Emory Law Journal

All Things in Proportion? American Rights Review and the Problem of Balancing
Jud Mathews,
Alec Stone Sweet Jud Mathews is Visiting Assistant Professor, University of Illinois College of Law.Alec Stone Sweet is Leitner Professor of Law, Politics, and International Studies, Yale Law School. The authors would like to thank participants in the Yale Law School faculty workshop for helpful comments on a draft of this paper.

Abstract

This Article describes and evaluates the evolution of rights doctrine in the United States, focusing on the problem of balancing. In the current Supreme Court, deep conflict over whether, when, and how courts balance rights is omnipresent. Elsewhere, we find that the world’s most powerful constitutional courts have embraced a stable analytical procedure for balancing, known as proportionality. Today, proportionality analysis (PA) constitutes the defining doctrinal core of a transnational, rights-based constitutionalism. This Article critically examines alleged American exceptionalism, from the standpoint of comparative constitutional law and practice. Part I provides an overview of how constitutional judges in other systems use PA, assesses the costs and benefits of adopting it, and contrasts proportionality with American strict scrutiny. Part II recovers the foundations of proportionality in American rights review, focusing on two critical junctures: (1) the emergence of a version of PA in dormant Commerce Clause doctrine in the late nineteenth century, the core of which persists today; and (2) the consolidation of the strict scrutiny framework in the mid-twentieth century. Part III demonstrates that the “tiered review” regime chronically produces pathologies that have weakened rights protection in the United States and undermined the coherence of the Supreme Court’s rights jurisprudence. PA, while not a cure-all for the challenges faced by rights-protecting courts, avoids these pathologies by providing a relatively systematic, transparent, and trans-substantive doctrinal structure for balancing. We also show that all three levels of review—rational basis, intermediate review, and strict scrutiny—have, at various points in their evolution, contained core elements of proportionality. In Part IV, we argue that the Supreme Court can and should cultivate a version of PA rooted in American constitutional traditions and values.

Introduction

In the United States, disagreement runs deep over the proper role for balancing in constitutional rights review: over whether, when, and how courts should turn to balancing. Nor is the problem a new one. Past struggles over the merits of balancing are etched into our constitutional caselaw. 1 See T. Alexander Aleinikoff, Constitutional Law in the Age of Balancing, 96 Yale L.J. 943, 964 (1987); Kathleen M. Sullivan, Categorization, Balancing, and Government Interests, in Public Values in Constitutional Law 241, 241–44 (Stephen E. Gottlieb ed., 1993). American rights doctrines are a tangle of different tests, some requiring the court to “balance” or “weigh” factors, and others taking the form of categorical constitutional rules. 2 See Aleinikoff, supra note 1, at 964–71 (canvassing constitutional balancing tests). This mix reflects, in part, the changing fortunes of balancing, which have waxed and waned over the years. 3 See infra Part II; see also Richard H. Fallon, Jr., Strict Judicial Scrutiny, 54 UCLA L. Rev. 1267, 1288–89 (2007) (noting that the Supreme Court adopted a balancing approach to First Amendment claims that was quite speech protective in the 1940s, which was replaced by a more deferential balancing approach in 1949). Of course, the diversity of legal tests in American constitutional law also reflects, in part, the diversity in the structure of rights found in the Constitution. Some rights lend themselves easily to formulations as constitutional rules (e.g., “no quartering of soldiers”), and others are more easily read to invite balancing tests (e.g., due process). In the current Supreme Court, conflicts over its legitimacy regularly flare into view; today, a new struggle over balancing is coming to dominate the politics of rights in America. 4In District of Columbia v. Heller, 554 U.S. 570 (2008), Justices Scalia and Breyer engaged in an unusually acerbic argument about the appropriate methodology to employ in review of a District of Columbia measure that banned handguns in the home and required that other firearms be rendered inoperable. Writing for a five-member majority, Justice Scalia focused on the early history of the United States in order to show that the Second Amendment possessed a categorical, rule-like structure. Id. Once the Court determined that the Second Amendment guaranteed an individual right to bear arms for self-defense, the law was struck down. Id. at 635. On the issue of balancing, Justice Scalia stated that the Second Amendment, like the First, “is the very product of an interest-balancing by the people—which Justice Breyer would now conduct for them anew,” yet Justice Scalia also claimed: “We know of no other enumerated constitutional right whose core protection has been subjected to a freestanding ‘interest-balancing’ approach.” Id. at 634–35. For his part, Justice Breyer, joined by Justices Stevens, Souter, and Ginsburg in dissent, asserted that a “sort of ‘proportionality’” balancing approach is regularly used in “various constitutional contexts, including election-law cases, speech cases, and due process cases.” Id. at 690 (Breyer, J., dissenting). Advocating an “interest-balancing inquiry” that “asks whether the statute burdens a protected interest in a way or to an extent that is out of proportion to the statute’s salutary effects upon other important governmental interests,” Justice Breyer then adopted a relatively standard version of three-stage proportionality analysis to show why the District of Columbia’s ban should be upheld. Id. at 689–90, 693–719. For extended discussions of Heller’s debate over balancing, see Joseph Blocher, Categoricalism and Balancing in First and Second Amendment Analysis, 84 N.Y.U. L. Rev. 375 (2009); Moshe Cohen-Eliya & Iddo Porat, The Hidden Foreign Law Debate in Heller: The Proportionality Approach in American Constitutional Law, 46 San Diego L. Rev. 367 (2009).

The picture looks quite different elsewhere. In contrast to the United States, constitutional courts in legal systems around the world have converged on a method for adjudicating rights claims—proportionality analysis (PA)—an analytical procedure with balancing at its core. In the past half-century, PA has become a centerpiece of jurisprudence across the European continent, as well as in common law systems as diverse as Canada, South Africa, Israel, and the United Kingdom. 5 See Alec Stone Sweet & Jud Mathews, Proportionality Balancing and Global Constitutionalism, 47 Colum. J. Transnat’l L. 72 (2008). PA, which began as an unwritten set of general principles of law, has evolved into a standardized doctrinal framework that courts can apply across substantive areas of law. Today, judges have raised proportionality to the rank of a fundamental, constitutional principle, which they deploy to manage rights claims, including conflicts between constitutional rights. PA has also been adopted by the most powerful international courts, including the European Court of Justice, the European Court of Human Rights, and the Appellate Body of the World Trade Organization. 6 See id. at 138–59. In a previous article, we elaborated a theory of why judges are attracted to PA; we traced the framework’s global diffusion outward from Germany after World War II; and we showed that adopting PA serves to enhance the importance of rights, and of judicial authority, within policy processes otherwise dominated by nonjudicial officials. 7 See generally Stone Sweet & Mathews, supra note 5.

This Article takes a fresh look at alleged American exceptionalism in the field of rights review, from the perspective of proportionality. It complements our previous study of the global diffusion of PA by focusing comparative attention on American experience and practice. Our analysis shines new light on American doctrinal developments and challenges the image of the United States as an outlier. Strikingly, we find that U.S. courts did develop frameworks for rights review that resembled PA, starting in the nineteenth century. The Supreme Court first derived the functional equivalent of PA as a test for state restrictions on trade under the dormant Commerce Clause. And in the mid-twentieth century, strict scrutiny review emerged as a rights-favoring balancing framework with pronounced similarities to PA. It turns out that American judges chose proportionality in the past and introduced it into our doctrinal DNA.

But this heritage is sometimes obscured in current rights jurisprudence and scholarly discourse. The nearest analogue to PA today—the closest thing we have to a common rubric for reviewing claims across different substantive areas—is the set of standards that makes up tiered scrutiny. 8Canonically, the tiers are strict scrutiny, intermediate scrutiny, and rational basis review. The picture is complicated somewhat by less well-established gradations (e.g., “rational basis with bite”). Also, the Supreme Court does not always use these three labels. Kathleen Sullivan points out that tests for a number of constitutional claims, from privileges and immunities issues to public-forum speech regulations, amount to intermediate scrutiny, even when the Court does not apply such a label. Kathleen M. Sullivan, Post-Liberal Judging: The Roles of Categorization and Balancing, 63 U. Colo. L. Rev. 293, 297 (1992). Not only do PA and tiered review share certain core elements but, we argue, in a head-to-head comparison, PA has clear advantages. Far from balancing rights away, PA can protect rights more consistently and coherently than can tiered review. The American approach limits the flexibility of judges in the face of complexity, falsely portrays adjudication as a mechanical exercise in applying law that is akin to a “constitutional code,” 9 See Stone Sweet & Mathews, supra note 5, at 88. and creates unnecessary inconsistency and arbitrariness.

Our purpose is not to heap abuse on tiered review for its own sake; its shortcomings have already been amply canvassed by scholars 10 See, e.g., Suzanne B. Goldberg, Equality Without Tiers, 77 S. Cal. L. Rev. 481, 484–91 (2004); Jeffrey M. Shaman, Cracks in the Structure: The Coming Breakdown of the Levels of Scrutiny, 45 Ohio St. L.J. 161 (1984); Eugene Volokh, Freedom of Speech, Permissible Tailoring and Transcending Strict Scrutiny, 144 U. Pa. L. Rev. 2417 (1996). and judges. 11 See, e.g., City of Cleburne v. Cleburne Living Ctr., Inc., 473 U.S. 432, 478 (1985) (Marshall, J., concurring in the judgment in part and dissenting in part); Craig v. Boren, 429 U.S. 190, 211–12 (1976) (Stevens, J., concurring). Rather, we seek to show, constructively, how incorporating PA can remedy problems of our constitutional practice. Proportionality’s extraordinary success—its value to the judges who have adapted it for use in very different legal systems—lies in the fact that it provides a doctrinal anchor for principled balancing as a mode of rights protection. We argue that American rights review lacks such an anchor and suffers as a result. Nor does moving toward PA mean adopting an exotic foreign transplant. Instead, this move would reclaim and build on the foundations that already exist in our own history and doctrine. We seek to show how courts can construct a modern, distinctly American PA that squares with the practices of common law constitutionalism, including a respect for history and precedent.

Together, our claims add up to what we regard as the strongest case for recognizing and incorporating features of PA into American constitutional law. To be clear, we are not proposing that PA should necessarily govern every constitutional rights claim. Nor do we regard proportionality as a miracle cure-all that will make hard constitutional questions easier to answer. PA does not spare judges the hard work of theorizing the nature and scope of a right in play (a process that triggers PA) nor does it dictate correct answers. PA does require that judges—openly, routinely, and without embarrassment—engage in balancing, and many will find this posture uncomfortable. Nonetheless, to the extent that balancing is inevitable in rights adjudication, the proportionality framework offers the best available procedure for doing so.

The Article is organized as follows. In Part I, we introduce PA as an argumentation framework, summarize the pros and cons of adopting proportionality as a standardized mode of rights adjudication, and contrast PA with American approaches in broad, relatively abstract terms. Building on the theoretical treatment of proportionality in our last article, 12 See Stone Sweet & Mathews, supra note 5. we argue that PA offers a number of systemic advantages relative both to tiered review and to more rule-based, categorical approaches. In Part II, we uncover the foundations of proportionality in American rights review, focusing on two critical junctures: (1) the emergence of a form of proto-proportionality in dormant Commerce Clause doctrine in the late nineteenth century, and (2) the consolidation of the strict scrutiny framework in the mid-twentieth century. We demonstrate that both of these developments, which mirrored simultaneous stages in the evolution of PA in Germany, laid a foundation for a structured but suitably flexible approach to rights review. In Part III, we subject the major features of American rights doctrine, as it currently exists, to critical analysis in the light of PA principles. Part IV considers how to give PA principles greater expression in American constitutional law and responds to objections.

I. Proportionality and Strict Scrutiny: An Overview

In constitutional systems across the globe, proportionality balancing today constitutes the dominant, “best practice” judicial standard for resolving disputes that involve either a conflict between (a) two rights claims or (b) a rights provision and a legitimate government interest. In the latter, paradigmatic situation, the analysis proceeds step by step, as follows. In a preliminary phase, the judge considers whether a prima facie case has been made to the effect that a government act burdens the exercise of a right. 13 See id. at 75. By convention, the judge will use this occasion to discuss the jurisprudential theories that underpin the pleaded right, as well as prior rulings and other legal materials that will bear upon the court’s determination of the right’s scope and application in the case at hand. No important claim will ever be rejected at this stage. PA then proceeds through a sequence of three tests. A government measure that fails any one of these tests violates the proportionality principle and is therefore unconstitutional.

The first stage of PA mandates inquiry into the “suitability” of the measure under review. The government must demonstrate that the relationship between the means chosen and the ends pursued is rational and appropriate, given a stated policy purpose. 14 See id. This mode of scrutiny is broadly akin to what Americans call “rational basis” review, although under PA, the appraisal of government motives and choice of means is more searching. In most systems, few laws are struck down at this stage. 15Dieter Grimm, Proportionality in Canadian and German Constitutional Jurisprudence, 57 U. Toronto L.J. 383, 389 (2007).

The second step—“necessity”—embodies what Americans know as a “narrow tailoring” requirement. At the core of necessity analysis is a least-restrictive-means (LRM) test, through which the judge ensures that the measure at issue does not curtail the right more than is necessary for the government to achieve its goals. 16 See Ian Ayres & Sydney Foster, Don’t Tell, Don’t Ask: Narrow Tailoring After Grutter and Gratz, 85 Tex. L. Rev. 517, 520–24 (2007) (identifying LRM testing as a “central meaning” of the narrow tailoring requirement, criticizing an alternative approach to narrow tailoring in racial preference cases, and calling for a return to LRM testing). For many courts, including the Canadian Supreme Court 17 See Grimm, supra note 15, at 384. and the European Court of Justice, 18 Jürgen Schwarze, European Administrative Law 857 (2006). the necessity stage is the heart of the analysis, and the majority of laws struck down by these courts failed the LRM test. In practice, judges do not invalidate a measure simply because they can find one less restrictive alternative. 19 See, e.g., RJR-MacDonald Inc. v. Canada (Att’y Gen.), [1995] 3 S.C.R. 199, 342 (Can.) (“The tailoring process seldom admits of perfection and the courts must accord some leeway to the legislator. If the law falls within a range of reasonable alternatives, the courts will not find it overbroad merely because they can conceive of an alternative which might better tailor objective to infringement.”); Canada (Att’y Gen.) v. JTI-Macdonald Corp., [2007] 2 S.C.R. 610, 631 (Can.) (making a similar point). Instead, most courts, explicitly or implicitly, insist that policymakers have a duty to consider reasonably available alternatives and to refrain from selecting the most restrictive among them. Most courts will rarely if ever strike down a law without comparing it to a list of reasonably available alternatives.

The third step—balancing stricto sensu—is also known as “proportionality in the narrow sense.” 20Robert Alexy, Constitutional Rights, Balancing, and Rationality, 16 Ratio Juris 131, 135 (2003). In the balancing phase, the judge weighs, in light of the facts, the benefits of the act (already found to have been narrowly tailored) against the costs incurred by infringement of the right, in order to decide which side shall prevail. Most judges who use PA would not characterize balancing in such blunt, utilitarian terms. 21Following German practice, PA is not a jurisprudence of interests but of constitutional “principles” and “values.” Robert Alexy, A Theory of Constitutional Rights 86–93 (Julian Rivers trans., 2002); Donald P. Kommers, The Constitutional Jurisprudence of the Federal Republic of Germany 45–48 (1997). Instead, they would emphasize that the balancing stage allows them to “complete” the analysis, in order to ensure that no factor of significance to either side has been overlooked. In contrast to the practices of the high courts of Canada and the European Union, for example, the German Federal Constitutional Court and the Israeli Supreme Court tend to move more systematically to the final balancing stage, especially when they confront controversial “hard cases.” 22 See Grimm, supra note 15, at 393–95; Stone Sweet & Mathews, supra note 5, at 135–37. A court that strikes down a law as unconstitutional in the third stage of PA will typically use the first two steps to pay its respects to the importance of the purposes pursued by the government and to the quality of the government’s own deliberations on the proportionality of the law. 23 See, e.g., HCJ 2056/04 Beit Sourik Village Council v. Gov’t of Isr. 58(5) IsrSC 807 [2004], translated in Isr. L. Rep. 264 (2004) (striking down the proposed route of Israel’s security fence as disproportionate in the strict sense after finding that it satisfied the first two standards of proportionality).

Before turning to a more nuanced discussion of balancing within PA and strict scrutiny, two crucial comparative points deserve emphasis. First, in the United States, opponents of judicial balancing have largely built their case on the view that balancing is necessarily ad hoc, open-ended, and unprincipled from the standpoint of rights protection. 24 See Aleinikoff, supra note 1, at 984–95. The conclusion, often presented as an easy assumption, is that judicial balancing is an inherently undisciplined exercise of unbridled law making that deprives rights of their a priori normative status, for example, as “trumps” or “shields” against government action. 25 See, e.g., Ronald Dworkin, Rights as Trumps, in Theories of Rights 153 (Jeremy Waldron ed., 1984) (discussing situations in which individual rights may trump public goals). This characterization does not easily fit PA. PA is a highly formalized argumentation framework, the basic function of which is to organize a systematic assessment of justifications for government measures that would burden the exercise of a right. A government must explain such acts, which PA subjects to the highest standard of judicial scrutiny. In doing so, PA enhances the transparency of rights review, not least by making explicit the justifications for limiting rights the court has either accepted or rejected and at precisely what stage of the analysis. Second, as a formal doctrinal structure, PA is on its face no less intrusive, “strict,” or rights-prioritizing, than is American strict scrutiny. In the United States, however, strict scrutiny is only applied to a small number of rights, whereas PA is applicable to virtually all rights claims.

A. Balancing

In classic notions of separation of powers, the responsibility to balance the varied, multidimensional costs and benefits to society of any set of public policy options belongs to legislative authority. PA does not challenge that view. Instead, it subjects the balancing that inheres in policy making to judicial supervision. PA is thus directly implicated in the exercise of legislative power. Indeed, PA forges important causal connections between judicial balancing and the exercise of law-making authority, and between law making and the evolution of constitutional rights jurisprudence. Moreover, PA comprises a multi-stage balancing framework; that is, judicial balancing is not restricted to the final balancing-in-the-strict-sense stage, but takes place within each of the tests. And the tests are sequenced in order of increasing stringency, so that courts insert themselves into the legislative process no more than is necessary to defend rights. If a measure does not survive the inquiry into the means–ends fit, the court need not escalate to a more probing form of balancing analysis. Consider again the paradigmatic case, wherein a pleaded right (x) comes into conflict with the means employed by a given measure to achieve a stated government purpose (y). Under PA, the court must assess the harm (to value x) against the contribution (to value y) of a policy decision, but it does so in three different ways. In the first two stages, the court examines the means–ends nexus, assessing how lawmakers themselves weighed costs and benefits. The “suitability” test will normally capture laws where the mismatch between means and ends is most acute—irrational or grossly overbroad laws that exact a cost in terms of rights while accomplishing little to nothing. The second test, necessity analysis, enables judges to probe intentions of legislators in much more detail, not least to smoke out bad motives. 26 See Jed Rubenfeld, Affirmative Action, 107 Yale L.J. 427, 436–37 (1997). If the court finds that lawmakers had less-restrictive alternatives, it overturns the law as a disproportionate exercise of legislative authority. Thus, the court moves to balancing in the strict sense only after the measure has survived scrutiny into how the legislator has already balanced the values in tension. If the government is to prevail in the final phase of PA, the court must agree that the measure under review generates enough added benefit (to value y) to justify the harm (to value x).

Judges who would construct rights in absolute terms, or who would prefer to build fixed hierarchies of constitutional values, or who would seek to ban balancing from their repertoires for other reasons, will have no use for PA. In contrast, most judges who use PA presume that balancing cannot be avoided in rights adjudication. Modern systems of rights protection delegate massive law-making authority to constitutional judges, typically under institutional arrangements that we have characterized as “structural judicial supremacy.” 27Stone Sweet & Mathews, supra note 5, at 86 (“Modern constitutionalism is characterized by structural judicial supremacy, where the principals have, in effect, transferred a bundle of significant ‘political property rights’ to judges, for an indefinite duration.”); see also id. at 85–96 (discussing the concept of structural judicial supremacy with reference to balancing). In effect, judicial supremacy is a tax that a polity pays for maintaining a modern system of constitutional justice. In such systems, judges do not use PA to mask balancing, or to camouflage law making, but to rationalize both within the protective confines of a stable procedure that they see as inherently judicial.

B. Proportionality and Judicial Power

In our previous article, we developed a theory to explain why constitutional and supreme courts would find the proportionality framework attractive. Our explanation blended strategic (political) and legal (norm-governed) factors and logics, theorized in particular ways. 28 Id. at 80–97. In a nutshell, we argued that judges are attracted to PA because: (a) it neatly “fits” the structure of qualified rights; (b) it prioritizes rights protection while giving judges flexibility to tailor outcomes to highly charged political contexts; and (c) it provides a stable, defensible framework for argumentation and justification that judges can deploy to reduce uncertainty and to enhance consistency and predictability. 29 Id. at 87–90. We then charted the global diffusion of PA to the most powerful constitutional and supreme courts in the world. As we demonstrated, these courts adopted PA in order to deal with the most politically salient and controversial cases that they could be expected to face. In every system examined, we also found that the move to PA served to enhance the status of rights and the judiciary’s role in both law-making processes and the overall process of constitutional development. The institutionalization of the framework places nonjudicial actors in an ever-deepening shadow of rights adjudication. In consequence, policymakers become increasingly careful to build records of the proportionality of their own decision making, knowing full well that their decisions will be subject to review by the courts under PA. Further, all relevant actors in the system, including future litigants and their lawyers, government officials, and legal scholars, gradually begin to think of their roles in terms of proportionality, further consolidating PA’s centrality.

Proportionality balancing necessarily exposes judges as lawmakers and raises the classic legitimacy dilemmas associated with such exposure. 30 Id. at 80–87. At the same time, PA provides various means of coping with these dilemmas, which partly accounts for the framework’s popularity. Here we highlight four strategic advantages that PA offers the balancing, rights-protecting judge.

First, adopting the proportionality framework comprises an effective procedural response to an intractable substantive problem. Constitutional rights are famously imprecise, open-ended, and incomplete in other important ways, and these qualities are amplified when provisions are qualified by limitation clauses. While a shift to balancing is, arguably, an appropriate response to incompleteness, it can only reinforce the perception that rights adjudication is outcome indeterminate. Nonetheless, judges can bring a semblance of determinacy to balancing by subjecting it to a fixed procedure, the most formalized and well tested of which is PA. As important, PA offers judges the possibility of building trans-substantive coherence, since it can be applied across the board, to virtually all disputes involving rights.

Second, PA bestows a sheen of politico-ideological neutrality on a court, across time and circumstances. In any dispute, one party—or constitutional value—will ultimately prevail against the other, but only with regard to a specific context or set of facts. Under PA, it is not the law that varies from case to case, but the facts or decision-making context. In a future case involving a conflict between the same two values, the other side may well prevail if circumstances lead the judge to weigh the values in tension differently. PA maximizes the court’s flexibility vis-à-vis all potential litigants in future cases and gives the court a structured setting in which to pay equal respect and honor to each constitutional value on its own, and in competition with one another, before determining a winner.

Third, PA is specifically designed to reduce the harm to the losing party as much as possible. The point, formalized by Robert Alexy, follows from the view that (a) rights, and the constitutional values that must be balanced against rights, are “legal principles,” as distinguished from “legal rules”; and (b) a conflict between two principles requires that both be “optimized” through balancing. However rights are conceptualized, necessity analysis requires narrow tailoring, which permits the judge to access the legitimizing logics of Pareto optimality. 31 Alexy, supra note 21, at 40–110. Under PA, a government measure that restricts a right more than is necessary to achieve a legitimate state purpose can never be justified, since the right claimant can be made better off with no additional cost to the value being pleaded by the government.

As a practical matter, we can expect PA to constrain judicial law making in ways that can be tracked empirically, insofar as judges actually assess the proportionality of government measures with competence and in good faith. As mentioned, a court will rarely strike down a law in the necessity stage unless it can demonstrate that the law infringes more on the right than a range of other explicitly identified, “reasonably available alternatives.” In our view, judges gain an important strategic resource when they treat this (defensive) practice as a de facto constitutional duty. Similarly, in the balancing stricto sensu phase, many judges assume an obligation—we would call it a constitutional duty—to be as precise as possible about how they weigh the contribution of a government act to value y (the government’s legislative purpose) compared to the harm of that measure to value x (a pleaded right).

Fourth, adopting PA is a way for courts that seek to secure their bona fides in the field of rights protection to build their credibility. The recognition of PA as a “best-practice standard” of global constitutional law is the outcome of an ongoing social process of diffusion, and of legitimation, that has occurred on a global scale. This process has all the hallmarks of what sociological institutionalists call “institutional isomorphism,” in that PA’s diffusion has become subject to logics of mimesis and increasing returns (bandwagon effects). 32The classic treatment is Paul J. DiMaggio & Walter W. Powell, The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Fields, in The New Institutionalism in Organizational Analysis 63, 67–69 (Walter W. Powell & Paul J. DiMaggio eds., 1991). Faced with similar problems, judges copy what they take to be the emerging, high-prestige standard, thereby ensuring the result. For new constitutional courts—or for old courts charged with protecting a new charter of rights—embracing PA is a low-cost move, compared to the costs of developing an untested alternative on their own. Finally, PA is a simple but comprehensive doctrinal structure, which facilitates diffusion. Lawyers, law students, and judges can learn the basics quickly and deploy the framework with ease, which benefits constitutional judges in obvious ways.

Steadfast opponents of PA may well concede these points while maintaining their objections. Although critiques vary in subtlety and sophistication, most share a common hostility toward judicial balancing per se and a suspicion of judicial law making and supremacy. 33 See, e.g., Stavros Tsakyrakis, Proportionality: An Assault on Human Rights?, 7 Int’l J. Const. L. 468 (2009); Grégoire C. N. Webber, Proportionality, Balancing, and the Cult of Constitutional Rights Scholarship, 23 Can. J.L. & Jurisprudence 179 (2010). Consider the following provocative statement, with which we would agree: “If we ask the question—what does a person have by virtue of a right? 34 See Mattias Kumm, Political Liberalism and the Structure of Rights: On the Place and Limits of the Proportionality Requirement, in Law, Rights and Discourse: Themes from the Legal Philosophy of Robert Alexy 131, 131 (George Pavlakos ed., 2007). The answer given to Kumm’s question is ours, and not necessarily Kumm’s. —the answer, under PA, is that a right gives a right-bearer an entitlement to have her claim evaluated under the proportionality framework, and nothing more.”

Many who believe that rights express (or positivize as constitutional law) moral principles, 35 See, e.g., Ronald Dworkin, Taking Rights Seriously 184–205 (1977) (arguing that moral aspects of rights make balancing inappropriate except in emergency situations); Kai Möller, Balancing and the Structure of Constitutional Rights, 5 Int’l J. Const. L. 453, 458–61 (2007) (arguing that constitutional rights ought to be described from the standpoint of substantive morality). or that rights constitute shields (against government action), 36 See, e.g., Frederick Schauer, A Comment on the Structure of Rights, 27 Ga. L. Rev. 415, 429 (1993) (describing constitutional rights in the United States as “shields”). or that rights have a categorical, rule-like quality (rather than being “principles” to be optimized), 37 See, e.g., Jed Rubenfeld, The First Amendment’s Purpose, 53 Stan. L. Rev. 767 (2001) (arguing that the rule-like structure of the First Amendment requires inquiry into government purpose and forbids balancing); Antonin Scalia, The Rule of Law as a Law of Rules, 56 U. Chi. L. Rev. 1175, 1179–80 (1989) (arguing that rights as rules limit the balancing discretion of judges and commit them to judicial restraint). The broader literature is reviewed in Blocher, supra note 4. will also share a commitment to the belief that a more principled mode of rights protection is possible to achieve. We are skeptical that judges can or should dispense with balancing when they adjudicate rights claims, for reasons argued through this Article, but we do not deny that balancing may require some to rethink certain deeply held assumptions about the nature of rights and of rights adjudication.

With respect to separation of powers, PA is clearly not a neutral, analytical procedure. Compared to alternatives, PA is a highly intrusive standard of judicial review. Wherever it has been adopted, PA replaced more deferential standards. (Judges can build de facto deference into PA on an ad hoc basis, but deference, too, is an outcome of the analysis and, thus, a product of judicial decision making.) Those, like Jeremy Waldron, who have taken the view that rights adjudication adds nothing of value to the quality of policy deliberations, and who believe that principles of democracy are violated by systems in which courts have “the final word,” must attack PA as inherently undemocratic. 38 See Jeremy Waldron, Some Models of Dialogue Between Judges and Legislators, in 23 The Supreme Court Law Review 7 (Grant Huscroft & Ian Brodie eds., 2d Ser., 2004); Jeremy Waldron, The Core of the Case Against Judicial Review, 115 Yale L.J. 1346 (2006). It is obvious that PA positions constitutional judges as powerful lawmakers and that, under conditions of structural supremacy, judges will often dominate law-making processes and the evolution of the constitutional law. We see such outcomes as costs that any polity must inevitably pay if it wishes to maintain an effective system of constitutional justice. If a polity does not wish to pay such costs, it must not, within its constitutional law, include a charter of rights meant to be protected by a judicial body whose decisions are effectively insulated from reversal.

In contrast to the views of more fervent supporters of PA, 39 See, e.g., David M. Beatty, The Ultimate Rule of Law 159–76 (2004). we deny the claim that PA gives a unique, correct answer to complex legal questions involving the interpretation and application of rights. As we have already stated, PA is an analytical procedure, a framework for assessing justifications for government policy that would infringe upon rights. PA does not tell judges what weight to give constitutional values that are in tension. At best, when used properly, PA guides or constrains how judges balance once they have a sense of how the contending values are to be weighed. Put differently, balancing will always require some background notions or theories of the nature and scope of rights, the proper role of the state in the society, economy, or private life, and so on. PA does not supply these background ideas. Further, it is far from clear that any court can actually test the “necessity” of a government measure with any precision, let alone determine which policy options fall along a presupposed “Pareto frontier.” These are specific examples of the generic, potentially irresoluble problem with balancing. Balancing typically involves weighing two goods that are formally incommensurable in that the respective values cannot be measured on the same scale or metric.

In our view, these are valid points, though different conclusions can be drawn from them. An opponent of PA may well conclude that, at best, PA is little more than fancy, doctrinal window dressing for what is, in fact, generic law making by any other name. From the same facts, we conclude only that rights adjudication can never be dissociated from law making; that it will regularly involve difficult cases; that there exists no stable, compelling alternative to balancing; and that PA provides the most defensible balancing structure currently available. We will seek to defend the latter two points in the remainder of the Article.

C. Strict Scrutiny

Given the importance of strict scrutiny to rights protection in the United States, one might assume that American readers would need no introduction to it. Until recently, however, little systematic research on the origins, evolution, or law and politics of the doctrine existed. 40We do not mean to understate the importance of research that bears upon how the Court has used (and has changed how it uses) strict scrutiny in specific cases or areas of the law, on which there are dozens of important articles. Since 2005, several important pieces on the topic have appeared, including articles by Richard Fallon, 41 See Fallon, supra note 3. Stephen Siegel, 42 See Stephen A. Siegel, The Origin of the Compelling State Interest Test and Strict Scrutiny, 48 Am. J. Legal Hist. 355, 359–60 (2006). G. Edward White, 43G. Edward White, Historicizing Judicial Scrutiny, 57 S.C. L. Rev. 1 (2005). and Adam Winkler. 44Adam Winkler, Fatal in Theory and Strict in Fact: An Empirical Analysis of Strict Scrutiny in the Federal Courts, 59 Vand. L. Rev. 793 (2006). In this Article, we seek to contribute to this literature as well as to a rapidly emerging comparative research agenda. 45On the comparative agenda, see Beatty, supra note 39; Cohen-Eliya & Porat, supra note 4; Stephen Gardbaum, Limiting Constitutional Rights, 54 UCLA L. Rev. 789 (2007); Mattias Kumm, Constitutional Rights as Principles: On the Structure and Domain of Constitutional Justice, 2 Int’l J. Const. L. 574 (2004) (reviewing Alexy, supra note 21). In 2010, Law & Ethics of Human Rights, edited by Iddo Porat and Moshe Cohen-Eliya, among others, dedicated two symposium issues to the topic of “Rights, Balancing & Proportionality.See Symposium, Rights, Balancing & Proportionality, 4 Law & Ethics Hum. Rts. 1 (2010).

Although we will compare PA and strict scrutiny throughout the Article (most directly in Part III), we wish to emphasize four general points in advance of the analysis to follow. First, as noted, PA overlaps strict scrutiny in crucial respects. In addition to sharing constitutive components, in the form of tests, both frameworks displaced doctrines that were far more deferential to legislative power and government authority. In the United States, what makes the scrutiny “strict” is the fact that it negates the normal presumption that legislation will be treated as constitutionally valid unless the law fails basic rationality requirements. When the Supreme Court decides to protect a right under strict scrutiny, the Court’s de facto supremacy within law-making processes is fully realized.

Second, the Supreme Court developed strict scrutiny in order to deal with a set of specific strategic problems. 46Fallon puts it as follows:In the 1960s, the Warren Court was eager first to establish and then to consolidate a doctrinal structure sharply differentiating preferred from ordinary constitutional rights. With rational basis review established as the norm in run-of-the-mill cases, this strategy required the development of an implementing test or tests to protect preferred rights. Strict judicial scrutiny . . . furnished an attractive model . . . .Fallon, supra note 3, at 1335. Fallon describes the standard as “a judicially crafted formula for implementing constitutional values,” specifically, those rights determined to be “preferred” or most “fundamental.” 47 Id. at 1268. But he also insists that the Court needed something like strict scrutiny “to impose discipline, or at least the appearance of discipline, on judicial decisionmaking and thus to escape the taint both of Lochneresque second-guessing of legislative judgments and of flaccid judicial ‘balancing.’” 48 Id. at 1270. As we have argued throughout this section, PA too gives rights-protecting courts a means to rationalize and discipline judicial review, and balancing in particular. It is also true that, at times, the Court seeks to render balancing invisible, notably by treating a right as quasi-absolute.

Third, strict scrutiny contains within it space for balancing—indeed, strict scrutiny began as a structure for balancing—and the Court has never been able to banish balancing from it. 49 See Siegel, supra note 42, at 394–97. Siegel shows that strict scrutiny was not, in its original guise, designed to aid motive analysis by “smoking out” illicit motives. Id. at 394 (internal quotation marks omitted). Rather, the formula was a refinement of the balancing-of-interest approach that had long appeared in First Amendment cases. See id. (“When the Court introduced narrow tailoring and the compelling state interest standard into First Amendment analysis, it did so as part of its general ‘balancing/cost-benefit justification’ approach to First Amendment questions.”). Only later, once it migrated into the equal protection context, was strict scrutiny deployed as a tool for smoking out illegitimate motives. Id. at 394–97. As we describe in more detail in the next Part, the modern strict scrutiny formula was first consolidated in First Amendment cases, where it represented a more rigorous and rights-favoring update of an approach to expressive rights that the Court had long employed. As Stephen Siegel points out, the rights absolutists Black and Douglas concurred in the early First Amendment decisions applying strict scrutiny, rather than join the Court’s opinion, so as not to endorse a balancing approach to rights. 50 Id. at 394–95. Indeed, the structure is sometimes overtly deployed as a “weighted” or “all-things-considered balancing test.” 51Fallon, supra note 3, at 1306. Generally, however, the scope for balancing within the strict scrutiny framework contracted sharply by the end of the 1960s, owing to the framework’s embrace by an ascendant civil libertarian majority on the Supreme Court and its migration into equal protection jurisprudence, where it was typically deployed to a different end: to ferret out illegitimate motives. 52 See infra text accompanying notes 179–82.

The fourth point follows from the third. To allow proportionality analysis in domains where tiered review now occupies the field is not, as some critics would have it, somehow unprincipled or unsanctioned by precedent. To the contrary, restoring balancing is faithful to the foundational strict scrutiny precedents and recovers an aspect of our constitutional practice that has been obscured. This point is developed further in Parts II and IV.

II. Roots of Proportionality in U.S. Rights Review

Building on those final two points, this Part seeks to recover a lost history of proportionality, or perhaps “proto-proportionality,” in American rights doctrines. 53In a similar spirit, Thomas Sullivan and Richard Frase identify sites in American doctrine, constitutional and otherwise, where some form of proportionality principle operates. See E. Thomas Sullivan & Richard S. Frase, Proportionality Principles in American Law 53–168 (2009). This undertaking is important, but different from our own. Sullivan and Frase take a somewhat more capacious view of proportionality than we do, and they identify a large number of areas of law where some form of interest-balancing exercise takes place. See id. at 6–7. We are focused specifically on doctrinal structures that anticipate the multistep framework of modern proportionality analysis, where judicial scrutiny mounts through a prescribed sequence of stages, culminating in means–ends testing and then balancing stricto sensu. As we show below, this same sequence of steps emerges more than once, in different areas of constitutional doctrine. Our aim is not to offer a comprehensive or definitive history, but rather a high-speed and selective tour, focusing on those sites that hold particular interest from a comparative perspective on rights adjudication. We look not only at the consolidation of the strict scrutiny framework in the 1950s and 1960s, but also chart the development of dormant Commerce Clause doctrine in the late nineteenth century. In both instances, we find doctrinal structures with striking resemblances to proportionality, both formally and functionally. 54 See infra Part III.A (highlighting similarities between mid-century rational basis review and proportionality).

The picture that emerges stands in sharp contrast to the view of those who claim a fundamental incompatibility between proportionality and American constitutionalism. 55 See, e.g., District of Columbia v. Heller, 554 U.S. 570, 634–35 (2008). To the contrary: doctrinal structures that approximate proportionality are a recurrent feature of our constitutional practice, dating back more than a century. Given the importance of doctrinal continuity in American law, exposing this history subtly shifts the burden of justification: the move to reject a proportionality-based balancing framework requires explanation at least as much as the move to embrace it.

A. The Dormant Commerce Clause

The early dormant Commerce Clause cases deserve our attention for a number of reasons. It is in this domain that the LRM test first appears as a stable feature of American constitutional law, along with the “undue burden” standard. 56Guy Miller Struve claims that the “roots of the principle” go back at least to Lawton v. Steele, 152 U.S. 133 (1894). Guy Miller Struve, The Less-Restrictive-Alternative Principle and Economic Due Process, 80 Harv. L. Rev. 1463, 1464 n.4 (1967). In fact, Lawton absorbs the LRM test from prior dormant Commerce Clause cases.The first glimmers of LRM testing appear even earlier. In the 1821 case Anderson v. Dunn, challenging an exercise of the congressional contempt power, the Court ruled that Congress may only employ “the least possible power adequate to the end proposed.” 19 U.S. (6 Wheat.) 204, 230–31 (1821). But it took several decades before this test was employed systematically in an important doctrinal area, as we describe in the main text. Second, in asserting judicial authority to review the necessity of state “police acts,” the Court had to overcome the view, deeply held by many judges at both the state and federal levels, that assessing the necessity of public policy was a legislative, not a judicial, function. Third, the Court’s dormant Commerce Clause jurisprudence was part and parcel of a broader move to give rigorous constitutional protection to freedom of contract and property rights. The Court would retreat from this stance post-New Deal, of course, but not in the domain governed by the dormant Commerce Clause. Last, viewed comparatively, the Court’s doctrine in the area is all but indistinguishable from PA. We will explore each of these points in turn.

Since Chief Justice Marshall’s opinion in the 1829 case Willson v. Black Bird Creek Marsh Co., the Supreme Court has interpreted the Commerce Clause 57“The Congress shall have Power . . . [t]o regulate Commerce with foreign Nations, and among the several States, and with the Indian Tribes.” U.S. Const. art. I, § 8, cl. 3. as authorizing the states to “regulate commerce in its dormant state.” 5827 U.S. (2 Pet.) 245, 252 (1829). States in the Union possess a presumptive “right” to use their police powers to regulate market activity, in the absence of federal preemption, albeit under the supervision of the federal courts. Prior to the Civil War, the U.S. Supreme Court confronted few (if any) important cases in the area, and it established no lasting doctrine that is relevant here. 59The one case still cited, see, e.g., United Haulers Ass’n v. Oneida-Herkimer Solid Waste Mgmt. Auth., 550 U.S. 330, 338 (2007), is Cooley v. Board of Wardens, 53 U.S. (12 How.) 299 (1851), in which the Court considers the basis for permitting local governments to take measures that would affect commerce:Now the power to regulate commerce embraces a vast field, containing not only many, but exceedingly various subjects, quite unlike in their nature; some imperatively demanding a single uniform rule, operating equally on the commerce of the United States in every port; and some, like the subject now in question, as imperatively demanding that diversity, which alone can meet the local necessities of navigation.Id. at 319. During the period from 1875 to 1902, the Court faced a rising tide of litigation brought by merchants and traders seeking to invalidate state regulations whose effect was to prohibit or burden interstate commerce. It met this challenge by developing a full-fledged version of PA: the Court begins with an inquiry into the legitimacy and importance of state purposes; it then assesses the necessity of the regulations through the deployment of an LRM test; and it fashions a place for balancing, in the form of an unreasonable burden standard. During this period, the Court also derived, from the Commerce Clause, an individual right to buy and sell goods across borders. 60 See infra text accompanying notes 71–81.

The LRM test first surfaced in the 1875 Supreme Court decision Chy Lung v. Freeman. 6192 U.S. 275 (1875). The case involved the extortion of Chinese immigrants by California officials who claimed that impounding shipping vessels and their cargo—in this case, women alleged to be sex workers—was necessary to secure the public order. 62 Id. at 278–79. Although the facts are not typical for a dormant Commerce Clause case, the Supreme Court ritually cited Justice Miller’s decision in all cases throughout the seminal period:

We are not called upon by this statute to decide for or against the right of a State, in the absence of legislation by Congress, to protect herself by necessary and proper laws against paupers and convicted criminals from abroad; nor to lay down the definite limit of such right. . . . Such a right can only arise from a vital necessity for its exercise, and cannot be carried beyond the scope of that necessity. . . . The statute of California goes so far beyond what is necessary, or even appropriate, for this purpose, as to be wholly without any sound definition of the right under which it is supposed to be justified. 63 Id. at 280.

Two years later, in Railroad Co. v. Husen, the Court confronted a Missouri embargo on “Texas, Mexican, [and] Indian” cattle, which was put in place for eight months of the year. 6495 U.S. 465, 468–69 (1877). The State argued that the measure was necessary on health grounds, given that these particular cattle carried diseases that could infect and devastate Missouri’s herds. 65 Id. at 473. A unanimous Court struck down the law. 66 Id. In dicta, Justice Strong noted that the State had legitimate interests in regulating interstate commerce under a number of public policy headings, including “domestic order, morals, health, and safety.” 67 Id. at 470–71. He continued:

We are thus brought to the question whether the Missouri statute is a lawful exercise of the police power of the State. . . . It may . . . be admitted that the police powers of a State justifies the adoption of precautionary measures against social evils. . . .

. . . While we unhesitatingly admit that a State may pass sanitary laws, and laws for the protection of life, liberty, health, or property within its borders; while it may prevent persons and animals suffering under contagious or infectious diseases, or convicts . . . from entering the State; while for the purpose of self-protection it may establish quarantine, and reasonable inspection laws, it may not interfere with transportation into or through the State, beyond what is absolutely necessary for its self-protection. 68 Id. at 470–72.

In his ruling, Justice Strong noted that neither the Missouri legislature nor the Missouri Supreme Court (upholding the law) had considered less restrictive alternatives to an outright ban, such as a quarantine or animal inspections. 69 Id. at 472. Either option would have protected a valid state interest while reducing the burden on interstate commerce. This appears to be plain PA: Because Missouri chose the complete “destruction” of commerce 70 Id. at 470. when less intrusive options were reasonably available, the law could not stand.

Still missing from this emerging framework was a conception of individual constitutional rights. The Court first began to fill this lacuna explicitly in the 1890 case Minnesota v. Barber. 71136 U.S. 313 (1890). The litigation targeted a Minnesota statute that prohibited the sale of fresh meat unless the livestock was inspected—within the state—a maximum of twenty-four hours before its slaughter. 72 Id. at 318. The statute was facially neutral, applying to all such products regardless of origin, which Minnesota argued should insulate the law from censure. 73 Id. at 326. The Court rejected this assertion:To this we answer, that a statute may, upon its face, apply equally to the people of all the States, and yet be a regulation of interstate commerce which a State may not establish. A burden imposed by a State upon interstate commerce is not to be sustained simply because the statute imposing it applies alike to the people of all the States, including the people of the State enacting such statute.Id. But because of the twenty-four-hour time limitation, slaughterhouses in the surrounding states would find it impossible to transfer their cattle for inspection, ship them back to the slaughterhouse for processing, and then return the meat to Minnesota for sale. 74 Id. at 322. The case thus involved generally applicable measures that created differential burdens that, plaintiffs claimed, constituted discrimination. Writing for the Court, Justice Harlan accepted the state’s claim that the law “was enacted, in good faith, . . . to protect the health of the people of Minnesota,” but then struck it down. 75 Id. at 319, 329–30. The state’s defense of the law, Harlan complained,

ignores the right which the people of other States have in commerce between those States and the State of Minnesota. And it ignores the right of the people of Minnesota to bring into that State, for purposes of sale, sound and healthy meat, wherever such meat may have come into existence. 76 Id. at 329 (emphases added).

Moreover, the statute failed the LRM test, among other reasons, because it did not recognize the efficacy of the inspection and certification processes in place in home states of slaughterhouses. 77 Id. at 322 (“It will not do to say—certainly no judicial tribunal can with propriety assume—that the people of Minnesota may not, with due regard to their health, rely upon inspections in other states of animals there slaughtered for purposes of human food.”). Here we find the germ of what international trade law today calls the principle of mutual recognition, which the European Court of Justice (ECJ) articulated in its celebrated free movement of goods judgment, Cassis de Dijon. 78Case 120/78, Rewe-Zentral AG v. Bundesmonopolverwaltung fur Branntwein (Cassis de Dijon), 1979 E.C.R. 649. The ECJ explicitly speaks of the right of individuals to engage in intra-European Community trade, which it derived from the prohibition of nontariff barriers under Article 28 of the European Economic Community Treaty. There is another important parallel between Cassis de Dijon and Minnesota v. Barber. In Cassis de Dijon, the ECJ (controversially) extended proportionality review to national market regulations that applied equally to goods, regardless of their origin. Prior to Cassis de Dijon, the ECJ limited proportionality review to measures that directly regulated trade. See Miguel Poiares Maduro, We the Court: The European Court of Justice and the European Economic Constitution 106–09 (1998); Alec Stone Sweet, The Judicial Construction of Europe 109–45 (2004). The U.S. Supreme Court would also come to rely heavily on the principle of mutual recognition, when performing its own necessity analysis.

By the turn of the twentieth century, the Court had expressly recognized a constitutional right to buying and selling across state borders. In Reid v. Colorado, another case involving a state’s regulation of trade in livestock under the heading of protecting health, Justice Harlan stated for the Court:

Now, it is said that the defendant has a right under the Constitution of the United States to ship live stock from one State to another State. This will be conceded on all hands. But the defendant is not given by that instrument the right to introduce into a State, against its will, live stock affected by a contagious, infectious or communicable disease, and whose presence in the State will or may be injurious to its domestic animals. The State—Congress not having assumed charge of the matter . . . —may protect its people and their property against such dangers, taking care always that the means employed to that end do not go beyond the necessities of the case or unreasonably burden the exercise of privileges secured by the Constitution of the United States. 79 187 U.S. 137, 151 (1902) (first two emphases added).

Justice Harlan thus neatly expressed the constituent elements of PA. Although the standard of scrutiny is as robust as any in the Court’s repertoire, the Court nonetheless upheld the necessity of Colorado’s measure, 80 Id. at 152–53. as it had done in prior cases. 81 Compare Plumley v. Massachusetts, 155 U.S. 461 (1894) (upholding a state law regulating the sale of margarine on consumer protection grounds), with Schollenberger v. Pennsylvania, 171 U.S. 1 (1898) (striking down as unconstitutional a state law regulating the sale of margarine on LRM grounds).

There is an evident awkwardness to the Court’s derivation from the Commerce Clause of an individual, constitutional right to trade across state borders, let alone the treatment of such a right as impossible to impugn in a legal proceeding. Nowhere does the Federal Constitution proclaim such a right, and in formal terms, the right’s existence depends entirely upon the U.S. Congress not taking action under the Commerce Clause. Yet, by the time Reid v. Colorado was decided, the Court had already embarked on its venture to construct and defend property rights as “preferred freedoms” (in all but name). Not surprisingly, the dormant Commerce Clause jurisprudence of this period became imbued with the Court’s position on freedom of contract, and due process more generally.

Doctrinally, the Court prioritized economic liberty as a constitutional value by constructing a prototype of PA, or strict scrutiny. At its core is an LRM test. Although the LRM test was born in dormant Commerce Clause jurisprudence, it quickly spread to other areas in which the courts reviewed state police power regulations restricting economic freedom. 82Howard Gillman explains this development succinctly. See Howard Gillman, Preferred Freedoms: The Progressive Expansion of State Power and the Rise of Modern Civil Liberties Jurisprudence, 47 Pol. Res. Q. 623, 629–34 (1992). Nineteenth-century judges understood the legislative power to have built-in limits: legislation that interfered with private liberty or property interests was legitimate only when it promoted the general welfare of the whole. Id. at 629. When legislative acts could not be so justified—when they were arbitrary or class legislation, enacted for the benefit of certain groups—they would be held to be improper exercises of legislative power and, hence, invalid. Id. By the late nineteenth century, federal courts had come to police the boundaries of legislative power with a form of LRM test. Id. at 634. Lawton v. Steele, 152 U.S. 133 (1894), illustrates this approach in practice. The case concerned a New York statute that declared fish nets in the vicinity of two waterways to be a public nuisance and subject to abatement and destruction by any person. Id. at 135–36. This provision was alleged to exceed the state’s police power. The Court framed its approach to the question as follows:To justify the State in thus interposing its authority in behalf of the public, it must appear, first, that the interests of the public generally, as distinguished from those of a particular class, require such interference; and, second, that the means are reasonably necessary for the accomplishment of the purpose, and not unduly oppressive upon individuals. The legislature may not, under the guise of protecting the public interests, arbitrarily interfere with private business, or impose unusual and unnecessary restrictions upon lawful occupations. In other words, its determination as to what is a proper exercise of its police powers is not final or conclusive, but is subject to the supervision of the courts.Id. at 137. Ultimately, the Court sided with the state, while strongly voicing its support of property rights more generally. Id. at 142–43. Three Justices dissented on the grounds that, in their view, the State had violated the plaintiff’s constitutional right to property because necessity had not been met. Id. at 144 (Fuller, C.J., dissenting). It did so, in our view, because it broadly fit a general orientation of the judiciary to recognize a sphere of private liberty—in the form of vested economic rights—that was viewed by the judicial elite as constituting an inherent, but not quite absolute, restriction on the exercise of legislative authority. 83Moreover, congressional powers were limited to the exercise of those powers enumerated in Article I, Section 8 of the Federal Constitution.

In looking back over the foundational dormant Commerce Clause cases, it is important to note the variability of outcomes: sometimes the challenged measure was upheld, and sometimes it was struck down. Like PA and strict scrutiny, dormant Commerce Clause analysis imposes a burden of justification on the state, and the justifications offered will be subjected to searching scrutiny. But these frameworks do not contain, in and of themselves, anything like a strong presumption that the measures under review are unconstitutional. One of the virtues of proportionality balancing is that it allows a court to claim doctrinal consistency while retaining flexibility across time and cases. Courts commit a strategic error when they destroy this virtue by using the framework in an outcome-determinate way and, in effect, rigging outcomes in advance. The Supreme Court grew increasingly rigid in its jurisprudence of property rights and due process in the early twentieth century, just as it did years later with respect to strict scrutiny. 84 See infra text accompanying note 187. The discredit and analytical difficulties that ensued from these doctrinal moves add support to our case for PA.

PA and strict scrutiny are still balancing frameworks whose use will inexorably blur distinctions between legislative and judicial functions. In the dormant Commerce Clause cases of the pre-New Deal era, the Court routinely confronted resistance to LRM testing, on the part of state judges and some of its own members. As Justices Gray and Harlan charged, dissenting in Schollenberger v. Pennsylvania: necessity analysis cannot be separated from “questions of fact and of public policy, the determination of which belongs to the legislative department, and not to the judiciary.” 85171 U.S. at 29–30 (Gray, J., dissenting). We think that the PA-style framework that emerged in the late nineteenth century made sense in the dormant Commerce Clause domain, just as it does today. Dormant Commerce Clause cases involve important constitutional values that routinely conflict with one another in our federal polity, and being highly context-specific, they are difficult to resolve using a more rule-based approach. Further, insofar as the Court deploys LRM testing rigorously and effectively, states will be led to reduce reliance on measures that facially discriminate against out-of-state goods and services. Instead, they will turn to measures that are equally applicable to all goods, services, and traders. If state purposes are, in fact, protectionist in nature, some states will always be tempted to disguise these purposes by claiming some important or compelling public interest. This is, in fact, what happened. In response, the Court began to deploy LRM testing as a means of “smoking out” protectionist motives, using an undue burden test to deal with the residual interests at stake. This approach is defensible given the importance of an overarching value: the building and maintenance of market federalism.

By the end of the 1950s, the Supreme Court had abandoned the task of protecting economic liberties under the Due Process Clause. The basics of dormant Commerce Clause doctrine, however, remained largely intact, although two changes deserve mention. First, the post-New Deal era witnessed the virtual disappearance of reference to the constitutional right of individuals to engage in interstate commerce—although remnants of this right remain today, as Granholm v. Heald confirms. 86544 U.S. 460, 473 (2005) (“Laws of the type at issue . . . . deprive citizens of their right to have access to the markets of other States on equal terms. The perceived necessity for reciprocal sale privileges risks generating the trade rivalries and animosities, the alliances and exclusivity, that the Constitution and, in particular, the Commerce Clause were designed to avoid.”). Instead, the Court came to rely on one of several (largely compatible) theories of the Commerce Clause to justify jurisdiction. 87 Compare H. P. Hood & Sons, Inc. v. Du Mond, 336 U.S. 525, 539 (1949) (“Our system, fostered by the Commerce Clause, is that every farmer and every craftsman shall be encouraged to produce by the certainty that he will have free access to every market in the Nation, that no home embargoes will withhold his exports, and no foreign state will by customs duties or regulations exclude them.”), with Donald H. Regan, The Supreme Court and State Protectionism: Making Sense of the Dormant Commerce Clause, 84 Mich. L. Rev. 1091 (1986) (arguing that the dominant value protected by the dormant Commerce Clause is “national unity,” and that the Constitution’s dormant Commerce Clause corrects one of the main defects created by the Articles of Confederation—the ability of states to legislate benefits for their own citizens at the expense of the rest of the nation), and Hughes v. Oklahoma, 441 U.S. 322, 325–26 (1979) (noting that caselaw in this domain “reflect[s] a central concern of the Framers that was an immediate reason for calling the Constitutional Convention: the conviction that in order to succeed, the new Union would have to avoid the tendencies toward economic Balkanization that had plagued relations among the Colonies and later among the States under the Articles of Confederation”), and S.C. State Highway Dep’t v. Barnwell Bros., 303 U.S. 177, 185 n.2 (1938) (“Underlying the stated rule has been the thought, often expressed in judicial opinion, that when the regulation is of such a character that its burden falls principally upon those without the state, legislative action is not likely to be subjected to those political restraints which are normally exerted on legislation where it affects adversely some interests within the state.”). This change has had little, if any, practical effect on how parties litigate cases, or how the Court decides them. Not so with regard to the second change. In the 1950s, the Court began to treat state and local regulations that facially discriminate against interstate commerce as presumptively unconstitutional, 88 See, e.g., Dean Milk Co. v. City of Madison, 340 U.S. 349, 354 (1951). although the state or local authority could, in theory, rebut that presumption by showing that its interest was both sufficiently important and “unrelated to economic protectionism.” 89Wyoming v. Oklahoma, 502 U.S. 437, 454–55 (1992). Some Justices began explicitly using the phrase “strict scrutiny” to describe what the Court did in such cases. 90 See, e.g., United Haulers Ass’n v. Oneida-Herkimer Solid Waste Mgmt. Auth., 550 U.S. 330, 360 (2007) (Alito, J., dissenting) (“The Court has long subjected discriminatory legislation to strict scrutiny, and has never, until today, recognized an exception for discrimination in favor of a state-owned entity.”); cf. C & A Carbone, Inc. v. Town of Clarkstown, 511 U.S. 383, 392 (1994) (characterizing standard as “rigorous scrutiny”); id. at 422 (Souter, J., dissenting) (characterizing standard as “virtually fatal scrutiny”). The more difficult litigation, of course, concerns facially neutral market regulations that incidentally burden interstate commerce. The leading decision is Pike v. Bruce Church, Inc., which puts balancing at the core of the framework:

Although the criteria for determining the validity of state statutes affecting interstate commerce have been variously stated, the general rule that emerges can be phrased as follows: Where the statute regulates evenhandedly to effectuate a legitimate local public interest, and its effects on interstate commerce are only incidental, it will be upheld unless the burden imposed on such commerce is clearly excessive in relation to the putative local benefits. 91397 U.S. 137, 142 (1970).

Although dormant Commerce Clause doctrine is one of the most durable structural features of American constitutional law, it has also been subjected to relentless attacks both from within and beyond the Court, and these assaults have mounted steadily in the past two decades. The most recent dormant Commerce Clause case of significance, United Haulers, provides a good example. In his opinion for the Court, Chief Justice Roberts succinctly summarized the core elements of this established doctrine. 92550 U.S. at 338–39, 346 (“Discriminatory laws motivated by ‘simple economic protectionism’ are subject to a ‘virtually per se rule of invalidity,’ which can only be overcome by a showing that the State has no other means to advance a legitimate local purpose. . . . [But t]he Counties’ . . . ordinances are properly analyzed under the test set forth in Pike v. Bruce Church, Inc., which is reserved for laws ‘directed to legitimate local concerns, with effects upon interstate commerce that are only incidental.’” (citations omitted)). He then proceeded to balancing (weighing the benefits of a county ordinance in dealing with a waste management crisis against the costs to trucking companies of higher tipping fees), finding that “any arguable burden the ordinances impose on interstate commerce does not exceed their public benefits.” 93 Id. at 347. In his concurrence, Justice Scalia reiterated his view that the whole edifice of dormant Commerce Clause doctrine was “an unjustified judicial invention” with no constitutional foundation. 94 Id. at 348 (Scalia, J., concurring in part) (quoting General Motors Corp. v. Tracy, 519 U.S. 278, 312 (1997) (Scalia, J., concurring)). He further dismissed the “so-called ‘Pike balancing’” on the grounds that “[g]enerally speaking, the balancing of various values is left to Congress—which is precisely what the Commerce Clause (the real Commerce Clause) envisions.” 95 Id. at 348–49. Whereas Justice Scalia would adhere, restrictively, to the judicial enforcement of the dormant Commerce Clause “on stare decisis grounds,” 96 Id. at 348. Justice Thomas would “discard” the Court’s jurisprudence altogether, for similar reasons. 97 Id. at 349 (Thomas, J., concurring in the judgment) (“The negative Commerce Clause has no basis in the Constitution and has proved unworkable in practice. As the debate between the majority and dissent shows, application of the negative Commerce Clause turns solely on policy considerations, not on the Constitution. Because this Court has no policy role in regulating interstate commerce, I would discard the Court’s negative Commerce Clause jurisprudence.” (citations omitted)). Regarding balancing, Justice Thomas had this to say:To the extent that Congress does not exercise its authority to make that choice, the Constitution does not limit the States’ power to regulate commerce. In the face of congressional silence, the States are free to set the balance between protectionism and the free market. Instead of accepting this constitutional reality, the Court’s negative Commerce Clause jurisprudence gives nine Justices of this Court the power to decide the appropriate balance.Id. at 352. The attack on balancing has also been as ferocious on the part of leading academic commentators in the area, notably, Donald Regan. 98Simplifying a complex position, Regan argues that the courts should rarely engage in LRM testing and never engage in balancing, under Pike or any other authority, but should limit themselves exclusively to inquiry into “legislative purpose.” If the Court finds discriminatory purpose, then it must strike down the law. Regan further argues that the Supreme Court does not really balance at all, although it claims to do so. See Donald H. Regan, Judicial Review of Member-State Regulation of Trade Within a Federal or Quasi-Federal System: Protectionism and Balancing, Da Capo, 99 Mich. L. Rev. 1853, 1870–71 (2001); Regan, supra note 87. We find many of Regan’s readings of the Court’s caselaw strained at best, and his interpretation seems impossible to square with the two most recent Supreme Court rulings in the area, United Haulers and Granholm v. Heald.

We conclude on a comparative note. The Supreme Court’s approach to the dormant Commerce Clause, whether viewed from the vantage point of the year 1900 or the year 2000, would be immediately recognizable to any European as a familiar, remarkably straightforward version of proportionality. The proportionality framework first emerged as a principle of European public law in the German States, notably Prussia, during precisely the same period—the late nineteenth century. 99 See Stone Sweet & Mathews, supra note 5, at 101 (“[B]y the late nineteenth century, German administrative courts were striking down [state] police actions that violated proportionality, which was conceptualized at that time as an enforceable LRM test.”). In the 1970s, the ECJ developed a virtually identical framework to deal with “free movement of goods” litigation under the Treaty of Rome. 100 Stone Sweet, supra note 78, at 109–45. These cases, brought by traders challenging Member State market regulations in the national courts, dominated the ECJ’s docket throughout the 1970s and 1980s. 101 Id. The ECJ used PA, in this and related areas, to help jump-start efforts to complete the Single Market in the face of protectionism and collective action failures on the part of the Member States. 102 Id. In the WTO, the Appellate Body would later adopt PA to deal with the legal issues that most resemble those faced by the Supreme Court and the ECJ. 103 See Stone Sweet & Mathews, supra note 5, at 152–59. It may be that market integration can proceed across state borders without courts as commitment devices and without intrusive but flexible doctrines like proportionality, but we are skeptical. 104For a view of courts and judicial review as guarantors of credible commitments between federated states engaged in market-building projects, see Walter Mattli, The Logic of Regional Integration: Europe and Beyond (1999); Stone Sweet, supra note 78, at 7–9, 109–45 (with reference to Europe); Jenna Bednar & William N. Eskridge, Jr., Steadying the Court’s “Unsteady Path”: A Theory of Judicial Enforcement of Federalism, 68 S. Cal. L. Rev. 1447 (1995) (with reference to the United States); Martin Shapiro, The European Court of Justice, in The Evolution of EU Law 321 (Paul Craig & Gráinne de Búrca eds., 1999).

While the comparative parallels are compelling, it is also obvious that the Court’s dormant Commerce Clause jurisprudence has maintained an affinity to strict scrutiny that reaches back more than a century. While the familiar strict scrutiny formula, coupling a compelling governmental interest standard with an LRM test, first appeared in First Amendment 105 See Siegel, supra note 42 (arguing that strict scrutiny first emerged in First Amendment cases). and equal protection jurisprudence, 106 See Melvin I. Urofsky, Division and Discord: The Supreme Court Under Stone and Vinson,1941–1953, at 89–90 (1997) (locating strict scrutiny’s origins in equal protection doctrine). it is striking and significant that a functional equivalent developed in the nineteenth century within the Court’s approach to the Commerce Clause, and later, to property rights more generally. Indeed, a plausible case can be made for the view that the origins of strict scrutiny are neither in equal protection law nor in First Amendment jurisprudence, as is commonly argued, 107 See, e.g., Fallon, supra note 3 (arguing that strict scrutiny emerged almost simultaneously in a number of areas). but are found within the Court’s approach to the Commerce Clause. One might even conclude that the Court repurposed a doctrinal structure originally developed to protect one era’s set of “preferred freedoms”—property rights, in the pre-New Deal period—to serve the set of values favored in a later era—free speech and civil rights, in the post-war years.

B. Strict Scrutiny and Rights Protection

In the twentieth century, dormant Commerce Clause doctrine was an island of stability in a sea of constitutional upheaval. The heightened scrutiny of police power actions infringing economic liberties did not last. It was overtaken by the decisive changes in the Supreme Court’s constitutional jurisprudence during the New Deal, when the Court famously relaxed its scrutiny of social and economic legislation. 108 See, e.g., Nebbia v. New York, 291 U.S. 502 (1934) (upholding an order of the New York Milk Control Board that fixed milk prices); Shaman, supra note 10, at 162–63. But in a broader view, those changes of the New Deal Court were part of a still larger redistribution of judicial scrutiny that took two more decades to complete. The Court’s retreat from a generalized scrutiny of legislation preceded—and precipitated—a concentration of scrutiny around a set of “preferred” freedoms. 109 See Jones v. City of Opelika, 316 U.S. 584, 608 (1942) (Stone, C.J., dissenting) (“[T]he Constitution, by virtue of the First and the Fourteenth Amendments, has put those freedoms [of speech and religion] in a preferred position.”). The favored technique for testing restrictions on those freedoms, reached after years of experimentation and conflict, was strict scrutiny.

Drawing on recent scholarship, this section briefly describes the origins of strict scrutiny and corrects two common misconceptions. First, strict scrutiny emerged from First Amendment doctrine, not from equal protection doctrine. 110Siegel, supra note 42, at 364–80. Second, and more important for our purposes, the Supreme Court introduced strict scrutiny not as a rigid, outcome-determinative rule, but “as part of its general ‘balancing/cost-benefit justification’ approach to First Amendment questions.” 111 Id. at 394.

By the 1930s, the Supreme Court had begun to afford social and economic legislation a presumption of constitutionality 112 See Shaman, supra note 10, at 161–63. through application of a rational basis test: legislation forfeits the presumption only when it lacks a rational basis in the law. 113 Id. Time and again, the Court demonstrated just how weak a constraint this was. 114 See, e.g., Williamson v. Lee Optical of Okla., Inc., 348 U.S. 483 (1955) (upholding the constitutionality of an Oklahoma statute prohibiting persons other than optometrists or ophthalmologists from fitting or replacing lenses); see also infra Part III.A.

But another development paralleled this slackening of scrutiny. The Court’s new posture of deference heightened the concern of some Justices that important civil liberties remained underprotected. The idea that certain rights deserve increased protection had advocates on the Court for years, including Justices Harlan, Holmes, and Brandeis. 115 See Gillman, supra note 82, at 640–44. The presumption of constitutionality lowered the baseline of protection and made those concerns all the more pressing. 116 Id. at 640–47.Carolene Products’ famous footnote four gave voice to this anxiety. The presumption of constitutionality may have a “narrower scope,” Justice Stone submitted, when legislation runs up against specific constitutional prohibitions, notably those found in the Bill of Rights. 117United States v. Carolene Prods. Co., 304 U.S. 144, 153 n.4 (1938). Justice Stone also identified two other classes of statutes that deserve greater scrutiny, previewing two of the abiding preoccupations of the Court’s constitutional jurisprudence over the next half-century: legislation affecting the political process and legislation targeting minorities. 118 Id. This footnote in Carolene Products expressed the ambition of the bifurcated review project: to develop an elevated standard of review in order to complement rational basis review for those rights requiring special solicitude. 119The phrase “bifurcated review project” is G. Edward White’s. See G. Edward White, The First Amendment Comes of Age: The Emergence of Free Speech in Twentieth-Century America, 95 Mich. L. Rev. 299, 301 (1996). White argued that the start of this project predates Carolene Products and was motivated specifically by a desire to afford heightened protection to speech rights. Id. at 301–02, 308–23. But it was some time before the Court arrived at a stable formula for operationalizing this solicitude for “preferred” rights—namely, strict scrutiny analysis.

It is often asserted—not least, in the opinions of the Supreme Court—that strict scrutiny originated in equal protection cases dating from the 1940s, 120See, e.g., Simon & Schuster, Inc. v. Members of the N.Y. State Crime Victims Bd., 502 U.S. 105, 124–28 (1991) (Kennedy, J., concurring in the judgment); Goldberg, supra note 10; Greg Robinson & Toni Robinson, Korematsu and Beyond: Japanese Americans and the Origins of Strict Scrutiny, Law & Contemp. Probs., Spring 2005, at 29; Peter J. Rubin, Reconnecting Doctrine and Purpose: A Comprehensive Approach to Strict Scrutiny After Adarand and Shaw, 149 U. Pa. L. Rev. 1, 3 n.1 (2000).Skinner v. Oklahoma 121316 U.S. 535, 541 (1942). and Korematsu v. United States. 122323 U.S. 214 (1944). But while the phrases “strict scrutiny” and “most rigid scrutiny” appear in these decisions, 123 Skinner, 316 U.S. at 541; Korematsu, 323 U.S. at 216. the Court actually applied strict scrutiny in neither case. 124Indeed, the phrase strict scrutiny appears in isolated decisions dating back to the late nineteenth century. See Winkler, supra note 44, at 798 n.10 (citing, inter alia, Altschuler v. Coburn, 57 N.W. 836, 838 (Neb. 1894), and Booher v. Worrill, 57 Ga. 235, 238 (1876)). We follow Stephen Siegel in taking strict scrutiny to consist of three parts: (1) a burden shifting to the government to justify the constitutionality of its law, and the requirement that a statute must be (2) “narrowly tailored” to serve (3) “a compelling state interest.” See Siegel, supra note 42, at 359–60. As Stephen Siegel argues, the roots of strict scrutiny analysis are in fact found in First Amendment cases from the late 1950s and early 1960s. 125 See Siegel, supra note 42, at 361–80. And in looking there, we find that strict scrutiny originated not as a blunt hammer for striking down legislation, but a flexible instrument that combined new and old doctrinal elements to balance the benefits and costs of rights-infringing legislation. 126Good accounts of this development from different perspectives can be found in Fallon, supra note 3, at 1285–97, and Gillman, supra note 82, at 640–44.

The consolidation of strict scrutiny capped a decades-long struggle on the Court over how much protection to afford First Amendment rights, and by means of which doctrinal techniques. 127 See Martin Shapiro, Freedom of Speech: The Supreme Court and Judicial Review 46–107 (1966). This contest for the soul of the First Amendment, which is outlined only in broad strokes here, 128For a more comprehensive treatment, see id. generated a good deal of doctrinal innovation, as Justices angled for advantage within a densifying matrix of precedents by introducing new tests or altering existing ones. A narrow majority of civil libertarian Justices generally held the upper hand from the late 1930s through the early 1950s, upholding most First Amendment challenges to speech-restrictive statutes, often by applying the “clear and present danger” test. 129 Id. at 58. The test was originally articulated in Justice Holmes’s 1919 opinion for the Court in Schenck v. United States, 249 U.S. 47, 52 (1919) (“The question in every case is whether the words used are used in such circumstances and are of such a nature as to create a clear and present danger that they will bring about the substantive evils that Congress has a right to prevent.”). These Justices also employed a balancing analysis in a rights-protecting manner, by acknowledging the “preferred position” of First Amendment rights. 130 Shapiro, supra note 127, at 76–78 (describing balancing as an adjunct of the clear and present danger and preferred position doctrines). Marsh v. Alabama offers a concise statement of how the 1940s Court deployed balancing in the service of civil liberties: “When we balance the Constitutional rights of owners of property against those of the people to enjoy freedom of press and religion, as we must here, we remain mindful of the fact that the latter occupy a preferred position.” 326 U.S. 501, 509 (1946). The author is, of all people, Justice Black, who would later become balancing’s most persistent critic on the Court. For instance, in a case concerning municipal restrictions on the distribution of handbills, Justice Roberts explained:

In every case, therefore, where legislative abridgment of the rights is asserted, the courts should be astute to examine the effect of the challenged legislation. Mere legislative preferences or beliefs respecting matters of public convenience may well support regulation directed at other personal activities, but be insufficient to justify such as diminishes the exercise of rights so vital to the maintenance of democratic institutions. And so, as cases arise, the delicate and difficult task falls upon the courts to weigh the circumstances and to appraise the substantiality of the reasons advanced in support of the regulation of the free enjoyment of the rights. 131Schneider v. State, 308 U.S. 147, 161 (1939).

In the hands of the speech-favoring Justices, the balance was usually—but not always—struck in favor of the rights claimant. During this period, the Court also used narrow tailoring to protect speech rights, especially in the challenges to time, manner, and place restrictions on speech. 132The facts of Martin v. City of Struthers illustrate how narrow tailoring was used. 319 U.S. 141 (1943). At issue was a municipal ordinance that forbade the ringing of doorbells for the purpose of distributing “handbills, circulars or other advertisements.” Id. at 142. The Court concluded that the prohibition violated the First Amendment because the measure restricted far more communication than necessary to combat the ills targeted by the measure. Id. at 146–49.

The tide began to turn toward a less speech-protective jurisprudence starting in the late 1940s, under the leadership of Justice Frankfurter. Successive majorities limited the applicability of the clear and present danger test, watered it down, or adopted a highly deferential balancing test instead. 133 See, e.g., Am. Commc’ns Ass’n v. Douds, 339 U.S. 382 (1950). The decision in Dennis v. United States embodied the new approach. 134341 U.S. 494 (1951). Dennis, the head of the U.S. Communist Party, was convicted under the Smith Act for conspiring to advocate the overthrow of the government by force and violence. 135 Id. at 495–96. In upholding his conviction, a plurality of the Court adopted Learned Hand’s reformulation of the clear and present danger test: “In each case [courts] must ask whether the gravity of the ‘evil,’ discounted by its improbability, justifies such invasion of free speech as is necessary to avoid the danger.” 136 Id. at 510 (alteration in original) (quoting United States v. Dennis, 183 F.2d 201, 212 (2d Cir. 1950)) (internal quotation marks omitted). By eliminating recognition of the privileged position of rights, this formulation invites judges to “balance rights away”: restrictions may be permitted when the gravity of evil is high (as with attempts to overthrow the government), even when the probability of success is low. The repeated use of deferential balancing tests, along the lines of Dennis, to defeat civil liberties claims in the 1950s, gave balancing a bad name among civil libertarians, both on and off the Court. 137 See, e.g., Barenblatt v. United States, 360 U.S. 109, 139–44 (Black, J., dissenting); Laurent B. Frantz, The First Amendment in the Balance, 71 Yale L.J. 1424 (1962); Alexander Meiklejohn, The First Amendment Is an Absolute, 1961 Sup. Ct. Rev. 245. Additionally, the gap between “low-protectionist” and “high-protectionist” Justices widened, as Justices Black and Douglas increasingly endorsed an absolutist view of First Amendment rights. 138The “low-protectionist” and “high-protectionist” terminology is borrowed from Stephen Siegel. See Siegel, supra note 42, at 362.

Only when the Court arrived at a formula that combined balancing with a heightened regard for rights did the doctrine attain a measure of stability. The seeds of this approach were planted in Justice Frankfurter’s concurrence in Sweezy v. New Hampshire, which introduced a compelling state interest test for the first time. 139354 U.S. 234 (1957).Sweezy concerned the contempt conviction of a New Hampshire college professor who refused to answer questions from the State’s attorney general about subversive tendencies in his teaching. 140 Id. at 234–42. The Court decided the case on the narrow ground that the attorney general was acting in excess of his remit from the legislature, 141 Id. at 254–55. but questioned in dicta whether any state interest would justify infringing Sweezy’s First Amendment interest in lecturing free from state interference. 142 Id. at 251. Justice Frankfurter responded to this open question in his concurrence: the state could “intru[de] into this activity of freedom” but only “for reasons that are exigent and obviously compelling.” 143 Id. at 262 (Frankfurter, J., concurring in the judgment). Justice Frankfurter justified this high bar with a peroration to the importance of free academic discourse: “For society’s good,” he wrote, the interchange of ideas connected with the social sciences “must be left as unfettered as possible.” 144 Id.

It is clear from Justice Frankfurter’s discussion that he did not understand the compelling interest standard as imposing a rigid, one-size-fits-all requirement on state action infringing speech, but rather as creating an adaptable test that courts can calibrate to different circumstances. Whether the state’s interest qualifies as “compelling” is not determined in a vacuum but, rather, depends on the weight of the free expression interests at stake. Thus, the State’s reason for requiring Sweezy’s testimony was not compelling because it was insufficient relative to the costs imposed on First Amendment freedoms. The Court held that, “[w]hen weighed against the grave harm resulting from governmental intrusion into the intellectual life of a university, such justification for compelling a witness to discuss the contents of his lecture appears grossly inadequate.” 145 Id. at 261; see also id. at 265 (“But the inviolability of privacy belonging to a citizen’s political loyalties has so overwhelming an importance to the well-being of our kind of society that it cannot be constitutionally encroached upon on the basis of so meagre a countervailing interest of the State as may be argumentatively found in the remote, shadowy threat to the security of New Hampshire allegedly presented in the origins and contributing elements of the Progressive Party and in petitioner’s relations to these.”). In his Sweezy concurrence, Justice Frankfurter did not abandon his customary balancing approach to First Amendment cases. Rather, in light of the importance he placed on academic freedom, 146 See Siegel, supra note 42, at 367. he upgraded to a form of balancing more solicitous of the right at stake: a balancing with bite.

Justice Frankfurter’s “compelling state interest” language was invoked occasionally in the late 1950s, in cases where the Court struck the balance in the government’s favor. 147 See Barenblatt v. United States, 360 U.S. 109, 127 (1959); Uphaus v. Wyman, 360 U.S. 72, 79–81 (1959). But it was Justice Brennan who, in the early 1960s, incorporated the test into a strict scrutiny framework that would become a fixture of First Amendment jurisprudence. 148As Siegel notes, although courts applied strict scrutiny in several First Amendment cases in the 1960s, strict scrutiny only became a cornerstone of First Amendment doctrine in the 1970s, after it had been embraced in equal protection jurisprudence. Justice Brennan fashioned the compelling state interest test into an enduring doctrinal settlement that “mediated the ‘deadlock’ between the Court’s deferential-balancing and absolutist camps.” 149 Siegel, supra note 42, at 375. As Stephen Siegel explains:

On the one hand, as a form of balancing, the compelling interest standard recognized that constitutional rights could be subordinated to governmental needs in a particular case. On the other hand, Justice Brennan recognized that if the compelling standard could be such a rigorous criterion that its application in almost all cases upheld First Amendment claims without the need for additional weighing of interests in the particular case. 150 Id. at 375–76.

As an argumentation framework, strict scrutiny analysis combined a solicitude for rights with balancing in a package that the Court would turn to again and again.

Justice Brennan began to put the pieces together in his dissent to Braunfeld v. Brown, a 1961 free exercise case concerning Sunday closing laws. 151366 U.S. 599, 610 (1961) (Brennan, J., concurring and dissenting). Justice Brennan addressed the question of rights-review methodology head on, stating that “[t]he first question to be resolved . . . concerns the appropriate standard of constitutional adjudication in cases in which a statute is assertedly in conflict with the First Amendment . . . .” 152 Id. at 611. He invoked the clear and present danger test, as framed in West Virginia State Board of Education v. Barnette: “freedoms of speech and of press, of assembly . . . are susceptible of restriction only to prevent grave and immediate danger to interests which the State may lawfully protect.” 153 Id. at 612 (quoting W. Va. State Bd. of Educ. v. Barnette, 319 U.S. 624 (1943)). Justice Brennan then assessed the laws’ interference with free exercise rights: “Their effect is that no one may at one and the same time be an Orthodox Jew and compete effectively with his Sunday-observing fellow tradesmen.” 154 Id. at 613. He next invoked the compelling interest test to move toward balancing, asking, “What, then, is the compelling state interest which impels the Commonwealth of Pennsylvania to impede appellants’ freedom of worship? What overbalancing need is so weighty in the constitutional scale that it justifies this substantial, though indirect, limitation of appellants’ freedom?” 155 Id. at 613–14. The asserted interest is paltry: “It is the mere convenience of having everyone rest on the same day.” 156 Id. at 614. And hence, in Justice Brennan’s view, the plaintiffs should prevail.

Justice Brennan’s important conceptual move in Braunfeld was employing a balancing test (framed as whether the interest is compelling) as a way to give meaning to Barnette’s stringent standard. 157Barnette, 319 U.S. 624. And after 1962, when personnel changes on the Court cemented a majority with strong civil libertarian tendencies, Justice Brennan was increasingly in a position to stamp this perspective onto the Court’s caselaw.

Justice Brennan’s coalitions applied the standard aggressively, and the First Amendment claimant prevailed in each instance. These cases were not always precise or consistent in defining how closely the means chosen by the legislature had to fit the end in question. Finally, in the 1963 free exercise case Sherbert v. Verner, Justice Brennan yoked the compelling state interest standard to narrow tailoring, operationalized as an LRM test. 158374 U.S. 398 (1963). Even if the State had a compelling interest to condition unemployment benefits on a willingness to work on Saturdays, “it would plainly be incumbent upon [South Carolina] to demonstrate that no alternative forms of regulation would [serve such interest] without infringing First Amendment rights.” 159 Id. at 407. The pairing of a compelling state interest test with narrow tailoring stuck. Here was the mature strict scrutiny framework, which would come to be applied across a range of doctrinal areas in the coming years.

Another example illustrates how the Court sometimes employed the compelling state interest standard as a vehicle for balancing interests in First Amendment cases during the 1960s. Williams v. Rhodes concerned a freedom of association challenge to Ohio laws raising barriers for new parties to make it onto presidential election ballots in the state. 160393 U.S. 23 (1968). Ohio advanced a number of state interests served by the statutes, but the Court, in an opinion by Justice Black, found that none of these “justifie[d] imposing such heavy burdens on the right.” 161 Id. at 31. The right to vote was also implicated, but the Court’s analysis of these two rights was the same. Some of the State’s arguments were rejected because the tailoring of the means to the ends was not sufficiently tight. 162For instance, the State argued that promoting a two-party system encouraged compromise and political stability. But the Court noted that the measures did not just promote a two-party system; rather, they gave two particular parties a monopoly over citizens’ votes. Id. at 31–32. The Court also conceded that some of the State’s asserted interests—such as an interest in electing the candidate chosen by a majority of voters—would indeed be served by the laws. 163 Id. at 32. But the burdens on the right were too great to justify the benefit:

[T]o grant the State power to keep all political parties off the ballot until they have enough members to win would stifle the growth of all new parties working to increase their strength from year to year. Considering these Ohio laws in their totality, this interest cannot justify the very severe restrictions on voting and associational rights which Ohio has imposed. 164 Id.

In subsequent years, of course, the strict scrutiny framework migrated into equal protection analysis and became a mainstay of constitutional rights review in that and other doctrinal areas. 165 See, e.g., Shapiro v. Thompson, 394 U.S. 618 (1969) (applying strict scrutiny to strike down a one-year waiting period for welfare-benefit eligibility on equal protection grounds). And it was in the course of this expansion from its First Amendment roots that strict scrutiny gained a reputation for unbending stringency—for being, in Gerald Gunther’s famous phrase, “fatal in fact.” 166Gerald Gunther, The Supreme Court, 1971 Term—Foreword: In Search of Evolving Doctrine on a Changing Court: A Model for a Newer Equal Protection, 86 Harv. L. Rev. 1, 8 (1972). The balancing of interests, undertaken openly in the context of the “compelling state interest” requirement in some early cases of the 1960s, was less apparent in the 1970s. 167 See Rubin, supra note 120, at 4 (“In the courts, strict scrutiny is essentially invoked, not employed. Despite its name—strict ‘scrutiny’—it ordinarily amounts to a finding of invalidity, not a tool of analysis.”). More often, the Court would intensely scrutinize the tailoring of the statute, striking down legislation on the basis of its overinclusiveness or underinclusiveness, so as to avoid having to consider whether the interest at stake qualified as “compelling.” 168 First National Bank of Boston v. Bellotti, 435 U.S. 765 (1978), illustrates the point. In this case, Massachusetts advanced an interest in protecting shareholders to support its ban on corporate contributions to political campaigns. Id. at 766. But the Court found the statute both overinclusive and underinclusive with respect to those interests. Id. at 793. The Court “[a]ssum[ed], arguendo, that protection of shareholders is a ‘compelling’ interest under the circumstances of this case,” and struck down the statute for lack of a “substantially relevant correlation between the governmental interest” and the right to free speech. Id. at 795 (quoting Shelton v. Tucker, 364 U.S. 479, 485 (1960)).

C. Assessment

We have shown that PA has analogues and antecedents in our own constitutional history, both in dormant Commerce Clause and strict scrutiny. Strikingly, the evolution of rights review in the United States parallels developments in Germany. 169These events are described in some detail in our previous article. See Stone Sweet & Mathews, supra note 5, at 97–111. In both countries, judicial efforts to protect individual liberties passed through two distinct phases. 170 Id. In the late nineteenth century in Germany, as in the United States, courts devised techniques to test state encroachments on private freedoms that set standards for both the ends and means of government power. 171 Id. at 100–02. In Germany, the target of judicial scrutiny was administrative action, not legislation, and the courts were specialized administrative tribunals, the first of which was created in 1875. 172 Id. at 100. In the famous Kreuzberg decision of 1882, the Prussian higher administrative court circumscribed the reach of the police power: it only authorized measures to promote public safety—as opposed to, say, aesthetics. 173In Kreuzberg, the court struck down a regulation by the Berlin police that restricted the heights of buildings in order to afford a view of a military monument. Preußisches Oberverwaltungsgericht [PrOVG] [Prussian Higher Administrative Court] July 14, 1882, 9 Entscheidungen des preußischen Oberverwaltungsgerichts [PrOVGE] 353, reprinted in 100 Deutsches Verwaltungsblatt 219 (1985). That same decade, the court also began imposing an LRM test on administrative actions that curtailed private freedoms. 174For instance, in an 1886 case, the court ruled that the police could not require, on public safety grounds, a landowner to remove a post erected at the edge of his property. Rather, all that was necessary to protect the public was requiring the landowner to light the post after dark. As the court explained, “The protection from accidents . . . is indeed the task of the police; this task and the authority finds its limit, however, in that the chosen measures may not extend farther than they must to meet the goal of eliminating the danger.” PrOVG July 3, 1886, 13 PrOVGE 426, 427. While Germany’s courts lacked the capacity for constitutional review of statutes, it is striking that the techniques they derived so closely resemble those applied by the U.S. Supreme Court in cases like Lawton v. Steele and the dormant Commerce Clause cases. 175 See supra Part I.A.

There are similarities between the second phases of each country’s judicial efforts to protect individual liberties as well. Starting in the 1950s, Germany’s Constitutional Court, like the United States Supreme Court, began experimenting with techniques to give heightened protection to rights of constitutional stature. 176 See Stone Sweet & Mathews, supra note 5, at 104–11. An early form of proportionality analysis appeared in a 1958 case, Apothekenurteil, as discussed in Part III below. PA built on the LRM test that had been pioneered decades earlier, but strengthened the protection for rights by adding a balancing component. 177 Id. The balancing step ensured validation of even narrowly tailored measures only if their benefit exceeded their cost in terms of rights infringement. 178 Id. at 107–08. The eighteenth-century natural law scholars who originally conceptualized PA insisted upon this balancing stage, as did constitutional law scholars in the 1950s. 179 Id. at 99, 108–09. Once formalized in subsequent cases, the PA framework quickly became a fixture of German constitutional law, spreading into a number of different rights areas in the 1960s. 180 Id. at 109–10.

These similar stages of development, in the face of the manifold legal and political differences between Germany and the United States, testify to the powerful, functional imperatives that modern courts face to manage tensions between important values. And they suggest that PA offers real advantages to courts in that position. 181Nor should the parallels obscure important differences in the scope of review. Most importantly, in nineteenth-century Germany, judicial scrutiny was only applied against administrative action, not legislation, and in the twentieth century, proportionality persisted as an administrative doctrine even as it gained new life in constitutional law. Id. at 97–111. In both systems, when tasked with protecting rights against interventionist states in a manner suitable for modern liberal polities, courts came up with similar formulas that combine LRM testing and balancing. The attractiveness of proportionality-based doctrinal frameworks is understandable. They enable judges to protect rights at a high level without proclaiming any right to be absolute, in light of all important, contextual factors. Further, they accommodate judges with different views on the merits of particular disputes. PA channels disagreements on the merits into the substantive issues that matter most: whether measures are narrowly tailored, and whether the state’s asserted interest is sufficiently compelling to curtail a right.

The development toward convergence in the 1960s also casts the subsequent divergence between the United States and Germany in another light. Why did strict scrutiny in the United States gain the reputation as a rigid test, and new tiers of analysis evolve to handle different legal questions, 182 See infra Part III. while in Germany, PA came to be accepted as an all-purpose technique for adjudicating constitutional rights? Our response is threefold.

First, when strict scrutiny spread into equal protection, the “suspect class case” quickly became the paradigmatic application for strict scrutiny. 183 See Siegel, supra note 42, at 398–99. As Siegel points out, the Court was slow to apply the compelling state interest test in racial discrimination cases because the narrow tailoring inquiry was sufficient to ferret out illicit motives. But in that context, and particularly in cases involving racial discrimination, the strict scrutiny test assumed a different purpose. The intensive scrutiny on the connection between the means chosen and the state’s asserted interests was intended to “smok[e] out” invidious purposes—to test the state’s good faith. 184 See Fallon, supra note 3, at 1308–11; see also Rubenfeld, supra note 26, at 428. While the narrow tailoring requirement serves this purpose well, this use of the test tended to obscure its suitability for a balancing analysis. Second, the dominance of a liberal majority on the Court during the period in which strict scrutiny emerged helped to cement the idea, on the Court and in the legal academy, that state measures virtually always failed under strict scrutiny. This empirical regularity appears to have shaped the Court’s own conception of how strict scrutiny operates, so that the framework was defined more by an outcome than by the technique. 185As has been noted, however, strict scrutiny is by no means always “fatal in fact.” See Winkler, supra note 44. Winkler’s article does not cover cases before 1990, and no one has yet demonstrated how fatal strict scrutiny in fact was during the 1960s, 1970s, and 1980s. At a deeper level, the predictability of outcomes under strict scrutiny may have promoted a particular—absolutist—conception of rights. The rights-protecting majorities that applied strict scrutiny contained Justices who viewed some rights, at least, as trumps that admitted no infringement. 186As noted above, Justices Black and Douglas, who took this view of rights, originally declined to join opinions applying strict scrutiny, but did in time. See supra text accompanying note 50. Although this view was never endorsed by a majority of the Court, the outcomes of cases employing strict scrutiny appear to have subtly reinforced the underlying conception of absolutists.

Third, the retreat from open balancing in strict scrutiny reflects the Court’s attempt to protect itself from exposure as a law-making body. This motivation is not without its ironies. An open balancing of the relevant interests usurps the role of the legislator far less than the kind of probing underinclusiveness/overinclusiveness analysis that the Court undertook in Bellotti. 187 See supra note 168. And if the alternative to a principled balancing is the elaboration of constitutional rules as the basis for its decision, a court cannot hide from law making. The court that constructs a constitutional code—that derives a rule for every occasion—makes itself not only a lawmaker, but a supreme, constitutional lawmaker.

III. Pathologies of Tiered Review

We now shift focus from the past development of rights-review frameworks to the present state of rights doctrine, focusing on tiered review. The contemporary tiered review regime is the heir to the developments discussed in Part II.B. As the strict scrutiny framework gained its reputation as “fatal in fact,” bifurcated review splintered, and the Court improvised new forms of less stringent review to handle new classes of claims. Canonically, the Supreme Court employs three tiers of review: rational basis review, intermediate scrutiny, and strict scrutiny. But even counting the tiers is no easy matter. Some commentators have identified apparent deviations from these standards as establishing new tiers, such as “minimal scrutiny with bite” for laws directed at “quasi-suspect” classes, 188 See, e.g., Jeffrey M. Shaman, Constitutional Interpretation: Illusion and Reality 81–84 (2001) (discussing how the Court upgraded minimal scrutiny in certain contexts beginning in the 1970s). and others have construed inconsistencies in how the standards are applied as evidence that the whole edifice of tiered review is beginning to crumble. 189 See Calvin Massey, The New Formalism: Requiem for Tiered Scrutiny?, 6 U. Pa. J. Const. L. 945, 946 (2004) (“The flaccidity of the strict scrutiny employed in Grutter, coupled with the robust skepticism of the minimal scrutiny employed in Lawrence, suggests that the neat compartments of tiered scrutiny are beginning to collapse.”). In any case, the set of standards that together make up tiered review is our closest functional analogue to proportionality: a general rubric for reviewing rights claims in multiple substantive areas. But, we argue, there are serious pathologies associated with tiered review, including: (A) judicial abdication, in the form of rational basis review; (B) analytical incompleteness, in the absence of an explicit balancing stage; and (C) doctrinal instability. Compared to PA, each of these pathologies threatens rights protection in fundamental ways.

The discussion brings to light a peculiar and persistent dynamic in the evolution of our doctrinal standards. Elements of proportionality are scattered throughout American rights doctrines; indeed, as we will show, even rational basis review, as conducted until the mid-1950s, had affinities with classic PA. 190 See infra Part III.A. Over time, though, the Court worked to shut down the interest balancing internal to these standards, such that the outcome of rational basis review, and strict scrutiny, became practically a foregone conclusion. 191 See supra note 185. Faced with this rigidity, the Court has brought balancing back in, often in the form of a new ad hoc standard or test.

A. Abdication

Judges who use the proportionality framework typically begin by discussing the nature, scope, and purpose of the pleaded right—first in the abstract, then with reference to the claimant and the conflict between the right and another constitutional value, such as that pleaded by the government. 192 See supra Part I. When balancing under PA, the court typically considers how much of the “essence” of the right is implicated in any dispute: the greater the incursion into the essential core of a right, the weightier the public interest invoked by the government must be to justify the incursion. Thus, a court may recognize that an individual’s wish to consume child pornography falls within the rubric of free expression, while permitting the government to regulate the harms that result from the sexual exploitation of children, including to criminalize the possession of such pornography. 193In a leading child pornography case, R. v. Sharpe, the Canadian Supreme Court held that a provision of the Criminal Code, as applied to Mr. Sharpe, violated his freedom of expression but was justified as a proportional measure designed to protect children from exploitation. [2001] 1 S.C.R. 45. The approach thus contrasts with categorical, rule-based approaches to rights protection that seek to dispense with balancing once the nature and scope of the right have been defined. In this latter approach, a court might decide that Mr. Sharpe’s rights were not abridged, since child pornography is not a protected form of expression, per se. Id.; cf. New York v. Ferber, 458 U.S. 747 (1982) (holding that child pornography is beyond the limits of First Amendment protection). That same court might also consider political protest to be presumptively protected from state regulation. Only a small sliver of the right to free expression is implicated in child pornography, whereas the value of political speech constitutes part of the essential core of that freedom. The multistage proportionality framework facilitates the task of developing such distinctions in a principled, relatively transparent manner.

In contrast, as presently deployed, rational basis review leads American judges to abdicate their duty to protect rights, including property rights, that are expressly provided for by the Constitution. The underlying doctrinal justification for abdication is unclear: either the claimed “right” is, in fact, not a right; or the right is not important enough to deserve the judicial protection afforded to “fundamental” rights; or, through balancing, the Court has determined that the right can never outweigh any reasonable public purpose legitimately pursued by the government. Further, the Supreme Court often portrays and uses rational basis as a kind of deference doctrine—it covers policy domains in which legislators, not courts, are to do the balancing. As regards property rights, the Court traded one rigid position (aggressive protection of property in the Lochner era) for another (abdication in the post-Lochner era). We think it is a bad strategic move for the Court, or any rights-protecting court, to treat a rights provision as either de facto absolute or de facto without force. It is also indefensible, as a formal matter, to excise a right from the Constitution, yet that is what the Court has done. 194We are not the first to take this view. See, e.g., James W. Ely, Jr., The Guardian of Every Other Right: A Constitutional History of Property Rights (1992); Richard A. Epstein, Takings (1985).

A telling example is the well-known 1955 Supreme Court decision Williamson v. Lee Optical of Oklahoma, Inc. 195348 U.S. 483 (1955). The case involved an Oklahoma statute that would have forbade opticians (who grind and duplicate lenses, and fit them to frames and faces) from selling their services and products without the prior authorization of either an ophthalmologist (a medical doctor specializing in eye care) or an optometrist (a licensed professional who diagnoses but does not treat eye disease, and who writes prescriptions for lenses). 196 Id. at 485–86. Among other things, the law would have prohibited opticians from replacing old with new frames, or from duplicating existing lenses to insert into old frames, without the customer first obtaining a prescription from an ophthalmologist or optometrist. 197 Id. at 485. Ophthalmologists and optometrists would also possess the authority to designate which optician a customer could use. 198 Id. at 485 n.1. Days before the law was to enter into force, opticians asked a three-judge federal district court to strike it down on the grounds that the reform would violate property rights guaranteed by the Constitution. 199Lee Optical of Okla., Inc. v. Williamson, 120 F. Supp. 128, 137 (W.D. Okla. 1954), rev’d, 348 U.S. 483 (1955). The plaintiffs also claimed that the statute comprised an illegal delegation, to private parties, of the authority to determine who would do business as a dispensing optician in the state. 200 Id. Oklahoma claimed that the statute was designed to promote better eye care, not least by increasing the frequency of eye exams. 201 Lee Optical, 348 U.S. at 487.

After extensive testimony, the district court found that the effect of the law would be to put many opticians out of business 202 Lee Optical, 120 F. Supp. at 144 n.37 (“The inevitable result of the enforcement of the provisions found by this Court unconstitutional . . . will be to literally put said plaintiffs out of business . . . .”). —a total deprivation of rights—while providing no added benefit to public health. 203 Id. at 135 (“The evidence indicates, almost without variance, that written prescriptions issued by the professional examiner contain no directive data in regard to the manner in which the spectacles are to be fitted to the face of the wearer. In addition, the Court is satisfied that the mere fitting of frames to the face, where the old lenses are available, is in reality only an incident to what fundamentally is a merchandising transaction, that is, the sale of a pair of frames; and, in any event, the knowledge necessary to perform these services is strictly artisan in character and can skillfully and accurately be performed without the professional knowledge and training essential to qualify as a licensed optometrist or ophthalmologist.”). Under PA, such findings can only lead to one result: the invalidation of the law as a disproportionate exercise of legislative power. Judge Wallace, writing for the court, recognized that the dispute was to be resolved under a rationality standard. 204 Id. at 132–33. But he did not regard rational basis as a strict deference standard that would insulate the statute from judicial review of the rights claim. Instead, Judge Wallace understood rational basis to require verification that the state had not abridged the rights of the opticians in an arbitrary 205 Id. at 142 (“The dispensing optician, a merchant in this particular, cannot arbitrarily be divested of a substantial portion of his business upon the pretext that such a deprivation is rationally related to the public health.”). or unreasonable 206 Id. at 137–38 n.21 (“The action by the legislature in the instant act is as unreasonable as if pharmicists [sic] were divested of their right to do business by legislative action which delegated to physicians the absolute control and responsibility for accurate compounding of medical prescriptions for the reason that a physician alone is qualified to prescribe drugs for patients.”). manner:

It is recognized . . . that all legislative enactments are accompanied by a presumption of constitutionality; and, that the court must not by decision invalidate an enactment merely because in the court’s opinion the legislature acted unwisely. Likewise, where the statute touches upon the public health and welfare, the statute cannot be deemed unconstitutional class legislation, even though a specific class of persons or businesses is singled out, where the legislation in its impact is free of caprice and discrimination and is rationally related to the public good. A court only can annul legislative action where it appears certain that the attempted exercise of police power is arbitrary, unreasonable or discriminatory. 207 Id. at 132.

In this passage, Judge Wallace took pains to distinguish his approach from the kind of robotic formalism that had so discredited “reasonableness” review during the so-called Lochner era. His opinion incorporated elements of PA into its analysis. Citing what he considered stable precedent, Judge Wallace concluded that the state’s exercise of police powers would be justified only if the state could show, “[f]irst, that the interests of the public generally, as distinguished from those of a particular class, require such interference; and, second, that the means are reasonably necessary for the accomplishment of the purpose, and not unduly oppressive upon individuals.” 208 Id. at 137 (emphasis added) (quoting Lawton v. Steele, 152 U.S. 133, 137 (1894)). He then disposed of the case in ways that anyone versed in PA would find familiar 209Concurring with this portion of the decision, Judge Murrah’s opinion also has the feel of a standard proportionality analysis:The obvious result of this legislation is to appropriate a property right of one class to follow a legitimate calling or occupation and to give it to another class not shown to be more competent in the public interest. Indeed, the evidence shows that it would be inimical to the public interest to require the wearer of eye glasses to return to the ophthalmologist in order to obtain a prescription for the replacement or duplication of an eye glass, or to place the burden upon the ophthalmologist of merchandising frames and mountings and other appliances, or placing that purely commercial monopoly in the optometrist. There is no rational basis for depriving the optician of the right to perform this purely mechanical and commercial service to the public.Id. at 144 (Murrah, J., dissenting in part) (dissenting on provisions affecting advertising, which we do not discuss here). :

The means chosen by the legislature does not bear “a real and substantial relation” to the end sought, that is, better vision, inasmuch as . . . [ophthalmologists and optometrists] possess no knowledge or skill superior to a qualified practicing optician . . . and in fact [ophthalmologists and optometrists], as a class, are not as well qualified as opticians as a class to either supervise or perform the services here regulated. 210 Id. at 138.

The district court held that three provisions of the Oklahoma statute violated the Due Process Clause of the Fourteenth Amendment, and that one provision violated the Equal Protection Clause of that same Amendment.

It is crucial to stress that the district court did not find elements of PA alien or inimical to rational basis review. As we found with respect to earlier versions of strict scrutiny, 211 See supra Part II. American judges considered necessity analysis and balancing to be inherent parts of the judicial repertoire when they adjudicated rights.

The Supreme Court overturned the district court’s (much more nuanced) decision without so much as a single word concerning the rights being pleaded, and without serious consideration of facts or policy. Instead, Justice Douglas, writing for a unanimous Court, simply abdicated:

The Oklahoma law may exact a needless, wasteful requirement in many cases. But it is for the legislature, not the courts, to balance the advantages and disadvantages of the new requirement. . . . It is enough that there is an evil at hand for correction, and that it might be thought that the particular legislative measure was a rational way to correct it.

The day is gone when this Court uses the Due Process Clause of the Fourteenth Amendment to strike down state laws, regulatory of business and industrial conditions, because they may be unwise, improvident, or out of harmony with a particular school of thought. We emphasize again what Chief Justice Waite said in Munn v. Illinois, “For protection against abuses by legislatures the people must resort to the polls, not to the courts.” 212Williamson v. Lee Optical of Okla., Inc., 348 U.S. 483, 497–88 (1955) (citations omitted) (quoting Munn v. Illinois, 94 U.S. 113, 134 (1877)).

Abdication also entailed an extraordinary act of constitutional law making: the Supreme Court rewrote the U.S. Constitution so as to deprive individuals of the judicial protection of rights expressly guaranteed by that same Constitution. 213Justice Field’s dissent from Munn v. Illinois, which Justice Douglas cites as authority, applies equally well to Lee Optical:The construction actually given by . . . this court makes the provision, in the language of Taney, a protection to “a mere barren and abstract right, without any practical operation upon the business of life,” and renders it “illusive and nugatory, mere words of form, affording no protection and producing no practical result.”94 U.S. 113, 145 (1876) (Field, J., dissenting) (quoting Bronson v. Kinzie, 42 U.S. (1 How.) 311, 318 (1843)). Arguably, this decision is one of the most poorly reasoned by the Court in an important case. Indeed, it contains virtually no rights reasoning whatsoever. 214With respect to the equal protection claim, Justice Douglas had only this to say:The problem of legislative classification is a perennial one, admitting of no doctrinaire definition. Evils in the same field may be of different dimensions and proportions, requiring different remedies. Or so the legislature may think. Or the reform may take one step at a time, addressing itself to the phase of the problem which seems most acute to the legislative mind. The legislature may select one phase of one field and apply a remedy there, neglecting the others. The prohibition of the Equal Protection Clause goes no further than the invidious discrimination. We cannot say that that point has been reached here.Lee Optical, 348 U.S. at 489 (citations omitted).

The Supreme Court’s ruling in Lee Optical and its enduring effects deserve to be considered in light of what the German Federal Constitutional Court (GFCC) was doing at the same time. In Apothekenurteil, 215Bundesverfassungsgericht [BVerfG] [Federal Constitutional Court] June 11, 1958, 7 Entscheidungen des Bundesverfassungsgerichts [BVerfGE] 377 (Ger.). one of its earliest balancing cases, in 1958, a druggist challenged a Bavarian law regulating drugstores on the ground that it violated Article 12(1) of the Basic Law, 216 Article 12(1) states, “All Germans shall have the right freely to choose their occupation or profession, their place of work and their place of training. The practice of an occupation or profession may be regulated by or pursuant to a law.” Grundgesetz für die Bundesrepublik Deutschland [Grundgesetz] [GG] [Basic Law], May 23, 1949, BGBl. I, art. 12(1) (Ger.). which provides for occupational freedom. The law authorized the licensing of new pharmacies only when “in the public interest” and only when new stores would not destabilize the market by threatening the viability of existing pharmacies. 2177 BVerfGE 377 (380). In framing its analysis, the GFCC squarely confronted the tension between individual rights and public goals, which led it to embrace balancing:

The constitutional right should protect the freedom of the individual; the professional regulation should ensure sufficient protection of societal interests. The individual’s claim to freedom has a stronger effect . . . the more his right to free choice of a profession is put into question; the protection of the public becomes more urgent, the greater the disadvantages are, that come from the free practicing of professions. When one seeks to maximize both . . . equally legitimate . . . demands in the most effective way, then the solution can only lie in a careful balancing [Abwägung] of the meaning of the two opposed and perhaps conflicting interests. 2187 BVerfGE 377 (404–05).

In a rich and complex decision, the court then elaborated an early version of PA, tailored to the specifics of Article 12(1), 2197 BVerfGE 377 (396–12). The case was decided before the GFCC had fully developed the proportionality principle. Today there is one proportionality test for all fundamental rights. Stone Sweet & Mathews, supra note 5, at 110–11. and annulled the law on necessity grounds. In a case decided the next year, the German Court dismissed claims under Article 12(1), upholding a law that prohibited businesses other than licensed pharmacies from selling prepackaged, nonprescription drugs at retail. 220BVerfG Jan. 7, 1959, 9 BVerfGE 73 (82). It held that public health would be better served if all drugs were sold by trained professionals, and that pharmacies would be economically imperiled if they could no longer rely on this stream of income. 2219 BVerfGE 73 (76–81).

The German Court did not use these cases to lay a procrustean bed that would then force every set of facts to the same result. On the contrary, it developed a flexible balancing framework capable of adaptation to changing circumstances. While balancing may lead to inconsistent outcomes (a functional economic argument is rejected in the first case but accepted in the second, if on different facts 222 See Paul G. Kauper, The Constitutions of West Germany and the United States: A Comparative Study, 58 Mich. L. Rev. 1091, 1130 (1960). ), the GFCC did not abdicate its constitutional duties. Unlike the U.S. Supreme Court, the GFCC (a new jurisdiction whose political legitimacy remained to be tested) did not tell rights claimants: “go to the polls, not the courts.” Rather, it fully accepted the responsibility that attends the power of judicial review. Taking a wider perspective, the German Court’s commitment to proportionality balancing has helped it avoid the kinds of problems that the U.S. Supreme Court created for itself. In the pre-New Deal era, the American Court committed itself to defending freedom of contract and laissez-faire capitalism against market regulation, as if that defense was one of its central institutional missions; it then abandoned property rights altogether. In contrast, the German Court has consistently held that the Basic Law is “neutral” with respect to the kind of economic system or regulation fashioned by the legislator. 223The leading decision is the Investment Aid I case, BVerfG July 20, 1959, 4 BVerfGE 7. A translation and commentary can be found in Kommers, supra note 21, at 243–45. Yet, that position has never entailed abdication. 224The GFCC seeks to reduce, as much as possible, the tension between the pleaded property right and the public good resulting from a limitation of property rights; that is, the GFCC balances among contending constitutional values in light of the context of the case. In cases like Lee Optical, the U.S. Supreme Court has little use for either the Constitution or context. For a superb comparison of German and American approaches to property rights as constitutional rights, see Gregory S. Alexander, Property as a Fundamental Constitutional Right? The German Example, 88 Cornell L. Rev. 733 (2003). On the contrary, one of the major responsibilities of the GFCC has been to define the reciprocal limits of market regulation and property rights 225 The Basic Law provides for the right to property in Article 14, which qualifies the right to property in a manner that strongly implies the use of PA. GG art. 14 (“(1) Property and the right of inheritance shall be guaranteed. Their content and limits shall be defined by the laws. (2) Property entails obligations. Its use shall also serve the public good. (3) Expropriation shall only be permissible for the public good. It may only be ordered by or pursuant to a law that determines the nature and extent of compensation. Such compensation shall be determined by establishing an equitable balance between the public interest and the interests of those affected. In case of dispute concerning the amount of compensation, recourse may be had to the ordinary courts.”). as German society and the economy have evolved. While doing so, it has sometimes angered the Christian Democratic Right, sometimes the Social Democratic Left, but it has never, over more than five decades, lost its centrality to economic governance. The German Court would have failed, as did the American Supreme Court, had it embraced rigidity and absolutism.

B. Incompleteness

From the perspective of a judge who systematically uses PA to manage rights litigation, the Supreme Court’s use of variegated approaches can at times seem casual or worse—unprincipled. The American judge is much more comfortable omitting analytical steps that the PA judge would view as essential. The PA judge always begins with a general discussion of the right being pleaded as a matter of constitutional structure or theory, and in light of past rulings. Once a right has been so constructed, the judge then turns to PA. If the government measure under review survives necessity analysis, then the balancing-in-the-strict-sense stage will allow the judge to ensure that the law does not infringe too much on the right, given the polity’s constitutional commitments. The judge thus begins with a relatively abstract construction of the right and ends with analysis of the right in a specific conflict with an opposing value. The American judge may leave out one or both of these steps, even in strict scrutiny, as when the judge performs the compelling interest inquiry without serious consideration of the right, or balancing.

A well-known example of egregious incompleteness is United States v. O’Brien. 226391 U.S. 367 (1968). Mr. O’Brien had burned his draft card, in violation of federal statutes, while protesting the Vietnam War and military conscription. 227 See id. at 369–70, 376. Although the case raised important questions about the government’s use of a statutory amendment to stifle protest, 228In 1965, Congress adopted an amendment to the Universal Military Training and Service Act of 1948 in part to deter those who would burn their draft cards in protest of the war. See id. at 370, 385–86. The circuit court, noting that O’Brien’s action was already punishable under a separate regulation, declared the amendment unconstitutional on the grounds that it was expressly designed to limit speech. O’Brien v. United States, 376 F.2d 538, 540–41 (1st Cir. 1967), vacated, 391 U.S. 367. our focus here is on the Supreme Court’s treatment of the right to “wordless” or “symbolic” speech. The Court mentioned but took no firm position on the question of whether O’Brien’s act was to be considered “speech” under the First Amendment—the inquiry was simply left incomplete. 229 See O’Brien, 391 U.S. at 376 (“We cannot accept the view that an apparently limitless variety of conduct can be labeled ‘speech’ whenever the person engaging in the conduct intends thereby to express an idea. However, even on the assumption that the alleged communicative element in O’Brien’s conduct is sufficient to bring into play the First Amendment, it does not necessarily follow that the destruction of a registration certificate is constitutionally protected activity.”). Instead, the Court formulated the following test:

This Court has held that when “speech” and “nonspeech” elements are combined in the same course of conduct, a sufficiently important governmental interest in regulating the nonspeech element can justify incidental limitations on First Amendment freedoms. To characterize the quality of the governmental interest which must appear, the Court has employed a variety of descriptive terms: compelling; substantial; subordinating; paramount; cogent; strong. Whatever imprecision inheres in these terms, we think it clear that a government regulation is sufficiently justified if it is within the constitutional power of the Government; if it furthers an important or substantial governmental interest; if the governmental interest is unrelated to the suppression of free expression; and if the incidental restriction on alleged First Amendment freedoms is no greater than is essential to the furtherance of that interest. 230 Id. at 376–77 (footnotes omitted).

The Court did not distinguish a “compelling interest” from a “substantial interest”; on the contrary, it was implied that the “descriptive terms” listed were broadly synonymous. On the basis of this test, the Court upheld O’Brien’s conviction, since the draft card fulfilled a variety of administrative functions, such as communicating the rights and duties of the bearer. 231 Id. at 378–80, 382.

The judgment concluded in deafening silence on the question of whether the government’s interest did or did not outweigh O’Brien’s “alleged” speech right—the answer to which one might have learned if the framework contained a balancing-in-the-strict-sense phase. The Court might also have balanced the contending values in its discussion of the nature of the government’s “important or significant” interest; instead it marched through the test while saying nothing of any importance about speech. To resolve such a case under a standard version of PA, the Court would have been required to consider the nature and scope of the right at play upfront, and then to ensure that the right did not get lost in the analysis of the government’s position at the end.

O’Brien may illustrate something else about American rights review. As we mentioned, O’Brien did not subscribe to the literal strict scrutiny formula of “compelling” state interest; a “substantial” or “important” government interest would do. 232 See id. at 366–67. This position is subject to different interpretations. Some might view O’Brien as a strict scrutiny case in all but name; others might reject this conclusion, especially in light of the result—the government wins. 233 See id. at 386. But if O’Brien is not a strict scrutiny case, what is it?

O’Brien may well be a forerunner of the intermediate scrutiny tests that would proliferate starting in the 1970s. 234Kathleen Sullivan, for one, views the O’Brien test as a version of intermediate scrutiny. Sullivan, supra note 8, at 297. Like the standard in O’Brien, the canonical intermediate scrutiny test also requires an “important” state interest (although it does not impose a narrow tailoring test). 235 For example, see Craig v. Boren, 429 U.S. 190, 197 (1976), discussed in Part III.C.1 infra. Instead of narrow tailoring, the test required a “substantial” relation between the policy and the government’s objectives, again echoing O’Brien’s language. See id. To the extent that we regard the O’Brien standard as an early intermediate scrutiny test, the case follows another pattern. As with the strict scrutiny framework, we see the Court undertaking a doctrinal innovation with respect to the vanguard of civil liberties—First Amendment rights—which in short order spreads to other constitutional rights. We discuss the creation of intermediate forms of scrutiny immediately below.

C. Instability

The third pathology is related to the first two. In the classic “bifurcated review” regime, judges are asked to sort cases into two bins that represent extremes of stringency and deference. 236 See White, supra note 119, at 301 (describing the Court’s “bifurcated review project”). The inadequacy of these options generates a persistent instability in the doctrinal structure. Time and again, the Court has introduced an intermediate standard of review to govern its inquiry in different areas of law. 237Also, on occasion the Court has claimed to be applying one of the two canonical standards, but transparently not done so. See, e.g., City of Cleburne v. Cleburne Living Ctr., Inc., 473 U.S. 432 (1985) (recognizing that a law discriminating against mentally retarded individuals does not merit heightened scrutiny because mental retardation is not a quasi-suspect class, but finding that the law fails to rationally relate to any of its various justifications, suggesting that the Court was actually applying a higher standard than rational basis review). Forms of intermediate review represent efforts to make a space for balancing in the context of rights review. As such, they are also symptomatic of the dysfunctionality of classic two-tiered review.

But the ad hoc introduction of new standards of review does not “solve the problem” of two-tiered review; rather, it creates new problems. For all its faults, the two-tiered framework of review was at least rooted in a coherent distinction between economic liberties and “preferred” fundamental rights. 238 See Gillman, supra note 82, at 623–26; White, supra note 119, at 309. Adding new tiers registers as a deviation from this baseline and may come across as an unprincipled or illegitimate move. The dissenting opinions in the cases that first adopt intermediate scrutiny reliably paint them in this light. 239 See, e.g., Plyler v. Doe, 457 U.S. 202, 244 (1982) (Burger, C.J., dissenting) (criticizing the majority’s application of intermediate scrutiny to a statute that deprives illegal alien children of the right to public education); Craig, 429 U.S. at 220–21 (Rehnquist, J., dissenting) (criticizing the majority’s application of intermediate scrutiny to a law that imposes different drinking ages for men and women).

More generally, defections from bifurcated review undermine the stability of rights adjudication. Because we have no meta-theory of rights that maps certain classes of claims into standards of review, confusion and contention over the proper standard are common, particularly when new rights claims arise. The unpredictability and instability of tiered review are partly to blame for the fact that so much of the analytic firepower in our constitutional caselaw is aimed at the standard of review, rather than the substance of the claim. 240 See White, supra note 43, at 82–83.

We illustrate these points by looking at cases where courts elaborate new standards of review in an attempt to introduce some balancing into the analysis. We first discuss two developments well-known to students of constitutional law: the Supreme Court’s changing approach to sex discrimination under the Equal Protection Clause, and Planned Parenthood of Southeastern Pennsylvania v. Casey’s revisions to the Roe v. Wade framework for abortion rights. Our treatment of these exhaustively studied developments is brisk and focuses on the shifts in doctrine. We also consider in some detail a lesser known federal appeals court case, involving a constitutional challenge to a local juvenile curfew. This case shows a court trying to fit a new classification into the existing tiers, and the difficulties and conflicts that ensue. 241Another line of prominent cases in which the Court shifts the standard of review, with dramatic results, includes Metro Broadcasting, Inc. v. FCC, 497 U.S. 547 (1990) (applying intermediate scrutiny to affirmative action policies in broadcasting), and Adarand Constructors, Inc. v. Peña, 515 U.S. 200 (1995) (applying strict scrutiny to affirmative action policy in federal contracting program). We pair familiar and unfamiliar cases deliberately, to underscore the pervasiveness of the difficulties with tiered review. 242Tiered review is also a pervasive feature of state constitutional doctrine, which space prevents us from discussing in this Article. Notably, states have also been at the forefront of experimenting with rights review, and some have opted out of tiered review of their constitutional claims. See Goldberg, supra note 10, at 522 n.154.

1. Sex Equality and Equal Protection

The equal protection jurisprudence of sex classifications has famously followed a meandering course. The Supreme Court originally subjected sex classifications to rational basis review, then increased the scrutiny without formally changing the standard, even flirting with full suspect classification status before devising a new standard of intermediate scrutiny. Then, in time, the intermediate scrutiny standard drifted upward to become more stringent. From a comparative perspective, one cause of this well-known evolution seems clear. The inherited two-tier framework offered the Court no opportunity to measure the harms of gender distinctions against their benefits, and the Court has struggled ever since to find a workable formula for doing so.

In the mid-twentieth century, the Supreme Court subjected sex classifications to rational basis review and often upheld them. 243 See, e.g., Hoyt v. Florida, 368 U.S. 57 (1961) (upholding a Florida law that disallowed women from being selected for jury service unless they volunteered); Goesaert v. Cleary, 335 U.S. 464 (1948) (upholding a Michigan law that banned certain women from serving as bartenders). As traditional attitudes about gender roles began to fade, it became apparent that rational basis review provided women with no meaningful guarantee to equal protection of the laws. 244In Goesaert, for instance, the Court held that Michigan had a rational basis for banning women, except the wives or daughters of male bar owners, from tending bars. 335 U.S. at 466–67. In Reed v. Reed, Chief Justice Burger, writing for a unanimous Court, struck down an Idaho law giving preference to men in estate administration, ostensibly on a rational basis standard. 245404 U.S. 71, 76–77 (1971). But the opinion lacked the characteristic deference to legislative judgments that is associated with rational basis review. The state’s justification for the preference was to avoid controversy when more than one person applied to administer the will. 246 Id. at 76. The Court dismissed the rationale summarily, criticizing the policy as “the very kind of arbitrary legislative choice forbidden by the Equal Protection Clause of the Fourteenth Amendment.” 247 Id. The Court added, without further explanation, that “whatever may be said as to the positive values of avoiding intrafamily controversy, the choice in this context may not lawfully be mandated solely on the basis of sex.” 248 Id. at 76–77. Even as Reed pledged allegiance to rational basis review, it marked a sharp sub silentio departure from it.

The Reed approach did not long endure, however. Two years later, in Frontiero v. Richardson, a plurality of the Court pushed to shift gender from rational basis review into strict scrutiny by claiming gender as a suspect classification. 249411 U.S. 677, 682 (1973). The dispute in Frontiero concerned a federal policy that made it easier for male military personnel than for female military personnel to claim their spouses as dependents for benefits purposes. 250 Id. at 678–79. The appropriate standard of review for gender-based restrictions became the focal point of the case and fractured the Court.

The proposition that sex classifications should be subjected to strict scrutiny found only four votes: Justice Brennan was joined by Justices Douglas, White, and Marshall. 251 Id. at 678. Justice Powell, joined by Justice Blackmun and Chief Justice Burger, wrote an opinion that reached the same result on the authority of Reed without elevating sex classifications to suspect status. 252 Id. at 691–92 (Powell, J., concurring in the judgment). The concurrence declined to decide whether sex was a suspect class because the statute failed even under the lesser measure of Reed-style rational basis review. 253 See id. at 692. The Powell concurrence also argued that for the Court to decide the issue, it would effectively preempt democratic deliberations on precisely this issue, as the Equal Rights Amendment was then under consideration, and would have the effect of making sex a suspect classification. Id. Justice Stewart concurred separately, and Justice Rehnquist dissented, adopting the opinion of the district court. Id. at 691.

The status of sex classifications in equal protection analysis reached an equilibrium of sorts in 1976 with Craig v. Boren. 254429 U.S. 190 (1976).Reed and Frontiero had demonstrated that no majority of the Court was satisfied with analyzing sex classifications under rational basis review, and no majority was willing to apply strict scrutiny. In Craig, the Court resolved this impasse by creating a new option: an intermediate standard of review for classifications relating to sex.

The case concerned an Oklahoma statute prohibiting the sale of low-alcohol beer to women younger than age eighteen, and to men younger than age twenty-one. 255 Id. at 191–92. Writing for a bare majority, Justice Brennan articulated a new standard of review for gender distinctions as a distillation of recent practice. Citing to Reed, Frontiero, and other cases, he wrote: “To withstand constitutional challenge, previous cases establish that classifications by gender must serve important governmental objectives and must be substantially related to achievement of those objectives.” 256 Id. at 197. The government objective asserted here was public safety, and the state had sought to demonstrate the statute’s fit with statistical evidence showing, most pertinently, that men ages eighteen to twenty were much more likely to be arrested for drunk driving than women of the same age. 257 Id. at 199–200. The Court concluded that this evidence was insufficient to establish the requisite substantial relation: only 2% of the men in this cohort were found to drive drunk, and penalizing them all was unjustified. 258 Id. at 202.

Justice Brennan’s intermediate standard in Craig won the support of a majority on the Court—and was applied in subsequent cases—but fewer than half of the Justices on the Craig Court endorsed the standard without reservation. Justices Powell and Stevens both joined the opinion but wrote separately to express misgivings over the introduction of a new, intermediate standard for sex-based classifications. Justice Powell candidly discussed the problems with two-tiered analysis and Reed’s divergence from traditional rational basis review, but he preferred proceeding under Reed’s nominal rational basis review to “a further subdividing of equal protection analysis.” 259 Id. at 211 n.* (Powell, J., concurring). Justice Stevens, for his part, challenged the value of the tiered analysis altogether. He began:

There is only one Equal Protection Clause. It requires every State to govern impartially. It does not direct the courts to apply one standard of review in some cases and a different standard in other cases. Whatever criticism may be leveled at a judicial opinion implying that there are at least three such standards applies with the same force to a double standard. 260 Id. at 211–12 (Stevens, J., concurring).

In his own analysis of the statute, Justice Stevens emphasized the poorness of fit between the statute and the alleged purpose in his decision to strike it down. 261 Id. at 213–14.

Chief Justice Burger and Justice Rehnquist dissented. Justice Rehnquist in particular visited scorn on the creation of a new standard, emphasizing the arbitrariness and novelty of Justice Brennan’s formulation:

The Court’s conclusion that a law which treats males less favorably than females “must serve important governmental objectives and must be substantially related to achievement of those objectives” apparently comes out of thin air. The Equal Protection Clause contains no such language, and none of our previous cases adopt that standard. I would think we have had enough difficulty with the two standards of review which our cases have recognized—the norm of “rational basis,” and the “compelling state interest” required where a “suspect classification” is involved—so as to counsel weightily against the insertion of still another “standard” between those two. 262 Id. at 220–21 (Rehnquist, J., dissenting).

Justice Rehnquist also faulted the content of this standard—in particular, the qualifiers “important” and “substantial”—as not lending themselves to perspicuous judicial application. 263 Id. at 221. The dissent further faulted the majority for rejecting the statistical evidence proffered by the state, which showed higher drinking while driving among young men than women, in support of the policy. 264 Id. at 225–26.

In the years after Craig v. Boren, the Court accepted and applied an intermediate scrutiny standard to sex classifications in a number of cases. 265 See, e.g., Califano v. Westcott, 443 U.S. 76 (1979) (applying intermediate scrutiny to unemployment benefits provided by the Social Security Act to a father, but not a mother); Orr v. Orr, 440 U.S. 268 (1979) (applying intermediate scrutiny to Alabama statute under which husbands, but not wives, may be required to make alimony payments); Califano v. Webster, 430 U.S. 313 (1977) (applying intermediate scrutiny to benefits provided by the Social Security Act that favor female wage owners over similarly situated male wage owners). But even as the standard seemed settled, the formula for review proved less than completely stable. Starting with the 1979 case Personnel Administrator v. Feeney, 266442 U.S. 256 (1979). a number of decisions emphasized the stringency of the Craig v. Boren test by noting that it requires “an exceedingly persuasive justification” from the government. 267 See, e.g., Miss. Univ. for Women v. Hogan, 458 U.S. 718, 724 (1982); Kirchberg v. Feenstra, 450 U.S. 455, 461 (1981). In striking down Virginia Military Institute’s (VMI) exclusion of women in 1996, Justice Ginsburg for the Court interpreted “an exceedingly persuasive justification” not merely as a characterization of the standard, but as the standard itself. 268United States v. Virginia, 518 U.S. 515, 531 (1996). The groundwork for this move was laid by Hogan and J. E. B. v. Alabama ex rel. T. B., 511 U.S. 127 (1994). In these cases, the Court also characterized the defendant’s burden as providing “an exceedingly persuasive justification” for the discrimination. But, in the Court’s analysis, the way to meet this burden was, in essence, to show a substantial relation to an important governmental interest: the test derived from Craig. The Court found fault with Virginia’s proffered justification—that the exclusion of women was necessary to achieving VMI’s mission—because some women who would want to attend VMI could meet its standards. 269 Virginia, 518 U.S. at 550. The ratcheting up of intermediate scrutiny in the majority opinion did not escape the notice or criticism of Chief Justice Rehnquist, concurring, or Justice Scalia in dissent. Chief Justice Rehnquist faulted the new language as having less “content and specificity” than the older formula 270 See id. at 559 (Rehnquist, C.J., concurring in the judgment). —which he had previously criticized for its own opacity. 271Craig v. Boren, 429 U.S. 190, 219–21 (1976) (Rehnquist, J., dissenting). For his part, Justice Scalia excoriated the majority at length for its application of intermediate scrutiny in a manner that, in his view, amounted to strict scrutiny. 272 Virginia, 518 U.S. at 570–76 (Scalia, J., dissenting) (“Only the amorphous ‘exceedingly persuasive justification’ phrase, and not the standard elaboration of intermediate scrutiny, can be made to yield this conclusion that VMI’s single-sex composition is unconstitutional because there exist several women (or, one would have to conclude under the Court’s reasoning, a single woman) willing and able to undertake VMI’s program. Intermediate scrutiny has never required a least-restrictive-means analysis, but only a ‘substantial relation’ between the classification and the state interests that it serves.”).

As Kathleen Sullivan has written, “Intermediate scrutiny, unlike the poles of the two-tier system, is an overtly balancing mode.” 273Sullivan, supra note 8, at 297 (internal quotation marks omitted). The Supreme Court’s development of an intermediate scrutiny standard reflects how ill-equipped the bifurcated review regime is for evaluating the range of gender-based distinctions in public law. But creating this new standard hardly fixed the problems of tiered review. First, the new standard is ad hoc and improvised, and as such, lacks roots in constitutional theory. 274As Justice Rehnquist pointedly noted in dissent, the Craig test—that measures must be substantially related to the achievement of important government objectives—seems to “come[] out of thin air.” Craig, 429 U.S. at 220 (Rehnquist, J., dissenting). And second, introducing another tier of analysis may resolve a case, but it does not necessarily offer a workable solution to future cases with different questions and contexts. The post-Craig drift in the intermediate scrutiny standard shows that the formula is unstable and, if anything, has grown more contentious over time.

2. Undue Burdens and Abortion Rights

The “undue burden” standard in abortion cases is another doctrinal device introduced to allow some latitude for balancing in a field of law previously governed by a rigid tiered review standard. In Roe v. Wade, the Court held that restrictions on abortion before the point of viability would be subject to strict scrutiny. 275410 U.S. 113, 155 (1973). Given the Court’s rightward shift over the following two decades, Casey, which came before the Court in 1992, was widely expected to overturn Roe. 276 See David Margolick, Seeking Strength in Independence, Abortion-Rights Unit Quits A.C.L.U., N.Y. Times, May 21, 1992, at A20 (“It is in [Casey], argued last month before the Supreme Court, that abortion-rights advocates expect the Court to either overturn or to greatly weaken Roe v. Wade.”). The Court’s use of strict scrutiny in abortion cases had been under attack on the Court since 1989, with a plurality of Justices urging that restrictions on abortions should be subject only to rational basis review. 277Webster v. Reprod. Health Servs., 492 U.S. 490, 517–19 (1989) (plurality opinion). In fact, Casey’s lead plurality opinion, authored jointly by Justices O’Connor, Kennedy, and Souter, preserved the right to terminate pregnancies before the point of viability, but offered an undue burden standard to replace strict scrutiny as the framework for evaluating restrictions on that right. 278Planned Parenthood of Se. Pa. v. Casey, 505 U.S. 833, 877 (1992). The plurality explained that a state places an undue burden on the right when “a state regulation has the purpose or effect of placing a substantial obstacle in the path of a woman seeking an abortion of a nonviable fetus.” 279 Id. The plurality opinion conceded that Justices might agree on the standard and yet disagree on its application in a concrete case: “Even when jurists reason from shared premises, some disagreement is inevitable. That is to be expected in the application of any legal standard which must accommodate life’s complexity. We do not expect it to be otherwise with respect to the undue burden standard.” 280 Id. at 878 (citation omitted).

Although the Casey dissenters alleged that the undue burden standard was an unprecedented and unprincipled invention, 281 See id. at 964 (Rehnquist, C.J., concurring in the judgment in part and dissenting in part) (describing the undue burden standard as “created largely out of whole cloth by the authors of the joint opinion”). judges had previously undertaken undue burden analysis when asked to balance competing interests, in the abortion context and elsewhere. 282Abortion rights opinions that discuss some form of undue burden analysis include City of Akron v. Akron Center for Reproductive Health, Inc., 462 U.S. 416, 463 (1983) (O’Connor, J., dissenting); Harris v. McRae, 448 U.S. 297, 314 (1980); Bellotti v. Baird, 443 U.S. 622, 640 (1979); Maher v. Roe, 432 U.S. 464, 473 (1977); Bellotti v. Baird, 428 U.S. 132, 147 (1976). For more on the use of undue burden standards in and out of the abortion rights context, see Alan Brownstein, How Rights Are Infringed: The Role of Undue Burden Analysis in Constitutional Doctrine, 45 Hastings L.J. 867 (1994); Valerie J. Pacer, Note, Salvaging the Undue Burden Standard—Is It a Lost Cause? The Undue Burden Standard and Fundamental Rights Analysis, 73 Wash. U. L.Q. 295 (1995). One author notes that 256 Supreme Court decisions issued between 1945 and 1993 contain the phrase undue burden or the equivalent. Curtis E. Harris, An Undue Burden: Balancing in an Age of Relativism, 18 Okla. City U. L. Rev. 363, 423 (1993). Notably, the undue burden language mirrors almost exactly the “unreasonable burden” standard applied in early dormant Commerce Clause cases. 283 See supra Part II.A. In our view, the proliferation of undue burden standards in U.S. constitutional law is not accidental. This standard, like the move to intermediate scrutiny, is a means to allow courts to consider the interests on both sides of a constitutional controversy. The introduction of such devices in domains previously governed by the bifurcated review regime is symptomatic of the mismatch between the rigidity of the doctrinal structures a court has inherited and the demands placed on constitutional judges in complex disputes raising multiple important values.

3. Juvenile Curfew and Equal Protection

The discussion above illustrates how intermediate scrutiny came into being, as a result of the struggle to situate the review of sex classifications within equal protection doctrine. The cases below, which concern youth curfews, illustrate how the introduction of this new tier has added a measure of flexibility, but also a dose of confusion, to equal protection analysis when issues of first impression arise. Intermediate scrutiny makes it easier for courts to strike an appropriate balance between the poles of strict scrutiny and rational basis deference. But because intermediate scrutiny is not so much a theoretically justified standard as a “Goldilocks” solution—a happy midpoint between these extremes—courts seeking to sort new legal issues into the tiers have little guidance.

A number of appeals courts have encountered challenges to local youth curfews over the past two decades, and the results—both as to standard of review and as to outcome—have varied wildly. The Second Circuit held that a youth curfew should be examined under intermediate scrutiny and struck it down. 284Ramos v. Town of Vernon, 353 F.3d 171, 176, 186 (2d Cir. 2003). The D.C. Circuit and Fourth Circuit each applied intermediate scrutiny to a challenged curfew but went on to uphold it. 285Hutchins v. District of Columbia, 188 F.3d 531, 541, 548 (D.C. Cir. 1999) (en banc); Schleifer v. City of Charlottesville, 159 F.3d 843, 847, 855 (4th Cir. 1998). The Fifth Circuit determined that strict scrutiny was the appropriate standard and sustained a curfew under that standard. 286Qutb v. Strauss, 11 F.3d 488, 492, 496 (5th Cir. 1993). The Ninth Circuit also applied strict scrutiny and struck down a curfew. 287Nunez v. City of San Diego, 114 F.3d 935, 946, 949 (9th Cir. 1997). Of course, nothing is unusual about circuit splits and differences in approach, and courts always frame issues based on how litigants frame their claims. But we contend that the tiered review framework generates an extra measure of uncertainty and disagreement. (We note also that all of the cases mentioned above included dissents—an unusual feature in appeals court decisions and another measure of the discord that these cases produce.) We illustrate how the framework fails to provide adequate guidance by looking more closely at one case, the most recent, from the Second Circuit.

Youth curfews present a double challenge to equal protection analysis. They seem to implicate both the fundamental rights and the suspect classification prongs of equal protection doctrine, but a categorical approach to either prong would not subject curfews to any real scrutiny. The mobility right at issue is an important one, but not one conventionally included in the canon of “fundamental” rights. And while juveniles are a discrete, vulnerable group, age is not considered a suspect classification. Courts that believe youth curfews merit more than minimal scrutiny will therefore find themselves pushing against the categorical approach to equal protection analysis.

Ramos ex rel. Ramos v. Town of Vernon concerned a challenge to a nighttime curfew, established in 1994 by local ordinance, for persons under age eighteen. Angel Ramos, a minor resident of Vernon, Connecticut, and his parents sued the town in 1998, claiming, among other things, that the curfew impermissibly burdened the fundamental rights of minors. 288 See Ramos ex rel. Ramos v. Town of Vernon, 48 F. Supp. 2d 176, 188 (D. Conn. 1999), rev’d, 353 F.3d 171 (2d Cir. 2003). Ramos also raised vagueness, First Amendment overbreadth, and Fourth Amendment claims against the statute, as well as claims arising under the Connecticut constitution. Id. at 180–81. His parents also alleged infringement of their right as parents to direct the raising of their children. Id. at 181. The district court rejected all of the federal claims and certified the state law claims to the Connecticut Supreme Court, id. at 188, which rejected them. See Ramos v. Town of Vernon, 761 A.2d 705, 710 (Conn. 2000).

The district court’s analysis of this claim was something of a jumble. First, the court seemed to rule out a higher measure of scrutiny for this ordinance based on its disparate treatment of juveniles, reasoning that age is not a suspect classification. 289 Ramos, 48 F. Supp. 2d at 184. Then the court moved to take up the plaintiffs’ claim that the ordinance violated a fundamental right to travel. But without pausing to consider whether this ordinance implicated any such right, the court immediately circled back to the relevance of the minor plaintiff’s age to his constitutional claim. The court observed that “[t]he Supreme Court has never clearly indicated the appropriate level of scrutiny to apply to legislation that affects minors,” but noted reasons given in the Bellotti plurality decision for treating children’s constitutional rights differently than those of adults. 290 Id. In the next sentence the court concluded, “After careful consideration of the Bellotti factors and other Supreme Court jurisprudence involving minors,” 291 Id. the youth curfew should be subject to “less than the strictest level of scrutiny,” that is, intermediate scrutiny. 292 Id. (quoting Schleifer v. City of Charlottesville, 159 F.3d 843, 847 (4th Cir. 1998)) (internal quotation mark omitted). The court did not share its “careful consideration” of these issues with the reader, and the choice of intermediate scrutiny was presented as a fait accompli.

With the standard selected, the court’s analysis of the equal protection claim was brisk. Because the plaintiffs did not deny the importance of the town’s stated interests—protecting the community from crime, and the safety and welfare of minors—the analysis turned on the issue of fit: Were the ordinance’s means substantially related to these goals? The court concluded that they were. The court noted that crime in Vernon had dropped since enactment of the curfew, and while declining to assert a causal connection, the court deferred to the legislative judgment that the ordinance was effective. 293 Id. at 185–86. The court also noted that the ordinance’s limited hours were not unduly burdensome and that the numerous exceptions allowed most unobjectionable activities to continue after dark. 294 Id.

The plaintiffs appealed to the Second Circuit Court of Appeals, which reversed the district court, finding an equal protection violation. 295 Ramos ex rel. Ramos v. Town of Vernon, 353 F.3d 171, 187 (2d Cir. 2003). The bulk of the Second Circuit’s analysis concerned selecting the proper tier for evaluating the curfew. Although the appeals court’s inquiry was far more expansive than the district court’s, it too came up short, as the court arrived at the intermediate standard through a process of elimination and provided few affirmative reasons for why intermediate scrutiny was appropriate.

The court began its analysis by declaring that the curfew implicated the fundamental right to intrastate travel, a right previously recognized in the Second Circuit. 296Id. at 176. The court then asked whether this curfew, which would receive strict scrutiny if applied to adults, should be reviewed under a less stringent standard because it applies only to juveniles. 297 Id. The court noted that other circuits had reviewed curfews under each of the tiers and proceeded to consider the different approaches. 298 Id.

First the court considered, and rejected, rational basis review. The court found no justification for concluding that juveniles are excluded from the right to intrastate travel, although greater limits on their rights than adults’ rights may be permissible, owing to the vulnerabilities and other characteristics of minors. Since “denying the existence of a constitutional right is too blunt an instrument to resolve the question of juvenile rights to freedom of movement,” the court would “prefer to admit minors to the protected zone and then engage in a balancing of constitutional rights and children’s vulnerabilities.” 299 Id. at 178.

The court then considered, and rejected, strict scrutiny. “Strict scrutiny,” the court argued, “embodies a constitutional preference for ‘blindness,’” and should be deployed only where blindness as to classifications is desired. 300 Id. at 179. In the equal protection context, some rights—the fundamental rights—are so important that they should almost always be available without reference to any group classifications, and some classifications—the suspect classifications—are in themselves so pernicious that the state should almost never use them. But, the court noted, “youth-blindness is not a goal in the allocation of constitutional rights.” 301 Id. Hence, strict scrutiny was not appropriate. The court concluded, almost by default, that intermediate scrutiny was the appropriate standard: “Hence, strict scrutiny would appear to be too restrictive a test to address government actions that implicate children’s constitutional rights. Consequently, we choose the second of the three approaches described above and apply intermediate scrutiny.” 302 Id. at 180 (citation omitted). The court went on to argue that intermediate scrutiny is flexible enough to allow legislation properly drafted to address the needs and vulnerability of children, but probing enough to disallow legislation justified by facile and untested generalizations about the young. Id. at 181.

A couple of features are worth noting about the court’s selection of a standard of review. First, the analysis implicitly treats intermediate review as a residual category, to be adopted when neither of the traditional standards of review is suited to the issue at hand. This is consistent with the genesis of intermediate review, but hardly adequate as a justification for choosing a standard of review. Nowhere did the court build up an affirmative case for intermediate scrutiny as the appropriate standard of review (although the court did note that intermediate review has some advantages). Second, in choosing a standard, the court slipped from the fundamental rights prong of equal protection analysis into a kind of suspect class view. According to the court’s analysis, strict scrutiny is appropriate where “blindness” is the goal. 303 Id. at 179 (internal quotation marks omitted). But the court treated this case as belonging to the fundamental rights prong of equal protection analysis, not the suspect classification prong. 304 See id. at 181 n.4 (noting that the Supreme Court has not generally considered age to be a suspect classification). As the court presented the issue, the relevant question to determine if strict scrutiny was appropriate is whether freedom of movement is so important a right that it should be afforded to individuals without regard to group classifications. 305 See id. at 179–81. Instead, however, the court asked whether blindness to age is a goal in the allocation of constitutional rights. 306 Id.

Having settled on intermediate scrutiny, the court then proceeded to evaluate the curfew. The court identified two of the interests invoked by Vernon—protecting minors from harm and protecting the community from youth crime at night—as important. 307 Id. at 182. The question became whether the ordinance has a substantial relation to these goals. The court found it did not. 308 Id. at 185–87. Although the ordinance was prompted by a concern about crime, the town provided no good reasons for its particular features: for the hours it was in effect, or for its application to minors. Nor was there any proof that the curfew was responsible for a drop in crime. 309 Id. at 186. Therefore, the court concluded, the curfew violated equal protection. 310 Id.

In a wide-ranging dissent, Judge Ralph Winter took issue with, among other things, the court’s selection of intermediate scrutiny. Judge Winter emphasized the well-established legal recognition of children’s custodial control by their parents, and denied that minors had a right to travel independent of parental control. 311 Id. at 188 (Winter, J., dissenting). Hence, the curfew should be subjected only to rational basis review, and upheld. 312 Id. at 190–91. The dissent also considered, and rejected, the claim that the curfew burdened parental or guardian rights. Id. at 190.

4. Evaluation

The examples above demonstrate the persistent instability of tiered review. Some critics—including the authors of some of the dissents—might counter that this “instability” is really just a lack of judicial discipline. Nothing forces judges to adopt new standards such as intermediate scrutiny or the undue burden test; when judges do so, they are recklessly pushing a political agenda without respect for precedent. We note first that this charge cannot be leveled at the curfew ruling: when courts face questions of first impression, existing doctrine offers little or no guidance as to what standard should govern. 313 Cf. District of Columbia v. Heller, 554 U.S. 570, 628–29 (2008) (evaluating the firearms restriction without selecting a standard of review). And even if the pathologies of bifurcated review did not produce a Casey or a Craig in any deterministic sense, these rulings and many others are best understood as responses to the shortcomings of tiered review.

Consider the sex discrimination cases again. The two-tiered framework put judges in an impossible position: choosing between these dichotomous standards of review meant committing either to uphold, or to strike down, virtually every gender distinction in law. In this situation, the incentive to find, or create, some middle ground was powerful. That middle ground took the form of intermediate scrutiny. And the improvisatory character of intermediate scrutiny opened the door for its future evolution. If intermediate scrutiny had started life as a distinct standard of review with a deep grounding in constitutional theory rather than as a workaround to the difficulties of bifurcated review, the VMI majority might have found it more difficult to change the operative test. 314 See supra text accompanying notes 268–70. Of course, the members of the Court are responsible for the particular ways in which sex discrimination caselaw has developed. But the broadest outlines of this development are influenced by the doctrinal structures that confront the Justices and constrain their moves.

PA offers a more stable, principled alternative. As a unified framework of analysis that allows a court to tailor the stringency of review to the particulars of each claim, PA sidesteps the conflict and confusion surrounding the choice of a standard of review. Nor is this approach a far-fetched departure for American law. As we have noted, Justices Marshall and Stevens have proposed something similar for equal protection claims. Justice Marshall has argued that “the level of scrutiny employed in an equal protection case should vary with ‘the constitutional and societal importance of the interest adversely affected and the recognized invidiousness of the basis upon which the particular classification is drawn.’” 315City of Cleburne v. Cleburne Living Ctr., Inc., 473 U.S. 432, 460 (1985) (Marshall, J., concurring in the judgment in part and dissenting in part) (quoting San Antonio Indep. Sch. Dist. v. Rodriguez, 411 U.S. 1, 99 (1973) (Marshall, J., dissenting)). Adopting such an approach would not eliminate disagreement among judges, but it would refocus the disagreement onto the substantive issues of importance.

The jurisprudence of the GFCC gives a glimpse into what a proportionality-based approach to an equality right looks like. 316A good overview of the doctrine in English can be found in Susanne Baer, Equality: The Jurisprudence of the German Constitutional Court, 5 Colum. J. Eur. L. 249 (1999). A more detailed treatment, albeit in German, can be found in a constitutional commentary, such as Hans D. Jarass & Bodo Pieroth, Grundgesetz für die Bundesrepublik Deutschland 90–137 (2006). Naturally, important differences exist between the Basic Law’s Article 3, which guarantees equality before the law, and the Constitution’s Fourteenth Amendment. 317One notable difference is that Article 3 spells out what we would call “suspect classifications,” on the basis of which discrimination is not allowed. GG art. 3(3) (“No person shall be favoured or disfavoured because of sex, parentage, race, language, homeland and origin, faith, or religious or political opinions. No person shall be disfavoured because of disability.”); see also id. art. 3(2) (specifically addressing sex equality). And of course, adopting PA does not mean reaching the same outcomes as the German court—indeed, the Supreme Court could use PA and still replicate most of the judgments that they have arrived at by other means, and are worth preserving. Still, the GFCC’s approach to Article 3 shows how proportionality can provide a stable argumentation framework for dealing with the most difficult legal issues. 318PA is not used across the board in the GFCC’s Article 3 jurisprudence, but increasingly it is supplanting older approaches. See Baer, supra note 316, at 260. We focus our discussion on the “new” proportionality-based approach to the general equality guarantee in Article 3(1).

Since the 1980s, the GFCC has construed the general equality right in ways that are tailor-made for PA. The right amounts to a guarantee that any unequal treatment under the law be justified by a “sufficiently weighty reason.” 319BVerfG Apr. 28, 1999, 100 BVerfGE 138 (174); see also BVerfG May 30, 1990, 82 BVerfGE 126. What counts as a sufficient reason depends on the particulars of the claim. In effect, the more a distinction in the law threatens the core values of equality, the greater the state’s burden of justification. 320Jarass & Pieroth, supra note 316, at 100; Baer, supra note 316, at 262. For instance, when a challenged measure cuts across the exercise of a fundamental right, or when groups are disadvantaged based on personal characteristics rather than voluntary behavior, the Court insists on “a strict application of the proportionality requirement.” 321BVerfG Mar. 2, 1999, 99 BVerfGE 367 (388). On the other hand, when a measure implicates none of these factors, or falls within the particular competence of the legislature, the Court operates with a healthy degree of deference to distinctions drawn by the legislature. 322 Jarass & Pieroth, supra note 316, at 99, 100. The legislature has substantial latitude, for instance, in choosing the subjects and rates of taxation. Id. at 109. Proportionality is no silver bullet, and Germany’s equality jurisprudence has not been free of controversy. 323For a (now somewhat dated) overview of prominent critiques and controversies, see Albert Bleckmann, Die Struktur des Allgemeinen Gleichheitssatzes 1–2 (1995). But the basic framework, grounded in PA, has been stable and widely accepted.

With respect to abortion politics, which have been ferocious in Germany, 324 Alec Stone Sweet, Governing with Judges: Constitutional Politics in Europe 109–14 (2000). the GFCC’s jurisprudence has had a legitimizing and stabilizing effect. The GFCC has issued two landmark rulings on abortion, roughly contemporaneous with Roe and Casey. 325BVerfG May 28, 1993, 88 BVerfGE 203 (“Abortion II”); BVerfG Feb. 25, 1975, 39 BVerfGE 1 (“Abortion I”). These rulings established that, given Germany’s responsibility for the Holocaust and other crimes against humanity prior to and during World War II, the criminal law must always discourage abortion as a taking of life, but that abortion would nonetheless be permitted under certain specified conditions.

In its 1975 decision, and in contrast to the U.S. Supreme Court’s approach, the GFCC held that the right to life covers the fetus. 326A careful comparison of the two approaches can be found in Richard E. Levy & Alexander Somek, Paradoxical Parallels in the American and German Abortion Decisions, 9 Tul. J. Int’l & Comp. L. 109 (2001). An abridged translation of the two key German decisions is contained in Kommers, supra note 21, at 336–56. That finding comprised the critical starting point from which the balancing would proceed, given that the constitutional value represented by the life of the unborn child will, at times, conflict with the constitutional rights of the woman carrying the child, her human dignity, bodily integrity, and the free development of her personality. 32788 BVerfGE 203 (253). Summarizing a rich and sophisticated decision, the court held that to permit abortion without restriction, and without the express disapprobation of the state, would infringe too much on the right to life, and would therefore be unconstitutional; 32839 BVerfGE 1 (44–46). it also insisted, however, that the right to life did not extinguish the rights of pregnant women. 329 Id. at 47. Indeed, the court ruled, the latter’s rights must prevail in special circumstances. Although abortion could not be decriminalized per se, the procedure could nonetheless go unpunished after mandatory counseling and when justified by genetic, medical, or social hardship reasons, or when the pregnancy resulted from rape. 330Id. at 48–50.

In 1993, the GFCC upheld the basics of its earlier ruling, while creating an “unreasonable burden” (Unzumutbarkeit) test. The test is rooted in the logic of proportionality: abortions would be tolerated by the criminal law if carrying a fetus to term would impose on a pregnant woman “heavy and unusual burdens” that go well beyond “reasonable sacrifice.” 33188 BVerfGE 203 (255); 39 BVerfGE 1 (48). That is, “another interest”—of bodily integrity and self-determination—“equally worthy of constitutional protection asserts itself with such urgency that the state’s legal order cannot require the pregnant woman always to defer to the rights of the unborn.” 33239 BVerfGE 1 (50) (quoted in Kommers, supra note 21, at 341). Although scholars have debated whether the Unzumutbarkeit standard is identical to, or partially distinct from, proportionality, 333 See Rüdiger Konradin Albrecht, Zumutbarkeit als Verfassungsmassstab (1995) (distinguishing Unzumutbarkeit from proportionality). “the practical effect of the Unzumutbarkeit principle is to balance the unborn child’s right to life against the mother’s interest in bodily integrity and personal autonomy.” 334Levy & Somek, supra note 326, at 122.

Comparing the abortion jurisprudence of the U.S. Supreme Court and the GFCC, and its broader reception, reveals some striking commonalities and differences. The baseline constitutional rules in the two countries appear as opposites: the right to access abortion services prohibits a ban on abortions in the United States, while the unborn child’s right to life requires restrictions on abortion in Germany. But as both courts have balanced, they have presided over a convergence in abortion law. 335 Id. at 111 (“In both countries, abortion is generally available early in the pregnancy after various informational, counseling and waiting restrictions; and is available later in the pregnancy to protect the life or health of the mother, as well as in cases of serious fetal deformity, rape, or incest.”). Since Roe, however, abortion has become far and away the most controversial issue the Supreme Court touches, dominating the discourse during the confirmation of Justices and creating a social movement that challenges the political legitimacy of the Court itself. 336 See Robert Post & Reva Siegel, Roe Rage: Democratic Constitutionalism and Backlash, 42 Harv. C.R.-C.L. L. Rev. 373 (2007). The GFCC’s rulings on abortion were also politically controversial, but not because they were understood to be inconsistent, unprincipled, and inherently political. 337The difference in tone extends to the bench as well. The German abortion rulings produced dissents, but none that challenged the legitimacy of the majority’s ruling, or the good faith of the justices, like Justice Scalia’s blistering dissent in Casey. See Planned Parenthood of Se. Pa. v. Casey, 505 U.S. 833, 982–1002 (1992) (Scalia, J., dissenting).

We do not claim that proportionality is the key variable that explains this latter difference. But it plays a crucial role in political acceptance of the GFCC’s involvement in regulating abortion. In both abortion cases discussed above, the GFCC anchored its analysis in balancing, whose legitimacy had been built over previous decades. This technique allowed the GFCC to give due consideration to the competing interests at stake, and required that the judges seek to reduce harms as much as possible. Both German cases generated hard-hitting dissents. But no dissenter questioned the appropriateness of balancing; rather, they challenged the relative weights given to the values in tension by the majority when it balanced.

IV. All Things in Proportion?

A. Bringing Proportionality Back In

We have shown that the tiered review regime chronically generates pathologies that have weakened rights protection and undermined the coherence of rights jurisprudence in the United States. 338 See supra Part III. PA, while not a cure-all for the challenges facing constitutional courts, avoids these pathologies by providing a relatively systematic, transparent, and trans-substantive analytical procedure for the adjudication of virtually all rights claims. We also found that all three levels of review—rational basis, intermediate review, and strict scrutiny—have, at various points in their evolution, contained core elements of proportionality. In our view, this finding bolsters our claim that no rights-protecting court can, in practice, do without these elements unless it is willing to commit itself to a posture of abdication or absolutism.

Together, these findings lead to the conclusion that embracing proportionality may well enhance the consistency and transparency of U.S. constitutional rights adjudication in a manner that keeps faith with our constitutional traditions and commitments. No single, simple formula predicts how U.S. courts might go about amplifying and regularizing the elements of proportionality that already exist, sometimes partially buried, in our doctrines. 339As we noted before, a number of areas of American constitutional law have interest-balancing tests that resemble PA. Justice Breyer lists some of these in his Heller dissent. District of Columbia v. Heller, 554 U.S. 570, 690 (2008) (Breyer, J., dissenting); see also, e.g., Nixon v. Shrink Mo. Gov’t PAC, 528 U.S. 377, 403 (2000) (citing examples where the Court has taken such an approach); Thompson v. W. States Med. Ctr., 535 U.S. 357, 388 (2002) (Breyer, J., dissenting) (commercial speech); Burdick v. Takushi, 504 U.S. 428, 433 (1992) (election regulation); Mathews v. Eldridge, 424 U.S. 319, 339–49 (1976) (procedural due process); Pickering v. Bd. of Ed. of Twp. High Sch. Dist. 205, 391 U.S. 563, 568 (1968) (government employee speech). The Supreme Court has employed balancing tests outside of the constitutional context as well. Perhaps most famously, the “rule of reason” used to evaluate combinations and contracts in restraint of trade under section 1 of the Sherman Act is a balancing test: the court asks whether net effect of a restriction on trade is pro-competitive. See Bd. of Trade v. United States, 246 U.S. 231 (1918). Indeed, in recent years federal appeals courts have supplemented the Board of Trade balancing with an LRM test, so that the total effect closely resembles PA. See Gabriel A. Feldman, The Misuse of the Less Restrictive Alternative Inquiry in Rule of Reason Analysis, 58 Am. U. L. Rev. 561, 580 (2009). Election law, too, has turned with increasing frequency to a standard that balances the state’s interest against the burden on an individual’s right to vote in a manner resembling proportionality. See Joseph Fishkin, Equal Citizenship and the Individual Right to Vote, 86 Ind. L.J. (forthcoming 2011). While a detailed roadmap for this process is beyond the scope of this Article, we do suggest, in broadest strokes, some possibilities. We hope these remarks might spark a broader discussion of what the reconciliation of U.S. rights review and proportionality might look like. We conclude by noting and responding to objections to the arguments we have presented in this Article.

As a starting point, we note again that Justices Marshall and Stevens have proposed, in concurrences and dissents, approaches to equal protection analysis that would displace tiered review in favor of a “sliding scale” review that more nearly approximates proportionality. 340 See James E. Fleming, “There Is Only One Equal Protection Clause”: An Appreciation of Justice Stevens’s Equal Protection Jurisprudence, 74 Fordham L. Rev. 2301, 2305–08 (2006) (describing the positions of both Justices). Justice Marshall explained his alternative to the Court’s “rigidified approach to equal protection analysis” in San Antonio Independent School District v. Rodriguez: “As the nexus between the specific constitutional guarantee and the nonconstitutional interest draws closer, the nonconstitutional interest becomes more fundamental and the degree of judicial scrutiny applied when the interest is infringed on a discriminatory basis must be adjusted accordingly.” 341411 U.S. 1, 102–03 (1973) (Marshall, J., dissenting). For his part, Justice Stevens proposed a single standard, under which the state’s justificatory burden in establishing a “rational basis” for legislation is pegged to the classification employed—it is easier to provide a legitimate, neutral reason for treating people differently on the basis of profession, say, than on the basis of race. 342City of Cleburne v. Cleburne Living Ctr., Inc., 473 U.S. 432, 452–53 (1985) (Stevens, J., concurring).

Although the two standards are different, 343 See Goldberg, supra note 10, at 519–24. either could serve as a foothold for a PA-style, interest-balancing approach in equal protection analysis. Approaching equal protection in the spirit of Justice Marshall or Justice Stevens does not require the Court to reject existing doctrinal standards; rather, it provides a way to make sense of precedent by locating doctrine in a more expansive and coherent framework. For instance, the sliding scale perspective unlocks a reappraisal of the compelling state interest test: what qualifies as a “compelling” interest for overcoming a rights claim will vary based on the intensity of the right infringement. This perspective thus helps to recover the original role of strict scrutiny analysis as a form of rights-protective interest balancing. In this view, the other standards of review represent points along a continuous spectrum of scrutiny, within a common master methodology. 344 Cf. Craig v. Boren, 429 U.S. 190, 212 (1976) (Stevens, J., concurring) (arguing that cases in different tiers “actually apply a single standard in a reasonably consistent fashion”). Over time, the Court could elaborate these doctrinal structures into a more articulated framework where means–ends testing and interest balancing are handled in distinct phases. And this proportionality standard could be applied, not only in equal protection but wherever tiered review now is operative.

Pushing further, we see no reason to limit the adoption of PA to areas now governed by the three-tier regime. Justice Breyer, joined in his Heller dissent by Justices Stevens, Souter, and Ginsburg, forcefully argued that the American approach to rights adjudication has always included elements of proportionality and balancing. 345District of Columbia v. Heller, 554 U.S. 570, 693–719 (2008) (Breyer, J., dissenting). In a celebrated 2010 commencement address at Harvard, retiring-Justice Souter emphasized the inescapable conflicts between values involved in constitutional litigation. Justice David H. Souter, Commencement Address at Harvard University (May 27, 2010), available at http://news.harvard.edu/gazette/story/2010/05/text-of-justice-david-souters-speech/. He then went on to demonstrate how a version of PA should be applied to the Second Amendment. 346 Heller, 554 U.S. at 693–719. In his dissent, Justice Breyer regarded his approach—correctly—not as a great departure from the Court’s practice, but as a rationalization of techniques already in place, and of a piece with the standard of review employed in different areas. Id. The Second Amendment identifies a value to be maximized, subject to certain limitations. There is no question—and the majority did not suggest—that the right to bear arms implies that no restrictions (for instance, on any type of weapon, in any circumstance) are permitted. Once the Court determines that the Second Amendment covers an individual’s right to bear arms outside of a militia context, the crucial questions must concern limits, which PA is designed to address. In his dissent, Justice Breyer applied a version of PA to show that reasoning about limits could proceed in a principled, open way. Certainly PA would have provided a surer guide to future cases in this area than the majority’s approach. In declining even to pick a standard of review under which to evaluate state limitation of the right to bear arms, Justice Scalia provided no guidance for future cases.

But a case like Heller is a rarity in that it presents a question of first impression about the existence of a right under one of the Constitution’s first ten amendments. The twenty-first-century Court almost never writes on a blank slate. Although many other areas of American constitutional law could benefit from the introduction of the proportionality principle, we also recognize that good reasons may exist for the Court to treat some rights—such as the Third Amendment’s prohibition against quartering—as clear, per se rules. 347 U.S. Const. amend. III. Still, the Court has already determined, and consistently held, that most of the rights that are of importance to Americans are relative, not absolute. Given that the Court already subjects important rights—including those expressed in absolute terms (e.g., the First Amendment)—to a litany of limitations and qualifications, courts would be justified in asking the following two questions. First, why should every constitutional right not be protected by a necessity standard, that is, why should government be allowed to infringe more on any right than is necessary to achieve a declared public purpose? Second, do laws that have passed the narrow tailoring requirement nonetheless infringe more on the right than is tolerable given our constitutional commitments? In asking these questions, judges would begin the work of reconceptualizing rights doctrine in terms of PA. 348 Cf. Sullivan & Frase, supra note 53, at 173 (proposing that proportionality “be adopted as a general standard of review”). Failure to consider these questions will, in most cases, weaken rights protection.

That PA can fit with existing constitutional doctrine could be illustrated with reference to examples drawn from every area of the Court’s rights jurisprudence. Here we will provide one more, concerning procedural due process claims arising from reputational injuries caused by state action. Such claims are presently governed by Paul v. Davis. 349424 U.S. 693 (1976).Paul concerned a challenge to a flyer distributed by local police that included the claimant’s name and photograph in a list of “Active Shoplifters” on the basis of an arrest for shoplifting that was never prosecuted. 350 Id. at 695.Paul established that merely showing that stigmatizing state action abridges liberty interests protected by the due process guarantee is not enough to establish a due process violation. Rather, something else, some change in legal status, is also required, though the court did not explain what should happen once this “stigma plus” threshold is satisfied. 351 Id. at 701. Not surprisingly, the Paul decision has been roundly criticized for failing to set out standards for how courts should evaluate claims to procedural due process violations emerging from reputational injuries. 352 See Barbara E. Armacost, Race and Reputation: The Real Legacy of Paul v. Davis, 85 Va. L. Rev. 569, 571 (1999) (“Scholars have been relentlessly and uniformly negative in their reactions to the Supreme Court’s opinion and holding in Paul . . . .”); id. at 575–84 (collecting criticisms of Paul).

A PA approach would provide a neat solution to the problem. Once a triggering stigma is identified, the Court would move into proportionality mode, asking what state interest is at stake, and whether the stigma-causing means are suitable for achieving it. The Court would engage the LRM inquiry, and if that is satisfied, it would move to the final stage of analysis, where it asks whether the harm to the claimant is justified, all things considered. The final balancing stage could effectively do the work now done—poorly—by the nebulous “plus” factor, the work of filtering out de minimis injuries. This use of PA would also resonate with the Court’s balancing approach in Mathews v. Eldridge 353424 U.S. 319 (1976). and its associated line of due process cases. For our purposes, most important is that the basic structure of the right—the conditions that have to be satisfied to state a claim—can be established through the relevant precedents, and PA can solve the balancing problems that remain.

A necessary task of rights review is defining the scope of rights, that is, determining the range of conduct that is protected by the right. For instance, do the protections of the First Amendment extend to obscenity? 354 See, e.g., Miller v. California, 413 U.S. 15, 36 (1973) (affirming that obscene material is not protected by the First Amendment). PA is, of course, only an analytical procedure; it is otherwise without content. Everywhere PA flourishes, however, judges use it to structure a theory of the rights they adjudicate. As Mattias Kumm observes, a legal system’s approaches to the questions of a right’s scope and a right’s limitation are logically linked. 355 See Mattias Kumm, Who Is Afraid of the Total Constitution? Constitutional Rights as Principles and the Constitutionalization of Private Law, 7 German L.J. 341, 347 (2006). So it is in the United States as well. Strict scrutiny, for example, represents an extremely stringent approach to limiting rights: a measure infringing a right is permissible only if narrowly tailored to a compelling interest. The stakes of acknowledging a prima facie rights claim under strict scrutiny are huge, since the test for infringements stacks the deck in favor of the right. It makes sense, then, that the Supreme Court would be reluctant to hold, for instance, that the full measure of First Amendment protection extends to commercial speech when doing so would lock the Court into an analysis that steamrolls over limits on, say, liquor advertisements, without regard to relevant differences between these and forms of expression nearer to the essential core of the First Amendment. 356Cent. Hudson Gas & Elec. Corp. v. Pub. Serv. Comm’n, 447 U.S. 557, 566 (1980). A stingy approach to the limitation of rights goes hand in hand with a stingy approach to the scope of rights: when the decision that a right is in play is so consequential, the borders of that right’s protection must be carefully policed.

PA offers a different model. As a balancing framework, PA expressly allows for the possibility that prima facie rights claims may be outweighed by competing values, as long as the other steps of the proportionality analysis are satisfied. Because much less hinges on the prima facie determination that any right can be properly pleaded, judges can afford to define the scope of that right more broadly. An example can help illustrate how the adoption of PA would impact U.S. rights review. Free speech jurisprudence recognizes several carve-outs for expressive activities that merit less than the full measure of First Amendment protection, including fighting words, 357 See Chaplinsky v. New Hampshire, 315 U.S. 568 (1942). obscenity, 358 See Roth v. United States, 354 U.S. 476 (1957). child pornography, 359 See New York v. Ferber, 458 U.S. 747 (1982). commercial speech, 360 See 44 Liquormart, Inc. v. Rhode Island, 517 U.S. 484 (1996). and advocacy of imminent lawless action. 361 See Brandenburg v. Ohio, 395 U.S. 444 (1969). In a PA context, a court could arrive at these same outcomes through an explicit, rather than disguised, balancing process. In effect, the court would recognize that, in fact, some expressive interests are at stake in these forms of conduct, but to a lesser degree than in the heartland of First Amendment speech. Not only does PA all but require the court to construct a theory of speech rights, it also allows the court to sidestep some of the intractable border wars that the American penchant for rules and exceptions inevitably generates. We see few, if any, advantages to an approach that seeks to determine, once and for all, on which side of a line a particular case falls, or where to draw the lines separating rules from exceptions in the first place. In any event, the American version of this approach has never actually succeeded in banishing balancing from the judicial tool kit. Pushed out the front door, balancing comes in through the back, where it is used to create ever more nuanced rules and exceptions. 362 See Aleinikoff, supra note 1, at 963–71. In our view, the PA approach is less arbitrary than a binary “yes/no” response to the complexity of rights adjudication. It builds flexibility and a concern for context into rights review, and it lessens the likelihood that judges will paint themselves into doctrinal corners.

Finally, applying PA does not prevent courts from incorporating other modes of rights reasoning in their analysis. For instance, a court that uses PA may also categorically reject certain grounds of justification for state action. 363 See Joseph Raz, The Morality of Freedom 38–69 (1986). Courts may exclude reasons in this way in order to provide added protection for important constitutional values. 364 See Richard H. Pildes, Avoiding Balancing: The Role of Exclusionary Reasons in Constitutional Law, 45 Hastings L.J. 711, 713 (1994). Pildes demonstrates that defining excluded reasons was a dominant technique for adjudicating rights claims in the constitutional jurisprudence of the nineteenth century, id. at 712–13, and Mattias Kumm, supra note 34, at 144–45, emphasizes its role in contemporary constitutional analysis. Suppose, for instance, that the government seeks to ban speech promoting certain social or political theories, on the grounds that they are incorrect and harmful. 365 See Kumm, supra note 34, at 144–45. Under existing First Amendment doctrine, courts will not credit this rationale for government action. Courts reject the government’s position not because the theories are not false or harmful—they could be—but because our theory of freedom of expression excludes reasons for state action based on viewpoint discrimination. 366 Id. at 145. Similarly, a PA court sharing the same First Amendment values may likewise decline to credit the government’s rationale, or give it any weight in the analysis. In Mattias Kumm’s words, “Freedom of speech is not balanced against the harm done by proposing false ideas.” 367Id. Indeed, Kumm correctly notes that PA may exclude reasons at two stages of the analysis: in assessing the legitimacy of the state’s aim, and in the balancing phase. 368 Id. Thus, it is wrong to assume that a move to PA would mean abandoning the delicate constitutional architecture that American courts have constructed to give meaning to rights provisions.

B. Objections

We have tried to make the best case for adopting PA in the United States, but we recognize that there are countervailing considerations. Here, we briefly discuss some of the important institutional and historical reasons why PA might appear to be a problematic fit for the American courts.

Formal differences between the American system of judicial review and other contemporary systems of constitutional justice around the world are obvious and important. Most powerful supreme and constitutional courts in the world today understand their central mission to be the robust protection of fundamental rights, not least since their respective constitutions prioritize rights protection. 369 See Víctor Ferreres Comella, Constitutional Courts and Democratic Values 36–54 (2009). Modern constitutions typically announce rights before state structures are constituted; most rights are qualified by limitation clauses; and the rights-protecting court is expressly designated as the authoritative interpreter of the higher law. Further, the European constitutional courts—and the supreme courts of Canada, India, and South Africa to different degrees—are not limited to “case or controversy” jurisdiction. Cases come to them in diverse ways, often in the form of abstract constitutional questions about the meaning, scope, and application of rights. 370 Id. at 66–67; Stone Sweet, supra note 324, at 343–46. Constitutional judges have an obligation to answer these questions, no matter how controversial. Indeed, giving constitutional answers to controversial, deeply “political” questions about rights is basic to their job description. 371 See Comella, supra note 369, at 72–73. In short, all modern, rights-protecting courts perform an oracular function, a byproduct of which is law making (constitutional, legislative, administrative, and so on). 372In such systems, the classic distinctions made between the “judicial function” (dispute resolution under the law) and the “legislative function” (the elaboration of law) break down. In these systems, traditional notions of separation of powers become weak sources of systemic legitimacy. Instead, the legitimacy of the constitutional order is critically linked to the constitutional court’s capacities to defend rights.

The American Supreme Court is not a specialized constitutional court. In contrast to the modern constitutional court, the U.S. Constitution does not confer on the Court the power of constitutional judicial review. Added as supplementary “amendments,” rights only surface after the organs of government are established. Moreover, the American Constitution does not give interpretive primacy to the judiciary; in fact, perhaps the three branches, being coequal, have the same claim to competence and the same obligation to interpret and apply rights faithfully. 373Prominent contemporary defenders of the “departmentalist” position—that neither the Founders nor the American Constitution meant to grant the judiciary supremacy with respect to constitutional interpretation—include Akhil Amar and Larry Kramer. See Akhil Reed Amar, America’s Constitution (2005); Larry D. Kramer, The People Themselves: Popular Constitutionalism and Judicial Review (2004). In Marbury v. Madison, of course, the Court would derive its constitutional judicial review authority directly from its Article III “case or controversy” jurisdiction. 3745 U.S. (1 Cranch) 137 (1803). Taking a highly formalist view, all judicial law making in the United States (constitutional, legislative, and so on) can be understood as a byproduct of the basic judicial (dispute resolution) function. Of course, to claim that the Supreme Court remains principally a case or controversy court would do violence to reality, downplaying the Court’s oracular, law-making function. Although few serious observers would still make that claim, the Article III case or controversy limitation no doubt comprises an important part of the Supreme Court’s historical identity as a judicial organ, constraining the Court in meaningful ways.

These organic differences must impact how rights are understood and how rights claims are adjudicated. Across Europe, and in Canada and South Africa, constitutional judges face no serious or sustained challenge to their legitimacy when they protect rights, at least as a formal matter. Indeed, constitutional judges would seriously undermine their political legitimacy if they were to abdicate their rights-protecting role, say, by adopting a rational basis standard as the norm for protecting rights. In the United States, the specter of original sin (Marbury) 375Id. and its evil potentials (Lochner), 376Lochner v. New York, 198 U.S. 45 (1905). the defensive discourse of the counter-majoritarian difficulty, and the anxiety over what has become de facto judicial supremacy, has meant that any move by the courts away from deference and toward robust rights protection needs special justification. 377The most influential statements of the problem and (partial) justification of judicial supremacy remain, respectively, Alexander M. Bickel, The Least Dangerous Branch (1962), and John Hart Ely, Democracy and Distrust (1980). American judges and legal academics have engaged in a seemingly endless production and critique of these justifications, but reached little consensus. At the same time, counterpressure flows from the fact that many of the most important American rights are expressed in absolute terms, as rules to be enforced, which makes adopting a blanket rational basis posture deeply problematic, if not indefensible.

As this discussion implies, a legal system’s rights doctrines constitute, and will then embody, notions of what rights are in that system. Kumm explains that such notions tell us “what . . . you have in virtue of having a right.” 378Kumm, supra note 34, at 131. Although the topic raises jurisprudential issues far beyond the scope of this Article, it should be obvious that important structural differences will still distinguish doctrines that reference a conception of rights as “trumps” (in Dworkin’s sense), those that conceive of rights as defensive “shields” (insulating the sphere of the private from the reach of state action), or as “optimization requirements” (in Alexy’s sense). 379 For a careful discussion of this issue, see Mattias Kumm, The Idea of Socratic Contestation and the Right to Justification: The Point of Rights-Based Proportionality Review, 4 Law & Ethics Hum. Rts. 140, 144–52 (2010). The Supreme Court, of course, has not settled on any dominant “theory” of “what rights are,” which partly accounts for the systemic incoherence of its rights jurisprudence.

Europeans were forced to rethink and reconstruct their constitutional law after the horrors of the Holocaust and the destruction of World War II. 380Jed Rubenfeld argues that, due to this fact, the American and European conceptions of constitutionalism, rights, and democracy are fundamentally at odds with one another. Jed Rubenfeld, The Two World Orders, Wilson Q., Autumn 2003, at 22. The new Federal Republic of Germany firmly committed itself to protecting fundamental rights at the highest possible level, while the prestige of political parties and legislative authority was relatively low. The collapse of fascist-authoritarian regimes in Southern Europe in the 1970s, and then across Central and Eastern Europe and the Balkans in the 1990s, reproduced that situation in key respects, and the German approach to constitutionalism was copied and extended. During these latter episodes, those who drafted new constitutions saw no contradiction between democracy and rights protection. On the contrary, a robust system of rights protection was viewed as a precondition for democratic rule. Today, even after the consolidation of stable party systems, European citizenries continue to support constitutional courts—which they equate with rights protection—far more than they support legislatures. Where an ideology of fundamental rights has congealed as a kind of civic religion, rights jurisdiction may ground the legitimacy of constitutional review. Tellingly, this civic religion of rights now also permeates common law jurisdictions, including our neighbor to the north, Canada, which (like Israel) adopted PA from the Europeans.

In the United States, some judges and academics have always portrayed PA and balancing as antithetical to American notions of popular sovereignty and, therefore, to democracy itself. We reject this portrayal, not least, as unsupported by the facts. In the post-New Deal period, when the U.S. Supreme Court moved to protect fundamental rights more robustly, it too began to look like a modern, rights-protecting constitutional court and less like a case or controversy court. The Court relaxed standing doctrines, and abstract review—in the form of facial challenges and related constitutional remedies—emerged and became routine in some areas; the Court frankly assumed a more oracular role as a constitutional lawmaker. 381 See Martin Shapiro & Alec Stone Sweet, On Law, Politics, and Judicialization 347–75 (2002). One might say that, for the first time, a system of constitutional justice began to appear in America, with rights protection at its core. With the important exception of property rights, the Court reserved deference doctrines, such as rational basis, for only those rights judged not to be “fundamental.” Further, it began to theorize the rights being favored more explicitly, which went hand in hand with the development of doctrinal tests that one finds central to PA.

Conclusion

In this Article, we have sought to describe and assess the evolution of American rights doctrine, not just on its own terms, but with reference to constitutional law and practice elsewhere. Our first motivation was, in fact, comparative and empirical. In a first stage of this project, we developed a theoretical explanation of why rights-protecting judges would be attracted to PA: it provides the best-possible response to the challenges of adjudicating qualified rights, and it offers a (partial) solution to certain intractable legitimacy dilemmas generated by judicial law making and supremacy. 382 See Stone Sweet & Mathews, supra note 5, at 80–97. In a second stage, we tracked the emergence and global diffusion of PA. Today, all of the world’s most powerful constitutional courts have adopted a version of PA, which they deploy as an overarching analytical framework for adjudicating rights. Arguably, PA now constitutes the defining doctrinal core of a global, rights-based constitutionalism. Here, we have sought to bring these considerations to bear on the American case, not least because the United States is often characterized (especially by non-Americans) as an outlier, an island unto itself, a legal system that refuses to participate in the transnational conversation about rights adjudication that has exploded into prominence in recent years. While we would agree that the American constitutionalism is ill-equipped either to engage in constitutional dialogues across borders or to exercise positive influence on the evolution of global constitutionalism, it is not true that the American system has rejected either balancing or proportionality. Instead, as we have shown, elements of PA have deep roots in American constitutional law, and American courts, try as they sometimes do, have never been able to dispense with balancing when they adjudicate rights.

Our second motivation was normative. We have argued that the American courts would benefit, on balance, from standardizing rights doctrine under a version of PA. To take just one example, we have identified three endemic pathologies of American rights jurisprudence and shown the various ways in which PA would enable the courts to mitigate these pathologies, or to eliminate them altogether. Our normative claims are informed by both our theoretical priors and our comparative findings. We stressed that one of the virtues of PA is that it allows a court to maximize its own flexibility, with reference to the constitutional values and interests it protects, and with regard to present and future litigants. 383 Id. at 163 (“This flexibility, which we count as a virtue rather [than] a vice of PA, is never immune from attack by those who believe that a more determinate and principled approach to rights adjudication is possible, or that PA is just a fancy way to package judicial policy-making.”). Flexibility inheres in PA in another way: courts can adapt PA to fit their own purposes. In fact, how courts use PA varies widely across jurisdictional boundaries. 384 Id. at 163–64. To those who would claim that the proportionality framework is “foreign law” and therefore alien to American constitutionalism, we would reply that the United States has, at different times, evolved “homegrown” versions of PA. This Article shows as much. If the Supreme Court were to develop a more formalized version of PA, along the lines that we have argued, it would be an American creation, referencing existing caselaw and consistent with our own constitutional traditions and values. In any event, the problem of balancing in American rights adjudication lies once again at the top of the Supreme Court’s agenda. One hopes that, this time, the Justices will consider more deeply the merits of what is the most tried and tested approach to that problem: the proportionality principle. 385We close on a note of clarification. Though we believe that PA, as a doctrinal framework for adjudicating rights claims, performs better overall than any known competitor, we do not argue that it solves all of the legitimacy dilemmas faced by courts, including the problem of judicial law making. As we wrote:[T]he key to the political success of PA—its social logic—is that it provides a set of relatively stable, off-the-shelf, solutions to a set of generic dilemmas faced by the constitutional judge. If PA mitigates certain legitimacy problems, it also creates, or at least spotlights, an intractable, second-order, problem. PA does not camouflage judicial lawmaking. Properly employed, it requires courts to acknowledge and defend—honestly and openly—the policy choices that they make when they make constitutional choices. Proportionality is not a magic wand that judges wave to make all of the political dilemmas of rights review disappear. Indeed, waving it will expose rights adjudication for what it is: constitutionally-based lawmaking. Nonetheless, . . . PA offers the best position currently available for judges seeking to rationalize and defend rights review, given certain strategic considerations, the structure of modern rights provisions, and the precepts of contemporary constitutionalism.Id. at 77.

Footnotes

Jud Mathews is Visiting Assistant Professor, University of Illinois College of Law.Alec Stone Sweet is Leitner Professor of Law, Politics, and International Studies, Yale Law School. The authors would like to thank participants in the Yale Law School faculty workshop for helpful comments on a draft of this paper.

1 See T. Alexander Aleinikoff, Constitutional Law in the Age of Balancing, 96 Yale L.J. 943, 964 (1987); Kathleen M. Sullivan, Categorization, Balancing, and Government Interests, in Public Values in Constitutional Law 241, 241–44 (Stephen E. Gottlieb ed., 1993).

2 See Aleinikoff, supra note 1, at 964–71 (canvassing constitutional balancing tests).

3 See infra Part II; see also Richard H. Fallon, Jr., Strict Judicial Scrutiny, 54 UCLA L. Rev. 1267, 1288–89 (2007) (noting that the Supreme Court adopted a balancing approach to First Amendment claims that was quite speech protective in the 1940s, which was replaced by a more deferential balancing approach in 1949). Of course, the diversity of legal tests in American constitutional law also reflects, in part, the diversity in the structure of rights found in the Constitution. Some rights lend themselves easily to formulations as constitutional rules (e.g., “no quartering of soldiers”), and others are more easily read to invite balancing tests (e.g., due process).

4In District of Columbia v. Heller, 554 U.S. 570 (2008), Justices Scalia and Breyer engaged in an unusually acerbic argument about the appropriate methodology to employ in review of a District of Columbia measure that banned handguns in the home and required that other firearms be rendered inoperable. Writing for a five-member majority, Justice Scalia focused on the early history of the United States in order to show that the Second Amendment possessed a categorical, rule-like structure. Id. Once the Court determined that the Second Amendment guaranteed an individual right to bear arms for self-defense, the law was struck down. Id. at 635. On the issue of balancing, Justice Scalia stated that the Second Amendment, like the First, “is the very product of an interest-balancing by the people—which Justice Breyer would now conduct for them anew,” yet Justice Scalia also claimed: “We know of no other enumerated constitutional right whose core protection has been subjected to a freestanding ‘interest-balancing’ approach.” Id. at 634–35. For his part, Justice Breyer, joined by Justices Stevens, Souter, and Ginsburg in dissent, asserted that a “sort of ‘proportionality’” balancing approach is regularly used in “various constitutional contexts, including election-law cases, speech cases, and due process cases.” Id. at 690 (Breyer, J., dissenting). Advocating an “interest-balancing inquiry” that “asks whether the statute burdens a protected interest in a way or to an extent that is out of proportion to the statute’s salutary effects upon other important governmental interests,” Justice Breyer then adopted a relatively standard version of three-stage proportionality analysis to show why the District of Columbia’s ban should be upheld. Id. at 689–90, 693–719. For extended discussions of Heller’s debate over balancing, see Joseph Blocher, Categoricalism and Balancing in First and Second Amendment Analysis, 84 N.Y.U. L. Rev. 375 (2009); Moshe Cohen-Eliya & Iddo Porat, The Hidden Foreign Law Debate in Heller: The Proportionality Approach in American Constitutional Law, 46 San Diego L. Rev. 367 (2009).

5 See Alec Stone Sweet & Jud Mathews, Proportionality Balancing and Global Constitutionalism, 47 Colum. J. Transnat’l L. 72 (2008).

6 See id. at 138–59.

7 See generally Stone Sweet & Mathews, supra note 5.

8Canonically, the tiers are strict scrutiny, intermediate scrutiny, and rational basis review. The picture is complicated somewhat by less well-established gradations (e.g., “rational basis with bite”). Also, the Supreme Court does not always use these three labels. Kathleen Sullivan points out that tests for a number of constitutional claims, from privileges and immunities issues to public-forum speech regulations, amount to intermediate scrutiny, even when the Court does not apply such a label. Kathleen M. Sullivan, Post-Liberal Judging: The Roles of Categorization and Balancing, 63 U. Colo. L. Rev. 293, 297 (1992).

9 See Stone Sweet & Mathews, supra note 5, at 88.

10 See, e.g., Suzanne B. Goldberg, Equality Without Tiers, 77 S. Cal. L. Rev. 481, 484–91 (2004); Jeffrey M. Shaman, Cracks in the Structure: The Coming Breakdown of the Levels of Scrutiny, 45 Ohio St. L.J. 161 (1984); Eugene Volokh, Freedom of Speech, Permissible Tailoring and Transcending Strict Scrutiny, 144 U. Pa. L. Rev. 2417 (1996).

11 See, e.g., City of Cleburne v. Cleburne Living Ctr., Inc., 473 U.S. 432, 478 (1985) (Marshall, J., concurring in the judgment in part and dissenting in part); Craig v. Boren, 429 U.S. 190, 211–12 (1976) (Stevens, J., concurring).

12 See Stone Sweet & Mathews, supra note 5.

13 See id. at 75.

14 See id.

15Dieter Grimm, Proportionality in Canadian and German Constitutional Jurisprudence, 57 U. Toronto L.J. 383, 389 (2007).

16 See Ian Ayres & Sydney Foster, Don’t Tell, Don’t Ask: Narrow Tailoring After Grutter and Gratz, 85 Tex. L. Rev. 517, 520–24 (2007) (identifying LRM testing as a “central meaning” of the narrow tailoring requirement, criticizing an alternative approach to narrow tailoring in racial preference cases, and calling for a return to LRM testing).

17 See Grimm, supra note 15, at 384.

18 Jürgen Schwarze, European Administrative Law 857 (2006).

19 See, e.g., RJR-MacDonald Inc. v. Canada (Att’y Gen.), [1995] 3 S.C.R. 199, 342 (Can.) (“The tailoring process seldom admits of perfection and the courts must accord some leeway to the legislator. If the law falls within a range of reasonable alternatives, the courts will not find it overbroad merely because they can conceive of an alternative which might better tailor objective to infringement.”); Canada (Att’y Gen.) v. JTI-Macdonald Corp., [2007] 2 S.C.R. 610, 631 (Can.) (making a similar point).

20Robert Alexy, Constitutional Rights, Balancing, and Rationality, 16 Ratio Juris 131, 135 (2003).

21Following German practice, PA is not a jurisprudence of interests but of constitutional “principles” and “values.” Robert Alexy, A Theory of Constitutional Rights 86–93 (Julian Rivers trans., 2002); Donald P. Kommers, The Constitutional Jurisprudence of the Federal Republic of Germany 45–48 (1997).

22 See Grimm, supra note 15, at 393–95; Stone Sweet & Mathews, supra note 5, at 135–37.

23 See, e.g., HCJ 2056/04 Beit Sourik Village Council v. Gov’t of Isr. 58(5) IsrSC 807 [2004], translated in Isr. L. Rep. 264 (2004) (striking down the proposed route of Israel’s security fence as disproportionate in the strict sense after finding that it satisfied the first two standards of proportionality).

24 See Aleinikoff, supra note 1, at 984–95.

25 See, e.g., Ronald Dworkin, Rights as Trumps, in Theories of Rights 153 (Jeremy Waldron ed., 1984) (discussing situations in which individual rights may trump public goals).

26 See Jed Rubenfeld, Affirmative Action, 107 Yale L.J. 427, 436–37 (1997).

27Stone Sweet & Mathews, supra note 5, at 86 (“Modern constitutionalism is characterized by structural judicial supremacy, where the principals have, in effect, transferred a bundle of significant ‘political property rights’ to judges, for an indefinite duration.”); see also id. at 85–96 (discussing the concept of structural judicial supremacy with reference to balancing).

28 Id. at 80–97.

29 Id. at 87–90.

30 Id. at 80–87.

31 Alexy, supra note 21, at 40–110.

32The classic treatment is Paul J. DiMaggio & Walter W. Powell, The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Fields, in The New Institutionalism in Organizational Analysis 63, 67–69 (Walter W. Powell & Paul J. DiMaggio eds., 1991).

33 See, e.g., Stavros Tsakyrakis, Proportionality: An Assault on Human Rights?, 7 Int’l J. Const. L. 468 (2009); Grégoire C. N. Webber, Proportionality, Balancing, and the Cult of Constitutional Rights Scholarship, 23 Can. J.L. & Jurisprudence 179 (2010).

34 See Mattias Kumm, Political Liberalism and the Structure of Rights: On the Place and Limits of the Proportionality Requirement, in Law, Rights and Discourse: Themes from the Legal Philosophy of Robert Alexy 131, 131 (George Pavlakos ed., 2007). The answer given to Kumm’s question is ours, and not necessarily Kumm’s.

35 See, e.g., Ronald Dworkin, Taking Rights Seriously 184–205 (1977) (arguing that moral aspects of rights make balancing inappropriate except in emergency situations); Kai Möller, Balancing and the Structure of Constitutional Rights, 5 Int’l J. Const. L. 453, 458–61 (2007) (arguing that constitutional rights ought to be described from the standpoint of substantive morality).

36 See, e.g., Frederick Schauer, A Comment on the Structure of Rights, 27 Ga. L. Rev. 415, 429 (1993) (describing constitutional rights in the United States as “shields”).

37 See, e.g., Jed Rubenfeld, The First Amendment’s Purpose, 53 Stan. L. Rev. 767 (2001) (arguing that the rule-like structure of the First Amendment requires inquiry into government purpose and forbids balancing); Antonin Scalia, The Rule of Law as a Law of Rules, 56 U. Chi. L. Rev. 1175, 1179–80 (1989) (arguing that rights as rules limit the balancing discretion of judges and commit them to judicial restraint). The broader literature is reviewed in Blocher, supra note 4.

38 See Jeremy Waldron, Some Models of Dialogue Between Judges and Legislators, in 23 The Supreme Court Law Review 7 (Grant Huscroft & Ian Brodie eds., 2d Ser., 2004); Jeremy Waldron, The Core of the Case Against Judicial Review, 115 Yale L.J. 1346 (2006).

39 See, e.g., David M. Beatty, The Ultimate Rule of Law 159–76 (2004).

40We do not mean to understate the importance of research that bears upon how the Court has used (and has changed how it uses) strict scrutiny in specific cases or areas of the law, on which there are dozens of important articles.

41 See Fallon, supra note 3.

42 See Stephen A. Siegel, The Origin of the Compelling State Interest Test and Strict Scrutiny, 48 Am. J. Legal Hist. 355, 359–60 (2006).

43G. Edward White, Historicizing Judicial Scrutiny, 57 S.C. L. Rev. 1 (2005).

44Adam Winkler, Fatal in Theory and Strict in Fact: An Empirical Analysis of Strict Scrutiny in the Federal Courts, 59 Vand. L. Rev. 793 (2006).

45On the comparative agenda, see Beatty, supra note 39; Cohen-Eliya & Porat, supra note 4; Stephen Gardbaum, Limiting Constitutional Rights, 54 UCLA L. Rev. 789 (2007); Mattias Kumm, Constitutional Rights as Principles: On the Structure and Domain of Constitutional Justice, 2 Int’l J. Const. L. 574 (2004) (reviewing Alexy, supra note 21). In 2010, Law & Ethics of Human Rights, edited by Iddo Porat and Moshe Cohen-Eliya, among others, dedicated two symposium issues to the topic of “Rights, Balancing & Proportionality.See Symposium, Rights, Balancing & Proportionality, 4 Law & Ethics Hum. Rts. 1 (2010).

46Fallon puts it as follows:In the 1960s, the Warren Court was eager first to establish and then to consolidate a doctrinal structure sharply differentiating preferred from ordinary constitutional rights. With rational basis review established as the norm in run-of-the-mill cases, this strategy required the development of an implementing test or tests to protect preferred rights. Strict judicial scrutiny . . . furnished an attractive model . . . .Fallon, supra note 3, at 1335.

47 Id. at 1268.

48 Id. at 1270.

49 See Siegel, supra note 42, at 394–97. Siegel shows that strict scrutiny was not, in its original guise, designed to aid motive analysis by “smoking out” illicit motives. Id. at 394 (internal quotation marks omitted). Rather, the formula was a refinement of the balancing-of-interest approach that had long appeared in First Amendment cases. See id. (“When the Court introduced narrow tailoring and the compelling state interest standard into First Amendment analysis, it did so as part of its general ‘balancing/cost-benefit justification’ approach to First Amendment questions.”). Only later, once it migrated into the equal protection context, was strict scrutiny deployed as a tool for smoking out illegitimate motives. Id. at 394–97.

50 Id. at 394–95.

51Fallon, supra note 3, at 1306.

52 See infra text accompanying notes 179–82.

53In a similar spirit, Thomas Sullivan and Richard Frase identify sites in American doctrine, constitutional and otherwise, where some form of proportionality principle operates. See E. Thomas Sullivan & Richard S. Frase, Proportionality Principles in American Law 53–168 (2009). This undertaking is important, but different from our own. Sullivan and Frase take a somewhat more capacious view of proportionality than we do, and they identify a large number of areas of law where some form of interest-balancing exercise takes place. See id. at 6–7. We are focused specifically on doctrinal structures that anticipate the multistep framework of modern proportionality analysis, where judicial scrutiny mounts through a prescribed sequence of stages, culminating in means–ends testing and then balancing stricto sensu. As we show below, this same sequence of steps emerges more than once, in different areas of constitutional doctrine.

54 See infra Part III.A (highlighting similarities between mid-century rational basis review and proportionality).

55 See, e.g., District of Columbia v. Heller, 554 U.S. 570, 634–35 (2008).

56Guy Miller Struve claims that the “roots of the principle” go back at least to Lawton v. Steele, 152 U.S. 133 (1894). Guy Miller Struve, The Less-Restrictive-Alternative Principle and Economic Due Process, 80 Harv. L. Rev. 1463, 1464 n.4 (1967). In fact, Lawton absorbs the LRM test from prior dormant Commerce Clause cases.The first glimmers of LRM testing appear even earlier. In the 1821 case Anderson v. Dunn, challenging an exercise of the congressional contempt power, the Court ruled that Congress may only employ “the least possible power adequate to the end proposed.” 19 U.S. (6 Wheat.) 204, 230–31 (1821). But it took several decades before this test was employed systematically in an important doctrinal area, as we describe in the main text.

57“The Congress shall have Power . . . [t]o regulate Commerce with foreign Nations, and among the several States, and with the Indian Tribes.” U.S. Const. art. I, § 8, cl. 3.

5827 U.S. (2 Pet.) 245, 252 (1829).

59The one case still cited, see, e.g., United Haulers Ass’n v. Oneida-Herkimer Solid Waste Mgmt. Auth., 550 U.S. 330, 338 (2007), is Cooley v. Board of Wardens, 53 U.S. (12 How.) 299 (1851), in which the Court considers the basis for permitting local governments to take measures that would affect commerce:Now the power to regulate commerce embraces a vast field, containing not only many, but exceedingly various subjects, quite unlike in their nature; some imperatively demanding a single uniform rule, operating equally on the commerce of the United States in every port; and some, like the subject now in question, as imperatively demanding that diversity, which alone can meet the local necessities of navigation.Id. at 319.

60 See infra text accompanying notes 71–81.

6192 U.S. 275 (1875).

62 Id. at 278–79.

63 Id. at 280.

6495 U.S. 465, 468–69 (1877).

65 Id. at 473.

66 Id.

67 Id. at 470–71.

68 Id. at 470–72.

69 Id. at 472.

70 Id. at 470.

71136 U.S. 313 (1890).

72 Id. at 318.

73 Id. at 326. The Court rejected this assertion:To this we answer, that a statute may, upon its face, apply equally to the people of all the States, and yet be a regulation of interstate commerce which a State may not establish. A burden imposed by a State upon interstate commerce is not to be sustained simply because the statute imposing it applies alike to the people of all the States, including the people of the State enacting such statute.Id.

74 Id. at 322.

75 Id. at 319, 329–30.

76 Id. at 329 (emphases added).

77 Id. at 322 (“It will not do to say—certainly no judicial tribunal can with propriety assume—that the people of Minnesota may not, with due regard to their health, rely upon inspections in other states of animals there slaughtered for purposes of human food.”).

78Case 120/78, Rewe-Zentral AG v. Bundesmonopolverwaltung fur Branntwein (Cassis de Dijon), 1979 E.C.R. 649. The ECJ explicitly speaks of the right of individuals to engage in intra-European Community trade, which it derived from the prohibition of nontariff barriers under Article 28 of the European Economic Community Treaty. There is another important parallel between Cassis de Dijon and Minnesota v. Barber. In Cassis de Dijon, the ECJ (controversially) extended proportionality review to national market regulations that applied equally to goods, regardless of their origin. Prior to Cassis de Dijon, the ECJ limited proportionality review to measures that directly regulated trade. See Miguel Poiares Maduro, We the Court: The European Court of Justice and the European Economic Constitution 106–09 (1998); Alec Stone Sweet, The Judicial Construction of Europe 109–45 (2004).

79 187 U.S. 137, 151 (1902) (first two emphases added).

80 Id. at 152–53.

81 Compare Plumley v. Massachusetts, 155 U.S. 461 (1894) (upholding a state law regulating the sale of margarine on consumer protection grounds), with Schollenberger v. Pennsylvania, 171 U.S. 1 (1898) (striking down as unconstitutional a state law regulating the sale of margarine on LRM grounds).

82Howard Gillman explains this development succinctly. See Howard Gillman, Preferred Freedoms: The Progressive Expansion of State Power and the Rise of Modern Civil Liberties Jurisprudence, 47 Pol. Res. Q. 623, 629–34 (1992). Nineteenth-century judges understood the legislative power to have built-in limits: legislation that interfered with private liberty or property interests was legitimate only when it promoted the general welfare of the whole. Id. at 629. When legislative acts could not be so justified—when they were arbitrary or class legislation, enacted for the benefit of certain groups—they would be held to be improper exercises of legislative power and, hence, invalid. Id. By the late nineteenth century, federal courts had come to police the boundaries of legislative power with a form of LRM test. Id. at 634. Lawton v. Steele, 152 U.S. 133 (1894), illustrates this approach in practice. The case concerned a New York statute that declared fish nets in the vicinity of two waterways to be a public nuisance and subject to abatement and destruction by any person. Id. at 135–36. This provision was alleged to exceed the state’s police power. The Court framed its approach to the question as follows:To justify the State in thus interposing its authority in behalf of the public, it must appear, first, that the interests of the public generally, as distinguished from those of a particular class, require such interference; and, second, that the means are reasonably necessary for the accomplishment of the purpose, and not unduly oppressive upon individuals. The legislature may not, under the guise of protecting the public interests, arbitrarily interfere with private business, or impose unusual and unnecessary restrictions upon lawful occupations. In other words, its determination as to what is a proper exercise of its police powers is not final or conclusive, but is subject to the supervision of the courts.Id. at 137. Ultimately, the Court sided with the state, while strongly voicing its support of property rights more generally. Id. at 142–43. Three Justices dissented on the grounds that, in their view, the State had violated the plaintiff’s constitutional right to property because necessity had not been met. Id. at 144 (Fuller, C.J., dissenting).

83Moreover, congressional powers were limited to the exercise of those powers enumerated in Article I, Section 8 of the Federal Constitution.

84 See infra text accompanying note 187.

85171 U.S. at 29–30 (Gray, J., dissenting).

86544 U.S. 460, 473 (2005) (“Laws of the type at issue . . . . deprive citizens of their right to have access to the markets of other States on equal terms. The perceived necessity for reciprocal sale privileges risks generating the trade rivalries and animosities, the alliances and exclusivity, that the Constitution and, in particular, the Commerce Clause were designed to avoid.”).

87 Compare H. P. Hood & Sons, Inc. v. Du Mond, 336 U.S. 525, 539 (1949) (“Our system, fostered by the Commerce Clause, is that every farmer and every craftsman shall be encouraged to produce by the certainty that he will have free access to every market in the Nation, that no home embargoes will withhold his exports, and no foreign state will by customs duties or regulations exclude them.”), with Donald H. Regan, The Supreme Court and State Protectionism: Making Sense of the Dormant Commerce Clause, 84 Mich. L. Rev. 1091 (1986) (arguing that the dominant value protected by the dormant Commerce Clause is “national unity,” and that the Constitution’s dormant Commerce Clause corrects one of the main defects created by the Articles of Confederation—the ability of states to legislate benefits for their own citizens at the expense of the rest of the nation), and Hughes v. Oklahoma, 441 U.S. 322, 325–26 (1979) (noting that caselaw in this domain “reflect[s] a central concern of the Framers that was an immediate reason for calling the Constitutional Convention: the conviction that in order to succeed, the new Union would have to avoid the tendencies toward economic Balkanization that had plagued relations among the Colonies and later among the States under the Articles of Confederation”), and S.C. State Highway Dep’t v. Barnwell Bros., 303 U.S. 177, 185 n.2 (1938) (“Underlying the stated rule has been the thought, often expressed in judicial opinion, that when the regulation is of such a character that its burden falls principally upon those without the state, legislative action is not likely to be subjected to those political restraints which are normally exerted on legislation where it affects adversely some interests within the state.”).

88 See, e.g., Dean Milk Co. v. City of Madison, 340 U.S. 349, 354 (1951).

89Wyoming v. Oklahoma, 502 U.S. 437, 454–55 (1992).

90 See, e.g., United Haulers Ass’n v. Oneida-Herkimer Solid Waste Mgmt. Auth., 550 U.S. 330, 360 (2007) (Alito, J., dissenting) (“The Court has long subjected discriminatory legislation to strict scrutiny, and has never, until today, recognized an exception for discrimination in favor of a state-owned entity.”); cf. C & A Carbone, Inc. v. Town of Clarkstown, 511 U.S. 383, 392 (1994) (characterizing standard as “rigorous scrutiny”); id. at 422 (Souter, J., dissenting) (characterizing standard as “virtually fatal scrutiny”).

91397 U.S. 137, 142 (1970).

92550 U.S. at 338–39, 346 (“Discriminatory laws motivated by ‘simple economic protectionism’ are subject to a ‘virtually per se rule of invalidity,’ which can only be overcome by a showing that the State has no other means to advance a legitimate local purpose. . . . [But t]he Counties’ . . . ordinances are properly analyzed under the test set forth in Pike v. Bruce Church, Inc., which is reserved for laws ‘directed to legitimate local concerns, with effects upon interstate commerce that are only incidental.’” (citations omitted)).

93 Id. at 347.

94 Id. at 348 (Scalia, J., concurring in part) (quoting General Motors Corp. v. Tracy, 519 U.S. 278, 312 (1997) (Scalia, J., concurring)).

95 Id. at 348–49.

96 Id. at 348.

97 Id. at 349 (Thomas, J., concurring in the judgment) (“The negative Commerce Clause has no basis in the Constitution and has proved unworkable in practice. As the debate between the majority and dissent shows, application of the negative Commerce Clause turns solely on policy considerations, not on the Constitution. Because this Court has no policy role in regulating interstate commerce, I would discard the Court’s negative Commerce Clause jurisprudence.” (citations omitted)). Regarding balancing, Justice Thomas had this to say:To the extent that Congress does not exercise its authority to make that choice, the Constitution does not limit the States’ power to regulate commerce. In the face of congressional silence, the States are free to set the balance between protectionism and the free market. Instead of accepting this constitutional reality, the Court’s negative Commerce Clause jurisprudence gives nine Justices of this Court the power to decide the appropriate balance.Id. at 352.

98Simplifying a complex position, Regan argues that the courts should rarely engage in LRM testing and never engage in balancing, under Pike or any other authority, but should limit themselves exclusively to inquiry into “legislative purpose.” If the Court finds discriminatory purpose, then it must strike down the law. Regan further argues that the Supreme Court does not really balance at all, although it claims to do so. See Donald H. Regan, Judicial Review of Member-State Regulation of Trade Within a Federal or Quasi-Federal System: Protectionism and Balancing, Da Capo, 99 Mich. L. Rev. 1853, 1870–71 (2001); Regan, supra note 87. We find many of Regan’s readings of the Court’s caselaw strained at best, and his interpretation seems impossible to square with the two most recent Supreme Court rulings in the area, United Haulers and Granholm v. Heald.

99 See Stone Sweet & Mathews, supra note 5, at 101 (“[B]y the late nineteenth century, German administrative courts were striking down [state] police actions that violated proportionality, which was conceptualized at that time as an enforceable LRM test.”).

100 Stone Sweet, supra note 78, at 109–45.

101 Id.

102 Id.

103 See Stone Sweet & Mathews, supra note 5, at 152–59.

104For a view of courts and judicial review as guarantors of credible commitments between federated states engaged in market-building projects, see Walter Mattli, The Logic of Regional Integration: Europe and Beyond (1999); Stone Sweet, supra note 78, at 7–9, 109–45 (with reference to Europe); Jenna Bednar & William N. Eskridge, Jr., Steadying the Court’s “Unsteady Path”: A Theory of Judicial Enforcement of Federalism, 68 S. Cal. L. Rev. 1447 (1995) (with reference to the United States); Martin Shapiro, The European Court of Justice, in The Evolution of EU Law 321 (Paul Craig & Gráinne de Búrca eds., 1999).

105 See Siegel, supra note 42 (arguing that strict scrutiny first emerged in First Amendment cases).

106 See Melvin I. Urofsky, Division and Discord: The Supreme Court Under Stone and Vinson,1941–1953, at 89–90 (1997) (locating strict scrutiny’s origins in equal protection doctrine).

107 See, e.g., Fallon, supra note 3 (arguing that strict scrutiny emerged almost simultaneously in a number of areas).

108 See, e.g., Nebbia v. New York, 291 U.S. 502 (1934) (upholding an order of the New York Milk Control Board that fixed milk prices); Shaman, supra note 10, at 162–63.

109 See Jones v. City of Opelika, 316 U.S. 584, 608 (1942) (Stone, C.J., dissenting) (“[T]he Constitution, by virtue of the First and the Fourteenth Amendments, has put those freedoms [of speech and religion] in a preferred position.”).

110Siegel, supra note 42, at 364–80.

111 Id. at 394.

112 See Shaman, supra note 10, at 161–63.

113 Id.

114 See, e.g., Williamson v. Lee Optical of Okla., Inc., 348 U.S. 483 (1955) (upholding the constitutionality of an Oklahoma statute prohibiting persons other than optometrists or ophthalmologists from fitting or replacing lenses); see also infra Part III.A.

115 See Gillman, supra note 82, at 640–44.

116 Id. at 640–47.

117United States v. Carolene Prods. Co., 304 U.S. 144, 153 n.4 (1938).

118 Id.

119The phrase “bifurcated review project” is G. Edward White’s. See G. Edward White, The First Amendment Comes of Age: The Emergence of Free Speech in Twentieth-Century America, 95 Mich. L. Rev. 299, 301 (1996). White argued that the start of this project predates Carolene Products and was motivated specifically by a desire to afford heightened protection to speech rights. Id. at 301–02, 308–23.

120See, e.g., Simon & Schuster, Inc. v. Members of the N.Y. State Crime Victims Bd., 502 U.S. 105, 124–28 (1991) (Kennedy, J., concurring in the judgment); Goldberg, supra note 10; Greg Robinson & Toni Robinson, Korematsu and Beyond: Japanese Americans and the Origins of Strict Scrutiny, Law & Contemp. Probs., Spring 2005, at 29; Peter J. Rubin, Reconnecting Doctrine and Purpose: A Comprehensive Approach to Strict Scrutiny After Adarand and Shaw, 149 U. Pa. L. Rev. 1, 3 n.1 (2000).

121316 U.S. 535, 541 (1942).

122323 U.S. 214 (1944).

123 Skinner, 316 U.S. at 541; Korematsu, 323 U.S. at 216.

124Indeed, the phrase strict scrutiny appears in isolated decisions dating back to the late nineteenth century. See Winkler, supra note 44, at 798 n.10 (citing, inter alia, Altschuler v. Coburn, 57 N.W. 836, 838 (Neb. 1894), and Booher v. Worrill, 57 Ga. 235, 238 (1876)). We follow Stephen Siegel in taking strict scrutiny to consist of three parts: (1) a burden shifting to the government to justify the constitutionality of its law, and the requirement that a statute must be (2) “narrowly tailored” to serve (3) “a compelling state interest.” See Siegel, supra note 42, at 359–60.

125 See Siegel, supra note 42, at 361–80.

126Good accounts of this development from different perspectives can be found in Fallon, supra note 3, at 1285–97, and Gillman, supra note 82, at 640–44.

127 See Martin Shapiro, Freedom of Speech: The Supreme Court and Judicial Review 46–107 (1966).

128For a more comprehensive treatment, see id.

129 Id. at 58. The test was originally articulated in Justice Holmes’s 1919 opinion for the Court in Schenck v. United States, 249 U.S. 47, 52 (1919) (“The question in every case is whether the words used are used in such circumstances and are of such a nature as to create a clear and present danger that they will bring about the substantive evils that Congress has a right to prevent.”).

130 Shapiro, supra note 127, at 76–78 (describing balancing as an adjunct of the clear and present danger and preferred position doctrines). Marsh v. Alabama offers a concise statement of how the 1940s Court deployed balancing in the service of civil liberties: “When we balance the Constitutional rights of owners of property against those of the people to enjoy freedom of press and religion, as we must here, we remain mindful of the fact that the latter occupy a preferred position.” 326 U.S. 501, 509 (1946). The author is, of all people, Justice Black, who would later become balancing’s most persistent critic on the Court.

131Schneider v. State, 308 U.S. 147, 161 (1939).

132The facts of Martin v. City of Struthers illustrate how narrow tailoring was used. 319 U.S. 141 (1943). At issue was a municipal ordinance that forbade the ringing of doorbells for the purpose of distributing “handbills, circulars or other advertisements.” Id. at 142. The Court concluded that the prohibition violated the First Amendment because the measure restricted far more communication than necessary to combat the ills targeted by the measure. Id. at 146–49.

133 See, e.g., Am. Commc’ns Ass’n v. Douds, 339 U.S. 382 (1950).

134341 U.S. 494 (1951).

135 Id. at 495–96.

136 Id. at 510 (alteration in original) (quoting United States v. Dennis, 183 F.2d 201, 212 (2d Cir. 1950)) (internal quotation marks omitted).

137 See, e.g., Barenblatt v. United States, 360 U.S. 109, 139–44 (Black, J., dissenting); Laurent B. Frantz, The First Amendment in the Balance, 71 Yale L.J. 1424 (1962); Alexander Meiklejohn, The First Amendment Is an Absolute, 1961 Sup. Ct. Rev. 245.

138The “low-protectionist” and “high-protectionist” terminology is borrowed from Stephen Siegel. See Siegel, supra note 42, at 362.

139354 U.S. 234 (1957).

140 Id. at 234–42.

141 Id. at 254–55.

142 Id. at 251.

143 Id. at 262 (Frankfurter, J., concurring in the judgment).

144 Id.

145 Id. at 261; see also id. at 265 (“But the inviolability of privacy belonging to a citizen’s political loyalties has so overwhelming an importance to the well-being of our kind of society that it cannot be constitutionally encroached upon on the basis of so meagre a countervailing interest of the State as may be argumentatively found in the remote, shadowy threat to the security of New Hampshire allegedly presented in the origins and contributing elements of the Progressive Party and in petitioner’s relations to these.”).

146 See Siegel, supra note 42, at 367.

147 See Barenblatt v. United States, 360 U.S. 109, 127 (1959); Uphaus v. Wyman, 360 U.S. 72, 79–81 (1959).

148As Siegel notes, although courts applied strict scrutiny in several First Amendment cases in the 1960s, strict scrutiny only became a cornerstone of First Amendment doctrine in the 1970s, after it had been embraced in equal protection jurisprudence.

149 Siegel, supra note 42, at 375.

150 Id. at 375–76.

151366 U.S. 599, 610 (1961) (Brennan, J., concurring and dissenting).

152 Id. at 611.

153 Id. at 612 (quoting W. Va. State Bd. of Educ. v. Barnette, 319 U.S. 624 (1943)).

154 Id. at 613.

155 Id. at 613–14.

156 Id. at 614.

157Barnette, 319 U.S. 624.

158374 U.S. 398 (1963).

159 Id. at 407.

160393 U.S. 23 (1968).

161 Id. at 31. The right to vote was also implicated, but the Court’s analysis of these two rights was the same.

162For instance, the State argued that promoting a two-party system encouraged compromise and political stability. But the Court noted that the measures did not just promote a two-party system; rather, they gave two particular parties a monopoly over citizens’ votes. Id. at 31–32.

163 Id. at 32.

164 Id.

165 See, e.g., Shapiro v. Thompson, 394 U.S. 618 (1969) (applying strict scrutiny to strike down a one-year waiting period for welfare-benefit eligibility on equal protection grounds).

166Gerald Gunther, The Supreme Court, 1971 Term—Foreword: In Search of Evolving Doctrine on a Changing Court: A Model for a Newer Equal Protection, 86 Harv. L. Rev. 1, 8 (1972).

167 See Rubin, supra note 120, at 4 (“In the courts, strict scrutiny is essentially invoked, not employed. Despite its name—strict ‘scrutiny’—it ordinarily amounts to a finding of invalidity, not a tool of analysis.”).

168 First National Bank of Boston v. Bellotti, 435 U.S. 765 (1978), illustrates the point. In this case, Massachusetts advanced an interest in protecting shareholders to support its ban on corporate contributions to political campaigns. Id. at 766. But the Court found the statute both overinclusive and underinclusive with respect to those interests. Id. at 793. The Court “[a]ssum[ed], arguendo, that protection of shareholders is a ‘compelling’ interest under the circumstances of this case,” and struck down the statute for lack of a “substantially relevant correlation between the governmental interest” and the right to free speech. Id. at 795 (quoting Shelton v. Tucker, 364 U.S. 479, 485 (1960)).

169These events are described in some detail in our previous article. See Stone Sweet & Mathews, supra note 5, at 97–111.

170 Id.

171 Id. at 100–02.

172 Id. at 100.

173In Kreuzberg, the court struck down a regulation by the Berlin police that restricted the heights of buildings in order to afford a view of a military monument. Preußisches Oberverwaltungsgericht [PrOVG] [Prussian Higher Administrative Court] July 14, 1882, 9 Entscheidungen des preußischen Oberverwaltungsgerichts [PrOVGE] 353, reprinted in 100 Deutsches Verwaltungsblatt 219 (1985).

174For instance, in an 1886 case, the court ruled that the police could not require, on public safety grounds, a landowner to remove a post erected at the edge of his property. Rather, all that was necessary to protect the public was requiring the landowner to light the post after dark. As the court explained, “The protection from accidents . . . is indeed the task of the police; this task and the authority finds its limit, however, in that the chosen measures may not extend farther than they must to meet the goal of eliminating the danger.” PrOVG July 3, 1886, 13 PrOVGE 426, 427.

175 See supra Part I.A.

176 See Stone Sweet & Mathews, supra note 5, at 104–11.

177 Id.

178 Id. at 107–08.

179 Id. at 99, 108–09.

180 Id. at 109–10.

181Nor should the parallels obscure important differences in the scope of review. Most importantly, in nineteenth-century Germany, judicial scrutiny was only applied against administrative action, not legislation, and in the twentieth century, proportionality persisted as an administrative doctrine even as it gained new life in constitutional law. Id. at 97–111.

182 See infra Part III.

183 See Siegel, supra note 42, at 398–99. As Siegel points out, the Court was slow to apply the compelling state interest test in racial discrimination cases because the narrow tailoring inquiry was sufficient to ferret out illicit motives.

184 See Fallon, supra note 3, at 1308–11; see also Rubenfeld, supra note 26, at 428.

185As has been noted, however, strict scrutiny is by no means always “fatal in fact.” See Winkler, supra note 44. Winkler’s article does not cover cases before 1990, and no one has yet demonstrated how fatal strict scrutiny in fact was during the 1960s, 1970s, and 1980s.

186As noted above, Justices Black and Douglas, who took this view of rights, originally declined to join opinions applying strict scrutiny, but did in time. See supra text accompanying note 50.

187 See supra note 168.

188 See, e.g., Jeffrey M. Shaman, Constitutional Interpretation: Illusion and Reality 81–84 (2001) (discussing how the Court upgraded minimal scrutiny in certain contexts beginning in the 1970s).

189 See Calvin Massey, The New Formalism: Requiem for Tiered Scrutiny?, 6 U. Pa. J. Const. L. 945, 946 (2004) (“The flaccidity of the strict scrutiny employed in Grutter, coupled with the robust skepticism of the minimal scrutiny employed in Lawrence, suggests that the neat compartments of tiered scrutiny are beginning to collapse.”).

190 See infra Part III.A.

191 See supra note 185.

192 See supra Part I.

193In a leading child pornography case, R. v. Sharpe, the Canadian Supreme Court held that a provision of the Criminal Code, as applied to Mr. Sharpe, violated his freedom of expression but was justified as a proportional measure designed to protect children from exploitation. [2001] 1 S.C.R. 45. The approach thus contrasts with categorical, rule-based approaches to rights protection that seek to dispense with balancing once the nature and scope of the right have been defined. In this latter approach, a court might decide that Mr. Sharpe’s rights were not abridged, since child pornography is not a protected form of expression, per se. Id.; cf. New York v. Ferber, 458 U.S. 747 (1982) (holding that child pornography is beyond the limits of First Amendment protection).

194We are not the first to take this view. See, e.g., James W. Ely, Jr., The Guardian of Every Other Right: A Constitutional History of Property Rights (1992); Richard A. Epstein, Takings (1985).

195348 U.S. 483 (1955).

196 Id. at 485–86.

197 Id. at 485.

198 Id. at 485 n.1.

199Lee Optical of Okla., Inc. v. Williamson, 120 F. Supp. 128, 137 (W.D. Okla. 1954), rev’d, 348 U.S. 483 (1955).

200 Id.

201 Lee Optical, 348 U.S. at 487.

202 Lee Optical, 120 F. Supp. at 144 n.37 (“The inevitable result of the enforcement of the provisions found by this Court unconstitutional . . . will be to literally put said plaintiffs out of business . . . .”).

203 Id. at 135 (“The evidence indicates, almost without variance, that written prescriptions issued by the professional examiner contain no directive data in regard to the manner in which the spectacles are to be fitted to the face of the wearer. In addition, the Court is satisfied that the mere fitting of frames to the face, where the old lenses are available, is in reality only an incident to what fundamentally is a merchandising transaction, that is, the sale of a pair of frames; and, in any event, the knowledge necessary to perform these services is strictly artisan in character and can skillfully and accurately be performed without the professional knowledge and training essential to qualify as a licensed optometrist or ophthalmologist.”).

204 Id. at 132–33.

205 Id. at 142 (“The dispensing optician, a merchant in this particular, cannot arbitrarily be divested of a substantial portion of his business upon the pretext that such a deprivation is rationally related to the public health.”).

206 Id. at 137–38 n.21 (“The action by the legislature in the instant act is as unreasonable as if pharmicists [sic] were divested of their right to do business by legislative action which delegated to physicians the absolute control and responsibility for accurate compounding of medical prescriptions for the reason that a physician alone is qualified to prescribe drugs for patients.”).

207 Id. at 132.

208 Id. at 137 (emphasis added) (quoting Lawton v. Steele, 152 U.S. 133, 137 (1894)).

209Concurring with this portion of the decision, Judge Murrah’s opinion also has the feel of a standard proportionality analysis:The obvious result of this legislation is to appropriate a property right of one class to follow a legitimate calling or occupation and to give it to another class not shown to be more competent in the public interest. Indeed, the evidence shows that it would be inimical to the public interest to require the wearer of eye glasses to return to the ophthalmologist in order to obtain a prescription for the replacement or duplication of an eye glass, or to place the burden upon the ophthalmologist of merchandising frames and mountings and other appliances, or placing that purely commercial monopoly in the optometrist. There is no rational basis for depriving the optician of the right to perform this purely mechanical and commercial service to the public.Id. at 144 (Murrah, J., dissenting in part) (dissenting on provisions affecting advertising, which we do not discuss here).

210 Id. at 138.

211 See supra Part II.

212Williamson v. Lee Optical of Okla., Inc., 348 U.S. 483, 497–88 (1955) (citations omitted) (quoting Munn v. Illinois, 94 U.S. 113, 134 (1877)).

213Justice Field’s dissent from Munn v. Illinois, which Justice Douglas cites as authority, applies equally well to Lee Optical:The construction actually given by . . . this court makes the provision, in the language of Taney, a protection to “a mere barren and abstract right, without any practical operation upon the business of life,” and renders it “illusive and nugatory, mere words of form, affording no protection and producing no practical result.”94 U.S. 113, 145 (1876) (Field, J., dissenting) (quoting Bronson v. Kinzie, 42 U.S. (1 How.) 311, 318 (1843)).

214With respect to the equal protection claim, Justice Douglas had only this to say:The problem of legislative classification is a perennial one, admitting of no doctrinaire definition. Evils in the same field may be of different dimensions and proportions, requiring different remedies. Or so the legislature may think. Or the reform may take one step at a time, addressing itself to the phase of the problem which seems most acute to the legislative mind. The legislature may select one phase of one field and apply a remedy there, neglecting the others. The prohibition of the Equal Protection Clause goes no further than the invidious discrimination. We cannot say that that point has been reached here.Lee Optical, 348 U.S. at 489 (citations omitted).

215Bundesverfassungsgericht [BVerfG] [Federal Constitutional Court] June 11, 1958, 7 Entscheidungen des Bundesverfassungsgerichts [BVerfGE] 377 (Ger.).

216 Article 12(1) states, “All Germans shall have the right freely to choose their occupation or profession, their place of work and their place of training. The practice of an occupation or profession may be regulated by or pursuant to a law.” Grundgesetz für die Bundesrepublik Deutschland [Grundgesetz] [GG] [Basic Law], May 23, 1949, BGBl. I, art. 12(1) (Ger.).

2177 BVerfGE 377 (380).

2187 BVerfGE 377 (404–05).

2197 BVerfGE 377 (396–12). The case was decided before the GFCC had fully developed the proportionality principle. Today there is one proportionality test for all fundamental rights. Stone Sweet & Mathews, supra note 5, at 110–11.

220BVerfG Jan. 7, 1959, 9 BVerfGE 73 (82).

2219 BVerfGE 73 (76–81).

222 See Paul G. Kauper, The Constitutions of West Germany and the United States: A Comparative Study, 58 Mich. L. Rev. 1091, 1130 (1960).

223The leading decision is the Investment Aid I case, BVerfG July 20, 1959, 4 BVerfGE 7. A translation and commentary can be found in Kommers, supra note 21, at 243–45.

224The GFCC seeks to reduce, as much as possible, the tension between the pleaded property right and the public good resulting from a limitation of property rights; that is, the GFCC balances among contending constitutional values in light of the context of the case. In cases like Lee Optical, the U.S. Supreme Court has little use for either the Constitution or context. For a superb comparison of German and American approaches to property rights as constitutional rights, see Gregory S. Alexander, Property as a Fundamental Constitutional Right? The German Example, 88 Cornell L. Rev. 733 (2003).

225 The Basic Law provides for the right to property in Article 14, which qualifies the right to property in a manner that strongly implies the use of PA. GG art. 14 (“(1) Property and the right of inheritance shall be guaranteed. Their content and limits shall be defined by the laws. (2) Property entails obligations. Its use shall also serve the public good. (3) Expropriation shall only be permissible for the public good. It may only be ordered by or pursuant to a law that determines the nature and extent of compensation. Such compensation shall be determined by establishing an equitable balance between the public interest and the interests of those affected. In case of dispute concerning the amount of compensation, recourse may be had to the ordinary courts.”).

226391 U.S. 367 (1968).

227 See id. at 369–70, 376.

228In 1965, Congress adopted an amendment to the Universal Military Training and Service Act of 1948 in part to deter those who would burn their draft cards in protest of the war. See id. at 370, 385–86. The circuit court, noting that O’Brien’s action was already punishable under a separate regulation, declared the amendment unconstitutional on the grounds that it was expressly designed to limit speech. O’Brien v. United States, 376 F.2d 538, 540–41 (1st Cir. 1967), vacated, 391 U.S. 367.

229 See O’Brien, 391 U.S. at 376 (“We cannot accept the view that an apparently limitless variety of conduct can be labeled ‘speech’ whenever the person engaging in the conduct intends thereby to express an idea. However, even on the assumption that the alleged communicative element in O’Brien’s conduct is sufficient to bring into play the First Amendment, it does not necessarily follow that the destruction of a registration certificate is constitutionally protected activity.”).

230 Id. at 376–77 (footnotes omitted).

231 Id. at 378–80, 382.

232 See id. at 366–67.

233 See id. at 386.

234Kathleen Sullivan, for one, views the O’Brien test as a version of intermediate scrutiny. Sullivan, supra note 8, at 297.

235 For example, see Craig v. Boren, 429 U.S. 190, 197 (1976), discussed in Part III.C.1 infra. Instead of narrow tailoring, the test required a “substantial” relation between the policy and the government’s objectives, again echoing O’Brien’s language. See id.

236 See White, supra note 119, at 301 (describing the Court’s “bifurcated review project”).

237Also, on occasion the Court has claimed to be applying one of the two canonical standards, but transparently not done so. See, e.g., City of Cleburne v. Cleburne Living Ctr., Inc., 473 U.S. 432 (1985) (recognizing that a law discriminating against mentally retarded individuals does not merit heightened scrutiny because mental retardation is not a quasi-suspect class, but finding that the law fails to rationally relate to any of its various justifications, suggesting that the Court was actually applying a higher standard than rational basis review).

238 See Gillman, supra note 82, at 623–26; White, supra note 119, at 309.

239 See, e.g., Plyler v. Doe, 457 U.S. 202, 244 (1982) (Burger, C.J., dissenting) (criticizing the majority’s application of intermediate scrutiny to a statute that deprives illegal alien children of the right to public education); Craig, 429 U.S. at 220–21 (Rehnquist, J., dissenting) (criticizing the majority’s application of intermediate scrutiny to a law that imposes different drinking ages for men and women).

240 See White, supra note 43, at 82–83.

241Another line of prominent cases in which the Court shifts the standard of review, with dramatic results, includes Metro Broadcasting, Inc. v. FCC, 497 U.S. 547 (1990) (applying intermediate scrutiny to affirmative action policies in broadcasting), and Adarand Constructors, Inc. v. Peña, 515 U.S. 200 (1995) (applying strict scrutiny to affirmative action policy in federal contracting program).

242Tiered review is also a pervasive feature of state constitutional doctrine, which space prevents us from discussing in this Article. Notably, states have also been at the forefront of experimenting with rights review, and some have opted out of tiered review of their constitutional claims. See Goldberg, supra note 10, at 522 n.154.

243 See, e.g., Hoyt v. Florida, 368 U.S. 57 (1961) (upholding a Florida law that disallowed women from being selected for jury service unless they volunteered); Goesaert v. Cleary, 335 U.S. 464 (1948) (upholding a Michigan law that banned certain women from serving as bartenders).

244In Goesaert, for instance, the Court held that Michigan had a rational basis for banning women, except the wives or daughters of male bar owners, from tending bars. 335 U.S. at 466–67.

245404 U.S. 71, 76–77 (1971).

246 Id. at 76.

247 Id.

248 Id. at 76–77.

249411 U.S. 677, 682 (1973).

250 Id. at 678–79.

251 Id. at 678.

252 Id. at 691–92 (Powell, J., concurring in the judgment).

253 See id. at 692. The Powell concurrence also argued that for the Court to decide the issue, it would effectively preempt democratic deliberations on precisely this issue, as the Equal Rights Amendment was then under consideration, and would have the effect of making sex a suspect classification. Id. Justice Stewart concurred separately, and Justice Rehnquist dissented, adopting the opinion of the district court. Id. at 691.

254429 U.S. 190 (1976).

255 Id. at 191–92.

256 Id. at 197.

257 Id. at 199–200.

258 Id. at 202.

259 Id. at 211 n.* (Powell, J., concurring).

260 Id. at 211–12 (Stevens, J., concurring).

261 Id. at 213–14.

262 Id. at 220–21 (Rehnquist, J., dissenting).

263 Id. at 221.

264 Id. at 225–26.

265 See, e.g., Califano v. Westcott, 443 U.S. 76 (1979) (applying intermediate scrutiny to unemployment benefits provided by the Social Security Act to a father, but not a mother); Orr v. Orr, 440 U.S. 268 (1979) (applying intermediate scrutiny to Alabama statute under which husbands, but not wives, may be required to make alimony payments); Califano v. Webster, 430 U.S. 313 (1977) (applying intermediate scrutiny to benefits provided by the Social Security Act that favor female wage owners over similarly situated male wage owners).

266442 U.S. 256 (1979).

267 See, e.g., Miss. Univ. for Women v. Hogan, 458 U.S. 718, 724 (1982); Kirchberg v. Feenstra, 450 U.S. 455, 461 (1981).

268United States v. Virginia, 518 U.S. 515, 531 (1996). The groundwork for this move was laid by Hogan and J. E. B. v. Alabama ex rel. T. B., 511 U.S. 127 (1994). In these cases, the Court also characterized the defendant’s burden as providing “an exceedingly persuasive justification” for the discrimination. But, in the Court’s analysis, the way to meet this burden was, in essence, to show a substantial relation to an important governmental interest: the test derived from Craig.

269 Virginia, 518 U.S. at 550.

270 See id. at 559 (Rehnquist, C.J., concurring in the judgment).

271Craig v. Boren, 429 U.S. 190, 219–21 (1976) (Rehnquist, J., dissenting).

272 Virginia, 518 U.S. at 570–76 (Scalia, J., dissenting) (“Only the amorphous ‘exceedingly persuasive justification’ phrase, and not the standard elaboration of intermediate scrutiny, can be made to yield this conclusion that VMI’s single-sex composition is unconstitutional because there exist several women (or, one would have to conclude under the Court’s reasoning, a single woman) willing and able to undertake VMI’s program. Intermediate scrutiny has never required a least-restrictive-means analysis, but only a ‘substantial relation’ between the classification and the state interests that it serves.”).

273Sullivan, supra note 8, at 297 (internal quotation marks omitted).

274As Justice Rehnquist pointedly noted in dissent, the Craig test—that measures must be substantially related to the achievement of important government objectives—seems to “come[] out of thin air.” Craig, 429 U.S. at 220 (Rehnquist, J., dissenting).

275410 U.S. 113, 155 (1973).

276 See David Margolick, Seeking Strength in Independence, Abortion-Rights Unit Quits A.C.L.U., N.Y. Times, May 21, 1992, at A20 (“It is in [Casey], argued last month before the Supreme Court, that abortion-rights advocates expect the Court to either overturn or to greatly weaken Roe v. Wade.”).

277Webster v. Reprod. Health Servs., 492 U.S. 490, 517–19 (1989) (plurality opinion).

278Planned Parenthood of Se. Pa. v. Casey, 505 U.S. 833, 877 (1992).

279 Id.

280 Id. at 878 (citation omitted).

281 See id. at 964 (Rehnquist, C.J., concurring in the judgment in part and dissenting in part) (describing the undue burden standard as “created largely out of whole cloth by the authors of the joint opinion”).

282Abortion rights opinions that discuss some form of undue burden analysis include City of Akron v. Akron Center for Reproductive Health, Inc., 462 U.S. 416, 463 (1983) (O’Connor, J., dissenting); Harris v. McRae, 448 U.S. 297, 314 (1980); Bellotti v. Baird, 443 U.S. 622, 640 (1979); Maher v. Roe, 432 U.S. 464, 473 (1977); Bellotti v. Baird, 428 U.S. 132, 147 (1976). For more on the use of undue burden standards in and out of the abortion rights context, see Alan Brownstein, How Rights Are Infringed: The Role of Undue Burden Analysis in Constitutional Doctrine, 45 Hastings L.J. 867 (1994); Valerie J. Pacer, Note, Salvaging the Undue Burden Standard—Is It a Lost Cause? The Undue Burden Standard and Fundamental Rights Analysis, 73 Wash. U. L.Q. 295 (1995). One author notes that 256 Supreme Court decisions issued between 1945 and 1993 contain the phrase undue burden or the equivalent. Curtis E. Harris, An Undue Burden: Balancing in an Age of Relativism, 18 Okla. City U. L. Rev. 363, 423 (1993).

283 See supra Part II.A.

284Ramos v. Town of Vernon, 353 F.3d 171, 176, 186 (2d Cir. 2003).

285Hutchins v. District of Columbia, 188 F.3d 531, 541, 548 (D.C. Cir. 1999) (en banc); Schleifer v. City of Charlottesville, 159 F.3d 843, 847, 855 (4th Cir. 1998).

286Qutb v. Strauss, 11 F.3d 488, 492, 496 (5th Cir. 1993).

287Nunez v. City of San Diego, 114 F.3d 935, 946, 949 (9th Cir. 1997).

288 See Ramos ex rel. Ramos v. Town of Vernon, 48 F. Supp. 2d 176, 188 (D. Conn. 1999), rev’d, 353 F.3d 171 (2d Cir. 2003). Ramos also raised vagueness, First Amendment overbreadth, and Fourth Amendment claims against the statute, as well as claims arising under the Connecticut constitution. Id. at 180–81. His parents also alleged infringement of their right as parents to direct the raising of their children. Id. at 181. The district court rejected all of the federal claims and certified the state law claims to the Connecticut Supreme Court, id. at 188, which rejected them. See Ramos v. Town of Vernon, 761 A.2d 705, 710 (Conn. 2000).

289 Ramos, 48 F. Supp. 2d at 184.

290 Id.

291 Id.

292 Id. (quoting Schleifer v. City of Charlottesville, 159 F.3d 843, 847 (4th Cir. 1998)) (internal quotation mark omitted).

293 Id. at 185–86.

294 Id.

295 Ramos ex rel. Ramos v. Town of Vernon, 353 F.3d 171, 187 (2d Cir. 2003).

296Id. at 176.

297 Id.

298 Id.

299 Id. at 178.

300 Id. at 179.

301 Id.

302 Id. at 180 (citation omitted). The court went on to argue that intermediate scrutiny is flexible enough to allow legislation properly drafted to address the needs and vulnerability of children, but probing enough to disallow legislation justified by facile and untested generalizations about the young. Id. at 181.

303 Id. at 179 (internal quotation marks omitted).

304 See id. at 181 n.4 (noting that the Supreme Court has not generally considered age to be a suspect classification).

305 See id. at 179–81.

306 Id.

307 Id. at 182.

308 Id. at 185–87.

309 Id. at 186.

310 Id.

311 Id. at 188 (Winter, J., dissenting).

312 Id. at 190–91. The dissent also considered, and rejected, the claim that the curfew burdened parental or guardian rights. Id. at 190.

313 Cf. District of Columbia v. Heller, 554 U.S. 570, 628–29 (2008) (evaluating the firearms restriction without selecting a standard of review).

314 See supra text accompanying notes 268–70.

315City of Cleburne v. Cleburne Living Ctr., Inc., 473 U.S. 432, 460 (1985) (Marshall, J., concurring in the judgment in part and dissenting in part) (quoting San Antonio Indep. Sch. Dist. v. Rodriguez, 411 U.S. 1, 99 (1973) (Marshall, J., dissenting)).

316A good overview of the doctrine in English can be found in Susanne Baer, Equality: The Jurisprudence of the German Constitutional Court, 5 Colum. J. Eur. L. 249 (1999). A more detailed treatment, albeit in German, can be found in a constitutional commentary, such as Hans D. Jarass & Bodo Pieroth, Grundgesetz für die Bundesrepublik Deutschland 90–137 (2006).

317One notable difference is that Article 3 spells out what we would call “suspect classifications,” on the basis of which discrimination is not allowed. GG art. 3(3) (“No person shall be favoured or disfavoured because of sex, parentage, race, language, homeland and origin, faith, or religious or political opinions. No person shall be disfavoured because of disability.”); see also id. art. 3(2) (specifically addressing sex equality).

318PA is not used across the board in the GFCC’s Article 3 jurisprudence, but increasingly it is supplanting older approaches. See Baer, supra note 316, at 260. We focus our discussion on the “new” proportionality-based approach to the general equality guarantee in Article 3(1).

319BVerfG Apr. 28, 1999, 100 BVerfGE 138 (174); see also BVerfG May 30, 1990, 82 BVerfGE 126.

320Jarass & Pieroth, supra note 316, at 100; Baer, supra note 316, at 262.

321BVerfG Mar. 2, 1999, 99 BVerfGE 367 (388).

322 Jarass & Pieroth, supra note 316, at 99, 100. The legislature has substantial latitude, for instance, in choosing the subjects and rates of taxation. Id. at 109.

323For a (now somewhat dated) overview of prominent critiques and controversies, see Albert Bleckmann, Die Struktur des Allgemeinen Gleichheitssatzes 1–2 (1995).

324 Alec Stone Sweet, Governing with Judges: Constitutional Politics in Europe 109–14 (2000).

325BVerfG May 28, 1993, 88 BVerfGE 203 (“Abortion II”); BVerfG Feb. 25, 1975, 39 BVerfGE 1 (“Abortion I”).

326A careful comparison of the two approaches can be found in Richard E. Levy & Alexander Somek, Paradoxical Parallels in the American and German Abortion Decisions, 9 Tul. J. Int’l & Comp. L. 109 (2001). An abridged translation of the two key German decisions is contained in Kommers, supra note 21, at 336–56.

32788 BVerfGE 203 (253).

32839 BVerfGE 1 (44–46).

329 Id. at 47.

330Id. at 48–50.

33188 BVerfGE 203 (255); 39 BVerfGE 1 (48).

33239 BVerfGE 1 (50) (quoted in Kommers, supra note 21, at 341).

333 See Rüdiger Konradin Albrecht, Zumutbarkeit als Verfassungsmassstab (1995) (distinguishing Unzumutbarkeit from proportionality).

334Levy & Somek, supra note 326, at 122.

335 Id. at 111 (“In both countries, abortion is generally available early in the pregnancy after various informational, counseling and waiting restrictions; and is available later in the pregnancy to protect the life or health of the mother, as well as in cases of serious fetal deformity, rape, or incest.”).

336 See Robert Post & Reva Siegel, Roe Rage: Democratic Constitutionalism and Backlash, 42 Harv. C.R.-C.L. L. Rev. 373 (2007).

337The difference in tone extends to the bench as well. The German abortion rulings produced dissents, but none that challenged the legitimacy of the majority’s ruling, or the good faith of the justices, like Justice Scalia’s blistering dissent in Casey. See Planned Parenthood of Se. Pa. v. Casey, 505 U.S. 833, 982–1002 (1992) (Scalia, J., dissenting).

338 See supra Part III.

339As we noted before, a number of areas of American constitutional law have interest-balancing tests that resemble PA. Justice Breyer lists some of these in his Heller dissent. District of Columbia v. Heller, 554 U.S. 570, 690 (2008) (Breyer, J., dissenting); see also, e.g., Nixon v. Shrink Mo. Gov’t PAC, 528 U.S. 377, 403 (2000) (citing examples where the Court has taken such an approach); Thompson v. W. States Med. Ctr., 535 U.S. 357, 388 (2002) (Breyer, J., dissenting) (commercial speech); Burdick v. Takushi, 504 U.S. 428, 433 (1992) (election regulation); Mathews v. Eldridge, 424 U.S. 319, 339–49 (1976) (procedural due process); Pickering v. Bd. of Ed. of Twp. High Sch. Dist. 205, 391 U.S. 563, 568 (1968) (government employee speech). The Supreme Court has employed balancing tests outside of the constitutional context as well. Perhaps most famously, the “rule of reason” used to evaluate combinations and contracts in restraint of trade under section 1 of the Sherman Act is a balancing test: the court asks whether net effect of a restriction on trade is pro-competitive. See Bd. of Trade v. United States, 246 U.S. 231 (1918). Indeed, in recent years federal appeals courts have supplemented the Board of Trade balancing with an LRM test, so that the total effect closely resembles PA. See Gabriel A. Feldman, The Misuse of the Less Restrictive Alternative Inquiry in Rule of Reason Analysis, 58 Am. U. L. Rev. 561, 580 (2009). Election law, too, has turned with increasing frequency to a standard that balances the state’s interest against the burden on an individual’s right to vote in a manner resembling proportionality. See Joseph Fishkin, Equal Citizenship and the Individual Right to Vote, 86 Ind. L.J. (forthcoming 2011).

340 See James E. Fleming, “There Is Only One Equal Protection Clause”: An Appreciation of Justice Stevens’s Equal Protection Jurisprudence, 74 Fordham L. Rev. 2301, 2305–08 (2006) (describing the positions of both Justices).

341411 U.S. 1, 102–03 (1973) (Marshall, J., dissenting).

342City of Cleburne v. Cleburne Living Ctr., Inc., 473 U.S. 432, 452–53 (1985) (Stevens, J., concurring).

343 See Goldberg, supra note 10, at 519–24.

344 Cf. Craig v. Boren, 429 U.S. 190, 212 (1976) (Stevens, J., concurring) (arguing that cases in different tiers “actually apply a single standard in a reasonably consistent fashion”).

345District of Columbia v. Heller, 554 U.S. 570, 693–719 (2008) (Breyer, J., dissenting). In a celebrated 2010 commencement address at Harvard, retiring-Justice Souter emphasized the inescapable conflicts between values involved in constitutional litigation. Justice David H. Souter, Commencement Address at Harvard University (May 27, 2010), available at http://news.harvard.edu/gazette/story/2010/05/text-of-justice-david-souters-speech/.

346 Heller, 554 U.S. at 693–719. In his dissent, Justice Breyer regarded his approach—correctly—not as a great departure from the Court’s practice, but as a rationalization of techniques already in place, and of a piece with the standard of review employed in different areas. Id.

347 U.S. Const. amend. III.

348 Cf. Sullivan & Frase, supra note 53, at 173 (proposing that proportionality “be adopted as a general standard of review”).

349424 U.S. 693 (1976).

350 Id. at 695.

351 Id. at 701.

352 See Barbara E. Armacost, Race and Reputation: The Real Legacy of Paul v. Davis, 85 Va. L. Rev. 569, 571 (1999) (“Scholars have been relentlessly and uniformly negative in their reactions to the Supreme Court’s opinion and holding in Paul . . . .”); id. at 575–84 (collecting criticisms of Paul).

353424 U.S. 319 (1976).

354 See, e.g., Miller v. California, 413 U.S. 15, 36 (1973) (affirming that obscene material is not protected by the First Amendment).

355 See Mattias Kumm, Who Is Afraid of the Total Constitution? Constitutional Rights as Principles and the Constitutionalization of Private Law, 7 German L.J. 341, 347 (2006).

356Cent. Hudson Gas & Elec. Corp. v. Pub. Serv. Comm’n, 447 U.S. 557, 566 (1980).

357 See Chaplinsky v. New Hampshire, 315 U.S. 568 (1942).

358 See Roth v. United States, 354 U.S. 476 (1957).

359 See New York v. Ferber, 458 U.S. 747 (1982).

360 See 44 Liquormart, Inc. v. Rhode Island, 517 U.S. 484 (1996).

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

362 See Aleinikoff, supra note 1, at 963–71.

363 See Joseph Raz, The Morality of Freedom 38–69 (1986).

364 See Richard H. Pildes, Avoiding Balancing: The Role of Exclusionary Reasons in Constitutional Law, 45 Hastings L.J. 711, 713 (1994). Pildes demonstrates that defining excluded reasons was a dominant technique for adjudicating rights claims in the constitutional jurisprudence of the nineteenth century, id. at 712–13, and Mattias Kumm, supra note 34, at 144–45, emphasizes its role in contemporary constitutional analysis.

365 See Kumm, supra note 34, at 144–45.

366 Id. at 145.

367Id.

368 Id.

369 See Víctor Ferreres Comella, Constitutional Courts and Democratic Values 36–54 (2009).

370 Id. at 66–67; Stone Sweet, supra note 324, at 343–46.

371 See Comella, supra note 369, at 72–73.

372In such systems, the classic distinctions made between the “judicial function” (dispute resolution under the law) and the “legislative function” (the elaboration of law) break down. In these systems, traditional notions of separation of powers become weak sources of systemic legitimacy. Instead, the legitimacy of the constitutional order is critically linked to the constitutional court’s capacities to defend rights.

373Prominent contemporary defenders of the “departmentalist” position—that neither the Founders nor the American Constitution meant to grant the judiciary supremacy with respect to constitutional interpretation—include Akhil Amar and Larry Kramer. See Akhil Reed Amar, America’s Constitution (2005); Larry D. Kramer, The People Themselves: Popular Constitutionalism and Judicial Review (2004).

3745 U.S. (1 Cranch) 137 (1803).

375Id.

376Lochner v. New York, 198 U.S. 45 (1905).

377The most influential statements of the problem and (partial) justification of judicial supremacy remain, respectively, Alexander M. Bickel, The Least Dangerous Branch (1962), and John Hart Ely, Democracy and Distrust (1980).

378Kumm, supra note 34, at 131.

379 For a careful discussion of this issue, see Mattias Kumm, The Idea of Socratic Contestation and the Right to Justification: The Point of Rights-Based Proportionality Review, 4 Law & Ethics Hum. Rts. 140, 144–52 (2010).

380Jed Rubenfeld argues that, due to this fact, the American and European conceptions of constitutionalism, rights, and democracy are fundamentally at odds with one another. Jed Rubenfeld, The Two World Orders, Wilson Q., Autumn 2003, at 22.

381 See Martin Shapiro & Alec Stone Sweet, On Law, Politics, and Judicialization 347–75 (2002).

382 See Stone Sweet & Mathews, supra note 5, at 80–97.

383 Id. at 163 (“This flexibility, which we count as a virtue rather [than] a vice of PA, is never immune from attack by those who believe that a more determinate and principled approach to rights adjudication is possible, or that PA is just a fancy way to package judicial policy-making.”).

384 Id. at 163–64.

385We close on a note of clarification. Though we believe that PA, as a doctrinal framework for adjudicating rights claims, performs better overall than any known competitor, we do not argue that it solves all of the legitimacy dilemmas faced by courts, including the problem of judicial law making. As we wrote:[T]he key to the political success of PA—its social logic—is that it provides a set of relatively stable, off-the-shelf, solutions to a set of generic dilemmas faced by the constitutional judge. If PA mitigates certain legitimacy problems, it also creates, or at least spotlights, an intractable, second-order, problem. PA does not camouflage judicial lawmaking. Properly employed, it requires courts to acknowledge and defend—honestly and openly—the policy choices that they make when they make constitutional choices. Proportionality is not a magic wand that judges wave to make all of the political dilemmas of rights review disappear. Indeed, waving it will expose rights adjudication for what it is: constitutionally-based lawmaking. Nonetheless, . . . PA offers the best position currently available for judges seeking to rationalize and defend rights review, given certain strategic considerations, the structure of modern rights provisions, and the precepts of contemporary constitutionalism.Id. at 77.