Printer Friendly

Redistricting and the territorial community.

As the current redistricting cycle unfolds, the courts are stuck in limbo. The Supreme Court has held unanimously that political gerrymandering can be unconstitutional--but it has also rejected every standard suggested to date for distinguishing lawful from unlawful district plans. This Article offers a way out of the impasse. It proposes that courts resolve gerrymandering disputes by examining how well districts correspond to organic geographic communities. Districts ought to be upheld when they coincide with such communities, but struck down when they unnecessarily disrupt them.

This approach, which I call the "territorial community test," has a robust theoretical pedigree. In fact, the proposition that communities develop geographically and require legislative representation has won wide acceptance for most of American history. The courts have also employed variants of the test (without scholars previously having noticed) in several related fields: reapportionment, racial gerrymandering, racial vote dilution, etc. The principle of district-community congruence thus animates much of the relevant case law already. The test is largely unscathed, furthermore, by the unmanageability critique that has doomed every other potential redistricting standard. The courts have shown for decades that they can compare district and community boundaries, and the social science literature confirms the feasibility of such comparisons. Finally, the political implications of the test's adoption would likely be positive. My empirical analysis suggests that partisan bias would decrease, relative to the status quo, while electoral responsiveness and voter participation would rise.

It is true that the territorial community test does not directly address partisan motives or outcomes. But the Court has made clear that it views these issues as doctrinal dead ends. Ironically, the only way left to combat gerrymandering might be to strike at something other than its heart.
     A. The Underlying Theory
     B. Other Approaches
        1. Compactness
        2. Competition
        3. Partisan Fairness
     A. Ascendance
     B. Decline and Fall
     C. Comeback
        1. Reapportionment
        2. Racial Vote Dilution
        3. Racial Gerrymandering
        4. Political Gerrymandering
        5. State Redistricting Law
     A. Defining the Territorial Community
     B. Doctrinal Details
     C. Relation to Other Domains
     A. Sidestepping the Problem
     B. The Supreme Court's Experience
     C. The State Courts' Experience
     D. The Political Science Literature
     E. Lingering Ambiguity
     A. Empirical Analysis
        1. Bias and Responsiveness
        2. Minority Representation
        3. Voter Engagement
     B. The Political Science Literature
        1. Bias and Responsiveness
        2. Minority Representation
        3. Voter Engagement


The decennial bloodsport of redistricting is now underway. Across America, state legislatures are busy drawing new electoral district lines based on the results of the 2010 Census. These new district lines, of course, will produce both winners and losers. Some political parties will gain seats while others will lose them. Some incumbents will have their districts fortified while others will be thrown to the wolves. Some minority groups will be able to elect the candidates of their choice while others will be engulfed by the surrounding majority.

In typical American fashion, many of the losers of the redistricting wars are seeking redress in court. (1) For better or worse, the doctrine that governs most of their claims--unequal district population, racial vote dilution, racial gerrymandering, retrogression, etc.--is relatively clear. But in one crucial area, that of political gerrymandering, (2) the case law is in chaos, at the levels of both theory and practice. The relevant scholarly literature is less confused but equally fragmented. There is thus an urgent doctrinal and academic need for new ideas, as well as for some coherence where now there is mostly upheaval. The "territorial community test" that this Article introduces is an attempt to meet that need.

A generation ago, in Davis v. Bandemer, the Supreme Court recognized a cause of action for political gerrymandering for the first time) However, the standard the plurality announced, focusing on the "consistent[] degrad[ation] ... of voters' influence on the political process as a whole," (4) proved hopelessly unworkable. Scholars puzzled over what values the standard sought to capture, (5) while lower courts struggled to apply it in actual cases. (6) A few years ago, in Vieth v. Jubelirer, the Court tackled gerrymandering again, with even worse results. A plurality would have reversed Bandemer and declared the whole field nonjusticiable. (7) Three dissents proposed separate (and conflicting) approaches for determining when gerrymanders cross the constitutional line. (8) As for Justice Kennedy, ever the Court's agonist, he was unpersuaded by the plurality, but also unpersuaded by any of the dissents, leaving him (and us) in a limbo where a standard for identifying unlawful gerrymanders might exist but has yet to be discovered. (9)

In the literature, as one might expect, potential approaches abound for resolving political gerrymandering disputes. One important camp, led by Samuel Issacharoff and Richard Pildes, argues that courts should emphasize electoral competition and intervene when districts are deliberately drawn to be uncompetitive. (10) Political scientists such as Andrew Gelman, Bernard Grofman, and Gary King have devised quantitative measures that show how fairly (or unfairly) a given district plan treats the two major parties. (11) Other scholars contend that the judicial inquiry should center on partisan intent, (12) district compactness, (13) the loss of democratic legitimacy, (14) or a series of factors derived from traditional districting criteria. (15) Still other scholars assert that gerrymandering is not particularly harmful and should not be dealt with by the courts at all. (16)

As voluminous as this literature is, much of it favors approaches that already have been spurned by the Court or that are in tension with the principles underlying the American electoral system. For instance, an outright majority of the Vieth Court rejected Justice Stevens's partisan-intent standard, Justice Souter's five-part test based on traditional districting criteria, and Justice Breyer's minority-entrenchment approach. (17) Nor, despite repeated invitations, has the Court ever embraced competition as the linchpin of its election law jurisprudence; indeed, even one of the dissenters in Vieth declined to adopt "[t]he analogy to antitrust." (18) A competition-centered approach also would be difficult to reconcile with the American commitment to geographic districting, which intrinsically produces many uncompetitive constituencies. Similarly, quantitative measures of partisan fairness have been appraised skeptically by the Court, (19) and have limited relevance for a districting regime that is organized around localized constituencies rather than statewide seat and vote tallies.

In this Article, I present an approach for curbing political gerrymandering that promises to avoid some of these pitfalls. The approach, in brief, is that electoral districts should be required to correspond to underlying territorial communities. To the extent possible, the boundaries of districts and organic geographic communities should be required to coincide--and the courts should be prepared to intervene when communities are unnecessarily fused, fragmented, or subverted, and the state can offer no reasonable explanation for the communal disruption.

A few points of clarification: First, by "territorial community," I mean (1) a geographically defined group of people who (2) share similar social, cultural, and economic interests and (3) believe they are part of the same coherent entity. Under this definition, territorial communities sometimes, but not always, mirror political subdivisions such as towns and counties. Territorial communities also are not quite the same thing as "communities of interest" (a common term in the redistricting case law), which are not necessarily geographically rooted and can form on the basis of any shared concern. Rather, territorial communities arise from the unique combinations of geography, interests, and identity that characterize particular places.

Second, district and community boundaries should coincide "to the extent possible" because the one-person, one-vote rule makes perfect congruence impossible. When communities must be disrupted, however, the disruption should be minimized--for instance, by joining groups that are as similar in their interests and affiliations as is practicable. Third, "fusion" and "fragmentation" refer, respectively, to the unnecessary merger of disparate communities and division of unified communities. By "subversion," I mean the drawing of districts that diverge sharply from the defining characteristics of the larger communities in which they are located.

Lastly, when I say that the judiciary should be ready to "intervene," I primarily have in mind constitutional law as the source of the courts' authority. But the power could, of course, be derived from other wellsprings, too: an act of Congress, state legislation, or, as I have discussed elsewhere, popular initiatives and referenda. (20) I also do not claim that adherence to community boundaries should be the sole criterion for determining whether a district plan is valid. My goal here is to call attention to an intriguing but underdeveloped doctrinal possibility--not to steal the thunder of other methods for combating gerrymandering.

Little in law is ever entirely new, but this territorial community test clearly runs against the grain of contemporary scholarship. In three recent articles, for example, Richard Briffault, Richard Hasen, and Richard Pildes list an array of potential standards for political gerrymandering cases, but do not even mention adherence to community boundaries as a possibility. (21) When scholars have addressed the territorial community, they typically have done so in cursory fashion (22) and with a hint of dismissiveness. (23) As John Hart Ely once wrote, "'[C]ommunity' is a concept so squishy that we should hesitate to entrust its specific application to either judges or politicians...." (24) Just about the only academics to delve into the subject at any length are Bernard Grofman (who suggests that districts should be "cognizable" to voters), (25) Henry Chambers (who advocates "enclave districting"), (26) and James Gardner (who has researched state law on communities of interest). (27)

The territorial community test thus bucks the academy's conventional wisdom. What does it have going for it beyond its contrariness? One significant (and previously unnoticed) advantage is that it has been employed by the Supreme Court, at least implicitly, in a range of recent redistricting decisions. In the racial gerrymandering context, for example, districts that correspond to geographic communities universally have been upheld by the Court, while districts that do not invariably have been struck down. The Court's reasoning is that districts coinciding with communities cannot have been crafted primarily on the basis of race.

Similarly, in the realm of racial vote dilution, the Court's governing standard asks whether a minority group is "geographically compact" and "politically cohesive"--in other words, whether the group is a territorial community. (28) In its most prominent recent dilution case, the Court both objected to the breakup of an old district that contained a "cohesive and politically active Latino community," and invalidated a new district that "combined two groups of Latinos, hundreds of miles apart, that represent different communities." (29) It is notable that the author of this decision was Justice Kennedy, the Hamlet of Vieth. That he placed such weight on district-community congruence indicates that he may be open to a similar approach in the political gerrymandering arena.

A renewed emphasis on the territorial community also would be consistent with the theory and historical practice of American districting. From colonial times until the reapportionment revolution of the 1960s, a core premise of American democracy was that communities arise on the basis of geography, possess distinctive political interests, and require representation in the legislature. Although this perspective has become less prevalent over the past two generations, it remains well-regarded among political theorists, and it dovetails nicely with the enduring American commitment to geographic districting. Consistent with this theory, the practice of most American jurisdictions, over most of the country's history, was to base representation at least partly on the territorial community. At the state legislative level, the most common constituency until the 1960s consisted of one or more towns or counties. At the federal level, the county was typically the building block for the congressional district.

The classic objection to the territorial community test is that it is too difficult to determine where community boundaries are and whether they coincide with district lines. (80) The absence of "judicially discernible and manageable standards," of course, was the precise reason the Vieth plurality gave for rejecting all of the dissenters' suggested approaches. (31) In my view, however, there are several reasons why the test would actually be workable. First, the Court has shown repeatedly, in the racial gerrymandering and racial vote dilution contexts, that it can identify community boundaries and evaluate district-community congruence. There is no reason why this task would be significantly harder in the political gerrymandering context.

Second, an array of state courts already have done very nearly what I propose: pursuant to state constitutional and statutory provisions, they have resolved political gerrymandering disputes based on the degree to which districts coincided with underlying communities. The rich body of case law produced by these decisions--which scholars have largely ignored--certainly does not seem arbitrary or unprincipled. Third, political scientists have developed several quantitative measures that could be used to define communities and to assess how well districts correspond to them. Some measures employ socioeconomic data or people's own geographic affiliations to identify community boundaries. Other techniques gauge the level of congruence between districts and political subdivisions or media markets, both of which are decent proxies for geographic communities. (32)

The second common criticism of the territorial community test is that it would have undesirable political consequences. Skeptics claim that it would harm Democrats and racial minorities (who supposedly live in particularly concentrated communities) and make elections less competitive. Concerns of this sort have limited bearing on the test's legal merit, and they turn out to be empirically unfounded as well. I analyzed how partisan bias, electoral responsiveness, and minority representation differed during the last redistricting cycle between states that paid heed to community boundaries when they redrew their district maps and states that did not. (33) I found that bias was markedly lower in the community-respecting states (5.4% versus 9.4%), responsiveness was markedly higher (1.43 versus 1.03), and minority representation was essentially unchanged. These findings are largely confirmed by an extensive political science literature--and suggest that the political implications of the territorial community test's adoption actually would be quite positive.

The Article proceeds as follows: Part I sets forth the theory of representation that justifies drawing electoral districts on the basis of territorial communities. Part II describes the tumultuous history of the territorial community as a factor in redistricting practice and case law. Part III considers how the territorial community test might translate into doctrine. Part IV addresses concerns about the test's judicial manageability. Finally, Part V discusses the likely political consequences of tighter district-community congruence.


Drawing district lines is no mere clerical task. To the contrary, the way in which districts are delineated has profound implications for the public's engagement with politics, for the character of political representation, for electoral competition, and for partisan fairness. One might even say that our democracy itself hinges, to some degree, on our system of districting. (34)

This means that no potential standard for drawing district lines (or for thwarting political gerrymandering) can be evaluated properly without taking into account its theoretical underpinnings. On what theory of representation is the standard based? How compelling is this theory? How consistent is it with the American democratic tradition? In this Part, I outline the theory--that communities arise along geographic lines and should be represented in the legislature--in which the territorial community test is rooted. I argue that this theory is plausible on its face and, more importantly, in harmony with the American commitment to geographic districting (which I take here as a given). I also contend that other prominent redistricting approaches either lack a clear theoretical foundation or proceed from theories that cannot easily be reconciled with core premises of American democracy.

Because this Article is more concerned with doctrinal matters than with issues of democratic theory, this Part is relatively succinct. Still, in the realm of redistricting, theory cannot be ignored--and, indeed, warrants some discussion at the very outset. (35)

A. The Underlying Theory

A distinct theory, which I call the theory of communal representation, underlies the territorial community test. (36) The theory makes two central claims. The first is that meaningful communities do in fact develop on the basis of geography. Subjectively, people feel connected to (and affiliated with) other people who live close to them: neighbors, members of the same school district, fellow residents in town, other center-city dwellers, folks on this side of the mountains, etc. Of course, powerful bonds can exist between people on opposite corners of the globe, but proximity at least fosters a sense of subjective solidarity. (37)

Objectively, people who live nearby tend to have common interests. They may have similar income levels, educational backgrounds, or housing situations. They may work for the same local employer or industry. They may have moved to the area for related reasons: the prevalence of certain cultural values, the availability of particular leisure pursuits, the presence of ethnic or racial compatriots, etc. Or they may simply care about the effective governance of the place they call home. Either way, the point is that people's spatial closeness correlates with, and helps generate, shared interests. (38) As Gardner puts it, "[P]eople who live in close physical proximity inevitably share certain kinds of activities, and ... the bonds created through these shared activities give rise to a community ... that is deeply connected to, built upon the matrix of, even induced by, the particular locality in which the members live." (39)

The theory's second claim is that territorial communities should be represented as such in the legislature. They are important enough entities, capturing interests and values with sufficient political salience, that they should serve as the basic building blocks of legislative representation. Geographic affinity--not party, race, profession, or any other factor--should be the criterion on the basis of which people are placed into electoral districts. As Justice Stewart once wrote, elected officials represent "people with identifiable needs and interests ... which can often be related to the geographical areas in which these people live." (40) The significance of these "geographical areas"--territorial communities, in this Article's parlance--"carries with it an acceptance of the idea of legislative representation of regional needs and interests." (41)

The theory of communal representation also has implications for how districts should be constituted and how politicians should represent their constituents. First, if districts are drawn to coincide with geographic communities, then they should be easily understandable by voters--"cognizable," to use Grofman's term. (42) Voters should be less confused and more politically engaged when district boundaries also demarcate groups of people with common interests and affiliations. (43) Second, districts should be relatively homogeneous in their social, cultural, and economic complexions. When people who share key attributes and who feel a kinship with one another are placed into the same district, it follows that the district too should possess a comparatively uniform texture. (44) Not a perfectly uniform texture, since there is often substantial diversity even within a single community, (45) but certainly a more consistent profile than a district that fuses together unrelated groups of people.

Third, because of this homogeneity, it should be relatively straightforward for elected officials to identify and advance their districts' interests. Representatives should be able fairly easily to become "acquainted with the interests and circumstances of [their] constituents"--as Madison put it in The Federalist--when those traits are broadly similar. (46) In contrast, heterogeneous districts should pose a greater representational challenge since they make it trickier both to discern districts' needs and to satisfy them effectively. (47) Fourth, elected officials should tend toward the delegate side of the delegate-trustee spectrum of representation. (48) Given coherent constituencies with distinct interests, it should be simpler (and more electorally beneficial) for officials to promote those interests than to exercise their own independent discretion. (49)

The theory's final implication is that, while districts should be relatively homogeneous, the legislature should be relatively heterogeneous. (50) With each community in the state (or nation) represented, the legislature should reflect the full diversity of views held by the general population--not just the preferences of the median voter. As John Adams famously wrote, the legislature "should be an exact portrait, in miniature, of the people at large." (51) On this view, legislative politics should be inherently pluralistic, as varied communities jockey for position and negotiate over the formulation of public policy. Legislative politics also should be less dominated by partisan cleavages, since party affiliation presumably has lower salience in a system in which politicians identify strongly with entities (i.e., territorial communities) other than political parties.

On balance, I find the theory's two key claims--that territorial communities exist and should be represented in the legislature--at least plausible. While some politically salient interests do not correlate with geography, many others plainly do, and it is sensible, in my view, to capture these territorial concerns and assure them a legislative hearing. I also think the theory's implications for districting and representation are relatively attractive. It does seem desirable for voters to be able to make sense of their districts' boundaries, for representatives to be able to identify and promote their constituents' interests, and for legislatures to reflect accurately the diverse views of the public. In an era riven by partisanship, it is also hard to quarrel with an approach that might weaken the grip of political parties.

Much more could be said, of course, about the theory of communal representation. In this Article, however, I am less interested in fleshing out the theory than in pointing out that it is consistent with--and arguably compelled by--the American commitment to geographic districting. This commitment, as many scholars have noted (often in frustration), is ironclad (52) and dates all the way back to the Framing. (53) With very few exceptions, American jurisdictions have always elected representatives from geographically defined constituencies. At times, jurisdictions have used at-large elections, multimember districts, and even voting rules other than first-past-the-post, but almost always in the context of districts that were spatially demarcated. There are also no signs that this practice is likely to change anytime soon. As Peter Schuck has observed, geographic districting "remains a firmly embedded feature of American political life, one that reformers' criticisms have utterly failed to dislodge." (54)

Given the premise of geographic districting (which this Article does not challenge (55)), the theory that territorial communities should be represented in the legislature follows fairly naturally. If district lines are to be drawn on the basis of geography, those lines must correspond to something, and there are few candidates for what that something should be other than people's territorially linked interests and affiliations. Certainly, it would not be sensible to draw spatial boundaries at random, or in order to create pretty shapes on a map, or to unite people who have no geographic connection to one another. Districting based on geography (rather than party, race, profession, or any of the other myriad possibilities) entails districting based on the value that geography is meant to capture. And that value is adherence to the territorial community--and it is hard to see how it could be anything else. In the words of one political theorist, "[T]he claim for geographic [districting] ... usu-ally rests ... on some concept of community, on the relationships among people living in relative proximity to one another." (56)

There is a lively academic debate as to whether territorial communities are as meaningful today as they once were. Some scholars assert that advances in transportation and telecommunications, combined with the public's greater mobility and social fluidity, have undermined all geographic groupings. (57) In contrast, other observers contend that people continue to have a "deep concern for [their] locality or region" and a "fundamental desire ... for a sense of place, [for] belonging to a community." (58) On this account, recent developments have not undercut territorial communities, but rather increased their appeal for the atomized inhabitants of the modern world.

This Article is not the place for me to venture into this controversy. My claim, rather, is that whether or not the importance of territorial communities has waned, they ought to remain the focus of districting as long as it is carried out geographically. If territorial communities indeed no longer matter to people, then perhaps it is time for America to abandon geographic districting. But as long as districts continue to be drawn along spatial lines, I see no theoretically justifiable alternative to drawing them to correspond to territorial communities (weakened though they may be). Territorial communities remain the only game in town.

B. Other Approaches

There are, of course, many possible ways to combat political gerrymandering other than by requiring district-community congruence. To cite some of the most prominent options, traditional districting criteria such as compactness could be enforced more stringently, highly uncompetitive districts could be invalidated, or statewide measures of partisan fairness could be applied. (59) There is much to like about these approaches, and any of them would represent a marked improvement over the status quo. Here, though, I am interested in evaluating the theoretical underpinnings of redistricting standards. And on this criterion, none of the above options is entirely satisfying. The first lacks a clear theoretical foundation, while the others are based on theories that are not easy to square with basic assumptions of the American electoral system.

1. Compactness

Beginning with compactness, there is no reason to think that people are represented adequately only when they are placed in districts whose shapes are aesthetically pleasing. No theory holds that people's engagement with the political process, the relationships between constituents and their elected officials, or politicians' performance in office, is optimal in a regime where districts are compact. Neither the jaggedness of districts' boundaries, nor the dispersion of their territories, (60) has any inherent connection to the caliber of districts' political life. As Robert Dixon has noted, "Shape requirements focus on form rather than the substance of effective political representation." (61)

Claims are occasionally advanced that compact districts facilitate communication between constituents and representatives, improve voters' knowledge of elected officials, or convey political legitimacy. (62) But it should be obvious that compactness itself does not necessarily produce these benefits. A district may be a perfect circle, but have a mountain range running down the middle of it, or combine two towns with nothing in common. (63) In these (and many other) cases, attractive district shape is not correlated with good communication, high voter knowledge, or enhanced legitimacy. Moreover, when compactness is correlated with these attributes, it is often because the compact district corresponds to a compact territorial community. Within a community, of course, it is relatively easy for voters and elected officials to communicate and for constituents to inform themselves about their representatives. Districts that reflect underlying communities are also usually perceived as politically legitimate. Accordingly, there is no theoretical basis for focusing on compact districts--whose principal appeal is that they are weak proxies for territorial communities--instead of on the communities themselves. (64)

2. Competition

As compactness has lost some of its luster in recent years, scholars have turned their attention to the argument (most associated with Issacharoff and Pildes) that courts should prioritize structural values in election law cases, electoral competition chief among them. (65) In the redistricting context, this argument means that highly uncompetitive districts should be invalidated (or at least regarded with severe skepticism), particularly when the lack of competition is dehberate. (66) As its advocates make clear, a particular theory of democracy underlies this approach. "[D]emocratic politics [is] akin in important respects to a robustly competitive market.... Only through an appropriately competitive partisan environment can one of the central goals of democratic politics be realized: that the policy outcomes of the political process be responsive to the interests and views of citizens." (67) In other words, competition is the essence of democracy, because it is only through competition that the government becomes responsive and accountable to voters.

This is a compelling conception, and it has significant (and largely positive) implications for how courts should tackle an array of election law controversies. (68) In the redistricting realm, though, the theory runs into some difficulty. The problem that is particularly relevant here (and that has received little attention in the literature) is the tension between the theory and the American commitment to geographic districting. Different geographic areas, of course, have different political profiles. In some areas (e.g., blue-collar ethnic communities, certain suburbs), the two major parties enjoy similar levels of support. But in other areas (e.g., rural farmland, urban centers), one party is far more popular than the other. In this latter category of places, it is extremely difficult to design competitive districts. Any district that accurately reflects the area's politics inevitably will be safe for one party or another. The only way to create a swing district is to fuse part of one area (a center city, say) with part of another (for instance, a faraway farming community). (69)

The conflict between the competition-centered approach and geographic districting thus arises because many geographic areas simply are not politically competitive. In these areas, either uncompetitive districts must be drawn--despite the harms that allegedly ensue for democratic responsiveness and accountability--or competitive districts must be created by connecting groups of people with little in common. Surprisingly, advocates of the competition-centered approach do not appear to favor the latter option. Samuel Issacharoff and Pamela Karlan write that "It]here will always be Berkeley and Orange County," where "[i]t would take a radical gerrymander ... to bring them to a contested balance between the major parties." (70) Richard Pildes similarly distinguishes between "safe districts that arise naturally" and "safe districts that arise because political insiders have grossly manipulated district designs." (71)

But if competitiveness can be sacrificed wherever it does not develop organically, then the competition-centered approach is not all that different from this Article's call for district-community congruence. In that case, both approaches recommend competitive districts where the broader community is competitive, and uncompetitive districts where the broader community is uncompetitive. Likewise, both approaches condemn uncompetitive districts that are drawn within a competitive community--on one account because responsiveness and accountability are needlessly undermined, and on the other because a coherent community is needlessly disrupted.

This may seem like an elegant theoretical convergence, but it is actually more akin to a surrender. The logical implication of the competition-centered approach is that competitive districts indeed should be drawn, where necessary, by merging disparate communities. By resisting this implication, the approach's proponents concede that the principle of adherence to territorial communities takes precedence over their desire for competitive districts. They concede (to quote Issacharoff and Karlan) that "there is ... normative force to the idea that districts should reflect some reality on the ground." (72)

If advocates of the competition-centered approach did not make this concession--that is, if they insisted on competitive districts under all circumstances--then their goals would be met most readily by nonterritorial constituencies. (73) Districts defined by their balanced partisan composition (rather than by their spatial boundaries) are the obvious way to maximize electoral competitiveness. But even if such a districting regime were attractive in the abstract, (74) it would bear little resemblance to our actual system. We would have to renounce our commitment to territoriality in order to optimize political competition, because geographic districts cannot fully accommodate non. geographic values. (75)

3. Partisan Fairness

The final approach that is often recommended for curbing political gerrymandering is some quantitative measure of partisan fairness. For instance, the partisan symmetry test (the most sophisticated of these measures) focuses on whether each major party would win the same share of seats given a particular share of the statewide vote. The disjunction, if any, between the seat shares--e.g., if Democrats would win sixty percent of the seats if they received fifty percent of the votes, but Republicans would win sixty-five percent of the seats with that vote share--constitutes the partisan asymmetry. (76) The theory underlying this approach is clear: "[T]he electoral system [should] treat similarly-situated parties equally," (77) so that "candidates of each political party ... have equal opportunity in translating voter support into the division of legislative seats between the parties." (78)

This theory too is appealing in principle but difficult to square with the practice of geographic districting. Given the different political profiles of different areas, there is little reason to think that districts drawn on the basis of geographic considerations will typically yield the same results for both major parties. For example, one party's support in a state might be heavily concentrated while the other party's backers are more evenly (and effectively) dispersed. In this case, the parties would not win the same share of seats given a particular share of the statewide vote--at least not as long as districts are drawn on the basis of the usual redistricting criteria. (79) It might be possible to devise districts that ensure partisan symmetry in such a state, but these districts would likely form strange shapes and pay little heed to subdivision or community boundaries. (80) If constituencies are to be defined not just spatially, but also in accordance with underlying geographic realities, then partisan symmetry cannot be guaranteed. (81) Indeed, the best way to ensure symmetry is to adopt a non-territorial electoral system--proportional representation, for instance, which intrinsically treats all major parties identically.

The inherent symmetry of non-territorial regimes points to the core problem with applying metrics of partisan fairness to American elections: those metrics look only to statewide shares of seats and votes, while the character of geographic districting is irreducibly local. As discussed above, the very reason for drawing geographic districts is to capture something unique about each particular place--to enable each distinct locality to have its voice heard (and its interests advanced) in the legislature. (82) Under this conception of representation, concerns about statewide seats and votes are largely irrelevant. What matters, instead, is that each district make sense (because it corresponds to a territorial community), not that the consolidated votes in all the districts across the state bear some relationship to the seats controlled statewide by each party. As Justice O'Connor once noted, "[V]oters cast votes for candidates in their districts, not for a statewide slate of legislative candidates.... Consequently, efforts to determine party voting strength presuppose a norm that does not exist-statewide elections for representatives along party lines." (83)

Accordingly, the principal alternatives to the territorial community test are theoretically problematic. While preferable to the status quo and appealing in many respects, they either lack an appropriate theoretical foundation or conflict with the enduring American commitment to geographic districting.


It is not only theory that demands consideration in the redistricting context. History, too, has important implications for which approaches to political gerrymandering can and should be adopted. That a particular approach has long been used by many American jurisdictions suggests that it fits well with our electoral institutions and values. (84) Conversely, there is greater uncertainty associated with standards that rarely or never have been employed (such as competitiveness and partisan fairness). The ecosystem of election law is intricate, and it is difficult to predict what consequences new seedlings might have.

Accordingly, this Part examines the rich history of the territorial community in American redistricting practice and doctrine. I first discuss the period prior to the reapportionment revolution, during which districts generally corresponded to communities and gerrymandering was understood as the absence of such congruence. I next explain how the momentous one-person, one-vote decisions of the 1960s dethroned the territorial community and replaced it with a singleminded focus on population equality. Lastly, in what is this Part's most novel contribution to the literature, I argue that respect for territorial communities has returned to the fore in recent years. State courts often focus on adherence to community boundaries when they evaluate district plans, and the Supreme Court's intuition that districts and communities should coincide now animates--and gives coherence to--a good deal of its redistricting case law. Rumors of the territorial community's demise, like Mark Twain's, seem to have been greatly exaggerated.

A. Ascendance

Almost since the inception of Anglo-American democracy, electoral districts and territorial communities usually have coincided. More than seven hundred years ago, when England's House of Commons came into being, it was geographic communities that received representation. Counties, shires, and boroughs all sent delegates to inform the monarch of their views and to promote their particular interests. Because it was distinct localities that were represented, rather than the public at large, it was "more accurate to call the representative portion of the Parliament a House of Communes than a House of Commons." (85)

Colonial America adopted a similar model. In all thirteen colonies, local communities (such as towns, counties, plantations, and parishes) directly elected representatives to the various assemblies. Though more populous communities were sometimes assigned more elected officials, representation was always provided to some communal unit. (86) After independence, states continued to organize their politics around the territorial community. Their original constitutions all designated towns (in New England) or counties (everywhere else) as the entities from which representatives were to be elected to the state legislature. (87) "[T]he basic unit of legislative representation was widely understood throughout the eighteenth and nineteenth centuries to be the [town or] county." (88)

As time passed and the country grew, a number of states switched from town or county representation to electoral districts drawn roughly on the basis of population. According to Gardner's survey of state constitutions, this trend took hold at the state senate level around 1800, and at the state house level around 1850. (89) Rosemarie Zagarri reports that it was mostly the larger and newer states that switched to districted elections in the 1800s. (90) Crucially, these state legislative districts, unlike their modern analogues, almost never disrupted town or county boundaries. To the contrary, state constitutions frequently prohibited districts from dividing political subdivisions, and occasionally defined districts outright as combinations of smaller subunits. (91) At the federal level too, congressional districts in the 1800s typically were composed of whole towns and counties and rarely crossed their boundaries. (92) In this era, the "principle that local government units should be kept intact for purposes of representation" was "almost universally accepted" for both state and federal elections. (93)

This regime endured mostly unchanged through the first half of the twentieth century. In 1955, on the eve of the reapportionment revolution, nine states still elected at least one chamber purely by town or by county. (94) Only twelve states required districts to be drawn solely on the basis of population. (95) The "most common practice" was to allocate representatives to state legislative districts, composed of one or more whole counties, in "rough proportion to their respective populations." (96) Most congressional districts also continued to respect the boundaries of political subdivisions (though the number of infractions inched higher). (97) Just before the Supreme Court entered the fray, then, representation in America largely remained as it had always been: founded first on districts that kept territorial communities intact, and only afterward on the concept of equal district population.

Not surprisingly, the reason why American representation took this form prior to the 1960s was that Americans tended to accept the theory of communal representation (discussed above in Part I). By and large, that is, American jurisdictions agreed that territorial communities existed and ought to be represented as such in the legislature. Gordon Baker notes that electoral arrangements in this period "reflect[ed] ... the force of localism, the view that every community should have ... a distinct and substantial[] voice in the state legislature." (98) Similarly, state courts often explained their commitment to districts that respected community boundaries by waxing eloquent about the "common interests and objects" of each county's residents, (99) "the community of interests in the respective counties," (100) and the "right" of counties "to be represented by their own members of the legislature." (101) It was thus no coincidence that districts usually coincided with territorial communities in this era; rather, this was the natural result of the theory of representation to which most Americans adhered.

Under this theory, gerrymandering was conceived as not only the undue pursuit of political advantage, but also the disruption of organic geographic communities. The original Massachusetts gerrymander of 1812 provoked such ire (according to a scholar writing in 1907) because "[t]owns were separated and single towns were isolated from their proper counties." (102) Another early twentieth-century observer defined gerrymandering as "a violation of the geographic unity of regions," with "[t]he amount of divergence of electoral boundaries from geographic boundaries" serving as "a measure of their fairness." (103) Consistent with this definition, many states prohibited districts from dividing territorial communities. The first such ban was adopted by Pennsylvania in 1790, and the idea eventually became one of the most common state constitutional techniques for combating gerrymandering. (104) As the Indiana Supreme Court explained in 1892, by "prohibiting the division of counties in the formation of ... districts," the state constitution's drafters "intended to put it beyond the power of the general assembly" to engage in gerrymandering. (105)

B. Decline and Fall

This coherent conception of representation--in which districts corresponded to territorial communities and gerrymandering was understood as the absence of such congruence--came under fierce attack during the reapportionment revolution of the 1960s. In a series of landmark decisions, the Supreme Court held that reapportionment disputes are justiciable, (106) and that the one-person, one-vote rule (which requires districts to have the same population) applies to federal, (107) state, (108) and local (109) elections. Crucially, the Court refused to relax the one-person, one-vote rule so that district and community boundaries could continue to coincide. Instead, the Court insisted on near-perfect district equipopulation even at the cost of widespread community disruption.

In the foundational 1964 case of Reynolds v. Sims, for instance, the Court rejected the argument that some discrepancies in district population could be justified by ajurisdiction's policy of keeping communities intact. The Court asserted that political subdivisions are merely "subordinate governmental instrumentalities" and that "[l]egislators are elected by voters, not farms or cities or economic interests." (110) Community-oriented "[c]onsiderations of area" were therefore "an insufficient justification for deviations from the equal-population principle." (111) Similarly, in the 1969 case of Kirkpatrick v. Preisler, the Court invalidated a Missouri district plan whose population variances (of less than four percent) were motivated by "regard for such factors as the representation of distinct interest groups [and] the integrity of county lines." (112) The Court declared that "to accept population variances, large or small, in order to create districts with specific interest orientations is antithetical to ... the constitutional command." (113)

Perhaps the most striking example of the Court's preoccupation with the one-person, one-vote rule, no matter what the communal cost, came in Lucas v. Forty-Fourth General Assembly. (114) In this 1964 case, Colorado's voters decisively rejected an initiative that would have reapportioned both houses of the state legislature solely on the basis of population. (115) Instead, majorities in every county endorsed a rival initiative that would have reapportioned the state house based only on population, and the state senate based on both population and "a variety of geographical, historical, topographic and economic considerations." (116) The Court nevertheless struck down the voters' preferred plan. (117) It was irrelevant to the Court that the plan's population deviations were quite small, that they were the product of "geography, ... accessibility, observance of natural boundaries, and conformity to historical divisions"--and even that they had been approved by the very voters whose rights the Court held had been violated. (118)

Not surprisingly, the dissenters in this period recognized the transformation wrought by the Court. In Baker v. Carr, the 1962 decision that began the reapportionment revolution, Justice Frankfurter decried the Court's "massive repudiation of the experience of our whole past." (119) At great length, he reviewed the history of American districting, which previously had revolved around the territorial community, and argued that it was this history that the Court was abandoning. (120) In Reynolds, similarly, Justice Harlan condemned the rupture with the past represented by the one-person, one-vote decisions. He noted that the Court had dismissed in turn every non-population factor once thought relevant to districting: "(1) history; (2) economic or other sorts of group interests; (3) area; [and] (4) geographical considerations...." (121) In Lucas too, Justice Stewart contended that the Court's "draconian" adherence to equipopulation conflicted with the "strongly felt American tradition" that "many diverse interests" should be "ex

pressed by a medley of component voices" in the legaslature. (122)

As all of the Justices expected (some eagerly, others apprehensively), the reapportionment revolution swiftly stripped the territorial community of its centrality in American districting. Once federal, state, and local districts were all required to possess the same population, the number of communities that had to be divided or merged increased radically. In almost all areas, communities were not equipopulous, meaning that the only way to comply with the one-person, one-vote rule was to countenance their wholesale disruption. As Micah Altman found in a detailed study, "[T]he frequency of violations of 'traditional boundaries' skyrocket[ed] following the population requirements imposed by the Court in Reynolds...." (123) Counties, cities, and wards were broken up at triple their earlier rate, and the splitting of rural communities became widespread as well. (124)

The reapportionment revolution also undermined the territorial community in a second, subtler way. Under the Court's one-person, one-vote doctrine, districts must be redrawn each decade after the latest Census figures are released. (125) As a result, there is now less opportunity for communities to coalesce on the basis of district boundaries.

Every ten years, those borders are unpredictably shaken and stirred, destabilizing any nascent communities that have begun to develop around them. As one scholar has explained, "A boundary that is continually moving is one that is unlikely to serve ... as a dividing line between genuinely distinct political communities. In this way, one person, one vote continually impedes the formation ... of meaningful local political identity." (126)

C. Comeback

According to the conventional narrative, this is where the tale of the territorial community ends. Once at the heart of the American system of representation, it now has been eclipsed for good by the one-person, one-vote rule. (127) In this Section, I argue that this narrative is simplistic and in many ways incorrect. While the territorial community may play a relatively small role in contemporary redistricting practice, it remains very much alive in both federal and state doctrine. In fact, in areas as diverse as (1) reapportionment, (2) racial vote dilution, (3) racial gerrymandering, (4) political gerrymandering, and (5) state redistricting law, the courts' intuition that districts and communities should coincide still commonly drives (and gives coherence to) judicial outcomes. This Section discusses the relevant case law and presents the surprisingly strong case for the territorial community's continued doctrinal relevance. The Section errs on the side of thoroughness because this story is an important one that has not previously been told.

1. Reapportionment

As noted above, by the end of the 1960s, the Supreme Court refused to permit district population variances even if they were quite small and resulted from a good faith effort to respect the integrity of territorial communities. (128) Just a couple years later, the Court commenced a fairly dramatic retreat from this absolutist position. Over a series of decisions in the 1970s and 1980s, the Court accepted moderate--and, in one case, extreme--population deviations that were the product of state policies promoting the representation of political subdivisions. In the very field in which district-community congruence decisively had been rejected, it made a swift and improbable comeback.

The comeback began with the 1971 case of Abate v. Mundt, in which the Court upheld districts for a county legislature that coincided perfectly with the county's five towns. (129) Though these districts were just as divergent in population as those struck down in earlier decisions, the Court was suddenly flexible where before it had been unyielding. "[T]he particular circumstances and needs of a local community ... may sometimes justify departures from strict equality," the Court remarked, adding that "exact [town-district] correspondence" was desirable because it "encourage[d] town supervisors to serve on the county board." (130) These were the same arguments about distinct local interests and the quality of representation that the Court previously had dismissed out of hand. (131)

In a 1973 case, the Court likewise endorsed a Virginia district plan that carefully avoided crossing city or county boundaries. (132) Though the plan's population deviations again were substantial, the Court deferred to the state's policy of "avoid[ing] the fragmentation of such subdivisions" and "afford[ing] them a voice in Richmond." (133) The Court further argued, in a mode entirely absent from its 1960s decisions, that adverse consequences would follow if counties were divided among multiple districts: "The opportunity of [their] voters to champion local legislation [would be] virtually nil," and their "representation [would be] no representation at all so far as local legislation is concerned." (134)

More remarkable still was the Court's 1983 ratification of a Wyoming district, corresponding to a small county, whose population diverged by sixty percent from the ideal. (135) The Court approvingly cited the state legislature's position that "the needs of each county are unique and the interests of each county must be guaranteed a voice." (136) The Court also relied on the trial court's findings that "[t]he people within each county have many interests in common" and that "to deny these people their own representative borders on abridging their right to be represented." (137) Reasoning of this sort, of course, would have been unthinkable during the Warren Court's one-person, one-vote heyday. In that era, nothing was allowed to interfere with the Court's quest for perfect population equality; indeed, claims that communities were unique and required their own representation routinely were rejected. (138) Fifteen years later, however, the Court was willing to tolerate extreme population inequality as long as it was the result of a "legitimate policy of preserving county boundaries." (139) District-community congruence had returned with a vengeance.

2. Racial Vote Dilution

District-community congruence has only ever played a starring role in the realm of racial vote dilution. In this field, under both the Constitution and the Voting Rights Act (VRA), (140) minority plaintiffs may bring claims that their votes have been unlawfully diluted. They may argue, in other words, that they have been denied sufficient political influence by district lines that "pack" or "crack" minority groups or by multimember districts in which minority voices are drowned out. Crucially, whether a dilution claim is constitutional or statutory, it requires in effect that the minority group constitute a coherent territorial community. Under the Equal Protection Clause, both dilution cases in which the Supreme Court has granted relief have involved distinct minority communities. Similarly, under section 2 of the VRA, the Court's operative test includes explicit prongs for geographic compactness and political cohesiveness.

The Court first ruled in favor of plaintiffs advancing a constitutional claim of racial vote dilution in the 1973 case of White v. Regester. (141) Mexican Americans in San Antonio argued that they were effectively excluded from representation in the state legislature by a countywide multimember district. The Court agreed, and replaced the multimember district with multiple single-member districts, in large part because it viewed the city's Mexican Americans as a discrete and underprivileged geographic community. (142) The Court observed that almost all of the Mexican Americans lived close together in the Barrio, "an area of poor housing" whose "residents have low income and a high rate of unemployment" and "suffer[] a cultural and language barrier." (143) It was precisely in order "to bring the community into the full stream of political life" that the Court authorized the creation of single-member districts in which the Mexican Americans could elect their preferred candidates. (144)

The logic of the Court's other decision granting relief to plaintiffs for constitutional vote dilution was nearly identical. In a 1982 case, the Court found that African Americans in rural Burke County, Georgia, were a "cohesive political group" with a "depressed socioeconomic status," "less formal education," and "less pay." (145) Accordingly, the Court dismantled the at-large election system that had prevented a single African American from being elected to the county commission, again replacing it with multiple single-member districts. (146)

The Court has not decided a constitutional vote dilution case since 1982 because, in that same year, Congress amended section 2 of the VRA to make it easier to bring statutory dilution claims. (147) In its 1986 decision interpreting the revised statute, Thornburg v. Gingles, the Court set forth a test composed of three prongs, two of which focus directly on whether a minority group comprises a territorial community. (148) "First, the minority group must be able to demonstrate that it is sufficiently large and geographically compact to constitute a majority in a single-member district," and "[s]econd, the minority group must be able to show that it is politically cohesive." (149) A minority group that cannot establish that it is geographically compact and politically cohesive--in other words, that it is a territorial community--cannot make out a dilution claim.

Pursuant to this test, the Court has repeatedly ruled against minority groups that failed to prove compactness or cohesiveness. In a 1993 case, for example, the Court concluded that a motley set of minority plaintiffs from Minneapolis (including African Americans, Hispanics, Asian Americans, and Native Americans) was not politically cohesive. (150) An "agglomeration of distinct minority groups," lacking shared interests or a common identity, could not amount to a community or prevail in a vote dilution challenge. (151) Similarly, in a pair of 1996 cases, the Court held that highly non-compact districts combining disparate minority groups could not remedy alleged section 2 violations. (152) Districts like the one in Texas that "reache[d] out to grab small and apparently isolated minority communities" could never be required by the VRA. (153)

Conversely, in the important 2006 case League of United Latin American Citizens (LULAC) v. Perry, the Court both found vote dilution when a coherent Hispanic community in Texas was fragmented, and rejected a proposed remedy that involved merging dissimilar Hispanic communities. (154) The vote dilution took place when a "cohesive Latino community" that previously had been placed in a single district, and that had developed an "efficacious political identity," was dispersed among multiple districts in order to protect an imperiled incumbent politician. (155) "The State not only made fruitless the Latinos' mobilization efforts [against the incumbent] but also acted against those Latinos who were becoming most politically active, dividing them with a district line through the middle of Laredo." (156) This sort of deliberate community disruption, the Court held (with Justice Kennedy writing), was prohibited. (157)

Analogously, the problem with the State's proposed remedial district was that, while it had a Hispanic majority, it fused together two very different Hispanic communities. "[T]here was a 300-mile gap between the Latino communities [in the district], and a similarly large gap between the needs and interests of the two groups." (158) Hispanics along the Mexican border and in Austin were "distant, disparate communities," with major "differences in socio-economic status, education, employment, health, and other characteristics." (159) A district that combined these "two farflung segments of a racial group with disparate interests" thus could not cure the section 2 violation identified by the Court. (160) A better example of how the territorial community test would operate (at least with regard to racial minorities) is hard to imagine. District-community congruence had become, in LULAC, the Court's measure for both vote dilution and any effort to remedy the wrongdoing. (161)

3. Racial Gerrymandering

Though the literature has not yet recognized it, the Court also has employed something akin to the territorial community test in the context of racial gerrymandering. In this field of equal protection doctrine, which originated with the 1993 decision of Shaw v. Reno (Shaw I), a plaintiff may challenge a district on the ground that race was the predominant motive for the district's creation. (162) In a series of decisions since Shaw I, the Court uniformly has struck down districts that combined disparate communities or deviated substantially from the larger communities in which they were located. On the other hand, the Court invariably has upheld districts that coincided with underlying geographic communities. District-community congruence therefore seems to be an ironclad defense to the charge that a district is a racial gerrymander.

In both Shaw I and its 1996 sequel, the Court objected to a majority-black North Carolina district that joined together "tobacco country, financial centers, and manufacturing areas." (163) The Court commented that districts that combine minority populations from different geographic communities are constitutionally suspect. "A reapportionment plan that includes in one district individuals who belong to the same race, but who are otherwise widely separated by geographical and political boundaries ... bears an uncomfortable resemblance to political apartheid." (164) In a 1995 case, the Court likewise invalidated a Georgia district that lumped together blacks in Atlanta and coastal Chatham County. (165) These communities were "260 miles apart in distance and worlds apart in culture," and they were characterized by "fractured political, social, and economic interests." (166) Their fusion plainly was not motivated by concern for community integrity, and indeed could be explained only by prohibited racial considerations.

The communal harm was somewhat different in the 1996 case of Bush v. Vera. (167) The three Texas districts at issue did not merge different communities, but they did markedly concentrate what were relatively diffuse minority populations. One Dallas district meandered around the city in order "to connect dispersed minority population[s]," (168) using "far-reaching tentacles that intricately and consistently maximize[d] the available ... African-American population." (169) Analogously, the "interlocking shapes" of two Houston districts were "almost exclusively[] the result of an effort to create, out of largely integrated communities, both a majority-black and a majority-Hispanic district." (170) The Court found all three districts unconstitutional, in large part because they did not reflect accurately the characteristics of the broader communities in which they were drawn. Again, the disjunction between constituency and community signaled that race had played too large a role in the districts' creation.

Conversely, the Court's more recent racial gerrymandering decisions all have upheld districts that did correspond to underlying territorial communities. In one 1997 case, the Court approved a Georgia district plan that avoided splitting counties, noting that "[t]hese small counties represent communities of interest" that should not be needlessly "chopp[ed] ... in half." (171) In another 1997 case, the Court endorsed a Tampa Bay district whose residents "regard[ed] themselves as a community" and that "comprise[d] a predominantly urban, low-income population ... whose white and black members alike share[d] a similarly depressed economic condition and interests that reflect[ed] it." (172) And in the 1999 and 2001 chapters of North Carolina's Shaw saga, the Court twice upheld a revised version of the district it previously had invalidated. (173) One reason for the Court's change of heart was that the updated district "joined three major cities in a manner ... reflecting a real commonality of urban interests, [such as] inner city schools, urban health care[, and] public housing problems." (174) The updated district, in other words, possessed the coherent communal identity that its predecessor had lacked.

4. Political Gerrymandering

In the political gerrymandering arena, unlike in the above areas, a majority of the Court has never embraced the proposition that electoral districts should correspond to territorial communities. At the same time, the Court has never rejected the proposition either. The Court has rebuffed all sorts of other potential standards for political gerrymandering cases: for example, that voters should not experience a consistent degradation of their political influence, that the predominant motive for a district's creation should not be partisan advantage, and that statewide minority parties should not be able to entrench themselves in office. (175) But the territorial community test proposed by this Article has never been definitively assessed.

However, the notion that districts and communities should coincide has animated several Justices' separate opinions in political gerrymandering cases. A template for analyzing district-community congruence-which could be adopted in the future without raising any stare decisis concerns--thus exists already in the case law. In the 1983 case of Karcher v. Daggett, for instance, Justice Stevens's concurrence argued that New Jersey's district plan was unconstitutional because, among other things, it brazenly disrupted communities across the state. (176) The plan "wantonly disregard[ed] county boundaries" and evinced "little effort to create districts having a community of interests." (177) The residents of several districts were served by "different television and radio stations, different newspapers, and different transportation systems." (178) Two districts were particularly offensive because they merged, respectively, "New York suburbs [and] the rural upper reaches of the Delaware River" and "industrial Elizabeth, liberal, academic Princeton and largely Jewish Marlboro." (179)

Similarly, in the 1986 case of Davis v. Bandemer, Justice Powell's separate opinion endorsed district-community congruence at the levels of both theory and practice. (180) From a theoretical standpoint, "[a]dherence to community boundaries ... both deter[s] the possibilities of gerrymandering, and allow[s] communities to have a voice in the legislature that directly controls their local interests." (181) But districts that do not correspond to communities, in Justice Powell's view, foster voter uncertainty and apathy. "Confusion inevitably follows ... when a citizen finds himself or herself forced to associate with several artificial communities," and "the potential for voter disillusion and nonparticipation is great, as voters are forced to focus their political activities in artificial electoral units." (182)

Applying these principles, Justice Powell found much to criticize in Indiana's district plan (which he would have invalidated). As a whole, the plan showed "no concern for any adherence to principles of community interest," as it "carved up counties, cities, and even townships" into "strange shapes lacking in common interests." (183) One especially objectionable district combined "blacks in Washington Township and white suburbanites in Hamilton and Boone Counties," while another merged "Allen and Noble County farmers with residents of downtown Fort Wayne." (184) Fort Wayne itself was cut in two, and each half was linked to faraway rural counties. (185) Around Indianapolis as well, an irregular C-shaped district "include[d] portions of the urban south-westside of the city, the airport and suburban area ... on the west side, and the Meridian Hills area at the northern part of the county." (186) With its strong theoretical mooring and detailed factual analysis, this opinion is the best example in the Court's case law of how the territorial community test would function (particularly with regard to nonracial communities).

In the Court's two most recent political gerrymandering decisions, Vieth in 2004 and LULAC in 2006, the dissenters once again called attention to districts that failed to respect community boundaries. In Vieth, Justice Stevens argued that a Pennsylvania district that "loom[ed] like a dragon descending on Philadelphia from the west, splitting up towns and communities throughout Montgomery and Berks Counties" should be struck down. (187) In LULAC, likewise, he would have held unconstitutional four Texas districts that were formed when a "minority community ... was splintered and submerged into majority Anglo districts in the Dallas-Fort Worth area." (188) The issue of district-community congruence was not as fully developed in these opinions as in Karcher and Bandemer (in part because of the parties' litigation strategies (189)), but the communal strand still remained salient. Were the Court so inclined, it could easily weave this strand into its future gerrymandering decisions.

5. State Redistricting Law

It is not just in federal doctrine that the territorial community has made a comeback in the last few decades. In state law too, it is now recognized in a number of constitutions and statutes, and in even more non-binding redistricting guidelines. State courts also tend to conceive of gerrymandering as the disruption of organic geographic communities and the creation of artificial political cleavages--precisely the definition implicit in the territorial community test. In recent years, indeed, a number of state courts have used the test to strike down districts that did not adhere to community boundaries and to uphold districts that did.

The constitutions of five states (Alaska, Arizona, California, Colorado, and Hawaii) expressly require districts to reflect community interests. These provisions all were adopted between 1959 and 2010, and they include such formulations as Alaska's mandate that districts contain "a relatively integrated socio-economic area," (190) California's declaration that "[t]he geographic integrity of any ... local community of interest shall be respected in a manner that minimizes [its] division," (191) and Hawaii's admonition that communities not be "sun merge[d]" within areas where "substantially different socio-economic interests predominate." (192) Seven additional states have statutory requirements that districts correspond to communities where possible, and a further twelve states adopted non-binding guidelines including similar provisions during the last redistricting cycle. (193) These statutes and guidelines typically use the term "community of interest," and when the term is defined, it usually refers to the shared social, cultural, and economic interests of people living in a particular area. (194)

Consistent with these provisions, state courts tend to embrace the theory of communal representation (discussed above in Section I.A) and to understand gerrymandering as the violation of territorial communities. James Gardner recently completed a thorough survey of the state constitutional law on redistricting, in which he reached the following three conclusions about the views of the state courts: (195) First, they generally consider the entity that is represented in the legislature to be not the individual person (or any arbitrary set of people), but rather a geographic area that corresponds to particular interests. "[T]erritory and interest ... typically are thought to coincide. People who live in the same place ... are thought to have similar interests ... [that] justify representing them in the legislature by territorial groupings." (196) Second, the reasons why state courts believe that geographic proximity generates a sense of community are (1) that people who live nearby "share a common local economy and economic life" and (2) that they "participate together in the public life of a shared unit of political and governmental administration." (197) Community flows, in other words, from shared commercial and civic experience.

Third, and most important here, state courts usually define gerrymandering as the needless disruption of geographic communities. Districts are supposed to correspond to such communities in order to produce a political life that is "harmonious at home due to the unity of local economic interest" and "conflictual far away in the state legislature due to the ... diversity of the interests represented." (198) Gerrymandering, alas, interferes on both sides of the equation. At the district level, it divides communities and combines people with conflicting interests, thus fostering acrimony. (199) At the state legislative level, it encourages partisanship and ideological extremism instead of hearty pluralism. "[P]recisely because it disregards the 'natural' territorial cleavages ... that divide the state populace," gerrymandering enables the ascendance of party and ideology. (200) Or so, at least, say the state courts.

Based on this conception of gerrymandering, courts in ten states (Alaska, (201) Arizona, (202) Colorado, (203) Idaho, (204) Kansas, (205) Montana, (206) North Carolina, (207) Oregon, (208) Rhode Island, (209) and Vermont (210)) have invalidated districts that did not correspond to geographic communities and upheld districts that did. These decisions--which vividly illustrate the territorial community test in action--all were handed down between 1974 and 2006, with their frequency increasing in recent years. In Alaska, Arizona, Colorado, North Carolina, and Rhode Island, the decisions were based on state constitutional requirements that districts and communities coincide, while the Idaho, Kansas, Montana, Oregon, and Vermont cases relied on analogous statutory and guideline provisions. In sum, courts have assessed more than fifty districts based on their adherence to community boundaries, of which a solid majority have been upheld. (211)

Examples of districts that have been struck down include: an Alaska district that joined the hub of an agricultural region with coastal fishing villages; (212) another Alaska district that merged suburban Wasilla with urban Anchorage; (213) a Colorado district that combined very different neighborhoods within Denver; (214) two more Colorado districts that divided Aspen from neighboring Snowmass; (215) three Idaho districts that split the state's rugged southeast corner; (216) and a Vermont district that lumped together towns on opposite sides of the Green Mountains. (217) Conversely, districts have been upheld where there was evidence that their residents worked in similar industries, shared a racial or ethnic heritage, interacted in community organizations, received information from the same media sources, or were connected by transport links. (218) In these circumstances, the courts could not conclude that any community disruption--with its negative implications for harmony in the district and partisanship in the legislature--had occurred.


The above Parts demonstrate that the principle of district-community congruence has a robust theoretical, historical, and doctrinal pedigree. It is rooted in a theory of representation that is particularly consistent with the American commitment to geographic districting. It has functioned as a crucial touchstone for district drawers for most of American history. And, though neither courts nor scholars previously have noticed, it continues to drive judicial outcomes in a host of election law domains.

Accordingly, there is at least a plausible basis for the Supreme Court to constitutionalize the territorial community test. In Philip Bobbitt's terminology, the test is supported by ethical, historical, and doctrinal modalities of constitutional argument. (219) To be sure, an explicit textual hook is missing, but similar concerns did not prevent the Court from endorsing the one-person, one-vote rule (the only Equal Protection doctrine that ignores intent altogether) or the prohibition on racial gerrymandering (the only standard that does not require any particularized harm). Moreover, as Issacharoff and Pildes have noted, the Constitution addresses electoral practices in such oblique terms that the Court has long had to fashion the law of democracy mostly by reference to overarching structural principles. (220) Justices and academics alike have therefore sought to combat political gerrymandering without worrying much about the specific textual bases for their efforts. (221) The territorial community test fits squarely within this tradition.

This Part, then, considers how the test might operate if it were converted by the Court--or enacted by Congress, the states, or the people themselves--into actual doctrine. (222) I first define the territorial community as clearly as possible, drawing on state and federal case law as well as state constitutional provisions. I next explain how judges might employ the territorial community test to assess district plans, again using doctrinal examples and focusing on the three ways in which communities can be disrupted: fusion, fragmentation, and subversion. Lastly, I discuss the test's links to adjacent election law domains, arguing that it would promote doctrinal coherence, reduce violations of other redistricting rules, and lessen the salience of divisive racial rhetoric.

A. Defining the Territorial Community

As used in this Article, a territorial community is (1) a geographically defined group of people who (2) share similar social, cultural, and economic interests and (3) believe they are part of the same coherent entity. The first element, geographic demarcation, is necessary because of the American commitment to geographic districting. While non-geographic communities certainly exist, they cannot easily be captured by districts that are drawn spatially. The second element, shared interests, is mostly objective in character and gives rise to groups of people who are affected in similar ways by (and usually have similar positions on) the gamut of governmental actions. The common concerns that are most relevant here are those that bear on some matter of actual or potential governmental policy. Lastly, the third element, a feeling of communal affiliation, is subjective in nature. It ensures that members of a community actually understand themselves to be part of the same cognizable unit. (223)

This definition draws from earlier judicial and state constitutional efforts. The Supreme Court's Gingles factors, for example, require that a minority group be geographically compact, politically cohesive, and, at least as construed in LULAC, socioeconomically and culturally unified. (224) My conception of the territorial community is not very different from the application of the Gingles factors to all groups rather than only to racial minorities. Similarly, state and lower federal courts often have drawn districts so that they correspond to "[t]he social and economic interests common to the population of an area which are probable subjects of legislative action" (225) or to "distinctive units which share common concerns with respect to ... demography, ethnicity, culture, socio-economic status or trade." (226) While these standards do not take into account people's subjective sense of affiliation, they are otherwise analogous to my approach. So too are state constitutional provisions that define the territorial community in objective terms: e.g., "a relatively integrated socio-economic area" (Alaska), (227) or "a contiguous population which shares common social and economic interests" (California). (228)

I use the term "territorial community" instead of the more common "community of interest" because of certain connotations that the latter phrase has acquired. For one thing, a community of interest does not have to be spatially bounded, meaning that it coexists uneasily with the American system of geographic districting. In addition, a community of interest can be deemed to arise on the basis of any common concern, making the term notably imprecise and malleable. (229) With its strong geographic valence and emphasis on the full array of interests and affiliations that people share, the concept of a territorial community seems substantially more determinate.

A territorial community also is not quite the same thing as a political subdivision such as a town or county. The two may sometimes be functionally identical, both because subdivisions tend to be inhabited by people with similar socioeconomic characteristics, (230) and because civic ties can foster a sense of kinship. But communities and subdivisions are often different as well--as when people's interests and affiliations do not follow subdivision lines, or when subdivisions contain within them more than one (or only part of a) community. (231) There is thus nothing wrong with judges beginning their inquiry with subdivision boundaries, but the task of identifying territorial communities cannot end there. (232)

Furthermore, territorial communities exist, and should be represented in the legislature, at different levels of generality. As the geographer David Knight has written, "[W]e have ties to different scales of territory and ... we can operate at several levels of abstraction at any one time--from personal to small group, to a parochial localism ... to a broader regionalism." (233) Quite specific groups can therefore form state house (or even smaller-scale) districts, while broader communities can be captured by state senate districts, and yet more diffuse groups by congressional districts. The particularity of the community that comprises a given district typically varies in accordance with the district's size. (234)

Examples illustrating these definitional points abound in state and federal law. The significance of geographic demarcation (the first element of a territorial community) is conveyed by the Supreme Court's racial redistricting cases. According to the Court, far-flung, noncontiguous groups of minority voters--like the Hispanics along the Mexican border and in Austin in LULAC, and the African Americans scattered across North Carolina in Shaw I--do not constitute cognizable communities. (235) Because of their lack of geographic connectedness, they do not warrant their own districts under section 2 of the VRA, and the Equal Protection Clause may well be violated if they are placed within the same constituencies. (236)

These decisions also demonstrate some of the shared interests that can give rise to genuine communities (the second element). Race alone is never enough, in the Court's view, but race plus cultural isolation (as with the Hispanics in San Antonio's Barrio in White), or race plus urban poverty (as with the Tampa Bay African Americans in Lawyer), can suffice. (237) Outside the Court's case law, it is common socioeconomic interests that are most often thought to generate communities. California's constitution, for instance, specifies that a territorial community can be "an urban area, a rural area, an industrial area, or an agricultural area," or an "area[] in which the people share similar living standards ... [or] have similar work opportunities." (238) Analogously, lower courts have recognized communities such as Alaskan fishing villages, (239) rural farmland in Oregon, (240) ski towns in the Rockies, (241) industrial areas in California, (242) bedroom suburbs of Denver and San Francisco, (243) and the urban cores of several major cities. (244) Additional factors the courts have deemed relevant to community formation include transportation links between areas, (245) membership in regional organizations, (246) the level of residents' commercial interaction," (247) and their degree of reliance on the same media outlets. (248)

Finally, with regard to subjective affiliation (the third element of a territorial community), a number of cases have highlighted its importance and shown that it does not always coincide with objective interests. In Arizona, for example, the neighboring Navajo and Hopi tribes share many socioeconomic concerns but are also historical adversaries with clashing identities; accordingly, they have long been placed in different congressional districts. (249) Similarly, in Colorado, adjacent Pueblo and El Paso Counties resemble each other in several respects, but nevertheless do not comprise an authentic community because they are "commercial rivals" and "hereditary enemies." (250) Conversely, the Hispanics in the original Texas district in LULAC (which the Court endorsed) were just as heterogeneous as the Hispanics in the new district (which the Court struck down). (251) The first group, however, was far more "cohesive and politically active"--that is, a far more subjectively unified community. (252)

B. Doctrinal Details

Armed with a clearer understanding of what one is, how should a court go about deciding whether a district does or does not disrupt a territorial community? In my view, there are three types of disruption that the court should consider: fusion, fragmentation, and subversion. Any of these should be enough to invalidate a district, but the court should be relatively deferential toward any explanation offered by the State as to how the district in fact corresponds to a community. The court should also keep in mind that the one-person, one-vote rule makes inevitable a certain amount of community disruption--but that this disruption can and should be minimized by intelligent district drawing.

The first kind of disruption is fusion: the needless placement of different territorial communities within the same district. Fusion is what occurred in LULAC when Hispanics along the Mexican border and in Austin were joined together even though they were "distant, disparate communities." (253) It was also fusion when African Americans from North Carolina's "tobacco country, financial centers, and manufacturing areas" were merged in Shaw I, (254) and when blacks in Atlanta, Savannah, and coastal Chatham County--"260 miles apart in distance and worlds apart in culture"--were lumped together in Miller. (255) In all three cases, of course, the Court struck down the offending districts. (256)

The second type of disruption is fragmentation: the unnecessary division of a territorial community among multiple districts. Fragmentation took place in LULAC when the cohesive Hispanic community that previously had been placed in a single district was dispersed among several new districts in order to protect an endangered incumbent politician. (257) Fragmentation also occurred in the 1977 case of United Jewish Organizations of Williamsburgh, Inc. v. Carey when a group of Hasidic Jews in Brooklyn was split between two state assembly and two state senate districts. (258) While a divided Court ruled against the Hasidim (in a period prior to the Court's recognition of a cause of action for racial gerrymandering), (259) Chief Justice Burger wrote in dissent that "members of an ethnic community" have "the constitutional right not to be carved up" for another group's benefit. (260)

The final kind of disruption is subversion: the creation of a district that diverges sharply from the defining characteristics of the larger community in which it is located. Communities are not perfectly uniform in their spatial composition, meaning that it is sometimes possible to draw districts that fit entirely within them but that deviate dramatically from their overall tenor. In Bush, for example, a largely integrated area around Houston was sliced by highly convoluted lines into one majority-black district and one majority-Hispanic district. (261) These boundaries subverted what was a genuinely mixed community into two unrepresentative halves. Similarly, in the 1960 case of Gomillion v. Lightfoot, the city border of Tuskegee, Alabama was converted from a square into an "uncouth twenty-eight-sided figure" that carefully separated white and black voters. (262) A biracial area was thus transformed into a white city and a black hinterland. The Court invalidated the schemes in both cases. (263)

Presented with a claim that a territorial community has been disrupted, a court should display reasonable deference toward the State's justification for the challenged district. Such respect is appropriate, first, because there is--notwithstanding this Article's efforts in the previous Section--a fair amount of fuzziness in the concept of a territorial community. When a State is able to mount a solid case that a district corresponds to a community, relying on evidence contemporaneous with the district's creation, it follows that judges should uphold the district even if they disagree personally as to the degree of district-community congruence. Judicial restraint is also necessary in order to reduce the potential invasiveness of the territorial community test. Were judges free to strike down every district that, in their considered view, failed to adhere to a community, a large number of districts would end up on thin ice. Limiting the test to cases where the State cannot present a credible defense preserves legislative authority over redistricting and focuses the doctrine on severe instances of community disruption.

For precisely these reasons, state courts that employ something akin to the territorial community test universally have adopted quite deferential standards of review. They invalidate districts only when they find clear constitutional or statutory violations, and not when they merely quarrel with the legislature's reasoned explanations for its district drawing choices. In Vermont, for instance, the state supreme court has stated that "it is primarily the Legislature, not this Court, that must make the necessary compromises to effectuate state constitutional goals." (264) The court is willing to strike down districts only when plaintiffs demonstrate "the absence of a rational or legitimate basis for the challenged plan []." (265) Similarly, Colorado's supreme court has characterized its review as "narrow" and commented that "we should not substitute our judgment for the Commission's unless we are convinced the Commission departed from constitutional criteria." (266) These statements are consistent with my view of how the territorial community test should operate.

A related issue for a court applying the test is that districts and communities can never coincide perfectly thanks to the one-person, one-vote rule. Communities, of course, do not come in populations that are tidy multiples of the ideal district size, meaning that some degree of community disruption is inevitable whenever districts are drawn. This disruption, however, can and should be minimized by savvy district drawing. For example, when smaller communities must be combined in order to form a sufficiently populous district, groups that are as similar as possible in their interests and affiliations should be joined. Analogously, when a larger community must be divided into multiple districts, these districts should either correspond to more specific subcommunities, if they exist, or else mirror the characteristics of the broader community. (267) The point is that the needless fusion, fragmentation, and subversion of communities should be avoided, so that the one-person, one-vote rule does not result in more community ruptures than are necessary. (268)

The case law provides numerous illustrations of how this can be done. Presented with an island borough that was too small to support its own state house district, for instance, the Alaska Supreme Court held that it should be combined with a nearby lake region rather than an adjacent peninsula. The court's rationale was that the island and the lake region were more closely related thanks to their municipal ties and involvement in commercial fishing. (269) Similarly, a federal district court in Colorado that was forced to divide Denver because it was too large for a single congressional district severed the city's southwest corner and thereby left it "more compact" and with its "minority and neighborhood communities" intact. (270) The court also merged small Denver suburbs and split counties (where it had no other option) based on "the stark contrast between the concerns of the expanding municipalities and the outlying rural areas." (271) This is the sort of sensible district drawing, attentive to both population figures and communal ties, that courts should expect (and require) from legislatures.

A final point about the territorial community test is that it need not serve as the judiciary's only tool for combating political gerrymandering. In fact, because it focuses on specific districts' congruence with communities, it could be complemented effectively by standards that aim to cure statewide pathologies. The two obvious candidates, both discussed earlier, are measures of district competitiveness and gauges of how fairly a district plan treats the two major parties. (272) While my position remains that the territorial community test is more consistent with the American commitment to geographic districting than these approaches, I certainly have no objection to their adoption alongside it. Political gerrymandering is a wily enough foe that it may well take a set of standards to subdue.

C. Relation to Other Domains

It should be clear by now that there are many similarities between the territorial community test and the existing bodies of state and federal law that regulate redistricting. State legislative districts that correspond to political subdivisions (which themselves often coincide with territorial communities) are permitted to deviate substantially from the population otherwise dictated by the one-person, one-vote rule. (273) Minority groups that constitute territorial communities are commonly entitled to districts in which they can elect the candidates of their choice under section 2 of the VRA. (274) Majority-minority districts that mirror territorial communities are essentially immune from claims that they are unconstitutional racial gerrymanders. (275) And many states already abide by constitutional or statutory requirements that their districts respect territorial community boundaries. (276)

Accordingly, were the territorial community test adopted as a means of curbing political gerrymandering, much of the law of redistricting would be elegantly harmonized. The same core inquiry--the degree to which electoral districts and territorial communities coincide--would help determine: (1) whether districts can diverge from perfect equipopulation; (2) whether section 2 of the VRA has been violated; (3) whether districts are unlawful racial gerrymanders; (4) whether certain state redistricting rules have been followed; and (5) whether impermissible political gerrymandering has taken place. Judges and scholars who value coherence in the law should welcome such doctrinal convergence, particularly given the confusion that has long reigned in the redistricting realm. (277)

The benefits of convergence also would not be merely aesthetic. Parties responsible for crafting districts, who often complain about the plethora of applicable requirements, would have a single, straight-forward directive with which to comply (in addition to the one-person, one-vote rule): that districts should correspond as closely as possible to territorial communities. As long as this mandate was satisfied, state legislative districts could deviate somewhat from perfect population equality, and all districts would be largely insulated from VRA section 2, racial gerrymandering, certain state law, and political gerrymandering challenges. The legal uncertainty that surrounds most contemporary district plans would be substantially reduced.

As for the courts, they would presumably find fewer breaches of the redistricting rules once district drawers fully grasped the need for district-community congruence. Districts designed to comply with the territorial community test, of course, would typically violate neither it nor any other requirement. Moreover, when the courts did find breaches, they would often be able to avoid framing their decisions in inflammatory racial terms. (278) Much of the redistricting case law currently revolves around racial issues because several key causes of action (e.g., racial vote dilution and racial gerrymandering) are race-related. But since districts that offend section 2 of the VRA or the prohibition on racial gerrymandering also frequently run afoul of the territorial community test, they could be struck down, in many cases, on communal grounds alone. Incendiary claims about discrimination and racial motivation might thus be limited at the same time that the doctrine's overall coherence would be enhanced.


The Supreme Court's rationale in Vieth and LULAC for rejecting all of the potential standards that it considered was not that it deemed them deficient on the merits. The Court's justification, rather, was that none of the standards was sufficiently "judicially discernible and manageable." (279) None of the standards, in other words, could give rise to decisions that would be "principled, rational, and based upon reasoned distinctions"--and therefore the whole field of political gerrymandering, according to the Vieth plurality, represented a nonjusticiable political question." (280)

Vieth and LULAC impose a heavy burden on anyone who proposes a new standard for curbing political gerrymandering. The standard not only must be sound as a matter of constitutional law, but, in contrast to the array of approaches the Court already has evaluated, it also must be judicially workable. This burden is particularly weighty for the territorial community test since the objection most commonly posed to it is that community boundaries (let alone how well they match up with district lines) cannot be reliably determined. (281)

This Part, then, takes on the challenge of demonstrating that the territorial community test is manageable. I first argue that the test avoids the problem that has led the Court to find every other approach wanting: the judiciary's alleged inability to assess partisan motives and outcomes. I next discuss the courts' experiences to date with standards similar to my own. My conclusion is that both the Supreme Court and the state courts have had no particular difficulty ascertaining community boundaries or comparing them to district configurations. I then explore the relevant political science literature, which also shows that communities can be identified and that district-community congruence can be measured. Finally, I contend that any remaining vagueness is actually beneficial because it would encourage risk-averse district drawers to avoid provocative district plans.

A. Sidestepping the Problem

In both Vieth and LULAC, the Court dismissed as unmanageable all the standards that it considered for the same two reasons: (1) they required the courts to make impossible evaluations of the political motives underlying district plans; or else (2) they asked the courts to decide whether specific electoral outcomes were sufficiently "fair." For instance, the predominant-partisan-intent tests advanced by appellants in both cases, as well as by Justice Stevens, were deemed unworkable because the Court could see no way to determine whether partisanship in fact outweighed all other redistricting considerations. (282) Similarly, Justice Souter's emphasis on partisans' "packing" or "cracking," and Justice Breyer's minority-party-entrenchment standard, were considered untenable because people's political affiliations are changeable and there is no way to tell how much power a party should have. (283) "'Fairness,'" declared the Vieth plurality, "does not seem to us a judicially manageable standard." (284)

Whatever one may think of these critiques, they plainly leave the territorial community test unscathed. Unlike the standards assessed in Vieth and LULAC, it does not require the courts to determine how partisan district drawers' motives are--or even to adjudge intent at all. Also unlike those standards, it does not necessitate any analysis of voters' political affiliations or parties' levels of electoral success. Instead, the territorial community test focuses exclusively on where geographic communities are located and how well electoral districts correspond to them. These sorts of questions, which avoid partisan motives and outcomes altogether, have never been held by the Court to be judicially unanswerable. They are nowhere to be found in the "sea of imponderables" bemoaned by the Vieth plurality. (285)

B. The Supreme Court's Experience

Not only has the Court never declared the territorial community test to be unworkable, but it has successfully employed something similar to it in a wide range of cases. In these decisions, the Court has frequently, and without any obvious hardship, identified the boundaries of geographic communities and reached well-reasoned conclusions as to whether they were disrupted by district plans. While these conclusions have not always been unanimous, no Justice has ever complained that the inquiry itself is somehow beyond the judiciary's ability. The Court's own experience thus provides compelling evidence that the territorial community test is manageable.

In the reapportionment context, first, at least five of the Court's decisions have determined that towns or counties comprised distinct communities, and that districts corresponding to these units could, for this reason, diverge substantially from perfect population equality. (286) While the Justices have disagreed as to whether population deviations should be allowed in the first place, no Justice has claimed that it is impossible to tell whether political subdivisions amount to genuine communities. Even Justice Brennan, the Court's most ardent champion of the one-person, one-vote rule, conceded in a 1983 opinion that a "longstanding policy of using counties as the basic units of representation" could be applied "rational[ly]" and without "arbitrariness or discrimination." (287)

Similarly, the Court has considered whether minority groups constitute territorial communities (and hence require districts of their own) in close to ten racial vote dilution cases. (288) These decisions have split almost evenly between holdings in favor of and holdings against the minority groups. In some of the decisions in their favor, the Court recognized territorial communities such as Mexican American residents of San Antonio's Barrio, (289) poor African Americans in rural Georgia, (290) and politically mobilized Latinos in southwestern Texas. (291) In some of the adverse decisions, the Court held that assorted minorities around Minneapolis, (292) and geographically scattered African Americans in Georgia, (293) North Carolina, (294) and Texas, (295) did not qualify as coherent communities. Never in any of these cases did the Court express any existential angst about the inquiry it was conducting. Indeed, in a recent vote dilution decision, the Justices were able both to ascertain the boundaries of a Texas Latino community, and to conclude that the old district lines corresponded to them while the new lines did not. (296) If this sort of analysis is unworkable, there is no hint of it in the case law.

Nor is there any sign of trouble in the Court's racial gerrymandering decisions, of which another ten or so have examined whether districts coincided with territorial communities in order to assess claims that the districts were created for primarily racial reasons. (297) In these decisions too, the Court repeatedly has shown itself capable of evaluating community boundaries as well as district-community congruence. For instance, the Court struck down a North Carolina district that combined African Americans from across the state, (298) a Georgia district that joined blacks in inland cities and rural coastal areas, (299) and racially homogeneous Texas districts carved out of racially integrated regions. (300) On the other hand, the Court upheld a Tampa Bay district composed of poor urban African Americans (301) and a revised North Carolina district that confined itself to urban centers in the Piedmont. (302) Nothing would seem to differentiate these cases (at least in terms of judicial manageability) from the rest of the Court's equal protection docket. Their outcomes, reflecting the consistent application of something akin to the territorial community test, appear just as "principled, rational, and based upon reasoned distinctions" (303) as those of any other constitutional cases.

To be sure, the Justices have sometimes disagreed about questions of community in these decisions. In Bush, for example, Justice Stevens argued in dissent that the Dallas district rejected by the Court in fact had been "drawn to align with certain communities of interest, such as land use, family demographics, and transportation corridors." (304) The plurality was unpersuaded because the legislature had not considered this information when it drew the district and the district correlated better with racial than community lines. (305) In LULAC, similarly, Chief Justice Roberts dissented on the ground that the State's proposed remedial district (which the majority disallowed) was just as attentive to Texas Latino communities as the original district that the state dismantled. (306) The majority conceded that the districts were similar in several respects, but held that the original district corresponded to a more politically cohesive Latino group. (307)

For present purposes, it is unimportant who had the better of these disputes. The key point, rather, is that they are no different from the disagreements that characterize all non-unanimous Court decisions. As with any standard less determinate than one-person, one-vote, judges applying the territorial community test will occasionally reach different conclusions with regard to the same sets of facts. But this reality does not make the test unmanageable; it simply makes it par for the constitutional course.

C. The State Courts' Experience

A skeptic might point out that almost all of the Court's experience with the territorial community test has involved racial communities. Perhaps these communities are easier to identify and to compare to district lines than communities oriented along other axes. The skeptic might add that district-community congruence is not the central issue in either racial vote dilution or racial gerrymandering doctrine. Perhaps the Court would have more trouble with the territorial community test if it were the dispositive inquiry rather than a subsidiary consideration.

These (perfectly valid) concerns are largely allayed by the experiences of the state courts. In a series of cases spanning four decades, these courts have frequently recognized geographic communities that were defined by non-racial criteria. They also have repeatedly examined whether districts corresponded to these communities--not as a subprong of some other standard, but rather as the core element of their political gerrymandering analysis. The body of doctrine produced by these decisions is substantively rich; it has led to both the affirmation and invalidation of challenged districts (usually unanimously); it has evolved, in classic common law fashion, through the refinement of earlier precedent; and it has almost never been denounced as unworkable.

By my count, courts in at least ten states, in at least twenty-two decisions over the past four redistricting cycles, have evaluated at least fifty districts based on their congruence with territorial communities. (308) About two-thirds of these districts were upheld, while about one-third were struck down. (309) About two-thirds of the decisions also were unanimous, while only three were decided even in part by bare majorities (310) That so many districts have been assessed under the territorial community test, yielding both affirmations and invalidations, and typically by lopsided margins, strongly suggests that the test is judicially manageable. (311) An unmanageable approach, in all likelihood, would not have survived as long or generated as much consensus with regard to results on both sides of the doctrinal line. (312)

A closer examination of the state case law sheds light on the factors that most often prove decisive for districts' fates. That such generalizations are possible itself bolsters the case for the test's workability. On the one hand, districts typically were struck down when it was clear to the reviewing courts that they merged unrelated communities or divided what plainly were unified communities. Some examples of community fusion and fragmentation were noted earlier. (313) Others include an Alaska district that joined the "physically and economically segregated" town of Cordova with the Inside Passage region; (314) another Alaska district that "mixe[d] small, rural, Native communities with the urban areas of Ketchikan and Sitka"; (315) two Denver districts that "split[] the Five Points community along its main business route"; (316) and two North Carolina districts that separated a county seat from the rest of the jurisdiction. (317) In many of these cases, the state was unable to produce any proof that community boundaries had been respected (or even considered). (318) Often, it was apparent that the basis for the state's district drawing choices had been political advantage rather than community preservation. (319)

Conversely, the state courts have generally upheld districts when there was reasonable evidence of district-community congruence and this evidence was actually taken into account by the redistricting body. Some common types of evidence have included shared economic pursuits, similar income levels, significant social and commercial interaction, good transportation links, reliance on the same media outlets, and common membership in regional organizations. (320) Presented with such evidence, the courts have typically rejected gerrymandering challenges even when the plaintiffs were able to mount plausible cases of community disruption. (321) The courts have emphasized, however, that the evidence should have been compiled and consulted at the time the districts were formed--not introduced later in the heat of litigation. It was significant in a number of cases that the district drawers themselves considered community boundaries and gave reasons for their actions. (322) Absent such a contemporaneous record, the districts may well have been struck down.

A different kind of confirmation of the territorial community test's manageability can be gleaned from the case law of Alaska (the state with the most extensive experience with it). Over four decades, in quintessential common law fashion, the Alaskan courts have cited their earlier precedents, refined them, and relied on them to determine the outcomes that different fact patterns should produce. That the territorial community test can generate such a familiar sort of doctrine shows that it is no more vague or indeterminate than most legal standards.

The Alaska Supreme Court first used the test in 1974, permitting population deviations for districts that coincided with communities and prohibiting them for districts that did not. (323) In the court's next redistricting decision, in 1983, it looked to its earlier case for the applicable standard of review as well as the definition of the relevant constitutional provision. (324) Four years later, the court carefully analyzed its earlier precedents to decide how to resolve challenges to three districts. Noting that district validity "can be determined by way of comparison with districts which we have previously [examined]," the court concluded that a district combining two Anchorage suburbs was "[u]nlike the district linking Cordova and the Southeast which we invalidated in [1983]," but "[l]ike the Juneau District upheld in [1974]." (325) By 1992, the court had an even deeper pool of precedents to draw from--deep enough that it could start specifying the hallmarks of sound and unsound districts. "In our previous reapportionment decisions we have identified several specific characteristics of [permissible districts]," the court explained, cataloging factors such as "common major economic activity," "the predominantly Native character of the populace," "historical economic links," and "transportation ties." (326) Through the alchemy of the common law, a delphic constitutional command had become a detailed doctrinal infrastructure for deciding cases.

Some further support for the territorial community test's workability stems from the kinds of disagreements that it has produced. As noted above, courts applying the test have not disagreed much; most decisions have been unanimous and only a handful have deeply divided the reviewing judges. (327) When judges have come to different conclusions, moreover, they usually have done so for the most ordinary of reasons: diverging views on whether certain facts satisfy a legal standard. An Alaska judge, for example, objected to a district's invalidation because, while a "close question," he thought "it seem[ed] relatively reasonable to include Cordova with the other waterlocked fishing communities." (328) Similarly, Colorado judges quarreled over whether the division of certain communities was unavoidable due to the one-person, one-vote rule (while agreeing that the division was regrettable). (329) Only once, in the case law that I surveyed, did I find any claim that the territorial community test is particularly difficult to apply. A Vermont judge labeled it "vague and tentative" and criticized its "breadth and imprecision"--but nevertheless used it to analyze the validity of six districts, dissenting only as to one. (330)

D. The Political Science Literature

While the courts' own experiences are the most persuasive proof of the territorial community test's manageability, the political science literature offers some additional confirmation. Political scientists have managed to identify geographic communities both directly, through the analysis of large volumes of socioeconomic and survey data, and indirectly, through the use of proxies such as political subdivisions and media markets. Political scientists also have successfully quantified district-community congruence, again utilizing the subdivision and market proxies. These techniques demonstrate that the territorial community test can be employed not just qualitatively but also with some social scientific rigor. (331)

An important effort to identify communities directly was recently completed by Dante Chinni and James Gimpel. (332) Chinni and Gimpel compiled extensive demographic and socioeconomic data for every county in America: population size, population density, income, occupation, education, race, ethnicity, age distribution, religion, etc. (333) They then used a statistical procedure to assign each county to one of twelve community types. (334) These community types, which capture much of the country's diversity, include fast-growing "Boom Towns," university-focused "Campus and Careers," deeply religious "Evangelical Epicenters," Hispanic-heavy "Immigration Nation," urban "Industrial Metropolis," struggling "Service Worker Centers," and agrarian "Tractor Country." (335) That counties can methodically be classified in this manner helps show that the task of community identification is tractable.

Equally promising is the CommonCensus Map Project, which asks respondents to provide their home addresses as well as the "local community," "local area," and major city with which they identify most closely. (336) Complex algorithms then convert the information into fascinating maps (available at local, regional, and national scales) of America's geographic communities, as specified and experienced by the people themselves. (337) While more data is necessary to create maps that are detailed enough for redistricting, (338) the idea of ascertaining community boundaries through people's own subjective affiliations is potentially very powerful.

Political scientists have also used less precise proxies to determine community borders: political subdivisions such as towns, (339) counties," (340) and school districts, (341) and media markets for television stations (342) and newspapers. (343) These entities have the advantage of being easily discernible; political subdivision boundaries can be seen on any map, while groups such as Nielsen Media Research and Arbitron define media markets in spatial terms. The entities have the disadvantage, of course, of not coinciding perfectly with territorial communities; as noted earlier, the patterns of people's interests and affiliations do not always match up with lines drawn for other reasons. (344) Still, both political subdivisions and media markets are useful starting points for any attempt to determine community borders, and it is unsurprising that courts conducting such an inquiry often have relied on--but not limited themselves to--them. (345)

Unfortunately, Chinni and Gimpel did not try to analyze how well electoral districts correspond to their twelve community types, nor has the CommonCensus Map Project compared constituencies to people's self-identified communities. Helpfully, however, many of the scholars who used political subdivisions or media markets as proxies for geographic communities also assessed district-subdivision or district-market congruence. Using a variety of statistical techniques, they managed to quantify the level of congruence of congressional districts across the country, typically on a scale from 0 (low) to 1 (high). (346)

That this sort of measurement can be done is quite significant. It means that a challenged district's congruence can be calculated reliably, and then compared to the congruence of other districts in the state (or nation). While political subdivisions and media markets are not identical to territorial communities, particularly incongruent districts are still more likely to disrupt communities, and particularly congruent districts are still more likely to respect them. At the very least, it is probative whether a district does or does not correspond to subdivisions or markets as well as most of its peers. Courts employing the territorial community test, then, could use comparative congruence data to bolster their qualitative analysis. The result, presumably, would be case outcomes that are more consistent and predictable--and doctrine that is more judicially manageable." (347)

E. Lingering Ambiguity

Of course, neither the courts' experiences nor the efforts of the political scientists can eliminate all uncertainty regarding the application of the territorial community test. No matter how many cases are decided under it, and no matter how sophisticated the social scientific measures become, judges will still need to make tough judgment calls about community boundaries and district-community congruence. This remaining imprecision, however, is not necessarily undesirable. If the Court's history with racial gerrymandering is any guide, some lingering ambiguity could actually prove useful by prompting district drawers to play it safe and avoid provocative district plans.

When the Court first announced in 1993 that districts are invalid if they are drawn for predominantly racial reasons, (348) the new standard was attacked at once for its vagueness. (349) Critics noted that legislative intent is very difficult to discern, and that it is more challenging still to determine whether race overwhelmed all other redistricting considerations. Early returns seemed to hear out these fears. The Court took case after case in the new doctrinal field, dividing bitterly almost every time and reaching outcomes that were not easy to reconcile with one another. By the end of the 1990s, as Justice Sourer put it, there was widespread "confusion in statehouses and courthouses" thanks to the absence of a "practical standard for distinguishing between the lawful and unlawful use of race." (350)

But something funny happened in the next redistricting cycle: the number of racial gerrymandering challenges plummeted, and almost every such challenge--including the only one considered by the Court--failed. (351) As Richard Pildes has explained, risk-averse district drawers, hoping to avoid litigation, largely stopped forming strangely shaped majority-minority districts. They "internalized Shaw, not as barring them from intentionally creating [majority-]minority districts, but as imposing general, extrinsic limits on the extent to which districts could be noncompact." (352) The aesthetics of majority-minority districts thus improved, litigation declined precipitously, and "vague law was transformed into settled practice." (353)

Though it is impossible to say for sure, a similar story could unfold if the territorial community test were adopted. At first, court decisions might be highly controversial, many districts might seem to be in jeopardy, and complaints about the test's unworkability might abound. But over time, as in the racial gerrymandering context, risk-averse district drawers would likely internalize the new rule and learn how to steer clear of litigation. They might form districts that adhere closely to political subdivisions, or assemble detailed data in order to identify community boundaries, or document in writing the reasons for their redistricting choices--all steps not strictly necessary for avoiding liability. In this manner, the territorial community test might end up both sparsely litigated and strictly enforced thanks to its residual ambiguity. And while heavy enforcement has its drawbacks in other areas, it is relatively unproblematic when the activity being curtailed (without excessive judicial involvement) is political gerrymandering. (354)


If the most common objection to the territorial community test is that it is judicially unmanageable, (355) claims that it would harm Democrats and racial minorities and reduce competitiveness come in a close second. (356) The assumptions underlying these criticisms are that Democratic and minority voters are especially spatially concentrated, and that most geographic communities are skewed in a particular party's favor. The formation of districts that coincide with communities, then, would allegedly "pack" Democrats and minorities and render most districts uncompetitive.

Arguments of this sort, focusing on statewide electoral consequences, have only limited bearing on whether the territorial community test should be adopted. As laid out in the preceding Parts, my case for the test is theoretical, historical, and doctrinal--but not overtly political. That the test may unintentionally help or hurt a party or racial group, or make elections more or less competitive, is largely irrelevant to its legal merit. Electoral impact, of course, is not a recognized modality of constitutional interpretation. (357)

Nevertheless, it is plainly good policy for district plans to treat parties and racial groups fairly and for election outcomes to be responsive to changes in public opinion. Accordingly, this Part considers the likely political implications of the territorial community test's adoption-and concludes that they are actually quite positive. I first investigate how partisan bias, electoral responsiveness, minority representation, and other key variables relate to district-community congruence. Contrary to critics' expectations, preliminary empirical analysis shows that bias is lower, responsiveness is higher, and minority representation is unchanged in states that respect community boundaries when they redistrict. I then examine the relevant political science literature, which largely confirms these findings and extends them in several interesting ways. Thus not only does the territorial community test not injure any party or racial group or reduce competitiveness, but, based on the available evidence, it appears to do the opposite.

A. Empirical Analysis

In the last redistricting cycle (2001-2010), almost half the states were required or encouraged to draft state legislative district plans that adhered to territorial community boundaries. (358) Of the twenty-two community-respect provisions that were in force, four were constitutional, seven were statutory, and eleven were included in non-binding guidelines. (359) Courts in six states had also issued decisions specifically addressing these provisions before the last cycle began. (360) In this Section, I analyze how states that paid heed to community boundaries over the last decade differed, along several important metrics, from states that did not. My results help rebut many of the politically rooted criticisms of the territorial community test.

It is true, of course, that states with community-respect provisions on their books do not necessarily abide by them. Districts might fail to correspond to underlying geographic communities even if correspondence is nominally urged or even mandated. (361) Still, the presence of a community-respect requirement is at least a decent proxy for actual district-community congruence. Particularly when such a requirement is legally binding (because it is included in a constitution or statute rather than a hortatory guideline), and even more so when it previously has been judicially enforced, one would expect it to have a discernible effect on how closely districts and communities in a state coincide. (362)

It is also true that the territorial community test does not directly target bias, responsiveness, or minority representation. The goal of the test is concededly to optimize district-community congruence, not to accomplish other laudable policy objectives. However, that the approach may influence the electoral system indirectly (by sharply limiting the discretion of district drawers (363)) and somewhat coincidentally (because districts that coincide with communities often happen to correlate with desirable statewide attributes) is no reason to discount these consequences. Positive side effects are still positive.

1. Bias and Responsiveness

Beginning with bias and responsiveness, political scientists Bruce Cain and John Hanley calculated both metrics for fifty legislative chambers in twenty-six states based on the results of the 2002 elections. (364) As noted earlier, partisan bias refers to the divergence in the share of seats that each party would win given the same share of the statewide vote. (365) For example, if Democrats would win 48% of the seats with 50% of the vote (in which case Republicans would win 52% of the seats), then a district plan would have a bias of 2%. Electoral responsiveness refers to the rate at which a party gains or loses seats given changes in its statewide vote share. For instance, if Democrats would win 10% more seats if they received 5% more of the vote, then a plan would have a responsiveness of 2.0. (366) In general, the lower a plan's bias, and the higher its responsiveness, the better the plan is. (367)

As Figures 1 and 2 illustrate, bias was markedly lower and responsiveness was markedly higher in the states that paid heed to territorial communities in the last redistricting cycle. Specifically, average bias was 4.0 percentage points lower (5.4% versus 9.4%), while average responsiveness was higher by 0.40 (1.43 versus 1.03). (368) All the subcategories of states with community-respect provisions--i.e., states with constitutional or statutory requirements, states with non-binding guidelines, and states with prior relevant court decisions--also scored better than the states with no such provisions. The differences in bias and responsiveness were statistically significant as well. (369) And while adherence to community boundaries did not remain a significant predictor of bias and responsiveness when I controlled for other common redistricting criteria (370) and aspects of states' redistricting environments, (371) the magnitude and direction of the variable's coefficients continued to be consistent with the above results. Controlling for all these factors, a shift from no community-respect provision at all to the most stringent (i.e., judicial) enforcement of such a provision was associated with a fall in bias of 3.3 percentage points and a rise in responsiveness of 0.26.

This analysis, while preliminary, tends to refute the claims that the territorial community test would disadvantage either major party or make elections less competitive. In fact, the test's implications for bias and responsiveness seem to be not just neutral but actually somewhat favorable. In the last redistricting cycle, states that respected community boundaries when they redrew their district maps treated the major parties more fairly, and held elections that were more responsive to changes in public opinion, than states that did not. These findings should help allay misgivings about the potential impact of district-community congruence on the American electoral system. Based on the available evidence, such congruence would likely make the system better--a bit less skewed in either party's favor and a bit more responsive to the views of the public--not worse. (372)

2. Minority Representation

Turning next to minority representation, I used Census data to estimate the number of majority-minority districts in each state legislative chamber during the last redistricting cycle. (373) I defined a majority-minority district as one in which any racial minority made up more than 50% of the population. (374) I then calculated the difference between each state's proportion of majority-minority districts (averaged over both its chambers) and its minority population percentage. (375) For example, if a state had 15% majority-minority districts and a minority population of 20%, then the gap for the state was 5%. The smaller this gap, the better representation minorities presumably received. (376)

As Figure 3 indicates, states that paid heed to community boundaries when they last redistricted had a slightly smaller difference between their proportions of majority-minority districts and their minority population percentages. Specifically, the gap in these states was 2.7 percentage points lower (10.8% versus 13.5%). (377) In other words, minorities appeared to receive slightly better legislative representation, relative to their share of the overall state population, in community-respecting states. The difference between the two categories of states was not large; in fact, it just failed to rise to the level of statistical significance. (378) Nor was any clear pattern discernible in the subcategories of states that enforced community-respect provisions more or less stringently. Still, these results offer no support for the claim that district-community congruence systematically reduces minority representation. During the last redistricting cycle, such congruence had either no link, or a slightly positive link, with states' relative proportions of majority-minority districts.

3. Voter Engagement

Though they are relatively favorable, the implications of the territorial community test for bias, responsiveness, and minority representation are somewhat fortuitous. As noted above, the test is not designed to make district plans less biased, more responsive, or better for minorities. That it appears to do so, relative to the status quo, is a product of the particular political and racial geography of contemporary America--as well as an illustration of how flawed the status quo is.

However, the territorial community test is aimed at making voters more invested in the political process and more likely to participate in it. An important prediction of the theory of communal representation (discussed above in Part I) is that relations between voters and elected officials, as well as voters' political engagement, should improve when districts and communities coincide. (379) To assess this prediction, I compiled data on voter turnout (380) and trust in government (381) for all available states during the last redistricting cycle. Turnout is the proportion of the voting-eligible population that casts a ballot in a given election. Trust in government is measured by survey on a scale from 0 to 100 points.

As Figures 4 and 5 show, both turnout and trust in government were substantially higher in the states that paid heed to community boundaries when they last redistricted. Specifically, average turnout was 4.9 percentage points higher (55.5% versus 50.6%), while average trust in government was 6.1 points higher (39.3 points versus 33.2 points). (382) All the subcategories of states with community-respect provisions also scored better than the states without them. The differences in turnout and trust in government were statistically significant as well. (383) And adherence to community boundaries remained a significant predictor of turnout and trust in government even when I controlled for other common redistricting criteria (384) and demographic variables. (385) In fact, controlling for all these factors, a shift from no community-respect provision to the most stringent enforcement of such a provision was associated with a turnout increase of 4.4 percentage points and a trust-in-government increase of 6.2 points.

This analysis indicates that, as predicted by the theory of communal representation, the territorial community test is linked to higher voter turnout and greater trust in government. This is a notable finding that suggests that voters indeed respond favorably to improved district-community congruence. Where districts and communities coincided more closely over the last decade, voters were both more likely to go to the polls and more trustful of their elected representatives. Democratic participation improved, in other words, while the usual electoral pathologies were somewhat less virulent. (386)

B. The Political Science Literature

The political science literature largely confirms the above findings and extends them in several interesting ways. The literature thus provides further evidence that the political implications of the territorial community test's adoption would likely be positive.

1. Bias and Responsiveness

To begin with, an array of studies, employing a variety of methods, suggest that district-community congruence would give rise to less biased and more responsive district plans than the status quo. First, Roland Fryer and Richard Holden recently redrew the congressional districts of four large states (California, New York, Pennsylvania, and Texas) using an algorithm that minimized the spatial distance between voters in each district. (387) This redrawing, which the authors analogized to reliance on territorial communities, (388) resulted in substantially lower bias and higher responsiveness scores. Average bias in the four states fell from 4.9% to 2.4%, while average responsiveness doubled from 0.9 to 1.8. (389) In other words, the revised (and more community-attentive) district plans were about twice as fair and responsive as the states' actual maps.

Michael McDonald carried out a similar study for five Midwestern states (Illinois, Michigan, Minnesota, Ohio, and Wisconsin), redrawing their districts using several different techniques. (390) He did not calculate bias or responsiveness for his new maps, but, when he minimized congressional districts' splits of Census places (i.e., political subdivisions), he found notable improvements in other metrics of fairness and competitiveness. In particular, the average proportion of Democratic-leaning districts increased from 40% to 49%, and the average proportion of competitive districts increased from 44% to 51%. (391) While not as reliable as bias and responsiveness calculations, these findings also illustrate the potential positive impact of district community congruence. (392)

So too does recent scholarship on the geographic distribution of the major U.S. parties' supporters, which suggests that the territorial community test would be unlikely to produce partisan imbalances. Democrats, it turns out, are not more spatially concentrated than Republicans. Rather, "[b]oth Republicans and Democrats live in counties where about fifty percent of the voters share their own party," and the "isolation index," a measure of partisan segregation, is actually slightly higher for Republicans than for Democrats. (393) Accordingly, neither party has much reason to fear that its voters would be particularly "packed" if districts were drawn to better coincide with geographic communities.

The capacity of the territorial community test to curb gerrymandering is further confirmed by the strong record of requirements that districts adhere to political subdivisions. One scholar found that highly partisan plans in the 1980s disregarded more subdivision boundaries than bipartisan or nonpartisan plans. (394) If district drawers had been obligated to follow subdivision lines, many heavily biased plans would not have been possible. Similarly, Jonathan Winburn determined that, in a range of states over the last decade, "the principle against splitting political subdivisions play[ed] a key function in constraining the re-mappers from gerrymandering." (395) Where such a provision existed, district plans were generally fair to the major parties; where it was absent, inequitable plans could be (and usually were) implemented. (396) The same would likely be true for a district-community congruence requirement.

Lastly, a number of studies have established that challengers have better odds of success in congressional districts that coincide with political subdivisions or media markets. The explanation for the improvement is that challengers are better able to convey their messages (and names) to the public when districts have a cognizable identity and media channels are efficiently structured. (397) The boost is also surprisingly large. One study estimated that challengers receive 8% more of the vote in high-congruence distracts, (398) while another study found that a challenger is 16% more likely to win a voter's support if the voter can recall the challenger's name. (399) District-community congruence thus not only correlates with, but also helps foster, increased electoral competitiveness.

2. Minority Representation

There is less political science literature on the territorial community test's effect on minority representation, but the available studies suggest that minorities (just like Democrats) would not end up inefficiently overconcentrated. One study examined the relationship between political subdivision preservation requirements and the numbers of majority-minority and minority-influence districts within states. The study found that such requirements have no impact on the number of majority-minority districts in a state, but increase the number of minority-influence districts. (400) Similarly, another study analyzed how often districts with varying proportions of minority voters cross county lines. The study determined that districts with the highest percentages of African American and Hispanic voters split counties more frequently than all other districts. (401) As long as political subdivisions are a decent proxy for geographic communities, the upshot of these studies is that higher district-community congruence would likely result in more minority-influence districts, about the same number of majority-minority districts, and fewer supermajority-minority (i.e., "packed") districts. This is not an outcome that should alarm proponents of minority representation.

3. Voter Engagement

Finally, the political science literature indicates that both voter knowledge and voter turnout tend to increase as districts and communities coincide more closely. With regard to voter knowledge, a series of studies have shown that voters are better informed about politics when they live in high-congruence districts. One study found that voters in districts that correspond well to media markets are 8% more likely to recognize incumbent politicians' names, and 19% more likely to recognize challengers' names. (402) An analogous study determined that voters in districts that mirror political subdivisions are 8% more likely to recall incumbents' names, and 12% more likely to recall challengers' names. (403) Taking a slightly different tack, another study concluded that voters in high-congruence districts are 14% more likely to state correctly their representatives' votes in Congress than voters in low-congruence districts. (404)

The story with turnout is similar (though not quite as clear-cut). One study found that district congruence with media markets is strongly linked to higher turnout, but that district-county congruence has no effect on it. (405) Another study reported that neither media market nor county congruence has a statistically significant impact on turnout. (406) A further study determined that districts that are more compact tend to have higher turnout. (407) A Canadian study, lastly, concluded that, after Ontario's districts were redrawn in the 1980s, turnout rose in the districts that corresponded best to geographic communities, and fell in the districts that corresponded worst. (408) These results are therefore inconclusive, but they do at least hint at a positive relationship between district-community congruence and voter turnout.


As the current redistricting cycle continues to unfold, America's courts find themselves in a deeply problematic position. They are as by a unanimous Supreme Court that political gerrymandering can sometimes be unconstitutional--but majorities of that same Court have also rejected every standard suggested to date for distinguishing valid from invalid district plans. This Article has sought to offer a way out of this judicial limbo: the territorial community test, under which courts would assess electoral districts based on how well they correspond to underlying geographic communities. As the Article has argued, the test has a robust theoretical and historical pedigree, it already animates much of the redistricting case law, it could be administered capably by the judiciary, and its political implications would be quite positive.

It is true that the territorial community test does not aim directly at the heart of what is typically understood today to be the problem with political gerrymandering: the deliberate manipulation of district lines in order to help or harm particular parties or candidates. But the test does attack another aspect of gerrymandering that has been bemoaned since the days of Elbridge Gerry: the lack of congruence between gerrymandered districts and organic geographic communities. The test also does (even though it is not designed to) make elections fairer for both major parties and more responsive to changes in public opinion. But perhaps the most important point is this: the Supreme Court has already ruled out just about every standard that would explicitly tackle partisan unfairness or incumbent entrenchment. At this point, more oblique measures are all that are left. And of these judicial bank shots, there seems to be none more promising than the territorial community test. Second-best may now be the best we can do.

Table 1. Partisan Bias Regressions

                                       Model 1:        Model 2: Other
                                   Community Respect   Redistricting
Variables                                Alone            Criteria

Community Respect                  -0.02 (0.01) **     -0.01 (0.01)
Compactness                               --           -0.04 (0.03)
Subdivision Preservation                  --            0.02 (0.03)
Prior District Core Preservation          --            0.001 (0.03)
Redistricting Commission Use              --                 --
Unified Party Control                     --                 --
Court-Drawn Districts                     --                 --
Constant                            0.09 (0.01) ***     0.10 (0.03)
N                                         26                  26
Adjusted R-Squared                       0.14                0.14

                                      Model 3:
                                    Redistricting       Model 4:
Variables                            Environment           All

Community Respect                  -0.01 (0.01)      -0.01 (0.01)
Compactness                              --          -0.02 (0.03)
Subdivision Preservation                 --           0.03 (0.03)
Prior District Core Preservation         --          -0.02 (0.03)
Redistricting Commission Use       -0.06 (0.02) **   -0.06 (0.03) **
Unified Party Control              -0.01 (0.02)      -0.008 (0.02)
Court-Drawn Districts              -0.04 (0.03)      -0.04 (0.04)
Constant                            0.12 (0.02)       0.11 (0.03) **
N                                       26                26
Adjusted R-Squared                     0.22              0.16

Entries for variables take form: coefficient (standard error).

N = 26 states for which data is available

* P< 0.10

** P< 0.05

*** P< 0.01

Table 2. Electoral Responsiveness Regressions

                                   Model 1:         Model 2: Other
                               Community Respect    Redistricting
Variables                            Alone             Criteria

Community Respect               0.18 (0.08)         0.08 (0.10)
Compactness                           --            0.31 (0.21)
Subdivision Preservation              --            0.18 (0.27)
Prior District Core                   --           -0.29 (0.21)
Redistricting Commission Use          --                  --
Unified Party Control                 --                  --
Court-Drawn Districts                 --                  --
Constant                        1.06 (0.11) ***     0.84 (0.25) ***
                                      26                  26
Adjusted R-Squared                    0.13               0.18

                                   Model 3:          Model 4:
                                Redistricting           All
Variables                        Environment

Community Respect               0.11 (0.09)        0.09 (0.10)
Compactness                           --           0.15 (0.25)
Subdivision Preservation              --           0.13 (0.28)
Prior District Core                   --          -0.40 (0.25)
Redistricting Commission Use    0.33 (0.20)        0.12 (0.26)
Unified Party Control           0.17 (0.17)        0.19 (0.18)
Court-Drawn Districts           0.50 (0.27)        0.46 (0.32)
Constant                        0.80 (0.17) ***    0.79 (0.26) **
                                      26                26
Adjusted R-Squared                   0.19               0.18

Entries for variables take form: coefficient (standard error).

N = 26 states for which data is available

* P< 0.10

** P< 0.05

*** P< 0.01

Table 3. Voter Turnout Regressions

                               Model 1:         Model 2: Other
                           Community Respect    Redistricting
Variables                        Alone             Criteria

Community Respect           0.03 (0.01) ***     0.02 (0.01) ***
Compactness                       --            0.003 (0.02)
Subdivision Preservation          --            0.009 (0.03)
Prior District Core               --           -0.04 (0.02) *
Household Income                  --                  --
% Population over 65              --                  --
% Population Black                --                  --
% Population Hispanic             --                  --
% Population with                 --                  --
  Bachelor's Degree
Constant                    0.51 (0.01) ***     0.51 (0.02) ***
N                                 50                  50
Adjusted R-Squared               0.19                0.20

                                Model 3:              Model 4:
                              Socioeconomic              All
Variables                         Data

Community Respect            0.01 (0.007) *          0.02 (0.01) *
Compactness                        --               -0.002 (0.02)
Subdivision Preservation           --                0.009 (0.02)
Prior District Core                --                -0.01 (0.02)
Household Income           0.000001 (0.000002)   0.000001 (0.000002)
% Population over 65           0.44 (0.48)           0.44 (0.49)
% Population Black           -0.17 (0.09) **        -0.14 (0.10)
% Population Hispanic       -0.20 (0.07) ***        -0.20 (0.08) **
% Population with              0.32 (0.26)           0.32 (0.27)
  Bachelor's Degree
Constant                     0.35 (0.10) ***       0.35 (0.11) ***
N                                  50                    50
Adjusted R-Squared                0.43                  0.40

Entries for variables take form: coefficient (standard error).

N= all 50 states

* P< 0.10

** P< 0.05

*** P< 0.01

Table 4. Trust in Government Regressions

                                   Model 1:         Model 2: Other
                               Community Respect    Redistricting
Variables                            Alone             Criteria

Community Respect               2.67 (0.94) ***     2.41 (1.03) **
Compactness                           --            0.49 (2.13)
Subdivision Preservation              --            2.28 (3.00)
Prior District Core                   --           -2.79 (2.16)
Household Income                      --                  --
% Population over 65                  --                  --
% Population Black                    --                  --
% Population Hispanic                 --                  --
% Population with Bachelor's          --                  --
Constant                       33.86 (1.14) ***    32.50 (2.81) ***
N                                     47                  47
Adjusted R-Squared                   0.13                0.12

                                   Model 3:
                                 Socioeconomic         Model 4:
Variables                            Data                 All

Community Respect                2.13 (1.03) *       2.07 (1.16) *
Compactness                           --            -0.42 (2.00)
Subdivision Preservation              --             1.90 (2.80)
Prior District Core                   --            -0.60 (2.41)
Household Income                 0.0002 (0.0003)     0.0002 (0.0003)
% Population over 65            57.56 (66.30)       60.14 (68.71)
% Population Black             -16.17 (10.93)      -15.04 (13.12)
% Population Hispanic          -23.92 (9.39) **    -23.55 (9.74) **
% Population with Bachelor's   -16.14 (37.55)      -16.60 (39.13)
Constant                        23.75 (13.66) *      21.79 (14.45)
N                                     47                  47
Adjusted R-Squared                   0.29                0.24

Entries for variables take form: coefficient (standard error).

N=47 states for which data is available

* P< 0.10

** P< 0.05

*** P< 0.01

(1) For more on the ongoing litigation over electoral district lines since the 2010 Census, see Justin Levitt, Litigation in the 2010 Cycle, ALL ABOUT REDISTRICTING, (last visited Feb. 15, 2012).

(2) Political gerrymandering is usually defined as the drawing of electoral districts in order (1) to advance one major party's interests at the other's expense or (2) to protect the incumbents of both parties. Vieth v. Jubelirer, 541 U.S. 267, 271 n.1 (2004) (plurality opinion) (citing BLACK'S LAW DICTIONARY 696 (7th ed. 1999)). While not quarreling with this definition, this Article also conceives of political gerrymandering in a third way: as the drawing of district lines in such a way that organic territorial communities are disrupted. Cf. Whitcomb v. Chavis, 403 U.S. 124, 177 (1971) (Douglas, J., dissenting in part and concurring in the result in part) ("The problem of the gerrymander is how to defeat or circumvent the sentiments of the community.").

(3) 478 U.S. 109, 113, 125 (1986) (plurality opinion).

(4) Id. at 132.

(5) See, e.g., Samuel Issacharoff, Judging Politics: The Elusive Quest for Judicial Review of Political Fairness, 71 TEX. L. REV. 1643, 1684 (1993) ("Bandemer is a mass of confusion on what the Court actually believes is the constitutional harm."); Daniel Hays Lowenstein, Bandemer's Gap: Gerrymandering and Equal Protection, in POLITICAL GERRYMANDERING AND THE COURTS 64, 74-77 (Bernard Grofman ed., 1990).

(6) See Vieth, 541 U.S. at 282 (plurality opinion) ("In the lower courts, the legacy of the plurality's test [in Bandemer] is one long record of puzzlement and consternation.").

(7) See id. at 305-06.

(8) See id. at 339 (Stevens, J., dissenting) (proposing inquiry focused on partisan in tent); id. at 346-52 (Souter, J., dissenting) (proposing burden-shifting approach imported from employment discrimination context); id. at 360-65 (Breyer, J., dissenting) (emphasizing risk of "unjustified entrenchment" by political minority).

(9) See id. at 316-17 (Kennedy, J., concurring in the judgment). In the subsequent case of League of United Latin American Citizens (LULAC) v. Perry, the Court explicitly declined to answer any of the questions left hanging by Vieth. See 548 U.S. 399, 414 (2006) ("We do not revisit [Vieth's] justiciability holding...."). Unsurprisingly, the scholarly response to Vieth has been sharply negative. See, e.g., Richard Briffault, Defining the Constitutional Question in Partisan Gerrymandering, 14 CORNELL J.L. & PUB. POL'Y 397, 399 (2005); James A. Gardner, A Post-Vieth Strategy for Litigating Partisan Gerrymandering Claims, 3 ELECTION L.J. 643, 643 (2004) (noting that "It]he result [in Vieth] was disappointing" and "left the law a shambles"); Heather K. Gerken, Lost in the Political Ticket: The Court, Election Law, and the Doctrinal Interregnum, 153 U. PA. L. REV. 503, 528 (2004) ("[T]he Court at this moment seems to be at sea when deciding election law cases."); Richard L. Hasen, Looking for Standards (in All the Wrong Places): Partisan Gerrymandering Claims After Vieth, 3 ELECTION L.J. 626, 627-28, 642 (2004); Daniel H. Lowenstein, Vieth's Gap: Has the Supreme Court Gone from Bad to Worse on Partisan Gerrymandering?, 14 CORNELL J.L. & PUB. POL'Y 367, 368, 393-94 (2005).

(10) See, e.g., Samuel Issacharoff, Gerrymandering and Political Cartels, 116 HARV. L. REV. 593, 615-17 (2002); Samuel Issacharoff & Pamela S. Karlan, Where To Draw the Line?: Judicial Review of Political Gerrymanders, 153 U. PA. L. REV. 541, 570-74 (2004); Samuel Issacharoff & Richard H. Pildes, Politics as Markets: Partisan Lockups of the Democratic Process, 50 STAN. L. REV. 643, 680-81 (1998); Richard H. Pildes, The Constitution and Political Competition, 30 NOVA L. REV. 253, 262-63 (2006).

(11) See, e.g., Andrew Gelman & Gary King, A Unitized Method of Evaluating Electoral Systems and Redistricting Plans, 38 AM. J. POL. SCI. 514, 517-21 (1994); Bernard Grofman & Gary King, The Future of Partisan Symmetry as a Judicial Test for Partisan Gerrymandering After LULAC v. Perry, 6 ELECTION L.J. 2, 6-9 (2007); Richard G. Niemi, The Relationship Between Votes and Seats: The Ultimate Question in Political Gerrymandering, 33 UCLA L. REV. 185, 195-200 (1985).

(12) See, e.g., Mitchell N. Berman, Managing Gerrymandering, 83 TEX. L. REV. 781, 825-28 (2005); Briffault, supra note 9, at 416.

(13) See, e.g., Richard G. Niemi et al., Measuring Compactness and the Role of a Compactness Standard in a Test for Partisan and Racial Gerrymandering, 52 J. POL. 1155, 1176-78 (1990); Daniel D. Polsby & Robert D. Popper, The Third Criterion: Compactness as a Procedural Safeguard Against Partisan Gerrymandering, 9 YALE L. & POL'Y REV. 301, 302, 326-51 (1991); cf. Richard H. Pildes & Richard G. Niemi, Expressive Harms, "Bizarre Districts," and Voting Rights: Evaluating Election-District Appearances After Shaw v. Reno, 92 MICH. L. REV. 483, 536 (1993) (analyzing racial gerrymandering doctrine in compactness terms).

(14) See, e.g., Guy-Uriel E. Charles, Democracy and Distortion, 92 CORNELL L. REV. 601, 607-16 (2007).

(15) See, e.g., Gordon E. Baker, The "Totality of Circumstances" Approach, in POLITICAL GERRYMANDERING AND THE COURTS, supra note 6, at 203, 205-10; Bernard Grofman, Criteria for Districting: A Social Science Perspective, 33 UCLA L. REV. 77, 171-72 (1985).

(16) See, e.g., Larry Alexander, Lost in the Political Thicket, 41 FLA. L. REV. 563, 577-78 (1989); Bruce E. Cain, Simple vs. Complex Criteria for Partisan Gerrymandering: A Comment on Niemi and Grofman, 33 UCLA L. REV. 213, 225-26 (1985); Daniel H. Lowenstein & Jonathan Steinberg, The Quest far Legislative Districting in the Public Interest: Elusive or Illusory?, 33 UCLA L. REV. 1, 73-75 (1985); Nathaniel Persily, Reply, In Defense of Foxes Guarding Henhouses: The Case for Judicial Acquiescence to Incumbent-Protecting Gerrymanders, 116 HARV. L. REV. 649, 679-81 (2002); Peter H. Schuck, The Thickest Thicket: Partisan Gerrymandering and Judicial Regulation of Politics, 87 COLUM. L. REV. 1325, 1330, 1384 (1987).

(17) See Vieth v. Jubelirer, 541 U.S. 267, 292-301 (2004) (plurality opinion); id. at 308 (Kennedy, J., concurring in the judgment) (agreeing with plurality's criticisms of dissenters' proposed standards); see also Davis v. Bandemer, 478 U.S. 109, 138-43 (1986) (plurality opinion) (rejecting Justice Powell's totality-of-circumstances approach).

(18) Vieth, 541 U.S. at 350 n.5 (Souter, J., dissenting).

(19) See LULAC, 548 U.S. 399, 419-20 (2006) (opinion of Kennedy, J.).

(20) See Nicholas Stephanopoulos, Reforming Redistricting: Why Popular Initiatives to Establish Redistricting Commissions Succeed Fail, 23 J.L. & POL. 331, 332-37 (2007).

(21) See Briffault, supra note 9, at 418-21; Hasen, supra note 9, at 637-39; Richard H. Pildes, The Supreme Court, 2003 Term--Foreword: The Constitutionalization of Democratic Politics, 118 HARV. L. REV. 28, 66 (2004).

(22) See, e.g., JEREMY BUCHMAN, DRAWING LINES IN QUICKSAND: COURTS, LEGISLA TURES, & REDISTRICTING 204 (2003); BRUCE E. CAIN, THE REAPPORTIONMENT PUZZLE 63-66 (1984); Dean Alfange, Jr., Gerrymandering and the Constitution: Into the Thorns of the Thicket at Last, 1986 SUP. CT. REV. 175, 215-16; Charles Backstrom et al., Establishing a Statewide Electoral Effects Baseline, in POLITICAL GERRYMANDERING AND THE COURTS, supra note 5, at 145, 153; Pildes & Niemi, supra note 13, at 577.

(23) See, e.g., John Hart Ely, Gerrymanders: The Good, the Bad, and the Ugly, 50 STAN. L. REV. 607, 616 (1998); Richard L. Engstrom, The Post-2000 Round of Redistricting: An Entangled Thicket Within the Federal System, PUBLIUS, Fall 2002, at 51, 66-67; Lowenstein & Steinberg, supra note 16, at 32-33; Nathaniel Persily, When Judges Carve Democracies: A Primer on Court-Drawn Redistricting Plans, 73 GEO. WASH. L. REV. 1131, 1160-61 (2005).

(24) Ely, supra note 23, at 616 (emphasis omitted).

(25) See Grofman, supra note 15, at 90; Bernard Grofman, Would Vince Lombardi Have Been Right if He Had Said: "When It Comes to Redistricting, Race Isn't Everything, It's the Only Thing"?, 14 CARDOZO L. REV. 1237, 1262-63 (1993).

(26) See Henry L. Chambers, Jr., Enclave Districting, 8 WM. & MARY BILL RTS. J. 135, 176-82 (1999).

(27) See James A. Gardner, Foreword: Representation Without Party: Lessons from State Constitutional Attempts To Control Gerrymandering, 37 RUTGERS L.J. 881, 894-97 (2006) [hereinafter Gardner, Representation Without Party]; James A. Gardner, One Person, One Vote and the Possibility of Political Community, 80 N.C. L. REV. 1237, 1241-46 (2002) [herein after Gardner, Possibility of Community]; James A. Gardner, What Is "Fair" Partisan Representation, and How Can It Be Constitutionalized? The Case for a Return to Fixed Election Districts, 90 MARQ. L. REV. 555, 578-82 (2007) [hereinafter Gardner, Fixed Election Districts].

(28) Thornburg v. Gingles, 478 U.S. 30, 50-51 (1986).

(29) LULAC, 548 U.S. 399, 441 (2006).

(30) See supra notes 22-24 and accompanying text.

(31) Vieth v. Jubelirer, 541 U.S. 267, 281 (2004) (plurality opinion).

(32) In a sequel to this Article, I use newly available Census data as well as specialized mapping software to identify geographic communities directly. See Nicholas O. Stephanopoulos, Spatial Diversity, 125 HARV. L. REV. (forthcoming 2012).

(33) Partisan bias refers to the divergence in the share of seats that each party would win given the same share of the statewide vote. Electoral responsiveness refers to the rate at which a party gains or loses seats given changes in its statewide vote share. See infra notes 76, 365-67 and accompanying text. A sequel to this Article also finds that the political consequences of tighter district-community congruence are positive. See Stephanopoulos, supra note 32 (manuscript at 31-35, 48-51).

(34) See Daniel R. Ortiz, Got Theory?, 153 U. PA. L. REV. 459, 475-76 (2004) (noting that redistricting "raises issues about the purpose and function of the [legislature], the value of political competition, and how voters should be represented").

(35) See Richard Briffault, Lani Guinier and the Dilemmas of American Democracy, 95 COLUM. L. REV. 418, 421 (1995) (reviewing LANI GUINIER, THE TYRANNY OF THE MAJORITY: FUNDAMENTAL FAIRNESS IN REPRESENTATIVE DEMOCRACY (1994)) (arguing, in election law context, that "[t]he theoretical and the lawyerly are closely intertwined").

(36) As is evident from the many sources cited in this Section, this theory is not new. But even though it is well-established, it is rarely set forth at length, and it is not associated with any specific thinker. Accordingly, there is value in clearly presenting the theory's central claims and consequences.

(37) See, e.g., NANCY L. SCHWARTZ, THE BLUE GUITAR: POLITICAL REPRESENTATION AND COMMUNITY 54 (1988); Alfange, supra note 22, at 216 ("Voters do identify with the place in which they live, and do have a feeling of sharing concerns with others who live in that place...."); Charles H. Backstrom, Problems of Implementing Redistricting ("People think of themselves as belonging together in counties in rural areas, in cities or sectors of metropolitan areas, and in neighborhoods of central cities."), in REPRESENTATION AND REDISTRICTING ISSUES 43, 47 (Bernard Grofman et al. eds., 1982); Harold M. Proshansky et al., Place Identity: Physical World Socialization of the Self ("Individuals do indeed define who and what they are in terms of strong affective ties to ... neighborhood and community."), in GIVING PLACES MEANING 87, 91 (Linda Groat ed., 1995).

(38) See, e.g., Richard Briffault, Race and Representation After Miller v. Johnson, 1995 U. CHI. LEGAL F. 23, 41 ("Many of the most important interests and concerns people have relate to ... their immediate geographic environment."); Chambers, supra note 26, at 163 ("[P]eople who live in close proximity share similar political interests...."); Benjamin Forest, Mapping Democracy: Racial Identity and the Quandary of Political Representation, 91 ANNALS ASS'N AM. GEOGRAPHERS 143, 157 (2001) ("[P]ropinquity is important because political interests, preferences, and identity ... are formed and defined by local geographic communities....").

(39) Gardner, Representation Without Party, supra note 27, at 950.

(40) Lucas v. Forty-Fourth Gen. Assembly, 377 U.S. 713, 750 (1964) (Stewart, J., dissenting).

(41) Id.; see also Whitcomb v. Chavis, 403 U.S. 124, 159 (1971) (discussing the "general preference for legislatures reflecting community interests as closely as possible"); Reynolds v. Sims, 377 U.S. 533, 623-24 (1964) (Harlan, J., dissenting) ("[L]egislators can represent their electors only by speaking for their interests ... many of which do reflect the place where the electors live."); Briffault, supra note 35, at 431 ("[P]lace-based interests are the ones that ought to be guaranteed representation in the legislature."); Gardner, Fixed Election Districts, supra note 27, at 574 ("[T]he interests that ought to be represented in the legislature are those that are held in common by people living in a particular place."); C.O. Sauer, Geography and the Gerrymander, 12 AM. POL. SCI. REV. 403, 404-05 (1918).

(42) Grofman, supra note 25, at 1262-63.

(43) See, e.g., Davis v. Bandemer, 478 U.S. 109, 173 n.13 (1986) (Powell, J., concurring in part and dissenting in part) ("Confusion inevitably follows ... when a citizen finds himself or herself forced to associate with several artificial communities...."); BUCHMAN, supra note 22, at 203 ("Plans that do not split [community] boundaries ostensibly help voters identify their representatives, and vice versa."); Martin Shapiro, Gerrymandering, Unfairness, and the Supreme Court, 33 UCLA L. REV. 227, 243 (1985).

(44) See, e.g., CAIN, supra note 22, at 64 ("The argument that districts should not divide communities ... is in effect a plea for homogeneous districts."); RICHARD F. FENNO, JR., HOME STYLE: HOUSE MEMBERS IN THEIR DISTRICTS 6 (1978) ("[T]he most homogeneous districts ... incorporate pre-existing communities of interest."); Richard Morrill, A Geographer's Perspective, in POLITICAL GERRYMANDERING AND THE COURTS, supra note 5, at 212, 216.

(45) See SCHWARTZ, supra note 37, at 54 ("[C]ommunity ... need not necessarily[] mean 'demographic homogeneity,' sameness of social situation.").

(46) THE FEDERALIST NO. 56, at 346 (James Madison) (Clinton Rossiter ed., 1961); see also MALCOLM E. JEWELL, REPRESENTATION IN STATE LEGISLATURES 115 (1982) ("In a district that is basically homogeneous ... the task of representation is relatively easy."); Chambers, supra note 26, at 149 ("A representative can most easily advance the interests of her constituency when those interests are clearly or narrowly defined.").

(47) See, e.g., Karcher v. Daggett, 462 U.S. 725, 787 (1983) (Powell, J., dissenting) ("A legislator cannot represent his constituents properly ... when a voting district is nothing more than an artificial unit...."); Prosser v. Elections Bd., 793 F. Supp. 859, 863 (W.D. Wis. 1992) ("To be an effective representative, a legislator must represent a district that has a reasonable homogeneity of needs and interests...."); CAIN, supra note 22, at 63-64; Chambers, supra note 26, at 149; Morrill, supra note 44, at 216.

(48) According to the traditional delegate-trustee dichotomy, representatives who are delegates abide by the expressed preferences of their constituents, while representatives who are trustees make their own autonomous policy decisions. See generally HANNA FENICHEL PITKIN, THE CONCEPT OF REPRESENTATION 144-67 (1967).

(49) See, e.g., DAVID BUTLER & BRUCE CAIN, CONGRESSIONAL REDISTRICTING: COM PARATIVE AND THEORETICAL PERSPECTIVES 71 (1992) ("[A] communities of interest approach necessarily implies a delegate theory of representation."); ROYCE HANSON, THE POLITICAL THICKET: REAPORTIONMENT AND CONSTITUTIONAL DEMOCRACY 130-31 (1966); SCHWARTZ, supra note 37, at 135; Gardner, Representation Without Party, supra note 27, at 955-60.

(50) Heather Gerken has labeled this pattern of intra-district homogeneity and interdistrict heterogeneity as "second-order diversity." Heather K. Gerken, Second-Order Diversity, 118 HARV. L. REV. 1099, 1107-08 (2005). Of course, even under this approach, only views that are held by sufficiently large and geographically defined groups of people will be represented in the legislature.

(51) Letter from John Adams to John Penn (Jan. 1776), in THE POLITICAL WRITINGS OF JOHN ADAMS 491, 493 (George W. Carey ed., 2000); see also Lucas v. Forty-Fourth Gen. Assembly, 377 U.S. 713, 749 (1964) (Stewart, J, dissenting) (arguing for "effective representation in the State's legislature ... of the various groups and interests making up the electorate"); J.R. POLE, POLITICAL REPRESENTATION IN ENGLAND AND THE ORIGINS OF THE AMERICAN REPUBLIC 184 (1966) ("[T]he representative Assembly must be merely a small-scale replica of the whole people, drawn directly from it and reproducing it without the slightest distortion."); Phil C. Neal, Baker v. Carr: Politics in Search of Law, 1962 SUP. CT. REV. 252, 277 ("[T]he general interest, and the innumerable separate interests of which it is composed, will be better expressed in a medley of voices from minor fractions of the population than by any monolithic majority.").

(52) See, e.g., EDMUND S. MORGAN, INVENTING THE PEOPLE: THE RISE OF POPULAR SOVEREIGNTY IN ENGLAND AND AMERICA 41-43 (1988); ANNE PHILLIPS, ENGENDERING DEMOCRACY 63 (1991) (discussing "near universal practice of electing representatives according to geographical constituencies"); ANDREW REHFELD, THE CONCEPT OF CONSTITUENCY: POLITICAL REPRESENTATION, DEMOCRATIC LEGITIMACY, AND INSTITUTIONAL DESIGN 56 (2005) ("[T]erritorial districting in the United States is such a habit of mind that it is not seriously challenged...."); Lani Guinier, Groups, Representation, and Race-Conscious Districting: A Case of the Emperor's Clothes, 71 TEX. L. REV. 1589, 1603-05 (1993); Pildes & Niemi, supra note 13, at 483 (noting "long-standing Anglo-American commitment to organizing political representation around geography").

(53) See Polsby & Popper, supra note 13, at 307 n.29 ("[T]he Founders valued the faction-diluting character of representation by place.... [T]he authors of the Constitution deliberately chose geographical representation."); James Thomas Tucker, Redefining American Democracy: Do Alternative Voting Systems Capture the True Meaning of "Representation"?, 7 MICH. J. RACE & L. 357, 369 (2002) ("[T]he Constitutional Framers intended that Representatives be elected from [geographic] districts.").

(54) Schuck, supra note 16, at 1360.

(55) I agree with many of the critiques of geographic districting: that people's interests are often non-territorial, that it can produce unfair electoral outcomes (especially in combination with plurality voting and single-member districts), that other systems are better at reflecting the public's views, etc. In this Article, however, I take geographic districting as a given. My goal here is only to explore how political gerrymandering can be curbed in a manner consistent with the practice and underlying theory of geographic districting. The debate over geographic districting itself I leave for another day.

(56) ELAINE SPITZ, MAJORITY RULE 52 (1984); see also BUCHMAN, supra note 22, at 197 (discussing "the assumed linkage between geography and politically salient interests that conceptually underpins a reliance on district-based elections"); Gardner, Fixed Election Districts, supra note 27, at 584 ("[T]erritorial representation ... must rest on the fact that the residents of represented territories comprise distinct and coherent local communities...."); Guinier, supra note 52, at 1610-12; Timothy G. O'Rourke, Shaw v. Rent: The Shape of Things to Come, 26 RUTGERS LJ. 723, 770 (1995) ("[T]he commitment to geographic districting requires ... the creation of districts that ... constitute identifiable constituencies.").

I can think of one value other than adherence to the territorial community that geographic districting might plausibly aim to capture: administrability. But while geographically demarcated districts are indeed easily administrable, so too are non-spatial districts, particularly given the rise of modern information technology.

(57) See, e.g., BUCHMAN, supra note 22, at 160; MELISSA S. WILLIAMS, VOICE, TRUST, AND MEMORY: MARGINALIZED GROUPS AND THE FAILINGS OF LIBERAL REPRESENTATION 73 (1998); T. Alexander Aleinikoff & Samuel Issacharoff, Race and Redistricting: Drawing Constitutional Lines After Shaw v. Reno, 92 MICH. L. REV. 588, 637 (1993); Gardner, Possibility of Community, supra note 27, at 1261-62.

(58) POCHARD L. MORRILL, POLITICAL REDISTRICTING AND GEOGRAPHIC THEORY 63 (1981); see also BILL BISHOP, THE BIG SORT: WHY THE CLUSTERING OF LIKE-MINDED AMERICA IS TEARING US APART 5 (2008) (arguing that Americans now "cluster[] in communities of sameness, among people with similar ways of life, beliefs, and, in the end, politics"); Richard T. Ford, Law's Territory (A History of Jurisdiction), 97 MICH. L. REV. 843, 914 (1999) ("Even in the highly urban, industrialized and culturally polyglot societies of the Western capitalist democracies, the assertion of organic cultural community is appealing."); David B. Knight, Identity and Territory: Geographical Perspectives on Nationalism and Regionalism, 72 ANNALS ASS'N AM. GEOGRAPHERS 514, 526 (1982).

(59) See Briffault, supra note 9, at 400-02 (listing possible gerrymandering standards); Hasen, supra note 9, at 637-39 (same); Pildes, supra note 21, at 66-74 (same).

(60) The regularity of a district's perimeter and the dispersion of a district's area are two well-established measures of compactness. See BUCHMAN, supra note 22, at 200; Pildes & Niemi, supra note 13, at 554-56.

(61) Robert G. Dixon, Jr., Fair Criteria and Procedures for Establishing Legislative Districts, in REPRESENTATION AND REDISTRICTING ISSUES, supra note 37, at 7, 16; see also Shaw v. Hunt (Shaw II), 517 U.S. 899, 936 n.13 (1996) (Stevens, J., dissenting) (positing that voters' "interest in being in a district whose members share similar interests and concerns ... often is not[] vindicated by drawing districts with attractive shapes"); MARK MONMONIER, BUSHMANDERS AND BULLWINKLES: HOW POLITICIANS MANIPULATE ELECTRONIC MAPS AND CENSUS DATA TO WIN ELECTIONS 71 (2001) ; Grofman, supra note 15, at 89-90 ("[T]he usefulness of requiring that districts be compact has been vastly overrated."); Issacharoff, supra note 5, at 1693 (noting "absence of an independent normative foundation for a compactness requirement").

(62) See, e.g., BUCHMAN, supra note 22, at 198; BUTLER & CAIN, supra note 49, at 72-73; Pildes & Niemi, supra note 13, at 538 n.177.

(63) See Persily, supra note 23, at 1158 ("One could draw compact districts that group unrelated communities on different sides of a mountain or river...."). Conversely, noncompact districts may sometimes correspond to territorial communities and be cognizable to voters. See Briffault, supra note 38, at 44 n.104; Grofman, supra note 25, at 1263.

(64) See Grofman, supra note 15, at 92 (arguing that compactness is "a useful criterion only to the extent that it happens to coincide with other features"); Pildes & Niemi, supra note 13, at 538 ("[C]ompactness might be associated with relevant substantive districting values, like preserving communities that shared common political interests, but ... compactness [is] a poor proxy for those values."). The other common argument for compactness is instrumental: imposing an additional constraint on linedrawers makes it more difficult for them to pursue partisan advantage or incumbent protection. See, e.g., Polsby & Popper, supra note 13, at 332 ("[W]here compactness is a constraint, a gerrymanderer's job is noticeably harder."). This argument plainly is not based on any theory of representation. Moreover, there is reason to doubt that compactness is a meaningful constraint for redistricters, see Lowenstein & Steinberg, supra note 16, at 22-23, and similar indirect limitations can be imposed in other ways--such as through adherence to territorial communities.

(65) See supra note 10 and accompanying text.

(66) See, e.g., Issacharoff, supra note 10, at 620 (arguing that bipartisan gerrymanders should be presumptively invalid); Pildes, supra note 10, at 271-76.

(67) Issacharoff & Pildes, supra note 10, at 646; see also Issacharoff, supra note 10, at 623 ("[D]emocracy is defined primarily by the accountability of the elected to the electors, an accountability that is in turn shaped through competitive elections."); Richard H. Pildes, Competitive, Deliberative, and Rights-Oriented Democracy, 3 ELECTION L.J. 685, 688 (2004) (reviewing RICHARD A. POSNER, LAW, PRAGMATISM, AND DEMOCRACY (2003)).

(68) See, e.g., Issacharoff & Pildes, supra note 10, at 652-87 (applying competition centered approach to White Primary Cases, write-in ballots, and fusion candidacies); Richard H. Pildes, The Theory of Political Competition, 85 VA. L. REV. 1605, 1613-15 (1999) (applying approach to threshold vote requirements for parties in proportional representation systems).

(69) See Adam B. Cox, Partisan Gerrymandering and Disaggregated Redistricting, 2004 SUP. CT. REV. 409, 425 (2004) ("[I]n places where there are ... politically homogeneous groups of voters, there may be no reasonable redistricting arrangement that is capable of ... produc[ing] competitive general elections.").

(70) Issacharoff & Karlan, supra note 10, at 574; see also Samuel Issacharoff, Surreply, Why Elections?, 116 HARV. L. REV. 684, 692 (2002) ("There will always be a Utah and a Massachusetts. The question is not whether districts should be homogenized ... but whether districts may be rigged so as to diminish or eliminate competition that would otherwise emerge....").

(71) Pildes, supra note 10, at 266.

(72) Issacharoff & Karlan, supra note 10, at 552.

(73) See, e.g., REHFELD, supra note 52, at 177 (endorsing non-territorial districts that each resemble the nation as a whole); Pamela S. Karlan, A Bigger Picture (commenting that since "many citizens' most pressing interests ... are not primarily defined by where they live ... we probably would pick a different system if we were starting from scratch"), in REFLECTING ALL OF US: THE CASE FOR PROPORTIONAL REPRESENTATION 73, 74-76 (Joshua Cohen &Joel Rogers eds., 1999).

(74) For a powerful critique of such a regime, see generally Persily, supra note 16.

(75) See Gardner, Fixed Election Districts, supra note 27, at 572-73 ("[P]artisan competition and territorial districting ... are conflicting and indeed incommensurable principles upon which to base a system of legislative representation."); cf. Michael P. McDonald, Redistricting and Competitive Districts (noting that in Arizona, where competitiveness is a redistricting criterion, competitive districts often can be drawn only by "placing dissimilar communities together"), in THE MARKETPLACE OF DEMOCRACY: ELECTORAL COMPETITION AND AMERICAN POLITICS 222, 239 (Michael P. McDonald & John Samples eds., 2006).

(76) See Grofman & King, supra note 11, at 6-9. A cruder approach would simply ask whether a party receives the same share of seats and votes. This is another way of asking whether representation is proportional--clearly a value that is inconsistent with geographic graphic districting. See Alfange, supra note 22, at 221-22; Schuck, supra note 16, at 1361-77.

(77) Grofman & King, supra note 11, at 6.

(78) Id. at 8.

(79) See LULAC, 548 U.S. 399, 420 (2006) (opinion of Kennedy, J.) ("The existence or degree of asymmetry may in large part depend on conjecture about where possible vote-switchers will reside."); CAIN, supra note 22, at 75 ("If there are enclaves of homogeneous support for one or the other party, then it will be necessary to distinguish between partisan inefficiency caused by residential segregation and that caused by redistricting per se.").

(80) See Alfange, supra note 22, at 223 ("To achieve [partisan symmetry] in a districting system, it would probably be necessary to engage in a process of reverse gerrymandering, creating meandering districts that violate the compactness criterion...."); cf. Jenni Newton-Farrelly, From Gerry-Built to Purpose-Built: Drawing Electoral Boundaries far Unbiased Election Outcomes, 45 REPRESENTATION 471, 476 (2009) (U.K.) (noting that in South Australia, the only .jurisdiction in the world with an explicit partisan fairness requirement, "geographic districts ... [are] split where required by fairness or equity").

(81) See Gardner, Fixed Election Districts, supra note 27, at 576 ("[O]ur ability to achieve fair partisan representation is consistently thwarted by our commitment to territorial representation."); Grofman & King, supra note 11, at 29 (conceding that "partisan bias [can be] necessitated by the state's compliance with neutral districting principles").

(82) See supra notes 50-51 and accompanying text.

(83) Davis v. Bandemer, 478 U.S. 109, 159 (1986) (O'Connor, J., concurring); see also Alfange, supra note 22, at 224 ("Individual [district] elections are often intensely personal matters, turning not in the slightest degree on which party the voter wants to control the legislature.... It just cannot be assumed that a vote for a particular candidate in a particular district is a vote for that candidate's party statewide.").

(84) Cf. Vieth v. Jubelirer, 541 U.S. 267, 308 (2004) (Kennedy, J., concurring in the judgment) (expressing interest in learning more about "principles of fair districting discussed in the annals of parliamentary or legislative bodies" in order to determine "well-accepted rules of fairness that should govern districting").

(85) A.H. BIRCH, REPRESENTATION 27 (1971); see also Baker v. Cart, 369 U.S. 186, 307 (1962) (Frankfurter, J., dissenting) (referring to "the basic English principle of apportioning representatives among the local governmental entities ... rather than among units of approximately equal population"); GORDON E. BAKER, THE REAPPORTIONMENT REVOLUTION: REPRESENTATION, POLITICAL POWER, AND THE SUPREME COURT 15-16 (1966); JOHN C. COURTNEY, COMMISSIONED RIDINGS: DESIGNING CANADA'S ELECTORAL DISTRICTS 206 (2001) ("[T]he House of Commons was originally the House of Communitates, the counties and boroughs of England."); ROSEMARIE ZAGARRI, THE POLITICS OF SIZE: REPRESENTATION IN THE UNITED STATES, 1776-1850, at 37 (1987) ("[E]ach community sent a designated number of delegates to the House of Commons.").

(86) See Gardner, Possibility of Community, supra note 27, at 1244-45; see also Baker, 369 U.S. at 307 (Frankfurter, J., dissenting) (noting that British approach "had early taken root in the colonies"); GORDON E. BAKER, RURAL VERSUS URBAN POLITICAL POWER 7 (1955); HOWARD BALL, THE WARREN COURT'S CONCEPTIONS OF DEMOCRACY: AN EVALUATION OF THE SUPREME COURT'S APPORTIONMENT OPINIONS 52 (1971); REHFELD, supra note 52, at 72-77; Guinier, supra note 52, at 1604-05 ("This link between political representation and ... geographic ties was later carried over to the United States during the Colonial period."); Tucker, supra note 53, at 366-69.

(87) See ROBERT G. DIXON, JR., DEMOCRATIC REPRESENTATION: REAPPORTIONMENT IN LAW AND POLITICS 62-63 (1968) (listing apportionment methods of original thirteen states); ELMER C. GRIFFITH, THE RISE AND DEVELOPMENT OF THE GERRYMANDER 30-31 (1907). For their upper houses, about half the states elected senators from individual counties, while the other half elected them from districts composed of multiple whole counties. See DIXON, supra, at 62; Tucker, supra note 53, at 367-68. 88 Gardner, Representation Without Party, supra note 27, at 918.

(89) Id. at 900-02.

(90) See ZAGARRI, supra note 85, at 57-59; see also BAKER, supra note 85, at 20-21; DIXON, supra note 87, at 66-70 (observing that newer midwestern and western states typically used districts but also often guaranteed at least one representative for each county and barred districts from dividing counties).

(91) See Baker, 369 U.S. at 311 n.85 (Frankfurter, J., dissenting); Gardner, Representation Without Party, supra note 27, at 914-15 ("At no time before the [reapportionment revolution] did state constitutions authorize the creation of ... districts that crossed local government boundaries."); Micah Altman, Districting Principles and Democratic Representation 130-32 (Mar. 31, 1998) (unpublished Ph.D. dissertation, California Institute of Technology), available at

(92) See MONMONIER, supra note 61, at 4; Micah Altman, Traditional Districting Principles: Judicial Myths vs. Reality, 22 SOC. SCI. HIST. 159, 168-71 (1998) (analyzing congressional districts' correspondence to political subdivisions throughout U.S. history); Altman, supra note 91, at 150-51, 163 n.112 (same).

(93) Gardner, Representation Without Party, supra note 27, at 914.

(94) BAKER, supra note 86, at 11.

(95) Id. at 12; DIXON, supra note 87, at 76-77.

(96) Gardner, Fixed Election Districts, supra note 27, at 578.

(97) See Altman, supra note 92, at 181 fig.7; Tucker, supra note 53, at 376.

(98) BAKER, supra note 85, at 27; see also REHFELD, supra note 52, at 57; JOHN PHILLIP REID, THE CONCEPT OF REPRESENTATION IN THE AGE OF THE AMERICAN REVOLUTION 133 (1989) (arguing that American system of representation was based on "the local necessity, the local advantage, the benefit of local representatives ... who knew the electors, their circumstances, needs, and desires"); Gardner, Representation Without Party, supra note 27, at 936, 939 (describing "fixture in American political thought" that "inhabitants of a county or town ... comprise a community," which "justif[ies] the constitutional designation of local governments ... as the fundamental units of legislative representation").

(99) Denney v. State ex rel. Basler, 42 N.E. 929, 937 (Ind. 1896).

(100) Brown v. Saunders, 166 S.E. 105, 108 (Va. 1932).

(101) State ex rel. Att'y Gen. v. Cunningham, 51 N.W. 724, 730 (Wis. 1892); see also People ex rel. Baird v. Bd. of Supervisors, 33 N.E. 827, 830 (N.Y. 1893) (discussing "certain community of interest among the inhabitants of a county").

(102) GRIFFITH, supra note 87, at 17; see also POLE, supra note 51, at 247 (noting that in Massachusetts "old connections had been sundered" and "[g]enuine 'interests' had been divided"); James A. Gardner, Wandering Lonely as a Cloud: National Citizenship and the Case for Non-Territorial Election Districts, 5 ELECTION L.J. 210, 211 (2006) (reviewing REHFELD, supra note 52)(explaining that 1812 Massachusetts gerrymander was "shocking" because "it conspicuously flouted ... county boundaries").

(103) Sauer, supra note 41, at 404-05. Sauer examined districts in Kentucky, Missouri, and Tennessee and identified several gerrymanders under his definition. One district in Missouri, for example, "pair[ed] the cotton farmer of the Southeastern Lowlands with the native of the remote White River hills in Stone and Taney counties, most Ozarkian of the Ozarks. A more ill-matched group would be hard to find." Id. at 413.

(104) See Gardner, Representation Without Party, supra note 27, at 916-20; see also GRIF FITH, supra note 87, at 95-97, 111-13, 123 (noting several states that adopted ban on division of communities).

(105) Parker v. State ex rel. Powell, 32 N.E. 836, 842 (Ind. 1892); see also Cunningham, 51 N.W. at 730 (ruling district plan invalid in part because of "breaking up of the lines and boundaries of counties by the new assembly districts").

(106) See Baker v. Carr, 369 U.S. 186, 187-88, 237 (1962).

(107) See Wesberry v. Sanders, 376 U.S. 1, 2-4 (1964).

(108) See Reynolds v. Sims, 377 U.S. 533, 568 (1964).

(109) See Avery v. Midland Cnty., 390 U.S. 474, 476, 485-86 (1968).

(110) Reynolds, 377 U.S. at 562, 575.

(111) Id. at 580; see also id. at 579-80 ("[N]either history alone, nor economic or other sorts of group interests, are permissible factors in attempting to justify disparities from population-based representation. Citizens, not history or economic interests, cast votes." (footnote omitted)).

(112) 394 U.S. 526, 528-30 (1969).

(113) Id. at 533; see also Alfange, supra note 22, at 197 ("Kirkpatrick's message to the gerrymanderer was clear: there is absolutely no reason to be hesitant about splitting communities among various districts...."). Other Supreme Court cases in the 1960s reinforced Kirkpatrick's message. See Wells v. Rockefeller, 394 U.S. 542, 546-47 (1969) (invalidating New York congressional plan that tried to "keep regions with distinct interests intact"); Kilgarlin v. Hill, 386 U.S. 120, 122-23 (1967) (per curiam) (invalidating Texas congressional plan that "resulted from a bona fide attempt ... to respect county boundaries"); Davis v. Mann, 377 U.S. 678, 686, 691-93 (1964) (invalidating Virginia congressional plan that followed "tradition of respecting the integrity of the boundaries of cities and counties").

(114) 377 U.S. 713 (1964).

(115) Id. at 717.

(116) Id. at 726-28, 738.

(117) Id. at 738-39.

(118) Id. at 719-20.

(119) 369 U.S. 186, 267 (1962) (Frankfurter, J., dissenting).

(120) See, e.g., id. at 269 ("[G]eography, economics, [and] urban-rural conflict.., have throughout our history entered into political districting...."); id. at 301 (noting that one-person, one-vote "was not the English system, it was not the colonial system, it was not the system chosen for the national government by the Constitution").

(121) Reynolds v. Sims, 377 U.S. 533, 622-23 (1964) (Harlan, J., dissenting) (footnotes omitted) (internal quotation marks omitted); see also Gray v. Sanders, 372 U.S. 368, 384 (1963) (Harlan, J., dissenting) (contending that "It]he Court's holding surely flies in the face of history"); Baker, 369 U.S. at 330 (Harlan, J., dissenting) (discussing "abrupt departure the majority makes from judicial history").

(122) Lucas, 377 U.S. at 746, 751 (Stewart, J., dissenting).

(123) Altman, supra note 92, at 187.

(124) Id. at 180; see also Bruce E. Cain et al., From Equality to Fairness: The Path of Political Reform Since Baker v. Carr ("The equal population criterion inevitably wreaked havoc on geographic representation since in many instances homogeneous communities of interest had to be split or combined in order to achieve population equality...."), in PARTY LINES: COMPETITION, PARTISANSHIP, AND CONGRESSIONAL REDISTRICTING 6, 8 (Thomas E. Mann & Bruce E. Cain eds., 2005); Michael W. McConnell, The Redistricting Cases: Original Mistakes and Current Consequences, 24 HARV. J.L. & PUB. POL'Y 103, 112 (2000) ("[L]egislators used to be extremely reluctant to violate city, county, and township lines. Now, under 'one person, one vote,' they are required to do so.").

(125) See Reynolds, 377 U.S. at 583-84.

(126) Gardner, Possibility of Community, supra note 27, at 1242.

(127) Cf. e.g., id. at 1238-43 (failing to consider post-1960s doctrinal history of territorial community); Pamela S. Karlan, Maps and Misreadings: The Role of Geographic Compactness in Racial Vote Dilution Litigation, 24 HARV. C.R.-C.L. L. REV. 173, 173-74 (1989) (same); McConnell, supra note 124, at 108, 112 (same); Richard H. Pildes, Formalism and Functionalism in the Constitutional Law of Politics, 35 CONN. L. REV. 1525, 1525 (2003) (same).

(128) See supra notes 107-18 and accompanying text.

(129) 403 U.S. 182, 184-87 (1971). These districts had a maximum deviation of 11.9%. Id. at 184.

(130) Id. at 185, 187.

(131) See NANCY MAVEETY, REPRESENTATION RIGHTS AND THE BURGER YEARS 46 (1991) (describing Abate as "the first Burger decision that is clearly consistent with the conceptualization of territorial representation"); Alfange, supra note 22, at 198 (noting that with Abate "the Court began a retreat from the extreme rigidity of its 1969 position").

(132) Mahan v. Howell, 410 U.S. 315, 324-28 (1973). The plan had a maximum deviation of 16A%. Id. at 319.

(133) Id. at 323.

(134) Id. at 324; see also Voinovich v. Quilter, 507 U.S. 146, 161-62 (1993) (overturning district court's decision to hold Ohio district plan unconstitutional that had substantial population deviations caused by "policy in favor of preserving county boundaries"); White v. Regester, 412 U.S. 755, 764 & n.8 (1973) (largely upholding Texas plan with sizeable population variations that were precipitated by "policy against cutting county lines in forming representative districts").

(135) Brown v. Thomson, 462 U.S. 835, 843, 848 (1983).

(136) Id. at 839 n.4 (quoting 1981 Wyo. Sess. Laws, ch. 76, [section] 3).

(137) Id. at 841 n.5 (quoting Brown v. Thomson, 536 F. Supp. 780, 784 (D. Wyo. 1982), aff'd, 462 U.S. 835); see also id. (noting district court's finding that small county's interests would be "virtually unprotected" if it were merged with larger county); Shapiro, supra note 43, at 234 ("[Brown] recognized what Chief Justice Warren had steadfastly refused to recognize--that political geography exists.").

(138) See supra notes 110-18 and accompanying text.

(139) Brown, 462 U.S. at 847. Brown is unusual in that the plaintiffs challenged only the over-representation of a single small county, and not the validity of the district plan as a whole. See id. at 846-47. Had they advanced a broader challenge, they may well have prevailed. See id. at 848-50 (O'Connor, J., concurring). Moreover, the Court has shown no willingness to relax the one-person, one-vote rule for congressional (as opposed to state or local) districting. See, e.g., Karcher v. Daggett, 462 U.S. 725, 727, 744 (1983) (striking down congressional district plan with maximum population deviation of less than one percent).

(140) Pub. L. No. 89-110, 79 Stat. 437 (codified as amended at 42 U.S.C. [section][section] 1973 1973bb-1 (2006)).

(141) 412 U.S. 755, 769-70 (1973).

(142) Id. at 769.

(143) Id. at 768.

(144) Id. at 769. For similar reasons, the White Court also invalidated a Dallas multi member district that disadvantaged an African American community. See id. at 765-67.

(145) Rogers v. Lodge, 458 U.S. 613, 626-27 (1982).

(146) See id. at 627-28.

(147) See Voting Rights Act Amendments of 1982, Pub. L. No. 97-205, sec. 3, [section] 2, 96 Stat. 131, 134 (codified at 42 U.S.C. [section] 1973 (2006)).

(148) 478 U.S. 30, 50-51 (1986).

(149) Id. (emphasis added). The Court further explained that "members of geographically insular racial and ethnic groups" are entitled to statutory protection because they "frequently share socioeconomic characteristics" that can give rise to communal ties, such as "income level, employment status, amount of education, housing and other living conditions, religion, [and] language." Id. at 64.

(150) Growe v. Emison, 507 U.S. 25, 38-41 (1993).

(151) Id. at 41.

(152) See Bush v. Vera, 517 U.S. 952, 979 (1996) (plurality opinion); Shaw II, 517 U.S. 899, 916-17 (1996).

(153) Bush, 517 U.S. at 979; see also Shaw II, 517 U.S. at 916 (holding that district that did not contain geographically compact minority population "could not remedy any potential [section] 2 violation"). The Court also stated explicitly in Bush that the section 2 compactness inquiry should take into account communal considerations. See 517 U.S. at 977.

(154) 548 U.S. 399, 430-31, 435 (2006).

(155) Id. at 435, 439.

(156) Id. at 440; see also id. at 441 ("The State chose to break apart a Latino opportunity district to protect the incumbent congressman from the growing dissatisfaction of the cohesive and politically active Latino community in the district.").

(157) Id. at 440-41.

(158) Id. at 432.

(159) Id. at 432, 434 (internal quotation marks omitted).

(160) Id. at 433.

(161) Cf. Daniel R. Ortiz, Cultural Compactness, 105 MICH. L. REV. FIRST IMPRESSIONS 48, 50 (2006),; Richard H. Pildes, The Decline of Legally Mandated Minority Representation, 68 OHIO ST. L.J. 1139, 1159 (2007) (noting that in LULAC "[t]he touchstone appears to be the concept of a 'naturally arising' minority district, one that exists or would exist due to the geographic concentration of minority voters whose proximity also reflects common socioeconomic and other interests").

(162) 509 U.S. 630, 642, 649 (1993).

(163) Id. at 635; see also Shaw II, 517 U.S. 899, 903, 917-18 (1996).

(164) Shaw I, 509 U.S. at 647; see also id. (objecting to districts "in which a State concentrated a dispersed minority").

(165) Miller v. Johnson, 515 U.S. 900, 917-20, 922 (1995).

(166) Id. at 908, 919; see also id. at 908 ("[T]he social, political, and economic makeup of the Eleventh District tells a tale of disparity, not community.").

(167) 517 U.S. 952 (1996) (plurality opinion).

(168) Id. at 966.

(169) Id. at 971 n.*.

(170) Id. at 1033 (Stevens, J., dissenting); see also id. at 975 (plurality opinion) (describing "intricacy of the lines drawn, separating Hispanic voters from African-American voters on a block-by-block basis").

(171) Abrams v. Johnson, 521 U.S. 74, 100 (1997).

(172) Lawyer v. Dep't of Justice, 521 U.S. 567, 581 (1997) (citations omitted) (internal quotation marks omitted).

(173) See Easley v. Cromartie, 532 U.S. 234, 257-58 (2001); Hunt v. Cromartie, 526 U.S. 541, 552-54 (1999).

(174) Easley, 532 U.S. at 250 (internal quotation marks omitted); see also Hunt, 526 U.S. at 544 (describing revised district).

(175) See supra note 17 and accompanying text.

(176) 462 U.S. 725, 762 (1983) (Stevens, J., concurring). The majority opinion in Karcher solely addressed one-person, one-vote issues.

(177) Id. at 762, 764 n.33.

(178) Id. at 764 n.33.

(179) Id. at 762-63; see also id. at 789 (Powell, J., dissenting) (noting that many districts did not "reflect any attempt to follow natural, historical, or local political boundaries").

(180) 478 U.S. 109, 167 (1986) (Powell, J., concurring in part and dissenting in part).

(181) Id. (internal quotation marks omitted).

(182) Id. at 173 n.13, 177 (internal quotation marks omitted).

(183) Id. at 176; see also id. at 177 ("IT]he mapmakers gave no consideration to the interests of communities."); id. ("[T]he manner in which the districts divide established communities ... illustrate that community interests were ignored....").

(184) Id.

(185) Id. at 180.

(186) Id. at 180 n.21 (internal quotation marks omitted).

(187) Vieth v. Jubelirer, 541 U.S. 267, 340 (2004) (Stevens, J., dissenting) (internal quotation marks omitted).

(188) LULAC, 548 U.S. 399, 479 (2006) (Stevens, J, concurring in part and dissenting in part). It is clear that Justice Stevens would have invalidated these districts on political gerrymandering grounds, not because of racial vote dilution. See id. at 475.

(189) The petitioners in Vieth focused on statewide rather than district-specific claims. See 541 U.S. at 355 (Souter, J., dissenting). The petitioners in LULAC emphasized the mid-decade timing of Texas's redistricting. See 548 U.S. at 413-23 (opinion of Kennedy, J.).

(190) ALASKA CONST. art. VI, [section] 6.

(191) CAL. CONST. art. XXI, [section] 2(d)(4).

(192) HAW. CONST. art. IV, [section] 6(8); see also ARIZ. CONST. art. IV, pt. 2, [section]1(14) (D) ("District boundaries shall respect communities of interest to the extent practicable...."); COLO. CONST. art. V, [section] 47(3) ("[C]ommunities of interest, including ethnic, cultural, economic, trade area, geographic, and demographic factors, shall be preserved...."). In addition, the county-preservation provision in North Carolina's constitution and the compactness provision in Rhode Island's constitution have been judicially interpreted to require consideration of communities of interest. See Stephenson v. Bartlett, 562 S.E.2d 377, 397 (N.C. 2002); Parella v. Montalbano, 899 A.2d 1226, 1255 (R.I. 2006).

(193) The seven states with statutory requirements are Idaho, Maine, Oregon, South Dakota, Vermont, Washington, and Wisconsin. The twelve states with similar guidelines are Alabama, Kansas, Kentucky, Maryland, Minnesota, Missouri, Montana, Nebraska, New Mexico, South Carolina, Virginia, and Wyoming. See NAT'L CONFERENCE OF STATE LEGISIATURES, REDISTRICTING LAW 2010, at 172-217 (2009) [hereinafter NCSL GUIDE] (describing redistricting process in each state).

(194) See id.

(195) See Gardner, Representation Without Party, supra note 27.

(196) Id. at 933-34.

(197) Id. at 939.

(198) Id. at 963-64.

(199) Id. at 964.

(200) Id. at 965.

(201) In re 2001 Redistricting Cases, 44 P.3d 141, 149-50 (Alaska 2002); Hickel v. Se. Conference, 846 P.2d 38, 44-54 (Alaska 1993); Kenai Peninsula Borough v. State, 743 P.2d 1352, 1359-61 (Alaska 1987); Carpenter v. Hammond, 667 P.2d 1204, 1214-15 (Alaska 1983); Groh v. Egan, 526 P.2d 863, 879-80 (Alaska 1974).

(202) Ariz. Minority Coal. for Fair Redistricting v. Ariz. Indep. Redistricting Comm'n, 121 P.3d 843, 867-70 (Ariz. Ct. App. 2005).

(203) In re Reapportionment of Colo. Gen. Assembly, 46 P.3d 1083, 1090-91 (Colo. 2002); In re Reapportionment of Colo. Gen. Assembly, 45 P.3d 1237, 1247-49, 1251-52 (Colo. 2002); In re Reapportionment of Colo. Gen. Assembly, 828 P.2d 213, 215-17 (Colo. 1992); In re Reapportionment of Colo. Gen. Assembly, 828 P.2d 185, 194-200 (Colo. 1992); In re Reapportionment of Colo. Gen. Assembly, 647 P.2d 209, 212-13 (Colo. 1982); In re Reapportionment of Colo. Gen. Assembly, 647 P.2d 191, 194 (Colo. 1982).

(204) Bingham Cnty. v. Idaho Comm'n for Reapportionment, 55 P.3d 863, 865-71 (Idaho 2002).

(205) Graham v. Thornburgh, 207 F. Supp. 2d 1280, 1294-95 (D. Kan. 2002); In re Petition of Stephan, 836 P.2d 574, 581-83 (Kan. 1992).

(206) McBride v. Mahoney, 573 F. Supp. 913,915-18 (D. Mont. 1983).

(207) Stephenson v. Bartlett, 582 S.E.2d 247, 252-54 (N.C. 2003).

(208) Hartung v. Bradbury, 33 P.3d 972 (Or. 2001); Ater v. Keisling, 819 P.2d 296, 304-06 (Or. 1991).

(209) Parella v. Montalbano, 899 A.2d 1226, 1255-56 (R.I. 2006).

(210) In re Reapportionment of Towns of Woodbury & Worcester, 861 A.2d 1117 (Vt. 2004); In re Reapportionment of Towns of Hartland, Windsor & W. Windsor, 624 A.2d 323, 345-46 (Vt. 1993).

(211) Both state and federal courts have also drawn districts based in large part on communal considerations. See, e.g., Baumgart v. Wendelberger, Nos. 01-0121, 02-0366, 2002 WL 34127471, at *3, *7 (E.D. Wis. May 30, 2002); Smith v. Clark, 189 F. Supp. 2d 529, 542-46 (S.D. Miss. 2002); Arizonans for Fair Representation v. Symington, 828 F. Supp. 684, 688 (D. Ariz. 1992); Prosser v. Elections Bd., 793 F. Supp. 859, 863 (W.D. Wis. 1992); Wis. State AFL-CIO v. Elections Bd., 543 F. Supp. 630, 636-38 (E.D. Wis. 1982); Carstens v. Lamm, 543 F. Supp. 68, 89-93 (D. Colo. 1982); LaComb v. Growe, 541 F. Supp. 145, 148-50 (D. Minn. 1982); O'Sullivan v. Brier, 540 F. Supp. 1200, 120306 (D. Kan. 1982); Egan v. Hammond, 502 P.2d 856, 894, 896 (Alaska 1972); Wilson v. Eu, 823 P.2d 545, 551-53 (Cal. 1992); Legislature of Cal. v. Reinecke, 516 P.2d 6, 10, 16 (Cal. 1973); Beauprez v. Avalos, 42 P.3d 642, 647, 651-52 (Colo. 2002); In re 2003 Apportionment of State Senate, 827 A.2d 844, 847-48 (Me. 2003) ; Mellow v. Mitchell, 607 A.2d 204, 224 (Pa. 1992).

(212) See Hickel v. Se. Conference, 846 P.2d 38, 52-53 (Alaska 1993).

(213) See id. at 53.

(214) See In re Reapportionment of Colo. Gen. Assembly, 647 P.2d 209, 212 (Colo. 1982).

(215) See In re Reapportionment of Colo. Gen. Assembly, 828 P.2d 185, 195-96 (Colo. 1992).

(216) See Bingham Cnty. v. Idaho Comm'n for Reapportionment, 55 P.3d 863, 869-71 (Idaho 2002). Because the court struck down the entire district plan on one-person, one-vote grounds, its discussion of certain communities' division was technically dicta.

(217) See In re Reapportionment of Towns of Hartland, Windsor & W. Windsor, 624 A.2d 323, 331 (Vt. 1993).

(218) See infra notes 237-48.

(219) See PHILIP BOBBITT, CONSTITUTIONAL INTERPRETATION 12-13 (1991) (summarizing six modalities of constitutional interpretation). The territorial community test is also supported by Bobbitt's structural and prudential modalities of argument. Cf. infra Parts IV-V (discussing test's manageability and political consequences); infra text accompanying note 220 (elaborating on structural value of participatory democracy).

(220) See Issacharoff & Pildes, supra note 10, at 713, 716 (discussing "the great silences of the Constitution regarding the structure of electoral politics" and the "Court's efforts to fill the gaps of [its] framework for democratic politics"); see also Issacharoff, supra note 70, at 687-88. The structural principle that is most relevant here is participatory democracy, which is rooted most deeply in the Republican Guarantee Clause, see U.S. CONST. art. IV, [section] 4, and which has been stressed by writers including John Hart Ely, see JOHN HART ELY, DEMOCRACY AND DISTRUST: A THEORY OF JUDICIAL REVIEW 116-25 (1980), and Justice Breyer, see STEPHEN BREWER, ACTIVE LIBERTY: INTERPRETING OUR DEMOCRATIC CONSTITUTION 21-34 (2005). As both a theoretical and an empirical matter, districts that correspond to territorial communities tend to promote democratic participation. See supra note 43 and accompanying text (discussing theoretical reasons why this is so); see also infra subsections V.A.3, V.B.3 (presenting empirical confirmation).

(221) None of the Justices who proposed political gerrymandering standards in Karcher, Bandemer, Vieth, and LULAC attempted to provide any textual foundation for their approaches. Similarly, Issacharoff and Pildes freely admit that their competition-centered test is not textually grounded. See Issacharoff, supra note 70, at 687 (conceding that he can offer "no narrow textual justification" for his approach); Issacharoff & Pildes, supra note 10, at 716.

(222) Because the territorial community test is not constitutionally compelled, it might be preferable for it to be adopted by legislation or popular initiative rather than by judicial declaration. In addition, even if the test is not embraced as a self-standing bar on gerrymandering, it could be treated as a key component of a broader inquiry. For instance, a district plan's needless disruption of geographic communities is powerful evidence that partisan advantage (or incumbent protection) motivated the line-drawers. Cf. Vieth v. Jubelirer, 541 U.S. 267, 339 (2004) (Stevens, J., dissenting) (advocating partisan intent as standard for identifying unconstitutional gerrymanders).

(223) Unsurprisingly, this definition is closely related to the theory of communal rep resentation discussed in Section I.A. See supra notes 36-39 and accompanying text. It is also similar to the definitions advanced by various scholars. See, e.g., Richard Briffault, The Local Government Boundary Problem in Metropolitan Areas, 48 STAN. L. REV. 1115, 1127 (1996) (referring to community as "a place with a distinctive history, identifiable characteristics, and a unique identity"); Ford, supra note 58, at 859 ("An organic community may be united primarily by economy or by culture."); David M. Hummon, Community Attachment: Local Sentiment and Sense of Place (discussing community's "complex sources in both subjectively perceived and objective aspects of the local environment"), in PLACE ATTACHMENT 253, 253 (Irwin Altman& Setha M. Low eds., 1992).

It is worth noting as well that territorial communities can shift over time as people's interests and affiliations change. Courts should keep in mind the possibility of such shifts when they seek to ascertain community boundaries. Lastly, it is important to acknowledge that there are inevitably exceptions and outliers in any territorial community. The concept is meant to capture the interests and affiliations of most--not all people in a geographically defined area. See supra notes 4445 and accompanying text.

(224) See Thornburg v. Gingles, 478 U.S. 30, 50-51 (1986); see also supra notes 147-61 and accompanying text.

(225) Legislature v. Reinecke, 516 P.2d 6, 16 (Cal. 1973); see also Order Stating Redistricting Principles and Requirements for Plan Submissions at 3, Zachman v. Kiffmeyer, No. C0-01-160 (Minn. Special Redistricting Panel Dec. 11, 2001) (defining communities of interest as "groups of Minnesota citizens with clearly recognizable similarities of social, geographic, political, cultural, ethnic, economic, or other interests").

(226) Carstens v. Lamm, 543 F. Supp. 68, 91 (D. Colo. 1982).

(227) ALASKA CONST. art. VI, [section] 6.

(228) CAL. CONST. art. XXI, [section] 2(d)(4). Colorado, Kansas, Missouri, Montana, Vermont, and Virginia include similar definitions in their state constitutions, statutes, or redistricting guidelines. See NCSL GUIDE, supra note 193, at 175-76, 183-84, 194-96, 210-14.

(229) See In re Legislative Redistricting of State, 475 A.2d 428, 445 (Md. 1982) (characterizing concept of "communit[y] of interest" as "nebulous" and "unworkable"); Gardner, Representation Without Party, supra note 27, at 937-38 (noting that "community of interest" is used "so broadly and indiscriminately as to include virtually any group of people who share some trait or characteristic that has the potential to be salient politically" and "is not linked in any particular way to a specific piece of territory").

(230) See GREGORY R. WEIHER, THE FRACTURED METROPOLIS: POLITICAL FRAGMENTA TION AND METROPOLITAN SEGREGATION 100-07, 190 (1991); Richard Briffault, Our Localism: Localism and Legal Theory (pt. 2), 90 COLUM. L. REV. 346, 353-54 (1990).

(231) See BAKER, supra note 85, at 102 ("Economic and social interests usually transcend county and even state lines."); BUTLER & CAIN, supra note 49, at 70; DANIEL J. ELAZAR, THE AMERICAN MOSAIC: THE IMPACT OF SPACE, TIME, AND CULTURE ON AMERICAN POLITICS 289 (1994) ("[I]n many cases, the limits of specific civil communities may be less easily discovered within formal political boundaries."); HANSON, supra note 49, at 130 ("[P]olitical subdivisions ... are quite likely to fail to reflect ... a community of interest....").

(232) Cf., e.g., Wilson v. Eu, 823 P.2d 545, 575 (Cal. 1992) (giving "precedence to keeping geographically compact minority groups together rather than maintain[ing] city boundaries").

(233) Knight, supra note 58, at 515; see also ELAZAR, supra note 231, at 3; MORRILL, supra note 58, at 23; Robert J. Chaskin, Perspectives on Neighborhood and Community: A Review of the Literature, 71 SOC. SERVICE REV. 521, 535 (1997) ("[T]he units in which the circumstances and activities of daily life inhere can be 'nested,' where each member of a community is simultaneously a member of others.").

(234) See White v. Weiser, 412 U.S. 783, 793 (1973) (noting that "congressional districts are not so intertwined and freighted with strictly local interests as are state legislative districts"); In re Reapportionment of Towns of Hartland, Windsor & W. Windsor, 624 A.2d 323, 345 (Vt. 1993) (observing that it is "difficult, if not impossible, to achieve the same level of common interests among individual towns within Senate districts that is attainable among towns within House districts"). The implication is that legislatures composed of smaller districts (e.g., state houses) would focus on specific localized concerns, while legislatures composed of larger districts (e.g., Congress) would devote more attention to broader and more diffuse issues.

(235) See LULAC, 548 U.S. 399, 442 (2006); Shaw I, 509 U.S. 630, 647 (1993).

(236) See supra subsections II.C.2-3; see also Carpenter v. Hammond, 667 P.2d 1204, 1218 (Alaska 1983) (Matthews, J., concurring) (noting that town belonged to different community than rest of district, despite similar economic interests, where it was located "700 miles and two time zones away"); In re Reapportionment of Colo. Gen. Assembly, 828 P.2d 185, 195 (Colo. 1992) (en banc) (criticizing district that combined unconnected counties on opposite sides of Continental Divide).

(237) See Lawyer v. Dep't of Justice, 521 U.S. 567, 581 (1997); White v. Regester, 412 U.S. 755, 769-70 (1973). These cases also underscore the distinction between a community of interest and a territorial community. People of the same race are clearly a community of interest, but they are not a territorial community (which may prevail in a section 2 challenge and defeat a racial gerrymandering claim) unless they also live near one another, share non-racial interests, and feel subjectively unified.

(238) CAL. CONST. art. XXI, [section] 2(d)(4).

(239) See Groh v. Egan, 526 P.2d 863, 879 n.65 (Alaska 1974).

(240) See Ater v. Keisling, 819 P.2d 296, 303 n.10 (Or. 1991).

(241) See Beauprez v. Avalos, 42 P.3d 642, 652 (Colo. 2002).

(242) See Legislature of Cal. v. Reinecke, 516 P.2d 6, 29 (Cal. 1973).

(243) See Carstens v. Lamm, 543 F. Supp. 68, 93 (D. Colo. 1982); Reinecke, 516 P.2d at 31.

(244) See, e.g., Carstens, 543 F. Supp. at 96 (Denver); Groh, 526 P.2d at 879-80 (Anchorage); Wilson v. Eu, 823 P.2d 545, 581, 583, 587, 588 (Cal. 1992) (San Francisco, Sacramento, Los Angeles, and San Diego).

(245) See, e.g., Kenai Peninsula Borough v. State, 743 P.2d 1352, 1363 (Alaska 1987); Hartung v. Bradbury, 33 P.3d 972, 982-83 (Or. 2001).

(246) See, e.g., In re 2001 Redistricting Cases, 44 P.3d 141, 145 (Alaska 2002); In re Re apportionment of Towns of Hartland, Windsor & W. Windsor, 624 A.2d 323, 340 (Vt. 1993).

(247) See, e.g., Kenai, 743 P.2d at 1362-63; In re Reapportionment of Towns of Wood bury & Worcester, 861 A.2d 1117, 1122 (Vt. 2004).

(248) See, e.g., Bingham Cnty. v. Idaho Comm'n for Reapportionment, 55 P.3d 863, 870 (Idaho 2002); Hartung, 33 P.3d at 982.

(249) See Arizonans for Fair Representation v. Symington, 828 F. Supp. 684, 690 (D. Ariz. 1992); Ariz. Minority Coal. for Fair Redistricting v. Ariz. Indep. Redistricting Comm'n, 121 P.3d 843, 868, 871 (Ariz. Ct. App. 2005).

(250) Carstens v. Lamm, 543 F. Supp. 68, 92 (D. Colo. 1982); see also Carpenter v. Hammond, 667 P.2d 1204, 1215 (Alaska 1983) (invalidating district that combined communities similar in their interests but "completely separate" in their "social activity").

(251) LULAC, 548 U.S. 399, 432 (2006).

(252) Id. at 441.

(253) Id. at 434.

(254) Shaw I, 509 U.S. 630, 635 (1993).

(255) Miller v. Johnson, 515 U.S. 900, 908 (1995). According to my survey of the case law, fusion is the kind of community disruption that has most often led state courts to invalidate districts.

(256) See LULAC, 548 U.S. at 442; Miller, 515 U.S. at 927; Shaw I, 509 U.S. at 649.

(257) LULAC, 548 U.S. at 435, 439.

(258) 430 U.S. 144, 152 (1977).

(259) Id. at 167-68.

(260) Id. at 186 (Burger, C.J., dissenting). The Hasidim also lost because they did not advance an explicit community-based claim to their own district. See id. at 154 n.14 (majority opinion). In the state courts, fragmentation is a less common basis for invalidating a district than fusion.

(261) Bush v. Vera, 517 U.S. 952, 975 (1996) (plurality opinion).

(262) 364 U.S. 339, 340 (1960); see also Wright v. Rockefeller, 376 U.S. 52, 59 (1964) (Douglas, J., dissenting) (criticizing "zigzag, tortuous lines [that were] drawn to concentrate Negroes and Puerto Ricans in Manhattan's Eighteenth Congressional District and practically to exclude them from the Seventeenth"). I have not found any state cases invalidating districts because of community subversion.

(263) See Bush, 517 U.S. at 976, 983; Gomillion, 364 U.S. at 347-48.

(264) In re Reapportionment of Towns of Hartland, Windsor & W. Windsor, 624 A.2d 323, 326 (Vt. 1993).

(265) Id. at 327.

(266) In re Reapportionment of Colo. Gen. Assembly, 647 P.2d 191, 194, 197 (Colo. 1982); see also Graham v. Thornburgh, 207 F. Supp. 2d 1280, 1292 (D. Kan. 2002) (adopting deferential review standard); Carpenter v. Hammond, 667 P.2d 1204, 1214-15 (Alaska 1983) (same); Ater v. Keisling, 819 P.2d 296, 303 (Or. 1991) (same); Parella v. Montalbano, 899 A.2d 1226, 1256 (R.I. 2006) (same).

(267) See Backstrom et al., supra note 22, at 153 (discussing how principle of community preservation can be deployed both when communities must be split and when they must be combined).

(268) The larger population variances allowed by the Court for state legislative districts would reduce the community disruption caused by the one-person, one-vote rule at this level. See supra subsection II.C.1. The disruption would be greater at the congressional redistricting level, at which essentially no population deviations are permitted. See Karcher v. Daggett, 462 U.S. 725, 732-33 (1983).

The disruptive impact of the one-person, one-vote rule could also be reduced through the use of floterial districts. These are districts that overlay (or "float above") multiple conventional districts, providing additional representation to voters whose districts would otherwise be overpopulated. For instance, a community whose population entitles it to 1.5 representatives could be kept intact in a single conventional district with a single representative, while also being included in a larger floterial district that includes other overpopulated conventional districts. See generally Gary F. Moncrief, Floterial Districts, Reapportionment, and the Puzzle of Representation, 14 LEGIS. STUD. Q. 251 (1989).

(269) In re 2001 Redististricting Cases, 44 P.3d 141, 145 (Alaska 2002); see also id. (ensuring that each portion of community that was divided between two districts was "integrated, as nearly as practicable, with the district in which it [was] placed").

(270) Carstens v. Lamm, 543 F. Supp. 68, 87, 96 (D. Colo. 1982).

(271) Id. at 96-97; see also O'Sullivan v. Brier, 540 F. Supp. 1200, 1204 (D. Kan. 1982) (labeling as its "lodestar" the goal of "grouping together ... as many major communities ... as possible" while still complying with one-person, one-vote rule); Beauprez v. Avalos, 42 P.3d 642, 652 (Colo. 2002) (choosing most logical Colorado west slope counties to merge with Denver suburbs); Ater, 819 P.2d at 303 n.10 (using "best of the available choices" when forced to split an Oregon county); In re Reapportionment of Towns of Hartland, Windsor & W. Windsor, 624 A.2d 323, 336 (Vt. 1993) (noting approvingly that district line that divided Vermont town "was drawn to separate [its] most urban areas from [more rural] neighborhoods"); CAIN, supra note 22, at 63-64, 72 (discussing examples of California communities that could sensibly be combined by districts); Gardner, Possibility of Community, supra note 27, at 1258-59 (suggesting that if a community must be divided, its "core" should be preserved and its "hinterlands" sacrificed).

(272) See supra subsections I.B.2-3. To the extent that these approaches conflict with the territorial community test--if, for example, lower bias and higher responsiveness could be achieved only by disrupting more communities--I would want the value of community preservation to take precedence over other redistricting goals. But I would have no problem with a district plan that complies with the territorial community test being struck down for other reasons.

(273) See supra subsection II.C.1.

(274) See supra subsection II.C.2.

(275) See supra subsection II.C.3.

(276) See supra subsection II.C.5.

(277) See supra notes 2-19 and accompanying text.

(278) See Issacharoff, supra note 10, at 638-41 (discussing "perverse incentives" created by Supreme Court decisions "encourag[ing] the racialization of all claims of improper manipulation of the redistricting process").

(279) Vieth v. Jubelirer, 541 U.S. 267, 281 (2004) (plurality opinion); see also LULAC, 548 U.S. 399, 417-20 (2006) (opinion of Kennedy, J.).

(280) Vieth, 541 U.S. at 278.

(281) See supra notes 22-24 and accompanying text.

(282) See LULAC, 548 U.S. at 417-19 (opinion of Kennedy, J.); Vieth, 541 U.S. at 284-86, 292-95 (plurality opinion); id. at 316 (Kennedy, J., concurring in the judgment) ("[C]ourts must be cautious about adopting a standard that turns on whether the partisan interests in the redistricting process were excessive.").

(283) See Vieth, 541 U.S. at 286-90, 296-97, 299-301 (plurality opinion).

(284) Id. at 291.

(285) Id. at 290.

(286) See supra subsection II.C.1.

(287) Brown v. Thomson, 462 U.S. 835, 853 (1983) (Brennan, J., dissenting).

(288) See supra subsection II.C.2.

(289) See White v. Regester, 412 U.S. 755, 768-70 (1973).

(290) See Rogers v. Lodge, 458 U.S. 613, 626-27 (1982).

(291) See LULAC, 548 U.S. 399, 435, 438-41 (2006).

(292) See Growe v. Emison, 507 U.S. 25, 41-42 (1993).

(293) See Abrams v. Johnson, 521 U.S. 74, 91-92 (1997).

(294) See Shaw II, 517 U.S. 899, 916-18 (1996).

(295) See Bush v. Vera, 517 U.S. 952, 979-81 (1996) (plurality opinion).

(296) See LULAC, 548 U.S. at 435,438-41.

(297) See supra subsection II.C.3.

(298) See Shaw I, 509 U.S. 630, 647-49 (1993).

(299) See Miller v. Johnson, 515 U.S. 900, 908-09, 919-20 (1995).

(300) See Bush, 517 U.S. at 966, 971 n.*, 975.

(301) See Lawyer v. Dep't of Justice, 521 U.S. 567, 581-83 (1997).

(302) See Hunt v. Cromartie, 526 U.S. 541, 544, 553-54 (1999).

(303) Vieth v. Jubelirer, 541 U.S. 267, 278 (2004) (plurality opinion).

(304) Bush, 517 U.S. at 1026 (Stevens, J., dissenting).

(305) See id. at 966-67 (plurality opinion).

(306) LULAC, 548 U.S. 399, 504-05 (2006) (Roberts, C.J., concurring in part, concur ring in the judgment in part, and dissenting in part).

(307) See id. at 435, 438-41 (majority opinion).

(308) See supra subsection II.C.5.

(309) See supra subsection II.C.5.

(310) See Carpenter v. Hammond, 667 P.2d 1204 (Alaska 1983) (3-2 majority); In re Reapportionment of Colo. Gen. Assembly, 828 P.2d 185 (Colo. 1992) (one out of seven challenged districts upheld by 4-3 majority); In re Reapportionment of Towns of Hartland, Windsor & W. Windsor, 624 A.2d 323 (Vt. 1993) (one out of five challenged districts struck down by 3-2 majority).

(311) As does the fact that both state and lower federal courts have frequently drawn districts so that they correspond to territorial communities. See supra note 211.

(312) Cf. Vieth v. Jubelirer, 541 U.S. 267, 280-81 (2004) (plurality opinion) (finding Bandemer standard unmanageable in part because it always resulted in districts being upheld).

(313) See supra notes 212-17 and accompanying text.

(314) Carpenter, 667 P.2d at 1215.

(315) Hickel v. Se. Conference, 846 P.2d 38, 51 (Alaska 1993).

(316) In re Reapportionment of Colo. Gen. Assembly, 647 P.2d 209, 212-13 (Colo. 1982).

(317) Stephenson v. Bartlett, 582 S.E.2d 247, 253-54 (N.C. 2003).

(318) See, e.g., Hickel, 846 P.2d at 53-54 (noting that district was drawn with "little consideration of the relative socio-economic integration of the people who live there"); Carpenter, 667 P.2d at 1215 (pointing out that "record [was] simply devoid of evidence" that communities were related); In re Reapportionment of Colo. Gen. Assembly, 828 P.2d 185, 195-96 (Colo. 1992) (concluding that Colorado Reapportionment Commission could not provide "adequate factual showing" for its districting choices).

(319) See, e.g., Hickel, 846 P.2d at 51 (asserting that district plan raised "specter of gerrymandering"); In re Reapportionment of Colo. Gen. Assembly, 647 P.2d 209, 213 (Colo. 1982) (noting "partisan political nature of the Commission's action").

(320) See supra notes 245-48 and accompanying text.

(321) See, e.g., Kenai Peninsula Borough v. State, 743 P.2d 1352, 1362-63 (Alaska 1987) (upholding district despite constituent areas' "minimal" interaction because of their economic ties, transport links, and shared media outlets); In re Reapportionment of Towns of Woodbury & Worcester, 861 A.2d 1117, 1122 (Vt. 2004) (same where "petitioners' preferences and the bulk of their interests may indeed fall outside their current district" but "findings also include substantial ties" within district).

(322) See, e.g., In re Reapportionment of Colo. Gen. Assembly, 828 P.2d at 197; In re Petition of Stephan, 836 P.2d 574, 581-82 (Kan. 1992); Hartung v. Bradbury, 33 P.3d 972, 981-85 (Or. 2001); In re. Reapportionment of Towns of Hartland, Windsor & W. Windsor, 624 A.2d 323, 339-45 (Vt. 1993).

(323) See Groh v. Egan, 526 P.2d 863, 879-80 (Alaska 1974).

(324) See Carpenter v. Hammond, 667 P.2d 1204, 1214 (Alaska 1983).

(325) Kenai, 743 P.2d at 1363; see also id. at 1360 (contrasting unified Juneau district with "the division of relatively similar districts in Anchorage that we rejected in Groh"); id. at 1361 ("Like the Juneau District upheld in Groh, District 2 effectuates a rational state policy....").

(326) Hickel v. Se. Conference, 846 P.2d 38, 46-47 (Alaska 1992); see also In re 2001 Redistricting Cases, 44 P.3d 141, 145 (Alaska 2002) (citing Hickel, 846 P.2d at 51 n.20).

(327) See supra note 310 and accompanying text.

(328) Carpenter, 667 P.2d at 1222 (Compton, J., dissenting).

(329) See In re Reapportionment of Colo. Gen. Assembly, 828 P.2d 185, 210-11 (Colo. 1992) (Mullarky, J., concurring in part and dissenting in part).

(330) In re Reapportionment of Towns of Hartland, Windsor & W. Windsor, 624 A.2d 323, 349 (Vt. 1993) (Dooley, J., concurring and dissenting). It is worth noting that the Vermont statute is more "vague" and "imprecis[e]" than my conception of the territorial community test. See VT. STAT. ANN. tit. 17 [section] 1903(b)(2) (2002) (requiring districts to be drawn to achieve "recognition and maintenance of patterns of geography, social interaction, trade, political ties and common interests").

(331) Cf. Richard H. Pildes, Is Voting-Rights Law Now at War with Itself? Social Science and Voting in the 2000s, 80 N.C.L. REV. 1517, 1518 (2002) ("Law and social science are perhaps nowhere more mutually dependent than in the voting-rights field.").

(332) See DANTE CHINNI & JAMES GIMPEL, OUR PATCHWORK NATION: THE SURPRISING TRUTH ABOUT THE "REAL" AMERICA (2010). In a sequel to this Article, I also attempt to identify communities directly using newly available Census data as well as specialized mapping software. See Stephanopoulos, supra note 32.

(333) See CHINNI & GIMPEL, supra note 332, at 222-23.

(334) See id. at 223-25.

(335) See id. at 9-10. The twelve composite variables used to sort the country's counties explain seventy to eighty percent of the variance in the underlying data. Id. at 224; see also Brian A. Mikelbank, A Typology of U.S. Suburban Places, 15 HOUSING POL'Y DEBATE 935, 961 (2004) (using similar statistical technique to assign U.S. suburbs to ten different categories); Thomas J. Vicino et al., Megalopolis 50 Years On: The Transformation of a City Region, 31 INT'L J. URB. & REGIONAL RES. 344, 361-64 (2007) (assigning Census places in Northeast to five categories).

(336) About, COMMONCENSUS MAP PROJECT, php (last visited Feb. 15, 2012).

(337) See The Maps Page, COMMONCENSUS MAP PROJECT, http://www.commoncensus. org/maps.php (last visited Feb. 15, 2012); see also Matt Chisholm & Ross Cohen, NEIGHBORHOOD PROJECT, (similar San Francisco project) (last visited Feb. 15, 2012); cf. Claudia J. Coulton et al., Mapping Residents' Perceptions of Neighborhood Boundaries: A Methodological Note, 29 AM. J. COMMUNITY PSYCHOL. 371, 371-72 (2001) (discussing pilot study mapping Cleveland neighborhoods by surveying their residents).

(338) See FAQ COMMONCENSUS MAP PROJECT, (last visited Feb. 15, 2012).

(339) See Richard G. Niemi et al., The Effects of Congruity Between Community and District on Salience of U.S. House Candidates, 11 LEGIS. STUD. Q. 187, 198-99 (1986).

(340) See Richard N. Engstrom, District Geography and Voters, in REDISTRICTING IN THE NEW MILLENNIUM 65, 74-75 (Peter F. Galderisi ed., 2005); Jonathan Winburn & Michael W. Wagner, Carving Voters Out: Redistricting's Influence on Political Information, Turnout, and Voting Behavior, 63 POL. RES. Q. 373, 375 (2010).

(341) Cf. William A. Fischel, Neither "Creatures of the State" nor "Accidents of Geography": The Creation of American Public School Districts in the Twentieth Century, 77 U. CHI. L. REV. 177, 199 (2010).

(342) See PAUL GRONKE, THE ELECTORATE, THE CAMPAIGN, AND THE OFFICE: A UNIFIED APPROACH TO SENATE AND HOUSE ELECTIONS 165-69 (2000); C. DANIELLE VINSON, LOCAL MEDIA COVERAGE OF CONGRESS AND ITS MEMBERS: THROUGH LOCAL EYES 25 (2003); James E. Campbell et al., Television Markets and Congressional Elections, 9 LEGIS. STUD. Q. 665, 668-69 (1984); Dena Levy & Peverill Squire, Television Markets and the Competitiveness of U.S. House Elections, 25 LEGIS. STUD. Q. 313, 316 (2000); Daniel Lipinski, The Effect of Messages Communicated by Members of Congress: The Impact of Publicizing Votes, 26 LEGIS. STUD. Q. 81, 88-91 (2001); Timothy S. Prinz, Media Markets and Candidate Awareness in House Elections, 1978-1990, 12 POL. COMM. 305, 307 (1995).

(343) See Brian F. Schaffner & Patrick J. Sellers, The Structural Determinants of Local Congressional News Coverage, 20 POL. COMM. 41, 4647 (2003).

(344) See supra note 231 and accompanying text.

(345) See supra subsection II.C.5 and Section IV.C (discussing relevant state case law); see also Jason c. Miller, Community as a Redistricting Principle: Consulting Media Markets in Drawing District Lines, 86 IND. L.J. SUPP. 1, 6 (2010), 86/86_JMiller.pdf (arguing that courts should consult media market boundaries when evaluating district plans).

(346) See GRONKE, supra note 342, at 165-69; Campbell et al., supra note 342, at 668-70; Engstrom, supra note 340, at 74; Levy & Squire, supra note 342, at 316-17; Niemi et al., supra note 339, at 192-99; Schaffner & Sellers, supra note 343, at 43; Winburn & Wagner, supra note 340, at 376-77.

(347) Cf. Pildes & Niemi, supra note 13, at 563-74 (compiling comparative compactness data for congressional districts that was later relied on by Supreme Court racial gerrymandering decisions). A sequel to this Article attempts to measure district-community congruence directly rather than through proxies such as political subdivisions or media markets. See Stephanopoulos, supra note 32.

(348) See Shaw I, 509 U.S. 630, 657 (1993).

(349) See, e.g., id. at 670-74 (White, J., dissenting); id. at 685-87 (Souter, J., dissenting); Aleinikoff & Issacharoff, supra note 57, at 645.

(350) Bush v. Vera, 517 U.S. 952, 1045 (1996) (Souter, J., dissenting).

(351) See Pildes, supra note 21, at 67 & n.174; see also Easley v. Cromartie, 532 U.S. 234, 257-58 (2001).

(352) Pildes, supra note 21, at 68.

(353) Id. at 67-69.

(354) See id. at 70 ("[U]nlike gerrymandering in the Shaw context, partisan gerrymandering raises no concern about overenforcement costs."); cf Richard L. Hasen, The Benefits of "Judicially Unmanageable" Standards in Election Cases Under the Equal Protection Clause, 80 N.C.L. REV. 1469, 1489, 1503 (2002) (calling for more judicially unmanageable election law standards).

(355) See supra notes 22-24 and accompanying text.

(356) See, e.g., Vieth v. Jubelirer, 541 U.S. 267, 289 (2004) (plurality opinion); BUCHMAN, supra note 22, at 202; Issacharoff, supra note 5, at 1693; Lowenstein & Steinberg, supra note 16, at 23; Shapiro, supra note 43, at 238. It should be noted, however, that most of these criticisms are directed at district compactness, which is not the same thing as district-community congruence. See supra subsection I.B.1.

(357) See BOBBITT, supra note 219, at 12-13 (summarizing modalities of constitutional argument).

(358) I examine state legislative rather than congressional elections because (1) many community-respect provisions apply only to the former; and (2) the smaller number of congressional districts in most states makes congressional bias and responsiveness calculations less reliable.

(359) See supra notes 190-94 and accompanying text. California is not counted here because it has not redrawn districts since it adopted its community-respect provision in 2008. Kentucky is not counted because its community-respect provision applies only to congressional redistricting (though I do count it in the turnout and trust-in-government analyses, which are not limited to state elections). North Carolina and Rhode Island are not counted because they do not have explicit community-respect provisions. See supra note 192. The states that are counted are reasonably representative of the country as a whole, covering all major geographic regions as well as a wide array of demographic and socioeconomic environments.

(360) They are Alaska, Colorado, Kansas, Montana, Oregon, and Vermont. See supra notes 201, 203, 205-06, 208, 210 and accompanying text.

(361) See Michael P. McDonald, Redistricting Institutions and Competition in U.S. House Districts ("It may be that despite these criteria being on the books, there is little enforcement of them...."), in DESIGNING DEMOCRATIC GOVERNMENT: MAKING INSTITUTIONS WORK 141, 153 (Margaret Levi et al. eds, 2008).

(362) See, e.g., Jason Barabas & Jennifer Jerit, Redistricting Principles and Racial Representation, 4 ST. POL. & POL'Y Q. 415, 422 (2004) (treating states' compactness and subdivision preservation requirements as proxies for actual district compactness and respect for subdivisions); Pildes & Niemi, supra note 13, at 529-31 (same for compactness). It would be even better, of course, to assess the impact of district-community congruence directly rather than through proxies such as state legal provisions. A sequel to this Article attempts such direct assessment. See Stephanopoulos, supra note 32.

(363) In a recent article, Adam Cox and Richard Holden argue that the optimal gerrymandering strategy for a party is "to match slices of voters from opposite tails of the [partisan] distribution" such that the party's own die-hard supporters in a district slightly outnumber the opposing party's committed partisans. Adam B. Cox & Richard T. Holden, Reconsidering Racial and Partisan Gerrymandering, 78 U. CHI. L. REV. 553, 567 (2011). This "matching slices" strategy plainly cannot be executed when districts must correspond to geographic communities. A requirement of district-community congruence thus renders unavailable the most potent political gerrymanders.

(364) This data is on file with the author. The 2002 elections were the first to be held during the 2000s redistricting cycle, and are thus particularly relevant for purposes of bias and responsiveness. The twenty-six states that Cain and Hanley analyzed account for about 75% of the country's population.

(365) See supra notes 33, 76 and accompanying text.

(366) See Andrew Gelman & Gary King, Enhancing Democracy Through Legislative Redistricting, 88 AM. POL. SCI. REV. 541, 544-45 (1994) (defining bias and responsiveness).

(367) Reducing bias all the way to zero is unproblematic. However, very high rates of responsiveness are undesirable because they result in large changes in seat shares despite only small shifts in vote shares. Fortunately, the responsiveness scores reported here are not nearly high enough to raise such concerns.

(368) These figures are for mean bias and responsiveness. Median bias was 2.5 percentage points lower (5.9% versus 8.4%), while median responsiveness was 0.33 higher (1.34 versus 1.01).

(369) Differences were assessed at the 5% significance level.

(370) I coded community respect as an ordinal variable with the following possible values: 0 if a state had no community-respect provision, 1 if a state had a non-binding community-respect guideline, 2 if a state had a constitutional or statutory community-respect provision, and 3 if a state had a prior applicable court decision. Because I expected (and found) largely linear relationships between community respect and the various dependent variables, I did not convert community respect into dummy variables. The redistricting criteria for which I controlled were compactness, preservation of political subdivisions, and preservation of the cores of prior districts. I obtained data on these criteria from the NCSL and coded them all as dummy variables. See NCSL GUIDE, supra note 193, at 106-08 & tbl.8. All the regressions I ran for this Article used ordinary least squares. For regression results, see the Appendix.

(371) The aspects for which I controlled were whether a state used a redistricting commission, whether a party had unified control of a state government at the time of redistricting, and whether courts ended up drawing a state's district maps. I coded all of these aspects as dummy variables. I obtained commission data from NCSL GUIDE, supra note 193, at 161-62; data on unified party control from Carl Klarner, Klarner Data, ST. POL. & POL'Y Q. (2007), html; and data on court-drawn plans from Jonathan Winburn, Comparing Redistricting Outcomes Across the States: A Comparison of Commission, Court, and Legislative Plans 24 tbl.4 (Mar. 3-4, 2006) (unpublished manuscript), available at http://www.

(372) Data for the remaining twenty-four states and for the rest of the 2000s redistricting cycle would help confirm these conclusions. So too would analysis based directly on district-community congruence rather than proxies for it. See supra note 362; see also Stephanopoulos, supra note 32 (manuscript at 55-57) (finding that district-community congruence has a curvilinear relationship with bias and a positive relationship with responsiveness).

(373) See United States Census 2000: State Legislative District Summary. Files, U.S. CENSUS BUREAU (Jan. 4, 2007), I did not consider coalition or influence districts because of the fierce ongoing debates over how to identify them and how to evaluate them relative to majority-minority districts. Cf. Bartlett v. Strickland, 129 S. Ct. 1231, 1249 (2009) (holding that only majority-minority districts call ever be required under section 2 of the VRA).

(374) Unfortunately, the Census data did not specify minority groups' shares of the voting-eligible (as opposed to total) population in each district. However, these metrics are usually highly correlated.

(375) I separately calculated the proportion of majority-minority districts in each state legislative chamber, and then averaged the state house and state senate figures to obtain a single value for each state.

(376) Cf. Johnson v. De Grandy, 512 U.S. 997, 1014 (1994) (evaluating minority representation in similar fashion). I do not claim that the gap should necessarily be zero. I only compare states' proportions of majority-minority districts to their minority population percentages so that I can have a benchmark by which to assess those proportions. In addition, my analysis addresses descriptive representation only; it has nothing to say about the caliber of minorities' substantive representation.

(377) The gap increases to 3.2 percentage points if states with less than a 10% minority population (in which it is very difficult to draw majority-minority districts) are excluded from the analysis. The gap changes to 2.3 percentage points if median figures are used.

(378) The difference was assessed at the 5% significance level.

(379) See supra notes 4243 and accompanying text.

(380) See Michael McDonald, Voter Turnout, U.S. ELECTIONS PROJECT, http://elections. (last visited Feb. 15, 2012). I used turnout data for all fifty states, for the highest office on the ballot, and from 2002 to 2008, in my analysis.

(381) See Codebook Variable Documentation File, AM. NAT'L ELECTION STUD. 616 (Aug. 25, 2011), Data was available for forty-seven states and for the years 2002, 2004, and 2008.

(382) These figures are for mean turnout and trust in government. Median turnout was 6.6 percentage points higher (55.9% versus 49.3%), while median trust in government was 5.1 points higher (38.4 points versus 33.3 points).

(383) Differences were assessed at the 5% significance level.

(384) See supra note 370.

(385) The demographic variables for which I controlled were household income, percent of population over 65, percent of population that is black, percent of population that is Hispanic, and percent of population with a bachelor's degree. I coded them all as continuous variables. I obtained income data from State Median Income, U.S. CENSUS BUREAU, (last modified Nov. 1, 2011); age data from Age Data of the United States, U.S. CENSUS BUREAU,; race data from Population by Sex, Race, and Hispanic Origin, U.S. CENSUS BUREAU, http://www.census. gov/popest/data/state/asrh/2009/SC-EST2009-03.html (last visited Feb. 15, 2012); and education data from SARAH R. CRISSEY, U.S. CENSUS BUREAU, EDUCATIONAL ATTAINMENT IN THE UNITED STATES: 2007 (2009), available at prod/2009pubs/p20-560.pdf. For regression results, see the Appendix.

(386) See supra note 362 (noting that conclusions would be bolstered by direct analysis of district-community congruence); see also Stephanopoulos, supra note 32 (manuscript at 35-36) (finding that rate of voter roll-off is lower in districts that correspond more closely to communities).

(387) See Roland G. Fryer, Jr. & Richard Holden, Measuring the Compactness of Political Districting Plans (July 16, 2009) (unpublished manuscript), available at http://

(388) See id. at 2 n.5.

(389) See id. at 51 tbl.2. But see Jowei Chen & Jonathan Roddeu, Using Legislative Districting Simulations To Measure Electoral Bias in Legislatures 6, 14 (July 15, 2010) (unpublished manuscript), available at florida.pdf (finding that redrawing Florida districts on basis of compactness would result in significant pro-Republican bias).

(390) See MICHAEL P. MCDONALD, MIDWEST MAPPING PROJECT, available at http://

(391) I calculated these averages myself using the data compiled by McDonald in the Midwest Mapping Project.

(392) See also DOUGLAS JOHNSON, THE ROSE INST. OF STATE & LOCAL GOV'T, COMPETITIVE DISTRICTS IN CALIFORNIA: A CASE STUDY OF CALIFORNIA'S REDISTRICTING IN THE 1990s 7-9 (2005), available at pdf/rose_ca_case_study.pdf (finding high level of competitiveness when California's districts were drawn in the 1990s to correspond to geographic communities).

(393) Edward L. Glaeser & Bryce A. Ward, Myths and Realities of American Political Geography 6 (Harvard Inst. of Econ. Research Discussion Paper No. 2100, Nov. 23, 2005), available at; see also Philip A. Klinkner, Red and Blue Scare: The Continuing Diversity of the American Electoral Landscape, 2 FORUM, no. 2, 2004, at 1, 7, (finding that "index of exposure" in 2000 was almost identical for Democrats and Republicans, "mean[ing] that the average Democrat and the average Republican lived in a county that was close to evenly divided"); cf. MORRILL, supra note 58, at 21 (noting that compactness requirement would have random partisan effects because both parties' supporters are equally spatially concentrated); Polsby & Popper, supra note 13, at 334-35 (same).

(394) See Morrill, supra note 44, at 225-26.


(396) Id. at 200-02.

(397) See GRONKE, supra note 342, at 23, 97; Campbell et al., supra note 342, at 671-73; Levy & Squire, supra note 342, at 317-19; Niemi et al., supra note 339, at 193; Prinz, supra note 342, at 310-12.

(398) Campbell et al., supra note 342, at 673-74.

(399) Levy & Squire, supra note 342, at 321-23.

(400) Barahas & Jerit, supra note 362, at 423. In contrast, compactness requirements reduced the numbers of both majority-minority and minority-influence districts. Id. at 429; see also Carmen Cirincione et al., Assessing South Carolina's 1990s Congressional Districting, 19 POL. GEOGRAPHY 189, 201-02 (2000) (finding that redrawing South Carolina's congressional districts so as to minimize county splits would likely result in one minority-influence district and no majority-minority districts).

(401) See W. Mark Crain, The Constitutionality of Race-Conscious Redistricting: An Empirical Analysis, 30 J. LEG. STUD. 193, 207 fig.3 & 208 fig.4 (2001); see also id. at 209-10 (finding strong positive relationship between minority percentage in district and district's proportion of split Census places).

(402) Campbell et al., supra note 342, at 672. The larger increase for challengers ex plains why they do better in high-congruence districts. See supra notes 397-99 and accompanying text.

(403) See Niemi et al., supra note 339, at 192 tbl.1.

(404) Lipinski, supra note 342, at 93; see also VINSON, Supra note 342, at 44; Engstrom, supra note 340, at 78; Levy & Squire, supra note 342, at 319; Prinz, supra note 342, at 310-12; Schaffner & Sellers, supra note 343, at 52-53; Winburn & Wagner, supra note 340, at 378. These quantitative findings are corroborated by the dozens of in-depth interviews that Malcolm Jewell conducted with state legislators across the country. The politicians told Jewell that voters are often confused by district lines that disregard community boundaries, and that it is difficult for representatives to learn about and address their constituents' problems when districts and communities do not correspond. See JEWELL, supra note 46, at 55-60, 168.

(405) Engstrom, supra note 340, at 74-75, 77.

(406) Winburn & Wagner, supra note 340, at 382.

(407) Altman, supra note 91, at 333-34.

(408) COURTNEY, supra note 85, at 210-11; see also DAVID E. CAMPBELL, WHYWE VOTE: HOW SCHOOLS AND COMMUNITIES SHAPE OUR CIVIC LIFE 23-24, 42 (2006) (finding that turnout is higher in economically, ethnically, politically, and racially homogeneous communities); Michael S. Kang, Race and Democratic Contestation, 117 YALE L.J. 734, 78486 (2008) (summarizing political science literature showing that minority turnout is higher in majority-minority districts).

NICHOLAS O. STEPHANOPOULOS, Associate-in-Law, Postdoctoral Research Scholar, and Lecturer-in-Law, Columbia Law School. I am deeply indebted to all the people who read drafts of this Article and, at various stages, offered helpful comments and criticism: Jessica Amunson, Joseph Blocher, Richard Briffault, Bruce Cain, Adam Cox, Christopher Elmendorf, Joey Fishkin, Owen Fiss, David Fontana, James Gardner, Heather Gerken, Bernard Grofman, Philip Hamburger, Rick Hasen, Gerry Hebert, Samuel Issacharoff, Sanford Levinson, Justin Levitt, Trevor Morrison, Daniel Ortiz, Nathaniel Persily, Dean Robert Post, Paul Smith, Matthew Stephenson, et al. My thanks also to the workshop participants at Columbia Law School, Yale Law School, and the Law and Society Conference, where I presented earlier versions of the Article.
Figure 1: Average Partisan Bias by State Category

All Non-Community States       9.4%

Non-Binding Guideline States   7.7%

All Community States           5.4%

Constitution/Statute States    4.1%

Previous Case States           3.6%

Note: Table made from bar graph.

Figure 2: Average Electoral Responsiveness by State Category

All Non-Community States       1.03

Previous Case States           1.30

Non-Binding Guidelines States  1.33

All Community States           1.43

Constitution/Statute States    1.49

Note: Table made from bar graph.

Figure 3: Average Discrepancy Between Proportion of Majority-Minority
Districts and Minority Population
Percentage by State Category

All Non-Community States       13.5%

Constitution/Statute States    12.7%

Previous Case States           11.8%

All Community States           10.8%

Non-Binding Guideline States    8.9%

Note: Table made from bar graph.

Figure 4: Average Voter Turnout by State Category

All Non-Community States       50.6%

Non-Binding Guideline States   54.2%

All Community States           55.5%

Constitution/Statute States    56.9%

Previous Case States           58.3%

Note: Table made from bar graph.

Figure 5: Average Trust in Government by State Category

All Non-Community States       33.2

Non-Binding Guideline States   37.8

Previous Case States           37.9

All Community States           39.3

Constitution/Statute States    41.3

Note: Table made from bar graph.
COPYRIGHT 2012 University of Pennsylvania, Law School
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2012 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Stephanopoulos, Nicholas O.
Publication:University of Pennsylvania Law Review
Date:Apr 1, 2012
Previous Article:Adaptable due process.
Next Article:Toward a constitutional Chevron: lessons from Rapanos.

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters