of waste disposal sites clustered in an area that is geologically suitable. In that case, there is no "political injustice" to remedy; those affected would be best served by a siting process or remedy that simply protected communities from such disparate consequences. Or some decision-makers, such as environmental permitting officials, may have operated under statutory directives that were blind to potential distributional consequences. n100 They cannot be blamed for intentional discrimination. n101 Therefore, rules that [*1057] require them to take distribution into account in the permitting process, and hold the government liable for the failure to do so, might better address distributional disparities that are not caused by intentional wrongdoing. Nor are the environmental statutes facially discriminatory. Thus, remedies based upon distributive inequities would provide protection against detrimental disparities that occur even in the absence of political or clear social injustice.

In those cases where distributional injustice is attributable to political injustice, that injustice may be very difficult to remedy. It could be the consequence of historical rather than present discrimination, such as the legacy of discriminatory zoning and residential housing patterns that have left their mark for generations. n102 To the extent present discrimination is suspected, the "intent" requirement makes legal redress very difficult to obtain under modern interpretations of the Equal Protection Clause. n103 Discrimination is rarely explicit, n104 and the multiplicity of factors that could plausibly influence a siting process, such as land costs, site characteristics, or other logistical factors, provide many opportunities for decision-makers to explain what might have been a discriminatory decision on nondiscriminatory grounds. n105 Discrimination is not only rarely explicit, some contend that it is often unconscious - the product of repeated but unexamined social practices and deeply ingrained cultural assumptions. n106 If current legal rules require a demonstration of discriminatory intent to address political injustice, that intent will be hard to find, and justice difficult to achieve. While the political and social injustices that cause some distributional disparities should [*1058] be addressed where possible, the difficulty of doing so suggests that communities could be better off if they had direct mechanisms for improving distributional disparities.

Professor Sheila Foster has cautioned that inquiries into the nature of environmental justice should not be limited to distributional analyses, but should instead look behind the distributions to determine the range of political, social, and economic factors that have led to the disparities. n107 She argues that focusing on distributive justice "often obscures our consideration of the social structure and institutional context that play a role in determining the patterns of distribution." n108 Professor Foster acknowledges the relevance of distributive patterns, but sees them as a starting point for a more in-depth consideration of the "social processes underlying distributional patterns." n109

I agree with Professor Foster that a full understanding of the nature of environmental justice requires consideration of the intricate linkages between distributions and their social, political, and economic causes. I simply have a different focus from Professor Foster. My focus is not on what is necessary to understand the full nature of environmental injustice; that indeed requires a full inquiry into political and social forces. Instead, my focus is on what information is necessary to trigger the need for a remedy. My argument is simply that information about distributional disparities alone is a sufficient predicate to regulatory action; a community should not have to demonstrate any sort of suspect causation before being entitled to attention.

Writing in political philosophy, Professor Iris Marion Young has suggested a deeper critique of a focus on distributive justice. Like Professor Foster, she argues that focusing on distributive justice could fail to address the deeper social problems that cause disparities to arise n110 because it presupposes rather than scrutinizes institutional structures and processes. n111 Ultimately, a primary focus on distribution could implicitly support unjust institutions, since it takes [*1059] them as given. n112 Her concern is not just with the study of discrimination, however, but the development of distributionally-based remedies. For example, affirmative action efforts tend to focus on distributions: who gets employment or opportunities for higher education. That focus fails to address such critical underlying structural issues as who decides who is "qualified" for employment and why some have the means to attain these qualifications while others do not. n113 A focus on distribution is ultimately depoliticizing, Professor Young argues, because potential challenges to the existing systems of power and control become rechanneled into distributive "solutions" that dissipate the thrust of critical social movements. n114 Professor Young does not argue that distributive justice is irrelevant, n115 but she does argue that issues of political and social justice should be the primary focus.

Professor Young's concern that a preoccupation with distributive justice could lead to a failure to question and challenge unjust social and institutional structures is an important caution. If the environmental justice movement were reduced to simply counting how many facilities end up here and there, then critical aspects of the movement would indeed be lost. Many environmental justice leaders are not simply challenging the number of facilities to which they are subject, but also the fairness of decision-making and underlying power structures. n116 Challenging environmental decisions is one step in a broader engagement over the nature of economic and political power. The goal of sustained challenge is a greater political voice - a voice that may transcend particular disputes over particular facilities. n117 It is also important to identify the widespread inequities [*1060] that may lie behind current land use distributions. Furthermore, addressing political and social processes will also likely improve distributive justice given their role in causing disparities. Nonetheless, given the difficulty of devising effective remedies for many past and present forms of political and social injustices, n118 and the real world consequences of distributional inequities, I argue that it is appropriate for the movement to direct at least some of its efforts toward distribution-focused remedies.

I conclude this Section on the "importance of distributive justice" with a significant qualification about how distributive justice fits into the broader picture. Notwithstanding the critical importance I attribute to distributive justice, I recognize that it may not be the only factor that is ultimately relevant in the context of facility siting. n119 It is possible that the outcome of a siting decision could be distributively unjust, but, nonetheless, be justified overall. For example, there may be very important reasons, such as safety considerations, why a facility needs to be placed in an area that is already subject to a disproportionate number of other facilities. I argue that distributive justice should be an important factor, but recognize that it cannot always be the sole factor that determines facility siting.

Nonetheless, the presence of a worthwhile overriding justification for a distributively unjust outcome does not erase the distributional injustice. We do not have "justice" simply because we can explain the disparity. In such situations, distributional injustice must be recognized as a tradeoff, even if ultimately an acceptable tradeoff; it is not erased by the existence of other justifications for the distributively unjust outcome. Determining how distributive justice should be balanced with other goals presents one of the major challenges that specific policy initiatives will have to address. n120

[*1061]

III. Conceptions of Distributive Justice

To this point, I have primarily conceptualized distributive justice in comparison with political justice. In this Section, I focus on distributive justice itself, and consider how theories of justice in the LULU context relate to and emerge from broader concepts of distributive justice. n121

Most distributive theories of justice are inherently consequentialist; in other words, they focus on outcomes. n122 Utilitarianism is the quintessential consequentialist theory. According to classic utilitarianism, the best distribution is that which will lead to "the greatest good for the greatest number." n123 As many critics of utilitarianism have noted, however, this theory does not provide a satisfying answer to the question of how goods (or bads) ought to be distributed among people. For example, the greatest amount of good (e.g., the highest possible Gross National Product) might be achieved by distributing a great deal of wealth to many, but leaving some in dire poverty. n124 Given its potential for serious [*1062] inequality, the potential for some to suffer for the greater good, Rawls deems the "principle of utility ... to be inconsistent with the idea of reciprocity implicit in the notion of a well-ordered society." n125 Our search for a compelling theory of distributive justice thus goes beyond utilitarianism to find a theory that resonates more strongly with notions of fairness.

The potential for undesirable inequality in a utilitarian distribution leads us to a theory of equality as another potential basis for evaluating whether a distribution is just. n126 Isaiah Berlin has stated:

no reason need to be given for ... an equal distribution of benefits - for that is "natural' - self-evidently right and just, and needs no justification ... . The assumption is that equality needs no reasons, only inequality does so ... . n127

Mr. Berlin's comment is applicable to the distribution of harms, not just benefits. The critical question then becomes: what does equality mean? It quickly becomes clear that strict equality as such is not necessarily consistent with distributive justice, since it is arguably unjust to treat people in the same way if they are different. n128 Instead, the concept can be articulated as: "People who are alike in all relevant respects deserve the same things." n129

The next critical issue becomes, if equality requires that we treat different people differently, and like people alike, then what differences matter, and how much? A number of factors are frequently mentioned. For example, a theory of justice might distribute goods on the basis of need. n130 Another set of factors turns on the issue of "desert": of distribution accomplished in proportion [*1063] to what people receiving the goods (or bads) deserve. n131 Factors implicating "desert" include: one's innate ability; the level of effort expended; the actual achievements attained; and moral merit. Philosophers differ over the appropriate role of these factors: which are defensible and which not; n132 how much weight they should have relative to one another; and how one should measure and compare within and across categories. n133 As the discussion below makes clear, however, these thorny dilemmas need not be resolved where the distribution of LULUs is at issue.

In the context of siting LULUs, I argue that none of these factors justify a deviation from a general principle of equal distribution. In explaining the operation of "desert-based" factors, most theorists assume the distribution of "goods" - who deserves to have more money, more things, or more jobs. These factors are less compelling where the distribution of a "bad," like a LULU, rather than a "good," is at stake. No community "deserves" a LULU due to some defect in ability, effort, achievement, or moral standing. n134 Principles related to [*1064] "just deserts" therefore do not justify inequalities in distribution in the LULU context, however much these principles might be applicable in other contexts.

The role of community need as a justification for differing distributions is somewhat more complex, but ultimately similar. As indicated above, at this stage in the analysis we are assuming that LULUs are generally considered "bads." A community would not have a "need" for something that is considered a "bad." n135 Thus, so long as LULUs are treated as "bads," community need does not present a factor justifying an unequal distribution of LULUs. (The preference theory of distributive justice, discussed further below, n136 addresses the possibility that a community would find a LULU a good, not a bad.)

The factors of need and desert focus on the recipients of a given distribution. Other factors that could justify unequal distributions turn on extrinsic societal considerations unrelated to the person or persons receiving the distribution. n137 The most important of these echoes in the utilitarian theory discussed above: distributions should serve the common good. n138 While I argue that the classic utilitarian conception of the greatest good for the greatest number is an unsatisfying theory of distributive justice, there are other formulations along these lines that are more compelling. John Rawls's famous "difference principle" falls into this category. n139 Unlike the traditional utilitarians, Rawls uses equality as a starting point for conceptualizing distributive justice. He justifies inequalities under limited circumstances, however, if they will result "in compensating benefits for everyone, and in particular for the least [*1065] advantaged members of society." n140 While Rawls does justify inequality on behalf of the general welfare, he differs from classic utilitarianism because he would not accept a distribution that allows some to endure a greater disadvantage for the benefit of others. n141

The difference principle could be applied to LULU siting both narrowly and broadly. Applied narrowly and directly, the issue would be whether the inequality of imposing more LULUs on poor and minority areas would be justified because it assists "the least advantaged." Assuming that LULUs are "bads," not "goods," such inequality would clearly violate the difference principle because it worsens rather than betters the condition of these less advantaged communities.

The difference principle is also violated if one attempts to apply it more broadly. One could argue that siting LULUs in poor neighborhoods helps society at large because such areas are presumptively the cheapest, and therefore economically efficient. But in order to satisfy the difference principle, one must be able to show not only that there is a general benefit to society, but that no one is made worse off in the process, particularly the least advantaged. It is highly implausible that any societal efficiency gain from siting a LULU on cheap land would trickle down and benefit the impacted community enough to offset the negative impact of the LULU. n142 (Again, the presumption here is that the LULU is a "bad" that is not desirable to the impacted community.) Thus, the inequality of siting LULUs in poor neighborhoods is not justified under the difference principle because, even if it has some benefits for society overall, those benefits are not likely to compensate the impacted communities for the LULU's net negative impact.

Where LULUs are at issue, concepts of "need," "desert," and theories like Rawls's difference principle do not justify deviations from equality as a distributional goal. Equality is therefore a guiding [*1066] principle in determining whether the distribution of LULUs is just. It should be emphasized that I speak here only of distributive justice. As indicated above, there may, at times, be justifications for siting facilities unequally. n143 But that does not make the outcome distributively just; distributive injustice must be viewed as a real tradeoff, however acceptable a tradeoff we may find it.

Assuming that equality is the relevant touchstone for determining distributive justice, the question remains: how does one define equality in the LULU siting context? I argue that there are two, sometimes competing, theories of distributive justice that are both based upon equality but define the term differently: the "equal division" model and the "community preferences" model. n144 I sketch each theory briefly here, to be developed further in the following Sections.

The most common assumption about what we mean by distributive justice is that of a physically equal distribution of undesirable land uses. n145 The assumption is that if there are more LULUs in some neighborhoods than others, then the distribution is not equal.

The community preferences model has been raised to challenge the assumption that the unequal division of land uses is necessarily unjust. This alternative model of distributive justice flows from the theory of distributive justice developed above: strict physical equality might not be the appropriate measure if there are relevant differences. Professor Lynn Blais has suggested that one such [*1067] difference is communities' relative preferences for land uses commonly referred to as "undesirable." n146 Under the community preferences model, the relevant barometer for determining equality would be whether people's preferences are met equally, not whether LULUs are equally divided. Where there are different preferences, there should be different distributions. Thus, what might be considered distributively unjust under the equal division model would not necessarily be considered distributively unjust under the community preferences model.

IV. Distributive Justice under the Equal Division Model

A. Distributive Justice as an Equal Division of LULUs

The conception of distributive justice that most readily comes to mind is of an equal division of LULUs. Considered narrowly, distributional injustice would exist wherever a particular community was subject to more or worse LULUs than other communities. n147 Thus, a community could argue that it is unfairly subject to more industrial uses than other neighborhoods. Similarly, a community could argue that, even though other communities contain some LULUs, it hosts LULUs having more adverse impacts than the LULUs located in other communities. Considered broadly, distributive injustice would exist if more and worse LULUs are located in poor and minority neighborhoods. The idea of equal division often, although not necessarily, assumes that LULUs have objective burdens that will be viewed by everyone in the same way. n148 If everyone evaluates the benefits and burdens of various land uses in the same way, then distributive justice would be achieved by equal distribution.

As Professor Vicki Been has made clear, the problems with measuring fairness pursuant to this theory and devising efforts to [*1068] satisfy the theory are formidable. n149 It is difficult, for example, to craft an objective measure for comparing the relative harm from different types of LULUs. n150 Furthermore, the issue of how to define the identity or boundaries of a "neighborhood" being affected by LULUs is complicated and contested. n151 These definitional challenges do not, however, impugn the coherence of the theory itself as a model for conceptualizing a "just" distribution. n152

Some would argue that the equal division model of justice could be satisfied without actually equalizing the distribution of LULUs; instead, communities enduring a disproportionate burden could be compensated. n153 The compensation could be in the form of money, or could involve community benefits like health facilities or recreational centers. n154 Under this theory, the LULUs might not be evenly dispersed, but the communities having more LULUs would have the disparity equalized by compensating benefits. Because each community is being equalized, though by differing means, this theory falls under an "equal division" theory. However, the compensation model presents numerous ethical and practical difficulties not raised [*1069] by the pure equal-division-of-LULUs' model. n155 Any future efforts to develop specific policies to pursue distributive justice under the equal division model will have to confront the issue of whether, and to what extent, compensation is an appropriate mechanism.

B. Evidence of Distributive Injustice Under the Equal Division Model

This Section considers the environmental justice movement's "broad" claim of distributive injustice: that, overall, LULUs have not been distributed evenly among demographic groups, and are instead more heavily concentrated in poor and minority communities. n156

There is some evidence that social service LULUs are disproportionately located. Services for the homeless, group homes for the mentally disabled, and community correctional facilities, like halfway houses, tend to be clustered in minority and low-income communities. n157 Moreover, the more undesirable the social service, the greater the concentration in poor or minority neighborhoods. For example, halfway houses are correlated with higher percentages of minorities and the poor than homes for the developmentally disabled. n158

Most of the studies in the environmental justice context have concerned polluting LULUs. n159 A recent book, Environmental Injustice in the United States: Myths and Realities, summarizes an [*1070] extensive review of the existing literature. n160 The authors note that many case studies in particular regions have identified localized distributional disparities, with some studies focusing on impacts to communities of color and others focusing on impacts to low-income communities. n161 The authors also review more quantitative statistical analyses of distributional outcomes. They observe that pre-1992 studies:

generally determined that race (either in the presence or absence of a control for social class) was associated with higher rates of exposure to environmental hazards for a variety of geographic areas, such as regions, counties, [Standard Metropolitan Statistical Areas], or ZIP codes, and for a variety of environmental harms, such as air pollution, solid waste, pesticides, hazardous waste, and toxins ... . n162

Other reviewers of this literature have likewise concluded that it reveals widespread inequities. n163

Myths and Realities states that post-1992 studies generally confirmed the trend of earlier studies. n164 Those that considered the [*1071] correlation of environmental harms with race and class found that race tended to be "the most important predictor of risk." n165 Many of the post-1992 studies analyzed correlations of environmental problems with not only race and class, but other factors as well. These studies tended to find that race and class were still important predictors of risk, although not always the most important ones. n166

Myths and Realities also presents the results of the authors' own research. Unlike many of the studies they reviewed, which focused on a single unit of analysis (e.g., zip codes or census tracts), their study focused on three different geographic levels: state, county, and city. n167

In comparing states, they considered a wider range of environmental harms than had been considered by other individual studies, which tended to focus on one or another harm, but not a range of harms together. n168 These harms included air pollution (two types), n169 hazardous waste (as in generation and disposal sites), solid waste, toxic waste (as in certain indicators of toxic releases), n170 and water pollution (two types). n171 Comparing levels of inequity among states, they found that African-American concentrations were correlated with five of the seven types of environmental harms they researched, n172 and Hispanics were correlated with four of the seven in some or most regions of the country. n173 Social class, measured by income and education levels, n174 was correlated with two of the seven [*1072] environmental harms n175 when associated with additional factors that also affected the degree of environmental harm. n176

In comparing counties, the study focused on one important form of environmental harm: toxic air emissions. n177 They found that, to a greater or lesser extent, both African-American and Hispanic counties were exposed to higher levels of toxic emissions than other counties. n178 For African Americans, the correlation was particularly strong in the Sunbelt states. n179 For Hispanics, the correlation was not present in many cases, but was strong where combined with "low social class, western regionalism, and low county fiscal capacity." n180

In comparing cities, the study compared cities with populations over 50,000, n181 and once again focused on toxic air emissions. n182 The study found a strong correlation between African-American cities and toxic releases, but did not find a strong correlation between Hispanic demographics and toxic releases. n183 The study also found a strong correlation between social class and toxic releases, n184 a correlation that was stronger on the city level than on the state and county levels.

Considering all of the analyses done of all of the respective factors at the state, county, and city level, Myths and Realities concludes that "evidence of race-based environmental injustice is evident ... in nearly 86 percent of the equations that focus on the percent black population." n185 For Hispanics, they report an incidence of environmental injustice in fifty percent of their analyses. n186 Social [*1073] class proved to be correlated with environmental harm in thirty-five to fifty percent of the equations, and was at times particularly relevant where class was analyzed together with race. n187 While the study does not reveal inequities at every geographic level with regard to every type of pollution analyzed, it confirms many of the widespread disparities observed in previous studies, particularly with regard to race, and especially with regard to African Americans.

An issue that is likely to arise in connection with the Myths and Realities study is the relatively large size of the units of analysis: the smallest unit, cities, is larger than that used by most other studies, which typically consider zip codes or census tracts. n188 Large units create potential "aggregation" problems: they may not reveal significant variations within the chosen unit. n189 For example, while there may be a high minority population in a county, that population is not distributed evenly. The environmental harms identified in the county may or may not be located near minority residents. The aggregation objection is apt if the primary issue is actual exposure to harm. Where broader issues are at stake, however, such as understanding the political, social, or economic dynamics of large-scale distributional patterns, larger units of analysis are appropriate. In general, it is worthwhile to have data about distributional equity at a number of different levels, and focusing on cities, counties and states reveals patterns that might not be revealed with a smaller unit of analysis.

To provide a fuller range than that provided by the Myths and Realities study I will briefly discuss the most significant nationwide studies on the distribution of hazardous waste sites that use smaller units of analysis. I focus on hazardous waste studies because hazardous waste site distribution is one of the few issues for which there are national studies.

The United Church of Christ Commission for Racial Justice ("UCC Report") conducted the first major national survey on the distribution of hazardous waste sites. n190 Using zip codes as the unit of analysis, the study found that the percentage of minorities near controlled hazardous waste facilities was higher than in communities [*1074] without such facilities. n191 Based upon a descriptive study, the UCC Report concluded that uncontrolled toxic waste sites are concentrated in minority communities. n192 The economic status of a community was also correlated with the presence of hazardous waste facilities, but was not as strong an indicator as race. n193 A follow-up report completed in 1994 concluded that the average percentage of minorities living near hazardous waste facilities had increased since the original 1987 UCC Report. n194

Some commentators have contended that, for various reasons, the UCC Report's reliance on zip codes was problematic, and that census tracts should have been used instead. n195 Two similar studies on the distribution of hazardous waste sites have been done using census tracts. The University of Massachusetts's Social and Demographic Research Institute ("SADRI") evaluated demographic data regarding the distribution of commercial hazardous waste facilities. n196 Unlike most of the studies referenced above, the SADRI study concluded that there was no statistically significant disparity between the minority populations in census tracts with hazardous waste facilities and those without. n197 The SADRI study did, however, find a correlation between facility distribution and certain socioeconomic factors. n198

[*1075] A subsequent study n199 using census tracts did, however, find a link between race and distribution, and criticized some of the pivotal assumptions made in connection with the SADRI study. Professor Been argues that the SADRI researchers' inappropriate elimination of many relevant census tracts from consideration "reduced the differences between the racial and ethnic composition of host and non-host tracts." n200 When the relevant census tracts were included in the analysis, Professor Been found that the percentage of African Americans or Hispanics in a census tract was a significant predictor of whether the tract hosted a hazardous waste facility. n201 Professor Been found that income, particularly very low income, was not a significant predictor of hazardous waste facilities. n202 The national hazardous waste studies thus present somewhat mixed results.

The results of the national hazardous waste studies, the Myths and Realities study, and the many additional studies that have been conducted depend upon important, but debated, methodological choices. n203 As indicated above, the unit of analysis is critical: state, county, city, zip code, census tract, or "neighborhood." I would argue that there is value to evidence from all of these levels, and that the appropriate unit of analysis depends upon the nature of the facilities being studied and the purpose of the study. A study evaluating the health effects from facilities having highly localized impacts would ideally use a small unit of analysis. Facilities having more widespread impacts would suggest the need for larger units of analysis. On the other hand, a study focusing on the equity of siting decisions, that is [*1076] not necessarily limited to evaluating the direct physical risks from the facility, could use a variety of units of analysis, including very large units, to identify the political dimension of siting decisions. n204

Other important issues include how to define a "minority" community, n205 and whether the baseline for determining disparate impact should be the national average or the average for the state or region under study. Researchers must also consider whether studies should focus on the presence of facilities versus actual risk from those facilities. n206 If the relevant inquiry concerns political and social equity, then the presence of facilities may be the appropriate factor. If instead, the inquiry concerns physical harm, then risk may be the suitable factor. n207 The science of statistical analysis is enormously complex, and there are many variations in the manner in which analysis is conducted that can affect outcomes. Researchers differ in their choices and the importance they attach to them. n208 The validity and conclusions of studies are regularly challenged. Although some studies are undoubtedly better than others, in many instances the controversies turn on differences of opinion for which there are few right answers.

Notwithstanding the disputes about methodology, the vast majority of the studies demonstrate some degree of inequity in the distribution of LULUs on the basis of race and/or income, with race being the more frequently relevant factor. n209 While these studies do not prove anything about the presence of distributive injustice in every locality, they do support the "broad" claim of distributive injustice - that the overall pattern of LULU distribution is unequal. This corresponds with the visual image of urban America: clean, [*1077] green, wealthy and white suburbs juxtaposed with poor and minority residents living near industrial facilities, homeless shelters, bus stations, and the like.

The evidence from the aforementioned studies tells us that the existing distribution of land uses does not satisfy the dictates of an "equal division" theory of distributive justice. Although the evidence says nothing about the causes of these inequities - whether they are neutral factors, discrimination in siting, broader social injustices, or post-siting housing dynamics - identification of causes is not necessary to prove distributive injustice. n210 Focusing on what communities now experience, the evidence reveals that the existing division of LULUs is pervasively unequal.

V. Distributive Justice Under the Community Preferences Model

A. Distributive Justice as the Equal Satisfaction of Community Preferences

Professor Lynn Blais has argued that this evidence of disproportionate distribution does not necessarily tell us whether the disparities are unjust. n211 As discussed above, the equal division model generally rests on the assumption that all communities measure the benefits and burdens of various land uses in the same way. If, however, communities value different land uses differently, then a different model of distributive justice may be appropriate: one that evaluates justice by whether people's preferences are equally met, not by whether they are all subject to the same number of disagreeable land uses. As indicated above, I term this the "community preferences" model of distributive justice. Under this model, disproportionate distributions would be just if the disparities matched differing preferences.

In addition to Professor Blais, others have implicitly suggested that an assessment of distributive justice ought to consider relative [*1078] preference satisfaction. n212 Implicit evidence of the theory is pervasive in many of the ongoing debates about environmental justice. For example, in response to proposed EPA guidance documents suggesting that environmental permitting agencies must take the disparate impact of permits into consideration or risk violating Title VI of the Civil Rights Act, n213 many commentators questioned whether this disparate impact approach would interfere with the ability of host communities to obtain desired facilities. n214 They thus assumed that a disparate impact - an unequal division - would not be unjust if communities desired the LULU at issue in a permitting decision notwithstanding its disparate impact. Coming from the left, Professors Eric Yamamoto and Jen-L Lyman have critiqued the environmental justice movement for failing to recognize the critical differences among various ethnic communities in their attitudes about particular land uses, and cautioned against absolute positions that ignore such distinctions. n215 They thus implicitly adopt the view that justice should be considered in light of the unique preferences of particular communities. n216

[*1079] Under the community preferences model, the critical question would therefore be the extent to which communities are equally satisfied with the land uses that surround them, not the extent to which the land uses and their impacts are equally distributed. n217 Different communities may have differing degrees of tolerance for the burdens imposed by the various types of LULUs. What might be a highly objectionable LULU to one community may be only moderately annoying to another. Some communities might rise up in arms against a small clinical hospital; others might be merely indifferent. Some communities may be more attuned to and concerned about environmental quality than others. n218

Moreover, many LULUs bring a mix of benefits and burdens. For some communities, the benefits could outweigh the burdens, with the net result that an arguably "undesirable" land use becomes, overall, a desirable land use. n219 Professor Blais reflects this possibility in her use of the term "environmentally sensitive land uses" rather than "locally undesirable land uses," n220 and in referring to possible "differentials" in facility distribution patterns rather than "disproportions." n221 The benefits could include direct services, such as medical care needed within the community. n222 Many LULUs, both industrial and service-oriented, could bring significant employment [*1080] opportunities to a community. n223 A significant enterprise could improve an area's tax base. n224 Moreover, some communities are plagued by abandoned properties, industrial and otherwise, that degrade the community environment. n225 A LULU might be considered an improvement on the existing environment, notwithstanding certain undesirable features. The movement to develop "brownfields" - former industrial properties - reflects the judgment that communities could improve their lot by encouraging the development of new facilities on existing abandoned, industrial properties. n226

The community preferences model is rooted in insights stemming from neoclassical economic theory. In a nutshell, neoclassical economic theory assumes that individuals have particular preferences regarding all key aspects of life, n227 and that they have the capacity to act rationally in seeking to satisfy those preferences. n228 The "market" is the medium through which individuals and other actors make choices, consistent with their preferences, about what to do and where to be. Under the neoclassical model, the ability of private actors to realize their preferences through the market is critical to [*1081] maximizing individual welfare. Maximizing individual welfare is, in turn, critical to maximizing overall social welfare. n229 Many fundamental allocative and distributional consequences flow from the realization of private preferences through the invisible hand of the market. n230 Social welfare is maximized by respecting the allocation. n231 Professor Blais's analysis assumes there is a "market" that operates to distribute land uses and/or residents in response to resident preferences. n232

In general, under this view, the market is superior to government-led distributional efforts. For example, the market preserves individual liberty by allowing individuals to make choices for themselves rather than having choices imposed by the government. n233 It is also the most efficient mechanism for insuring that people's differing tastes are satisfied, since residents know what they want better than the government, and are more likely to have their preferences met if they, rather than the government, make key decisions. n234

[*1082] One might argue that, even if free markets provide the ideal mechanism for sound organization, the concept is simply inapplicable in the land use siting context. As will be discussed in detail below, the land use siting process is not an unencumbered market in which residents choose their desired land uses, or even in which facility siters and communities negotiate freely toward preference-matching outcomes. n235 In many instances, LULU siting involves a political decision, whether by local, state, or other political entities. n236 Nonetheless, preference theory remains relevant. The neoclassical economic theories that have informed the market preference-based view of social welfare have a political analog: public choice theory. n237 Under public choice theory, the political forum is similar in function to the market: both serve as venues where private preferences are weighed and decisions maximizing private preferences could, at least theoretically, be realized. n238 While not everyone's preferences could [*1083] be satisfied by every decision, this model envisions an on-going sequence of decisions involving trade-offs and compromises that lead, in the long run, to an expectation that "everyone would enjoy a net balance of political gains in excess of losses." n239

As with the pure market model, market advocates would likely argue that the political choice model would perform its preference-meeting functions best if allowed to operate freely in response to each community's preferences. Government efforts to resolve distributional issues more broadly, through, for example, general restrictions on disparate impacts, would interfere with communities' capacity to meet preferences in individual cases. n240 Thus, whether the private or the public "political" market is at stake, community preferences will most likely be met and liberty will be preserved if the government does not impose an overarching distributional program.

Assuming a community preferences model, how does one measure whether its dictates have been met? Clearly, some degree of preference satisfaction is implicit in the model, or its very basis of justification - maximizing social welfare through maximizing varying individual preferences - would be lost. Even a strong proponent of the model, however, is unlikely to measure justice under this theory by requiring one hundred percent resident satisfaction. Having all LULU siting decisions hinge upon the unanimous consent of all members of affected communities would make for poor public policy. There may be a variety of factors that could justify a siting decision notwithstanding community opposition. There are undoubtedly more LULUs than there are people who want them. Furthermore, if a LULU is undesirable to all, it may be impossible to find a consenting community. Where the undesirable LULU is necessary, requiring an [*1084] unattainable community assent could interfere with the ability to resolve important social and environmental problems. n241

Thus, while some degree of preference satisfaction is necessary to the attainment of justice under the community preferences model, the critical issue is whether the ratio of satisfaction to dissatisfaction is equivalent. If some communities were able to have more of their preferences met than other communities, then we would not have distributive justice under the community preferences model.

As with the equal division model of distributive justice, the challenge of measuring fairness pursuant to this theory is formidable. This theory shares all of the practical difficulties in defining "communities" that the equal division model presents. n242 In addition, preferences are frequently individual and subjective. As such, the task of determining what a "community" prefers presents a daunting task. It is not clear how one would identify a community's "preference," or how the differing views within a community would be assimilated into an overall level of preference, or who would speak for a community in attempting to gauge a community's relative satisfaction with its LULUs or the prospect of future LULUs. n243 Although these are major challenges that must be faced by anyone choosing to design an implementing policy, they do not undermine the coherence of the overall theory.

To what extent do advocates of the community preferences model believe that its dictates have been met? Professor Blais implies that the community preferences model is sufficiently satisfied - or, at least, satisfied sufficiently to preclude the need for government involvement. She has concluded that:

a community decision to act as host for an environmentally sensitive land use, or a private determination to live near [*1085] one, can be viewed as rational and informed, and the political and market determinations leading to the current distribution of environmentally sensitive land uses can be understood to accurately reflect rational and legitimate private preferences and collective judgments. n244

To be sure, Professor Blais, like most neoclassical economists, acknowledges that the market does not always work perfectly to maximize the satisfaction of private preferences. n245 Inequalities in wealth impact individuals' abilities to realize their preferences, n246 as do persistent problems of discrimination. n247 Even if it is not perfect, however, neoclassical economists believe it is often better to let the market operate than to interfere in its operation through an across-the-board government effort at improving distribution. n248 Government intervention poses a threat to the liberty and efficiency goals achieved by the market. n249 Government efforts to control siting would be paternalistic, questioning communities' ability to choose desired land uses for themselves. n250 Moreover, the government [*1086] suffers from so many of its own functional failures that there is no assurance that it will function better than the market. n251 Professor Blais concludes that "while disquieting vestiges of social injustice remain in our society, no convincing argument has been made for rejecting measured preferences concerning environmentally sensitive land uses." n252

This conclusion raises the critical question of whether the market in fact meets community preferences sufficiently to preclude consideration of other initiatives to guide LULU distribution. If it does not, then the market's goals of achieving "liberty" and "efficiency" are not being furthered because residents are not able to have their preferences met, or at least met equally. Neither individual nor social welfare is being maximized. n253 If preferences are not being met, or are being met unequally under current market conditions, then exclusive reliance on the market will not serve to meet the goals of the community preferences model.

Before analyzing the likelihood that preferences are met, a few words on the preferences model itself are in order. Although this Article takes the community preferences model as a given, it is worth noting that fundamental critiques could be leveled against it. If unequal distributions of LULUs were to cause serious harm to a community, it is not clear that preferences for the LULUs should be honored. n254 More broadly, preferences are a function of social context, and if that context is unjust, the resulting preferences may not form a legitimate basis for social decision-making. n255 A community's preferences and, more particularly, its weighing of the benefits and costs of LULUs, may be significantly shaped by factors such as relative wealth and race. n256 On a fundamental level, poorer [*1087] communities may place a lower value on environmental quality than wealthier communities because other needs may appear more pressing n257 and because they may not have the knowledge or resources to recognize the full harms associated with certain LULUs. n258 Poor and minority communities may also value the purported benefits of LULUs more highly due to their need for employment. n259 They could be more likely to welcome the siting of a LULU, or move toward one, than those with greater economic options and security. It is thus questionable whether it is "just" to allow a distribution that places more LULUs, with their inevitable burdens, in poor and minority communities when the differing preferences are borne of disparities in wealth and a legacy of discrimination.

The preceding suggests why it may be inappropriate to accept certain preferences for LULUs. It may also be inappropriate to accept certain preferences against LULUs. Opposition to low-income housing, group homes for the disabled, and the like may be born of prejudices that the legal system should not simply accept as given. n260 More broadly, one could argue that community preferences against LULUs are obstacles to be overcome, not satisfied. n261 Widespread [*1088] local opposition to unwanted land uses has come to be known as the "Not-In-My-Back-Yard," or "NIMBY" phenomenon. n262 Communities who cry NIMBY are often considered selfish, short-sighted, and an obstacle to providing needed social services and other necessary but undesirable land uses. n263

This Article is not, however, the place to determine the ultimate legitimacy of theories of distributive justice. The benefits and drawbacks to the equal division and community preferences models, and their role in specific public policies addressing environmental justice, will be the subject of future work. Instead, this Article addresses an important preliminary inquiry: whether the evidence of distributive injustice under the equal division model is moderated or eliminated by using a community preferences rather than an equal division model of justice.

B. Assessing Distributive Injustice Under the Community Preferences Model: The Need to Review Process to Assess Distribution

Under the community preferences model of distributive justice, an unequal distribution of LULUs would not necessarily constitute distributional injustice. Everyone might be equally satisfied, notwithstanding different exposures to LULUs. n264 The question, then, is how to determine whether we have justice under the community preferences model.

Direct evidence on relative satisfaction of preferences is impossible to obtain; it is not feasible to research the extent to which siting decisions actually meet community preferences for every [*1089] LULU around the nation. The closest one can get is to analyze the likelihood that preferences have been met, and met equally. To analyze this likelihood, one must evaluate the processes impacting land use distributions to determine whether the processes are likely to have generated equitable distributions under the community preferences model. While this Article is primarily devoted to the question of fair distribution, it considers processes to help predict likely distributions, not to evaluate the processes in and of themselves.

The inquiry first focuses on the land use siting process to determine what it tells us about the likelihood that the resulting distributions meet preferences. An important threshold question is whether there is evidence that land use siting decisions had a disparate impact when they were made; n265 the data about the disparate distribution of LULUs discussed above address only the current demographics, n266 not the demographics at the time of siting. However, these studies are limited, and they tell us only about distribution, not about preferences.

To evaluate the likelihood that preferences are equally met, the Article identifies the factors that generally govern the initial distribution of land uses, evaluate the extent to which community preferences are likely to play a role, and, to the extent community preferences do play a role, evaluate whether they are satisfied equally. The analysis first considers market forces, regulatory requirements, and other types of "objective" factors that may influence the placement of LULUs. n267 The Article then analyzes the political process associated with siting decisions, n268 considering both general zoning provisions n269 and the site-specific political determinations that are often made in connection with controversial land uses. n270 Finally, it considers whether statutes that explicitly require public participation in the siting process or in permitting decisions are likely to increase the extent to which community preferences are taken into account equally. n271

The Article then evaluates whether the post-siting housing market is likely to lead to distributions that satisfy community [*1090] preferences equally. n272 Regardless of what occurs at the siting stage, the market could, conceivably, remedy any distributional inequities that arise from the siting process. Residents could move toward or away from LULUs depending upon their preferences, thus leading to an equal satisfaction of preferences. The analysis also considers whether post-siting housing market dynamics have the opposite effect: whether they are likely to exacerbate rather than improve disparities in the extent to which the distribution of LULUs meets community preferences equally.

Although the primary thesis of this Article is that distributive injustice is worthy of concern in its own right, political and social injustice provide additional (though not necessary) arguments for addressing disparities. Since the discussion of siting processes and the housing market will inevitably expose many of the political and social causes of distributive injustice, I also note below how they provide additional grounds for addressing distributional disparities. n273

VI. Is the Land Use Siting Process Likely to Satisfy Community Preferences Equally?

A. Evidence of Siting Decisions' Disparate Impact

A first issue to consider in assessing the land use siting process is the extent of the disparities resulting from that process. While the presence or absence of disparities does not indicate whether they are the result of differing preferences or not, it would at least provide a starting point for analysis. The demographic evidence presented above suggests that many LULUs are currently disproportionately located in poor and minority neighborhoods. n274 But since the current demographics could have evolved after an initially neutral siting decision, evidence of current demographics does not prove that the original siting decisions had a disparate impact. Unfortunately, however, the data on the impact of actual siting decisions are too limited to be of significant use.

Professor Been has conducted several analyses of landfill sitings. n275 Two small and fairly localized studies did indicate [*1091] disproportionate impacts on African-American neighborhoods. n276 Professor Been's largest study, of 608 hazardous waste sites sited between 1970 and 1994, n277 revealed that, during this time period, the locations selected were not disproportionately African-American or poor, but that they were disproportionately Hispanic. n278 Nevertheless, the study noted that African-American communities are more likely than others to host a hazardous waste landfill. n279 The current disparity may be a consequence of disproportionate siting or post-siting housing market dynamics that took place prior to 1970, the beginning of the study period. n280 The study observed that working class and lower-middle income neighborhoods were more likely to be [*1092] chosen for hazardous waste facilities than poor or wealthy communities. n281 Although the studies provide some evidence of disparate siting decisions, they are limited in extent and do not indicate whether disparities were a function of differing preferences.

B. Overview of Land Use Siting Processes

The following discussion of land use siting processes is intended to determine whether they are likely to meet community preferences equally. The analysis does not address whether land use siting processes should address community preferences; instead, it critically evaluates the contention that they do so fairly.

If communities and LULU proponents negotiated on an equal footing about where to put various land uses, then, at least theoretically, the resulting distribution might reflect community preferences and interests. n282 In fact, that image of an affected community's role does not correspond with the complex reality of land use decision-making. In general, the primary players are facility promoters and regulatory entities, with some role for underlying municipal zoning decisions. The communities in which facilities are to be placed often play only a tangential role in the siting process.

Where private LULUs are at issue, such as various types of industry and many hazardous and other waste facilities, the industry itself plays a key role in selecting site locations. n283 State or local governments are critical in siting the many LULUs that are public facilities, such as landfills, prisons and jails, and social services like hospitals and homeless shelters. Other high-profile LULUs, like some hazardous waste facilities, may involve a mix of private and public control. n284

But site selection does not occur in a vacuum; it involves not only a search for those characteristics deemed desirable by the facility [*1093] proponent, but meeting whatever local, state, and/or federal legal requirements will ultimately be imposed. Siting processes thus often involve a strong political or regulatory component, regardless of whether they are initiated by a private or public entity.

Not surprisingly, siting processes vary tremendously depending upon the nature of the facilities and the infinite variation of federal, state, a