Article body


Political theorists have recently turned their attention to the city. Professor Daniel Weinstock, for example, has argued that we can profitably examine well-established issues in political theory in light of city-specific concerns. The city as an object of inquiry is distinct, argues Weinstock, because of its unique spatial circumstances. The city has a scale, and city dwellers live in a degree of proximity, that transforms sometimes abstract questions of political theory into more concrete and specific inquiries.[1] In this paper, I join this general project of theorizing the city but do so from the perspective of a public law scholar, and I believe that this perspective opens up lines of analysis that may not obviously present themselves to political theorists.

Two assumptions—that law and political theory are distinguishable and that legal arguments in public law raise institutional questions—inform the analysis undertaken in this paper.[2] I begin with the assumption that a legal theory of the city is not identical to a political theory of the city. The distinction arises because what the law is or can be in relation to a particular city is not identical to what is or can be consistent with the demands of a particular political theory. This distinction exists whether one conceives of cities as communities of interest or as specific jurisdictions defined by law. Scholars who argue for the first conception have proposed various measures, such as the census metropolitan area, to define a community of interest.[3] But a city can also be understood as, and identified with, a municipality. For example, we might understand the city of Victoria to be that entity which has been legally defined as a municipality, with all the powers that have been delegated to it by the province of British Columbia. Whether one conceives of a city broadly as a community of interest, or more specifically, as a municipality, laws shape cities, and the relevant issues of law are distinct from those of political theory. For instance, the question of what kinds of instruments and institutions should govern cities, when they are conceived of as communities of interest, is a central preoccupation of legal scholars, as we shall see in Part II, and this design preoccupation is distinctively legal in nature. The significance of law to cities is also striking when we consider cities, defined as municipalities.

As a matter of Canadian constitutional law, municipalities are creatures of the provinces, and therefore a province may create or destroy a city when a city is understood to be identical to its legal definition as a municipality.[4] Of course, this fact does not preclude one from asking whether the cities that are created by law are just or fair, nor does it prevent one from prescribing conditions for cities that are different from those of actually existing cities. What this fact about the legal nature of cities does suggest is that if one wants to advance a criticism or analysis of cities, one should be cognizant of the legal institutional context within which they exist. To criticize a city for not being what it cannot possibly be (where law sets the limits of possibility) strikes me as a somewhat futile exercise. In addition, the fact that cities exist within a legal context suggests that prescriptions for changes to cities should exhibit awareness of this context. Because cities are creatures of law, there is limited utility, in my view, in offering prescriptions that are ignorant of what is legally possible.

These claims may seem self-evident, but authors in disciplines other than law have made arguments that either misrepresent the legal status of cities in Canada or offer prescriptions that go beyond the range of what is legally possible. For instance, nonlegal scholars have argued that Canadian cities should benefit from home-rule status because, they argue, such status would prevent the provinces from intervening in city affairs.[5] These scholars draw their conception of home rule from the American experience.[6] This argument is flawed because it is mistaken as to what cities are and what they can be as a matter of law. The first thing to note is that these scholars do not understand what home rule is. Home rule, in its various forms, does not completely insulate cities from intervention; rather, it sets out a requirement that a higher-order government be explicit when it intervenes in the regulation of a municipality,[7] or it delineates the conditions under which municipalities can exercise authority that is not explicitly delegated to them from a higher political order.[8] Moreover, far from being radically different from home rule, legislation in Canada yields effects that are identical to those of home-rule provisions in the United States.[9] Finally, the general aim of completely insulating cities from provincial control is constitutionally impossible because, as we have seen, cities are creatures of the provinces.

I should be clear about the limits and the potential of my claims. I do not mean to say that we cannot study cities from the perspective of disciplines other than law. Indeed, given the complex nature of cities, it is probably futile to attempt even a legal analysis that ignores the insights of other disciplines. I only mean to point out that any serious analysis of cities should be cognizant of the relevant legal constraints. Cities can be many things, but they are necessarily creations of the law. This fact suggests that if one is to offer prescriptions to change cities, one should be sensitive to the legal implications of such prescriptions, and this awareness requires, at a minimum, an understanding of what is legally possible. But sensitivity to legal considerations involves more than simple awareness of what the law permits; it also involves an understanding of the kinds of institutional concerns that are omnipresent in public law scholarship. Let me turn now to these concerns, as they will motivate the arguments that follow.

The title of this paper refers to a federal legal theory of the city because I believe that cities have aspects that open themselves to a variety of legal arguments and theories. By a “legal argument”, I mean an argument that informed participants in legal discourse would recognize as valid, and by a “legal theory”, I mean any theory that supports such an argument.[10] Public law scholars have noted that these arguments and theories tend to involve judgments about legal institutions and prescriptions for changing these.[11] There are a variety of possible legal arguments and theories that can be applied to cities. One might conceive of cities in property law terms. For instance, Professor William Fischel has argued that one key feature of regulation in cities—namely, zoning regulation—should be understood as a collectively held property right.[12] Specific kinds of theoretical claims support such an argument, and prescriptions that flow from the argument focus on the institutions of property law.[13]

One might also conceptualize cities in administrative law terms. In another paper, I have argued that municipal institutions should be examined in light of standard administrative law concerns.[14] Administrative law scholars have long been concerned with how to ensure that public agencies are structured in ways that are democratically legitimate and in ways that are responsive to citizens who are regulated by those agencies.[15] I have argued that municipal governments and their institutions—which as a matter of law are bodies exercising delegated authority—should be designed in ways that are responsive to citizens. The existing legal debate in administrative law provided me with the argument’s terms and underlying theories, and the prescriptions I offered were resolutely institutional in nature. In this paper, I advance a different kind of legal argument about the design of public institutions and a different set of institutional proposals. I will claim that debates about cities and metropolitan regions can be and have been conceived in federalism terms, and I will offer institutional design prescriptions that are responsive to the concerns that underlie these terms.

By a “federalism argument”, I do not mean an argument about the place of cities within a federation, although this is an interesting legal question that I will discuss in other work. Rather, I mean an argument that addresses issues typically raised in legal debates in the federalism context. Such debates turn on theoretical questions of economic efficiency, regulatory coordination, and democratic legitimacy, and they arise in circumstances where political authority is divided, typically along overlapping geographic lines.[16] Furthermore, federalism debates often raise questions of institutional design, including some that involve the configurations of legislative or administrative bodies, and this paper will share that institutional focus. I am not alone in claiming that there is a conceptual relationship between debates about federalism and conceptions of the city. Indeed, law and economics scholars have regularly transposed theoretical insights that were originally generated through analyses of local governments to examinations of the relationships between states or provinces, and between those political orders and federal governments.[17] In addition, political theorists and legal scholars have applied questions of institutional design and conceptions of political obligation that were originally articulated in theories applied to federal states to analyses of local governments.[18]

In Part I, we will see that debates in the local-government law literature between regionalists and localists involve disagreements that sound in the language of federalism disputes. The underlying theoretical claims of the positions in those debates will be subject to close examination. In Part II, I will argue for a particular kind of institution that accommodates and is responsive to the range of concerns expressed in the localist-regionalist debate. I will argue that British Columbia’s regional district system resolves many of the contested issues in the localist-regionalist debate and that that system can be conceived of in federalism terms. The paper as a whole reflects the public law scholar’s sensitivity to legal argument and to legal institutions. In order to set the stage for the analysis, I turn now to examine the motivating question and the terms of the localist-regionalist debate.

I. Localism vs. Regionalism

How can metropolitan regions be governed effectively, legitimately, and equitably? This question and its variants have long been the focus of the local-governance debate in Canada and the United States.[19] Most previous responses have devalued either legitimacy or fairness norms while privileging the others, and the participants in the debate have offered competing definitions of effectiveness. The present paper departs from this trend by positing a theoretical and institutional framework that balances and affirms all three values and both conceptions of effectiveness.

The local-governance debate in Canada and the United States has, for the most part, coalesced around two poles. On one side are localists and on the other, regionalists. The former are proponents of multiple, autonomous local governments, while the latter advocate for metropolitan-wide or regional governments.[20] Localists stress the values of local-governance legitimacy, measured in terms of citizen participation in governance, and effectiveness, measured by the extent to which public-good provision matches consumer preferences.

Regionalists stress the values of distributive fairness and effectiveness, which is defined in terms of metropolitan regions’ global competitiveness. The localists tend to underplay the significance of distributive concerns, while the regionalists tend to downplay legitimacy issues, and each side privileges its own conception of effectiveness. By contrast, the present paper advocates for an institutional framework in which governance bodies of various scales coordinate to deliver effective and fair distributions of social and material resources, as well as the participatory goods associated with participation in local governance. Let us turn now to examine with greater care the localists’ arguments.

A. American Localism: Efficiency and Legitimacy

The starting point for the efficiency argument in the United States is Tiebout’s seminal 1956 article “A Pure Theory of Local Expenditure.”[21] Tiebout’s paper has been interpreted to support claims for local-government autonomy and was a response to Samuelson’s claim that there is no rational, market-driven means of allocating public goods.[22] According to Samuelson, because public goods are nonrival and nonexclusive, there is no way to allocate them to match consumer preferences.[23]

Tiebout countered that through the mechanisms of exit and voice, individuals can express their preferences for public goods, and local governments can respond to these. This allocative system requires a plurality of governments that offer competing sets of goods. Individuals in such a system will gravitate toward those local governments that reflect their preferences and exit those that do not. In turn, individuals will voice their preferences through voting, and fashion polities that reflect and reinforce these preferences.[24] The net result is a collection of local governments that are internally homogeneous but that collectively express a diverse set of preferences. This argument resonates with arguments made in the federalism context that conceive of subfederal units as market actors and that valorize decentralization in order to ensure an efficient marketplace of government services and a matching of government services with citizen preferences.[25]

Tiebout’s arguments have been interpreted to support the claim that local-government systems characterized by strong local governmental autonomy are preferable to systems with regional-oversight institutions.[26] Tiebout-influenced localists argue that competition among local governments maximizes efficiency as it permits the largest number of preferences to be expressed and satisfied. In the United States, this efficiency argument has sometimes been joined by normative claims about the inherent democratic value of local government. Such claims flow from a political tradition that values local control over, and participation in, government institutions.[27] And arguments in this tradition have often been framed in organic terms: some normative localists claim that moral communities, which pre-exist the drawing of jurisdictional boundaries, should be protected against centralizing forces.[28] Similar arguments about the virtues of self-government have been forcefully articulated in the American writing on federalism.[29]

Perhaps the most convincing articulation of a position that marries efficiency and legitimacy claims comes from Fischel. According to Fischel, American homeowners are passionately engaged in local government because they seek, through zoning regulations, “insurance” for their homes. Home values depend on their surroundings, and it is difficult, if not impossible, to purchase private insurance to protect against neighbourhood change. As a consequence, local government becomes the vehicle through which homeowners control their surroundings in order to protect home value.[30] Under Fischel’s account, legitimacy and efficiency concerns converge as citizen participation becomes an efficient market mechanism.[31]

Professor Edward A. Zelinsky adds to this defence of localism an affirmative case against regionalism. According to Zelinsky, if regional government were truly efficient, regional institutions in the United States would already exist due to citizens’ preferences, and regionalist policies would already have been put in place by states.[32] Zelinsky also argues that although the idea of regional governments finds favour among intellectual elites, it achieves no purchase among the populace, and is therefore antidemocratic.[33] Moreover, Zelinksy challenges the empirical studies that have been used to support regionalism and argues that there is no factual basis for the claim that regional governments reduce inequities within a metropolitan region.[34] Finally, he argues that even if regional governments were to come into being, they would be dominated by elites who are distant from local populations and, as a consequence, would be both less responsive and less legitimate than local governments.[35] This final concern about the possibility of special-interest capture also finds expression in public-choice-inflected theories of federalism. Scholars in that theoretical context examine the incentives that influence official decision making at the state and federal levels of government.[36]

B. Canadian Localism: Efficiency and Legitimacy

Canadian localists echo the arguments of their American counterparts. The Canadian resemblance to American efficiency localism is perhaps most evident in the work of Professors Andrew Sancton and Robert Bish. Sancton suggests that market competition among localities disciplines local government bureaucracies and justifies a multiplicity of local governments; he therefore counsels against the adoption of single-tier, regional governments.[37] Bish similarly rejects regional governments on efficiency grounds. He accepts the premise that competition among localities and the choices of local populations will yield more efficient governance outcomes than regional governments but is careful to note that the extent of the efficiencies gained depends on the kinds of services and decisions at issue. Bish argues that some services require regional co-operation, but he favours loose, interlocal, co-operative measures and rejects regional governments.[38] This approach to intergovernmental relations resonates with some public-choice-influenced writing in Canadian federalism.[39]

As in the United States, arguments in favour of local governments, based on democratic grounds, have been made within the Canadian political tradition. A tradition of thought strongly supporting local governmental autonomy has long thrived in the Canadian context and was evident in the earliest moments of Canadian national history.[40] New England loyalists fleeing the American Revolution were deeply familiar with the tradition of local governance exemplified by the town hall, and they sought to import that tradition into their new surroundings.[41] The canonical source for this tradition is Lord Durham’s Report, which advocated for the creation of a strong municipal level of government for the colony in British North America and extolled the democratic virtues of local government.[42] For the authors of Lord Durham’s Report, as for de Tocqueville, a municipality was the level of government that most closely expressed the democratic will of the governed and functioned best as the training ground for democratic participation. Through participation in local governance, citizens could learn about democratic processes and could apply those lessons to larger political institutions. These kinds of arguments in favour of provincial autonomy resonate to this day.[43]

C. The Regionalist Response

The arguments against localism and in favour of regionalism have functionalist and equality variants. The functionalist regionalist argues that metropolitan regions require coordinating levels of government, the extent of whose authority is defined by those regions’ boundaries, and that the governance challenges facing metropolitan regions are co-extensive with those boundaries.[44] Once again, these arguments resonate with those made in the federalism context. In the federalism literature, federal interventions are sometimes justified as a means of ensuring policy coordination where only federal regulation can achieve such coordination.[45] In the local government context, these challenges include environmental protection from the effects of urban sprawl[46] and coordinated economic development.[47] The latter is especially important, it is claimed, because metropolitan regions have surpassed states, provinces, and national governments as focal points of economic activity. Metropolitan regions require coordinated regional governance, the regionalist argues, to compete with their global peers.[48] At least to the extent necessary to meet these functional challenges, the public-choice model of multiple local governments surveyed above is rejected.

A second set of arguments in favour of regionalism is framed primarily in equality terms. Authors setting out this variant of the regionalist argument argue that regional mechanisms should effect redistribution or at least reduce coordination problems.[49] The federalism echoes are particularly striking here: in the federalism context, authors similarly argue for economic redistribution to redress inequalities among subfederal units.[50] In the metropolitan context, the equality regionalist argues that a system comprised of multiple, autonomous local governments leads to two kinds of inequalities. The first has primarily an economic dimension and results directly from legal regulation. This species of inequality has similar effects in the United States and Canada. In both jurisdictions, wealthier municipalities enact land-use policies that exclude lower-income families and individuals through the regulation of the density of uses.[51] Such policies have multiple negative effects. Poverty is typically concentrated in urban or older or “inner” suburban municipalities.[52] These municipalities have higher demands for social services than their wealthier suburban counterparts, and since local governments in some American jurisdictions assume at least part of the burden of providing these services, higher costs are imposed on them.[53] However, because their residents have low incomes, these municipalities require higher tax rates to meet these costs.[54] As a consequence, lower-income individuals bear a higher burden than their wealthier counterparts, and lower-income municipalities are unable to attract higher-income residents, who are deterred by these higher tax rates.[55] In addition, since school districts can cross municipal boundaries and school-district taxes tend to be constant across the district, the effect of such taxes on lower-income families is regressive.[56] In Canada, an analogous policy concern is to ensure that differential tax rates among municipalities do not result in burdens that are concentrated in one municipality or several municipalities when considerations of fairness may require that the burdens be shared across a region.[57]

Further, because in the United States lower-income localities have greater need of tax revenues and employment opportunities than do their wealthier counterparts, they are more likely to give incentives to manufacturing industries to establish plants in their neighbourhoods and engage in destructive interlocal competition to attract them.[58] And even without such competition, these plants are disproportionately sited in lower-income localities, which have less political power than their wealthier neighbours to resist them. As a consequence, lower-income residents suffer disproportionately from resulting environmental and health harms.[59] In Canada, authors have argued that government coordination has reduced the risk of this kind of interlocal competition.[60]

A second, racial or ethnic-identity dimension to intermunicipal inequality results indirectly from judicial and legislative endorsement of unconstrained local autonomy. Since class distinctions in the United States have a tendency to map onto racial and geographic ones, typically it is minority populations who bear the burden of wealthier locales’ exclusionary policies.[61] Judicial remedies for this form of inequality are limited. Given that the United States Supreme Court’s equal-protection jurisprudence requires a finding of governmental intention to discriminate for there to be a breach of the Fourteenth Amendment, there is no remedy under the federal Constitution for the effective segregation that results from exclusionary zoning laws.[62] The Court has held that zoning laws that aim to create agreeable neighbourhood environments for predominantly white, middle-class families do not evidence an intention to discriminate against nonwhites, and therefore do not violate equal protection. By contrast, state constitutional “general-welfare” provisions and federal fair-housing legislation have been the basis of challenges to unequal distribution of low-income housing and have incidentally aimed at remedying racial inequality.[63] The extent and success of such litigation has, however, been mixed.[64]

The Canadian context, born of a different history, gives rise to a different picture of minority interests in the context of urban disadvantage. Some have argued that Canadian provincial governments have been comparatively open to regional governance mechanisms because the racial divisions in Canada are less severe and do not typically manifest in concentrations of disadvantaged minority populations in urban centers.[65] In particular, authors point to differences between the United States and Canada in the history of race relations to explain the relatively less segregated Canadian urban landscape. And in a less economically and racially segregated environment, it is argued, there are fewer incentives for exclusionary zoning, and concomitantly, less resistance to regionalist institutions. This descriptive claim about the lack of spatial segregation should be nuanced, as empirical studies have shown that urban Canadian Aboriginal populations have experiences similar to those of minority populations in the United States.[66] Much like disadvantaged urban minority communities in the United States, urban Aboriginal communities in Canada suffer from a history of severe disadvantage and from the consequences of exclusive land-use policies. In urban Aboriginal neighbourhoods, particularly in some western Canadian cities, low-income housing is concentrated. Some scholars have, however, argued that regional governments have been introduced in Canada primarily because provincial governments have sought to attain cost-savings,[67] a motivation that is unrelated to racial and economic segregation. Whether one accepts the explanation grounded in racial and class segregation, or the arguments about provincial cost-saving motivations, the relative openness of Canadian metropolitan centres to regionalism is accurate. Now that we have surveyed the terrain of the localist-regionalist debate in North America, we are in a position to assess that debate.

II. Resolving the Debate: New Regionalism and Regional Districts

At the core of the localist-regionalist debate is a set of incommensurable normative terms. As we have seen, the localist asserts the value of an efficient marketplace of local governments and claims the normative significance of local participation, whereas the regionalist asserts the value of effective regional co-operation and claims the normative significance of inequality in the metropolitan centres. Some counter-arguments are sent across the divide. For instance, Professor Lee Anne Fennell argues that the Tiebout model and its adherents downplay the social and economic costs of externalities generated by exclusionary zoning policies;[68] Zelinsky argues that the empirical evidence upon which the regionalist case for equality rests is weak;[69] and Bish argues that there is no basis for the claim that regional governments are necessary for effective regional coordination.[70]

Moreover, the regionalist arguments about the economic importance of regional coordination can be articulated in localist terms. If we accept the premise that improvements to a region will benefit “favoured quarter” municipalities, it may be that although the favoured quarters, as well as the entire region, will benefit from regionalist measures, the favoured quarters face a prisoner’s dilemma. For any particular favoured quarter municipality, under this scenario, the benefit to be gained from free-riding is greater than that to be gained through co-operation, and the loss to be suffered from participating and bearing the costs of others’ free-riding outweighs the cost of nonparticipation.[71] Alternatively, it may be that collectively, the less-favoured quarters have more to gain from co-operation than the favoured quarters do from opposition but that individually, each less-favoured quarter has less to gain by participating than each favoured quarter has to gain from defection.[72] In either case, the incentives for the favoured quarters point to non-co-operation and in both cases, the outcome of non-co-operation is inefficient as this leads to outcomes that are sub-optimal, in the sense that municipalities are blocked from achieving preferred outcomes.

These kinds of arguments, which bridge the localist-regionalist divide are plausible. There is, however, a deeper disagreement that is sometimes made evident in the exchanges between localists and regionalists. After making his affirmative case, the localist adopts the stance of the public-choice realist: even if regionalist proposals are normatively desirable, they are antidemocratic because the polity does not, in fact, desire them.[73] As a concession to the normative force of regionalist arguments, some localists offer weak other-regarding palliatives.[74] After making her affirmative case, the regionalist adopts the stance of the moral idealist: even if regionalist proposals are not practicable given the polity’s resistance to them, the ends sought are “the right thing to do.”[75] And as a concession to the accuracy of the localist description, some authors prescribe weak regionalist medicine that concedes much to localist impulses.[76]

A. Beneath the Impasse

Underlying the debate between the regionalist and the localist is a disagreement over the nature of democratic legitimacy. On the one hand, democratic choices are imagined to be legitimate only when parties take a disinterested stance, abstracting themselves from their narrow self-interests. Professor Richard Thompson Ford has called this view “civic republicanism”. On the other hand, democratic choices are imagined to be just precisely when they give the fullest expression to parties’ or groups’ self-interested desires. Ford has called this view “interest-group pluralism”.[77]

The first, civic republican position emerges in the regionalist literature as authors appeal to institutional forms as means of shaping and realizing the deep commonalities that parties in an urban region share and as they appeal to equality norms, even though they acknowledge that there is no popular support for their proposals.[78] The localist vision of municipal governance is easily accommodated under the interest-group pluralist conception of democratic legitimacy: not only is it efficient for local governments to pursue their self-interest, it is just to do so.[79]

The problems with each of these positions are significant. The first position presupposes the existence of normative commitments that are shared among members of a metropolitan region and assumes that regional institutional forms can give concrete expression to these commitments.[80] These assumptions are subject to two lines of criticism. First, deep fissures typically run through norms that are characterized as communal.[81] Moral pluralism cannot be done away with by asserting the importance of norms or by simply positing institutional forms that assume a shared normative commitment.[82] The failure of such an approach to urban governance is evident in American regionalists’ resignation to the irrelevance of their arguments in the design and functioning of contemporary metropolitan organizations.[83]

A second, institutional set of criticisms can be directed against the regionalist position and against the regionalist struggles to engage the critique on its own terms. In the face of metropolitan fragmentation, the regionalist proposes regional governments, which can be either single tier or two tier in form.[84] The single-tier form of metropolitan government is most vulnerable to the legitimacy critique and to the charge of inefficiency.[85] Faced with single-tier government, especially where this precludes other smaller-scaled bodies, citizens are deprived of the opportunity to participate in de Tocqueville-style local democracy and to shop in the marketplace of local government service providers.

The proponent of two-tier forms of government can respond to this critique but is rendered vulnerable to another one. Two-tier governmental schemes allocate regional competences to a higher, regional level of government, and allocate local functions to lower levels of government. The two-tier regionalist can plausibly argue that such institutional forms are appropriately scaled to the size of metropolitan residents’ citizenship and consumer capacities. To the extent that a metropolitan resident is faced with regional-scale governance issues, she is a member of a regional community and the only appropriate set of government services to be consumed are those that have a regional scale. For all other issues the lower level of government promotes her roles as local government citizen and local government consumer.[86] Two-tier metropolitan governments are therefore not as vulnerable as are their single-tier brethren to the localist’s legitimacy and efficiency attacks. They are, however, open to indirect attack. Bish has noted that it is often difficult to demarcate clearly the lines between regional and local subject matters and, as a result, the levels of government are engaged in constant conflicts over jurisdiction. He notes that in cases where two-tier governments have been put in place in the United States and Canada, they have moved inexorably toward a single-tier form.[87] Bish concludes that two-tier governments are an unstable form of government that set metropolitan regions on the slippery slope toward single-tier forms, complete with the legitimacy and efficiency problems that these forms give rise to.[88]

The localist position and its vision of democratic legitimacy is similarly open to normative and institutional criticisms. If the regionalist model too easily explains away disagreement, the localist approach presumptively valorizes disagreement, and supports practices and outcomes that are widely thought to be objectionable. Any theory that asserts the value of self-interested rent seeking, undertaken to the detriment of vulnerable parties, ignores common moral intuitions.[89] Moreover, it is a serious task to propose institutional frameworks in which it is possible for those affected by existing institutional arrangements to question whether they desire them and their consequences, and even if they do desire existing arrangements, to determine whether that desire is normatively defensible.[90] If the regionalist cannot wave away moral disagreement, neither can the localist ignore morally problematic outcomes.

A second criticism can be directed against the factual assumptions of the localist position. The localist describes residents as consumers and communities as market providers of service and benefits packages.[91] But communities are constituted by legal regimes: consumer communities and, by extension, consumer residents do not pre-exist legal regulation. Although legal regimes do permit some public goods to be provided as if priced in the market, there is no reason that the law should do any such thing. Legal regimes can also provide that such public goods be provided on an equal basis.[92] Legal regimes constitute consumers, but they can also constitute good local government Samaritans. To the extent that the localist model rests on a description of residents as consumers, it occludes this constitutive function of law and falsely presents a normative claim as a descriptive one.[93]

But if neither the localist nor the regionalist account is completely satisfactory, what can take their place? The response to this question has particular descriptive and general normative aspects. The particular descriptive aspect of the response states that the question of how to engage in effective metropolitan governance is a classic example of a “wicked problem”.[94] Its resolution requires the participation of affected parties and the coordination of various and diverse centres of governance. That is, the particular facts of metropolitan challenges, including those related to spatially defined inequities, seem to call for an approach that eschews an appeal either to civic republicanism, with its attendant assumption of deep cohesion, or to public-choice inflected interest group pluralism, with its presumptive valuing of self-interested bargaining. Instead, what seems to be called for is a pragmatic recognition of diversity among participants, of the need for co-operation among disparate jurisdictions, and of the value of the participation of those affected by any institutional configuration.

The more general theoretical response is that an adequate normative grounding for metropolitan governance should avoid the conundrums of the two positions set out above.[95] Professor Iris Marion Young articulated just such a theory of metropolitan governance. She argued for a conception of “regional federalism” that is to be governed by an ideal of “differentiated solidarity”.[96] This ideal seeks to reconcile two normative claims. First, the idea of solidarity claims that those who live “together in complex causal relationships in metropolitan regions” have obligations toward one another and, specifically, obligations “to constitute and support institutions of collective actions organized to bring about relations of justice among persons.”[97] Second, the idea of differentiation “affirms a freedom to cluster, both in urban space and in religious, cultural, and other affinity group associations.”[98] These groupings, argued Young, allow members to express their “affinity attraction” toward one another;[99] the idea of differentiation implies that groups should enjoy a measure of self-determination.

However, this idea does not permit members to disregard the interests of those individuals who are outside of the group, but who participate in the complex causal relationships that arise by virtue of living together in a metropolitan region. Toward these others, Young argued, group members are under an obligation of “openness to listening ... and engaging with them in shared public spaces.”[100] Young included among prohibited acts of exclusion segregationist policies that inflict a variety of wrongs upon those with whom one shares a metropolitan region. For present purposes, the most significant wrong is that of impeding political communication. Policies that separate populations by locality will, in general, prevent residents of disparate localities from entering into political discussions with one another and preclude the sharing of political decision-making power.[101]

These ideas of differentiation and solidarity, which together constitute the ideal of differentiated solidarity, seem to enter into direct conflict when one considers how to give them institutional form in metropolitan regions. According to Young:

On the one hand, self-determination, cultural specificity,participation, and accountability seem best realized in relatively smallpolitical units. On the other hand, values of taking into account the needsand interests of differently situated others with whom local affinity groupsdwell are best realized in political units wide in scope, comprising atleast broad metropolitan regions.[102]

Professors David Barron and Gerald Frug have argued for institutionalproposals that can respond to these apparently conflicting demands. The formerargues that the opposition between home rule and regional equality undulysimplifies a complex legal doctrine,[103] whilethe latter points to European institutions to argue that the contrast betweenlocal autonomy and regional institutions ignores the extent to which carefuldesign can ensure that regional institutions affirm local autonomy.[104]

Young drew on the work of Frug to propose a governance regime that is comprised of two kinds of institutions. First, the proposed governance regime would consist of local governments that are sufficiently small to facilitate interactions robust enough to give effect to the goal of differentiation. Residents can exercise through these local governments a measure of self-determination. Second, this regime would create a regional institution that would facilitate interactions among representatives of residents who live in different localities in a metropolitan region, but who nonetheless share “complex causal relationships” and therefore are under a duty to consider one another’s interests. Such a regional institution, argued Young, should facilitate negotiations among local governments, and should not subordinate these governments to the will of the regional government.[105]

The literature of new regionalism has similarly come to recognize that the polar positions within the localist-regionalist debate are untenable, and advocates for voluntary agreements among localities to reduce coordination problems.[106] In the Canadian context, Bish has argued for the proposition that a variety of scales of government institutions is necessary to respond to particular local populations’ service needs: an ex ante privileging of either localist or regionalist forms of governance is, according to Bish, undesirable as a functional matter. In what follows, I will set out the new regionalist arguments and detail how Bish’s arguments play out in the context of a specific institutional form in British Columbia.[107] We shall see that that province’s regional district system accomplishes some of the objectives that motivate new regionalism, while responding to some objections to institutions for which new regionalist authors have advocated. We will further see that the regional district system instantiates Young’s federalist ideal of differentiated solidarity.

B. New Regionalism: Foundations and Criticisms

New regionalism rests on three insights. First, new regionalist scholars build on Ostrom, Tiebout, and Warren’s claim that the provision and production of local government services can be separated—that is, local governments can provide the resources and regulations necessary to ensure that a set of services is produced without the local governments themselves doing the production.[108] Services can be provided by either local governments or private actors. Once a local government has decided that it should provide a service, the government then decides how to raise and distribute revenues, and how to monitor the production process. And after the local government has chosen to provide the resources and regulation necessary for a service, it has various choices about how the good or service should be produced: it can choose to undertake production itself by creating an in-house production unit; it can contract with a private or public actor for production; or it can regulate private activity to achieve public ends.[109]

Underlying and flowing from this picture of the provision and production of local goods and services are a normative and a descriptive claim. These are, respectively, the second and third insights that underlie the new regionalist model. New regionalists reject command and control measures on normative grounds. They reject regional governments and their bureaucracies, advocating instead for regional institutions that are created by agreements among the relevant parties. For the new regionalist, such institutions permit citizens to act as political agents. By contrast, argues the new regionalist, centralized regional governments alienate citizens from the political process.[110] This rejection of regional governments, however, does not mean that the new regionalist denies the importance of addressing concerns that extend beyond the boundaries of a municipality. Indeed, the new regionalist places a premium on mechanisms that regulate issues of varying geographic scales.

New regionalists note that because provision and production are not necessarily coterminous, governments can facilitate the production of services for needs that are either broader or narrower than an existing local government’s territorial boundaries. New regionalists derive from this description of the metropolitan landscape institutional recommendations that emphasize flexibility. For instance, the new regionalist praises the increasing presence of sublocal units of governments, such as residential community associations (RCAs) and business improvement districts (BIDs).[111] These sublocal units, argues the new regionalist, provide a close match between property owners’ preferences and payments. But critics have argued that these institutions are antidemocratic. BIDs subject non-property-owning residents to their regulations without giving them a role in determining the content of those regulations.[112] Furthermore, it seems that instead of promoting true democratic participation, RCAs foster in residents a perversely privatized vision of democracy.[113]

In addition to favouring these sublocal institutions, new regionalist writers view as a positive development the proliferation of special districts. Special districts can have three geographic scales, and can be constituted by a variety of means. The first type of special district is smaller than the local government boundary in which it is located. The second is coterminous with an existing local boundary. The third type of special district crosses existing boundaries.[114] Special districts can be created through a local initiative, through state enabling legislation, or through an agreement between general purpose local government units.[115] The new regionalist approves of each of these institutional forms and all these means of institution formation, for three reasons.

First, the New regionalist argues that because these local government institutions arise out of the choices of those who benefit from them, there is fiscal equivalence: the set of end payers matches that of end users.[116] Second, not only does this matching of service needs and services reduce waste, but it creates an efficient market of services. There is competition among a greater number of units that offer a more diverse set of services than is the case in situations where the market only or predominately consists of general purpose local governments.[117] Third, the new regionalist argues that special districts are created by those who benefit from them. They are therefore subject to close democratic oversight and inculcate democratic values in the relevant populace.[118]

Professor Kathryn Foster has subjected each of these claims about the advantages of special districts in the American context to vigorous criticism. Empirical studies, she notes, belie the new regionalist’s efficiency claims. The motivations for creating special districts are diverse, and not reducible to a desire for an efficient matching of preferences with services.[119] If special districts were actually designed to match service needs, their boundaries would be defined with reference to those service needs, but these boundaries are more often coterminous with municipal boundaries.[120] This suggests that their creation is not efficiency motivated and, moreover, that their design does not efficiently match the service to the relevant need.

Equally serious criticisms are directed against the democratic arguments for special districts. Critics note that it is rare for local populations to be intensely involved in the formation of special districts. More typically, special districts are created in response to the demands of local elites.[121] In addition, special districts and their activities are invisible. There is little public participation in either their formation or in the selection of their members, and therefore they do not inculcate in citizens democratic values.[122] Taken together, the efficiency and normative arguments against special districts yield a final criticism. Because special districts are typically outside the purview of normal democratic processes, they can impede efficient and democratically desired policy choices. Special districts are not subject to the normal give and take of democratic debate about programs. Special districts rather pursue their objectives outside the parameters of that debate and often have service monopolies. Because the fiscal capacities of the relevant polity are limited, general purpose local governments have to account for special district charges in the formation of their policies.[123] General purpose local governments, in which market preferences are expressed and prioritized through democratic debate, are therefore constrained by undemocratic and unresponsive special districts. Critics argue that new regionalist horizontal institutions frustrate Tiebout-style preference satisfaction.[124]

C. Regional Districts as New Regionalist and Federalist Institutions

The regional district system in British Columbia offers an institutional resolution of the localist-regionalist debate that draws on new regionalist insights, and offers a response to the various criticisms of new regionalism. Bish and Clemens have noted that regional districts perform three distinct functions. First, they are the general purpose governments for areas of the province that have no municipal corporations. Second, they provide a forum for co-operation among municipalities and unincorporated areas that facilitates agreements for the provision of services, with cost recovery, among municipalities. Third, they function as regional governments for region-wide voluntary and mandatory services.[125] It is the second and third functions of regional districts that are most directly relevant to this paper, and to explain these, it is perhaps helpful to lay out the history of the regional district system, and to delineate their structure and powers. Let me begin by briefly setting out the history of that system, which I hope will in part vindicate this essay’s decision to conceive of metropolitan governance in federalism terms.

Authors have argued that with the rapid population growth of the post–World War II era, inadequacies in British Columbia’s municipal organizations became evident. According to James E. Brown: “[i]n the metropolitan regions, the dividing line between adjacent municipalities was becoming completely obliterated.”[126] Informal means of regulating relationships among these municipalities were “too slow and too uncertain” to be relied upon, and it became clear that the proliferation of single purpose districts to address intermunicipal regulatory problems was “going to inhibit any proper co-ordination of activities at the metro level and preclude the establishment of priorities as among functions.”[127] In nonmetropolitan regions, with the exception of school board districts, there were no significant local government organizations.[128] In an article written shortly after the coming into force of the regional district system, Professor Paul Tenant and David Zirnhelt argue that the system was a response to these absences in the province’s local government institutional structure.[129] The authors argue that, beginning in 1964, provincial officials were the primary drivers behind the effective coming into being of this system.[130] What is of particular interest to this paper is that the accounts of Tenant and Zirnhelt, and Brown, which were all contemporaneous with the creation of the regional district system, describe it in federalism terms. Tennant and Zirnhelt write of the regional district that governed Greater Vancouver, “[i]n structure the new government follows the local federation model pioneered on this continent in Toronto and later introduced in Winnipeg.”[131] This language is echoed by Brown: “Essentially regional districts are a federation of a number of member areas.”[132] These commentators conceive of the regional district system as a kind of federation.

In what follows, I will (1) show how the system responds to some of the criticisms of new regionalist institutions; and (2) rebut some potential localist and regionalist critiques of the regional district system. In concluding this part, I will argue that the regional district system is justifiably characterized in federalism terms and that Young’s conception of regional federalism provides a particularly apt characterization.

1. The Regional Districts as New Regionalist Institutions

There are twenty-seven regional districts in British Columbia; all areas of the province, except for an area in the northwest, fall within the jurisdiction of these regional districts. Each regional district is a corporation whose governing body is its board, and the powers of the board are limited to the jurisdictional boundaries of the district, unless otherwise expressly provided for by legislation.[133] The members of the board are either directly elected from electoral (or unincorporated) areas, or they are municipal officials who are appointed to the board by their municipal councils.[134] Each political unit (electoral area or municipality) is represented by at least one member on the board, and the number of votes and members accorded to each political unit is proportional to that unit’s population.[135] There are different voting rules for different kinds of issues:[136] on issues that affect the entire region, each member of the board has one vote;[137] for matters that affect a sub-area of the region, only members from political units in the affected sub-areas vote and the voting is weighted; and for some financial matters, all members of the board can vote, and the voting is again weighted.[138] Moreover, a board may, through a referendum, seek the opinion of affected electors concerning a service that is or may be operated by the regional district.[139] Bylaws establishing most services in a regional district require the approval of the participating area, which can be obtained through a variety of means, including the assent of electors or, if the participating area is an entire municipality, the relevant municipal council.[140]

Regional districts are granted broad powers.[141] Indeed, the legislation provides that “[s]ubject to the specific limitations and conditions established by or under this or another Act, a regional district may operate any service that the board considers necessary or desirable for all or part of the regional district.”[142] In order to provide a service, a board must pass a bylaw that describes the service and the boundaries of the service area, identifies all the municipalities and electoral areas within the service area, and establishes the cost recovery method for the service.[143] The bylaw governing the service may itself set out the cost-apportionment formula; otherwise, the costs are shared according to the property-tax base of the participating areas.[144] Examples of regional services that are provided to the entire region include regional parks, regional planning, water supply, sewage treatment and disposal, and emergency 911 services. Recreation centres and parks are examples of services governed by agreement that are provided to some but not all political units.[145]

British Columbia’s regional district system meets some of the aims of new regionalism and answers some of new regionalism’s critics. The system’s mechanisms for intraregional agreements facilitate the matching of citizen preferences with government services, and therefore achieve the new regionalist aim of fiscal equivalence. Citizens of political units within the system in general only receive and pay for those services to which they, through their representatives, have consented.[146] In addition, through the mechanisms of political representation and intraregional negotiation and agreements, the regional district system avoids the risk of top-down regulation by a regional government that is distant from citizens. The system also provides for a degree of intraregional equity, since the general cost allocation formula makes reference to the relevant political units’ property-tax bases, and because the weighting of votes and representation on the boards is determined by population.

The system further evades the most damaging critiques of new regionalist institutions’ susceptibility to special-interest capture and lack of democratic accountability. Unlike special districts, regional districts are not established at the behest of special interests. They are, rather, pre-existing political entities whose board members are themselves elected officials who are therefore subject to democratic controls.[147] Finally, in addition to the direct accountability mechanism of the referendum, the regional district system provides for a range of consultative processes, regular reporting, public hearings and meetings, newsletters, televised board proceedings and advisory committees and commissions, and regional planning processes that require extensive interactions with municipal governments.[148] The regional district system therefore captures the main benefits of the new regionalist position while avoiding the costs of other new regionalist institutions, such as special districts, and offers a compelling institutional resolution of the localist-regionalist debate.

The regional district system does not, of course, resolve all disagreements. Conflict among political units is a real possibility that is foreseen in the legislative provisions governing dispute resolution.[149] In addition, a regionalist might argue that because of the consensual nature of the agreements, there is a real risk that inequalities in bargaining power among political units may yield unequal allocations of services within a given region. Indeed, for regional fiscal equity mechanisms, such as affordable housing, the municipalities in a regional district must overcome collective action problems. An example of such a co-operative effort is the Regional Housing Trust Fund, to which municipalities in the Capital Regional District contribute voluntarily. Most, but not all of the member municipalities (that is, twelve of the sixteen) in the regional district contribute to the fund.[150] It is perhaps important to recognize that the province retains its power to regulate, and can exercise this power with respect to politically sensitive matters such as the siting of affordable housing.[151] Moreover, the province can mandate that regional districts perform some region-wide functions.[152] In the end, although the regional system provides for some measure of redistribution and can address some inequalities, the provincial government acts as the ultimate redistributive backstop.[153]

2. Responses to Localist and Regionalist Concerns About the Regional DistrictSystem

This regionalist concern for the capacity of the regional district system to achieve redistributive ends gives rise to issues that are analogous to those raised by some contemporary legal scholars in their analyses of redistributive policies that are undertaken by municipalities, generally. Professor Clayton Gillette has focused the discussion both by offering a concise definition of redistribution in the local government context, and by providing an account for when local redistribution is malign. Gillette describes as “redistributive” any policy that “confers on a subgroup of residents benefits that are substantially disproportionate to the related local costs that the same subgroup bears.”[154] Under this definition, homeless shelters are the result of redistributive policies, because they concentrate costs that are not paid for by the beneficiaries of the shelters, while well-maintained roads, which spread benefits equally among the population who in general contribute equally to their maintenance, do not result from redistributive policies. In addition, Gillette defines as malign those redistributive measures that are the result of interest group pressures, and which do not reflect the views of the majority (where those views could be ascertained through a well-functioning political market).[155] For Gillette, non-malign or benign redistribution can and does occur. He argues, for instance, “that some localities expressly embrace a reputation for being more redistributive than others and that potential residents use that variable in selecting among jurisdictions.”[156]

Gillette considers various reasons advanced in the literature in favour of such redistribution, and three seem, on their face, to be plausible explanations for redistributive policies undertaken by regional districts. First, local governments (as opposed to higher levels of government) may redistribute because residents enjoy a sense of well-being that results from taking care of fellow residents, where the intensity of this sense dissipates across geographical space.[157] Second, citizens in localities that choose redistributive policies may have a strong sense of camaraderie with their fellow residents despite disparities in their respective economic status.[158] Third, localities may undertake redistribution to avoid social and economic costs associated with the poor. In this scenario, local redistribution functions as “a form of ‘bribe’ to the local poor to maintain a certain level of social peace that the relatively wealthy may otherwise believe will be threatened.”[159]

Although each of these explanations is subject to critique, Gillette argues, they can nonetheless offer plausible explanations for local redistribution, under two conditions: (1) if there is an institution through which decisions are made “in which all affected parties are effectively represented;”[160] and (2) if redistributive decisions made by this institution are visible to groups who monitor it. In the absence of these two conditions, the “benign” explanations for redistribution suggested above are less plausible, and special interest group capture arguments gain in plausibility.[161] Arguably, regional districts satisfy these two conditions. As we have seen above, regional districts create political forums in which representatives of affected citizens in a region make politically visible decisions. Moreover, the structure of the regional district system can reduce the incidence of the standard conditions that give rise to the risk of malign local redistribution. One such condition arises when politically powerless local minorities disproportionately bear the costs of local regulation.[162] The fact that intraregional agreements allow for geographic tailoring of redistributive measures minimizes the risk that costs will be imposed on a municipality that is politically powerless. This fact does not, of course, remove the risk of intramunicipal exploitation, but it does reduce the risk that costs will be concentrated in a municipality that is politically powerless and that, in the absence of the regional district system, would have no means of advancing its interests.[163]

If the regionalist is concerned about the capacity of the regional district system to be attentive to issues of distributive justice, the localist might criticize the system for failing to adequately respond to the preferences of citizens. The localist might be concerned that because the regional system relies on existing boundaries to demarcate service provision areas, it is insufficiently responsive to preferences that exist at a submunicipal level. Yet the provincial legislative regime provides for the creation of improvement districts that can provide services to these kinds of areas, when a majority of landowners in an area being considered for incorporation vote in favour of forming a district and the Lieutenant Governor in Council incorporates the improvement district.[164] Moreover, regional districts are empowered by governing legislation to enter into agreements for services[165] and therefore enjoy a measure of the flexibility for which new regionalists advocate, and which they believe yields efficiencies in the delivery of services.

In short, the regional district system is supplemented by regulatory regimes whose application is either larger or smaller than the jurisdiction of regional districts, and the governing legislation builds in a degree of flexibility for how services can be delivered. These mechanisms of provincial oversight, sublocal regulation, and flexible service delivery suggest a final point about the nature of legal institutional design in the urban context. Given the regulatory complexity of that context, it is unlikely that any single institutional design intervention will adequately respond to any given set of normative concerns, including regionalist and localist ones. The regional district system is one promising institutional form that offers a plausible response to the debate between regionalists and localists. Yet the effectiveness of the design intervention hinges on the larger institutional context within which it operates, and any prescription offered should be sensitive to the complexity of that context.

3. The Regional District System as a Federalist Institution

In this section, I have argued that the regional district system in British Columbia offers a plausible institutional response to the concerns underlying the localist-regionalist debate. Moreover, I have argued that the system shares some of the advantages claimed for institutions that authors have characterized as new regionalist, while avoiding some of the sharpest criticisms that have been directed against those institutions. Finally, I have defended the system against potential localist and regionalist objections. I conclude this section by attempting to vindicate authors who, at the time of the regional district system’s coming into being, described the system as federal in nature. In particular, I claim that the system conforms to Young’s conception of “regional federalism”.

Recall that Young argued for an ideal of “differentiated solidarity” that would govern relationships among residents of metropolitan regions. According to this ideal, residents of localities in metropolitan regions can, through involvement in their local governments, express their affinity for those who live in closest physical proximity to them. We have seen that the regional district system aims to provide a range of local institutions, including municipalities, through which residents of relatively small jurisdictions can engage in self-governance. Yet, the idea of differentiated solidarity does not permit these residents to disregard the interests of others, who live in localities elsewhere, in a shared metropolitan region. Instead, according to this idea, these other residents are owed an obligation to have their interests heard. The regional district system also responds to this demand of the idea of differentiated solidarity. The regional districts enable representatives of municipalities in metropolitan regions to enter into negotiations with one another and to address issues of common concern that arise by virtue of those municipalities existing in a network of complex causal relationships. Finally, we have seen that the province of British Columbia functions as a central government that can provide for redistribution within a particular regional district, if processes of negotiation ultimately fail to arrive at effective and just redistributive policies.

One might ask whether this possibility of provincial intervention potentially places municipalities and regional districts in a position of subordination vis-à-vis the province. Such a concern is significant for any attempt to apply Young’s theory of regional federalism because, as we saw above, this theory requires that a regional federation be governed by negotiations among political units and does not permit some political units to be subordinate to others. My response to this concern is a qualified one. The regional district system enables representatives of municipalities to decide amongst themselves what the relevant geographical contours of governance challenges in a regional district are, and to determine the appropriate responses to these challenges. The system, in general, is governed by Young’s norm of negotiation. Yet if the participants in a particular regional district system fail to respond to some legitimate redistributive claims, the province can act on these claims. I have argued elsewhere that provinces have a plausible claim to being able to overcome failures of deliberation in and among municipalities.[166] Municipalities and a regional district may be subordinate to a province when it acts in this supervisory capacity, but this condition of subordination arises because the province is in general in a position to consider interests that representatives of municipalities, deliberating amongst themselves, may neglect. This departure from Young’s regional federalist norm of nonsubordination is therefore justified for reasons that are standard in the federalism literature.


This paper has situated itself within a particular legal debate about cities and has offered a resolution of that debate that prescribes a specific institutional form. I have claimed that the approach taken in this paper is a distinctively legal one, and I have argued for the necessity of such an approach. In my view, because cities are necessarily creatures of the law, if one is to make claims about cities, one must be cognizant of the relevant legal context, and if one is to offer prescriptions, one must have facility with the relevant legal materials. These claims are, I believe, uncontroversial. Perhaps more controversial is this article’s approach to public law. In this paper, I have taken seriously a particular genre of arguments that authors in law have made about cities, namely, federalism arguments. Moreover, I have focused my arguments on the institutional claims advanced by participants in that debate and I have proposed specific institutional responses to them. A critic might contend that I have erred in at least one of two ways. A critic might first charge that I have accepted without good reasons the terms of the debate, and second, charge that my claim about the institutional nature of public law arguments is mistaken. I close this paper with a response to these criticisms and with some thoughts about directions for future research.

Consider first the critic’s claim about my acceptance of the federalism terms of the debate. As a preliminary matter, I should note that I do not take the terms of the localist-regionalist debate to be the only terms under which the discussion of cities can be undertaken. Indeed, as I noted in the introduction, there are a variety of ways in which one can characterize the legal nature of cities: one can conceive of cities in terms of property law or administrative law and other legal characterizations are possible. What these characterizations enable one to do, however, is to have a discussion that takes seriously the legal nature of cities. It may, of course, be the case that a particular instance of legal characterization is misguided, but in my view, the plausibility of a characterization is best tested in the process of making arguments about it. This paper has tested the plausibility of a federalism characterization of the city by engaging seriously arguments that have been advanced in federalism terms. Given the fact that in Canada, legal and political authority in major urban centres is divided along overlapping geographic boundaries, the federalism characterization that drives the localist-regionalist debate does not strike me as implausible, and therefore I see no reason to reject it.

Consider next a challenge to my claim that public law arguments necessarily implicate institutional concerns. One might articulate this challenge by claiming that when we address issues in public law, we should advance arguments that directly and exclusively address questions of political philosophy. One might, for instance, argue that the only relevant question is whether cities, as currently constituted, are just or unjust: once one has appealed to and defended a particular theory of justice and applied it to cities, one’s work is done.[167] My response to this challenge relies on a particular pragmatic understanding of legal reasoning.

Legal pragmatists argue that legal ends and means exist in a relationship of iterative interdependence. According to pragmatists, normative principles and social objectives are inevitably subjected to questioning when legal means are chosen to implement them.[168] In turn, the institutional means chosen will themselves be open to examination and recasting in light of shifts in understandings of principles and objectives that arise once specific institutional choices are made.[169] In this pragmatic conception of legal reasoning, theories are inevitably tested when they are implemented through institutional design choices, and do not stand apart from such choices. In this pragmatic conception, legal reasoning necessarily raises institutional questions because law, in all of its manifestations, is an institution that is persistently open to critiques in light of experience. I believe that this pragmatic view accurately describes law and legal institutions, and I do not think that there are convincing counterexamples to challenge it. This view of law yields one conclusion and suggests future lines of inquiry.

In the introduction, I claimed that arguments about cities require some knowledge of the legal contexts of cities, because cities are creations of law. I make a similar claim to conclude this paper: any theory of public law must address institutional questions. In this light, the ambition for this paper has been relatively modest. I hope to have advanced an argument with respect to a theory of cities that has satisfied this minimum requirement of public law theory. A future research program might be less modest. If cities are, as I have claimed, legal institutions that invite continuing reflection and contestation, the challenge for future institutional design projects is to propose and create forums in which these kinds of deliberative activities can be productively undertaken. Scholars from outside of law schools have examined institutions that facilitate these kinds of activities.[170] The challenge for legal scholars is to draw on these nonlegal scholars’ insights to devise institutional frameworks that will enable such deliberation, and that will be sensitive to the wider legal context of the city.