Emory Law Journal

Partisan Gerrymandering and the Constitutionalization of Statistics
Jacob Eisler *Associate Professor of Public Law, University of Southampton Law School; Research Associate, Centre for Business Research, University of Cambridge. PhD, Harvard University Department of Government; JD, Harvard Law School. Many thanks to Michael Kang, George Brown, Jim Gardner, Rick Pildes, Richard Briffault, Adriaan Lanni, Bruce Mann, Eric Beerbohm, Billy Magnuson, Greg Conti, Jeremy Green, Adam Lebovitz, Evan Miller, and Sam Zeitlin. The ideas and flaws in this Article are my own.

Data analysis has transformed the legal academy and is now poised to do the same to constitutional law. In the latest round of partisan gerrymandering litigation, lower courts have used quantitative tests to define rights violations and strike down legislative districtings across the country. The Supreme Court’s most recent opinion on partisan gerrymandering, Gill v. Whitford, hinted that quantitative tests may yet define the constitutionality of partisan gerrymandering. Statistical thresholds thus could be enshrined as constitutional protections and courts recast as agents of discretionary policy.

This Article describes how excessive dependence on metrics transforms judicial decision-making and undermines rights enforcement. Courts enforce constitutional law to ensure governmental compliance with rights, not to advance alternative policy arrangements. Yet the core of rights is moral principle, not descriptive conditions in the world. If quantitative outcomes are used to define rights, the moral character of judicial rights enforcement is undermined, and courts act as quasi-regulatory entities that compete with democratically elected branches. Arguably the most condemned decision of the twentieth century, Lochner, reflected such a quasi-regulatory approach to rights enforcement; excessive reliance on statistics threatens to repeat that mistake.

The law of partisan gerrymandering needs a new principle, not new metrics. The best principle to identify partisan gerrymandering is the right to fair representation, which is violated when legislatures seize partisan advantage in democratic process. Quantitative analysis should have the sole function of proving that alleged partisan gerrymanders seek such advantage.

This Article thus identifies a novel and troubling trend in constitutional law and describes how it dominates a topic of immediate practical importance. It then offers a general framework for conceptualizing rights protection and applies it to this pressing doctrinal issue.

 

Introduction

It is the age of statistics. The quantitative analysis of data is transforming domains as diverse as romance, 1See, e.g., Eli J. Finkel et al., Online Dating: A Critical Analysis from the Perspective of Psychological Science, 13 Psychol. Sci. Pub. Int. 3, 6 (describing the role of algorithmic matching in online dating). finance, 2See, e.g., William Magnuson, Regulating Fintech, 71 Vand. L. Rev. 1167, 1188 (2018) (describing the systemic risks of various quantitative approaches to trading). and professional sports, 3See, e.g., Michael Lewis, Moneyball: The Art of Winning an Unfair Game (2003) (the seminal account of how sabermetrics and quantitative analysis in baseball transformed the approach to management strategy and tactics). and law cannot resist its influence. 4 Adam Chilton & Eric Posner, The Rise of Statistics in Law, Eric Posner (July 24, 2015), http://ericposner.com/the-rise-of-statistics-in-law/ (detailing one of the simplest metrics: the increased prevalence of the phrase “statistical significance” and its variant in legal settings). While quantitative metrics have long played a central role in areas of law such as contracts, 5See, e.g., Eric A. Posner, Economic Analysis of Contract Law After Three Decades: Success or Failure?, 112 Yale L.J. 829, 829 (2003) (“Traditional doctrinal analysis [in contract law] exerts less influence . . . and enjoys little prestige [since 1970].”). securities regulation, 6See, e.g., James Bohn & Stephen Choi, Fraud in the New-Issues Market: Empirical Evidence on Securities Class Actions, 144 U. Pa. L. Rev. 903, 912 (1996) (statistical analysis of impact of pursuing different types of suits); Bradford Cornell & R. Gregory Morgan, Using Finance Theory to Measure Damages in Fraud on the Market Cases, 37 UCLA L. Rev. 883, 885–87 (1990) (doctrinal relevance of quantitative methods to identify appropriate measure of damages in Rule 10b-5 suits). and antitrust, 7See, e.g., Einer Elhauge, Defining Better Monopolization Standards, 56 Stan. L. Rev. 253, 343 (2003); Thomas G. Krattenmaker & Steven C. Salop, Anticompetitive Exclusion: Raising Rivals’ Costs to Achieve Power over Price, 96 Yale L.J. 209, 250 (1986) (providing tests for anticompetitive behavior that weaves together principles of competition with certain quantified levels of market power to identify actionable harm). enterprising scholars have applied data analysis to constitutional law and related topics. 8 For an exemplar of such scholarship, see Nicholas O. Stephanopoulos & Mila Versteeg, The Contours of Constitutional Approval, 94 Wash. U. L. Rev. 113, 115–16, 143–45, 147–53 (2016) (using statistical analysis of surveys to assess the effect and validity of different types of constitutional designs). For a demonstration of such methods, see William Baude et al., Making Doctrinal Work More Rigorous: Lessons from Systematic Reviews, 84 U. Chi. L. Rev. 37 (2016) (using empirical methods to assess the validity of legal approaches to doctrinal interpretation). For a prescient anticipation of the current trends, see Michael Heise, The Past, Present, and Future of Empirical Legal Scholarship: Judicial Decision Making and the New Empiricism, 2002 U. Ill. L. Rev. 819 (2002). As constitutional law has typically been the domain of textual analysis, moral philosophy, and legal history, 9See, e.g., Antonin Scalia, Originalism: The Lesser Evil, 57 U. Cin. L. Rev. 849, 852 (1989) (describing analysis of history and context as the appropriate basis of constitutional analysis and rejecting moral intuitions as the alternatives); Seana Valentine Shiffrin, Inducing Moral Deliberation: On the Occasional Virtues of Fog, 123 Harv. L. Rev. 1214, 1216 (2010) (intervening in this debate through the use of deliberative moral philosophy); Kathleen M. Sullivan, Foreword: The Justices of Rules and Standards, 106 Harv. L. Rev. 22, 112 (1992) (observing the basic struggle on the bench of constitutional interpretation to be standards versus rules, based in differing moral visions of the rule of law). this shift is demonstrative of the prevalence of data analytic methods.

The quantitative treatment of constitutional law is now poised to jump from the academy to the highest levels of legal precedent. With the current round of partisan gerrymandering litigation, anti-gerrymandering reformers have relied heavily on novel quantitative tools to provide standards that can identify illegal partisan gerrymanders. Partisan gerrymandering allocates voters to legislative districts by the voters’ political affiliations. 10 For a seminal description of this practice, see Samuel Issacharoff & Pamela S. Karlan, Where to Draw the Line?: Judicial Review of Political Gerrymanders, 153 U. Pa. L. Rev. 541, 551–53 (2004). Critics of partisan gerrymandering have long argued that the practice undercuts representative fairness, impairs democratic accountability, and allows for corrupt entrenchment by the party in power. 11See, e.g., Samuel Issacharoff, Gerrymandering and Political Cartels, 116 Harv. L. Rev. 593, 600 (2002) (arguing that partisan gerrymandering can be a practice that inhibits “accountability to shifting voter preferences”); Michael J. Klarman, Majoritarian Judicial Review: The Entrenchment Problem, 85 Geo. L.J. 491, 516 (1997) (partisan gerrymandering is “indefensibly antimajoritarian”); Daryl Levinson & Benjamin I. Sachs, Political Entrenchment and Public Law, 125 Yale L.J. 400, 416–18 (2015) (partisan gerrymandering distorts electoral outcomes). However, since Vieth v. Jubelirer, a plurality of Justices have denied that courts can identify coherent standards for prohibiting partisan gerrymandering, and thus that the legitimacy of gerrymanders is a nonjusticiable political question. 12 541 U.S. 267 (2004). Reformers offer a methodological solution in the latest round of litigation, arguing that quantitative tools can provide courts with dispositive indications of when illicit gerrymandering occurs. 13 Jowei Chen & Jonathan Rodden, Cutting Through the Thicket: Redistricting Simulations and the Detection of Partisan Gerrymanders, 14 Election L.J. 331 (2015) (using simulations to assess partisan gerrymandering); Bernard Grofman & Gary King, The Future of Partisan Symmetry as a Judicial Test for Partisan Gerrymandering After LULAC v. Perry, 6 Election L.J. 1, 5 (2007) (using partisan symmetry, which compares the ratio of votes received to the number of seats per party, to assess partisan gerrymandering); Nicholas O. Stephanopoulos & Eric M. McGhee, Partisan Gerrymandering and the Efficiency Gap, 82 U. Chi. L. Rev. 831, 834 (2015) (using “wasted votes” as the basis for assessing partisan gerrymandering); see also Expert Report and Affidavit Assessing the Current Wisconsin State Legislative Districting Plan of Simon Jackman, Whitford v. Nichol, No. 315CV00421, 2015 WL 10091020 (W.D. Wis. July 7, 2015). For a description of the role quantitative methods played in the Pennsylvania Supreme Court decisions, see Issie Lapowsky, The Geeks Who Put a Stop to Pennsylvania’s Partisan Gerrymandering, Wired (Jan. 20, 2018, 11:21 AM), https://www.wired.com/story/pennsylvania-partisan-gerrymandering-experts/.

Reformers have achieved some resounding successes in the lower courts in the current round of partisan gerrymandering litigation. 14 Common Cause v. Rucho, 279 F. Supp. 3d 587, 608 (M.D.N.C.), vacated, 138 S. Ct. 2679 (2018); Whitford v. Gill, 218 F. Supp. 3d 837, 927 (W.D. Wis. 2016); League of Women Voters of Pa. v. Commonwealth, 178 A.3d 737, 741 (Pa. 2018). However, lower court judicial holdings reveal a remarkable pattern: where courts find statistics convincing, they find gerrymanders unconstitutional. 15See infra Section I.B. Where claimants have not relied on statistics, or where courts have found statistics to have little bearing on the constitutional questions, courts have not found gerrymanders to be illegal. 16See, e.g., Benisek v. Lamone, 266 F. Supp. 3d 799, 801 (D. Md. 2017). Moreover, even where reformers have used statistical indicia to convince courts that a gerrymander is illegal, neither the courts nor the reformers have clearly linked the metrics to constitutional doctrine. When the first of these cases, Gill v. Whitford, reached the Supreme Court, the Justices left open the question of whether statistics could inform constitutional rights. 17 138 S. Ct. 1916 (2018). The majority opinion stated that “two threshold questions remain: what is necessary to show standing in a case of this sort, and whether those claims are justiciable. Here we do not decide the latter question because the plaintiffs in this case have no shown standing under the theory upon which they based their claims for relief.” Id. at 1929. The law of partisan gerrymandering has thus evolved into a referendum on whether metrics can define constitutional law.

Judicial adoption of a radically new definition of rights as quantitative outcomes would be novel and problematic. It would transform the role of statistical analysis from providing evidence of rights violations to defining the content of rights. 18 Of course, even the evidentiary role of statistics in Supreme Court jurisprudence has been a matter of dispute, most notably in the case McCleskey v. Kemp. 481 U.S. 279, 280 (1987) (disregarding statistics regarding racial bias in sentencing); see Ruth E. Friedman, Statistics and Death: The Conspicuous Role of Race Bias in the Administration of the Death Penalty, 11 La Raza L.J. 75, 82 (1999) (observing the role of statistics in constitutional cases related to employment discrimination and jury selection but not death penalty sentencing). Government conduct might be lawful or unlawful depending upon (non)conformity to metrical tests. This would distort the role and nature of constitutional law. Rights are best understood as creating zones of protection that provide non-conditional weight to certain characteristics or activities. 19 The specific language of “trumps” is taken from Ronald Dworkin, Rights as Trumps, in Theories of Rights 154–59 (Jeremy Waldron ed., 1984). The link to Dworkin’s broader theory is described in his books Law’s Empire and Taking Rights Seriously. Ronald Dworkin, Law’s Empire 221–24 (1986) [hereinafter Dworkin, Empire]; Ronald Dworkin, Taking Rights Seriously 91, 108 (1977) [hereinafter Dworkin, Rights] (describing the distinction between rights and goals, and their implications for judicial lawmaking in constitutional interpretation); see infra Part II. For example, the right to free speech or of freedom from illicit racial classification are enforced by courts whenever the government action in question infringes upon the zone of constitutional protection. 20See infra Section III.B. The invocation of such right does not hang upon the effect of the challenged government action, 21 The classic doctrinal expression of this is the Court’s refusal to independently find conduct unconstitutional solely on the basis of disparate impact without indication of discriminatory intent, and to “leav[e] the choice whether to impose disparate impact standards to legislators.” Richard A. Primus, Equal Protection and Disparate Impact: Round Three, 117 Harv. L. Rev. 494, 495 (2003). The critical case for determining this issue is Washington v. Davis, 426 U.S. 229, 239 (1976) (“We have never held that the constitutional standard for adjudicating claims of invidious racial discrimination is identical to the standards applicable under Title VII, and we decline to do so today. . . . [O]ur cases have not embraced the proposition that a law or other official act, without regard to whether it reflects a racially discriminatory purpose, is unconstitutional solely because it has a racially disproportionate impact.”). Were there a law prohibiting partisan gerrymandering, the legal field would be far different. The most salient expression of this in the doctrine may be Arizona State Legislature v. Arizona Independent Redistricting Commission, 135 S. Ct. 2652, 2659 (2015), which held that a referendum implementing districting through an independent commission was constitutional. but rather only requires that the government action intersects a protected characteristic. 22 In its most general form, the principle of rule of law is a very broad one of governmental impartiality in all its actions. Dworkin, Empire, supra note 19, at 174 (“[O]fficials acting in their official capacity . . . must, we say, treat all members of their community as equals, and the individual’s normal latitude for self-preference is called corruption in their case.”). In the context of constitutional protection, this takes the form of “impartial justice. Courts should be available to enforce the law and should employ fair procedures.” Richard H. Fallon, Jr., “The Rule of Law” as a Concept in Constitutional Discourse, 97 Colum. L. Rev. 1, 9 (1997).

Conversely, rights defined by quantitative outcomes would turn courts into enforcers of policy outcomes. If courts identify constitutional wrongs whenever certain metrical thresholds are breached, they act as regulators who have concluded that certain outcomes are desirable. 23Chevron, U.S.A., Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837 (1984), formalized the judicial deference to legislative agencies. See Thomas W. Merrill & Kristin E. Hickman, Chevron’s Domain, 89 Geo. L.J. 833, 833–35 (2001). For example, if courts were to determine that only certain tax rates were constitutional (as opposed to merely requiring that taxes be applied in a non-discriminatory manner), 24See Harper v. Va. Bd. of Elections, 383 U.S. 663, 668 (1966) (poll taxes held illegal, because, like race, overt government categorization based on wealth is disfavored). they would act as an agency that set tax policy. In the partisan gerrymandering context, if courts were to conclude that fair representation required that districtings reflect the composition of voters by party identification, this would be a judicially enforced policy of proportional representation. 25 Nathaniel Persily, Reply: In Defense of Foxes Guarding Henhouses: The Case for Judicial Acquiescence to Incumbent-Protecting Gerrymanders, 116 Harv. L. Rev. 649, 650 (2002); Peter H. Schuck, The Thickest Thicket: Partisan Gerrymandering and Judicial Regulation of Politics, 87 Colum. L. Rev. 1325, 1350 (1987). That the current litigation has invoked more complex quantitative indicia does not make the use of metrics to define constitutional rights any less a form of judicial policy enforcement.

Respecting the nature of rights while giving due respect to the usefulness of quantitative methods requires careful cautious constitutional innovation. It is the intersection with protected categories, not the practical effect of the government action, that activates rights. Quantitative data can only serve as evidence that the government conduct touches upon protected categories. Because the role of quantitative data is limited to evidence for a wrong, it cannot be constitutive of the right itself.

Partisan gerrymandering provides an ideal case study for interpreting rights in a manner that recognizes the role of such quantitative evidence. As Professor Richard Briffault has observed, “[t]he Justices have bounced back and forth between the question of justiciability and the standards of proving partisan gerrymandering without addressing, at any length, which constitutional provision or norm gerrymandering might violate.” 26 Richard Briffault, Defining the Constitutional Question in Partisan Gerrymandering, 14 Cornell J.L. & Pub. Pol’y 397, 399 (2005); see also Davis v. Bandemer, 478 U.S. 109, 157–58 (1986) (O’Connor, J., concurring) (“In order to implement the plurality’s standard, it will thus be necessary for courts to adopt an analogous norm” to the proportionality norm applied in the racial districting jurisprudence, which would comprise “the use of the Equal Protection Clause as the vehicle for making a fundamental policy choice that is contrary to the intent of its Framers and to the traditions of this Republic.”). Without a clear relationship to a legal principle, the efficiency gap, partisan symmetry, or any other metric is constitutionally empty. If courts define illegal partisan gerrymandering by metrics, it would produce fragile precedent and betray the rule of law. 27See Luis Fuentes-Rohwer, Doing Our Politics in Court: Gerrymandering, “Fair Representation” and an Exegesis into the Judicial Role, 78 Notre Dame L. Rev. 527, 538 (2003) (“When courts intervene, they simply take sides in highly politicized debates. They do very little else.”); Schuck, supra note 25. As Heather K. Gerken, Lost in the Political Thicket: The Court, Election Law, and the Doctrinal Interregnum, 153 U. Pa. L. Rev. 503, 514 (2004), has observed, courts strain to avoid doing so explicitly, preferring a “studied agnosticism” in election law cases even when their decisions reveal clear substantive impacts. Yet disregarding the information provided by the new quantitative methods would deprive courts of powerful evidentiary tools. It would also allow those executing partisan gerrymanders to continue to deploy increasingly sophisticated methods without effective judicial oversight.

This Article culminates by advancing a principled right against partisan gerrymandering that allows for appropriate use of quantitative methods. The legislative intention of partisan gerrymandering—to achieve a distortive advantage in elections—enables identification of the state conduct that violates the constitutional right to fair representation: predomination of districting in the pursuit of partisan advantage in future elections. Such a districting offends the principle of political self-determination that courts have defended since Baker v. Carr28 369 U.S. 186, 208 (1962) (“A citizen’s right to a vote free of arbitrary impairment by state action has been judicially recognized as a right secured by the Constitution . . . .”); see Guy-Uriel E. Charles, Constitutional Pluralism and Democratic Politics: Reflections on the Interpretive Approach of Baker v. Carr, 80 N.C. L. Rev. 1103, 1104 (2002) (observing that Baker v. Carr has been recognized as successfully constitutionalizing democracy). This analysis also indicates the correct role of quantitative evidence in partisan gerrymandering litigation. Quantitative evidence is valuable as an indication that a districting was predominated by a governmental intention to achieve a partisan advantage in elections, rather than by service to neutral districting criteria. 29 The neutral districting criteria are described at Karcher v. Daggett, 462 U.S. 725, 740 (1983), and are central to the methods used by Chen & Rodden, supra note 13.

This Article proceeds as follows. Part I provides an overview of the state of the law on partisan gerrymandering. It describes the centrality of quantitative analysis to the current litigation, the remarkable sequence of events by which such methods have entered the legal debate, and how the appropriate search for a legal principle has been neglected. Part II describes the underlying error of identifying rights by quantitative outcomes. It demonstrates that legal rights are necessarily a matter of principle, whereas outcomes are a matter of policy best left to elected branches. Part III explores how evidence should be used given the nature of rights as principled and explores the pitfalls of misconstruing this characteristic of rights in the partisan gerrymandering context. Part IV uses this theory of rights and quantification to review the existing principles that have been leveraged against partisan gerrymandering and then innovates a new justification for judicial intervention, a principle against excessive partisan competition.

I. Technical Gerrymanders and Unclear Precedent

The fevered condition of partisan gerrymandering litigation can be attributed to the intersection of three factors: the application of sophisticated data analysis to ruthlessly gerrymander legislative districts; the use of similar tools by leading anti-gerrymandering reformers and courts to identify and classify such politicized districting; and the confused state of the partisan gerrymandering precedent thanks to Davis v. Bandemer 30 478 U.S. 109 (1986). and Vieth v. Jubelirer31 541 U.S. 267 (2004). Lack of clear standards regarding when partisan gerrymanders are illicit resulted in lower courts hesitating to strike down politicized districting plans at all. This in turn seemingly gave carte blanche to parties in power to aggressively deploy statistical methods to implement increasingly effective gerrymanders. 32 Common Cause v. Rucho, 279 F. Supp. 3d 587, 642–43 (M.D.N.C.) (noting the effectiveness of the gerrymander in question), vacated, 138 S. Ct. 2679 (2018); Whitford v. Gill, 218 F. Supp. 3d 837, 847–50 (W.D. Wis. 2016). Yet because the law remains that sufficiently egregious gerrymanders might be illegal (a point that, remarkably, received no additional illumination in Whitford), reformers have turned to similar data analysis to demonstrate the unfairness of gerrymanders. The missing piece of this puzzle, however, is the principle that explains how this analysis relates to legal rights. This section reviews the state of the doctrine and the new developments, particularly the technological arms race between legislatures and reformers.

A. The Modern History of Partisan Gerrymandering

Partisan gerrymandering consists of the drawing of political districts in a manner that takes the party identity of voters into account. In its most typical form, 33 A less blatantly partisan form of the practice is “shacking,” or the use of districting to entrench incumbents in their home districts. Issacharoff & Karlan, supra note 10, at 551–52; cf. Gaffney v. Cummings, 412 U.S. 735, 738 (1973) (suggesting that it is the illicit role of partisanship, not the entrenchment effect, that is actionable). it is done to benefit the party that controls the legislature, and thus has the power to draw district lines. In such cases, the party uses tactics such as the “cracking” of districts to give the gerrymandering party a disproportionate number of thin majorities and the “packing” of voters from the minority party into a few districts where they have disproportionate super-majorities. 34See Bruce E. Cain, The Reapportionment Puzzle 148–50 (1984); Issacharoff & Karlan, supra note 10, at 551. The result is a set of representatives elected by districts that have thin majorities from the gerrymandering party and large majorities from the disadvantaged party, with the gerrymandering party receiving disproportionate representation. Parties often use partisan gerrymandering to seek prolonged entrenchment. 35League of United Latin American Citizens v. Perry (LULAC), 548 U.S. 399 (2006), provides one of the most vivid illustrations of how parties use gerrymandering to attempt to hold on to political power even after popular opinion has changed.

1. Bandemer and Partisan Gerrymandering as a New Frontier

Though partisan gerrymandering has a long history as a practice in America—it is named after an 1812 districting by the eponymous Eldridge Gerry that favored Republicans over Federalists 36See Edward B. Foley, Due Process, Fair Play, and Excessive Partisanship: A New Principle for Judicial Review of Election Laws, 84 U. Chi. L. Rev. 655, 711–12 (2017) (briefly describing the history of the gerrymander).—it is a relative newcomer as a constitutional wrong. The prospective illegality of partisan gerrymandering was first substantively addressed in Davis v. Bandemer37 While Bandemer did not weave the idea that discriminatory politicized districting was illegal out of thin air, none of the prior cases provided detailed guidance on how districting on the basis of party affiliation should be managed. Davis v. Bandemer, 478 U.S. 109, 119–20 (1986) (quoting Fortson v. Dorsey, 379 U.S. 433, 439 (1965)). The case evaluated the redistricting that followed the 1980 census in Indiana which adopted a partisan gerrymander that disproportionately benefited Republicans. 38Id. at 113.

A plurality of the Supreme Court concluded that partisan gerrymandering could violate the Equal Protection Clause, but that the districting in question did not “visit[] a sufficiently adverse effect on the appellees’ constitutionally protected rights to make out a violation of the Equal Protection Clause.” 39Id. at 129. The Bandemer plurality indicated that both discriminatory intent and discriminatory effect are necessary to establish an illegal partisan gerrymander. 40Id. at 127. While there has been some debate over the precise contours of the intent prong, 41 The Vieth plurality raised the question of whether the intent must be to gerrymander a given district, or the entire state, and the degree of predominance required to gerrymander. Vieth v. Jubelirer, 541 U.S. 267, 284–85 (2004). Conversely, Justice Powell’s dissent in Bandemer, 478 U.S. at 179–80 (Powell, J., dissenting), suggests that intent alone may be enough to establish a constitutional violation, a sentiment shared by Justice Souter’s dissent in Vieth, 541 U.S. at 350 (Souter, J., dissenting). However, as described in Section III.A.2 infra, a pure intent test, while perhaps easier to operationalize, would affect dramatically wide-ranging judicial regulation of politics. the crux of the debate has been whether discriminatory effect of partisan gerrymandering can be consistently identified. 42See Vieth, 541 U.S. at 282–84 (reviewing lower court opinions and academic sources, and reflecting their perception of the discriminatory effects test as deeply problematic). The plurality offered a test that was both demanding and vague:

[A]n equal protection violation may be found only where the electoral system substantially disadvantages certain voters in their opportunity to influence the political process effectively . . . . [S]uch a finding of unconstitutionality must be supported by evidence of continued frustration of the will of a majority of the voters or effective denial to a minority of voters of a fair chance to influence the political process. 43Bandemer, 478 U.S. at 128, 133.

This has led the law into tangled territory. The Bandemer plurality primarily states its claims regarding how to identify the effects of illicit partisan gerrymandering in the negative. Districting becomes illicitly political when the electoral process is unresponsive to certain trends in voter preference. Yet since no party has the right to proportional representation, nor the right to an equal chance to select its desired representative, 44Id. at 131–32. the readily identifiable form of such unresponsiveness cannot be the discriminatory effect. 45 For a theorized explanation of why this must be so, see id. at 158 (O’Connor, J., concurring) (observing that such a move would dictate significant elements of design of the political system beyond the judicial ken); James A. Gardner, Madison’s Hope: Virtue, Self-Interest, and the Design of Electoral Systems, 86 Iowa L. Rev. 87, 104 (2000) (observing that the absence of such features might be part of a chosen political design, particularly in a first-past-the-post system); Schuck, supra note 25 (same).

The deficiency of this formulation is that it fails to link equal protection rights to the circumstances it deems pathological. It suggests that the Equal Protection Clause guarantees, as a general matter, some right to a baseline of political influence based on party identity. However, unlike the idea of racial vote dilution (which provides protection against reduction of voting power based on racial discrimination to any degree46 Gray v. Sanders, 372 U.S. 368, 379 (1963) (stating that “all who participate in the election are to have an equal vote—whatever their race . . . .This is required by the Equal Protection Clause of the Fourteenth Amendment.”). or one-person, one-vote (which universalizes the concept of personal equality of citizens and thus requires equipopulous voting districts), 47See Wesberry v. Sanders, 376 U.S. 1, 4 (1964); Reynolds v. Sims, 377 U.S. 533, 577 (1964); cf. Adam Raviv, Unsafe Harbors: One Person, One Vote and Partisan Redistricting, 7 U. Pa. J. Const. L. 1001, 1012–13 (2005) (describing how the 10% safe harbor for state districting plans has snuck into the jurisprudence). it is unclear what is required for “fair representation” 48Bandemer, 478 U.S at 123 (“[T]he Court characterized the question posed by election districts of disparate size as an issue of fair representation.”). in the context of partisan competition over districting. Because proportional representation is not a principle enshrined in the U.S. representative system, a lack of proportionality by party identity does not form a sufficient basis for concluding that a politicized districting is illicit. Conversely, because partisan identity is not an attribute that activates a heightened level of constitutional scrutiny like race, courts cannot identify an illicit partisan gerrymander simply by intent to advance a partisan goal (or disadvantage a partisan opponent), at least not without the dramatic act of innovating a new protected class. 49 The main arguments against this involve the claim that this class identity is fluid and thus related to the basic content of political contestation. For the latter claim, see Vieth v. Jubelirer, 541 U.S. 267, 288–89 (2004) (noting the non-determinativeness of party identity). For the argument that it is a function of political contestation, see generally Daniel H. Lowenstein & Jonathan Steinberg, The Quest for Legislative Districting in the Public Interest: Elusive or Illusory, 33 UCLA L. Rev. 1, 74–76 (1985). As the Bandemer plurality’s vague test suggests, without clear touchstones as to when a certain level of partisan discrimination is too much, courts risk making arbitrary or intuitive judgments.

2. Vieth and the Disintegration of the Doctrine

Twenty years of lower court litigation did little to produce consensus regarding the evaluation of partisan gerrymanders, a fact made painfully apparent by the split of opinions in Vieth v. Jubelirer50 In Vieth, a plurality, a concurring opinion, and three dissenting opinions were written. 541 U.S. at 271 (majority opinion); id. at 306 (Kennedy, J., concurring); id. at 317 (Stevens, J., dissenting); id. at 343 (Souter & Ginsburg, JJ., dissenting); id. at 355 (Breyer, J., dissenting). The case assessed a partisan gerrymander of Congressional districts in Pennsylvania based on the 2000 census which had the unabashed purpose and dramatic effect of benefiting Republican congressional representation. 51Id. at 271 (majority opinion).

Justice Scalia’s plurality opinion not only found the Pennsylvania gerrymander at hand to violate no constitutional right but rejected the justiciability of partisan gerrymandering altogether. 52Id. at 281. The opinion invokes one of the political question criteria identified in Baker v. Carr—“lack of judicially discoverable and manageable standards” 53Id. at 277–78 (quoting Baker v. Carr, 369 U.S. 186, 217 (1962)).—to deny that courts can prohibit partisan gerrymandering. It argued that Justice O’Connor’s speculation that the Bandemer plurality opinion would become “unmanageable and arbitrary” 54Id. at 282 (quoting Davis v. Bandemer, 478 U.S. 109, 155 (1986) (O’Connor & Rehnquist, JJ., concurring)). if it did not become a vehicle for proportionality had been proven correct, and it observed that any concrete effects test would either enforce a specific vision of political identity or mandate an individual right to proportional representation. 55Id. at 287–88.

The plurality opinion emphasizes the inefficacy of the post-Bandemer litigation as a vehicle for policing partisanship, and thereby evades the true cause of the confused state of the law: failure of previous courts or litigants to identify the normative principle that partisan gerrymandering offends. If courts were to deem some partisan gerrymanders illegal in a consistent and predictable fashion (the features of the rule of law that underlie the political question criteria the plurality invokes), what value would need to underlie such a reliable test? The Vieth plurality concludes that unless courts wish to impose a system of proportional representation, the proportionality test or some variant upon it cannot be the basis for identifying illicit gerrymanders. 56Id. at 287–89. While this may be true, the majority ignores that there are bases other than proportionality by which courts could assess partisan gerrymanders. 57 See infra Sections IV.A.2–A.3, describing both the innovation of the First Amendment right and the Foley and Kang theories.

Thus, the Vieth plurality opinion is analytically premature in its conclusion that partisan gerrymandering could never comprise a constitutional wrong. For this to be the case, the legal principle that would provide the most incisive evaluation of partisan gerrymandering would also have to fail to provide adequate guidance to courts. The hesitancy to reach this conclusion led the swing vote in Vieth, Justice Kennedy, to decline holding that partisan gerrymandering was nonjusticiable. 58 541 U.S. at 306. Kennedy conceded that the dramatic intervention proposed by the dissents would comprise an “unprecedented intervention in the American political process” and that the Court must avoid “substantial intrusion into the Nation’s political life.” 59Id. Yet as the Court had entered the political thicket generally, Kennedy was unwilling to “bar all future claims of injury from a partisan gerrymander.” 60Id. at 309. The four liberal Justices agreed that partisan gerrymanders in general—and the Pennsylvania gerrymander in particular—were illegal, but offered three divergent accounts of how these conclusions should be reached. 61Id. at 337 (Stevens, J., dissenting) (arguing for neutrality as a matter of principle); id. at 345–46 (Souter & Ginsburg, JJ., dissenting) (adapting the McDonnell Douglas burden-shifting test); id. at 360 (Breyer, J., dissenting) (arguing for an “unjustified entrenchment” standard).

The fractured Vieth opinions left the law of partisan gerrymandering in a sorry state. 62 Michael S. Kang, When Courts Won’t Make Law: Partisan Gerrymandering and a Structural Approach to the Law of Democracy, 68 Ohio St. L.J. 1097, 1111 (2007) (“Justice Kennedy’s ambivalence leaves it bizarrely unclear where the law of partisan gerrymandering stands.”). Two years later, in League of United Latin American Citizens v. Perry (LULAC), 63 548 U.S. 399 (2006). a split Supreme Court opinion on a messy race-and-party-conflict inflected districting in Texas only reinforced the divisions of Vieth. The liberal Justices suggested a “bloodfeud” gerrymander was grounds for legal intervention 64Id. at 456 (Stevens, J., dissenting). and proposed tests for intervention; 65Id.; id. at 483 (Souter, J., dissenting). Kennedy indicated he still saw no workable standard; 66Id. at 419–20 (Kennedy, J., concurring). and the conservative Justices ignored the topic altogether. As described infra, however, dicta from LULAC, perhaps unwittingly, inspired the content of the current debate.

B. The Infiltration of Quantitative Methods

The current round of Supreme Court partisan gerrymandering litigation is defined by a series of linked developments across academia, gerrymandering technology, and lower court decisions. As the tools for statistical modelling and prediction of voter behavior have grown more sophisticated, 67Whitford v. Gill, 218 F. Supp. 3d 837, 847–50 (W.D. Wis. 2016), describes the application of some of these techniques; see also Samuel S.-H. Wang, Three Tests for Practical Evaluation of Partisan Gerrymandering, 68 Stan. L. Rev. 1263, 1267 (2016). they have been applied to make partisan gerrymanders more effective. In particular, the partisan gerrymanders currently receiving scrutiny in the federal courts have been enabled by two developments: (1) more precise data regarding voter behavior (with regards to time of behavior, geographic specificity, and focusing on actual past voting patterns instead of merely asserted party affiliation), 68 Common Cause v. Rucho, 279 F. Supp. 3d 587, 601, 603 (M.D.N.C.), vacated, 138 S. Ct. 2679 (2018) (reliance on detailed data of past citizen voting behavior from 2008 to 2014); Whitford, 218 F. Supp. 3d at 848 (development of “partisan score[s]” as a type of “customized demographic data” for all districts that were being redrawn). and (2) more powerful and multifaceted tools for arranging districts in a way that favors the gerrymandering majority party while still complying with neutral redistricting criteria, such as compactness and continuity. 69Rucho, 279 F. Supp. 3d at 601–04 (Republican expert drew plans that maximized partisan advantage while also taking into account traditional districting criteria); Whitford, 218 F. Supp. 3d at 849 (designing districts that offered different levels of aggressiveness in terms of partisan advantage while also seeking, to the degree possible, to comply with traditional redistricting criteria). These methods allow for the creation of maps that favor the gerrymandering party with both greater precision and greater durability, and reduce the likelihood that changes in demographics or voter behavior would disrupt the partisan advantage. 70Cf. Cain, supra note 34, at 165 (describing that the risks of partisan gerrymandering, in particular that it should erode over time as demographics change); Jacob Eisler, Partisan Gerrymandering and the Illusion of Unfairness, 67 Cath. U. L. Rev. 229 (2018) (arguing that free political will should enable parties and voters to adapt). Reformers, however, have used parallel technical advances to invent robust new tools to identify gerrymanders. As these reformers have led the latest round of anti-gerrymandering litigation, courts have incorporated them into the law. This trend reflects a transmission of methodology from technical social sciences to political practice and academic thought, and finally into court-made law itself.

1. From Conceptual to Statistical Treatments of Gerrymandering

The legal scholarship surrounding the Vieth litigation focused on classical questions of constitutional doctrine: did the principle of fair representation impose limits upon politicized districting, and was the equal protection clause the best way to approach it? 71 Mitchell N. Berman, Managing Gerrymandering, 83 Tex. L. Rev. 781, 838 (2005); Fuentes-Rohwer, supra note 27, at 562; see Heather K. Gerken, Playing Cards in a Hurricane: Party Reform in an Age of Polarization, 54 Hous. L. Rev. 911, 912 (2017) (characterizing the seminal article of this approach as the finest ever written in election law); Richard L. Hasen, Looking for Standards (in All the Wrong Places): Partisan Gerrymandering Claims After Vieth, 3 Election L.J. 626, 627 (2004) (describing the influence of the method); cf. Michael S. Kang, Gerrymandering and the Constitutional Norm Against Government Partisanship, 116 Mich. L. Rev. 351 (2017) (arguing for a similar demand for government neutrality, but based on a broader reading of the doctrine). Even as the legal scholarship around Vieth remained focused on constitutional rights and politics, however, quantitative methods were gaining momentum. 72 Gary King, Representation Through Legislative Redistricting: A Stochastic Model, 33 Am. J. Pol. Sci. 787 (1989); see Gary King & Robert X. Browning, Democratic Representation and Partisan Bias in Congressional Elections, 81 Am. J. Pol. Sci. 1251, 1252 (1987). Notably, such a system does not demand proportionality, as King notes; it simply requires that each party receive the same number of seats when it receives the same number of votes. Thus, if both parties in a two-party system receives 67% of representatives when each receives 55% of the vote, then the system is unbiased, even though each is overrepresented relative to its proportion of votes when it does win. Id. The legal academy was initially dismissive of the idea that quantitative analysis could resolve legal treatment of politicized districting. 73See, e.g., Jeanne C. Fromer, An Exercise in Line-Drawing: Deriving and Measuring Fairness in Redistricting, 93 Geo. L.J. 1547, 1578 (2005) (observing that while partisan symmetry is accepted by political scientists, including King, as the measure of fairness, it suffers “ugly complications”); Heather K. Gerken, Understanding the Right to an Undiluted Vote, 114 Harv. L. Rev. 1663, 1680 n.63 (2001) (observing that political scientists accept the partisan symmetry standard, but reject its meaning for racial fairness); Samuel Issacharoff, Judging Politics: The Elusive Quest for Judicial Review of Political Fairness, 71 Tex. L. Rev. 1643, 1682 (1993) (observing that, to date, such quantitative tests failed in the district courts). The attention the partisan symmetry test received in LULAC transformed scholars’ perspective. While LULAC did nothing to alter the fractured state of the jurisprudence, it revealed that the Justices recognized the usefulness of quantitative tests. In a perhaps unsurprising recurrence of the coalitions from Vieth, Justice Kennedy acknowledged the partisan bias test, 74 LULAC, 548 U.S. 399, 419–20 (2006). but expressed skepticism toward it, while the liberal Justices hailed it as answering the justiciability challenge by providing a clear standard. 75Id. at 466 (Stevens, J., dissenting); id. at 484 (Souter, J., dissenting) (“[F]urther attention could be devoted to the administrability of such a criterion at all levels of redistricting and its review.”). While this dicta did not change the law, it inspired scholars to conclude that quantitative measures might provide one avenue for engineering a falsifiable test that at least some Justices might accept. 76 For King’s own assessment of the partisan symmetry test and his attempt to introduce it as the test for gerrymandering, see Grofman & King, supra note 13, at 4. Their conclusion that a “a majority of [J]ustices now appear to endorse our view that the measurement of partisan symmetry can be used as part of a broader test in resolving partisan gerrymandering” appears optimistic, given that Kennedy, despite his initial seeming warmth toward technological progress, appears skeptical of the test. Id. Moreover, a comparison with the distribution of Vieth may get the causality backwards—liberal Justices wish to strike down partisan gerrymanders and are willing to adopt any tool that will aid them in doing so. Stephanopoulos and McGhee, meanwhile, seem to have been inspired to design the efficiency gap to specifically address Kennedy’s concerns in LULAC. Supra note 13, at 842.

The suggestion that the right quantitative test will provide a definitive test for partisan gerrymandering has captured the imaginations of scholars. The method that has received the most attention, 77 In addition to receiving attention in the courts, the efficiency gap has been the subject of dedicated academic treatment. See Benjamin Plener Cover, Quantifying Partisan Gerrymandering: An Evaluation of the Efficiency Gap Proposal, 70 Stan. L. Rev. 1131, 1138–39 (2018). the efficiency gap metric championed by Nicholas Stephanopoulos and Eric McGhee, moves in the direction of greater simplicity rather than greater complexity. The formula compares the number of votes cast for a party with the number of representatives it receives, using a formula to calculate the number of votes cast for a party with the number of seats received, 78 Stephanopoulos & McGhee, supra note 13, at 851 (stating that the efficiency gap is “the difference between the parties’ respective wasted votes, divided by the total number of votes cast in the election. Wasted votes include both ‘lost’ votes (those cast for a losing candidate) and ‘surplus’ votes (those cast for a winning candidate but in excess of what she needed to prevail)” (emphasis omitted)). and thereby calculates the number of “wasted” votes produced by a given districting. From a quantitative standpoint, the efficiency gap is simpler than many of the partisan symmetry tests or the simulation approaches. This is by design; the efficiency gap, for example, does not require consideration of a hypothetical state of affairs, thus circumventing one objection raised by Justice Kennedy. 79Id. at 845. Another approach that has received attention in both the academy and in the courts is the simulation approach advanced by Jowei Chen and Jonathan Rodden, which uses traditional redistricting criteria to create multiple simulations of ‘neutral’ maps. 80 Chen & Rodden, supra note 13, at 332. These simulations are then compared to real districtings to detect the likelihood of a partisan gerrymander. 81Id. at 338–39.

2. The Battleground in Lower Federal Courts: Quantification as the Pivot of Legality

The current round of litigation over partisan gerrymanders has seen quantitative analysis transform from an academic darling into a critical aspect of the judicial decision-making. At the district court level, judicial attitudes toward quantitative indications have strongly correlated to courts’ willingness to declare them illegal. Oral argument before the Supreme Court has suggested that Gill v. Whitford could have been expected to produce a referendum on the validity of these methodologies, but the Court declined to directly confront the issue. 82 Transcript of Oral Argument at 42–44, Gill v. Whitford, 138 S. Ct. 1916 (2018) (No. 16-1161). Regardless, the treatment of quantitative data within the lower court decisions themselves reveal an important pattern: where courts conclude that the outcomes presented by a new metric can establish discriminatory effect, courts identify an illegal gerrymander.

The district court opinion in Whitford v. Gill is exemplary. After describing the sophisticated methods Wisconsin Republicans used to develop self-serving districting plans, 83 Whitford v. Gill, 218 F. Supp. 3d 837, 847–50 (W.D. Wis. 2016). the majority concluded that both the First Amendment right to association and the Equal Protection Clause right to fair representation prohibit partisan gerrymanders that impair the efficacy of citizens’ votes. 84Id. at 883. As discriminatory intent has never been the sticking point for the litigation of partisan gerrymandering, 85See Vieth v. Jubelirer, 541 U.S. 267, 281 (2004) (intent of a legislature is not difficult to establish); Davis v. Bandemer, 478 U.S. 109, 129 n.11 (1986) (while discriminatory intent must be established for a successful claim, it may not be difficult to do so). the revolutionary move is the Whitford majority’s willingness to accept statistical evidence as sufficient to establish discriminatory effect. 86Whitford, 218 F. Supp. 3d at 902–04 (relying on “S” curve analysis, “swing analysis,” and the efficiency gap). The North Carolina opinion, Common Cause v. Rucho, follows a similar pattern: following a high-flying celebration of the principle that self-rule is a critical part of republican governance, 87 279 F. Supp. 3d 587, 685–88 (M.D.N.C.), vacated, 138 S. Ct. 2679 (2018). it then laconically concludes that partisan gerrymanders are an affront to voting rights because they effect entrenchment. 88Id. at 689–90. The Court then turns to statistical evidence, accepting it as proof that the partisan gerrymander had this intent and effect, relying on both statistical simulations 89Id. at 640, 659. and efficiency gap analysis.

The latest opinion to strike down a partisan gerrymander on the back of metrical analysis is League of Women Voters of Pennsylvania v. Commonwealth in the Pennsylvania Supreme Court. While decided on the basis of the Pennsylvania state constitution and thus of limited precedential value for the current federal litigation, its reasoning reflects the same pattern. The Court concluded that the districting plan “cannot, as a statistical matter, be a plan directed at complying with traditional redistricting requirements” and therefore “is sufficient to establish that it violates the Free and Equal Elections Clause.” 90 League of Women Voters of Pa. v. Commonwealth, 178 A.3d 737, 820 (Pa. 2018) (emphasis added); see also id. at 818 (“[T]he compelling expert statistical evidence presented before the Commonwealth Court, in combination with and illustrated by an examination of the Plan itself and the remainder of the evidence presented below, demonstrates that the Plan cannot plausibly be directed at drawing equally populous, compact, and contiguous districts which divide political subdivisions only as necessary to ensure equal population.”). The legal right against partisan gerrymandering could be entirely defined by formulaic correspondence between the asserted neutral district criteria and statistical indications that partisan interest lead to deviation from these criteria.

Remarkably, the refusal to acknowledge the dispositive effect of such evidence, or the absence of plaintiffs to rely on it, has correlated to unfavorable judicial treatments of claims to illegal partisan gerrymanders. In the ongoing partisan litigation in Maryland, Benisek v. Lamone91 266 F. Supp. 3d 799, 801 (D. Md. 2017). the federal district court declined to identify a First Amendment retaliation claim based on a partisan gerrymander, in light of unclear Supreme Court precedent. 92 The decision was, however, stayed in order to await the Supreme Court decision in Gill v. Whitford. Id. at 808. The case differs from Rucho and Whitford in that, perhaps because of the nature of the claim, 93Id. it only involved plaintiffs from a single district asserting that their district’s design inflicted a specific associational harm, and thus did not centrally revolve around the same type of statewide statistical evidence. Rather, the Maryland district court indicated that ambiguity regarding the clarity of partisan gerrymandering demanded that they wait for instruction from the Supreme Court 94Id. at 806–07.—a form of a statement, in effect, that no clear principle regarding when a gerrymander is illegal is yet established in the jurisprudence (even though, ironically enough, recent Fourth Circuit precedent addressing the same redistricting concluded that partisan gerrymandering claims are justiciable). 95 Shapiro v. McManus, 203 F. Supp. 3d 579, 597 (D. Md. 2016). Even more compellingly, a lower Pennsylvania state court concluded that while a variety of statistical analyses (including efficiency gap and simulated districting analyses) were credible and had a Republican bias, 96 Recommend Findings of Fact and Conclusions of Law at 95–96, League of Women Voters of Pa. v. Commonwealth, No. 261 M.D. 2017 (brief filed with Pa. Commw. Ct. on Dec. 29, 2017). they had little bearing on the constitutionality of gerrymandering because there was no clear baseline of legally required neutrality against which they could be compared. Thus, in the absence of precedent indicating a workable standard, 97Id. at 126–27. the claim was dismissed.

These lower court decisions show a remarkable correlation: when quantitative analysis of partisan gerrymanders is taken as facially indicative of discriminatory state action, courts will find them illegal. When the statistical evidence is absent or identified as failing to adequately bear on the relevant legal standard, courts will decline to so identify illegal partisan gerrymanders. This suggests that the current partisan gerrymandering jurisprudence is less a judgment of the legality of gerrymanders themselves than a judicial evaluation of the ability of quantitative methodologies to inform a type of constitutional claim.

The opinions that do strike down the partisan gerrymanders share two features: a commitment to the sense that sufficiently extreme partisan gerrymandering is an affront to democracy; and the assertion that quantitative analysis can serve as the critical element to define when such districtings are such an affront by indicating the appropriate degree of extremity. What these opinions fail to do is build a clear principled link between the discriminatory effect of politicized districting and the constitutional rights. No articulated constitutional principle mandates a particular relationship between votes and representation such that courts have a justification for intervention. 98 Vieth v. Jubelirer, 541 U.S. 267, 288–90 (2004); Lowenstein & Steinberg, supra note 49; Schuck, supra note 25. Nor is there any legal or constitutional indication that establishes party identity as a generally protected class. 99 Davis v. Bandemer, 478 U.S. 109, 152 (1986) (O’Connor, J., concurring). Thus, the opinions produce a conceptual lacuna: how, precisely, is the right to vote affronted by a politicized districting that favors one party over another? Or, stated in the alternative, how must the right to vote be conceived such that districting that favors one party comprises a constitutional wrong, and is this conception consistent with past doctrine? By failing to offer a principled basis for the decision to nullify a legislatively chosen districting plan, these opinions risk imposing judicial policy, with associated Lochnerite concerns as described infra.

C. Gill v. Whitford: Amidst the Evasion, Hints of the Constitutionalization of Statistics

Many critics expected Whitford to be a referendum on the legality of partisan gerrymanders and the authority of statistical tests (particularly the efficiency gap). 100See, e.g., Stephen Ansolabehere, Gill v. Whitford and the Methods for Uncovering Discriminatory Partisan Gerrymandering, Harv. L. Rev. Blog (June 26, 2018), https://blog.harvardlawreview.org/gill-v-whitford-and-the-methods-for-uncovering-discriminatory-partisan-gerrymandering/. Yet the opinions smacked of evasion more than substantive review, 101 Early commentary has supported the perception that Whitford was a fundamentally evasive opinion. See David Gans, Symposium: The Fight to Vindicate Our Constitution’s Promise of Democracy is Far from Over, SCOTUSblog (June 19, 2018, 11:14 AM), http://www.scotusblog.com/2018/06/symposium-the-fight-to-vindicate-our-constitutions-promise-of-democracy-is-far-from-over/; G. Michael Parsons, Dodge, Duck, Dip, Dive, and Dodge: The 5 D’s of the Gill v. Whitford Decision, Mod. Democracy (June 19, 2018), https://moderndemocracyblog.com/2018/06/19/dodge-duck-dip-dive-and-dodge/. with the majority opinion explicitly declining to “decide the latter question because the plaintiffs . . . have not shown standing under the theory upon which they based their claims for relief.” 102 Gill v. Whitford, 138 S. Ct. 1916, 1929 (2018). The lead plaintiff resided in a district that was solidly Democratic under both the plan at issue and in ostensibly neutral alternatives offered by the plaintiffs. 103Id. at 1924–25. However, the other plaintiffs who did live in districts that were “cracked” and thus had their votes allegedly diluted by gerrymandering had failed “to prove that [they] live[d] in a cracked or packed district” at trial. 104Id. at 1932. The Court concluded that the plaintiffs’ argument and their reliance on the efficiency gap offered only a generalized statewide observation, rather than one that spoke to the personal rights of individual voters. 105Id.

The very question the Court throws back as a standing query—whether plaintiffs adequately alleged individual harm 106Id. at 1929–31.—is at root the same that has continuously vexed the jurisprudence: What is the nature of the harm of partisan gerrymandering? Proving that a particular plaintiff lives in a “cracked” district (and has thus suffered harm so as to have standing) requires defining when partisan gerrymandering has occurred at all. 107 In some respects, therefore, the Court’s references to racial gerrymandering and one-person, one-vote, id. at 1930–31, are deceptively unhelpful. There is relatively little debate as to what comprises racial vote dilution or malapportionment at a conceptual level, merely a requirement that individuals must demonstrate the given illegal act harmed them personally. Conversely, as Whitford fails to draw out, the unanswered question in partisan gerrymandering is what is the harm? This is the very question that the Vieth plurality deemed impossible for courts, and which the various quantifying efforts were meant to answer. Moreover, the opinions in Whitford track the partisan divide of Vieth¸ albeit with less visible rancor. Roberts’ majority opinion only indicated that the plaintiffs must show effect on individual citizens to establish standing. 108Id. at 1924–25. This might, under a demanding theory of standing, require a full-blown conception of the nature of partisan gerrymandering—precisely the question Vieth deemed courts of incapable of answering and which the Court evaded addressing in Whitford. Conversely, Justice Kagan’s concurrence (joined by the liberal Justices) described the problem as “readily fixable” 109Id. at 1936 (Kagan, J., concurring). and provided a road map for doing so, as well as providing an alternative First Amendment ground for the claim. 110Id. at 1938. Thus Whitford translates the great judicial challenge of partisan gerrymandering—and the partisan divide that has riven it—into a new legal vocabulary, but does little to substantively advance it.

Even as Whitford defers a definitive statement regarding substantive judicial review of partisan gerrymandering, however, it seems to implicitly concede that statistics may be at the core of the solution to the standing problem and, thereafter, providing standards for judicial review of partisan gerrymandering. The majority opinion implies that if it is possible to provide an assessment of the effect of gerrymanders on individual voters, then the standing problem would be resolved (though it does at a minimum require that even statistically defined rights still retain the traditional legal focus on particularized harms). 111See id. at 1931–33 (majority opinion). Justice Kagan’s concurrence goes so far as to describe how this might be done using the available methodologies. 112Id. at 1934–36 (Kagan, J., concurring). Both, in effect, suggest that at least the standing problem could be resolved through more particularized metrics. If resolving the standing problem requires providing a legally adequate universal conception of gerrymandering, then this would also provide standards for the substantive review of its legality. Thus, if statistical evidence of effects can solve the standing problem, it could also solve the substantive-definition problem, and clearly define the right against partisan gerrymandering.

D. The Substitution of Evidence for Principles

This account of the partisan gerrymandering litigation has revealed that the judiciary now stands at the critical juncture with ramifications both for election law and for how courts interpret statistical data. That story is characterized by a subtle shift in the conceptualization of the problem facing partisan gerrymandering. The Supreme Court precedent first, in Bandemer, asserted that partisan gerrymandering could comprise an infringement of constitutional rights, but failed to articulate precisely how partisan gerrymandering violated a right—or, alternatively stated, to indicate what constitutionally meaningful norm it contravened. 113See supra Section I.A.1. By the time of Vieth, the failure of lower courts to successfully apply a consistent standard led a slim majority of the Supreme Court to conclude that there was no justiciable standard 114 Vieth v. Jubelirer, 541 U.S. 267, 281 (2004). (in Justice Kennedy’s view, yet), 115Id. at 311 (Kennedy, J., concurring) (“That no such standard has emerged in this case should not be taken to prove that none will emerge in the future.”). while liberal Justices fragmented in their efforts to articulate one. 116Id. at 337 (Stevens, J., dissenting) (arguing for neutrality as a matter of principle); id. at 345–46 (Souter, J., dissenting) (adapting the McDonnell Douglas burden-shifting test); id. at 360 (Breyer, J., dissenting) (arguing for an “unjustified entrenchment” standard). At the same time, social scientists were developing increasingly sophisticated ways to identify (and to implement) districtings that favored one party. In LULAC, the liberal Justices seized upon the most prominent of these standards to argue they could clarify the right in a manner that had so far eluded courts. 117See supra Sections I.A.2, I.B.1. Inspired by the result in LULAC, anti-gerrymandering reformers redoubled the attention they dedicated to quantitative identification and classification of gerrymanders. This narrative reveals a remarkably responsive set of interactions between social science, legal analysis, and judicial opinions, one that Justice Kennedy longed for in Vieth 118Vieth, 541 U.S. at 312–13 (Kennedy, J., concurring) (“Technology is both a threat and a promise. On the one hand, if courts refuse to entertain any claims of partisan gerrymandering, the temptation to use partisan favoritism in districting in an unconstitutional manner will grow. On the other hand, these new technologies may produce new methods of analysis that make more evident the precise nature of the burdens gerrymanders impose on the representational rights of voters and parties. That would facilitate court efforts to identify and remedy the burdens, with judicial intervention limited by the derived standards.”). and Judge Wynn lionized in his Rucho opinion. 119See Common Cause v. Rucho, 279 F. Supp. 3d 587, 650–64 (M.D.N.C.), vacated, 138 S. Ct. 2679 (2018). The Whitford majority opinion, with its statement that the standing problem would be resolved by statistical indication of “the effect a gerrymander has on the votes of particular citizens[,]” 120 Gill v. Whitford, 138 S. Ct. 1916, 1933 (2018). provisionally accepted the ability of statistics to contribute to the definition of constitutional queries. Whitford thus intimates that the definitive challenge to the constitutionalization of statistics is matching types of claims to types of quantitative tests, though it does little to indicate how many new types of claims might be generated. 121 Justice Kagan’s concurrence offers another possibility, by suggesting the statewide evidence of the efficiency gap might have been sufficient to support standing on a First Amendment theory had the Democratic Party been party to the suit. Id. at 1939 (Kagan, J., concurring).

It is, however, a narrative that also reveals how the law has gone astray. The current focus on statistical methods as providing a standard expands their proper role beyond that of evidence, and displaces the real question facing judicial reviews of partisan gerrymandering: What is the nature of the wrong inflicted by districting by political identity, and is it a wrong that courts ought to address? Only after this question has been answered will it be possible to ascertain if the numerical output from a given test can serve to indicate that a wrong has been inflicted. 122See Briffault, supra note 26, at 401 (listing the four constitutional arguments that might support a claim against partisan gerrymandering). While courts have suggested that the Equal Protection Clause and the First Amendment associational right are potentially infringed by partisan gerrymandering, they have done nothing to explain how these rights, specifically, are impaired by partisan gerrymandering. Instead, the law now proceeds by general assertions regarding the nature of justice, followed by descriptive indicators of a social phenomenon effected by government action. 123See supra Section I.B.2.

The plurality opinion in Vieth may bear much of the responsibility for this digression, and Whitford seems to threaten to perpetuate it. By suggesting that the essential problem with legal intervention against partisan gerrymandering was one of consistent identification of the practice rather than one of the characters of partisan gerrymandering, Vieth began a race to identify partisan gerrymanders in a way pleasing to a majority of the Supreme Court (and in particular to Justice Kennedy in light of his suggestion that technology could solve the justiciability problem). Whitford could well lead the Court on a similarly tangential quest to identify a test for when a voter is individually harmed by a gerrymander. This project could easily produce a profusion of standing theories similar to the way Vieth produced a profusion of justiciability theories.

The questions of whether partisan gerrymandering is an appropriate subject of judicial assessment, and what the conceptual basis of that assessment should be, have fallen from vogue, 124 Stephanopoulos and McGhee, supra note 13, have virtually no discussion of how the metric intersects with constitutional principles. The normative discussion would, if extended as a principle, seem to mandate proportional representation. Id. at 852–53 (“The gerrymandering party enjoys a political advantage not because of its greater popularity, but rather because of the configuration of district lines.”). Even if it did not, it would seem to clash with Schuck’s observation that in the absence of a mandate to proportional representation, determining the degree of benefit conferred on a winning party is a question of democratic self-determination. Schuck, supra note 25. yet have never been answered. Yet as such quantitative outcomes can provide evidence of a state of affairs, they cannot indicate to courts if that state of affairs is one that comprises a legal wrong. The turn toward statistical evidence without fully addressing this question, a move that seems prompted by the Court itself, seeks to solve the wrong problem.

II. The Distinct Domain of Law: Principled Rights

To understand the inadequate condition of the partisan gerrymandering litigation, it is necessary to develop the assertion that legal rights are principled in nature. 125 This proposition and its implications are drawn primarily from two tests. First, a judge acts as a sort of “philosophical referee” who relies on a “constitutional theory” (with regard to constitutional provisions) and a “doctrine of fairness” (with regard to case precedent) to interpretively generate a “seamless web” of principle. See Dworkin, Rights, supra note 19, at 105–30. For a rigorous and vindicatory analysis, see J. Raz, Professor Dworkin’s Theory of Rights: A Critical Review of “Taking Rights Seriously”, 1977 Bull. Austl. Soc’y Legal Phil., no. 6, 1977, at 1, 7–13. For a broad overview and a justification of its relevance to American law, see Peter Gabel, Book Review, 91 Harv. L. Rev. 302, 302 (1977) (reviewing Ronald Dworkin, Taking Rights Seriously) (“Taking Rights Seriously is a justification of contemporary American legal practice expressed in abstract and universal terms.”). The second test further expands the theory of rights presented therein. See Dworkin, Empire, supra note 19, at 210–28. For a playful take on the relationship between the two works, see generally Allan C. Hutchinson, Indiana Dworkin and Law’s Empire, 96 Yale L.J. 637 (1987). They reflect the bedrock commitments of a polity, and offer protections of a universal character. 126Dworkin, Empire, supra note 19, at 214. One consequence of rights’ principled character is that when a party can successfully assert a legal right, courts must weigh that right independent of any other circumstances (such as the beneficial or pernicious consequence of recognizing the right). That rights are principled rather than consequential differentiates them from policy decisions, which are the domain of the elected branches of government and parallels the distinction between the judiciary as beholden to law and the political branches as beholden to popular will. This means judicial review of partisan gerrymandering must offer a clear principle, not merely select certain metrics that, unless in evidentiary service to such a principle, are legally arbitrary.

A. The Principled Nature of Rights

In legal application, rights are principled in nature. The invocation of a right before a court enables that party to demand protection or privilege 127 Jeremy Waldron argues, with circumspection, that political rights are distinct from personal rights in that they demand a positive rather than negative actualization, though he observes it is possible to provide a fundamentally negative account of all rights. Jeremy Waldron, Law and Disagreement 232–34 (1999). There is a powerful insight in his argument but, for better or worse, the U.S. courts have imagined rights largely in negative terms (as protection from illicit government action), and so will this Article. Further, it will identify rights as largely personal in nature. Baker v. Carr, 369 U.S. 186, 286 (1962) (Both standing generally and the political question doctrine specifically require “infringement of an interest particular and personal to [the plaintiff].”); Dworkin, Rights, supra note 19, at 133 (U.S. Constitution is designed to prevent individuals from suffering majoritarian abuses); cf. Gerken, supra note 73, at 1671–72 (arguing that vote dilution can only be understood as an aggregate right). on the basis that the governmental action intersects with the asserted right. 128 Given the focus on partisan gerrymandering, this Article focuses on, and uses the language of, constitutional rights that private parties assert against the government. Of course, individuals may assert private rights against one another as well, Arthur Ripstein, Private Wrongs 4 (2016), and at a certain level of abstraction the same concerns are present there. This might throw into doubt the universality or completeness of the ability of empirical analysis to displace principled analysis even in the context of other domains of law, but that is beyond the ken of this Article. When a court determines that an individual has a right that must be taken into account, it is not because of the beneficial consequences of recognizing the right, but rather because there is (for the purposes of the court’s legal analysis) an independent weight attached to the claim. When a right is applicable, the legal force attached to that right needs no further justification. An example of this would be the principled claim that the government ought not prevent individuals from engaging in speech. The particular instance or category of government restriction of speech activity may be harmless or even beneficial, but that is irrelevant for recognizing the weight of the right (requiring that the government conduct must pass strict scrutiny). 129 Cass Sunstein makes this point nicely: though such a law would likely have little practical impact, the First Amendment would require courts to strike down legislation that prohibited individuals from criticizing the President while they shower. Cass R. Sunstein, Democracy and the Problem of Free Speech 168 (1993). The right stands based on principle that is foundational to the polity, 130Dworkin, Empire, supra note 19, at 211. and the asserter of the right need offer no further justification for its legal weight. 131 This character of rights, and the differentiation from policy, can be attributed to the fact that rights are “individuated” rather than “collective” in character. Dworkin, Rights, supra note 19, at 95. Dworkin’s differentiation applies to all rights; legal rights have the feature of being enforced in the judicial framework. Id. at 106.

The pithiest expression of this idea may be Ronald Dworkin’s description of “rights as trumps.” 132Id. at 154–59 (arguing a concept of principled rights must have an “independence” that cannot be justified by utilitarianism). Even when there is a claim that a particular course of action is the right one for a society—either because it is allegedly more socially beneficial, or because it reflects democratic will—if that course of action violates a party’s right, that party can appeal to the courts to intervene. The court may not conclude that the right justifies interdicting the government action, but once the party can demonstrate that the action does infringe upon the right, the right receives weight independent of the considerations of popular will or social benefit. 133 Robert Nozick provides a generalized form of this attribute of rights (focused, perhaps unsurprisingly, on property)—a right has force regardless of the consequences of enforcing it. Robert Nozick, Anarchy, State, and Utopia 269–70 (1974). This is reflective of the Nozickian equivalent of the Dworkinian trump, the “side-constraints.” Id. at 29. The burning question of partisan gerrymandering can thus be stated simply enough (though it rarely has been): What is the “trump” that those who assert they have been harmed by a partisan gerrymander are asserting? That is, what protected quality or independently valuable right do they claim is harmed, even when that harm comes from democratic action? 134 As described in Part IV, infra, the clearest right they seem to assert is one to be free from harm by political competition; the typically adduced rights of fair representation under equal protection or a general right to association are so vague in this context as to potentially be undermined by the Vieth justiciability claim.

Given that a claim of right is not a claim of beneficial consequence, it fits most intuitively with a deontological understanding that deploys rights to protect certain essential human goods (the opportunity for political participation, to be free from illicit government classification, and so forth). 135 In American legal doctrine, the deontological character of rights (at least in intermediary terms) is probably most apparent in their individual character. See Baker v. Carr, 369 U.S. 186, 286–92 (1962); cf. Dworkin, Rights, supra note 19, at 166 (observing such rights must be located in granular individual equality, not a composite of collective welfare). Yet any theory that takes rights as a category seriously 136 Thus the great difficulty a committed utilitarian—that is to say, wholly consequence-focused theory—has accommodating a theory of rights. See David Lyons, Utility and Rights, in Theories of Rights, supra note 19, at 129. will recognize their principled treatment in judicial application. For example, Richard Posner has outlined a theory of law oriented toward wealth maximization, 137Richard J. Posner, Economics of Justice 62 (1981). Interestingly, Posner attacks Benthamite utilitarianism as an approach to law that struggles to incorporate rights. Dworkin, Rights, supra note 19, at 232–38. an unequivocally consequentialist approach. Yet he still identifies a place for “absolute rights” 138Posner, supra note 137, at 70–71. that serves the broader economic purpose of the law. For Posner such rights “are not transcendental or ends in themselves,” 139Id. at 71. but rather serve as fixed points upon which individuals can rely in the process of wealth-maximizing conduct. However, to achieve this effect, courts must treat rights in practice as principled in character and trumping in effect.

However, the successful invocation of the right does not entail that the invoker of the right will win. There may be other rights against which it must be balanced; there may be other interests that are not rights that can override it. A governmental racial classification must pass strict scrutiny; 140 A classically controversial example would be affirmative action programs at universities, which apply strict scrutiny to racial classifications but permit them as supported by a compelling state interest and narrowly tailored. See generally Grutter v. Bollinger, 539 U.S. 306 (2003). deviations from the equal right to vote must be supported by other “legitimate considerations” a state may have in districting. 141 Reynolds v. Sims, 377 U.S. 533, 579, 581 (1964). That a right has “trumping” capacity merely asserts that the right necessarily has weight regardless of other considerations. This is perhaps most classically apparent in free speech jurisprudence—except where speech is of one of the narrow preexisting categories denied legal protection, it cannot be restricted for content merely because the court believes the speech lacks value or is socially harmful. 142 Brandenburg v. Ohio, 395 U.S. 444 (1969). If the conduct is speech, then it receives an automatically significant weight, and may only be restricted by a compelling interest—and even that restriction must occur neutrally. 143See Elena Kagan, The Changing Faces of First Amendment Neutrality: R.A.V. v. St. Paul, Rust v. Sullivan, and the Problem of Content-Based Underinclusion, 1992 Sup. Ct. Rev. 29, 34, 40. See generally Sunstein, supra note 129, at 169 (describing the logic of content neutrality).

This explains why, even if a robust principle that supports one of the metrics offered to identify partisan gerrymandering is accepted, hitting a particular threshold would not automatically mean illegality. 144Cf. Raviv, supra note 47. Rather, upon identifying the existence of the right-offending districting, a court would then have to determine if the underlying government action survived, in a comprehensive analysis, the appropriate level of scrutiny. This in turn requires clear identification of the nature of the wrong inflicted, and of how the principle the wrong offends should be weighed against competing interests (in particular in partisan gerrymandering, the general right to political self-determination realized as legislative autonomy).

The judicial enforcement of rights is further clarified by a comparison with the role of political branches in advancing policy. 145 This Article again uses Dworkin as the foundation for this analytic distinction. See Dworkin, Rights, supra note 19, at 90 (“Principles are propositions that describe rights; policies are propositions that describe goals.”). That a court may protect speech that it deems to actively make society worse further clarifies the nature of rights, and in particular how the courts’ role as the protector of rights distinguishes the judiciary as a branch of government. A court could simultaneously conclude that an instance of speech is socially harmful, or entirely without real-world benefit—and thus unequivocally bad policy—and yet protect it from governmental interference. Such a holding, furthermore, might be sound, even obligatory, from a legal perspective. Yet for the legislative or executive branches of government to enact or enforce such deliberately bad policy would be a very defiance of their very mission to serve the public good. 146Id. at 91 (“Collective goals encourage trade-offs of benefits and burdens within a community in order to produce some overall benefit for the community as a whole.”).

The distinction of rights from politics can ultimately be traced to the role of the judiciary itself. Courts, unlike the political branches, are beholden to uphold the law rather than popular opinion or political interest. The rule of law is defined by impartial consistency in the application of rules 147See Dworkin, Empire, supra note 19, at 174 (“We apply the strictest standards of impartiality even to officials whose power is relatively slight and substantially less than that of many private citizens; we have no sense that an official’s duty of equal concern wanes as his power diminishes.”); see also Lon L. Fuller, The Morality of Law 46–49 (1964) (observing generality as the first essential quality of rule of law).—“treat like cases alike”—and rights comprise the fixed points of consistency to which a party can advert before a court (though, as described above, why such fixed points are valuable at all is itself a hotly disputed question). Certainly, a party must establish that it can properly invoke the relevant right. To invoke the protections of the right of free speech, it must be established the given activity is, in fact, speech; for a partisan gerrymandering, it must be established that a partisan gerrymander was intended and effected. However, this is the only requirement to gain the weight of the right. This concept is present in the idea of rights as “trumps.” 148 This feature is present in the very character of rights as trumps. See Dworkin, Rights, supra note 19, at 106. It has manifested variously in constitutional law. For example, in the application of the Fourteenth Amendment, the right—and subsequent strict scrutiny—is activated by any use of a racial classification, in order “to ‘smoke out’ illegitimate uses of race by assuring that the legislative body is pursuing a goal important enough to warrant use of a highly suspect tool.” City of Richmond v. J.A. Croson Co., 488 U.S. 469, 493 (1989).

That courts are the branch of government that is beholden only to law qua law reinforces this distinction. Democratically elected branches of the government may make the law and enforce the laws (through legislation and executive action), but law is a tool, to be deployed malleably and instrumentally to serve goals that particular political actors wish to achieve. Were rights’ principled character not central to rights enforcement, the distinction between the elected branches as policymakers and the courts as lawmakers would break down, and democratic autonomy itself (as well as the limited role of the courts) would be impaired.

B. The Risk of Legislating from the Bench and the Gerrymandering Morass

The abstract claims above speak to the general nature of judicial enforcement of rights, yet the principled character of rights protection manifests throughout the law. Perhaps its most general form is the threshold requirement of standing—because the sole role of courts is to “decide on the rights of individuals,” a party may only make a claim before a court through asserting its legal rights are violated. 149 Antonin Scalia, The Doctrine of Standing as an Essential Element of the Separation of Powers, 17 Suffolk U. L. Rev. 881, 884–85 (1983). Scalia’s vision has been challenged, however, by those who argue that standing itself is often used to achieve political ends and is not an intrinsically coherent doctrine. See Heather Elliott, The Functions of Standing, 61 Stan. L. Rev. 459, 501 (2008); Richard J. Pierce, Jr., Is Standing Law or Politics?, 77 N.C. L. Rev. 1741, 1775 (1999). Yet if a violation of a right is established, the court must give the right its due weight, rather than engage in a separate consequential consideration of the benefits of recognizing the right. Standing, in effect, reflects the crisp delineation that is the practical effect of rights’ principled nature, and the subsequently narrow character of judicial assessment. Courts only effect social change when a legal right is successfully invoked, but the invocation of that right is insulated from, and independent of, other factors in the broader context: either a party has had a legal right infringed and has standing, or has not, and has none (however unlucky or unfortunate the party’s circumstance). 150 The most concrete form of this may be the requirement for an injury-in-fact, Lujan v. Defs. of Wildlife, 504 U.S. 555, 563 (1992), which prevents courts from serving as “a forum for generalized grievances.” See Gill v. Whitford, 138 S. Ct. 1916, 1929 (2018) (quoting Lance v. Coffman, 549 U.S. 437, 439 (2007)). The Court in Whitford emphasized this when it observed that standing “ensures that we act as judges, and do not engage in policymaking properly left to elected representatives” 151 138 S. Ct. at 1923 (quoting Hollingsworth v. Perry, 570 U.S. 693, 700 (2013) (emphasis in original)). to justify its decision to remand the case back to the district court.

The most seminal manifestation of the idea that legal enforcement of rights should reflect principled rather than political judgments is the aptly named (and fiercely disputed) political question doctrine. 152 There is a well-founded dispute over whether the political question doctrine is itself principled, an assessment that would support a Dworkinian view of both the doctrine and judging more generally (the “classical” view), or is instrumental in the service of maintaining the courts’ institutional position (the “prudential” view). Herbert Wechsler, Toward Neutral Principles of Constitutional Law, 73 Harv. L. Rev. 1, 15–16 (1959) (laying out the classical view and describing judging as “genuinely principled, resting with respect to every step that is involved in reaching judgment on analysis and reasons quite transcending the immediate result that is achieved”); cf. Alexander M. Bickel, The Least Dangerous Branch 183–98 (1986) (laying out the prudential view); Fritz W. Scharpf, Judicial Review and the Political Question: A Functional Analysis, 75 Yale L.J. 517, 520 (1966) (“[A]voidance [of deciding a case] . . . need not rest on constitutional principle; it may legitimately express the Court’s prudential estimate of the desirability of deciding a given constitutional issue under the particular circumstances.”). The “classical” view thereby identifies the principled character of the legal defense of rights, as distinct from the consequential focus of policies, as ontologically intrinsic to lawmaking. Conversely, the “prudential” view identifies the Court’s focus on principled, non-consequential decisions as ultimately strategic in nature, inspired by a desire to avoid upsetting the allocation of institutional power. In effect, however, both the classical and prudential understandings of the political question doctrine produce judicial treatment of rights that in legal practice emphasizes principle rather than policy outcomes. The “classical” view identifies the fulcrum of principle as inhering within the Constitution itself, whereas the “prudential” view identifies it as expressive of the practical realities of intrabranch hydraulics. If the political question doctrine applies, and a dispute is political rather than legal in nature, it lies beyond the ken of courts. In the single case that perhaps birthed the role of the federal judiciary, Justice Marshall “provided key guiding factors for identifying a political question: ‘The subjects are political. They respect the nation, not individual rights . . . .’” 153 Rachel E. Barkow, More Supreme than Court? The Fall of the Political Question Doctrine and the Rise of Judicial Supremacy, 102 Colum. L. Rev. 237, 249 (2002) (quoting Marbury v. Madison, 5 U.S. (1 Cranch) 137, 165 (1803)). Many, including Barkow, have argued that the political question doctrine is on the retreat, id. at 336, particularly in the domestic arena. See, e.g., Harlan Grant Cohen, A Politics-Reinforcing Political Question Doctrine, 49 Ariz. St. L.J. 1, 12 (2017) (arguing that with a few exceptions, including most notably partisan gerrymandering, the political question doctrine in modern practice has become restricted to international law). Some have suggested this indicates judicial overreach, a claim with which this Article broadly agrees. See, e.g., Barkow, supra, at 336. The modern 154 Some have argued that the historical form of the political question doctrine is fundamentally unrelated to the contemporary form, and that it ultimately grants power to the court, rather than restricts it. See Tara Leigh Grove, The Lost History of the Political Question Doctrine, 90 N.Y.U. L. Rev. 1908, 1911 (2015). The former issue is tangential to the argument here, and the second imputes a tactical insincerity to judicial thinking with which this Article need not engage. expression of the doctrine has its most seminal formulation in Baker v. Carr’s indication that the political question doctrine instructs courts to abjure opining on issues where there is a “lack of judicially discoverable and manageable standards.” 155 369 U.S. 186, 217 (1962). While Baker articulates several different criteria that can indicate that the political question doctrine removes an issue from the Court’s ken, this has been identified as the “trigger . . . for nonjusticiability under political question rules,” Aziz Z. Huq, Removal as a Political Question, 65 Stan. L. Rev. 1, 22 (2013), and, most relevantly for the purposes of this analysis, is the basis of the conservative rejection of the justiciability of partisan gerrymandering. See Vieth v. Jubelirer, 541 U.S. 267 (2004).

The political question doctrine highlights parallels between rights as matters of principle, and the process of judging as characterized by “neutrality and generality.” 156 Wechsler, supra note 152, at 16. The Court only opines where the legal claim brought by a party is one that is not ultimately a matter of democratic discretion, but is of a general character such that a court can opine on it using principled reason, a type of decision-making distinct from discretionary assessment of the desirability of a particular outcome. 157See Baker, 369 U.S. at 217. The explicit invocation of a constitutional right by a party is an appeal to the judicial branch to engage in its characteristic reasoning that is distinctly principled. Having been failed by governance as politics, either in a highly granular way (such as by suffering at the hands of a low-level government agent, as with a typical § 1983 suit against a police officer), or an explicitly political way (such as a challenge to the legality of legislation which was the direct function of a democratic process), the party hopes that the court will enforce the overriding principle in the face of competing branches of government.

The relationship between rights, principles, and the political question doctrine clarifies the germane aspects of the Vieth plurality’s assertion that partisan gerrymandering is nonjusticiable. One claim of the Vieth plurality is that judicial identification of gerrymanders is impossible “in practicality” because there are no easily administered tests (as there are, for example, for one-person, one-vote conformity). 158Vieth, 541 U.S. at 290. If technically sound, the quantitative innovations resolve this problem by providing objective standards for defining partisan gerrymanders. Courts could simply find a breach of any threshold of a given metric, such as the efficiency gap, indicative of an illegal districting. 159See Stephanopoulos & McGhee, supra note 13, at 887 (setting thresholds for the identification of partisan gerrymanders that courts should strike down). Thus such standards would be “discernible” and “manageable” 160Vieth, 541 U.S. at 290. insofar as courts could consistently enforce them. 161 That is not to say their consistent enforcement would not raise its own puzzles. Id. at 288 (problem of identifying a majority party to describe partisan gerrymanders); see also Eisler, supra note 70 (describing the fluidity of political identity). Indeed, reliance on a statistical metric to resolve the standing question of Whitford would have the same effect; if the nature of the harm were defined by its granular impact on individual voters by party, it would be possible to define a particularized harm, and perhaps therefrom to provide a comprehensive description of partisan gerrymandering. 162 Gill v. Whitford, 138 S. Ct. 1916, 1932–33 (2018). Yet neither of these innovations would provide an identifiable legal principle that supports the enforcement of such thresholds. 163Vieth, 541 U.S. at 288 (standard advanced by appellants is a de facto argument for proportionality unsupported by equal protection). The remaining question is why courts should enforce such a standard. To do so without justification by a legal principle would comprise the type of “policy determination” 164Whitford, 138 S. Ct. at 1916 (describing the purpose of the standing requirement as preventing use of courts to raise general grievances about governance); Baker v. Carr, 369 U.S. 186, 217 (1962) (describing the features of a political question). that lies outside of judicial discretion, and thus is within the political question doctrine. 165Baker, 369 U.S. at 217. What is required is the identification of a principle that links the metrical standards to some constitutional right enforced by the courts.

III. Rights Protection and the Limits of Quantification

The standing requirement and the political question doctrine express a general feature of rights protection by courts: such rights protection cannot mandate specific outcomes, but rather must serve neutrally protected norms. 166See Richard H. Fallon, Jr., Strict Judicial Scrutiny, 54 UCLA L. Rev. 1267, 1316 (identifying the centrality of “triggering” in rights analysis as a background condition to protection of a right). This activation of the right is independent of the effect of the violation, by examining if the conduct or classification is of the type that right protects. Whether conduct is of the right categorical type is an analysis linked to the very principles the right is meant to defend. For example, whether the First Amendment protects obscenity or flag burning are questions that can only be answered by looking to the purposes of free speech, and asking if the practices in question are defended by the principles that emerge from consideration of the right. 167Id. at 1317–18. Once the judicial mandate to defend a right is established, metrical analysis may serve as evidence that the government conduct in question offends it. However, the role of such metrical analysis is delimited by the very nature of judicial rights enforcement: since rights provide ‘trumps’ independent of other considerations, the only question for the court is if the government conduct is offensive in character. As this section establishes, the ‘trumping’ nature of rights means that courts are ultimately concerned with the character of government action, rather than its policy effects. Effects may be relevant to establishing the presence of a rights violation, but they do not define the wrong itself. 168 For example, in the racial districting context, evidence of “a district’s shape and demographics” may help establish the illicit race-based intent that offends the Equal Protection Clause. Shaw v. Hunt, 517 U.S. 899, 905 (1996). This section demonstrates the implications of this argument for the gerrymandering litigation.

A. Judicial Rights Enforcement as Principled

That judicial rights enforcement serves principle rather than policy means that identifying a state of affairs in the world (and thus empirical data) alone cannot be the essence of a rights violation. Rather, judges must ask if that state of affairs has the moral quality that activates the right. The assessment of the action or condition itself—ultimately an evidentiary assessment—is only an intermediate step in the assessment. It is only relevant if the state of affairs establishes that government action has violated a principle that is protected by a right. If the violation is unjustified, then the assertion of the right will render the government action illegal. The assertion of the right itself, however, is based on the character of the government action, not a claim regarding the desirability of its effects. 169 This principle is apparent in the fact that even functionally harmless government action that violates rights is illicit. To borrow an example from Cass Sunstein: “There is no legitimate reason for government to say that people may not criticize the President while they are showering. The only possible reason is to insulate the President from criticism, and this is illegitimate. The restriction is therefore unconstitutional even if it is minor.” Sunstein, supra note 129. This is the essential divide between policy—the domain of the legislature in creating desirable states of affairs in the world—and the courts, as protecting rights from government intrusion.

Thus, a court must ultimately look to the ex ante character of the government action—independent of its consequences—to determine if a constitutional right can be successfully invoked. 170 In the partisan gerrymandering literature, this view has been most coherently advanced by Lowenstein and Steinberg. Lowenstein & Steinberg, supra note 49, at 15–16. Others have explicitly rejected this assertion, arguing that the practice of the Supreme Court reveals much more ad hoc intervention to prevent seeming unfairness. Martin Shapiro, Gerrymandering, Unfairness, and the Supreme Court, 33 UCLA L. Rev. 227, 256 (1985). Insofar as courts engage in such unprincipled intervention to prevent their intuitions regarding fairness, they undermine the rule of law. See Fuller, supra note 147, at 49–51 (promulgation of laws essential for the rule of law). This is the practical (but general) corollary of the philosophical claim regarding the nature of rights described infra: since rights are principled, it is the character of the action, not its consequences (the domain of legislative policy), that condemn it as a rights violation. The legal identification of a rights violation thus expresses the commitment to neutrality that is essential to good judging—the legal invocation of a right does not vary by any circumstances beyond those that are necessary to establish its presence. Therefore, once it is established that speech is restricted, the right to free speech is activated; once it is established that a government action uses racial classifications, the right to equal protection is activated. The commitment to judicial neutrality means that no other factors—such as the factual setting or empirical consequences of the government action that activates the right—are relevant to the abstract availability of the rights protection (though they may, of course, be relevant to subsequent judicial analysis, such as identifying basic activation of the right and if the government action passes the appropriate level of scrutiny). This neutrality also serves to maintain the differentiation between judging and politics that manifests through the political question doctrine. While the political branches may base their actions largely around what they believe the consequential outcomes of their policies to be, and how they expect the citizenry to react to them, the judiciary is beholden only to the law as a disinterested set of instructions for neutral enforcement of the law. 171Bickel, supra note 152.

The case that initiated modern judicial intervention into district line drawing, Gomillion v. Lightfoot, is exemplary in demonstrating that rights themselves cannot be defined by empirical states or outcomes. 172 Gomillion v. Lightfoot, 364 U.S. 339, 347–48 (1960). A white-majority state legislature, responding to rising levels of black voter registration, had drawn the Tuskegee city boundaries in a contorted manner to exclude black voters. 173Id. at 340. The Court concluded that the twenty-eight-sided figure was clearly indicative of racial discrimination in districting, and Gomillion was a watershed case in the development of the law demanding that legislative districting avoiding illicit classification by race. 174Id. at 346–47. Yet it was not the existence of the figure itself or any intrinsic geometric attribute that validated judicial rights enforcement, but the illicit moral character of the action—blatant racial animus—that motivated the drawing of it. 175Id. That claim is normative rather than evidentiary in character, standing as a matter of principle that demands neutrality in government decision-making. 176Id. at 347–48. No empirical evidence can generate this; its nature is that of a value. That rights are principled in nature, and thus ultimately cannot be informed by empirical evidence, is thus part of their very ontology. 177Dworkin, Rights, supra note 19, at 90–91. That rights have such unique ontology would be rejected by certain significant strands of thinking, such as legal realism. See generally Brian Leiter, Legal Formalism and Legal Realism: What Is the Issue?, 16 Legal Theory 111 (2010) (describing in general the sociological theory of judging that would undercut such a unique status held by legal rights).

This means, then, that a court’s decisions about rights cannot be founded upon an empirical claim about a state of affairs in the world, nor can empirical or statistical evidence delineate a new right or define its nature. Rather, a claim of right must have at its core the assertion that courts should prevent a certain category of governmental action. If the category asserted is one that is established as protected, the onus upon the claiming party is, broadly speaking, to demonstrate that the contested governmental action infringes upon it. If the category is arguably novel, then the onus upon the claiming party is to demonstrate by analogy that the type of attribute for which protection is sought ought, through conceptual likeness, to be understood as akin to a feature that is already protected by a right. 178 The development of rights related to sexuality provide one of the most famous mechanisms by which U.S. courts have come to innovate rights through conceptual likeness. Griswold v. Connecticut, 381 U.S. 479, 484. (“[S]pecific guarantees” of constitutional rights have “penumbras” and “emanations.”). See generally Paul G. Kauper, Penumbras, Peripheries, Emanations, Things Fundamental and Things Forgotten: The Griswold Case, 64 Mich. L. Rev. 235 (1965) (describing constitutional innovation of rights up to Griswold). A more recent expression of this may be in Obergefell v. Hodges, 135 S. Ct. 2584 (2015). For both a general theory of how this process unfolds and its specific application to the evolution of racial attitudes, see Michael J. Klarman, From Jim Crow to Civil Rights 5 (2006) (methodologically, asserting that “constitutional law is generally quite indeterminate [so] constitutional interpretation almost inevitably reflects the broader social and political context of the times.”). Klarman’s approach is distinct from Dworkin’s, who would argue that constitutional law, done right, reflects a process of interpretation that weaves together precedent, social values, and abstract fairness. Regardless, Klarman’s account demonstrates the significant role that social norms might play in the development of the substantive content of constitutional rights. Interestingly, Klarman’s focus on social facts might show how he is conceptually more closely aligned with the empirical legal scholars. See, e.g., Stephanopoulos and Versteeg, supra note 8 (using popular approval to assess the validity of constitutions). Claims that argue for a new rights interpretation operate by attempting to demonstrate how the disputed governmental action violates principles the Court already enforces. To make an argument for rights protection that simply seeks to make the state of practical affairs in the world more desirable—as might an argument based purely on the empirical effects of a decision—is to turn the Court into merely an organ of policy decision-making.

B. Rights as “Shadow” Policies and the Risk of Lochnerization

As courts decide cases rather than describe broad theories of judging, they have not explicitly articulated the theory above. However, the judicial obligation to serve principle rather than to effect policies, and consequent exclusion of empirical evidence of the effects of government action from judicial rights analysis, manifests across the constitutional jurisprudence. In particular, judicial assessments of whether rights violations have occurred consider government action as evidence—the character, and in particular the intentionality, of the action—rather than engaging in their own independent review of the desirability of the effects. If courts do simply strike down legislative action without a principled basis, they fail to respect the judicial role as well as separation of powers. 179See Bickel, supra note 152, at 16 (observing the basic normative paradox in judicial nullification of democratic government action). See generally David E. Pozen, Self-Help and the Separation of Powers, 124 Yale L.J. 2, 75 (2014) (“[T]he aim of separation of powers is . . . to preclude the exercise of arbitrary government.”).

The most notorious example of a case that engages in such judicial policymaking is Lochner v. New York, perhaps the most roundly condemned decision of the twentieth century. 180 198 U.S. 45 (1905). Given its effect of thwarting various New Deal reforms and the development of the modern welfare state, it is tempting to classify Lochner’s failure as a normative one: in siding with employers and the well-off rather than permitting government intervention to protect the less socially powerful, Lochner seemed to be on the wrong side of history. Yet as Cass Sunstein has observed, 181 Cass R. Sunstein, Lochner’s Legacy, 87 Colum. L. Rev. 873, 875 (1987). the substantive logic of Lochner—the idea that the Court should enforce certain distributions of resources—has persisted, albeit through different means of rights enforcement.

The real weakness of Lochner thus may not be its substantive effects and its location on the wrong side of history, 182 Indeed, some have challenged the analytic principle underlying Sunstein’s claim. Einer R. Elhauge, Does Interest Group Theory Justify More Intrusive Judicial Review?, 101 Yale L.J. 31, 56 (1991) (characterizing Sunstein’s analysis as baseline reinforcing). but rather that its reasoning relied on what amounted to enforcing rights as a means of achieving a judicially sanctioned policy outcome. 183See Fallon, Jr., supra note 166, at 1270 (Lochner is condemned as “second-guessing of legislative judgments”). Justice Holmes’s dissent captures this quality: the Lochner majority did not advance a principled conception of liberty from government interference so much as a particularized vision of what types of economic policies would prove most socially beneficial. 184Lochner, 198 U.S. at 74 (Holmes, J., dissenting). The majority opinion’s extensive discussion of the actual conditions faced by bakers as the fulcrum of its decision to enforce the employer’s rights likewise exemplifies a type of (crudely) outcome-assessing lawmaking that deviates from the principled nature of judicial rights enforcement. 185Id. at 59–60 (majority opinion). If there is a right to liberty of contract, its enforcement should be driven by the sole question of if the government action is of the quality that activates that right, not if the implementation of the right will have desirable consequences. 186 Indeed, while Sunstein and others, see, e.g., Amanda Shanor, The New Lochner, 2016 Wis. L. Rev. 133, 182 & n.207 (2016), characterized Buckley v. Valeo as a new type of Lochnerism, and there has long been anticipation of Buckley’s demise, Burt Neuborne, Money and American Democracy, in Law and Class in America 37, 47 (Paul D. Carrington & Trina Jones eds., 2006) (“Buckley is a rotten tree just waiting to be pushed over . . . .”), Buckley has persisted. That may be because the basic differentiation between contributions and expenditures, while crude and even pathological, is at least facially principled. Samuel Issacharoff & Pamela S. Karlan, The Hydraulics of Campaign Finance Reform, 77 Tex. L. Rev. 1705, 1711 (1999); Michael S. Kang, The End of Campaign Finance Law, 98 Va. L. Rev. 1, 6 (2012). Had the majority adopted a principled justification for its holding that could be the basis for consistent neutral application by courts, Lochner’s logic may merely have become a (legally) tenable new baseline. 187 This is precisely Sunstein’s point. Sunstein, supra note 181. This tenability would have been founded in the fact that a neutral principled application would have allowed elected branches of government the opportunity to advance its policy goals through legislation that satisfied any requirement of neutrality. By basing its holding in the near-explicit judicial promulgation of a policy goal, the Court may have instigated one of the most famous conflicts in American history over the judicial role. 188See Sujit Choudhry, The Lochner Era and Comparative Constitutionalism, 2 Int’l. J. Constitutionalism L., no. 1, 2004, at 1, 14–15.

C. The Twin Dangers of the Quantitative Turn for Partisan Gerrymandering

The need for rights to be founded in principles, rather than merely assert the undesirability of certain outcomes, has powerful implications for the current review of partisan gerrymandering. The arc of the current debate has drawn quantitative analysis of legislative action from social science into judicial dicta and legal scholarship, and then into the core of the current case law. Yet if metrical analysis of legislative action becomes functionally constitutive of rights, the law will develop badly: either the courts will emulate technocratic regulation and canonize certain metrics as dispositive of the legality of a given districting; or the courts will reject quantitative metrics altogether, thus declining to employ a potentially useful piece of evidence.

1. The Danger of Quantitative Fetishization

Courts and reformers have, of course, offered various constitutional rights that might be infringed by partisan gerrymandering, with a focus on the Equal Protection Clause and First Amendment associational rights. Yet these analyses have never been clearly particularized or elaborated in a manner that links them to the evidentiary role of the new statistical metrics. 189 These deficiencies are discussed in more detail in Section IV.A, infra. The Equal Protection Clause’s guarantee of fair representation—itself less than clear in the context of one-person, one-vote 190See Heather K. Gerken, The Costs and Causes of Minimalism in Voting Cases: Baker v. Carr and Its Progeny, 80 N.C. L. Rev. 1411 (2002); Sanford Levinson, One Person, One Vote: A Mantra in Need of Meaning, 80 N.C. L. Rev. 1269, 1274 (2002) (arguing that Baker’s minimalism has undermined its robustness).—entails that no specific threshold of partisan advantage from districting comprises a constitutional wrong, given that there is no guarantee of proportionality. 191 Davis v. Bandemer, 478 U.S. 109, 155 (1986) (O’Connor, J., concurring); Schuck, supra note 25, at 1362–63. Likewise, the First Amendment associational rights identify no such threshold, particularly in a robust two-party system where both parties have access to significant resources and support, as well as periodic political success. 192Bandemer, 478 U.S. at 152 (O’Connor, J., concurring) (“[P]olitical parties are the dominant groups” in American politics.); cf. Briffault, supra note 26, at 407 (difference of “‘pariah’ groups” as requiring protection). The lack of a principle that has specific meaning to a given degree of partisan gerrymandering makes appropriate treatment of the quantitative metrics difficult.

In practice, this has forked into two extreme and undesirable possibilities. The first is that courts’ holdings will be determined solely by quantitative assessments of partisan gerrymandering. In effect, courts would simply consider metrics, and subsequently opine if the partisan gerrymander was so extreme it offended their sensibilities and if so, deem it illegal. Most aggressively, courts might simply assert violation of a certain given metric comprises a prima facie constitutional violation. 193 There is one existing area of election law that has arguably developed in this direction, the 10% safe harbor for state legislatures. See Raviv, supra note 47, at 1063. This rule was cited by the majority opinion in Rucho with approval as justification for reliance on quantitative metrics in the contemporary partisan gerrymandering. See Common Cause v. Rucho, 279 F. Supp. 3d 587, 658 (M.D.N.C.), vacated, 138 S. Ct. 2679 (2018). Yet without a principled framework that contextualizes why the metrical qualities of a gerrymander comprise a constitutional wrong, such judgments would comprise a de facto form of judicial legislation. The form of that lawmaking would be a rule that the presence of certain falsifiable qualities in a districting will render it illegal. Typically, such an instruction would be supported either by a legislated rule that implemented the metrical requirements directly, or by a legislated standard that instructed the courts to specify the appropriate metrics. 194See Eric A. Posner, Standards, Rules, and Social Norms, 21 Harv. J.L. & Pub. Pol’y 101, 106 (1997). However, current partisan gerrymandering reformers argue that such districting should be illegal based on constitutional principles. Yet, as described above, this would be analogous to a political version of Lochner, dictating as a matter of law that certain distributions of (political) resources are “natural.” 195 Sunstein, supra note 181, at 885.

Were the relevant constitutional principle that defended a “trumping” individual right adequately specified and linked to the adduced quantitative analysis, however, the risk that the partisan gerrymandering jurisprudence would become such an output-assessing political Lochner would be greatly ameliorated. In addition, such a conceptual move would expose the nonjusticiability claim advanced by the Vieth plurality as fallacious. However, the contemporary litigation has failed to undertake this move. Instead, a troubling pattern has emerged that suggests the constitutionalization of statistical tests: where courts find quantitative metrics compelling, they identify partisan gerrymanders as illegal, but where metrics are not advanced or judges are dubious of their power, they do not. 196See supra Part I. This pattern began with select opinions in LULAC, and has been characteristic of lower courts’ opinions in the present round of gerrymandering. 197See supra Section I.B. While courts have made uninspired efforts to link these metrics to the Equal Protection Clause and First Amendment association rights, the deficiencies of the efforts are recounted in Part IV infra. Whitford may reveal some leaning in this direction, as the opinions suggest that the standing—a bedrock foundational safeguard of the distinctive role of courts in the constitutional order 198 Massachusetts v. EPA, 549 U.S. 497, 547 (2007) (“The constitutional role of the courts, however, is to decide concrete cases—not to serve as a convenient forum for policy debates.”).—might be answered by more nuanced quantitative tests. Without a clearer principled stance, the jurisprudence might evolve into the courts simply suggesting that certain allocations of political resources are illicit (even if the underlying constitutional wrong is unclear) and support their decisions by committing to certain numerical thresholds.

2. Judicial Rejection of Technical Progress: “Gobbledygook” and Numerical Skepticism

That overreliance on numbers in the assessment of partisan gerrymandering could lead down such a path, however, might lead judges to suspect their general usefulness. Such a sentiment seemed apparent in Chief Justice Roberts’s comment at oral argument that the statistical methods adduced by the Whitford plaintiffs were “gobbledygook” 199 Transcript of Oral Argument at 40, Gill v. Whitford, 138 S. Ct. 1916 (2018) (No. 16-1161).—a general rejection of the value of quantitative analysis to assessing political districting. Likewise, Justice Alito expressed concern that using such a metric as the pivot of legality would be essentially arbitrary. 200Id. at 42–44. Both Justices appeared to reflect the concern at the center of this Article: that use of a metric as the core of a constitutional inquiry would deviate from rule of law. Yet their concerns took the form of questioning the value of quantitative analysis as a general matter. Such an attitude threatens to reject a potentially useful social science tool, and generally exacerbate the discomfort with statistical analysis that has at times plagued courts in the assessment of constitutional rights. 201See Friedman, supra note 18, at 82–83 (describing the inconsistent use of race in the context of the race and the death penalty).

Moreover, once the problematic moral quality that marks partisan gerrymandering is identified, the quantitative methods could prove critical for identifying illicit gerrymanders. As techniques for executing gerrymanders become more sophisticated, it would be likewise necessary to rely on equally sophisticated techniques to identify them, particularly if only politicized districting as deemed to be illegal only if other neutral criteria in districting were disregarded (the type of question the analysis that, for example, the simulation method of Chen and Rodden could identify). 202See Chen & Rodden, supra note 13, at 336. Further, such methods might become likewise important if legislatures sought to avoid claims of gerrymandering by becoming less explicit regarding their intent. In such cases, quantitative analyses might be critical to identify when a given districting was so improbably politicized that its benefit to one party could not be attributed to chance.

IV. Toward a Principled Right for Partisan Gerrymandering

The analysis provided above describes the steps that must be taken to define and enforce a right, and the appropriate role that evidence should play in this process. Most importantly, it establishes that it is first necessary to identify a foundational normative principle that supports the right. Next, the conditions or character of state action that contravene the right must be identified. Finally, the type of evidence that can prove this condition or character must be determined. This results in a coherent sequence: the principle that justifies the right, the character of the conduct that activates judicial rights protection, and the evidence that proves such character.

The rights that have been typically been advanced to explain why partisan gerrymandering is illegal—equal protection of vulnerable groups and political association—fail at various points in this analysis. In particular, they tend to fail to link the asserted normative principles with the classification to districts by political identity. 203 Vieth v. Jubelirer, 541 U.S. 267, 288–89 (2004); Lowenstein & Steinberg, supra note 49, at 74–75. Certain scholarly innovations have offered alternate suggestions for how partisan gerrymandering might be identified as illegal at the level of principle; this section reviews the particularly noteworthy innovations of due process based in historical identification of gerrymanders advanced by Edward Foley, and the synthetic development of a right against government discrimination advanced by Michael Kang. Building on Kang’s innovations, the best principled characterization of partisan gerrymandering is as a type of political competition, and thus the constitutional right to be a right against deprivation of political power from political competition.

A. Existing and Proposed Wrongs of Partisan Discrimination

If courts employ quantitative analyses to evaluate partisan gerrymanders, it must be in the service of a principle that provides guidance regarding why partisan influence in a districting are illicit. The case law has primarily focused upon two possible grounds for deeming partisan gerrymandering illegal: a right to fair representation based in the Equal Protection Clause; and a right to political association founded in the First Amendment. The appeal of each of these arguments is clear. The Equal Protection Clause, the basis of the plurality opinion in Bandemer204 Davis v. Bandemer, 478 U.S. 109, 122–25 (1986). has an intuitive appeal, because it has a long-established history in the racial context of preventing discriminatory action in districting. 205 Shaw v. Reno, 509 U.S. 630, 649 (1993) (use of race in districting must face strict scrutiny); Gomillion v. Lightfoot, 364 U.S. 339, 346 (1960) (discriminatory districting inexplicable other than by race illicit). However, the difference between race and party identification means the law on racial districting cannot be simply translated into the partisan gerrymandering context. Perhaps inspired by Justice Kennedy’s mention in Vieth206 541 U.S. at 314–15. the First Amendment right to association has become a popular alternative. 207See, e.g., Shapiro v. McManus, 203 F. Supp. 3d 579, 595–96 (D. Md. 2016). Such an approach relies on the well-established First Amendment principle that the state commits a constitutional wrong when it harms individuals on the basis of their political identity. 208Id. at 594 (quoting Elrod v. Burns, 427 U.S. 347, 356 (1976)); see also Kang, supra note 71, at 377 & n.165 (discussing Elrod as the best example of the norm against government partisanship in the First Amendment). The difficulty analogizing to these cases, however, is that partisan gerrymandering is inevitably a form of vote dilution and thus a systemic intervention in electoral process. However, the most constitutionally relevant First Amendment cases address patronage decisions and thus identify specific ex post deprivations inflicted on individual persons on their basis of their political identity. To characterize the effect of partisan gerrymanders as such a total deprivation thus does little to clarify the appropriate threshold of illegality.

1. The Equal Protection Clause and the Inaptness of the Racial Analogy

The Equal Protection Clause initially seems a logical port of call for explaining the constitutional wrong inflicted by partisan gerrymandering. The intrinsic nature of discriminatory districting is to treat voters unequally. The argument that such discrimination on the grounds of political affiliation is a constitutional wrong is further supported by the statement in Fortson v. Dorsey that an “apportionment scheme [may not] minimize or cancel out the voting strength of racial or political elements of the voting population.” 209 379 U.S. 433, 439 (1965). Bandemer provides the case lineage, 478 U.S. at 120, as does Justice Steven’s Vieth dissent, 541 U.S. at 320 (Stevens, J., dissenting). The application of this principle to identify districting schemes as illicit on the basis of political bias in a districting, however, is extraordinarily thin. Rather, the Court has typically looked to political realities to vindicate particular districting schemes. The most salient example of this is Gaffney v. Cummings210 412 U.S. 735, 752–53 (1973). which upheld a districting that deliberately preserved an even allocation of political strength—in effect, a bipartisan gerrymander or mutually agreed “shacking.” 211 Issacharoff & Karlan, supra note 10, at 551–52. More generally, the opinion asserted that political considerations will inevitably contribute to districting, and that attempting to exclude them is both impossible and potentially even perverse. 212Gaffney, 412 U.S. at 753. It also repeated the commonplace that a districting design that “fence[s] out” a racial or political group may be illicit. 213Id. at 754.

The lack of clear precedential guidance as to when purely political interests in districting are illicit is the foundation of the Vieth plurality’s assertion of nonjusticiability. Yet the lack of clear guidance has also had the jurisprudential effect of inducing judges opposed to partisan gerrymanders to turn to the racial gerrymandering jurisprudence. 214 Eisler, supra note 70, at 266. This analogical reasoning is embedded throughout liberal opinions on partisan gerrymandering, beginning with Bandemer 215 Davis v. Bandemer, 478 U.S. 109, 125 (1986) (“[T]hat the claim is submitted by a political group, rather than a racial group does not distinguish it in terms of justiciability.”). and continuing on to the Vieth 216 LULAC, 548 U.S. 399, 469–70 (2006) (Stevens, J., dissenting) (treating party affiliation as stable and analogizing to cases on racial gerrymandering to the discriminatory effect on Democrats); Vieth v. Jubelirer, 541 U.S. 267, 320 (2003) (Stevens, J., dissenting) (Gaffney supports the idea that racial and political discrimination are equivalently illicit); id. at 337–38 (a political gerrymander is as objectionable as a racial gerrymander); id. at 344 (Souter, J., dissenting) (observing that excessive presence of both race or partisanship can render a districting plan illicit); see also Whitford v. Gill, 218 F. Supp. 3d 837, 867–73 (W.D. Wis. 2016) (carefully reviewing the case law and relying on Bandemer’s conclusion that racial and political gerrymandering are equivalently justiciable). and LULAC 217 548 U.S. at 469–70 (Stevens, J., dissenting) (analogizing to cases on racial gerrymandering). dissents. Whitford reinforces this by turning to the cases on race to address the standing problem, 218 Gill v. Whitford, 138 S. Ct. 1916, 1930 (2018). despite the fact that the wrong of racialized districting is clearly defined in the law, while the wrong of partisan gerrymandering is not.

Yet this raw analogizing between race and party identification is unhelpful, in no small part because the racial gerrymandering jurisprudence relies on a distinct principle of constitutionally mandated equal government treatment. Drawn directly from the Equal Protection Clause, this is a classic “trumping” right forged in the normative principle that racial classifications are presumed to be undesirable. When a party can show that a government action is based on a racial classification, the government action must face the onerous burden of strict scrutiny; while a law that relies on racial classification is not necessarily illegal, the mere use of race (whether explicit or by the fact that race is the only application) will demand narrow tailoring to satisfy a compelling state interest. 219 Miller v. Johnson, 515 U.S. 900, 904 (1995). This case continues to provide the bedrock reasoning in the assessment or racial districting. See Bethune-Hill v. Va. State Bd. of Elections, 137 S. Ct. 788, 797 (2017). While the Court has acknowledged that management of racial classifications in districting requires special delicacy because there is almost always awareness of race when drawing boundary lines, 220Bethune-Hill, 137 S. Ct. at 797. this does not modify the basic principle—equal government treatment of races—that undergirds the right. 221Miller, 515 U.S. at 904. To account for the delicacy of legislative districting, while districts are illegal only when race “predominates” over neutral factors, the intrinsically suspect nature of racial purpose itself does not change, and mere conformity with traditional districting principles will not cleanse a districting plan that is motivated by race. 222Bethune-Hill, 137 S. Ct. at 798. Thus, the racial gerrymandering jurisprudence is firmly grounded in a principle of equal treatment of races. Moreover, equal treatment on the basis of race, due to the vulnerability of minority groups, has a well-developed tradition in legal philosophy and in jurisprudence. 223See John Hart Ely, Democracy and Distrust: A Theory of Judicial Review 75–76 (1980) (observing the tradition from United States v. Carolene Products Co., 304 U.S. 144, 152 n.4 (1938), onwards of protecting insular and discrete minority groups).

Applying the limitations upon discriminatory racial districting to discriminatory partisan districting results in two sets of linked problems. First of all, it is far less straightforward to extract the basic constitutional principle that supports a finding of illegality for party identity from basic interpretation of the constitutional principles. 224See Dworkin, Empire, supra note 19, at 87–88 (describing the role of interpretation of principles). That the Fourteenth Amendment generates a “trump” against racial classifications is unequivocal, based in American “constitutional and demographic history.” 225Miller, 515 U.S. at 904. No such principle exists for party identity, and indeed, attempting to instantiate one would create immediate insoluble difficulties: the very purpose of competitive democracy (i.e., democracy that does not require consensus) is to select, via some fair procedure, some group of policies favored by one segment of the population over that favored by another, thus presumably harming the loser. In a two-party system, the winners and losers are organized into blocs engaged in contestation. To characterize this disfavoring of one bloc as rights-violating would effectively realize a general principle against democratic choice. 226 Lowenstein & Steinberg, supra note 49, at 74–75.

Moreover, it is difficult to establish a general principle of nondiscrimination on the basis of political identity in districting. This is not merely because eliminating political interests from districting is challenging, but because even many of the neutral factors in districts, such as making districts compact or “avoiding contests between incumbents,” 227 Karcher v. Daggett, 462 U.S. 725, 740 (1983). do not only correspond to political interests, 228 As Bethune-Hill v. Va. State Bd. of Elections, 137 S. Ct. 788, 798 (2017), observes, mere consequential conformity with neutral districting principles will not rescue an otherwise illicitly motivated districting plan. but are inextricable from them. Incumbent protection, for example, is ultimately a feature of political identity, as is “preserving the core of prior districts.” 229Karcher, 462 U.S. at 740 (listing neutral factors in districting). The Karcher v. Daggett factors, in effect, define the very content of politics. The existence of these features, and their political implications, are specific aspects of a deeper feature of districting—it is itself constitutive of political values in a democracy. That is, a districting does neutrally divide voters into equally apportioned units, because there is no true baseline of “neutral.” There are merely differing balanced values. One of those values may, of course, be racial classifications, but the Fourteenth Amendment mandates that such discrimination must be disfavored through strict scrutiny. 230Miller, 515 U.S. at 904. Making such a similar determination regarding political identity, however, is not possible without enforcing some general principle of political neutrality, which would at a minimum involve a general judicial theory of political justice and may have the specific effect of enforcing proportional representation. 231See Davis v. Bandemer, 478 U.S. 109, 145 (1986) (O’Connor, J., concurring). As Peter Schuck has noted, for example, in the absence of a commitment to proportional representation, a system that offers a “bonus” to a party that wins an electoral victory in terms of the opportunity to set the next map does not seem necessarily unfair; such a judgment is one of substantive politics. 232 Schuck, supra note 25; see also Lowenstein & Steinberg, supra note 49, at 74.

2. Associational Rights and Equalities of Outputs

Some jurists have turned to First Amendment associational rights as a basis for identifying illegal partisan gerrymanders, 233See Shapiro v. McManus, 203 F. Supp. 3d 579, 588–89 (D. Md. 2016); see also Whitford v. Gill, 218 F. Supp. 3d 837, 875 (W.D. Wis. 2016). See generally Daniel P. Tokaji, Gerrymandering and Association, 59 Wm. & Mary L. Rev. 2159, 2160–62 (2018); Daniel P. Tokaji, Voting Is Association, 43 Fla. St. U. L. Rev. 763 (2016). The latter piece has some consonance with the ultimate proposal of this Article’s conclusion insofar as it identifies entrenchment as the basis for an associational right against gerrymandering. Tokaji, Gerrymandering and Association, supra, at 2190–91, 2199. However, as this Article argues, the First Amendment jurisprudence does not clearly enough articulate a definitive anti-entrenchment rationale, particularly as the Celebrezze/Williams line of cases is primarily concerned with depriving minority parties of any associational rights, rather than the give-and-take in a two-party system. Cf. Bandemer, 478 U.S. at 152 (O’Connor, J., concurring) (parties are dominant actors). a possibility enthusiastically adopted by Justice Kagan in her Whitford concurrence. 234 Gill v. Whitford, 138 S. Ct. 1916, 1938 (2018) (Kagan, J., concurring). This tack relies on the principle that the government may not infringe on the right of individuals to effectively form organizations, and a line of precedent that holds that this principle prohibits government action that reduces the ability of persons to organize politically, or punishes individuals for their protected right to associate. 235Whitford, 218 F. Supp. 3d at 875. The First Amendment associational precedent bearing on politics, however, is difficult to translate directly into a principle that indicates when a constitutional wrong occurs in the context of two dominant parties engaged in a political struggle. The cases that bear most directly on election rights, Williams v. Rhodes 236 393 U.S. 23 (1968). and Anderson v. Celebrezze237 460 U.S. 780 (1983); cf. Crawford v. Marion Cty. Election Bd., 553 U.S. 181, 190 (2008) (stating that Anderson’s balancing test is now the guiding precedent); Foley, supra note 36, at 674–76 (describing the “Anderson–Burdick balancing test,” which uses a balancing test to assess whether a state’s regulation of the right to vote violates the Equal Protection Clause). This balancing test, however, merely inquires as to when the right to vote trumps the validity of government regulation of the electoral process. Thus, unlike partisan gerrymandering, which requires a threshold assessment of the severity of the practice itself, this balancing test simply weighs a “trump” against competing interests. The right to vote, as it were, is absolute. were concerned with situations where an electoral setup impaired the ability of an organization (marginal third parties) to gain access to the ballot at all. The constitutional wrong was the impairment of the ability of an organization to realize effective expression in the form of political voice238See Celebrezze, 460 U.S. at 788 (“[E]ach provision of these schemes . . . inevitably affects . . . the individual’s right to vote and his right to associate with others for political ends”); Williams, 393 U.S. at 38–39 (arguing that barriers hinder the ability to organize as a third party at all); cf. Shapiro, 203 F. Supp. 3d at 605 (Bredar, J., dissenting) (arguing that partisan gerrymandering does not infringe the type of First Amendment associational rights in a manner germane to the type of conduct that such rights are meant to protect, because partisan gerrymandering does not impair voters’ ability to “affiliate with the party of their choice, to vote, to run for office if they wish, and to participate in vibrant political debate wherever they find themselves”). Likewise, the patronage cases 239 Rutan v. Republican Party of Ill., 497 U.S. 62 (1990). stand for the proposition that individuals may have their political freedom silenced if they can be dismissed from a government job on the basis of political affiliation. 240Id. at 75–77. The First Amendment right to associate protects the fundamental right to access the political process through organization or affiliation. 241Celebrezze, 460 U.S. at 787. Conversely, partisan gerrymandering inevitably impacts not the presence of political participation, but rather the efficacy of that participation. However, as with equal protection, 242 Davis v. Bandemer, 478 U.S. 109, 131–32 (1986) (stating that a given districting reduces the likelihood a given group will be able to elect its representative does not comprise a rights violation); Shapiro, 203 F. Supp. 3d at 591 (“[C]itizens have no constitutional right to reside in a district in which a majority of the population shares their political views and is likely to elect their preferred candidate.”). the associational right cannot be a guarantee of success or equality of political outcomes (as this would again devolve into a guarantee of proportionality).

Thus, First Amendment associational rights case law on elections indicates government conduct may not prevent associations altogether. The “trumping” right is a right to exercise one’s political voice; the relevant character of the state action is “chilling” of speech through a prohibitive limitation upon political participation, either by raising a process barrier or imposing a penalty. Conversely, the problem of partisan gerrymandering is one of threshold identification: a partisan gerrymander does not exclude a party from participation, but rather limits its success. The challenge—as with relying on the law of fair representation—is identifying where this limitation becomes a constitutional wrong. Concluding that the First Amendment associational rights entail that there can be no deleterious effect from districting upon representative success due to party affiliation would mandate proportionality (though this is likely the logical implication of the First Amendment analysis in Kagan’s Whitford concurrence). 243 Gill v. Whitford, 138 S. Ct. 1916, 1939 (2018) (Kagan, J., concurring) (characterizing a First Amendment claim against gerrymandering as one that would “debilitate[] their party or weaken[] its ability to carry out its core functions and purposes”). Yet there is no clear way to adapt the First Amendment precedent to identify the appropriate level of baseline deleteriousness that is unconstitutional. 244Cf. Tokaji, Gerrymandering and Association, supra note 233, at 2199. Tokaji’s suggestion that the burden must be “substantial” and “enduring” does nothing to resolve the basic query raised by, inter alia, the Vieth majority: How enduring? How substantial? Thus, any attempt to identify evidence that a party’s chance of success or level of representation has been reduced—precisely the quantitative evidence offered in the new litigation—is of little bearing on First Amendment rights.

The emphasis on quantitative tests in gerrymandering litigation makes the need for innovation of a principle especially stark. The inaptness of equal protection and the First Amendment associational rights, as so far developed in the case law, is that they rely on principles that starkly prohibit certain types of state conduct. Governmental action may not rely on race without surviving strict scrutiny; it may not exclude citizens’ voices. The core cases of such action are readily identified: Was race used to classify voters? Were citizens prohibited from organizing in a manner that prevented their self-expression? While courts then must use less determinative tests, such as the question of if a racial classification passes strict scrutiny or if a nondiscriminatory association restriction is balanced by a state’s other interests, 245Celebrezze, 460 U.S. at 788–89. the initial identification has a clarity that it does not in the context of partisan gerrymandering. The bad character of a partisan gerrymander is a wrong of excessiveness itself, as Gaffney suggests. None of the grounds that have been central to the debate in the courts over partisan gerrymandering lend themselves, however, to such a test of excess.

3. Scholarly Innovations: The Prudent Turn Back to Constitutional Principle

Given the recent prominence of partisan gerrymandering, it is unsurprising that it has received diverse scholarly attention. While much of this scholarship has been dedicated to the development of new quantitative metrics or linear evolution of the First Amendment or Equal Protection Clause, some scholars have innovated new approaches.

a. Edward Foley: The Due Process Clause and Fair Play

Foley argues that a turn to the Due Process Clause, informed by the history of partisan gerrymandering, can provide a fair play principle that will indicate when partisan gerrymandering is unconstitutional. The foundation of Foley’s proposal is sympathetic to this project: after observing the failures of the Equal Protection Clause to adequately define partisan gerrymandering, he suggests an alternative principle may be necessary. 246 Foley, supra note 36, at 686. Foley’s analysis focuses on the balancing test from Anderson, and curiously does not engage extensively with the analogy to racial gerrymandering that has motivated liberal Justices. He then, however, makes a curious move, and turns to the history of the Fourteenth Amendment to argue that the Constitution contains a principle of “fair play” that “constrains partisan competition,” 247Id. at 688. and then argues that the original partisan gerrymander provides a guiding standard (even as he concedes that “the framers of the Fourteenth Amendment themselves may not have combined these two propositions”). 248Id. at 711. He then argues the original partisan gerrymander can serve as the benchmark that operationalizes judicial assessments of partisan gerrymanders. 249Id. at 720.

Foley’s approach is creative, yet it may distort the Dworkinian treatment of history and precedent. It is heavily dependent upon the appropriateness of applying two sets of historical norms to achieve law as integrity. Yet Dworkin himself disavows the need to have such an integral commitment to history:

Integrity . . . commands a horizontal rather than vertical consistency of principle across the range of the legal standards the community now enforces. It insists that the law . . . contains not only the narrow explicit content of . . . decisions but also, more broadly, the scheme of principles necessary to justify them. History matters because that scheme of principle must justify the standing as well as the content of these past decisions. 250Dworkin, Empire, supra note 19, at 227–28.

In essence, history is important because it provides context for what governmental practices and legal principles are legitimate in the present.

This is not to suggest that Foley’s analysis is illegitimate, but his approach is overly dismissive of the path of constitutional interpretation of the rights of districting and, even more explicitly, of the relevance of theoretical treatments of gerrymanders, including quantitative analysis. 251 Foley, supra note 36, at 726. The obligation of the Dworkinian judge is to balance competing principles that in the present have normative legitimacy, and provide an interpretation that builds on existing precedent. 252Dworkin, Empire, supra note 19, at 227. Dworkin thus analogizes the relationship of principled development of legal integrity over time to a “chain novel” in which each judge sees the next opinion as the best subsequent chapter. 253Id. at 228. Foley’s invocations of the Due Process Clause and the original partisan gerrymandering are original, but comprise new “works,” rather than contributions to the existing “novel.” His anchors are both too far in the past and too thoroughly separated from contemporary precedent on germane matters to offer the best path forward.

b. Michael Kang: The Invalidity of the Partisan Justification

In a methodologically more modest but doctrinally meticulous piece, Kang offers an alternative argument for why partisan self-interest may be generally illegitimate as a justification for partisan action, including in the context of districting. While Kang’s analysis may, if operationalized, generate undesirable levels of judicial intervention in political conflict, it suggests a principle that accommodates normative and legal realities of partisan gerrymandering while properly defining the role of quantitative information. Kang observes that the contradiction that has riven the law of partisan gerrymandering—that partisanship in districting is acceptable to a degree, but will be illegal if it goes too far 254 Kang, supra note 71, at 353.—dissolves once it is recognized that “[p]artisanship simply does not count . . . as a legitimate government interest to justify official government decisionmaking.” 255Id. at 354. In this respect, Kang’s analysis reaches the same legal point as the majority opinion in Rucho. Kang then engages in a careful and comprehensive parsing of constitutional doctrine to argue that Vieth is exceptional in asserting that courts ought not to enforce nonpartisanship as a principle. 256Id. at 403.

By the lights of the analysis presented in this Article, Kang’s approach has much to commend it. From a Dworkinian perspective, it effectively derives an operationalizable principle from a careful interpretation of prior case law, thus serving the end of law as integrity. Kang’s approach, in other words, can be defended as the implementation of the Dworkinian project of creating a more coherent and normatively justified jurisprudential universe. In addition, Kang satisfies the novel challenge identified for rights as implementing principles and requiring characteristic state action. Kang identifies partisan discrimination as an illicit characteristic, and thereby indicates what “trumping” right courts should identify when assessing partisan gerrymanders. 257Id. at 406. It thereby solves the puzzle that has motivated this argument—how can quantitative data be used by courts?—by allowing them to operate as evidentiary indicators of partisan purpose.

The problem facing Kang’s article as the final word, however, is the potential breadth of the principle. Kang’s analysis differentiates between permissible political criteria and impermissible partisan discrimination in districting. 258Id. at 352. He likewise makes a fine distinction between differing categories of governmental role when he observes that officials “advancing partisan priorities . . . define[s] democratic elections and public life,” 259Id. at 378–79. but that using state action to achieve partisan ends is illicit. Perhaps no single case makes the fragility of these distinctions more delicate than Gaffney. When Kang argues that a bipartisan gerrymander is political rather than partisan, 260Id. at 369. he reveals a degree of conceptual arbitrariness. Each party in Gaffney participated because it was to their partisan benefit; it was not a political act of general beneficence or public good, 261See Cain, supra note 34, at 159 (describing every seat under such a districting as “inefficient”); Issacharoff, supra note 11, at 598–99 (identifying the self-serving nature of such a districting). but simply a move of mutual self-interest. To call Gaffney political rather than partisan is interpretively arbitrary in a manner that requires examination of outputs—what is the effect of the state action?—in a manner that is suboptimal for rights.

Moreover, Kang’s approach, and the distinctions between individual official action and official state action, and between the political and the partisan, suggests a broader challenge to his principle: it would evolve into the courts setting a general baseline for appropriate government action. 262 It thus would run the risk of drifting toward the type of Lochnerism described in Section III.B.3, supra. Thus, it would force courts into a type of general substantive regulation of democratic action: courts would be forced to make a series of general line drawing differentiations between acceptable political behavior and unacceptable partisan behavior, or between conduct that officials took on behalf of themselves, and that which they took on behalf of the state. Drawing these lines would demand that courts develop a “root-and-branch” 263 Vieth v. Jubelirer, 541 U.S. 267, 285–86 (2004). theory of acceptable legislative conduct. Judicial enforcement of Kang’s principle would thus be to move courts from identifying discrete constitutional wrongs based on unacceptable government action to generally making a set of direct normative judgments regarding the propriety of governmental conduct. Because of the degree of interpretive breadth it requires, Kang’s principle would arrogate to the courts general review of government action, rather than merely identify when partisan gerrymandering inflicts a constitutional wrong upon distinct parties.

B. A Principle-Based Right Against Partisan Gerrymandering

Kang’s analysis, however, points the way to a principle that most precisely characterizes the values necessary to condemn the practice and precisely identifies how quantitative data should be used. His identification of partisan discrimination as the misconduct underlying a partisan gerrymandering claim describes the general content of the appropriate principle. As indicated above, however, his principle is overbroad. In order to avoid excessively general judicial review of democratic decision-making, it is necessary to more precisely characterize the legislative intentionality that makes the partisan gerrymandering offensive. This section argues for a principle prohibiting excessive partisan competition in the design of electoral procedures. Specifically, when a districting is predominated by serving the future electoral advantage of the dominant party, it should be deemed illegal. Such a principle can be realized through the right to fair representation, though this requires imputing substantive values to the right to equal voting power contained in the Equal Protection Clause. Evidence—including novel quantitative metrics—can aid in proving the intent to district is dominated by partisan competition by showing that alternative explanations are inadequate.

1. Preventing Democratic Distortion from Partisan Competition: A Predominant Purpose Standard

While partisan gerrymandering is typically identified as offensive to democratic sensibilities, 264 Klarman, supra note 11. it is helpful to go beyond aesthetic or moral intuitions. The actual deleterious impact of such politicized districting is the ostensible distortion of electoral outcomes. If courts are to avoid simply dictating the appropriate fundamental terms of political engagement by the electorate, judicial intervention to prevent such distortion is only appropriate when it impairs popular self-determination. 265 Lowenstein & Steinberg, supra note 49, at 75. This suggests that the appropriate principle should be directed toward electoral procedures, rather than a broader principle against partisan discrimination. Furthermore, given the seemingly inevitable role that political factors play in the drawing of district lines, seeking a partisan advantage in democratic procedures is better identified as the defining trait. Partisan competition regarding substantive policies to obtain the approval of voters is the lifeblood of responsive democracy; however, when the form of this partisan competition switches from competing over policies to changing the terms of democratic selection procedure, it is a type of legislative self-dealing. Given this, the most reliable indicator of this excess is not the effect of the partisan competition, but rather that the intention of the relevant action is purely to achieve distortive advantage in democratic decision procedures. Such a principle is normatively self-justifying, as partisan gerrymanders would be illegal only when they reflected the explicit governmental intention to undercut the self-determination that is constitutive of democracy. 266 Dworkin provides the philosophical justification for this by explaining how interpretation of a fair representative right in a mutually respecting community would emerge from a view of law as integrity. Dworkin, Empire, supra note 19, at 213, 225–26.

The right that serves as the best vehicle for this principle is the right to fair representation, the basis of the one-person, one-vote theory. 267 That the right is best understood as the right to fair representation suggests that, as the progeny of Baker v. Carr, it might best be understood as defended by the Equal Protection Clause. However, a due process argument similar to that advanced by Foley, supra note 36, could also provide a precedential lineage. Fair representation could be interpreted as entailing that individuals suffer a constitutional harm when their districts are drawn with the predominant intent to discriminate against them to favor the gerrymandering party, and the districting plan cannot be justified by other legitimate districting criteria. 268 The closest precedent to this test is Justice Stephen’s suggestion that the Shaw predomination test should apply. Vieth, 541 U.S. at 339.

This approach helpfully operationalizes the general idea that partisan discrimination is illicit while narrowing the likelihood that judicial intervention will displace appropriate self-determination by the electorate. The narrowness of the principle has both normative and functional aspects. Normatively, it reduces the likelihood that courts will nullify partisan gerrymanders or generally intervene in the democratic process in a manner that subordinates the electorate’s self-determination. It also reduces the likelihood that the principle would morph into prospectively antidemocratic judicial review of legitimate justifications for government conduct. Indeed, it also reduces the risk that, given the wide variety of (contested) possible configurations of party politics, that the Supreme Court would impose a specific vision of how party competition should unfold. 269 Indeed, the Court showed a form of such sensitivity in Gaffney v. Cummings, 412 U.S. 735 (1973), though whether the case is correctly decided is a separate question. Yet there are reasons to recognize that party leadership, rather than rank-and-file voters, ought to play a central role in designing terms of political contestation. See Sujit Choudhry, Resisting Democratic Backsliding: An Essay on Weimar, Self-Enforcing Constitutions, and the Frankfurt School, 7 Global Constitutionalism 54, 67 (2018).

The focus on legislative intention allows for adaptation of the predominant intent test for illicit districting from the racial gerrymandering cases. 270 Bethune-Hill v. Va. State Bd. of Elections, 137 S. Ct. 788, 798 (2017); Miller v. Johnson, 515 U.S. 900, 916 (1995). However, the main difference would be a focus on legislative intent: since partisan identity is not a protected class, a successful claim that a partisan gerrymander is a constitutional wrong would have to show that the legislature’s overriding motive in its design of districts was to materially favor one party. The purpose of such a focus on intent is to ensure that courts do not generally commandeer the substantive construction of the political process but intervene only when legislative action is a facial affront to procedural neutrality. This would helpfully differentiate the future test of partisan gerrymandering from the vague effects test proposed in Bandemer271See supra Section I.A.1.

2. The Bounded Role of Metrics in Assessing Intentionality

The focus on intent would also clarify the role of quantitative data. Legislatures might seek to conceal partisan advantage under the pretext of advancing legitimate districting criteria such as preserving district cores or protecting incumbents. Quantitative data could identify when the assertion that a district plan was guided by neutral principles was pretextual. Methods such as the Chen and Rodden simulation test would be particularly helpful for this purpose, as they could assess the likelihood that a purportedly neutral test was actually so. 272 Chen & Rodden, supra note 13, at 334, 340 (explaining the role of neutrality in the simulation approach and the results of running the simulations). Sustained efficiency gaps might likewise indicate the presence of partisan gerrymanders, but such a pattern would need to be considered in the presence of other possible explanations for such a gap. 273See Cover, supra note 77, at 1213 (a plan may have a low efficiency gap by sacrificing other desirable features). Such evidence would not deem a partisan gerrymander unconstitutional on its own, however, because it would only be in the service of allowing a court to confirm that the legislature exploited its districting power to achieve an illicit advantage as the predominant explanation for the districting. Such a bounded evidentiary role for quantitative data would ensure that its treatment remained principled, rather than the right being misconstrued as dependent upon the outputs of policy decisions.

However, because legislative intent would be the critical indicator for identifying excessive partisan competition, metrics could not act as the core of the right. From an evidentiary perspective, the legislative record—and exclusion of the competing party from the districting process—would be pivotal evidence along with metrics. At a conceptual level, they would never be sufficient to define the right, because the legal identification of the wrong would depend upon a conclusion regarding legislative intent. This approach would, however, serve to discourage courts from drifting into the realm of policymaking, and keep the judicial role focused on norms based in the rule of law rather than on advancing particular outcomes.

The focus on intentionality and the need to contextualize statistical indicia would enable parties to attempt to sneak in marginal partisan advantages through pretextual adherence to neutral districting factors. However, it is unclear that the rigidly defined exclusion of partisan interest from legislative districting by judicial means is desirable. 274Cf. Ariz. State Legislature v. Ariz. Indep. Redistricting Comm’n, 135 S. Ct. 2652, 2658–59 (2015) (confirming that a democratically selected referendum may perform just this purpose). Excessively precise judicial determination of appropriate terms of democratic engagement comprises a type of Lochnerism. Legislative acts, including districting, are ultimately expressive of democratic will; it is only necessary that the judiciary intervene when the legislative action threatens to derail democratic accountability. The Bandemer plurality test suggested this to be the case, but by focusing on the output effects of a districting, it placed courts in the untenable position of being forced to deem districtings so severe as to thwart democracy. 275See supra Section I.A.1. By focusing on the legislatures’ intention, the principle of excessive partisan competition enables courts to enforce a principle of legislative conduct rather than seek to achieve a policy. 276See supra Section III.B. But akin to the Equal Protection Clause right against racial classifications and the First Amendment right not to have society’s voices and associations chilled by governmental action, the principle against excessive partisan competition in electoral procedure monitors the character of government behavior, rather than its outcomes.

This principle would ultimately justify a fairly minimalist judicial intervention, potentially thwarting only the most egregious or extreme gerrymanders. Yet the claim that underlies much of the fervency of reformers is that gerrymanders are fundamentally distorting democracy and impairing general accountability; judicial nullification of only these gerrymanders might be sufficient. Moreover, more aggressive judicial intervention could only be justified by the presumption that the electorate is incapable of policing the government, suggesting a far more egregious failure of the democratic process. Should more effective forms of partisan manipulation of district lines emerge, a new principle might need to be engineered to address that particular malady. But given the current state of partisan gerrymandering jurisprudence, eliminating the most egregious and explicit examples of partisan opportunism in districting, while respecting the primacy of popular autonomy, is the favorable approach.

3. Solving the Standing Problem: Clear Harm from Clear Wrongs, and Clarifying the Consequences of Judicial Intervention

One of the by-products of the predominant purpose test for partisan gerrymandering is that it would allow for a simple test of when a plaintiff has standing. If a plaintiff is in a district whose shape was predominantly determined by the opposing party’s desire to achieve political advantage (whether the form of such shape was “cracking” or “packing”), the plaintiff would have standing. The injury-in-fact would be a vote whose meaning (systemic, but more importantly for standing, personal) is determined not by legitimate districting criteria, but by the illicit intent to serve partisan ends. Standing would thus be satisfied if a member of a minority party from a district that was illegally manipulated brought suit. This simple connection between the constitutional wrong and the standing test would minimize the potential for standing to become a vehicle of tactical manipulation or disguised merits review. 277See Elliott, supra note 149, at 466 n.35 (collecting criticism of standing as readily manipulated to surreptitiously serve judges’ substantive preferences).

The simplicity of standing under the predominant purpose test would also reduce the risk that the law on gerrymandering will be further waylaid by doctrinal tangents, as Whitford seems to threaten. The case shifted the terrain of the partisan gerrymandering debate into the realm of standing and suggested that this standing problem requires further metrical innovation to resolve. 278 Gill v. Whitford, 138 S. Ct. 1916, 1933 (2018) (suggesting the standing query requires inquiry into the effect on particular citizens); id. at 1936 (Kagan, J., concurring) (suggesting that the appropriate solution is further creation of metrics).Whitford thus threatens to further confuse analysis of partisan gerrymandering by adding another layer of procedural complexity that obscures cutting to the substantive question at issue, and to do so through recourse to quantitative methods that displace a constitutional question. 279See Scalia, supra note 149, at 894 (standing serves the end of separation of powers). The predominant purpose test obviates the need to numerically parse if a given voter has been harmed by partisan gerrymandering by identifying the harm with pure legislative intention.

The ease by which a plaintiff could achieve standing under a predominant purpose test, however, might intimate that the test undermines separation of powers. Under the test, individuals from minority parties could readily turn to the courts to seek districting changes, rather than needing to establish a highly particularized impact upon their own voting power. This may be a true observation—and it clarifies the broader implications of allowing courts to intervene in the politics of district line drawing. The harm inflicted by partisan gerrymandering is by its very nature diffused across groups; 280 Gerken, supra note 73, at 1671–72. the question of whether courts should strike down partisan gerrymanders is a substantive query regarding judicial involvement in the composition of parties and the terms of their conflict. 281See Eisler, supra note 70, at 233. The simplicity of the standing question under the predominant purpose test merely reveals that the predominant purpose test lays bare the substantive dispute at hand. It thus cuts through the haze of technicalities generated by both Vieth and Whitford.

Conclusion

This Article has explained why it is troubling if courts resort to legally arbitrary metrics to strike down partisan gerrymanders. Politicized districting reflects an outcome of the democratic process, 282 Lowenstein & Steinberg, supra note 49, at 76; Schuck, supra note 25. and courts do something exceptional when they deem expressions of democratic governance unconstitutional. 283See Bickel, supra note 152, at 16. Such exceptional judicial intervention is legitimate when it protects foundational values of the polity. 284Dworkin, Empire, supra note 19, at 213. To strike down gerrymanders without clearly identifying the value that gerrymanders offend displaces the electorate’s will with judges’ own policy preferences. This displacement is not merely undemocratic; it produces fragile law characterized by arbitrary and conditional reasoning, and which is unlikely to last as precedent. 285See supra Section III.B.3. This is true regardless of the technical sophistication of the evidence that informs courts’ substantive judgments.

Courts have shown an alarming tendency to over-rely on statistical metrics to inform constitutional rights in the partisan gerrymandering litigation. Judges have justified the nullification of districtings primarily by pointing to the outputs of novel statistical analyses. Yet, unless shown to contravene a legal right, the effect of government action is the unabashed domain of policy. The fruitlessness of the partisan gerrymandering litigation over the past thirty years has indicated a failure of legal imagination, not the inability to detect partisan interest in districting. What is needed is the innovation of a right—an explanation as to how districting by political identity offends a specific value foundational to the polity. This Article has thus sought to redirect attention on the partisan gerrymandering debate back toward the principled reasoning that is the proper domain of courts.

The broader infiltration of quantitative methods into constitutional law is only likely to accelerate. The sequence by which this transmission occurred in partisan gerrymandering anticipates how such infiltration will occur, as quantitative analysis jumped from social science to legal scholarship to judicial reasoning. 286See supra Section I.B.2. Preserving the rule of law and the norms of democratic accountability requires that such powerful tools be constrained to their appropriate role in judicial reasoning: as evidence. Yet properly defining the role of such metrics also increases the likelihood that they will receive due respect for what they can offer in legal decision-making, 287See supra Section III.C.2. rather than be disregarded or manipulated because of judges’ political views.

Preserving the uniquely principled character of judicial reasoning is of genuine practical importance. It not only ensures that courts remain dedicated to law instead of quasi-governance but prevents technocratic overdetermination of rights protection. If courts inform constitutional rights by quantitative methods, they give tremendous advantage to parties capable of advancing more sophisticated methods. Yet there is no reason to believe there will be a correlation between legitimate constitutional claims and access to sophisticated technical tools. 288 Einer Elhauge provides a structural parallel that explains the rule of lenity: interpretation should favor those who are least likely to be able to lobby to have their interests represented in drafting. Einer Elhauge, Statutory Default Rules: How to Interpret Unclear Legislation 169 (2008). A parallel principle suggests judicial caution toward metrics where more powerful members of society might have better access to them. Indeed, if any correlation were to be expected, it would be that social groups with the least need of judicial protection would have the most powerful tools. This reinforces the need of courts to found constitutional analysis on principled assessments of the character of government action. To do so not only retains the logical integrity of judicial action but will serve the interests of justice and social progress.

 

Footnotes

*Associate Professor of Public Law, University of Southampton Law School; Research Associate, Centre for Business Research, University of Cambridge. PhD, Harvard University Department of Government; JD, Harvard Law School. Many thanks to Michael Kang, George Brown, Jim Gardner, Rick Pildes, Richard Briffault, Adriaan Lanni, Bruce Mann, Eric Beerbohm, Billy Magnuson, Greg Conti, Jeremy Green, Adam Lebovitz, Evan Miller, and Sam Zeitlin. The ideas and flaws in this Article are my own.

1See, e.g., Eli J. Finkel et al., Online Dating: A Critical Analysis from the Perspective of Psychological Science, 13 Psychol. Sci. Pub. Int. 3, 6 (describing the role of algorithmic matching in online dating).

2See, e.g., William Magnuson, Regulating Fintech, 71 Vand. L. Rev. 1167, 1188 (2018) (describing the systemic risks of various quantitative approaches to trading).

3See, e.g., Michael Lewis, Moneyball: The Art of Winning an Unfair Game (2003) (the seminal account of how sabermetrics and quantitative analysis in baseball transformed the approach to management strategy and tactics).

4 Adam Chilton & Eric Posner, The Rise of Statistics in Law, Eric Posner (July 24, 2015), http://ericposner.com/the-rise-of-statistics-in-law/ (detailing one of the simplest metrics: the increased prevalence of the phrase “statistical significance” and its variant in legal settings).

5See, e.g., Eric A. Posner, Economic Analysis of Contract Law After Three Decades: Success or Failure?, 112 Yale L.J. 829, 829 (2003) (“Traditional doctrinal analysis [in contract law] exerts less influence . . . and enjoys little prestige [since 1970].”).

6See, e.g., James Bohn & Stephen Choi, Fraud in the New-Issues Market: Empirical Evidence on Securities Class Actions, 144 U. Pa. L. Rev. 903, 912 (1996) (statistical analysis of impact of pursuing different types of suits); Bradford Cornell & R. Gregory Morgan, Using Finance Theory to Measure Damages in Fraud on the Market Cases, 37 UCLA L. Rev. 883, 885–87 (1990) (doctrinal relevance of quantitative methods to identify appropriate measure of damages in Rule 10b-5 suits).

7See, e.g., Einer Elhauge, Defining Better Monopolization Standards, 56 Stan. L. Rev. 253, 343 (2003); Thomas G. Krattenmaker & Steven C. Salop, Anticompetitive Exclusion: Raising Rivals’ Costs to Achieve Power over Price, 96 Yale L.J. 209, 250 (1986) (providing tests for anticompetitive behavior that weaves together principles of competition with certain quantified levels of market power to identify actionable harm).

8 For an exemplar of such scholarship, see Nicholas O. Stephanopoulos & Mila Versteeg, The Contours of Constitutional Approval, 94 Wash. U. L. Rev. 113, 115–16, 143–45, 147–53 (2016) (using statistical analysis of surveys to assess the effect and validity of different types of constitutional designs). For a demonstration of such methods, see William Baude et al., Making Doctrinal Work More Rigorous: Lessons from Systematic Reviews, 84 U. Chi. L. Rev. 37 (2016) (using empirical methods to assess the validity of legal approaches to doctrinal interpretation). For a prescient anticipation of the current trends, see Michael Heise, The Past, Present, and Future of Empirical Legal Scholarship: Judicial Decision Making and the New Empiricism, 2002 U. Ill. L. Rev. 819 (2002).

9See, e.g., Antonin Scalia, Originalism: The Lesser Evil, 57 U. Cin. L. Rev. 849, 852 (1989) (describing analysis of history and context as the appropriate basis of constitutional analysis and rejecting moral intuitions as the alternatives); Seana Valentine Shiffrin, Inducing Moral Deliberation: On the Occasional Virtues of Fog, 123 Harv. L. Rev. 1214, 1216 (2010) (intervening in this debate through the use of deliberative moral philosophy); Kathleen M. Sullivan, Foreword: The Justices of Rules and Standards, 106 Harv. L. Rev. 22, 112 (1992) (observing the basic struggle on the bench of constitutional interpretation to be standards versus rules, based in differing moral visions of the rule of law).

10 For a seminal description of this practice, see Samuel Issacharoff & Pamela S. Karlan, Where to Draw the Line?: Judicial Review of Political Gerrymanders, 153 U. Pa. L. Rev. 541, 551–53 (2004).

11See, e.g., Samuel Issacharoff, Gerrymandering and Political Cartels, 116 Harv. L. Rev. 593, 600 (2002) (arguing that partisan gerrymandering can be a practice that inhibits “accountability to shifting voter preferences”); Michael J. Klarman, Majoritarian Judicial Review: The Entrenchment Problem, 85 Geo. L.J. 491, 516 (1997) (partisan gerrymandering is “indefensibly antimajoritarian”); Daryl Levinson & Benjamin I. Sachs, Political Entrenchment and Public Law, 125 Yale L.J. 400, 416–18 (2015) (partisan gerrymandering distorts electoral outcomes).

12 541 U.S. 267 (2004).

13 Jowei Chen & Jonathan Rodden, Cutting Through the Thicket: Redistricting Simulations and the Detection of Partisan Gerrymanders, 14 Election L.J. 331 (2015) (using simulations to assess partisan gerrymandering); Bernard Grofman & Gary King, The Future of Partisan Symmetry as a Judicial Test for Partisan Gerrymandering After LULAC v. Perry, 6 Election L.J. 1, 5 (2007) (using partisan symmetry, which compares the ratio of votes received to the number of seats per party, to assess partisan gerrymandering); Nicholas O. Stephanopoulos & Eric M. McGhee, Partisan Gerrymandering and the Efficiency Gap, 82 U. Chi. L. Rev. 831, 834 (2015) (using “wasted votes” as the basis for assessing partisan gerrymandering); see also Expert Report and Affidavit Assessing the Current Wisconsin State Legislative Districting Plan of Simon Jackman, Whitford v. Nichol, No. 315CV00421, 2015 WL 10091020 (W.D. Wis. July 7, 2015). For a description of the role quantitative methods played in the Pennsylvania Supreme Court decisions, see Issie Lapowsky, The Geeks Who Put a Stop to Pennsylvania’s Partisan Gerrymandering, Wired (Jan. 20, 2018, 11:21 AM), https://www.wired.com/story/pennsylvania-partisan-gerrymandering-experts/.

14 Common Cause v. Rucho, 279 F. Supp. 3d 587, 608 (M.D.N.C.), vacated, 138 S. Ct. 2679 (2018); Whitford v. Gill, 218 F. Supp. 3d 837, 927 (W.D. Wis. 2016); League of Women Voters of Pa. v. Commonwealth, 178 A.3d 737, 741 (Pa. 2018).

15See infra Section I.B.

16See, e.g., Benisek v. Lamone, 266 F. Supp. 3d 799, 801 (D. Md. 2017).

17 138 S. Ct. 1916 (2018). The majority opinion stated that “two threshold questions remain: what is necessary to show standing in a case of this sort, and whether those claims are justiciable. Here we do not decide the latter question because the plaintiffs in this case have no shown standing under the theory upon which they based their claims for relief.” Id. at 1929.

18 Of course, even the evidentiary role of statistics in Supreme Court jurisprudence has been a matter of dispute, most notably in the case McCleskey v. Kemp. 481 U.S. 279, 280 (1987) (disregarding statistics regarding racial bias in sentencing); see Ruth E. Friedman, Statistics and Death: The Conspicuous Role of Race Bias in the Administration of the Death Penalty, 11 La Raza L.J. 75, 82 (1999) (observing the role of statistics in constitutional cases related to employment discrimination and jury selection but not death penalty sentencing).

19 The specific language of “trumps” is taken from Ronald Dworkin, Rights as Trumps, in Theories of Rights 154–59 (Jeremy Waldron ed., 1984). The link to Dworkin’s broader theory is described in his books Law’s Empire and Taking Rights Seriously. Ronald Dworkin, Law’s Empire 221–24 (1986) [hereinafter Dworkin, Empire]; Ronald Dworkin, Taking Rights Seriously 91, 108 (1977) [hereinafter Dworkin, Rights] (describing the distinction between rights and goals, and their implications for judicial lawmaking in constitutional interpretation); see infra Part II.

20See infra Section III.B.

21 The classic doctrinal expression of this is the Court’s refusal to independently find conduct unconstitutional solely on the basis of disparate impact without indication of discriminatory intent, and to “leav[e] the choice whether to impose disparate impact standards to legislators.” Richard A. Primus, Equal Protection and Disparate Impact: Round Three, 117 Harv. L. Rev. 494, 495 (2003). The critical case for determining this issue is Washington v. Davis, 426 U.S. 229, 239 (1976) (“We have never held that the constitutional standard for adjudicating claims of invidious racial discrimination is identical to the standards applicable under Title VII, and we decline to do so today. . . . [O]ur cases have not embraced the proposition that a law or other official act, without regard to whether it reflects a racially discriminatory purpose, is unconstitutional solely because it has a racially disproportionate impact.”). Were there a law prohibiting partisan gerrymandering, the legal field would be far different. The most salient expression of this in the doctrine may be Arizona State Legislature v. Arizona Independent Redistricting Commission, 135 S. Ct. 2652, 2659 (2015), which held that a referendum implementing districting through an independent commission was constitutional.

22 In its most general form, the principle of rule of law is a very broad one of governmental impartiality in all its actions. Dworkin, Empire, supra note 19, at 174 (“[O]fficials acting in their official capacity . . . must, we say, treat all members of their community as equals, and the individual’s normal latitude for self-preference is called corruption in their case.”). In the context of constitutional protection, this takes the form of “impartial justice. Courts should be available to enforce the law and should employ fair procedures.” Richard H. Fallon, Jr., “The Rule of Law” as a Concept in Constitutional Discourse, 97 Colum. L. Rev. 1, 9 (1997).

23Chevron, U.S.A., Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837 (1984), formalized the judicial deference to legislative agencies. See Thomas W. Merrill & Kristin E. Hickman, Chevron’s Domain, 89 Geo. L.J. 833, 833–35 (2001).

24See Harper v. Va. Bd. of Elections, 383 U.S. 663, 668 (1966) (poll taxes held illegal, because, like race, overt government categorization based on wealth is disfavored).

25 Nathaniel Persily, Reply: In Defense of Foxes Guarding Henhouses: The Case for Judicial Acquiescence to Incumbent-Protecting Gerrymanders, 116 Harv. L. Rev. 649, 650 (2002); Peter H. Schuck, The Thickest Thicket: Partisan Gerrymandering and Judicial Regulation of Politics, 87 Colum. L. Rev. 1325, 1350 (1987).

26 Richard Briffault, Defining the Constitutional Question in Partisan Gerrymandering, 14 Cornell J.L. & Pub. Pol’y 397, 399 (2005); see also Davis v. Bandemer, 478 U.S. 109, 157–58 (1986) (O’Connor, J., concurring) (“In order to implement the plurality’s standard, it will thus be necessary for courts to adopt an analogous norm” to the proportionality norm applied in the racial districting jurisprudence, which would comprise “the use of the Equal Protection Clause as the vehicle for making a fundamental policy choice that is contrary to the intent of its Framers and to the traditions of this Republic.”).

27See Luis Fuentes-Rohwer, Doing Our Politics in Court: Gerrymandering, “Fair Representation” and an Exegesis into the Judicial Role, 78 Notre Dame L. Rev. 527, 538 (2003) (“When courts intervene, they simply take sides in highly politicized debates. They do very little else.”); Schuck, supra note 25. As Heather K. Gerken, Lost in the Political Thicket: The Court, Election Law, and the Doctrinal Interregnum, 153 U. Pa. L. Rev. 503, 514 (2004), has observed, courts strain to avoid doing so explicitly, preferring a “studied agnosticism” in election law cases even when their decisions reveal clear substantive impacts.

28 369 U.S. 186, 208 (1962) (“A citizen’s right to a vote free of arbitrary impairment by state action has been judicially recognized as a right secured by the Constitution . . . .”); see Guy-Uriel E. Charles, Constitutional Pluralism and Democratic Politics: Reflections on the Interpretive Approach of Baker v. Carr, 80 N.C. L. Rev. 1103, 1104 (2002) (observing that Baker v. Carr has been recognized as successfully constitutionalizing democracy).

29 The neutral districting criteria are described at Karcher v. Daggett, 462 U.S. 725, 740 (1983), and are central to the methods used by Chen & Rodden, supra note 13.

30 478 U.S. 109 (1986).

31 541 U.S. 267 (2004).

32 Common Cause v. Rucho, 279 F. Supp. 3d 587, 642–43 (M.D.N.C.) (noting the effectiveness of the gerrymander in question), vacated, 138 S. Ct. 2679 (2018); Whitford v. Gill, 218 F. Supp. 3d 837, 847–50 (W.D. Wis. 2016).

33 A less blatantly partisan form of the practice is “shacking,” or the use of districting to entrench incumbents in their home districts. Issacharoff & Karlan, supra note 10, at 551–52; cf. Gaffney v. Cummings, 412 U.S. 735, 738 (1973) (suggesting that it is the illicit role of partisanship, not the entrenchment effect, that is actionable).

34See Bruce E. Cain, The Reapportionment Puzzle 148–50 (1984); Issacharoff & Karlan, supra note 10, at 551.

35League of United Latin American Citizens v. Perry (LULAC), 548 U.S. 399 (2006), provides one of the most vivid illustrations of how parties use gerrymandering to attempt to hold on to political power even after popular opinion has changed.

36See Edward B. Foley, Due Process, Fair Play, and Excessive Partisanship: A New Principle for Judicial Review of Election Laws, 84 U. Chi. L. Rev. 655, 711–12 (2017) (briefly describing the history of the gerrymander).

37 While Bandemer did not weave the idea that discriminatory politicized districting was illegal out of thin air, none of the prior cases provided detailed guidance on how districting on the basis of party affiliation should be managed. Davis v. Bandemer, 478 U.S. 109, 119–20 (1986) (quoting Fortson v. Dorsey, 379 U.S. 433, 439 (1965)).

38Id. at 113.

39Id. at 129.

40Id. at 127.

41 The Vieth plurality raised the question of whether the intent must be to gerrymander a given district, or the entire state, and the degree of predominance required to gerrymander. Vieth v. Jubelirer, 541 U.S. 267, 284–85 (2004). Conversely, Justice Powell’s dissent in Bandemer, 478 U.S. at 179–80 (Powell, J., dissenting), suggests that intent alone may be enough to establish a constitutional violation, a sentiment shared by Justice Souter’s dissent in Vieth, 541 U.S. at 350 (Souter, J., dissenting). However, as described in Section III.A.2 infra, a pure intent test, while perhaps easier to operationalize, would affect dramatically wide-ranging judicial regulation of politics.

42See Vieth, 541 U.S. at 282–84 (reviewing lower court opinions and academic sources, and reflecting their perception of the discriminatory effects test as deeply problematic).

43Bandemer, 478 U.S. at 128, 133.

44Id. at 131–32.

45 For a theorized explanation of why this must be so, see id. at 158 (O’Connor, J., concurring) (observing that such a move would dictate significant elements of design of the political system beyond the judicial ken); James A. Gardner, Madison’s Hope: Virtue, Self-Interest, and the Design of Electoral Systems, 86 Iowa L. Rev. 87, 104 (2000) (observing that the absence of such features might be part of a chosen political design, particularly in a first-past-the-post system); Schuck, supra note 25 (same).

46 Gray v. Sanders, 372 U.S. 368, 379 (1963) (stating that “all who participate in the election are to have an equal vote—whatever their race . . . .This is required by the Equal Protection Clause of the Fourteenth Amendment.”).

47See Wesberry v. Sanders, 376 U.S. 1, 4 (1964); Reynolds v. Sims, 377 U.S. 533, 577 (1964); cf. Adam Raviv, Unsafe Harbors: One Person, One Vote and Partisan Redistricting, 7 U. Pa. J. Const. L. 1001, 1012–13 (2005) (describing how the 10% safe harbor for state districting plans has snuck into the jurisprudence).

48Bandemer, 478 U.S at 123 (“[T]he Court characterized the question posed by election districts of disparate size as an issue of fair representation.”).

49 The main arguments against this involve the claim that this class identity is fluid and thus related to the basic content of political contestation. For the latter claim, see Vieth v. Jubelirer, 541 U.S. 267, 288–89 (2004) (noting the non-determinativeness of party identity). For the argument that it is a function of political contestation, see generally Daniel H. Lowenstein & Jonathan Steinberg, The Quest for Legislative Districting in the Public Interest: Elusive or Illusory, 33 UCLA L. Rev. 1, 74–76 (1985).

50 In Vieth, a plurality, a concurring opinion, and three dissenting opinions were written. 541 U.S. at 271 (majority opinion); id. at 306 (Kennedy, J., concurring); id. at 317 (Stevens, J., dissenting); id. at 343 (Souter & Ginsburg, JJ., dissenting); id. at 355 (Breyer, J., dissenting).

51Id. at 271 (majority opinion).

52Id. at 281.

53Id. at 277–78 (quoting Baker v. Carr, 369 U.S. 186, 217 (1962)).

54Id. at 282 (quoting Davis v. Bandemer, 478 U.S. 109, 155 (1986) (O’Connor & Rehnquist, JJ., concurring)).

55Id. at 287–88.

56Id. at 287–89.

57 See infra Sections IV.A.2–A.3, describing both the innovation of the First Amendment right and the Foley and Kang theories.

58 541 U.S. at 306.

59Id.

60Id. at 309.

61Id. at 337 (Stevens, J., dissenting) (arguing for neutrality as a matter of principle); id. at 345–46 (Souter & Ginsburg, JJ., dissenting) (adapting the McDonnell Douglas burden-shifting test); id. at 360 (Breyer, J., dissenting) (arguing for an “unjustified entrenchment” standard).

62 Michael S. Kang, When Courts Won’t Make Law: Partisan Gerrymandering and a Structural Approach to the Law of Democracy, 68 Ohio St. L.J. 1097, 1111 (2007) (“Justice Kennedy’s ambivalence leaves it bizarrely unclear where the law of partisan gerrymandering stands.”).

63 548 U.S. 399 (2006).

64Id. at 456 (Stevens, J., dissenting).

65Id.; id. at 483 (Souter, J., dissenting).

66Id. at 419–20 (Kennedy, J., concurring).

67Whitford v. Gill, 218 F. Supp. 3d 837, 847–50 (W.D. Wis. 2016), describes the application of some of these techniques; see also Samuel S.-H. Wang, Three Tests for Practical Evaluation of Partisan Gerrymandering, 68 Stan. L. Rev. 1263, 1267 (2016).

68 Common Cause v. Rucho, 279 F. Supp. 3d 587, 601, 603 (M.D.N.C.), vacated, 138 S. Ct. 2679 (2018) (reliance on detailed data of past citizen voting behavior from 2008 to 2014); Whitford, 218 F. Supp. 3d at 848 (development of “partisan score[s]” as a type of “customized demographic data” for all districts that were being redrawn).

69Rucho, 279 F. Supp. 3d at 601–04 (Republican expert drew plans that maximized partisan advantage while also taking into account traditional districting criteria); Whitford, 218 F. Supp. 3d at 849 (designing districts that offered different levels of aggressiveness in terms of partisan advantage while also seeking, to the degree possible, to comply with traditional redistricting criteria).

70Cf. Cain, supra note 34, at 165 (describing that the risks of partisan gerrymandering, in particular that it should erode over time as demographics change); Jacob Eisler, Partisan Gerrymandering and the Illusion of Unfairness, 67 Cath. U. L. Rev. 229 (2018) (arguing that free political will should enable parties and voters to adapt).

71 Mitchell N. Berman, Managing Gerrymandering, 83 Tex. L. Rev. 781, 838 (2005); Fuentes-Rohwer, supra note 27, at 562; see Heather K. Gerken, Playing Cards in a Hurricane: Party Reform in an Age of Polarization, 54 Hous. L. Rev. 911, 912 (2017) (characterizing the seminal article of this approach as the finest ever written in election law); Richard L. Hasen, Looking for Standards (in All the Wrong Places): Partisan Gerrymandering Claims After Vieth, 3 Election L.J. 626, 627 (2004) (describing the influence of the method); cf. Michael S. Kang, Gerrymandering and the Constitutional Norm Against Government Partisanship, 116 Mich. L. Rev. 351 (2017) (arguing for a similar demand for government neutrality, but based on a broader reading of the doctrine).

72 Gary King, Representation Through Legislative Redistricting: A Stochastic Model, 33 Am. J. Pol. Sci. 787 (1989); see Gary King & Robert X. Browning, Democratic Representation and Partisan Bias in Congressional Elections, 81 Am. J. Pol. Sci. 1251, 1252 (1987). Notably, such a system does not demand proportionality, as King notes; it simply requires that each party receive the same number of seats when it receives the same number of votes. Thus, if both parties in a two-party system receives 67% of representatives when each receives 55% of the vote, then the system is unbiased, even though each is overrepresented relative to its proportion of votes when it does win. Id.

73See, e.g., Jeanne C. Fromer, An Exercise in Line-Drawing: Deriving and Measuring Fairness in Redistricting, 93 Geo. L.J. 1547, 1578 (2005) (observing that while partisan symmetry is accepted by political scientists, including King, as the measure of fairness, it suffers “ugly complications”); Heather K. Gerken, Understanding the Right to an Undiluted Vote, 114 Harv. L. Rev. 1663, 1680 n.63 (2001) (observing that political scientists accept the partisan symmetry standard, but reject its meaning for racial fairness); Samuel Issacharoff, Judging Politics: The Elusive Quest for Judicial Review of Political Fairness, 71 Tex. L. Rev. 1643, 1682 (1993) (observing that, to date, such quantitative tests failed in the district courts).

74 LULAC, 548 U.S. 399, 419–20 (2006).

75Id. at 466 (Stevens, J., dissenting); id. at 484 (Souter, J., dissenting) (“[F]urther attention could be devoted to the administrability of such a criterion at all levels of redistricting and its review.”).

76 For King’s own assessment of the partisan symmetry test and his attempt to introduce it as the test for gerrymandering, see Grofman & King, supra note 13, at 4. Their conclusion that a “a majority of [J]ustices now appear to endorse our view that the measurement of partisan symmetry can be used as part of a broader test in resolving partisan gerrymandering” appears optimistic, given that Kennedy, despite his initial seeming warmth toward technological progress, appears skeptical of the test. Id. Moreover, a comparison with the distribution of Vieth may get the causality backwards—liberal Justices wish to strike down partisan gerrymanders and are willing to adopt any tool that will aid them in doing so. Stephanopoulos and McGhee, meanwhile, seem to have been inspired to design the efficiency gap to specifically address Kennedy’s concerns in LULAC. Supra note 13, at 842.

77 In addition to receiving attention in the courts, the efficiency gap has been the subject of dedicated academic treatment. See Benjamin Plener Cover, Quantifying Partisan Gerrymandering: An Evaluation of the Efficiency Gap Proposal, 70 Stan. L. Rev. 1131, 1138–39 (2018).

78 Stephanopoulos & McGhee, supra note 13, at 851 (stating that the efficiency gap is “the difference between the parties’ respective wasted votes, divided by the total number of votes cast in the election. Wasted votes include both ‘lost’ votes (those cast for a losing candidate) and ‘surplus’ votes (those cast for a winning candidate but in excess of what she needed to prevail)” (emphasis omitted)).

79Id. at 845.

80 Chen & Rodden, supra note 13, at 332.

81Id. at 338–39.

82 Transcript of Oral Argument at 42–44, Gill v. Whitford, 138 S. Ct. 1916 (2018) (No. 16-1161).

83 Whitford v. Gill, 218 F. Supp. 3d 837, 847–50 (W.D. Wis. 2016).

84Id. at 883.

85See Vieth v. Jubelirer, 541 U.S. 267, 281 (2004) (intent of a legislature is not difficult to establish); Davis v. Bandemer, 478 U.S. 109, 129 n.11 (1986) (while discriminatory intent must be established for a successful claim, it may not be difficult to do so).

86Whitford, 218 F. Supp. 3d at 902–04 (relying on “S” curve analysis, “swing analysis,” and the efficiency gap).

87 279 F. Supp. 3d 587, 685–88 (M.D.N.C.), vacated, 138 S. Ct. 2679 (2018).

88Id. at 689–90.

89Id. at 640, 659.

90 League of Women Voters of Pa. v. Commonwealth, 178 A.3d 737, 820 (Pa. 2018) (emphasis added); see also id. at 818 (“[T]he compelling expert statistical evidence presented before the Commonwealth Court, in combination with and illustrated by an examination of the Plan itself and the remainder of the evidence presented below, demonstrates that the Plan cannot plausibly be directed at drawing equally populous, compact, and contiguous districts which divide political subdivisions only as necessary to ensure equal population.”).

91 266 F. Supp. 3d 799, 801 (D. Md. 2017).

92 The decision was, however, stayed in order to await the Supreme Court decision in Gill v. Whitford. Id. at 808.

93Id.

94Id. at 806–07.

95 Shapiro v. McManus, 203 F. Supp. 3d 579, 597 (D. Md. 2016).

96 Recommend Findings of Fact and Conclusions of Law at 95–96, League of Women Voters of Pa. v. Commonwealth, No. 261 M.D. 2017 (brief filed with Pa. Commw. Ct. on Dec. 29, 2017).

97Id. at 126–27.

98 Vieth v. Jubelirer, 541 U.S. 267, 288–90 (2004); Lowenstein & Steinberg, supra note 49; Schuck, supra note 25.

99 Davis v. Bandemer, 478 U.S. 109, 152 (1986) (O’Connor, J., concurring).

100See, e.g., Stephen Ansolabehere, Gill v. Whitford and the Methods for Uncovering Discriminatory Partisan Gerrymandering, Harv. L. Rev. Blog (June 26, 2018), https://blog.harvardlawreview.org/gill-v-whitford-and-the-methods-for-uncovering-discriminatory-partisan-gerrymandering/.

101 Early commentary has supported the perception that Whitford was a fundamentally evasive opinion. See David Gans, Symposium: The Fight to Vindicate Our Constitution’s Promise of Democracy is Far from Over, SCOTUSblog (June 19, 2018, 11:14 AM), http://www.scotusblog.com/2018/06/symposium-the-fight-to-vindicate-our-constitutions-promise-of-democracy-is-far-from-over/; G. Michael Parsons, Dodge, Duck, Dip, Dive, and Dodge: The 5 D’s of the Gill v. Whitford Decision, Mod. Democracy (June 19, 2018), https://moderndemocracyblog.com/2018/06/19/dodge-duck-dip-dive-and-dodge/.

102 Gill v. Whitford, 138 S. Ct. 1916, 1929 (2018).

103Id. at 1924–25.

104Id. at 1932.

105Id.

106Id. at 1929–31.

107 In some respects, therefore, the Court’s references to racial gerrymandering and one-person, one-vote, id. at 1930–31, are deceptively unhelpful. There is relatively little debate as to what comprises racial vote dilution or malapportionment at a conceptual level, merely a requirement that individuals must demonstrate the given illegal act harmed them personally. Conversely, as Whitford fails to draw out, the unanswered question in partisan gerrymandering is what is the harm?

108Id. at 1924–25.

109Id. at 1936 (Kagan, J., concurring).

110Id. at 1938.

111See id. at 1931–33 (majority opinion).

112Id. at 1934–36 (Kagan, J., concurring).

113See supra Section I.A.1.

114 Vieth v. Jubelirer, 541 U.S. 267, 281 (2004).

115Id. at 311 (Kennedy, J., concurring) (“That no such standard has emerged in this case should not be taken to prove that none will emerge in the future.”).

116Id. at 337 (Stevens, J., dissenting) (arguing for neutrality as a matter of principle); id. at 345–46 (Souter, J., dissenting) (adapting the McDonnell Douglas burden-shifting test); id. at 360 (Breyer, J., dissenting) (arguing for an “unjustified entrenchment” standard).

117See supra Sections I.A.2, I.B.1.

118Vieth, 541 U.S. at 312–13 (Kennedy, J., concurring) (“Technology is both a threat and a promise. On the one hand, if courts refuse to entertain any claims of partisan gerrymandering, the temptation to use partisan favoritism in districting in an unconstitutional manner will grow. On the other hand, these new technologies may produce new methods of analysis that make more evident the precise nature of the burdens gerrymanders impose on the representational rights of voters and parties. That would facilitate court efforts to identify and remedy the burdens, with judicial intervention limited by the derived standards.”).

119See Common Cause v. Rucho, 279 F. Supp. 3d 587, 650–64 (M.D.N.C.), vacated, 138 S. Ct. 2679 (2018).

120 Gill v. Whitford, 138 S. Ct. 1916, 1933 (2018).

121 Justice Kagan’s concurrence offers another possibility, by suggesting the statewide evidence of the efficiency gap might have been sufficient to support standing on a First Amendment theory had the Democratic Party been party to the suit. Id. at 1939 (Kagan, J., concurring).

122See Briffault, supra note 26, at 401 (listing the four constitutional arguments that might support a claim against partisan gerrymandering).

123See supra Section I.B.2.

124 Stephanopoulos and McGhee, supra note 13, have virtually no discussion of how the metric intersects with constitutional principles. The normative discussion would, if extended as a principle, seem to mandate proportional representation. Id. at 852–53 (“The gerrymandering party enjoys a political advantage not because of its greater popularity, but rather because of the configuration of district lines.”). Even if it did not, it would seem to clash with Schuck’s observation that in the absence of a mandate to proportional representation, determining the degree of benefit conferred on a winning party is a question of democratic self-determination. Schuck, supra note 25.

125 This proposition and its implications are drawn primarily from two tests. First, a judge acts as a sort of “philosophical referee” who relies on a “constitutional theory” (with regard to constitutional provisions) and a “doctrine of fairness” (with regard to case precedent) to interpretively generate a “seamless web” of principle. See Dworkin, Rights, supra note 19, at 105–30. For a rigorous and vindicatory analysis, see J. Raz, Professor Dworkin’s Theory of Rights: A Critical Review of “Taking Rights Seriously”, 1977 Bull. Austl. Soc’y Legal Phil., no. 6, 1977, at 1, 7–13. For a broad overview and a justification of its relevance to American law, see Peter Gabel, Book Review, 91 Harv. L. Rev. 302, 302 (1977) (reviewing Ronald Dworkin, Taking Rights Seriously) (“Taking Rights Seriously is a justification of contemporary American legal practice expressed in abstract and universal terms.”). The second test further expands the theory of rights presented therein. See Dworkin, Empire, supra note 19, at 210–28. For a playful take on the relationship between the two works, see generally Allan C. Hutchinson, Indiana Dworkin and Law’s Empire, 96 Yale L.J. 637 (1987).

126Dworkin, Empire, supra note 19, at 214.

127 Jeremy Waldron argues, with circumspection, that political rights are distinct from personal rights in that they demand a positive rather than negative actualization, though he observes it is possible to provide a fundamentally negative account of all rights. Jeremy Waldron, Law and Disagreement 232–34 (1999). There is a powerful insight in his argument but, for better or worse, the U.S. courts have imagined rights largely in negative terms (as protection from illicit government action), and so will this Article. Further, it will identify rights as largely personal in nature. Baker v. Carr, 369 U.S. 186, 286 (1962) (Both standing generally and the political question doctrine specifically require “infringement of an interest particular and personal to [the plaintiff].”); Dworkin, Rights, supra note 19, at 133 (U.S. Constitution is designed to prevent individuals from suffering majoritarian abuses); cf. Gerken, supra note 73, at 1671–72 (arguing that vote dilution can only be understood as an aggregate right).

128 Given the focus on partisan gerrymandering, this Article focuses on, and uses the language of, constitutional rights that private parties assert against the government. Of course, individuals may assert private rights against one another as well, Arthur Ripstein, Private Wrongs 4 (2016), and at a certain level of abstraction the same concerns are present there. This might throw into doubt the universality or completeness of the ability of empirical analysis to displace principled analysis even in the context of other domains of law, but that is beyond the ken of this Article.

129 Cass Sunstein makes this point nicely: though such a law would likely have little practical impact, the First Amendment would require courts to strike down legislation that prohibited individuals from criticizing the President while they shower. Cass R. Sunstein, Democracy and the Problem of Free Speech 168 (1993).

130Dworkin, Empire, supra note 19, at 211.

131 This character of rights, and the differentiation from policy, can be attributed to the fact that rights are “individuated” rather than “collective” in character. Dworkin, Rights, supra note 19, at 95. Dworkin’s differentiation applies to all rights; legal rights have the feature of being enforced in the judicial framework. Id. at 106.

132Id. at 154–59 (arguing a concept of principled rights must have an “independence” that cannot be justified by utilitarianism).

133 Robert Nozick provides a generalized form of this attribute of rights (focused, perhaps unsurprisingly, on property)—a right has force regardless of the consequences of enforcing it. Robert Nozick, Anarchy, State, and Utopia 269–70 (1974). This is reflective of the Nozickian equivalent of the Dworkinian trump, the “side-constraints.” Id. at 29.

134 As described in Part IV, infra, the clearest right they seem to assert is one to be free from harm by political competition; the typically adduced rights of fair representation under equal protection or a general right to association are so vague in this context as to potentially be undermined by the Vieth justiciability claim.

135 In American legal doctrine, the deontological character of rights (at least in intermediary terms) is probably most apparent in their individual character. See Baker v. Carr, 369 U.S. 186, 286–92 (1962); cf. Dworkin, Rights, supra note 19, at 166 (observing such rights must be located in granular individual equality, not a composite of collective welfare).

136 Thus the great difficulty a committed utilitarian—that is to say, wholly consequence-focused theory—has accommodating a theory of rights. See David Lyons, Utility and Rights, in Theories of Rights, supra note 19, at 129.

137Richard J. Posner, Economics of Justice 62 (1981). Interestingly, Posner attacks Benthamite utilitarianism as an approach to law that struggles to incorporate rights. Dworkin, Rights, supra note 19, at 232–38.

138Posner, supra note 137, at 70–71.

139Id. at 71.

140 A classically controversial example would be affirmative action programs at universities, which apply strict scrutiny to racial classifications but permit them as supported by a compelling state interest and narrowly tailored. See generally Grutter v. Bollinger, 539 U.S. 306 (2003).

141 Reynolds v. Sims, 377 U.S. 533, 579, 581 (1964).

142 Brandenburg v. Ohio, 395 U.S. 444 (1969).

143See Elena Kagan, The Changing Faces of First Amendment Neutrality: R.A.V. v. St. Paul, Rust v. Sullivan, and the Problem of Content-Based Underinclusion, 1992 Sup. Ct. Rev. 29, 34, 40. See generally Sunstein, supra note 129, at 169 (describing the logic of content neutrality).

144Cf. Raviv, supra note 47.

145 This Article again uses Dworkin as the foundation for this analytic distinction. See Dworkin, Rights, supra note 19, at 90 (“Principles are propositions that describe rights; policies are propositions that describe goals.”).

146Id. at 91 (“Collective goals encourage trade-offs of benefits and burdens within a community in order to produce some overall benefit for the community as a whole.”).

147See Dworkin, Empire, supra note 19, at 174 (“We apply the strictest standards of impartiality even to officials whose power is relatively slight and substantially less than that of many private citizens; we have no sense that an official’s duty of equal concern wanes as his power diminishes.”); see also Lon L. Fuller, The Morality of Law 46–49 (1964) (observing generality as the first essential quality of rule of law).

148 This feature is present in the very character of rights as trumps. See Dworkin, Rights, supra note 19, at 106. It has manifested variously in constitutional law. For example, in the application of the Fourteenth Amendment, the right—and subsequent strict scrutiny—is activated by any use of a racial classification, in order “to ‘smoke out’ illegitimate uses of race by assuring that the legislative body is pursuing a goal important enough to warrant use of a highly suspect tool.” City of Richmond v. J.A. Croson Co., 488 U.S. 469, 493 (1989).

149 Antonin Scalia, The Doctrine of Standing as an Essential Element of the Separation of Powers, 17 Suffolk U. L. Rev. 881, 884–85 (1983). Scalia’s vision has been challenged, however, by those who argue that standing itself is often used to achieve political ends and is not an intrinsically coherent doctrine. See Heather Elliott, The Functions of Standing, 61 Stan. L. Rev. 459, 501 (2008); Richard J. Pierce, Jr., Is Standing Law or Politics?, 77 N.C. L. Rev. 1741, 1775 (1999).

150 The most concrete form of this may be the requirement for an injury-in-fact, Lujan v. Defs. of Wildlife, 504 U.S. 555, 563 (1992), which prevents courts from serving as “a forum for generalized grievances.” See Gill v. Whitford, 138 S. Ct. 1916, 1929 (2018) (quoting Lance v. Coffman, 549 U.S. 437, 439 (2007)).

151 138 S. Ct. at 1923 (quoting Hollingsworth v. Perry, 570 U.S. 693, 700 (2013) (emphasis in original)).

152 There is a well-founded dispute over whether the political question doctrine is itself principled, an assessment that would support a Dworkinian view of both the doctrine and judging more generally (the “classical” view), or is instrumental in the service of maintaining the courts’ institutional position (the “prudential” view). Herbert Wechsler, Toward Neutral Principles of Constitutional Law, 73 Harv. L. Rev. 1, 15–16 (1959) (laying out the classical view and describing judging as “genuinely principled, resting with respect to every step that is involved in reaching judgment on analysis and reasons quite transcending the immediate result that is achieved”); cf. Alexander M. Bickel, The Least Dangerous Branch 183–98 (1986) (laying out the prudential view); Fritz W. Scharpf, Judicial Review and the Political Question: A Functional Analysis, 75 Yale L.J. 517, 520 (1966) (“[A]voidance [of deciding a case] . . . need not rest on constitutional principle; it may legitimately express the Court’s prudential estimate of the desirability of deciding a given constitutional issue under the particular circumstances.”). The “classical” view thereby identifies the principled character of the legal defense of rights, as distinct from the consequential focus of policies, as ontologically intrinsic to lawmaking. Conversely, the “prudential” view identifies the Court’s focus on principled, non-consequential decisions as ultimately strategic in nature, inspired by a desire to avoid upsetting the allocation of institutional power. In effect, however, both the classical and prudential understandings of the political question doctrine produce judicial treatment of rights that in legal practice emphasizes principle rather than policy outcomes. The “classical” view identifies the fulcrum of principle as inhering within the Constitution itself, whereas the “prudential” view identifies it as expressive of the practical realities of intrabranch hydraulics.

153 Rachel E. Barkow, More Supreme than Court? The Fall of the Political Question Doctrine and the Rise of Judicial Supremacy, 102 Colum. L. Rev. 237, 249 (2002) (quoting Marbury v. Madison, 5 U.S. (1 Cranch) 137, 165 (1803)). Many, including Barkow, have argued that the political question doctrine is on the retreat, id. at 336, particularly in the domestic arena. See, e.g., Harlan Grant Cohen, A Politics-Reinforcing Political Question Doctrine, 49 Ariz. St. L.J. 1, 12 (2017) (arguing that with a few exceptions, including most notably partisan gerrymandering, the political question doctrine in modern practice has become restricted to international law). Some have suggested this indicates judicial overreach, a claim with which this Article broadly agrees. See, e.g., Barkow, supra, at 336.

154 Some have argued that the historical form of the political question doctrine is fundamentally unrelated to the contemporary form, and that it ultimately grants power to the court, rather than restricts it. See Tara Leigh Grove, The Lost History of the Political Question Doctrine, 90 N.Y.U. L. Rev. 1908, 1911 (2015). The former issue is tangential to the argument here, and the second imputes a tactical insincerity to judicial thinking with which this Article need not engage.

155 369 U.S. 186, 217 (1962). While Baker articulates several different criteria that can indicate that the political question doctrine removes an issue from the Court’s ken, this has been identified as the “trigger . . . for nonjusticiability under political question rules,” Aziz Z. Huq, Removal as a Political Question, 65 Stan. L. Rev. 1, 22 (2013), and, most relevantly for the purposes of this analysis, is the basis of the conservative rejection of the justiciability of partisan gerrymandering. See Vieth v. Jubelirer, 541 U.S. 267 (2004).

156 Wechsler, supra note 152, at 16.

157See Baker, 369 U.S. at 217.

158Vieth, 541 U.S. at 290.

159See Stephanopoulos & McGhee, supra note 13, at 887 (setting thresholds for the identification of partisan gerrymanders that courts should strike down).

160Vieth, 541 U.S. at 290.

161 That is not to say their consistent enforcement would not raise its own puzzles. Id. at 288 (problem of identifying a majority party to describe partisan gerrymanders); see also Eisler, supra note 70 (describing the fluidity of political identity).

162 Gill v. Whitford, 138 S. Ct. 1916, 1932–33 (2018).

163Vieth, 541 U.S. at 288 (standard advanced by appellants is a de facto argument for proportionality unsupported by equal protection).

164Whitford, 138 S. Ct. at 1916 (describing the purpose of the standing requirement as preventing use of courts to raise general grievances about governance); Baker v. Carr, 369 U.S. 186, 217 (1962) (describing the features of a political question).

165Baker, 369 U.S. at 217.

166See Richard H. Fallon, Jr., Strict Judicial Scrutiny, 54 UCLA L. Rev. 1267, 1316 (identifying the centrality of “triggering” in rights analysis as a background condition to protection of a right).

167Id. at 1317–18.

168 For example, in the racial districting context, evidence of “a district’s shape and demographics” may help establish the illicit race-based intent that offends the Equal Protection Clause. Shaw v. Hunt, 517 U.S. 899, 905 (1996).

169 This principle is apparent in the fact that even functionally harmless government action that violates rights is illicit. To borrow an example from Cass Sunstein: “There is no legitimate reason for government to say that people may not criticize the President while they are showering. The only possible reason is to insulate the President from criticism, and this is illegitimate. The restriction is therefore unconstitutional even if it is minor.” Sunstein, supra note 129.

170 In the partisan gerrymandering literature, this view has been most coherently advanced by Lowenstein and Steinberg. Lowenstein & Steinberg, supra note 49, at 15–16. Others have explicitly rejected this assertion, arguing that the practice of the Supreme Court reveals much more ad hoc intervention to prevent seeming unfairness. Martin Shapiro, Gerrymandering, Unfairness, and the Supreme Court, 33 UCLA L. Rev. 227, 256 (1985). Insofar as courts engage in such unprincipled intervention to prevent their intuitions regarding fairness, they undermine the rule of law. See Fuller, supra note 147, at 49–51 (promulgation of laws essential for the rule of law).

171Bickel, supra note 152.

172 Gomillion v. Lightfoot, 364 U.S. 339, 347–48 (1960).

173Id. at 340.

174Id. at 346–47.

175Id.

176Id. at 347–48.

177Dworkin, Rights, supra note 19, at 90–91. That rights have such unique ontology would be rejected by certain significant strands of thinking, such as legal realism. See generally Brian Leiter, Legal Formalism and Legal Realism: What Is the Issue?, 16 Legal Theory 111 (2010) (describing in general the sociological theory of judging that would undercut such a unique status held by legal rights).

178 The development of rights related to sexuality provide one of the most famous mechanisms by which U.S. courts have come to innovate rights through conceptual likeness. Griswold v. Connecticut, 381 U.S. 479, 484. (“[S]pecific guarantees” of constitutional rights have “penumbras” and “emanations.”). See generally Paul G. Kauper, Penumbras, Peripheries, Emanations, Things Fundamental and Things Forgotten: The Griswold Case, 64 Mich. L. Rev. 235 (1965) (describing constitutional innovation of rights up to Griswold). A more recent expression of this may be in Obergefell v. Hodges, 135 S. Ct. 2584 (2015). For both a general theory of how this process unfolds and its specific application to the evolution of racial attitudes, see Michael J. Klarman, From Jim Crow to Civil Rights 5 (2006) (methodologically, asserting that “constitutional law is generally quite indeterminate [so] constitutional interpretation almost inevitably reflects the broader social and political context of the times.”). Klarman’s approach is distinct from Dworkin’s, who would argue that constitutional law, done right, reflects a process of interpretation that weaves together precedent, social values, and abstract fairness. Regardless, Klarman’s account demonstrates the significant role that social norms might play in the development of the substantive content of constitutional rights. Interestingly, Klarman’s focus on social facts might show how he is conceptually more closely aligned with the empirical legal scholars. See, e.g., Stephanopoulos and Versteeg, supra note 8 (using popular approval to assess the validity of constitutions).

179See Bickel, supra note 152, at 16 (observing the basic normative paradox in judicial nullification of democratic government action). See generally David E. Pozen, Self-Help and the Separation of Powers, 124 Yale L.J. 2, 75 (2014) (“[T]he aim of separation of powers is . . . to preclude the exercise of arbitrary government.”).

180 198 U.S. 45 (1905).

181 Cass R. Sunstein, Lochner’s Legacy, 87 Colum. L. Rev. 873, 875 (1987).

182 Indeed, some have challenged the analytic principle underlying Sunstein’s claim. Einer R. Elhauge, Does Interest Group Theory Justify More Intrusive Judicial Review?, 101 Yale L.J. 31, 56 (1991) (characterizing Sunstein’s analysis as baseline reinforcing).

183See Fallon, Jr., supra note 166, at 1270 (Lochner is condemned as “second-guessing of legislative judgments”).

184Lochner, 198 U.S. at 74 (Holmes, J., dissenting).

185Id. at 59–60 (majority opinion).

186 Indeed, while Sunstein and others, see, e.g., Amanda Shanor, The New Lochner, 2016 Wis. L. Rev. 133, 182 & n.207 (2016), characterized Buckley v. Valeo as a new type of Lochnerism, and there has long been anticipation of Buckley’s demise, Burt Neuborne, Money and American Democracy, in Law and Class in America 37, 47 (Paul D. Carrington & Trina Jones eds., 2006) (“Buckley is a rotten tree just waiting to be pushed over . . . .”), Buckley has persisted. That may be because the basic differentiation between contributions and expenditures, while crude and even pathological, is at least facially principled. Samuel Issacharoff & Pamela S. Karlan, The Hydraulics of Campaign Finance Reform, 77 Tex. L. Rev. 1705, 1711 (1999); Michael S. Kang, The End of Campaign Finance Law, 98 Va. L. Rev. 1, 6 (2012).

187 This is precisely Sunstein’s point. Sunstein, supra note 181.

188See Sujit Choudhry, The Lochner Era and Comparative Constitutionalism, 2 Int’l. J. Constitutionalism L., no. 1, 2004, at 1, 14–15.

189 These deficiencies are discussed in more detail in Section IV.A, infra.

190See Heather K. Gerken, The Costs and Causes of Minimalism in Voting Cases: Baker v. Carr and Its Progeny, 80 N.C. L. Rev. 1411 (2002); Sanford Levinson, One Person, One Vote: A Mantra in Need of Meaning, 80 N.C. L. Rev. 1269, 1274 (2002) (arguing that Baker’s minimalism has undermined its robustness).

191 Davis v. Bandemer, 478 U.S. 109, 155 (1986) (O’Connor, J., concurring); Schuck, supra note 25, at 1362–63.

192Bandemer, 478 U.S. at 152 (O’Connor, J., concurring) (“[P]olitical parties are the dominant groups” in American politics.); cf. Briffault, supra note 26, at 407 (difference of “‘pariah’ groups” as requiring protection).

193 There is one existing area of election law that has arguably developed in this direction, the 10% safe harbor for state legislatures. See Raviv, supra note 47, at 1063. This rule was cited by the majority opinion in Rucho with approval as justification for reliance on quantitative metrics in the contemporary partisan gerrymandering. See Common Cause v. Rucho, 279 F. Supp. 3d 587, 658 (M.D.N.C.), vacated, 138 S. Ct. 2679 (2018).

194See Eric A. Posner, Standards, Rules, and Social Norms, 21 Harv. J.L. & Pub. Pol’y 101, 106 (1997).

195 Sunstein, supra note 181, at 885.

196See supra Part I.

197See supra Section I.B.

198 Massachusetts v. EPA, 549 U.S. 497, 547 (2007) (“The constitutional role of the courts, however, is to decide concrete cases—not to serve as a convenient forum for policy debates.”).

199 Transcript of Oral Argument at 40, Gill v. Whitford, 138 S. Ct. 1916 (2018) (No. 16-1161).

200Id. at 42–44.

201See Friedman, supra note 18, at 82–83 (describing the inconsistent use of race in the context of the race and the death penalty).

202See Chen & Rodden, supra note 13, at 336.

203 Vieth v. Jubelirer, 541 U.S. 267, 288–89 (2004); Lowenstein & Steinberg, supra note 49, at 74–75.

204 Davis v. Bandemer, 478 U.S. 109, 122–25 (1986).

205 Shaw v. Reno, 509 U.S. 630, 649 (1993) (use of race in districting must face strict scrutiny); Gomillion v. Lightfoot, 364 U.S. 339, 346 (1960) (discriminatory districting inexplicable other than by race illicit).

206 541 U.S. at 314–15.

207See, e.g., Shapiro v. McManus, 203 F. Supp. 3d 579, 595–96 (D. Md. 2016).

208Id. at 594 (quoting Elrod v. Burns, 427 U.S. 347, 356 (1976)); see also Kang, supra note 71, at 377 & n.165 (discussing Elrod as the best example of the norm against government partisanship in the First Amendment).

209 379 U.S. 433, 439 (1965). Bandemer provides the case lineage, 478 U.S. at 120, as does Justice Steven’s Vieth dissent, 541 U.S. at 320 (Stevens, J., dissenting).

210 412 U.S. 735, 752–53 (1973).

211 Issacharoff & Karlan, supra note 10, at 551–52.

212Gaffney, 412 U.S. at 753.

213Id. at 754.

214 Eisler, supra note 70, at 266.

215 Davis v. Bandemer, 478 U.S. 109, 125 (1986) (“[T]hat the claim is submitted by a political group, rather than a racial group does not distinguish it in terms of justiciability.”).

216 LULAC, 548 U.S. 399, 469–70 (2006) (Stevens, J., dissenting) (treating party affiliation as stable and analogizing to cases on racial gerrymandering to the discriminatory effect on Democrats); Vieth v. Jubelirer, 541 U.S. 267, 320 (2003) (Stevens, J., dissenting) (Gaffney supports the idea that racial and political discrimination are equivalently illicit); id. at 337–38 (a political gerrymander is as objectionable as a racial gerrymander); id. at 344 (Souter, J., dissenting) (observing that excessive presence of both race or partisanship can render a districting plan illicit); see also Whitford v. Gill, 218 F. Supp. 3d 837, 867–73 (W.D. Wis. 2016) (carefully reviewing the case law and relying on Bandemer’s conclusion that racial and political gerrymandering are equivalently justiciable).

217 548 U.S. at 469–70 (Stevens, J., dissenting) (analogizing to cases on racial gerrymandering).

218 Gill v. Whitford, 138 S. Ct. 1916, 1930 (2018).

219 Miller v. Johnson, 515 U.S. 900, 904 (1995). This case continues to provide the bedrock reasoning in the assessment or racial districting. See Bethune-Hill v. Va. State Bd. of Elections, 137 S. Ct. 788, 797 (2017).

220Bethune-Hill, 137 S. Ct. at 797.

221Miller, 515 U.S. at 904.

222Bethune-Hill, 137 S. Ct. at 798.

223See John Hart Ely, Democracy and Distrust: A Theory of Judicial Review 75–76 (1980) (observing the tradition from United States v. Carolene Products Co., 304 U.S. 144, 152 n.4 (1938), onwards of protecting insular and discrete minority groups).

224See Dworkin, Empire, supra note 19, at 87–88 (describing the role of interpretation of principles).

225Miller, 515 U.S. at 904.

226 Lowenstein & Steinberg, supra note 49, at 74–75.

227 Karcher v. Daggett, 462 U.S. 725, 740 (1983).

228 As Bethune-Hill v. Va. State Bd. of Elections, 137 S. Ct. 788, 798 (2017), observes, mere consequential conformity with neutral districting principles will not rescue an otherwise illicitly motivated districting plan.

229Karcher, 462 U.S. at 740 (listing neutral factors in districting).

230Miller, 515 U.S. at 904.

231See Davis v. Bandemer, 478 U.S. 109, 145 (1986) (O’Connor, J., concurring).

232 Schuck, supra note 25; see also Lowenstein & Steinberg, supra note 49, at 74.

233See Shapiro v. McManus, 203 F. Supp. 3d 579, 588–89 (D. Md. 2016); see also Whitford v. Gill, 218 F. Supp. 3d 837, 875 (W.D. Wis. 2016). See generally Daniel P. Tokaji, Gerrymandering and Association, 59 Wm. & Mary L. Rev. 2159, 2160–62 (2018); Daniel P. Tokaji, Voting Is Association, 43 Fla. St. U. L. Rev. 763 (2016). The latter piece has some consonance with the ultimate proposal of this Article’s conclusion insofar as it identifies entrenchment as the basis for an associational right against gerrymandering. Tokaji, Gerrymandering and Association, supra, at 2190–91, 2199. However, as this Article argues, the First Amendment jurisprudence does not clearly enough articulate a definitive anti-entrenchment rationale, particularly as the Celebrezze/Williams line of cases is primarily concerned with depriving minority parties of any associational rights, rather than the give-and-take in a two-party system. Cf. Bandemer, 478 U.S. at 152 (O’Connor, J., concurring) (parties are dominant actors).

234 Gill v. Whitford, 138 S. Ct. 1916, 1938 (2018) (Kagan, J., concurring).

235Whitford, 218 F. Supp. 3d at 875.

236 393 U.S. 23 (1968).

237 460 U.S. 780 (1983); cf. Crawford v. Marion Cty. Election Bd., 553 U.S. 181, 190 (2008) (stating that Anderson’s balancing test is now the guiding precedent); Foley, supra note 36, at 674–76 (describing the “Anderson–Burdick balancing test,” which uses a balancing test to assess whether a state’s regulation of the right to vote violates the Equal Protection Clause). This balancing test, however, merely inquires as to when the right to vote trumps the validity of government regulation of the electoral process. Thus, unlike partisan gerrymandering, which requires a threshold assessment of the severity of the practice itself, this balancing test simply weighs a “trump” against competing interests. The right to vote, as it were, is absolute.

238See Celebrezze, 460 U.S. at 788 (“[E]ach provision of these schemes . . . inevitably affects . . . the individual’s right to vote and his right to associate with others for political ends”); Williams, 393 U.S. at 38–39 (arguing that barriers hinder the ability to organize as a third party at all); cf. Shapiro, 203 F. Supp. 3d at 605 (Bredar, J., dissenting) (arguing that partisan gerrymandering does not infringe the type of First Amendment associational rights in a manner germane to the type of conduct that such rights are meant to protect, because partisan gerrymandering does not impair voters’ ability to “affiliate with the party of their choice, to vote, to run for office if they wish, and to participate in vibrant political debate wherever they find themselves”).

239 Rutan v. Republican Party of Ill., 497 U.S. 62 (1990).

240Id. at 75–77.

241Celebrezze, 460 U.S. at 787.

242 Davis v. Bandemer, 478 U.S. 109, 131–32 (1986) (stating that a given districting reduces the likelihood a given group will be able to elect its representative does not comprise a rights violation); Shapiro, 203 F. Supp. 3d at 591 (“[C]itizens have no constitutional right to reside in a district in which a majority of the population shares their political views and is likely to elect their preferred candidate.”).

243 Gill v. Whitford, 138 S. Ct. 1916, 1939 (2018) (Kagan, J., concurring) (characterizing a First Amendment claim against gerrymandering as one that would “debilitate[] their party or weaken[] its ability to carry out its core functions and purposes”).

244Cf. Tokaji, Gerrymandering and Association, supra note 233, at 2199. Tokaji’s suggestion that the burden must be “substantial” and “enduring” does nothing to resolve the basic query raised by, inter alia, the Vieth majority: How enduring? How substantial?

245Celebrezze, 460 U.S. at 788–89.

246 Foley, supra note 36, at 686. Foley’s analysis focuses on the balancing test from Anderson, and curiously does not engage extensively with the analogy to racial gerrymandering that has motivated liberal Justices.

247Id. at 688.

248Id. at 711.

249Id. at 720.

250Dworkin, Empire, supra note 19, at 227–28.

251 Foley, supra note 36, at 726.

252Dworkin, Empire, supra note 19, at 227.

253Id. at 228.

254 Kang, supra note 71, at 353.

255Id. at 354. In this respect, Kang’s analysis reaches the same legal point as the majority opinion in Rucho.

256Id. at 403.

257Id. at 406.

258Id. at 352.

259Id. at 378–79.

260Id. at 369.

261See Cain, supra note 34, at 159 (describing every seat under such a districting as “inefficient”); Issacharoff, supra note 11, at 598–99 (identifying the self-serving nature of such a districting).

262 It thus would run the risk of drifting toward the type of Lochnerism described in Section III.B.3, supra.

263 Vieth v. Jubelirer, 541 U.S. 267, 285–86 (2004).

264 Klarman, supra note 11.

265 Lowenstein & Steinberg, supra note 49, at 75.

266 Dworkin provides the philosophical justification for this by explaining how interpretation of a fair representative right in a mutually respecting community would emerge from a view of law as integrity. Dworkin, Empire, supra note 19, at 213, 225–26.

267 That the right is best understood as the right to fair representation suggests that, as the progeny of Baker v. Carr, it might best be understood as defended by the Equal Protection Clause. However, a due process argument similar to that advanced by Foley, supra note 36, could also provide a precedential lineage.

268 The closest precedent to this test is Justice Stephen’s suggestion that the Shaw predomination test should apply. Vieth, 541 U.S. at 339.

269 Indeed, the Court showed a form of such sensitivity in Gaffney v. Cummings, 412 U.S. 735 (1973), though whether the case is correctly decided is a separate question. Yet there are reasons to recognize that party leadership, rather than rank-and-file voters, ought to play a central role in designing terms of political contestation. See Sujit Choudhry, Resisting Democratic Backsliding: An Essay on Weimar, Self-Enforcing Constitutions, and the Frankfurt School, 7 Global Constitutionalism 54, 67 (2018).

270 Bethune-Hill v. Va. State Bd. of Elections, 137 S. Ct. 788, 798 (2017); Miller v. Johnson, 515 U.S. 900, 916 (1995).

271See supra Section I.A.1.

272 Chen & Rodden, supra note 13, at 334, 340 (explaining the role of neutrality in the simulation approach and the results of running the simulations).

273See Cover, supra note 77, at 1213 (a plan may have a low efficiency gap by sacrificing other desirable features).

274Cf. Ariz. State Legislature v. Ariz. Indep. Redistricting Comm’n, 135 S. Ct. 2652, 2658–59 (2015) (confirming that a democratically selected referendum may perform just this purpose).

275See supra Section I.A.1.

276See supra Section III.B.

277See Elliott, supra note 149, at 466 n.35 (collecting criticism of standing as readily manipulated to surreptitiously serve judges’ substantive preferences).

278 Gill v. Whitford, 138 S. Ct. 1916, 1933 (2018) (suggesting the standing query requires inquiry into the effect on particular citizens); id. at 1936 (Kagan, J., concurring) (suggesting that the appropriate solution is further creation of metrics).

279See Scalia, supra note 149, at 894 (standing serves the end of separation of powers).

280 Gerken, supra note 73, at 1671–72.

281See Eisler, supra note 70, at 233.

282 Lowenstein & Steinberg, supra note 49, at 76; Schuck, supra note 25.

283See Bickel, supra note 152, at 16.

284Dworkin, Empire, supra note 19, at 213.

285See supra Section III.B.3.

286See supra Section I.B.2.

287See supra Section III.C.2.

288 Einer Elhauge provides a structural parallel that explains the rule of lenity: interpretation should favor those who are least likely to be able to lobby to have their interests represented in drafting. Einer Elhauge, Statutory Default Rules: How to Interpret Unclear Legislation 169 (2008). A parallel principle suggests judicial caution toward metrics where more powerful members of society might have better access to them.