Harm-Avoider Constitutionalism

Harm-Avoider Constitutionalism

How does the Supreme Court decide difficult questions of constitutional law? Standard accounts point to a range of interpretive approaches such as originalism, common law constitutionalism, political process theory, interest-balancing, and constitutional pluralism. And once the list of commonly used interpretive approaches is set, normative debates often follow over which is best.

In this Article, I argue that another theory belongs on the list. In a surprising number of cases spanning a range of doctrinal areas—such as Congress’s Article I power, equal protection, substantive due process, presidential immunity, and the dormant Commerce Clause—the Court has decided hard constitutional questions using a kind of argument that has evaded scholarly attention thus far. Rather than relying on original meaning, precedent, or other common tools for discerning the Constitution’s proper application, the Court has decided these cases based on a raw, second-order consideration: which group, if the Court rules against it, would be better able to avoid the harm it would suffer using public and private means? And in each case, the Supreme Court has consciously ruled against the best harm-avoider, trusting in that group’s superior ability to protect its interests outside the courts. I call this approach “harm-avoider constitutionalism.”

This Article explains how harm-avoider constitutionalism is a distinct—and powerful—theory by which the Supreme Court decides hard cases. To be sure, the theory does not help in every case. But where it does, harm-avoider constitutionalism furthers significant virtues. It curbs judicial partisanship. It bolsters the Court’s legitimacy by ensuring that losing groups have more effective responses to their defeats than attacking the Court. It encourages litigants to identify solutions rather than belittle their opponents. And it enables the Court to pursue an important yet underlooked objective in hard cases: to do the least harm possible.

 

 

Introduction

How does the Supreme Court actually decide difficult questions of constitutional law? Standard accounts point to a range of interpretive approaches such as originalism,[1] common law constitutionalism,[2] political process theory,[3] interest-balancing,[4] and constitutional pluralism.[5] At some level, each of these accounts is accurate because each finds support in the Court’s case law. Look around and originalist opinions are plentiful.[6] So, too, are common law constitutionalist, political process, interest-balancing, and pluralist ones.[7] And once the list of commonly used interpretive approaches is set, normative debates can follow over which is best.[8]

In this Article, I argue that another constitutional theory belongs on the list. In a surprising number of cases, the Court has decided hard constitutional questions using a kind of argument that has evaded scholarly attention thus far. Rather than relying on original meaning, precedent, or other common tools for discerning the Constitution’s proper application, the Court has ruled based on a raw, second-order consideration:[9] which group, if the Court rules against it, would be better able to avoid the harm it would suffer? I call this approach “harm-avoider constitutionalism.”

To see harm-avoider constitutionalism in action, one needn’t look far. Three cases from the Supreme Court’s 2019–20 Term show the theory in operation. Start with Trump v. Mazars USA, LLP.[10] The case concerned subpoenas issued by three House committees to President Trump’s bank and accounting firm seeking private financial records concerning the President, his children, and affiliated businesses.[11] The President challenged the subpoenas, arguing that the House lacked Article I power to issue them.[12] Several lower courts thought otherwise, reasoning that the subpoenas were sufficiently related to Congress’s interest in drafting legislation.[13] But the Supreme Court ruled in the President’s favor, remanding for the subpoenas to be reevaluated under a more demanding standard.[14]

How, as a matter of constitutional law, did the Mazars Court reach its conclusion? The answer cannot be found in the usual list of interpretive theories. The Court did not rest its ruling on the Constitution’s text or original meaning,[15] declined to find the case settled by precedent,[16] and made no reference to interest balancing or evolving social values. Instead, the Court ruled against the House because it—not the President—had easier options for avoiding the harms of a judicial defeat.

To see how, focus on the constitutional standard that Mazars actually announced. “[I]n assessing whether a subpoena directed at the President’s personal information” is permissible, the Court declared, certain “special considerations” must be analyzed.[17] First, “courts should carefully assess whether the asserted legislative purpose warrants the significant step of involving the President and his papers.”[18] “Congress may not [subpoena] the President’s information,” the Court cautioned, “if other sources could reasonably provide Congress the information it needs.”[19] Second, presidential subpoenas may be “no broader than reasonably necessary to support Congress’s legislative objective.”[20] And third, Congress must provide “detailed and substantial” evidence of an actual legislative purpose.[21]

Each consideration aims to smoke out whether Congress has an easy way to avoid harm to its legislative process if a given subpoena is blocked. Congress does not need the particular subpoena it has directed at the President, after all, if it can just get the information it needs somewhere else (the first factor), if it can send a more narrowly tailored request to the President (the second factor), or if it has no valid legislative aim to start with (the third factor).[22] Where these easy harm-avoidance strategies are available, Mazars holds, Congress should lose. By contrast, the Court found that presidents have limited options for avoiding their harm: harassment by a co‑equal branch.[23] Harm-avoider reasoning is thus the backbone of Mazars’ constitutional rule.

The Supreme Court relied on similar reasoning in Trump v. Vance.[24] Vance concerned a subpoena nearly identical to those at issue in Mazars, only this time a New York state prosecutor issued it as part of a criminal investigation.[25] The President sued, arguing that he should be absolutely immune from state criminal process.[26] Unlike in Mazars, however, the Court ruled against him and allowed the subpoena to proceed.[27]

Again, the usual interpretive tools offer no explanation. The majority did not ground its answer in originalism.[28] Precedent was also inconclusive given that Vance involved a claim of presidential immunity from state court criminal process “for the first time.”[29] Other modes of argument, such as interest balancing, were also conspicuously absent.[30]

Instead, harm-avoider reasoning pervaded the Court’s analysis. First, the Court recognized the harms that a ruling in either direction would cause. A ruling against the President would implicate three kinds of harms from compelled compliance with state court subpoenas: “diversion, stigma, and harassment.”[31] A ruling against New York, on the other hand, would undercut the “public interest in fair and effective law enforcement” by denying state prosecutors access to relevant evidence.[32]

Second, the Court examined whether the President or New York would be better able to avoid the harms of an adverse decision. It concluded that permitting state court subpoenas would hardly “leave Presidents with ‘no real protection.’”[33] For one thing, any stigma associated with presidential subpoenas would be mitigated by state law grand jury secrecy rules.[34] For another, presidents can avoid harassment and diversion via the “same protections available to every other citizen,” including state law challenges to subpoenas issued in bad faith or overbreadth.[35] Finally, the Court reasoned that presidents can also sue directly under the Supremacy Clause if a particular subpoena unconstitutionally “attempt[s] to influence the performance of [the President’s] official duties.”[36]

The Court then explained that New York lacked similarly effective options for avoiding its harm. It examined one alternative in particular: the retention of subpoenaed evidence for use after the President’s term. But even that, the Court reasoned, would deprive the State of “investigative leads that the evidence might yield,” denying innocent persons access to “exculpatory evidence” and allowing guilty parties to escape while “applicable statutes of limitations” lapse.[37] The President was thus the better harm avoider, and so the Court ruled against him.

Finally, consider June Medical Services L.L.C. v. Russo.[38] In that case, the Supreme Court struck down Louisiana’s Act 620, which required abortion providers to have admitting privileges at a hospital within thirty miles.[39] Like in Mazars and Vance, conventional interpretive theories do not explain the Court’s decision. The plurality did not attempt to ground its ruling in the Due Process Clause’s original meaning.[40] Nor did the plurality rest its conclusion solely on precedent[41] or interest balancing.[42]

What, then, did the case turn on? An assessment of which side—abortion providers and their patients, or the State—would be better able to avoid the harms of an adverse decision. The plurality began by considering whether abortion providers would be able to avoid the harm of an adverse ruling (the closure of their clinics and reduced access to abortion) by redoubling their efforts to comply with the law’s admitting privileges requirement.[43]  In a remarkably fact-intensive analysis spanning twelve pages—roughly 30 percent of the entire plurality opinion—the plurality explained why several doctors lacked any feasible way to do so.[44] The plurality then analyzed whether women seeking abortions might themselves be able to avoid the harm of an adverse ruling.[45] It rejected that possibility too, noting that many women would be unable to “spend nearly 20 hours driving back and forth to [the remaining clinic]” and that, in any event, the clinic would only be able to serve 30 percent of the “10,000 women annually who seek abortions in the State.”[46]

By contrast, the plurality reasoned that a ruling against the State would lead to harm that is far easier to avoid. The plurality recognized the State’s valid interest in protecting maternal health.[47] But after reviewing the trial court record, the plurality was convinced that the State had a readily available alternative for protecting that interest: eliminating the admitting privileges altogether.[48] Summing up the point, the plurality noted the absence of any evidence that abortion “patients have better outcomes when their physicians have admitting privileges.”[49] And because the State could resume implementing its medical regulations as they existed before Act 620 more easily than abortion providers and women could protect against their harms, the Court ruled against the State.

Mazars, Vance, and June Medical are not outliers. Although it has escaped scholarly attention so far, the decisions fit into a rich tradition of cases in which the Supreme Court has resolved difficult constitutional disputes by identifying and ruling against the best harm avoider. This tradition is not only lengthy but trans-substantive: the Court has used harm-avoider reasoning in a wide array of doctrinal areas such as Congress’s Article I power, presidential immunity, substantive due process, equal protection, and the dormant Commerce Clause.[50]

This Article’s foundational objective is to provide a descriptive account of this tradition, thereby laying the groundwork for a positive claim about its place in our law. By tracing the prevalence of harm-avoider reasoning in the Supreme Court’s own decisions, I aim to show that harm-avoider constitutionalism has a reasonable case for inclusion in our culture of constitutional interpretation. The question, in other words, is not whether the Court can decide hard constitutional cases by ruling against the better harm avoider. It already does. The question is whether it should.

The Article’s second objective is to offer an initial assessment of this normative question. The theory is not a panacea. It cannot deliver clear answers in every hard case, and it shouldn’t even apply in cases that can be resolved via the traditional tools of constitutional interpretation—cases that are not “hard” to begin with.[51] But in cases where the law “runs out,”[52] the theory serves four important virtues. First, harm-avoider constitutionalism reduces the harms of Supreme Court error by ensuring that losers have meaningful alternatives for safeguarding their interests. In this respect, the theory advances what may be thought of as a “least harm principle” for deciding difficult cases because optimizing the losing side’s ability to respond to defeat allows the Court to minimize the harm ultimately inflicted by its rulings.[53] Second, harm-avoider constitutionalism can mitigate concerns of judicial partisanship insofar as it leads to outcomes that are not ideologically predictable. Third, the theory can bolster the Court’s institutional legitimacy by offering losing groups ways to respond to defeat that are more effective than assailing the Court itself. And finally, harm-avoider constitutionalism can create incentives for a more generative mode of constitutional argument.[54]

The Article unfolds in four parts. Part I presents the positive case by showing that the approach used in Mazars, Vance, and June Medical is not some one-off (or three-off, as it were) outlier. Far from it. Harm-avoider decision rules pervade the judicial landscape. Part I starts by demonstrating how harm-avoider reasoning has its roots in private law judicial decisions. Part I then shows the prevalence of harm-avoider reasoning in an array of constitutional cases. Taken together, these decisions reveal how, in cases where original meaning, precedent, and other traditional modes of constitutional argument do not clearly determine the outcome, the Court often rules by identifying which group would be better able to avoid the harm it would suffer from an adverse decision. The Court does not always choose consciously to rule against the best harm avoider when confronted with difficult constitutional questions, to be sure.[55] Part I does not even claim that the Court typically does. But look closely enough, and harm-avoider reasoning emerges as a regular feature of our constitutional practice. That practice warrants the academy’s attention and scrutiny.

To engage in this scrutiny, however, first requires some groundwork to be laid. Despite using harm-avoider logic, the Court’s harm-avoider rulings to date have not utilized an express framework. The Court has instead engaged in harm-avoider reasoning on an ad hoc basis. To have a clear target for analysis, Part II develops a conceptual framework for how harm-avoider constitutionalism might function as a general theory of decision-making. It does so inductively, using the Court’s many harm-avoider opinions to tease out important lessons—as well as difficult questions—about the theory’s application.[56]

Armed with a clearer picture of how harm-avoider constitutionalism could be employed as a general framework, Part III presents the normative case. Harm-avoider constitutionalism is desirable, I argue, because it minimizes the harms of Supreme Court error, reduces polarization, enhances public trust in the Court, and has the potential to transform our polarizing culture of constitutional argument in new and constructive directions. Part IV considers some objections, including important arguments raised by Professors Charles Barzun and Mike Gilbert in a recent paper that proposes a closely related “conflict avoidance principle” for hard constitutional cases.[57] A conclusion follows.

I. The Positive Case: Harm-Avoider Reasoning’s Ubiquity in Law

Standard accounts of the Supreme Court’s constitutional practice predict that the Court should decide cases by examining a set of traditionally accepted modes of argument. Those modes include arguments over the Constitution’s original public meaning, structure, and purpose; the proper application of precedent; the balancing of competing interests; and more.[58] One might label these traditional arguments “first-order” interpretive theories[59] in the sense that they assist the Supreme Court in the ordinary task of ascertaining the Constitution’s “correct” operation as it applies on a given set of facts.

Yet first-order interpretive theories are sometimes unable to produce clear answers when the Court confronts hard cases.[60] Where that is so, first-order decision-making can entail significant information burdens and risk costly errors.[61] Ordinary decision-makers in such situations often seek refuge in second-order considerations—i.e., strategies that “reduc[e] the problems associated with making a first-order decision.”[62] The Supreme Court is no different. When faced with hard constitutional questions, the Supreme Court sometimes decides based on a very particular second-order consideration:[63] which group would be best able to avoid the harm of an adverse ruling.

Part I proceeds in two sections. Part I.A begins by showing how this approach has its roots in a private law tradition where American courts routinely use harm-avoider arguments to decide difficult questions of contract, tort, and property law. Part I.B then reveals how the Supreme Court uses harm-avoider reasoning to decide close questions of constitutional law.[64] Taken together, these Sections demonstrate that harm-avoider constitutionalism is not some mere esoteric theory. Indeed, it is already a regular part of our law.

A. Private Law

Any attempt to establish harm-avoider reasoning’s pedigree must start with its place in private law. At least since Abram Chayes’s influential article, The Role of the Judge in Public Law Litigation,[65] constitutional law scholars have had reason to remind themselves of the relative novelty of modern constitutional litigation in the broader landscape of what courts do. Constitutional cases are not the only game in court; in fact, they aren’t even the norm. As Chayes wrote, “[i]n our received tradition, the lawsuit is a vehicle for settling disputes between private parties about private rights.”[66]

Courts engage with private law disputes differently than public law disputes.[67] Notably, one such difference is how they approach uncertainty about the law. Rather than evincing a singular obsession with getting the law “right” in these situations, private law courts often think about how the parties might have prevented the harm in the first place. Which party, these courts often ask, was best situated to avoid the harm in question?

Start with contract law. The leading example is the rule of interpretation against the draftsman, or contra proferentem. The Restatement of Contracts describes this rule as follows: “In choosing among the reasonable meanings of a [contract], that meaning is generally preferred which operates against the party who supplies the words or from whom a writing otherwise proceeds.”[68] As Justice Gorsuch recently observed, we utilize this rule to “resolve contractual ambiguities” because “the drafter might have avoided the dispute by picking clearer terms.”[69] In other words, rather than fixating on the search for the single “right” meaning of an ambiguous provision, contra proferentem instructs courts to rule against the best harm avoider.

The Supreme Court first relied on contra proferentem’s harm-avoider reasoning in an 1806 case called Manella, Pujals, & Co. v. Barry.[70] There, Spanish plaintiffs alleged that an American defendant failed to follow the terms of the parties’ contract, resulting in the loss of a valuable shipment of cargo.[71] The defendant responded that because “any ambiguity” in the contract “was the fault of the plaintiffs” who had drafted it, “[t]he defendant, at 3,000 miles distance, could not consult them, and [thus] cannot be chargeable for an error.”[72] The plaintiffs, in other words, were the better harm avoider because they could have written the contract more clearly in the first place. Contra proferentem instructed that the plaintiffs should lose the case—and the Supreme Court agreed.[73]

Harm-avoider reasoning is prevalent in tort law, too. In The Costs of Accidents, Judge Guido Calabresi famously argued that the costs of automobile accidents could be minimized by “requir[ing] allocation of accident costs to those acts or activities . . . which could avoid the accident costs most cheaply.”[74]

Consider a well-known example. Suppose you are driving your car in a rush to get to an appointment. The car in front of you slows suddenly, and you cannot stop your vehicle in time. Who should be liable for the ensuing crash? As all graduates of drivers education classes ought to know, the general rule is that a driver who collides with the rear of a stopping vehicle will be held liable for the accident.[75] This is true even if the lead vehicle driver may be at some fault—for instance, if they fail to use a signal before they stop.[76]

Why is this the rule? Rather than engaging in the difficult task of determining who is at greater fault in any given accident, tort law imposes liability on the trailing driver because they are the easier (i.e., better) harm avoider.[77] Whereas trailing drivers can generally avoid collisions with ease by leaving ample distance to the vehicle in front of them, lead drivers may not be able to anticipate events that require them to stop suddenly (such as a deer leaping into the road).[78] Structuring liability in this way incentivizes drivers to trail at greater distances, maximizing the odds that all parties are better off because no accident happens in the first place.[79]

In sum, harm-avoider decision rules are common across private law fields. Their wisdom lies in the instinct that it is often better to structure liability in a way that encourages parties to avoid their harms altogether than to engage in a single-minded pursuit of the “right” outcome amidst legal and factual uncertainty. Indeed, this approach to legal decision-making is a staple of the law and economics movement.[80] Prominent scholars have argued, however, that for all of the advances this movement has brought to private law, law and economics has had little influence on constitutional law, and on constitutional interpretation in particular.[81]

The next Section suggests that this may be too quick a conclusion. It does so by presenting a systematic account of the Court’s harm-avoider reasoning across a significant number of constitutional cases. After closer inspection, it turns out, harm-avoider decision rules have already worked their way from private law to the Supreme Court’s constitutional jurisprudence.

B. Constitutional Law

The Supreme Court regularly utilizes harm-avoider reasoning to decide difficult constitutional questions. In several cases spanning diverse doctrinal areas—substantive due process, presidential immunity, equal protection, the dormant Commerce Clause, and Congress’s Article I power—the Supreme Court has decided hard cases based on which group would be better able to avoid the harm of an adverse decision.

This Section begins by describing a series of older, high-profile decisions that rely on harm-avoider logic. It then returns to the trio of harm-avoider rulings issued in the summer of 2020: Mazars, Vance, and June Medical. Under rudimentary precepts of legal positivism, this prolonged and trans-substantive practice matters when it comes to identifying the modes of constitutional interpretation that constitute a valid part of our law.[82] As Professor Jamal Greene has put it, “[i]dentifying a trend within [the Supreme Court’s] practices places a heavy burden on anyone who would argue against the trend’s significance within [our constitutional] culture.”[83]

One final caveat bears mentioning. In thinking about the cases canvassed below, readers may question the Supreme Court’s final judgment as to which group is the better harm avoider in this or that case (and thus which group should lose). This is a fair reaction. For present purposes, however, I do not seek to defend the particular outcomes of the Court’s harm-avoider opinions.[84] My objective here is instead to show that the Court engages in harm-avoider analysis at all. If that is true, then harm-avoider constitutionalism has a reasonable claim to being part of our law.

1. Older Cases

The Supreme Court’s use of harm-avoider reasoning in constitutional cases dates back more than four decades and encompasses a range of doctrinal fields. This Section canvasses these earlier rulings in the fields of substantive due process, presidential immunity, equal protection, and the dormant Commerce Clause.

a. Substantive Due Process & the Right to Die

Nancy Cruzan was twenty-five years old when she lost control of her car and suffered permanent brain damage in a horrific crash.[85] Rehabilitative efforts failed, and Nancy was left in a persistent vegetative state.[86] Unable to eat or drink on her own, Nancy laid unmoving in a Missouri hospital bed, connected to a feeding and hydration tube.[87] After it became clear that Nancy had “virtually no chance of regaining her mental faculties,” Nancy’s parents asked the hospital to remove the treatment.[88] The hospital refused, and Nancy’s parents went to court.[89] A Missouri trial court authorized the removal of treatment based on testimony that Nancy had once told a roommate she would not want to “face life as a ‘vegetable.’”[90] The Missouri Supreme Court reversed, finding this testimony insufficient to satisfy the clear and convincing evidence standard it deemed necessary to protect the State’s interest in preserving human life.[91]

Seven years after Nancy’s tragic car accident, the U.S. Supreme Court decided Cruzan v. Director, Missouri Department of Health. By then, the case had distilled to the following question: does the Due Process Clause of the Fourteenth Amendment forbid Missouri to condition removal of Nancy’s life‑sustaining treatment on the requirement to prove her desire by clear and convincing evidence?[92] In a 5-4 decision, the Supreme Court answered in the negative.[93] Significantly, the Court did not claim that this outcome was necessary because the Constitution’s original meaning, history, or even its own substantive due process precedents commanded it.[94] Instead, both the majority and Justice O’Connor’s concurring opinion used harm-avoider reasoning. If we err in ruling in either direction, the majority and Justice O’Connor asked, which side would be better able to avoid the resulting harm?

Consider the Court’s analysis of a ruling in Cruzan’s favor. The majority observed that striking down Missouri’s clear and convincing evidence standard would raise the risk of harms that are impossible to avoid.[95] After all, the choice to withdraw treatment is “final and irrevocable” once the patient dies, such that “[a]n erroneous decision to withdraw life-sustaining treatment . . . is not susceptible of correction.”[96] Ruling for Cruzan, in other words, would completely disable the State from protecting its interest in human life. And that would be so even if new evidence were to materialize indicating that a deceased patient actually wished to continue treatment all along, or new medical advances were made that would have enabled the patient to resume some cognitive function.

By contrast, the Court recognized that a ruling against Cruzan would lead to harms that can “eventually be corrected or . . . mitigated.”[97] In other words, allowing Missouri to enforce its clear and convincing evidence standard would increase the risk of “erroneous decision[s] not to terminate,” harming persons in Nancy’s position as well as her parents.[98] But these harms are neither final nor irrevocable. The majority recognized, for example, that loved ones such as Nancy’s parents may “discover[] new evidence regarding the patient’s intent” to terminate treatment, or wait for “changes in the law,” such as a state’s adoption of a lower, preponderance of evidence standard for withdrawing medical treatment.[99] Either of these developments would enable them to return to court seeking an order withdrawing treatment.[100] The majority also mentioned two other ways the harms of an adverse ruling might be avoided: “[A]dvancements in medical science,” by which the Court presumably meant the possibility of restoring a patient’s cognitive function, or “the unexpected death of the patient despite the administration of life-sustaining treatment.”[101]

Justice O’Connor joined the majority opinion but wrote separately to emphasize another strategy by which terminally ill patients and their loved ones may avoid the harm of an adverse ruling: the use of state statutes authorizing surrogate medical decision-making.[102] In Justice O’Connor’s view, the fact that individuals can designate legally recognized proxies to make end-of-life decisions on their behalf, thereby avoiding disputes like Nancy’s in the future, was a powerful reason to uphold Missouri’s clear and convincing evidence standard. “[P]rocedures for surrogate decisionmaking, which appear to be rapidly gaining in acceptance,” she wrote, are “a valuable additional safeguard of the patient’s interest.”[103]

In sum, the Court’s decision to uphold Missouri’s clear and convincing evidence standard was not predicated on the view that the Constitution’s text, history, or judicial precedent somehow required it.[104] Instead, the Court candidly recognized that striking down Missouri’s heightened evidentiary standard would raise the risk of harm that cannot be corrected—namely, the erroneous termination of life-saving treatment. Upholding the standard, by contrast, would lead to harm that is more easily avoided, both in Nancy’s own case (as the majority argued) and in all similar cases in the future (as Justice O’Connor observed). And so the Court took the latter path.

b. Presidential Immunity

A pair of famed presidential immunity cases also relied on harm-avoider reasoning: Nixon v. Fitzgerald and The Nixon Tapes Case. Much ink has been spilled about these opinions, and the common understanding is that the cases employed a kind of naked balancing inquiry that weighed generic, competing interests such as separation of powers and the needs of justice.[105] I do not dispute this. What I want to suggest is that the Supreme Court did not rest solely on balancing the interests on each side of the dispute. The Court also considered each side’s ability to protect its interests outside the Court, assuming it were to lose inside the Court. In other words, the Court identified—and ruled against—the better harm avoider.

Nixon v. Fitzgerald. In 1968, a Department of Defense engineer named A. Ernest Fitzgerald—a man later dubbed the “patron saint of government whistleblowers”[106]—dropped a bombshell in testimony before Congress: the Pentagon’s vaunted C‑5A aircraft program had overrun its budget by as much as $2 billion.[107] Fitzgerald lost his job and sued. Among other claims, Fitzgerald alleged that President Nixon had fired him in retaliation for his testimony in violation of the First Amendment and federal statutes.[108]

The question before the Supreme Court was whether Fitzgerald’s suit against Nixon should be dismissed on the ground that presidents enjoy a constitutionally rooted absolute immunity from lawsuits based on their official acts.[109] The Court ruled in President Nixon’s favor, 5-4.[110] Just as in Cruzan, however, the Court did not purport to locate this answer in anything close to a straightforward command from the Constitution, history, or precedent. The Court even admitted “the absence of explicit constitutional . . . guidance,” and the need to rule accordingly based on other sources including, with surprising candor, “concerns of public policy.”[111]

The Court relied, in substantial part, on harm-avoider reasoning to reach its conclusion. To start, the Court was clear eyed about the harm that a ruling against Fitzgerald would create: the risk of placing the President “above the law.”[112] Rather than downplaying the gravity of that harm, however, the Court pointed to strategies for avoiding it—or, in the Court’s words, the ready “existence of alternative remedies and deterrents” through which the people could secure “sufficient protection against misconduct on the part of the Chief Executive.”[113] “There remains the constitutional remedy of impeachment,” the Court noted, in addition to “scrutiny by the press,” and “[v]igilant oversight by Congress.”[114] There are also “[o]ther incentives” that render the President responsive to the public, such as the President’s “desire to earn reelection,” “maintain prestige,” and protect his “historical stature.”[115] The Court thus recognized that a ruling against Fitzgerald would implicate the risk of an unaccountable President. But it was persuaded by the fact that concerned citizens would retain other means to rein a President in.

By contrast, the Court expressed concern that a ruling against the President would inflict harms that could not be avoided. The President, after all, “must make the most sensitive and far-reaching decisions entrusted to any official under our constitutional system.”[116] Exposing the President to liability for those decisions would “render [him] unduly cautious in the discharge of his official duties.”[117] The result would be an enfeebled Chief Executive whose concerns about his “personal vulnerability” could distract him from official duties, producing profound “detriment” to “the Nation that the Presidency was designed to serve.”[118] Critically, the Court worried that if it opened the door to personal liability, there would be little way to seal it. In fact, the Court feared that the harms of imposing liability could increase over time given “the visibility of [the] office,” and the “effect of [the President’s] actions on countless people,” both of which make “the President . . . an easily identifiable target for suits.”[119] And because the President lacked effective strategies for avoiding this harm, the Court ruled in his favor.

The Nixon Tapes Case. In 1972, five men were caught breaking into the Democratic National Committee’s office at the Watergate Office Building in Washington, D.C.[120] Connected with the Nixon administration, the men attempted to eavesdrop on the President’s political opponents.[121] Two years later, a grand jury indicted seven of Nixon’s aides for offenses arising out of the break-in; the President was named as an unindicted co-conspirator.[122] The special prosecutor then subpoenaed a set of private White House tape recordings that implicated the defendants.[123] President Nixon moved to quash the subpoena, arguing that it violated his constitutional right to Executive Privilege.[124]

The Supreme Court ruled unanimously against the President[125]—a ruling that precipitated Nixon’s resignation from office just sixteen days later.[126] The Court’s opinion canvassed a number of issues, but the most significant was whether the President should enjoy a “presumptive privilege” against the subpoena in light of his general concern for protecting the confidentiality of his communications with his advisers.[127] The Court did not purport to locate the answer to this question solely in the Constitution’s original meaning, structure, or in the Court’s own precedent.[128] Instead, the Court resorted to a balancing test, writing: “[W]e must weigh the importance of the general privilege of confidentiality of Presidential communications in performance of the President’s responsibilities against the inroads of such a privilege on the fair administration of criminal justice.”[129]

One could characterize the Court’s ultimate judgment—that the “fair administration of criminal justice” should prevail[130]—as the product of direct balancing of the competing interests. This would be wrong. On closer inspection, the Court’s conclusion is appropriately understood as one predicated principally on harm-avoider reasoning. The Court did not conclude, in other words, that the needs of criminal justice are outright more important than the need for confidential presidential communications. Instead, the Court reasoned that a ruling for the President would impose harms on the criminal justice system that are difficult to avoid, whereas a ruling against him would lead to harms that are more easily avoided.

The Court began by recognizing the gravity of the harm itself: quashing subpoenas in criminal cases based on Executive Privilege would limit evidence and interfere with the justice system’s ability to discover the truth, “gravely impair[ing] the basic function of the courts.”[131] And critically, the Court argued that this harm could not be easily corrected.  The Court observed that “[w]ithout access to specific facts” contained in subpoenaed materials, “a criminal prosecution may be totally frustrated.”[132] Indeed, the Court took pains to emphasize that, in the case at bar, there was a “demonstrated, specific need”[133] for the requested evidence because the special prosecutor had already proven to the trial court’s satisfaction that the subpoenaed tape recordings were “essential to the justice of the [pending criminal] case.”[134] Put differently, ruling for the President would cause unavoidable prejudice to the ongoing criminal proceeding, allowing potentially guilty persons to go free.

A ruling against the President, conversely, would inflict more easily avoided harms. The Court began by displaying full awareness of the significant Article II interest in keeping confidential communications out of the public sphere: “The need for confidentiality even as to idle conversations with associates in which casual reference might be made concerning political leaders within the country or foreign statesmen is too obvious to call for further treatment.”[135] Rather than downplaying this interest, the Court argued that disclosure of many sensitive presidential conversations is preventable through careful in camera review of the subpoenaed material by the district court—a “scrupulous” procedure that should be especially protective in cases implicating the “high degree of respect due the President of the United States.”[136] The harms to the President could be avoided, in other words, so long as courts comply with their “very heavy responsibility” to keep private all “[p]residential conversations, which are either not relevant or not admissible.”[137] And given the President’s greater ability to avoid the harms of an adverse ruling, the Court ruled unanimously to enforce the subpoena.

c. Equal Protection

In Plyler v. Doe, a closely divided Court struck down a Texas law permitting local school districts to deny free enrollment to undocumented immigrant children,[138] holding that it violated the Equal Protection Clause.[139] Harm-avoider constitutionalism helps shed light on this decision, which has eluded explanation under the typical equal protection framework.

Scholars have wrung their hands over how to fit Plyler within equal protection law’s usual tiers of scrutiny.[140] For good reason. The Plyler majority conceded that “[u]ndocumented [persons] cannot be treated as a suspect class” and that “education [is not] a fundamental right”[141]—the two threshold inquiries that could trigger heightened scrutiny. The majority thus suggested that Texas’s law should be reviewed for “rationality.”[142] Yet, in the very next sentence, the majority appeared to change course by declaring that Texas’s law “can hardly be considered rational unless it furthers some substantial goal of the State.”[143]

A harm-avoider framing offers a plausible explanation for the majority opinion’s divergence from settled equal protection doctrine. In evaluating Texas’s law, the majority clearly identified the harms that both sides would suffer should the Court rule against them. And the majority conscientiously analyzed each side’s ability to avoid those harms, ruling against the better harm avoider.

First, the Court recognized the significant “costs involved to [undocumented] children” if it upheld the challenged law, including illiteracy and an increased risk of unemployment and crime.[144] Crucially, the Court emphasized how undocumented children lack the ability to avoid these harms. In one of the opinion’s most searing passages, the Court lamented the powerlessness of the affected children: unlike undocumented parents who “have the ability to . . . remove themselves from the State’s jurisdiction” in response to adverse legislation, “the children who are plaintiffs in these cases ‘can affect neither their parents’ conduct nor their own status.’”[145] Driving the point home still further, the Court criticized Texas’s law because it “imposes its discriminatory burden on the basis of a legal characteristic over which children can have little control.”[146] Put simply, if undocumented children were kicked out of Texas’s public schools, they would be able to do very little about it.[147]

The Court then reasoned that a ruling against the State would produce harms that are easier to avoid. The Court recognized the interest that other Texas parents had in ensuring that their children receive a “high-quality public education.”[148] But the Court credited the district court’s observation that Texas “failed to offer any ‘credible supporting evidence that a proportionately small diminution of the funds spent on each child [which might result from devoting some state funds to the education of the excluded group] will have a grave impact on the quality of education.’”[149] Put another way, the Court’s conclusion implied that concerned parents could protect their interest in a quality public education simply by leaving their children in the public schools. Or, as the Court summarized the trial court evidence, “the record in no way supports the claim that exclusion of undocumented children is likely to improve the overall quality of education in the State.”[150] Because undocumented children lacked any such strategy for avoiding their harms, the Court ruled in their favor.

d. The Dormant Commerce Clause

Harm-avoider reasoning has also figured prominently in the Court’s approach to the dormant Commerce Clause. Minnesota v. Clover Leaf Creamery is a good example.[151] The dispute arose when Minnesota enacted a law forbidding retail sale of milk in nonreturnable and nonrefillable containers.[152] Several companies involved in the milk industry—including a number of Minnesota firms—filed suit, arguing that the law imposed an unconstitutional burden on interstate commerce.[153] Minnesota responded that the law was a permissible effort to protect the environment.[154]

The Court ruled in the State’s favor.[155] But it did not do so based on anything approaching a clear constitutional command. In fact, the Court’s entire dormant Commerce Clause jurisprudence is founded on a “mere negative inference” from the Commerce Clause itself.[156] Instead, as is standard in these cases, the Court balanced “the burden imposed on . . . commerce” against the “putative local benefits” of the challenged law.[157]

While the Court’s opinion can partially be understood as an exercise in interest balancing, the force of that rationale is incomplete. The Court did base its conclusion in part on a direct weighing of the competing interests, concluding that the State’s environmental interests were “substantial,” whereas the burdens on interstate commerce were “relatively minor.”[158] But this logic is lacking in certain respects. In particular, asking whether the State’s environmental interests outweigh burdens on interstate commerce brings to mind Justice Scalia’s quip about the impossibility of “judging whether a particular line is longer than a particular rock is heavy.”[159] In other words, the Court could not simply rest its decision solely on its view that the environmental benefits of Minnesota’s law outweighed the law’s economic harm to the dairy industry.

Instead, the Court also utilized harm-avoider reasoning in its balancing analysis. Regardless of how the balance might be struck between the commercial and environmental interests at stake, the dairy industry was in a better position to avoid its harms than were the State and relevant environmental groups. The Court found it significant that “two of the three dairies, the sole milk retailer, and the sole milk container producer challenging the statute in this litigation are Minnesota firms.”[160] These “major in-state interests,” the Court reasoned, had the capacity to pursue alternative means to remedy any economic harms, most notably by supporting a legislative change within the State.[161] By contrast, the Court found no meaningful alternatives available to the State and concerned environmental groups if the law were invalidated. The suggested alternative was legislation “providing incentives for recycling.”[162] But that, the Court reasoned, was “less likely to be effective” at “promoting conservation of energy and other natural resources and easing solid waste disposal problems” than the outright ban on single-use milk containers.[163] Given the availability of legislative recourse for the Minnesota dairies and firms and the lack thereof for their opponents, the Court ruled in the environmental groups’ favor.[164]

2. The 2019-20 Term

Although the foregoing cases drew on harm-avoider reasoning, they did so in a largely episodic fashion over a period of years. This occasional usage changed in the summer of 2020. Within a period of three weeks, the Court issued a trio of decisions that squarely turned on an assessment of which side, if it lost the case, would be best positioned to avoid its harms.[165]

a. Congress’s Subpoena Power

The legal question in Trump v. Mazars USA, LLP was whether Congress had Article I authority to subpoena a private bank and accounting firm for financial records belonging to their client, President Donald Trump.[166] In actuality, the stakes of the case were far higher, for what was sought were records that might reveal (among other things) if “the President and his associates had been compromised by foreign actors or interest.”[167]

A string of lower federal courts ruled against the President. Congress, those courts reasoned, can subpoena information that serves a “valid legislative purpose”[168]—a test the Court had previously said would be satisfied (for non-presidential subpoenas) so long as the subpoena “concerned a subject on which ‘legislation could be had.’”[169] And because each congressional subpoena for President Trump’s financial records was accompanied by at least a generic reference to possible legislation, the lower courts found them permissible.[170]

The Supreme Court disagreed and vacated the lower court rulings.[171] In doing so, it rejected the “subject on which legislation could be had” standard in the context of subpoenas to the President,[172] remanding for the lower courts to conduct a more “careful analysis.”[173] In practical terms, the result was a victory for the President. Had the Court affirmed the lower court rulings, politically sensitive and potentially damaging information might have been revealed to the public. By vacating and remanding for the lower courts to reevaluate the subpoenas, the Court all but assured that the records would remain private until after the 2020 election.[174]

The political fallout of Mazars is, of course, significant. Equally significant, but much less attended to, are the implications for constitutional law. From that perspective, an important question warrants attention: what interpretive theory did the Court actually use to decide the case? Certainly not originalism; the Court candidly recognized the absence of any “enumerated power” in Article I that might serve as a textual hook for any originalist analysis.[175] Nor did the Court even engage with the original meaning arguments offered by Justice Thomas in dissent.[176] The Court did discuss precedent, but only to reject both the deferential precedential standard used in the lower courts and an especially heightened “demonstrably critical” standard proposed by the President.[177] And though the Court paid close attention to the “longstanding practice” by which previous presidents and Congress had amicably settled their disputes without judicial intervention, that practice was more a thumb on the scale than an actual source of a legal rule.[178]

The case turned instead on arguments over which side—Congress or the President—would have easier options for avoiding the harm of an adverse ruling. To appreciate how, one must focus on the reasoning the Court used to arrive at the constitutional rule that Mazars actually announced. Congressional subpoenas seeking the President’s personal information, the Court explained, require a “careful analysis” of four special considerations.[179] All four are aimed at identifying whether Congress or the President would be the better harm avoider in a particular subpoena dispute.

Mazars’s first three considerations are designed to ferret out whether Congress could easily avoid harm to its legislative efforts if a given subpoena is invalidated. First, courts should “carefully assess whether the asserted legislative purpose warrants the significant step of involving the President and his papers” at all.[180] For “if other sources could reasonably provide Congress the information it needs in light of its particular legislative objective,”[181] then Congress has a simple way to avoid any ill effects after a subpoena is blocked: it can use those other sources, rather than the President.[182]

Second, the Court declared that “courts should insist on a subpoena no broader than reasonably necessary to support Congress’s legislative objective.”[183] This, too, is aimed at identifying alternatives that Congress may use to achieve its legislative goals if a court refuses to enforce a congressional subpoena targeting the President. After all, Congress can avoid any harm to its legislative efforts if it can legislate just as well by using a narrower and less intrusive subpoena.

Third, a congressional subpoena will be deemed impermissible in certain cases concerning the President “unless Congress adequately identifies its aims and explains why the President’s information will advance its consideration of possible legislation.”[184] This is yet another factor that can be understood in terms of Congress’s ability to avoid the harms of an adverse ruling. For, if Congress truly needs the President’s information for a legitimate legislative undertaking but has failed to explain how, it can simply issue a new subpoena with a clearer description as to its legislative purpose and the need to target the President rather than some other source. By contrast, if Congress lacks a legislative goal in the first place or the President’s information is unnecessary to achieve that goal, then no lasting harm would be wrought by refusing to enforce a subpoena.

Mazars’ fourth consideration was a bit more ambiguous. The opinion exhorted courts to carefully “assess the burdens imposed on the President by a subpoena.”[185] But assess them for what?[186] The most sensible answer is to assess harm to the President in the same way Mazars analyzes harm to Congress: courts should consider whether the President has an easy way to avoid a given subpoena’s burdens. Indeed, that is how the Court itself analyzed the burdens inflicted by the very subpoenas at issue. Recall that one distinctive feature of the subpoenas was how they were directed at the President’s bank and accounting firm rather than the President personally.[187] Congress argued that this avoided the President’s burden of compliance, but the Court disagreed: allowing unlimited subpoenas of any presidential information “entrusted to a third party” would allow Congress to “declare open season” on all manners of private medical, financial, and educational records commonly held by third parties.[188] The Court did gesture, however, at other innovative solutions that might do more to mitigate the President’s harm. For example, the Court wrote approvingly of a prior compromise in which Congress permitted President Reagan to make a set of sensitive records “available, but only for one day with no photocopying, minimal notetaking, and no participation by non-Members of Congress.”[189]

In summary, Mazars made new constitutional law in the form of a four-part test for evaluating congressional subpoenas of private presidential information. The Court did not pretend, however, to derive this test from the Constitution’s text or original meaning, the Court’s own precedent, or other conventional interpretive tools. A different kind of consideration did the work: arguments over which side—the President or Congress—if it lost, would be best able to avoid the resulting harm.

b. Presidential Immunity in State Criminal Proceedings

Shortly after Congress issued the subpoenas challenged in Mazars, the New York County District Attorney’s office served a similar subpoena on the President’s accounting firm seeking information for a state criminal investigation.[190] The President sued to enjoin enforcement of the subpoena but lost in the lower courts. The Supreme Court affirmed in Trump v. Vance. [191]

Two aspects of Vance’s analysis warrant close attention. First, just as was true of Mazars, Vance was not grounded in traditional, first-order interpretive tools. The Vance majority engaged in a lengthy historical interlude describing the use of presidential subpoenas in prior criminal proceedings, including most notably the trial of Aaron Burr. The Court’s majority, however, ultimately declined to rest its decision on the “original understanding of the Constitution reflected in [the Burr trial].”[192] Precedent also failed to account for the Court’s ruling. As the Court acknowledged, the issue in Vance differed from those in prior cases involving criminal process directed at the President because it involved, “for the first time,” a state rather than federal court.[193] Other modes of reasoning, such as interest-balancing and historical practice, were absent too.

Vance is instead a decision rooted in straightforward arguments over which side—the President or New York[194]—would be better able to avoid its harm if the Court should rule against it. To perform this assessment, the Court identified the harms that both sides would suffer from defeat. For the President, the Court recognized that permitting enforcement of state criminal subpoenas might implicate “diversion, stigma, and harassment.”[195] For New York, disallowing the subpoena would prejudice the underlying criminal investigation, to the detriment of “fair and effective law enforcement.”[196]

The Court also considered whether either side would have easy strategies for avoiding its harm. As to the President, the Court specifically noted several “safeguards” that would continue to protect the president from diversion, stigma, and harassment even if the Court were to rule against him.[197] To start, New York already has on the books “longstanding rules of grand jury secrecy [that] prevent the very stigma the President anticipates.”[198] State law also affords the President the ability to avoid harassment by challenging subpoenas issued in “bad faith,” out of “malice, or [with] an intent to harass.”[199] Presidents are also entitled to special solicitude when it comes to scheduling, the Court observed, lest a President be diverted from his “ongoing discharge of his official responsibilities.”[200] Indeed, if a President can show that “an order or subpoena would significantly interfere with his efforts to carry out [his public] duties,” a court should “use its inherent authority to quash or modify the subpoena.”[201] And finally, presidents have a special constitutional tool in their arsenal. If a state court uses criminal process to “retaliate against,” “manipulate,” or otherwise harass a President, the President can file suit under the Supremacy Clause.[202] For all these reasons, the Court concluded, a ruling against the President would “not leave [the] President[] with ‘no real protection.’”[203]

The Court then explained why the opposite would be true of a ruling against New York. A loss for the State, the Court reasoned, would “hobble” the grand jury’s “ability to acquire all information” bearing on its investigation.[204] Critically, even the most sensible strategy for mitigating this harm—“preserv[ing]” any withheld evidence for use after “the conclusion of a President’s term”—would be ineffectual.[205] This was true, the Court explained, because “the State would be deprived of investigative leads that the evidence might yield” in the interim, which “could frustrate the identification, investigation, and indictment of third parties (for whom applicable statutes of limitations might lapse).”[206] And “[m]ore troubling, it could prejudice the innocent by depriving the grand jury of exculpatory evidence.”[207]

Thus, Vance, like Mazars, is a decision rooted firmly in arguments over which side is the better harm avoider. In both cases, the Court closely analyzed each side’s options for avoiding the harm of an adverse ruling, expressly recognized the superior options available to one side, and ruled against that side for that very reason. Of course, Mazars and Vance came out in opposite directions: the President won in Mazars and lost in Vance. But what led to these divergent outcomes was not the Constitution’s original meaning, precedent, or some other conventional constitutional argument. A close reading of Mazars and Vance reveals a different ground of distinction: the fact that the House had stronger options for avoiding its harm in Mazars, whereas the President had superior strategies for avoiding his harm in Vance.

c. Substantive Due Process & the Right to Abortion

For students of the Supreme Court’s abortion jurisprudence, it is difficult to read June Medical Services v. Russo without a sense of déjà vu. The Louisiana law at issue, Act 620, was identical to the Texas statute struck down four years earlier in Whole Woman’s Health v. Hellerstedt. Both regulations required doctors who perform abortions to maintain admitting privileges at a hospital within thirty miles.[208] Yet the Fifth Circuit upheld Louisiana’s law in June Medical.[209] Onlookers largely expected the Court to affirm that ruling, but Chief Justice John Roberts provided a surprising fifth vote to strike down Louisiana’s law.[210]

In reading the opinion, one is tempted to view June Medical as resting solely on an “ad hoc,” case-specific, interest-balancing approach to constitutional law.[211] After all, the plurality applied an undue burden standard that it described as requiring consideration of “the burdens a law imposes on abortion access together with the benefits those laws confer.”[212] And in applying that standard, the plurality concluded that while Act 620’s burden on abortion access was high, its benefits were low (indeed, non-existent) on the facts of the case. With respect to burdens, the plurality observed that Act 620 would result in four of the five abortion doctors in the state closing their practices.[213] That, in turn, would leave a single doctor (“Doe 5”) with a single clinic able to accommodate at most 30 percent of the roughly 10,000 abortions sought in the State each year. And any woman in the northern part of the State would need to travel a great distance to reach Doe 5’s clinic.[214] Those burdens, the plurality observed, outweighed the benefits offered by the law, insofar as “the State introduced no evidence . . . of any instance in which an admitting privileges requirement would have helped even one woman obtain better treatment.”[215]

But this reading of June Medical as a straightforward exercise in interest-balancing leaves unresolved a serious puzzle. For, as a matter of logic, a balancing approach alone does not explain the most curious thing about the plurality opinion and Justice Alito’s principal dissent: the incredible degree to which the two opinions argue over whether three particular doctors (Does 2, 5, and 6) might have tried even harder to obtain admitting privileges.

It is hard to overstate the degree to which these doctors’ respective efforts to obtain privileges was the focus of the dueling opinions. Combined, the plurality and principal dissent mentioned the three doctors 152 times—nearly twice as many times as the combined number of citations to Casey and Hellerstedt.[216] The two opinions also spend twenty-one pages and over 6,400 words debating whether the doctors could do more to obtain privileges. [217] For example, the opinions quarreled over Doe 2’s effort to obtain privileges at, among other places, the Willis Knighton Health Center. This effort stalled when the hospital requested data regarding Doe 2’s recent hospital admissions.[218] The plurality credited Doe 2’s testimony that responding to that request would have been pointless, since he hadn’t done any in-hospital work in ten years, precisely because abortions are so safe.[219] Justice Alito responded that Doe 2 should have requested “courtesy privileges” instead of active privileges at the hospital.[220] But when confronted with evidence that Doe 2 did exactly that, Justice Alito faulted him for doing so in a mere “three-paragraph e-mail” that “subsequent correspondence from [the hospital did] not acknowledge.”[221] Similarly, Justice Alito criticized Doe 5 for “not even call[ing] back to check on” applications he had sent to two hospitals, neither of which had responded to him in the first place.[222]

If June Medical rested solely on a balancing of Act 620’s burdens against its benefits, all of this debate would seem quite beside the point. After all, no one disputed that for “over a year and a half,” the doctors applied for and failed to obtain privileges from a total of thirteen hospitals, all under the district court’s supervision.[223] Act 620’s burdens were thus clear: the law would result in the reduction of in-state abortion providers from five to one.[224] Why should it matter, then, whether Doe 2 might have sent another email to follow up, or whether Doe 5 might have called to confirm that his applications had, in fact, been rejected?[225]

Put simply, the Court’s examination of the doctors’ ability to try harder to obtain privileges got at something quite different than a balancing of Act 620’s burdens and benefits. It identified whether a ruling against the doctors would produce harms that are easily avoidable. That would be true, for instance, if the doctors could respond to an adverse ruling simply by firing off a few emails to secure the necessary privileges. Of course, the plurality did not believe it was true. In concluding that any additional efforts by Does 2, 5, and 6 would be “an exercise in futility,”[226] the plurality held that the doctors lacked any readily available means to avoid the harm of an adverse ruling.

But the plurality did not stop its harm-avoider assessment there. It also asked whether “other doctors”—that is, “new abortion providers”—might “eventually replace” the various Does whom Act 620 would drive out of business.[227] The answer was no: the same barriers that stopped Does 2, 5, and 6 from obtaining privileges would likely block new providers from doing the same.[228] The plurality then asked whether women themselves might be able to avoid the harm of an adverse ruling by traveling to the State’s remaining clinic.[229] Again, the Court concluded no, as the remaining clinic would be able to accommodate a fraction of patients, at great difficulty and distance for many women.[230]

Finding no easy options for avoiding the harms of a ruling against the doctors, the plurality then considered whether a ruling against the State would lead to unavoidable harm. The plurality recognized the State’s valid interest in preventing harm to women’s health.[231] But it found a simple alternative by which the State could prevent that harm. The State could resume enforcing applicable medical regulations as they existed before Act 620. In other words, as the plurality put it, Act 620 “makes no improvement to women’s health ‘compared to prior law.’”[232] And because the State could more easily avoid harm to its underlying interests simply by enforcing existing law, the State lost the case.[233]

*  *  *

Taken together, the above cases paint what I hope is the following picture. When confronted with difficult constitutional questions where the usual, first-order modes of argument are indeterminate, the Court often relies on second-order arguments about which group, if it loses, would be best able to avoid the resulting harm. These arguments are not limited to a single doctrinal area in constitutional law. They appear in a wide range of decisions concerning Congress’s Article I power, substantive due process, presidential immunity, equal protection, and the dormant Commerce Clause. And in each case, the Court ultimately rules against the group it deems the best harm avoider.

In one sense, the Court’s effort to rule against the best harm avoider in these cases is similar to a traditional interest-balancing approach in which the Court weighs the interests on each side of a case. Both approaches play on the presence of asymmetric consequences. With interest balancing, the Court hopes to find more important interests on one side of the dispute than the other, such that it can rule for the former. But in controversial constitutional cases—such as those involving the right to die, presidential immunity, undocumented immigrant children, and abortion—those interests will often be incommensurable, with weights that are hotly contested.

Harm-avoider arguments implicate a different source of asymmetry. A court that decides against the better harm avoider can admit that the competing interests in constitutional cases are both important and, indeed, incommensurate—that reasonable people could disagree, for example, whether the House’s ability to draft legislation outweighs the President’s need for protection from harassment. Yet harm-avoider analysis can still function by attending to the asymmetry in the groups’ relative abilities to avoid their harms. Sometimes a group will have an easy way to protect itself from its harm, even if its harm is of high order—as was the case with the House’s ability to craft legislation using more tailored subpoenas or Louisiana’s capacity to protect maternal health without Act 620. To be clear, by focusing on the asymmetry in each side’s ability to respond to an adverse ruling, harm-avoider constitutionalism does not purport to make some utilitarian calculation regarding which outcome would produce “the greatest good.”[234] It pursues instead a least harm objective, trusting in the losing group’s ability to avoid its harm.

In focusing on asymmetric consequences, though, it should be evident that harm-avoider arguments do not reveal anything about what the Constitution means or requires. To some, that may seem a fatal concern. But it is a concern also implicated by other popular modes of constitutional argument—including modern originalism, which likewise recognizes that original public meaning is inconclusive in a range of cases.[235] In admitting this same constitutional under-determinacy, harm-avoider reasoning remains well within the mainstream of constitutional culture.

I do not want to overstate my case. The foregoing decisions may utilize harm-avoider logic, but they do not employ a harm-avoider label. Thus far, harm-avoider reasoning in Supreme Court constitutional opinions has been more informal and episodic than formal and systematic. Yet this is to the detriment of the Court and the groups who come before it. Harm-avoider reasoning ought to be systematized in a framework that belongs as a conscious and regular part of our constitutional practice. Showing that the Court already utilizes harm-avoider reasoning with some regularity helps to make out a prima facie case.[236] But convincing the legal community, especially the Court, to acknowledge and employ harm-avoider reasoning more broadly will require more. The next Part uses the Court’s existing harm-avoider decisions to construct a generalizable framework. Part III then presents a normative case that this framework is attractive, not just from Court’s perspective, but for scholars, lawyers, and the broader public too.

II. Constructing Harm-Avoider Constitutionalism

Part II proceeds inductively, building a generalizable framework for the consistent application of harm-avoider reasoning in constitutional law out of the just-discussed cases. Section A addresses the preliminary matter of what it means to “avoid harm” in the context of constitutional law. Once this concept is explained, Section B introduces and elaborates upon a framework for harm-avoider constitutionalism.

A. Avoiding Harms in Constitutional Law

What does it mean to avoid the harms of a Supreme Court constitutional ruling? In private law harm-avoider contexts, it’s fairly obvious. People avoid harm simply by changing their behavior: drivers should leave ample space to the vehicle in front of them, contract drafters should take care to avoid ambiguity, and so on. Yet, as Professor Barry Friedman put it, short of amending the Constitution or convincing five Justices to change an existing constitutional ruling, “[t]here is no overriding the Court.”[237] So talk of avoiding the harms of a Supreme Court decision may seem a bit beside the point. But this is only true of the judicial part of the story. Once it ends, the people can—and often do—write additional chapters.

Groups that come out on the losing end of Supreme Court cases can respond outside the judicial system in any number of ways. For present purposes, I want to draw a distinction between two kinds of responses: public and private avoidance. By public avoidance, I mean efforts to undo the harms of an adverse ruling by changing public law through democratic processes. By private avoidance, I mean efforts to counteract the harms of an adverse ruling that rely on private ordering rather than changes to the law. The following Sections explore these avoidance techniques in closer detail.

1. Public Avoidance: Harm Avoidance Through the Democratic Process

One way a group may try to avoid the harms of a Supreme Court loss is to change public law through the democratic process. Such harm avoidance comes in two flavors: constitutional amendments and ordinary legislation.

The most direct way for a group to counteract the harms it suffers when the Supreme Court adopts an adverse interpretation of the Constitution is to change the Constitution. The swift response by states after Chisholm v. Georgia offers the cleanest example.[238] Dismayed by the Supreme Court’s holding that they were amenable to suits brought by citizens of other states in federal court, the states rallied to propose the Eleventh Amendment to overrule Chisholm “at the first meeting of congress thereafter.”[239] Of course, amending the Constitution is impractical in all but the most extreme circumstances given the arduous Article V amendment process.[240] But constitutional amendments aren’t the only way for adversely affected groups to redress their harms.

Another tactic that also operates through public law is avoidance via legislation. At first blush, this might sound impossible: how can a group avoid a constitutional ruling by the Supreme Court via ordinary legislation, given that state and federal statutory law is subservient to the Constitution?[241] It turns out the harm-avoider decisions canvassed above suggest two possible legislative pathways, depending on the losing group’s relationship to the law initially under attack.

The first pathway involves a group that has failed in its effort to get some existing law or policy declared unconstitutional. In this scenario, public avoidance via legislation means going back to the democratic process to repeal the policy at issue. The aftermath of the Clover Leaf Creamery case is a good example. Recall that the Supreme Court rejected several Minnesota dairies’ and firms’ dormant Commerce Clause challenge in part because those firms possessed ample power to press their case in the legislature (whereas environmental groups lacked viable alternatives).[242] The Court’s analysis proved prescient. Less than five months after the Court announced its opinion, the in-state dairy industry prevailed upon the state legislature to repeal its plastic container ban and replace it with a far more limited intervention: a “marketing study” of the State’s milk packaging practices.[243]

Public avoidance of this sort is not always feasible, however. Consider the inability of disfranchised Blacks to change public law so as to repeal harmful Jim Crow laws after Plessy v. Ferguson.[244] Public avoidance by repealing existing legislation is thus largely a function of political power: the greater political clout a group has, the more likely it will be able to undo the harms of an adverse ruling through this route.[245]

The second pathway of public avoidance via ordinary legislation involves groups who originally prevailed in securing some policy in the democratic process, only for the Supreme Court to deem it unconstitutional. In this scenario, reenacting the same exact law is not an option, but there may exist other legislative solutions that do not run afoul of the Constitution. Alternative solutions of this kind play a substantial role in conventional constitutional analysis: the entire purpose of strict scrutiny’s “least restrictive alternative” test is to identify less harmful legislative means through which groups can attain their desired ends.[246]

Examples of this kind of public avoidance abound. Consider, for example, Plyler, where the Supreme Court struck down a Texas law denying undocumented children access to free public education in part because the Court believed that concerned Texas parents would be better able to protect their educational interests than would undocumented children.[247] This prediction was also borne out in time. Shortly after Plyler was issued, Texas enacted a major legislative overhaul of its school funding system aimed at enhancing the quality of public education offered throughout the state.[248] Not only did the new bill increase statewide school funding by 26 percent annually, it also pegged the amount of funds each school district would receive to actual daily student attendance (including undocumented children), rather than the prior system (which was based in part on the number of employed personnel).[249]

The aftermath of Nixon v. Fitzgerald provides another example. Although the Court rejected Fitzgerald’s attempt to sue the President in his personal capacity, the Court acknowledged the argument that absolute immunity might place the President “above the law” in the future.[250] The Court argued, however, that “alternative remedies and deterrents” exist to “protect[] against misconduct on the part of the Chief Executive,” including public law options such as “[v]igilant oversight by Congress” and “the threat of impeachment.”[251] Each of those alternatives proved instrumental just two years later in precipitating President Nixon’s resignation.[252]

Public avoidance techniques are important, but they are not the only strategies that groups have to avoid the harm of an adverse Supreme Court decision. The next Section explores private harm avoidance.

2. Private Avoidance: Harm Avoidance Outside the Democratic Process

The cases presented in Part I also show how defeated groups can use private ordering to avoid the harms of adverse Supreme Court decisions. This is one important reason to look to private law fields for lessons in constitutional adjudication. In those fields, the easiest strategies for parties to avoid harms involve shifts in private conduct rather than efforts to change public law. To wit, a driver can far more easily avoid liability for a rear-end collision by leaving a greater distance to the car in front of them than by lobbying the legislature to change the rear-end collision rule. Two categories of private avoidance are worth focusing on: case-specific changes in private conduct and foot voting.

The first and most important category involves case-specific changes in private conduct. These are changes that are tailored to the specific harms that a group will suffer should it lose a particular dispute. The aftermath of Cruzan provides a helpful illustration. The Court’s decision represented a clear defeat for Nancy and her parents, who were unable to obtain an order withdrawing Nancy’s life-sustaining treatment.[253] But recall one of the avoidance options the Court relied upon in the course of its decision: the possibility that Nancy’s parents might “discover[] new evidence regarding [Nancy’s] intent,” which would satisfy Missouri’s clear and convincing evidence standard.[254] Nancy’s parents took the Court up on this suggestion, finding three more of Nancy’s coworkers to testify that she would not have wanted to continue treatment.[255] A trial judge found this evidence clear and convincing and granted Nancy’s parents’ request to remove her feeding tube.[256] In December 1990, just six months after the Supreme Court decided Cruzan, Nancy passed away.[257]

Cruzan is a strong example of case-specific private avoidance for another reason. Separate and apart from Nancy’s own case, the decision was a defeat for similarly situated individuals who, at some future point in time, may wish to have life-sustaining treatment withdrawn but who will (like Nancy) find themselves unable to express that intention. In theory, such persons could have banded together in support of a public avoidance strategy, perhaps by lobbying for state legislation authorizing the removal of treatment upon a showing beyond a preponderance of the evidence. As Justice O’Connor’s concurring opinion emphasized, however, a far simpler private avoidance option was available to these people: the ex ante appointment of a proxy to make end-of-life decisions.[258] And at last count, more than one-third of American adults have executed some advanced directive to this effect—a nearly threefold increase from before Cruzan.[259]

A second category of private avoidance techniques involves voting with one’s feet. Whenever a group loses in the Supreme Court, there is always the option—at least in theory—to move. As Professor Ilya Somin has argued, “foot voting” is “a central feature of the American experiment and its relative success.”[260] The basic notion is that when the Supreme Court rules against you, you can always move to a jurisdiction where you will no longer suffer your harm, either because the law or facts on the ground are different. Plyler is a useful illustration of this strategy. There, the Court emphasized that undocumented children lack “the ability to remove themselves from the State’s jurisdiction,” leaving them unable to secure access to education if Texas’s law were upheld.[261] The Texas families who wished to exclude undocumented children, by contrast, had greater ability to move in pursuit of their particular vision of a quality education.[262] Of course, foot voting will be a less realistic avoidance strategy in some constitutional disputes than others, as some losing groups may lack the means to move.[263]

*  *  *

Two types of distinctions are worth noting. First, the constitutional law avoidance techniques just discussed differ significantly from the avoidance strategies in the private law contexts. In the private law examples, harm avoidance was purely retrospective given that the harm to be avoided had already occurred by the time the parties arrived in court. Private law courts use harm-avoider reasoning to rule against parties who could have avoided the harm before it happened (e.g., a car accident), but there is nothing the parties can do to avoid the harm after the fact (since the accident cannot be undone).

In constitutional law, the notion of harm avoidance is both retrospective and prospective. That is, the losing side in a Supreme Court case could have acted differently to avoid the Court’s intervention in the first place. Minnesota dairy industries, for example, could have lobbied for the repeal of the state’s plastic container law before filing suit. But the losing side can also avoid the harm of an adverse ruling after the Court rules. Once Clover Leaf was decided, the affected businesses did indeed press their case in the legislature.[264] This additional flexibility increases the attractiveness of harm-avoider reasoning in constitutional law, as it gives groups even greater ability to avoid their harms.[265]

A second distinction concerns the choice losing groups face between whether to pursue public or private avoidance strategies. Such groups may prefer private avoidance strategies over public avoidance for a simple reason: it is more within their control. Individuals worried about end-of-life decisions after Cruzan could appoint a surrogate in a living will without any legislator’s approval; the same was true of any Texas parent who wished to relocate to another school district after Plyler.

B. A Framework

With a fuller grasp of the various public and private techniques that Supreme Court decisions have identified in deciding which group is best suited to avoid the harms of an adverse decision, it is now possible to construct a general framework for harm-avoider constitutionalism. This Section presents that framework and offers an initial assessment of four sources of potential uncertainty, each of which is denoted below. Here is the framework in its most elemental form:

Under harm-avoider constitutionalism, the Supreme Court decides hard constitutional cases against the group that can best avoid the harm it would suffer from an adverse decision using public and private avoidance techniques.

1. What Counts as a Hard Case?

The first question concerns the scope of constitutional cases to which harm-avoider constitutionalism applies. The theory obviously has no role to play in at least some constitutional disputes. Suppose, for example, the State of California were to argue that it is entitled to three Senators in the next session of Congress. There would be no reason to identify a best harm avoider because the Constitution itself clearly resolves this question.[266] But in other cases, the Constitution does not speak with great clarity, leaving open the possibility of multiple understandings depending on one’s preferred interpretive theory.[267] These are the “hard” cases that would benefit from harm-avoider constitutionalism.

How is the Court to decide whether a constitutional question is “hard” enough for harm-avoider arguments to enter into its analysis? The Court’s approach in cases like Cruzan, Fitzgerald, The Nixon Tapes Case, Plyler, and Clover Leaf Creamery offers some insight. In each of these cases, the Court began its analysis with an honest effort to discern a resolution based on the usual, first-order modes of constitutional argument,[268] such as text,[269] structure,[270] history,[271] and precedent.[272] It was only after finding these sources indeterminate—whether because of the absence of meaningful guidance or the presence of conflicting, persuasive arguments—that the Court proceeded to consider arguments as to which group would be best able to avoid the harms of an adverse decision.

To describe this approach disguises another important question, however. How are the Justices to know when the first-order modes of argument are truly indeterminate? Different Justices will reasonably hold differing views as to how close to equipoise the evidence must be on either side of a constitutional issue before admitting uncertainty as to the correct answer. This implicates a notoriously difficult problem of proof that pervades law generally, and I cannot do it full justice here.[273] My own view is close to that of Professor Heidi Kitrosser, who argued that “[w]here more than one historically plausible meaning [of the Constitution] exists, interpreters should refrain from deeming one the ‘best’” because doing so creates “a false veneer of certainty.”[274] But I readily admit that this is a view predicated on my own normative judgment about the limits of judicial knowledge and the benefits of candidly admitting legal uncertainty.

Fortunately, regardless of one’s perspective on this issue, harm-avoider constitutionalism can function in the presence of pluralistic approaches to this threshold question.[275] For like other widely accepted legal doctrines, harm-avoider reasoning is a decision rule whose soundness judges can agree on even if they disagree as to the scope of the rule’s application.[276] The doctrine of contra proferentem is one example—judges agree that it is a valid tool for resolving ambiguous contracts, even though they often disagree over how quickly to find ambiguity in a contract.[277] Disagreement over a rule’s scope, in other words, need not impugn a rule’s wisdom or force.[278]

2. What Are the Relevant Groups?

The second question concerns the groups whose harm-avoidance abilities a court should be comparing. In the private law harm-avoider contexts, it is simple enough to know the identity of the comparators whose ease of avoidance courts are to consider: the people on either side of the case caption. Contracting Party A (the drafter) is sued by Party B (the non-drafter), so we compare the avoidance strategies possessed by A and B.

The Supreme Court has broadened the relevant comparators in its harm-avoider constitutional rulings. For example, in Plyler v. Doe the Court did not ask whether the named plaintiffs (a particular group of undocumented children) would be better able to avoid the harm of a Supreme Court defeat than the named defendants (a local school superintendent, a school board, and the State of Texas).[279] The Court instead broadened the relevant comparators in two important senses. First, the Court considered the consequences that would be suffered by not only the undocumented children who were named plaintiffs but also all similarly situated children.[280] Second, the Court analyzed the harm to the persons who actually supported the challenged law, not the harm to the government units that sought to defend it.[281] Plyler thus focused on whether striking down Texas’s law would impose educational burdens on “lawful residents” rather than the named superintendent or school district defendants.[282] It is, after all, the underlying groups who would stand to lose if their preferred laws are invalidated, and it is they who would work to avoid the harms of a loss moving forward.

Similarly, in Nixon v. Fitzgerald, the Court did not inquire merely into Fitzgerald’s options for avoiding his harm, but rather the options available to all concerned citizens who might seek “protection against misconduct on the part of the Chief Executive.”[283] Nor did the Court limit its assessment of the harm on the other side of the case to Richard Nixon alone, the named defendant.[284] The Court instead asked whether rejecting absolute immunity would produce a “detriment . . . not only [to] the President and his office but also the Nation that the Presidency was designed to serve.”[285]

The Court’s broader approach to identifying the relevant groups for harm-avoider analysis makes sense. Cases like Plyler and Fitzgerald implicated vital interests far beyond the named persons and entities. As Professor Chayes put it long ago, constitutional litigation implicates “a host of important public and private interactions—perhaps the most important in defining the conditions and opportunities of life for most people . . . [that] can no longer be visualized as bilateral transactions between private individuals.”[286] In such cases, it is “the group” that is “the real subject or object of the litigation,” not merely the named plaintiff and defendant.[287] And so comparison of avoidance strategies in constitutional law properly focuses on those groups, not just the named parties.[288]

3. How Are Each Group’s Harms to Be Defined?

The third question concerns the nature of the harms that groups would suffer from an adverse ruling. Up to this point, I have assumed that these harms are susceptible to uncontroversial identification. But that is not necessarily so. The harms caused by a Supreme Court loss might be characterized quite differently, depending on who one asks. Judges at some distance from the underlying controversy might describe the nature of a harm quite differently than those who suffer it firsthand, and opposing litigants will obviously have their own (likely less charitable) description. How should judges decide whose description of the “harm” of a given decision is the proper subject of harm-avoider analysis? The Court’s harm-avoider rulings suggest that the proper approach is largely to defer to the parties’ descriptions of their own harms.

A decision like The Nixon Tapes Case illustrates how the harm a group would suffer from defeat can be susceptible to competing descriptions—and how the Court itself responded. President Nixon presented one view of the harm that would be inflicted were the Court to reject his claimed privilege: that presidential advisors would henceforth “temper [their] candor with a concern for appearances and for their own interests to the detriment of the [presidential] decisionmaking process.”[289] But opponents of the privilege could quite plausibly characterize Nixon’s true interests quite differently—as a desire to evade public accountability for his role in the Watergate scandal or a desire to protect his criminally indicted aides.

The Court’s actual analysis in The Nixon Tapes Case reveals a particular response to competing harm descriptions. Rather than attempting to adjudicate what the President’s “real” harm would be were he to lose, the Court deferred to the President’s own description of his interests.[290] The Court thus acknowledged Nixon’s professed concern for “the necessity for protection of . . . candid . . . opinions in Presidential decisionmaking,” because “[a] President and those who assist him must be free to explore alternatives in the process of shaping policies and making decisions and to do so in a way many would be unwilling to express except privately.”[291] Other harm-avoider constitutional rulings display a similar kind of deference to the parties’ own harm descriptions.[292]

The Court’s deferential approach is sensible for at least two reasons. First, the Justices do not have any particular wisdom in defining the nature of harms that groups would suffer, especially in comparison to the injured groups themselves. Second, and just as importantly, paternalistic attempts by the Court to declare what a group’s harm really is would risk backlash.[293] The Court’s harm-avoider rulings thus wisely treat each group’s professed harms seriously and evenhandedly.

Deferring to each group’s description of its own harm leads to a potential problem, however. Might a group frame its harms in a way that leads to artificially difficult (or even impossible) avoidance options, thereby enhancing its ability to prevail? Groups that seek to defend a law as it exists might attempt to game the doctrine by characterizing their harm in a circular fashion, as the inability to have the law say what they want it to say. Groups in the converse position, who wish to have a law invalidated, may make a similar move by defining their harm as the loss of a constitutional right. Such descriptions would frustrate harm-avoider constitutionalism because they would define the relevant harms in a way that is impossible to avoid except by a favorable judicial ruling.[294]

In Cruzan, for example, defenders of Missouri’s clear and convincing evidence standard might describe the harm of an adverse ruling as the inability to have state law require clear and convincing evidence of a patient’s desire to terminate treatment. Conversely, Nancy Cruzan’s parents might have argued that the harm of ruling against them was the loss of their daughter’s substantive due process right to reject unwanted medical treatment. Were the Court to defer to these harm characterizations, the only way to avoid each side’s harm would be for the Court to adopt that side’s preferred interpretation of the Due Process Clause.

There is a clean answer to this kind of gamesmanship, however: judges should not consider the inability to have the law say what a group wants it to say as a credible harm. Not only are such claims tautological, as in “we adopt Smith’s interpretation of the law because Smith would like the law to say what he wants it to say,” but they also obscure what is really at stake in constitutional disputes. People support laws not because of some abstract utility they get from having certain words enshrined in a legal code but because of the real-life effects those laws have. It is those underlying effects that harm-avoider constitutionalism aims to grapple with.

A prohibition against circular harm descriptions draws support from the Supreme Court’s animus case law, which holds that moral disapproval is not a valid state interest.[295] The rationale for this rule is not that it is somehow illegitimate or impermissible for people to express moral disapproval of certain groups or behavior in society—indeed the First Amendment protects such speech.[296] The point may instead be that defending a statute by reference to a group’s desire to have the law express disapproval is circular and thus self-fulfilling: the only way to protect the supporters’ interest in having the law reflect their disapproval is to leave the law in force.

4. How Should the Ease of Avoidance Strategies Be Determined?

That brings us to a final source of possible uncertainty: how to determine the ease of a given harm-avoidance strategy. As an initial matter, it is important to emphasize the distinctive feature of the Court’s harm-avoider cases: the Court’s conscious effort to compare the ability of each group to avoid its harms, not the severity of the harms themselves.[297] The harms themselves are often incommensurable, such that any comparison would involve subjective, and quite likely contentious, value judgments. A judicial declaration that the Constitution permits Missouri’s clear and convincing evidence standard because protecting human life is more important than respecting the likely wishes of terminally ill patients would be hard to ground objectively and would inflame public sentiment, as would the inverse. The Court’s harm-avoider decisions instead acknowledge the gravity of the interests on both sides of the case,[298] sending the even-handed message to both groups that winners and losers will not be decided by a judicial guess as to which group really has more important interests at stake.

The ease of avoiding a given group’s harm, however, is often more objective. For example, one need not take a view on the relative importance of protecting human life and respecting the wishes of terminally ill patients to agree that it would be easier for the patients to avoid their harm than it would be for pro-life groups to avoid the harm of a mistaken withdrawal of life-saving treatment. This is not to suggest that comparing the ease of the parties’ avoidance strategies will never be hard. There are assuredly some close cases where the harms of an adverse ruling would be similarly difficult for either group to avoid.[299] But as a later Section aims to show, harm-avoider constitutionalism furthers important values nonetheless.[300]

With the relative ease of avoiding each group’s harms front and center, rather than the severity of the harms themselves, a final issue emerges: should the Court limit its comparison to each group’s avoidance strategies in a vacuum, or should it also consider each group’s actual ability to implement its strategies? The question matters because the ease of any given avoidance strategy is a product of two factors. First, there is the general degree of difficulty that the strategy itself entails: for example, enacting new legislation vs. a private change in behavior. Second, there is the actual means available to the group that would seek to employ it: for example, an affluent vs. disadvantaged group.

Here, again, the Court’s harm-avoider decisions shed light. The Court has cared not about the difficulty of avoidance strategies in the abstract but rather how the relevant groups would actually carry them out. For example, in Clover Leaf Creamery, the Court did not rest on some observation about the general difficulty facing groups who seek to repeal legislation at the state level. It instead relied on the fact that there were actually several “Minnesota firms” burdened by the law who were “major in-state interests” that could lobby against “legislative abuse.”[301] Likewise, in Plyler, the Court did not evaluate the avoidance strategies available to the undocumented children plaintiffs with a general observation that local school board policies are among the easier kinds of laws to change.[302] Such an argument would have elided reality. The Court instead emphasized that undocumented children are “special members of [an] underclass” who can “affect neither their parents’ conduct nor their own status.”[303] It was these children’s actual inability to pursue public and private avoidance strategies that tipped the Court’s harm-avoider analysis in their favor.

III. The Normative Case: Harm-Avoider Constitutionalism’s Virtues

With a conceptual framework set forth, it is now possible to deliberate on harm-avoider constitutionalism’s normative appeal. This Part argues that the Supreme Court should more broadly and consistently embrace harm-avoider constitutionalism as a tool for deciding difficult constitutional cases for four reasons: it reduces the harms of Supreme Court error, mitigates concerns over judicial partisanship, bolsters the Court’s institutional legitimacy, and creates incentives for a more generative mode of constitutional argument.

A. Reducing the Harms of Supreme Court Error

By definition, every Supreme Court case produces a losing party that in some sense suffers harm from its defeat. When the Supreme Court correctly decides a constitutional case, those harms are just a proper byproduct of our constitutional system. But when the Supreme Court reaches an erroneous conclusion—perhaps wrongly relegating free people into slavery or mistakenly leaving thousands of children to labor in dangerous conditions[304]—well, that is a matter of real regret. This is one significant reason to care about how the Supreme Court decides difficult constitutional questions: whenever the Court confronts a hard case, the risk arises that the Court’s first-order modes of decision making will lead to a mistaken outcome, wrongly inflicting human suffering on the losing side.[305]

Harm-avoider constitutionalism presents the Court with a meaningful response to this problem. If the Supreme Court is unable to get the Constitution right in all hard cases, perhaps a different, second-order objective can help guide decisions: reducing the harms produced by the Court’s rulings in difficult cases. This concept of harm prevention has its roots in work by British philosopher Karl Popper,[306] and American philosopher Tom Regan has described a similar objective in the field of animal rights: “[W]henever we find ourselves in a situation where all of the options at hand will produce some harm . . . we must choose that option that will result in the least total sum of harm.”[307] Harm-avoider constitutionalism activates this “least harm principle” by attending to each group’s relative ability to attain its desired ends through efforts outside the judiciary. In doing so, the theory maximizes the odds of an outcome in which neither group will suffer grave harm in the long run because the group that loses in court has the greatest ability to remedy its harm outside it.

Harm-avoider constitutionalism thus shares the intuition that underlies the medical profession’s Hippocratic Oath: “First, do no harm.”[308] Put another way, if the Supreme Court cannot reliably know how the Constitution resolves the next hard case that risks substantial human suffering, then one sensible response is to rule with an eye towards minimizing harm. On this view, the groups that are best positioned to mitigate their own harms will lose in the Supreme Court. But their losses would represent only the start of the story. Losing groups would remain able to prevent their harm using all of the private and public avoidance strategies at their disposal.

B. Mitigating Judicial Partisanship

A second virtue of harm-avoider constitutionalism is its ability to help counteract concerns over judicial partisanship. A recent survey revealed that 59 percent of Americans believe the Supreme Court is too influenced by politics, compared to just 35 percent who don’t.[309] As Professor Eric Segall argued, people increasingly believe that “the Supreme Court does not function as a true court and its Justices do not decide cases like true judges”; they are “politicians in robes.”[310] In other words, one explanation for the public’s growing concern about judicial partisanship is the appearance that the Supreme Court’s rulings are entirely predictable based on the political affiliations of the Justices themselves.[311]

Harm-avoider constitutionalism can help to reduce these perceptions because its focus on the respective sides’ ability to avoid their harms will lead to a mixture of progressive and conservative outcomes, thus avoiding case outcomes with any reliable ideological bent.[312] In some cases, it produces outcomes consistent with conservative political preferences. In Cruzan, for example, the conservative position, which opposes liberalization of the right to die,[313] prevailed.[314] Likewise, in Nixon v. Fitzgerald, the Court’s recognition of absolute presidential immunity insulated a Republican President from liability.[315] In other cases, the progressive position is more difficult to achieve outside the courts, such that harm-avoider constitutionalism results in progressive victories. Plyler’s recognition of an equal protection right for undocumented children to tuition-free public education is one example; [316] the outcomes of The Nixon Tapes Case and June Medical are others.[317] The 2019 Term’s subpoena cases are themselves an example of harm-avoider constitutionalism’s lack of ideological predictability: President Trump was the better harm-avoider and thus lost in Vance, but the opposite was true in Mazars.[318] As the Court’s own harm-avoider constitutional rulings show, the theory yields a mixed bag of outcomes from the perspective of partisan politics.

This outcome heterogeneity is no accident because the best harm avoider in any given dispute will typically depend on case-specific nuances that are not susceptible to broad generalization.[319] Unless one political party somehow systematically favors better harm avoiders than the other, harm-avoider constitutionalism will not produce politically predictable outputs. Broader and more consistent use of harm-avoider constitutionalism thus offers the Court a strategy for combating the perception of raw partisanship.

C. Bolstering Institutional Legitimacy

Few close followers of the Supreme Court would dispute that, from the standpoint of the public’s faith in the Court’s institutional legitimacy, the Court is “in a weaker position now than at nearly any point in modern history.”[320] Judicial partisanship plays a role in this development for reasons discussed above,[321] and so harm-avoider constitutionalism can help to bolster the Court’s legitimacy by encouraging the Court to decide cases in a less partisan manner. But harm-avoider constitutionalism also protects the Court’s legitimacy in another, more direct fashion: it provides losing groups more effective strategies for responding to their defeats than assailing the Supreme Court’s credibility.

Whenever the Supreme Court decides a difficult and politically sensitive case, the losing group finds itself with a choice. How should it react to its defeat? The cases that have most strained the Court’s public legitimacy are ones where losing groups have decided to respond by attacking the Court itself. Citizens United is one recent example.[322] Public disapproval of the Court’s decision—which struck down a federal statute banning corporate electioneering expenditures—is nearly 80 percent.[323] And opponents of the decision have voiced their displeasure with the Supreme Court; a series of protesters who wanted Citizens United overruled went so far as to disrupt a pair of recent oral arguments at the Court.[324]

What is notable about the Court’s harm-avoider decisions is the absence of any similar sense of public outrage. And one reason inheres in harm-avoider constitutionalism’s very approach to deciding hard cases: by identifying readily available alternatives that losing groups can use to avoid the harms of their defeat, harm-avoider decisions provide those groups with a better response than waging war on the Court’s credibility.[325]

Take the aftermath of Cruzan as one example. As Nancy Cruzan’s case—and later, the case of Terry Schiavo—demonstrates, the State’s forced provision of medical treatment against the wishes of a patient’s loved ones is incredibly contentious for freighted political and moral reasons.[326] How, then, did the Court manage to avert public backlash after Cruzan? It did so by providing Cruzan’s parents and the broader right-to-die movement with several effective alternatives to a frontal assault on the Court itself. Why would Cruzan’s parents rail against the Court when they could just gather additional evidence of Nancy’s wishes—as they in fact did, leading to the removal of treatment just months after the Court’s ruling?[327] Why would ordinary citizens take to the streets in opposition to Cruzan’s heightened evidentiary standard when they could protect their right to die far more easily by appointing a surrogate medical decision-maker? Indeed, in just the single month after the Cruzan decision, “the Society for the Right to Die answered nearly 300,000 requests for advanced‑directive forms.”[328] When presented with similar options, the losers in Fitzgerald, Plyler, and Clover Leaf Creamery seized upon them too.[329]

D. Towards A More Generative Supreme Court

Thus far, I have argued that harm-avoider constitutionalism produces three normative benefits—the reduction of harm from Supreme Court error, mitigation of judicial partisanship, and protection of institutional legitimacy—that exist by virtue of the theory’s reasoning in hard cases. These benefits were accordingly on display in Cruzan, Fitzgerald, The Nixon Tapes Case, Plyler, and Clover Leaf Creamery because those cases relied on harm-avoider reasoning,[330] even if not as part of an express framework publicly endorsed by the Court itself.

The fourth virtue I wish to advance—harm-avoider constitutionalism’s potential to create a more generative culture of constitutional argument—has not yet materialized from any of the Court’s decisions. That is because the Court has yet to expressly endorse harm-avoider constitutionalism as an argumentative framework to be used in close constitutional cases. What I want to argue now is that such an endorsement has the potential to transform today’s destructive culture of constitutional argument in profound ways.

Making out this argument requires me to first say a word about the current nature of constitutional discourse at the Supreme Court. That discourse, I am afraid, reflects—and worse yet, produces—pervasive polarization in society. Take Justice Scalia’s remarkable statement in his dissenting opinion in Obergefell v. Hodges: he would “hide [his] head in a bag” if he ever joined an opinion like Justice Kennedy’s majority opinion, which Scalia chided as possessing all the “mystical aphorisms” of a “fortune cookie.”[331] Or consider Justice Kagan’s accusation that the five conservative Justices were “black-robed rulers overriding citizens’ choices” in her dissent from Janus v. American Federation of State, County, and Municpal Employees, Council 31,[332] a case that struck down public sector unions’ decades-old practice of collecting fair-share fees from objecting workers.[333] When Supreme Court Justices use such bombastic language to discredit the legitimacy of their opponents’ views on the Constitution, it is no surprise that the litigants do the same.[334]

The Justices’ and litigants’ propensity to disparage even their opponents’ most reasonable constitutional claims follows naturally from the Supreme Court’s apparent position that it can conclusively pronounce the one and only true meaning of the Constitution, even in hard cases.[335] When those are the terms for engagement, victory can be won either by proving one’s position to be unassailably correct or by showing one’s opponent’s position to be worthy of ridicule. The result, as Professor Jamal Greene has observed, is a destructive manner of constitutional argument that “coarsens” and “diminishes” us as a people, “leaving us farther apart at the end of a dispute than we were at the beginning.”[336]

The Supreme Court’s open embrace of harm-avoider reasoning would improve this destructive constitutional discourse in two ways. First, the focus of the parties’ briefs (and ultimately, the Court’s opinions) would shift. When outcomes turn on each side’s ability to avoid the harms of an adverse ruling, there is suddenly no profit to belittling the constitutional status of one’s opponents. Rather than arguing that one’s opponents are unworthy of constitutional regard, harm-avoider constitutionalism incentivizes the parties to generate solutions to the underlying dispute—public and private avoidance strategies that their opponents can pursue to protect their interests.[337] The party who will prevail, after all, is the one who can point to the most readily-available strategy by which their opponent can avoid the harms of defeat. And so the entire litigation process—from discovery to trial to appellate briefing—would reorient towards the identification of such solutions.[338]

Second, harm-avoider constitutionalism creates incentives for competing groups to do more than argue over the availability of existing avoidance strategies; it incentivizes them to produce new solutions, too. That is, it encourages parties to actively seek out and agree to legislative and other compromises by which their opponents can protect their underlying interests. To see how, place yourself in the position of campaign finance supporters in Citizens United. Assume that the Court had explicitly endorsed harm-avoider constitutionalism’s decision rule, under which the Court rules against the group that has the most readily available strategies for avoiding the harms of a judicial defeat. The effect of that decision rule would be significant. Not only would it encourage you to identify alternative strategies that corporations already have for expressing their political views,[339] but it would also create incentives for you to propose new compromise legislation that would make it easier for corporations to avoid their harms.

To put a more concrete face on this idea, consider one of the private avoidance techniques available to corporations in Citizens United. If the federal corporate expenditure ban were upheld, one way corporations would still be able to participate in election-related speech is to engage in express advocacy through segregated political action committees (PACs) funded by voluntary contributions from persons affiliated with each corporation.[340] Federal law already permitted such segregated PAC expenditures, subject to strict limits on individual donations by persons affiliated with the corporation—such persons could only contribute $5,000 per year.[341] A Court committed to harm-avoider constitutionalism would create incentives for campaign finance reformers to propose a concession that would provide corporations greater ability to avoid their harms—thereby increasing the odds of a ruling in the reformers’ favor. For example, campaign finance reform proponents might agree to raise the $5,000 limit, making it easier for corporations to avoid the harm of an adverse ruling. Doing so would increase the reformers’ odds of prevailing at the Court. And it would accomplish this through a legislative compromise that, while perhaps not a first-best world from the reformers’ perspective, is surely more palatable than unlimited general corporate expenditures.

In sum, harm-avoider constitutionalism presents a solution-forcing approach to hard constitutional cases. It transforms constitutional argument into a constructive exercise where contestants fight to win by identifying solutions rather than denigrating their opponents. And it encourages the parties to strike mutually acceptable compromises that make it easier for their opponents to achieve their interests. These virtues—in addition to the theory’s other benefits[342]—are worth pursuing. I consider next whether there are drawbacks that might outweigh them.

IV. Counterarguments

Harm-avoider constitutionalism can be met with several important counterarguments. I consider them here under three broad headings. The first category encompasses concerns over harm-avoider constitutionalism’s workability. A second category argues that harm-avoider constitutionalism may be too simplistic insofar as it creates implicit advantages for certain kinds of litigants. A third set of counterarguments concerning harm-avoider constitutionalism’s administrability and objectives arises from important recent work by Professors Charles Barzun and Michael Gilbert.[343]

A. Is Harm-Avoider Constitutionalism Workable?

I have argued that harm-avoider constitutionalism offers an attractive approach that the Supreme Court should use—and in fact, has already used—to decide hard questions of constitutional law. An obvious rejoinder goes something like this: isn’t harm-avoider constitutionalism itself hard, perhaps even unworkably so? This rejoinder comes in different flavors, one rooted in the difficulty of harm-avoider analysis in some cases; a second rooted in the partisan motivations that all Justices may harbor; a third rooted in the uncertain role of precedent in harm-avoider constitutionalism.

The first version of the workability critique points out that harm-avoider constitutionalism requires courts to make contested judgments as to which group is best able to avoid its harm. Arguably, these judgments will sometimes be just as difficult as those required under the first-order tools of constitutional interpretation that harm-avoider constitutionalism seeks to supplement.

I agree wholeheartedly. There will assuredly be cases where a conclusion regarding the groups’ relative ability to implement avoidance strategies is too difficult to reach with any kind of confidence. One prominent example is the right to abortion.[344] Start with the harms implicated on each side of the debate. If Roe and Casey were overturned, pro‑choice groups would point to the dire harm to maternal health and child welfare that will result once women in many states no longer have access to safe procedures.[345] Pro-life groups, by contrast, would point to the loss of unborn life if Roe and Casey were upheld.[346] A court applying harm-avoider constitutionalism would recognize the gravity of the interests on both sides and proceed to evaluate whether one side is better situated to avoid its harm via public and private strategies.

To start, public avoidance seems a close question. Pro-choice groups may fight to oppose abortion bans in state legislatures—a route that, although difficult, is still easier than pro-life groups working to amend the Constitution. Pro-life groups, conversely, could avoid their harm by supporting laws that would prevent unwanted conception in the first place.[347] As for private avoidance, some women in states with abortion bans may travel to other states to obtain a procedure.[348] But that option would not be available to all women, especially those with fewer means.[349] It is also unclear whether pro-life groups can avoid their harm via any kind of private avoidance strategy. One could reasonably conclude that no clear winner can be gleaned from this analysis.

But the existence of cases where neither side boasts superior harm-avoidance strategies does not mean the theory should be altogether abandoned. It simply means that in some cases, the Court will need to look to other forms of argumentation for guidance—namely, to the same kinds of arguments the Court relies on currently when its usual first-order tools are indeterminate.[350] When a harm-avoider judgment is hard, in other words, the Court finds itself in no worse a position than it would have been otherwise.

Furthermore, admitting the existence of some hard harm-avoider cases that does not mean every case is hard. Just as there are some easy constitutional cases under first-order interpretive tools, there also easy harm-avoider cases. Moreover, the existence of cases like Mazars, Vance, June Medical, Cruzan, Fitzgerald, and Plyler,[351] where the Court has already utilized harm-avoider reasoning despite the absence of a generalized framework, suggests that the number of “easy cases” may be significant. And so long as the two axes of decision-making are not perfectly correlated—so long as some constitutional cases that are “hard” in the traditional first-order sense are “easy” in the harm-avoider sense—then the use of both approaches can increase the odds of the Court arriving upon a normatively desirable outcome.[352]

A second, more cynical version of the workability argument asserts that the Justices will approach the harm-avoider inquiry with the same degree of partisanship that they use in their first-order constitutional judgments, thus undermining any attempt to decide cases based on harm-avoider reasoning. Again, I confess my agreement. There is no reason to think motivated reasoning will suddenly disappear if Justices move beyond first-order questions of constitutional meaning to the second-order question of harm avoidance.

But motivated reasoning has its limits.[353] A case like Cruzan offers a good illustration. No reasonable harm-avoider analysis could produce a result in favor of the Cruzans. This is because the harm of erroneously terminating Nancy’s life‑saving treatment would be impossible to avoid.[354] And as the case’s history would later confirm, the avoidance strategies available to the Cruzans and other persons concerned with the right to die were well within reach.[355] Any Justice who harbored a partisan desire to rule in the Cruzans’ favor, in other words, would have been unable to disguise that preference in the objective language of harm-avoider reasoning.

Not all cases will present as clear cut a harm-avoider comparison as Cruzan, of course. In closer cases, I agree that individual Justices might manipulate the harm-avoider analysis by deeming their preferred side’s avoidance strategies less attainable. But even if this happens and the Court therefore gets the harm-avoider determination wrong in some meaningful set of cases, harm-avoider constitutionalism would still yield important virtues. Groups who lose on the basis of harm-avoider reasoning will have options for responding to defeat that are more productive than assailing the Court’s credibility.[356] And the process of deciding cases under harm-avoider constitutionalism would still produce a more generative culture of argument.[357]

A third workability argument concerns the role of precedent. What kind of guidance or precedential effect would an opinion based on harm-avoider reasoning provide to individual actors and courts in future cases? At a first cut, it’s clear that the answer must be “less”—or at least less than is offered by the Court currently given its “law-declarer” disposition.[358] The premise of a best harm-avoider opinion, after all, is that an underlying constitutional question is too hard to answer definitively without a significant risk of error, such that the case should be resolved narrowly against the best harm avoider.

In this sense, harm-avoider constitutionalism’s closest theoretical cousin may be Professor Cass Sunstein’s judicial minimalism, which advocates narrow rulings and shallowly reasoned opinions over broad and deeply reasoned ones.[359] For like minimalism, a court that decides cases under harm-avoider constitutionalism “is intensely aware of its own limitations,” “attempts to promote the democratic ideals of participation, deliberation, and responsiveness,” and therefore “leave[s] things undecided.”[360]

That is not to say, however, that the theory would result in decisions offering no precedential value. Harm-avoider decisions would still guide future courts in a fashion much like the common law: future cases raising the same legal question and operative facts (i.e., similar affected groups, harms, and avoidance options) would presumptively come out the same way. For example, the first time the Court confronted milk packaging law like the Minnesota statute challenged in Clover Leaf Creamery, it would explain that the in-state industries opposing the law are better harm avoiders than the environmental groups who support it.[361] If a similar law were challenged by businesses in another state on similar facts, the first decision would hold strong precedential weight. But if a case arose concerning a different economic regulation enacted for different reasons—perhaps a demonstrable desire to protect the in-state dairy industry to the detriment of out-of-state firms that lack influence over the state legislature[362]—then that might lead to a different harm-avoider analysis and outcome.[363]

This Section has argued that harm-avoider constitutionalism provides meaningful guidance in some cases that would be hard to decide on first-order principles. The following Section considers an attack from the opposite direction: does the theory provide too much guidance in ways that will consistently advantage certain kinds of litigants?

B. Does Harm-Avoider Constitutionalism Advantage Certain Groups?

A second kind of objection argues that harm-avoider constitutionalism directly advantages certain groups. One possible group is those who defend challenged laws. A second group includes politically powerless groups that lack social and political capital. I consider these concerns in turn.

One possibility is that harm-avoider constitutionalism stacks the deck in favor of deference to challenged legislation. From the perspective of public avoidance, it will typically be easier for the group that is challenging a law to move for the law’s repeal than it will be for the group that is defending the law to amend the Constitution to permit the law’s reenactment. Harm-avoider constitutionalism may thus reduce to Thayerian deference by another name.[364]

It is true that the public avoidance considerations will generally reflect a soft preference in favor of upholding challenged laws. But this is far from absolute; there will be occasions when the group seeking to defend a law is actually better positioned to avoid the harm of an adverse ruling. Plyler is a good example: even though lawfully present parents were the ones asking the Court to defer to Texas’s attempt to exclude undocumented children from public schools, the Court struck down the law because those parents had greater access to avoidance options than did the undocumented children.[365] Some of those avoidance options were private, such as leaving their children in public schools where the evidence suggested the quality of education would be unaffected by the enrollment of undocumented children,[366] or sending their children to different schools.[367] Other options were public in nature, including legislative alternatives far easier to secure than a constitutional amendment. For example, concerned Texas parents could, and eventually did, support an infusion of educational resources from the State.[368] Mazars is another example.[369] Rather than deferring to the various Congressional subpoenas, the Court recognized the availability of an easy avoidance strategy: more carefully crafted and thus less burdensome subpoenas (whether to the President or other sources) that would still enable Congress to pursue its professed legislative purposes.[370]

Another critique suggests that harm-avoider constitutionalism favors politically powerless groups. This is so, the critique runs, because in cases involving competing groups with appreciably different levels of political power, the more powerful group is likely to have greater access to public (and perhaps private) avoidance strategies.

I do not dispute this directionality.[371] But where is the problem? If the concern is that an advantage for powerless groups will translate into predictably partisan outcomes, this is mistaken for two reasons. To start, many cases will not implicate an obvious power imbalance. In Fitzgerald v. Nixon, for instance, it is hard to say if one group is less powerful than the other: individuals interested in checking a rogue President, or individuals interested in protecting the President’s ability to act vigorously without concern for private damages liability.[372] Furthermore, even in cases with a power imbalance, sometimes the less powerful group will align with progressive preferences, but other times it will align with conservative ones. The less powerful group in Plyler, for instance, was undocumented immigrant children—whose supporters are more likely to be progressive.[373] But the opponents of the right to die were the less powerful group in Cruzan—a quintessential conservative view.[374]

Finally, political power is far from absolute. A less powerful group will often be the better harm avoider in light of the particular nature of the underlying dispute and the availability of private avoidance options. Take a case like Illinois v. Wardlow, which held that a criminal suspect’s “unprovoked” and “[h]eadlong flight” upon seeing police officers raised sufficiently reasonable suspicion to warrant a Terry stop.[375] The Court did not decide the case using harm-avoider reasoning, but what if it had? While criminal suspects as a class are less politically powerful than police officers (and the general crime-fearing public), such suspects have an easy way to avoid the harm of an adverse ruling—don’t engage in headlong flight.[376] And so harm-avoider constitutionalism would weigh against them.

The defeasible role of political powerlessness underscores how harm-avoider constitutionalism differs from another theory to which it bears some surface-level similarities: John Hart Ely’s political process theory.[377] While harm-avoider constitutionalism shares Ely’s concern for situations in which majorities legislate to the disadvantage of powerless groups,[378] the underlying reason for that concern is distinct. Ely argued for heightened judicial intervention into laws burdening such groups because that is what he believed the Constitution requires.[379] Harm-avoider constitutionalism disclaims any ability to tell us what the Constitution requires in hard cases; indeed, it is precisely the difficulty and divisiveness of these cases that leads to the theory’s suggestion that courts should minimize harms instead of rendering some best guess as to the Constitution’s meaning.[380]

Because of these divergent normative underpinnings, the two theories also operate on different considerations. Ely conceived of powerlessness as an on/off switch: laws that burden groups that “keep[] finding [themselves] on the wrong end of the legislature’s classifications” are reviewed under strict scrutiny in all cases, while laws that do not burden such groups are presumed constitutional.[381] Harm-avoider constitutionalism looks at a wider range of factors in a more nuanced way. The theory cares, for example, whether a powerless group can easily avoid the harms of an adverse ruling through private ordering—a concern that escaped Ely’s attention. Harm-avoider constitutionalism is also sensitive to different levels of power: laws burdening powerful groups are especially worthy of our trust,[382] and even if a group is not so powerless as to qualify as a suspect class, its relative lack of influence still matters.[383] Finally, Ely’s theory has little to say about cases implicating no power differential between groups except to apply a presumption of constitutionality.[384] Under harm-avoider constitutionalism, these cases would turn on the availability of the strategies each group has to avoid its harms given the nature of the case at hand. The theory accordingly does not produce a categorical advantage for less powerful groups.

C. Avoiding Harm or . . . Conflicts?

A recent paper from Professors Charles Barzun and Michael Gilbert raised a final set of arguments that are comparative in nature. In Conflict Avoidance in Constitutional Law, Barzun and Gilbert forcefully advance a “conflict avoidance principle” under which a court should “decide hard cases against the party who could have more easily avoided the constitutional conflict in the first place.”[385] Barzun and Gilbert’s principle and harm-avoider constitutionalism build off many of the same major premises.[386] Both begin with hard cases, where the law runs out.[387] Both propose an approach to deciding these cases that diverges from the traditional, first-order arguments that are common in constitutional law.[388] And both attempt to justify this move away from the usual modes of argument by pointing to certain normative virtues, such as the promise of greater institutional legitimacy and positive litigation incentives.[389]

I want to underscore these similarities because they are important—far more important, I think, than the differences between our particular approaches. For unlike originalism, living constitutionalism, and other traditional theories that direct judges to render a best guess as to the Constitution’s proper meaning even amidst deep uncertainty, both of our approaches allow judges to openly and honestly admit when a legal question is indeterminate—and to pursue valuable second-order considerations when that is so. In this sense, our theories suggest the emergence of what might be thought of as a “second-order turn” in constitutional law.[390]

Harm-avoider constitutionalism and conflict-avoidance do have their differences, to be sure. One difference is that harm-avoider constitutionalism is arguably grounded in the Supreme Court’s case law;[391] Barzun and Gilbert do not contend that the Court already practices conflict-avoidance in its constitutional docket.[392] For their part, Barzun and Gilbert argue that the conflict-avoidance principle has the major advantage of being more easily administered by courts.[393]

I should note that I agree vigorously with Barzun and Gilbert’s concern: if our shared goal is to suggest a better, more honest theory for deciding hard cases, not much is accomplished by proposing an alternative that is just as hard to administer as the interpretive theories we already have. But there is strong evidence that harm-avoider constitutionalism is not so indeterminate: the fact that the Supreme Court has already used it effectively to resolve a significant number of disputes across a wide range of doctrinal fields, including Congress’s Article I power, presidential immunity, substantive due process, and equal protection, to name a few examples.[394]

That said, Barzun and Gilbert’s conflict-avoidance principle might well be easier still for judges to apply. If so, that would be a mark in their favor. Yet it isn’t clear if it is true. Barzun and Gilbert argue that conflict avoidance is more administrable because it only requires courts to inquire into the named parties’ avoidance strategies (rather than those of the groups they represent).[395] But I confess uncertainty as to whether they can cabin the inquiry in this way. For one thing, they openly admit that their named-party limit does not apply to half of the case caption: when it comes to government defendants, “[o]ur approach looks beyond broad assertions of ‘government interest’ to the real interests at hand.”[396] Thus, in Burwell v. Hobby Lobby Stores, Inc., the conflict-avoidance principle would ask courts to evaluate whether each of the “thousands”[397] of “employees to whom Hobby Lobby sought to deny coverage could achieve their interest through other means.”[398] Nor is it evident how their principle could save a court from analyzing avoidance costs for the whole group represented by the plaintiff’s side of the case caption, either. After all, plaintiffs in constitutional cases often bring their claims as class actions. A case like Brown v. Board of Education is emblematic. Given that the plaintiffs brought their claims as a class action,[399] a court applying Barzun and Gilbert’s principle would have to evaluate the avoidance costs of all members of the affected groups—precisely the same inquiry required under harm-avoider constitutionalism.

Finally, as Barzun and Gilbert recognized, one might wonder about the “core normative assumption” behind the conflict-avoidance principle, that “constitutional conflicts should be avoided where possible.”[400] To their great credit, Barzun and Gilbert grappled evenhandedly with this concern (do we really want to force same-sex couples to move to blue states to avoid conflict over red-state marriage bans before Obergefell?[401]), so I won’t press them on the point. I do, however, want to point out how harm-avoider constitutionalism aims to create a different set of incentives for competing groups in our pluralistic society: not to reduce interactions among people of different backgrounds and beliefs so as to avoid conflict, but to resolve our conflicts by working together in pursuit of mutually acceptable solutions.[402] In my view, what we need isn’t less conflict, but more mutual toleration—a renewed ability for people who disagree to recognize their opponents’ legitimacy, reach across the divide, and find workable compromises that avoid our harms.[403]

Conclusion

American courts have long recognized limits on their ability to find the single correct answer to difficult legal questions.[404] Rather than plowing ahead and issuing rulings that will often be erroneous, these courts often decide on a different basis: they rule against the best harm avoider.[405]

Until now, the assumption in the scholarly community has been that this approach is limited to the field of private law.[406] This Article has shown otherwise. Across a range of doctrinal areas—including a trio of blockbuster decisions in the Supreme Court’s most recent Term alone—the Court has used the same harm-avoider reasoning to resolve hard questions of constitutional law.[407] And so for those who wonder whether originalism,[408] living constitutionalism,[409] or some other interpretive theory is “our law,” it is fair now also to ask: do we have a harm-avoider Supreme Court?

The answer matters for reasons well beyond legal practice. As private law harm-avoider cases demonstrate, harm-avoider constitutionalism can have effects that reach far beyond the courts. Indeed, the theory may be capable of shaping the constitutional understandings and relationships of the American people writ large. After all, in the private law settings, the overarching purpose of harm-avoider rules is to encourage the parties to take ownership over their fates and prevent harms before they happen—be they car accidents, contract disputes, or property damage.[410] The same could be true of constitutional disputes. If hard constitutional questions were openly and consistently decided against the best harm avoider, litigants would find reason to candidly assess their options for securing their interests through alternative public and private means. Groups that enjoy easy avoidance options would do well to take them up instead of going to court, since litigation would be costly and fruitless. And groups less capable of avoiding their harms would be incentivized to make compromises that help their opponents achieve their interests outside of court.[411]

A full and open embrace of harm-avoider constitutionalism, in other words, could teach competing groups in our ever complex and polarized society to stop looking to the Supreme Court for answers to every dispute with constitutional dimensions. It could teach groups to reflect on how they and their opponents might attain their desired ends on their own, without ever stepping into a courtroom. Of course, the theory recognizes that some groups will be able to protect themselves more easily than others. Ruling against these groups because they are the best harm avoider isn’t likely to make them happy. But maybe, just maybe, it can show the people that we are the ones we’ve been waiting for—not nine unelected Justices on the Supreme Court.

 

Aaron Tang: Professor of Law, University of California, Davis, School of Law. I am grateful to Scott Altman, Charles Barzun, Mitchell Berman, Ash Bhagwat, Guido Calabresi, Katherine Mims Crocker, David Cruz, Bill Dodge, Chris Elmendorf, Richard Fallon, Barry Friedman, Michael Gilbert, Andrew Guzman, Daniel Hemel, Pam Karlan, Heidi Kitrosser, Larry Kramer, Carlton Larson, Sanford Levinson, Douglas Melamed, Frank Michelman, Manisha Padi, Richard Re, Stephen Rich, Eric Segall, Darien Shanske, Fred Smith, Larry Solum, Mark Tushnet, Ted Weinstein, and J. Harvie Wilkinson III for comments and ideas that helped shape this project. I thank the editors at the California Law Review for their tremendous assistance. I also thank workshop participants at the University of Southern California, Gould School of Law; the University of California, Hastings College of Law; the University of Virginia School of Law; the 2019 Constitutional Law Colloquium at Loyola University Chicago School of Law; and the AALS Section on Law & Interpretation’s New Voices in Interpretation Program.

           [1].     See, e.g., William Baude, Is Originalism Our Law?, 115 Colum. L. Rev. 2349, 2351–63 (2015).

           [2].     See, e.g., David A. Strauss, The Living Constitution 33–98 (2010).

           [3].     See John Hart Ely, Democracy and Distrust: A Theory of Judicial Review 105–80 (1980).

           [4].     See T. Alexander Aleinikoff, Constitutional Law in the Age of Balancing, 96 Yale L.J. 943, 963–72 (1987).

           [5].     See Phillip Bobbitt, Constitutional Interpretation 12–13 (1991); see also sources cited infra note 268 (identifying other pluralist accounts). This list of conventional accounts is of course only exemplary, not exhaustive.

           [6].     See, e.g., Ramos v. Louisiana, 140 S. Ct. 1390, 1395–97 (2020); District of Columbia v. Heller, 554 U.S. 570, 581–619 (2008).

           [7].     See supra notes 2–5 (identifying opinions that fit within these categories).

           [8].     To be sure, the normative debate need not be limited to those theories that the Court frequently uses. But the leading contenders in the debate are advantaged to the extent they can claim the Supreme Court’s regular use. See Jamal Greene, Rule Originalism, 116 Colum. L. Rev. 1639, 1690 (2016) (“In constitutional law, if not in other legal domains, ‘is’ implies ‘ought.’”).

           [9].     Cf. Cass R. Sunstein & Edna Ullman-Margalit, Second Order Decisions, 110 Ethics 5, 7 (1999) (defining a second-order decision as one “about the appropriate strategy for reducing the problems associated with making a first-order decision”).

         [10].     140 S. Ct. 2019 (2020).

         [11].     Id. at 2026.

         [12].     See id. at 2026, 2028.

         [13].     See id. at 2028–29 (describing lower court rulings).

         [14].     See id. at 2035 (remanding for lower courts to “perform a careful analysis” of “the separation of powers principles at stake”).

         [15].     See id. at 2031 (admitting there is no “enumerated constitutional power” in Article I to issue subpoenas).

         [16].     See id. at 2032 (rejecting the standard used in the Nixon Tapes case); id. at 2033–34 (rejecting the normal standard involving Congress’s powers to issue subpoenas due to separation-of-powers concerns).

         [17].     Id. at 2035.

         [18].     Id.

         [19].     Id. at 2035–36.

         [20].     Id. at 2036.

         [21].     Id. A fourth factor considers a subpoena’s burdens on the President. I discuss this infra Part I.B.5.

         [22].     See Mazars, 140 S. Ct. at 2035–36.

         [23].     See id. at 2034–35 (explaining how presidents cannot avoid harm by, for example, requiring the House to subpoena third-party custodians).

         [24].     140 S. Ct. 2412 (2020).

         [25].     Id. at 2420.

         [26].     Id.

         [27].     Id. at 2431.

         [28].     See id. at 2428–29 (acknowledging but declining to endorse Justice Thomas’s original meaning rationale).

         [29].     Id. at 2424–25. To be clear, the Court did closely consider precedent, including the use of a presidential subpoena in the treason trial of Aaron Burr. Id. at 2422–23. But as the Court pointed out, that case and all others relied on by the State of New York were not dispositive because they “involved federal criminal proceedings.” Id. at 2424.

         [30].     Indeed, the majority does not use the word “balance” a single time. Two other conventional modes of argument do play some role in the opinion—arguments based on historical practice and constitutional structure. I discuss them infra Part I.B.6.

         [31].     Vance, 140 S. Ct. at 2425.

         [32].     Id. at 2430.

         [33].     Id.

         [34].     Id. at 2427.

         [35].     Id. at 2430.

         [36].     Id.

         [37].     Id. (emphasis in original).

           [38].     140 S. Ct. 2103 (2020).

         [39].     Id. at 2112.

         [40].     See id. (no mention of original meaning).

         [41].     While the plurality did rely on the Court’s ruling four years earlier in Whole Woman’s Health v. Hellerstedt, 136 S. Ct. 2292 (2016), Louisiana argued (and the Fifth Circuit agreed) that the two cases were factually dissimilar. See June Med. Servs., 140 S. Ct. at 2116–17. What is more, Whole Woman’s Health was itself an example of harm-avoider constitutionalism. See discussion infra note 233.

         [42].     See infra Part I.B.7 (discussing interest-balancing understanding of June Medical).

         [43].     See infra Part I.B.7 (discussing interest-balancing understanding of June Medical).

         [44].     See id. at 2121–28 (discussing how several hospitals refused to afford admitting privileges for religious reasons and how others required significant in-hospital experience that abortion providers lack due to the safety of their procedures). Justice Alito’s principal dissent likewise focused on this issue, arguing that the doctors could have tried harder to obtain privileges at various hospitals. Id. at 2157–65 (Alito, J., dissenting). As I explain below, a harm-avoider analysis makes sense of this strangely exhaustive discussion. See infra Part I.B.7.

         [45].     June Med. Servs., 140 S. Ct. at 2129–30 (plurality opinion).

         [46].     Id.

         [47].     See id. at 2130–32.

         [48].     Id. at 2130–31.

         [49].     Id. at 2132.

         [50].     See infra Part I.B (discussing how the Court uses harm-avoider reasoning to decide difficult cases across these areas of law).

         [51].     See infra Part II.B (discussing the range of hard cases to which the theory applies).

         [52].     See H.L.A. Hart, The Concept of Law 128 (3d ed. 2012) (arguing that there exist difficult cases in which the law will “prove indeterminate” and “have what has been termed an open texture”).

         [53].     In the field of ethics, Tom Regan has advanced a similar concept known as the “minimize harm principle.”  Tom Regan, The Case for Animal Rights 302 (2004).

         [54].     See infra Part III (canvassing harm-avoider constitutionalism’s virtues).

         [55].     This is true of other theories of interpretation, too, of course. For instance, many cases are decided without mentioning the Constitution’s text, much less original meaning. See, e.g., Strauss, supra note 2, at 33 (“Pick up a Supreme Court opinion in a constitutional case, at random . . . the text of the Constitution will play, at most, a ceremonial role.”). Still others are decided without great reliance on precedent. See, e.g., NLRB v. Noel Canning, 573 U.S. 513, 526–49 (2014) (relying primarily on historical practice, not prior judicial interpretations, to resolve two major questions concerning the Recess Appointments Clause).

         [56].     The post hoc nature of my project is not fatal. Oftentimes a pattern in the Court’s decisions arises first, and theory comes second to describe it. See, e.g., Ely, supra note 3, at 105–80 (describing the political process theory jurisprudence that emerged during the Warren Court); Aleinikoff, supra note 4, at 948–63 (describing the emergence of interest balancing).

         [57].     See Charles Barzun & Michael Gilbert, Conflict Avoidance in Constitutional Law, 107 Va. L. Rev. 1 (2021).

         [58].     See supra notes 1–5; see also Jack Balkin, Arguing About the Constitution: The Topics in Constitutional Interpretation, 33 Const. Comment. 145, 181–83 (2018) (describing eleven modalities of argument, or “different ways that lawyers argue about the Constitution”).

         [59].     Cf. Sunstein & Ullmann-Margalit, supra note 9, at 6–7 (describing a “first-order decision” as an “ordinary decision-making situation” where one decides without the benefit of a second-order strategy for reducing information costs and other burdens of the decision).

         [60].     See Ronald Dworkin, Hard Cases, 88 Harv. L. Rev. 1057 (1975) (discussing the problem of what judges do in hard cases).

         [61].     See Sunstein & Ullmann-Margalit, supra note 9, at 11–12.

         [62].     Id. at 7.

         [63].     The application of second-order decision-making strategies in constitutional law is nothing new. One might think of all “implementing doctrine[s]” that the Court crafts to carry the Constitution into effect as a kind of second-order consideration insofar as such doctrines do not “plausibly reflect the Constitution’s true meaning.” Richard H. Fallon, Jr., Foreword: Implementing the Constitution, 111 Harv. L. Rev. 56, 57, 65 (1997). Some of what I have called “first-order” interpretive tools (like the use of implementing doctrines) could thus be conceptualized as second-order tools, too. Nothing turns on this distinction, though. My point is that the Court’s reliance on harm-avoider reasoning is different than other commonly accepted modes of argument to the extent it does not purport to tell us anything about the “correct” meaning of the Constitution.

         [64].     Although not the focus of this Article, the Court also uses harm-avoider reasoning outside its constitutional docket. For example, the Court’s decision to invalidate the Trump Administration’s effort to rescind DACA is plausibly understood as a cost-avoider ruling: whereas the Administration could avoid its harms easily by promulgating a new, more well-reasoned explanation for terminating DACA, a ruling against the plaintiffs would inflict harms that are far harder to avoid. Dep’t of Homeland Sec. v. Regents of the Univ. of Cal., 140 S. Ct. 1891 (2020). Harm-avoider arguments thus have the potential to apply outside constitutional law, such as to administrative law and statutory interpretation. See Aaron Tang, The Least Harm Principle 19–28 (July 12, 2021) (unpublished manuscript) (on file with author) (describing statutory and administrative law cases in which the Court makes harm avoider arguments).

         [65].     Abram Chayes, The Role of the Judge in Public Law Litigation, 89 Harv. L. Rev. 1281 (1976).

         [66].     Id. at 1282.

         [67].     Id. at 1282–83 (listing the “defining features” of private law civil adjudication).

         [68].     Restatement (Second) of Contracts § 206 (Am. L. Inst. 1981).

         [69].     Scenic Am., Inc. v. Dep’t of Transp., 138 S. Ct. 2 (2017) (Gorsuch, J., concurring in denial of certiorari).

         [70].     7 U.S. (3 Cranch) 415, 432 (1806).

         [71].     Id. at 439.

         [72].     Id. at 432.

         [73].     Id. at 448. A similar harm-avoider rationale explains the contract law rule governing property that is damaged at a particular moment in time: after a sale is executed, but before possession is transferred. Under the common law of contracts, the seller is liable for the property damage, not the buyer. See Unif. Vendor & Purchaser Risk Act § 1, 14 U.L.A. 471 (2005). As Samuel Williston explains, the reason is that “[t]he practical advantages of leaving the risk with the [seller] until transfer of possession are obvious . . . . [I]t is wiser to have the party in possession of property care for it at his peril.” Samuel Williston, The Risk of Loss After an Executory Contract of Sale in the Common Law, 9 Harv. L. Rev. 106, 122 (1895). Or put another way, the seller is the better harm avoider because she can more easily avoid damage to the property than a buyer who cannot yet enter the property.

         [74].     Guido Calabresi, The Cost of Accidents: A Legal and Economic Analysis 135 (1970). Following Calabresi’s lead, the private law literature typically refers to “harm-avoider” reasoning as “cost-avoider” reasoning. See, e.g., George L. Priest, The Rise of Law and Economics: A Memoir of the Early Years, in The Origins of Law and Economics 350, 369 (Francesco Parisi & Charles K. Rowley eds., 2005). The underlying concern, however, is the same: a focus on which side in the case could have avoided the underlying harm or cost (e.g., a contract dispute or traffic accident) more easily. Of course, in constitutional law, the theory cannot pursue economic efficiency in the same way as in private law. Instead, the theory must be justified by other normative objectives, which I lay out infra Part III.

         [75].     See, e.g., Filippazzo v. Santiago, 716 N.Y.S.2d 710 (N.Y. App. Div. 2000).

         [76].     Id.

           [77].     The same reasoning explains the common law rule that trains have the right of way at railroad crossings over wagons, cars, and like vehicles. See, e.g., Grand Trunk Ry. Co. v. Ives, 144 U.S. 408, 431 (1892) (“A traveler upon a highway, when approaching a railroad crossing, ought to make a vigilant use of his senses of sight and hearing, in order to avoid a collision . . . . If by neglect of this duty he suffers injury from a passing train, he cannot recover.”). As the Supreme Court explained long ago, given its heavy “character and momentum,” a train “cannot be expected [to] stop and give precedence to an approaching wagon to make the crossing first; it is the duty of the wagon to wait for the train.” Cont’l Improvement Co. v. Stead, 95 U.S. 161, 164 (1877). Car (or wagon) drivers are the best harm avoiders because they can stop more quickly, and so such vehicles must yield to oncoming trains.

         [78].     See, e.g., Cerilli v. Comm’r, Dep’t of Corr., No. HHDCV166068641S, 2018 WL 650305, at *1 (Conn. Super. Ct. Jan. 5, 2018).

         [79].     Harm-avoider rules have also been suggested in property law, most notably in academic discussions of private nuisance. See, e.g., Calabresi, supra note 74, at 254 (applying cheapest cost avoider reasoning to the problem of a “factory near a residential area [that] results in an air pollution problem”); R. H. Coase, The Problem of Social Cost, 3 J.L. & Econ. 1, 41 (1960) (suggesting nuisance rule against lower cost avoider); Guido Calabresi & A. Douglas Melamed, Property Rules, Liability Rules, and Inalienability: One View of the Cathedral, 85 Harv. L. Rev. 1089, 1118–19 (1972) (same).

         [80].     See, e.g., Calabresi, supra note 74; see also Priest, supra note 74, at 369 (noting that Guido Calabresi’s “best cost-avoider” theory of law and economics “has been widely adopted in judicial decisions”).

         [81].     See Robert D. Cooter & Michael D. Gilbert, Constitutional Law and Economics, in Research Methods in Constitutional Law: A Handbook (Malcolm Langford & David S. Law eds., forthcoming) (manuscript at 36) (arguing law & economics “has had little impact on constitutional law”), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3123253 [https://perma.cc/D2MG-A26S]. But see Richard M. Re, Beyond the Marks Rule, 132 Harv. L. Rev. 1942, 1969 (2019) (“[T]he law of precedent should place burdens on the ‘cheapest precedent creator . . . .’”); James E. Pfander, Alexander A. Reinert & Joanna C. Schwartz, The Myth of Personal Liability: Who Pays When Bivens Claims Succeed, 72 Stan. L. Rev. 561, 602 (2020) (describing the logic of cost-avoider reasoning in constitutional torts, such as the rule denying qualified immunity to municipalities).

         [82].     See Aaron Tang, Rethinking Political Power in Judicial Review, 106 Calif. L. Rev. 1755, 1775–76 (2018) (describing the importance of Supreme Court practice for purposes of determining the existence or content of law).

         [83].     Greene, supra note 8, at 1655.

         [84].     I do need to show, of course, that the task of identifying a best harm avoider is generally one that the Court can accomplish, or at least one that the Court can achieve in some meaningful number of cases where the first-order interpretive tools are inconclusive. Otherwise, harm-avoider constitutionalism would just replace one intractable puzzle with another. I consider this important critique below. See discussion infra Part IV.A.

         [85].     Cruzan v. Dir., Mo. Dep’t of Health, 497 U.S. 261, 268 (1990).

         [86].     Id. at 266.

         [87].     Id.

         [88].     Id. at 267.

           [89].     Id. at 268.

         [90].     Id. at 285.

           [91].     Id. at 286–87.

         [92].     Id. at 280.

         [93].     Id. at 282.

         [94].     See id. at 278–79 (canvassing the relevant text and precedent to recognize the existence of a fundamental liberty interest, but then admitting that this “does not end the inquiry,” such that the Court must “balanc[e] [the patient’s] liberty interests against the relevant state interests”).

         [95].     Id. at 283.

         [96].     Id.

         [97].     Id.

         [98].     Id. (emphasis added).

         [99].     See id.

      [100].     See id. at 274.

      [101].     Id. at 283.

      [102].     Id. at 289 (O’Connor, J., concurring).

      [103].     Id. at 291–92.

      [104].     See supra note 95 and accompanying text.

      [105].     See, e.g., Aleinikoff, supra note 4, at 947 & nn.22 & 25, 971 & n.181.

      [106].     Harrison Smith, A. Ernest Fitzgerald, Pentagon Whistleblower Fired by Nixon, Dies at 92, Wash. Post (Feb. 7, 2019), https://www.washingtonpost.com/local/obituaries/a-ernest-fitzgerald-pentagon-whistleblower-fired-by-nixon-dies-at-92/2019/02/07/2f3277f4-2afe-11e9-984d-9b8fba003e81_story.html [https://perma.cc/CZN8-A9GS].

      [107].     Nixon v. Fitzgerald, 457 U.S. 731, 734 (1982).

      [108].     Id. at 740, 740 n.20.

      [109].     Id. at 748–49, 748 n.27.

      [110].     Id. at 733, 749.

      [111].     Id. at 747.

      [112].     See id. at 758.

      [113].     Id. at 757–58.

      [114].     Id. at 757.

      [115].     Id. Recent events suggest limits on the effectiveness of these alternatives. My point presently, however, is that the Court relied on the availability of alternative ways concerned members of the public could avoid the harm of an unaccountable President—that is, it relied on arguments about the better harm avoider.

      [116].     Id. at 752.

      [117].     Id. at 752 n.32.

      [118].     Id. at 752–53.

      [119].     Id. at 753.

      [120].     See Bob Woodward & Carl Bernstein, GOP Security Aide Among Five Arrested in Bugging Affair, Wash. Post, June 19, 1972, at A1, https://www.washingtonpost.com/politics/gop-security-aide-among-five-arrested-in-bugging-affair/2012/06/07/gJQAYTdzKV_story.html?itid=lk_inline_manual_2 [https://perma.cc/TRH4-MA63].

      [121].     See id.

      [122].     United States v. Nixon, 418 U.S. 683, 687 (1974).

      [123].     Id. at 687–88.

      [124].     See id. at 688.

      [125].     Id. at 713.

      [126].     See James Hohmann, The Daily 202: Why U.S. v. Nixon Matters – Now More Than Ever, Wash. Post (July 24, 2018), https://www.washingtonpost.com/news/powerpost/paloma/daily-202/2018/07/24/daily-202-why-u-s-v-nixon-matters-now-more-than-ever/5b5678331b326b1e646954eb/ [https://perma.cc/9Y2T-RQQ4].

      [127].     Nixon, 418 U.S. at 708.

      [128].     See id. at 710–11 (declining to extend precedent involving the Executive Privilege). But see id. at 708 (noting that the Executive Privilege is “inextricably rooted in the separation of powers under the Constitution”).

      [129].     Id. at 711–12.

      [130].     Id. at 713.

      [131].     Id. at 712.

      [132].     Id. at 713.

      [133].     Id.

      [134].     Id. at 713 (emphasis added) (quoting United States v. Burr, 25 F. Cas. 187, 192 (C.C. Va. 1807) (No. 14,694)) (brackets in original).

      [135].     Id. at 715.

      [136].     Id. at 714–15.

      [137].     Id. at 715.

      [138].     See Plyler v. Doe, 457 U.S. 202, 205–06, 206 n.2 (1982) (describing the challenged school district policy as “requiring undocumented children to pay a ‘full tuition fee’ in order to enroll”).

      [139].     Id. at 230.

      [140].     See, e.g., Gerald L. Neuman, Aliens as Outlaws: Government Services, Proposition 187, and the Structure of Equal Protection Doctrine, 42 UCLA L. Rev. 1425, 1443 (1995) (“Plyler has been criticized for not fitting neatly into a one-dimensional account of the Equal Protection Clause under which heightened scrutiny is triggered by the presence of either a suspect classification or a fundamental right.”).

      [141].     Plyler, 457 U.S. at 223.

      [142].     See id. at 223–24.

      [143].     See id. at 224 (emphasis added).

      [144].     Id. at 230.

      [145].     Id. at 220 (quoting Trimble v. Gordon, 430 U.S. 762, 770 (1977)).

      [146].     Id.

      [147].     Id. at 230.

      [148].     See id. at 229.

      [149].     Id. at 229.

      [150].     Id.

      [151].     Minnesota v. Clover Leaf Creamery Co., 449 U.S. 456 (1981).

      [152].     Id. at 459.

      [153].     Id. at 470.

      [154].     See id. at 465 (“[T]he State argues that elimination of the popular plastic milk jug will encourage the use of environmentally superior containers.”).

      [155].     Id. at 470.

      [156].     Mitchell N. Berman, Coercion Without Baselines: Unconstitutional Conditions in Three Dimensions, 90 Geo. L.J. 1, 49 (2001).

      [157].     Clover Leaf Creamery Co., 449 U.S. at 471 (quoting Pike v. Bruce Church, Inc., 397 U.S. 137, 142 (1970)).

      [158].     Id. at 472–73.

      [159].     Bendix Autolite Corp. v. Midwesco Enters., 486 U.S. 888, 897 (1988) (Scalia, J., concurring).

      [160].     Clover Leaf Creamery Co., 449 U.S. at 473.

      [161].     Id. at 473 n.17. Or in the Court’s words, the fact that powerful Minnesota firms were “adversely affected by the Act is a powerful safeguard against legislative abuse.” Id.

      [162].     Id. at 473–74. Another alternative—outlawing paper and plastic containers alike—was impermissible because it would impose a greater burden on commerce. Id.

      [163].     Id. at 473.

      [164].     Clover Leaf Creamery is not the only dormant Commerce Clause case to rule against a given group because of its superior ability to avoid its harm via changes to public law. See, e.g., W. Lynn Creamery v. Healy, 512 U.S. 186, 200 (1994) (striking down a Massachusetts tax and subsidy in part “because one of the in-state interests which would otherwise lobby against the tax has been mollified by the subsidy”).

[165]. The Court utilized harm-avoider reasoning in another 2020 constitutional case, Ramos v. Louisiana. 140 S. Ct. 1390 (2020). In Ramos, the Court held that the Sixth Amendment right to a jury trial requires a unanimous verdict for a defendant to be convicted of a serious offense. Id. at 1394. Although Ramos was principally decided on the basis of the original meaning of the Sixth Amendment, see id. at 1395–97, the Court also noted how the losing side—states that would need to retry cases resolved via non-unanimous verdicts—would at least have some ability to avoid the harms of their defeat (unlike a ruling in the opposite direction). Compare id. at 1406 (recognizing that “retrying or plea bargaining [cases where defendants were convicted without unanimous juries] will surely impose a cost,” albeit one that is “less” than the cost inflicted by prior rulings), with id. at 1408 (arguing that a ruling in Louisiana’s favor would require the Court to “discard a Sixth Amendment right in perpetuity”).

      [166].     140 S. Ct. 2019, 2029 (2020). The subpoenas were directed nominally at the President’s bank and accounting firm. But the Court treated them as the President’s own records. See id. at 2035 (Thomas, J., dissenting) (“[I]t is, after all, the President’s information.”).

      [167].     Id. at 2027 (majority opinion).

      [168].     Id. at 2028.

      [169].     Eastland v. U.S. Servicemen’s Fund, 421 U.S. 491, 506 (1975) (quoting McGrain v. Daugherty, 278 U.S. 135, 177 (1927).

      [170].     See Mazars, 140 S. Ct. at 2028.

      [171].     Id. at 2036.

      [172].     See id. at 2033.

      [173].     See id. at 2035.

      [174].     See, e.g., Josh Chafetz, Opinion, Don’t Be Fooled, Trump Is a Winner in the Supreme Court Tax Case, N.Y. Times (July 9, 2020), https://www.nytimes.com/2020/07/09/opinion/trump-taxes-supreme-court-.html [https://perma.cc/EGS2-ZS82].

      [175].     See Mazars, 140 S. Ct. at 2034–35.

      [176].     See id. at 2037 (Thomas, J., dissenting).

      [177].     See id. at 2032, 2034 (majority opinion) (rejecting precedent developed and used in prior cases). Likewise, the Court referenced structural arguments such as the separation of powers—but only to reject one of the parties’ proposed rules and not to develop the correct constitutional standard. See id. at 2033 (explaining how the House’s preferred standard “fails to take adequate account of . . . separation of powers”).

      [178].     See id. at 2031 (describing prior practice as “a consideration of great weight”). See generally Will Baude, Constitutional Liquidation, 71 Stan. L. Rev. 1 (2019) (exploring the role of historical practice in constitutional law).

      [179].     See Mazars, 140 S. Ct. at 2035.

      [180].     Id.

      [181].     Id. at 2035–36. As was likely true of the subpoena issued by the House Financial Services Committee seeking President Trump’s financial records, for instance. See id. at 2027 (describing the purpose of the House Financial Services Committee’s subpoenas to the President as an effort “examine the implementation, effectiveness, and enforcement” of money laundering laws).

      [182].     Id. at 2036.

      [183].     Id.

      [184].     Id.

      [185].     Id.

      [186].     One possibility is that the Court envisioned some unspoken threshold beyond which a subpoena would impose burdens that are simply too great: a kind of direct weighing of the harm to the President. See Aleinikoff, supra note 4, at 945 (defining balancing to include the “construct[ion of] a rule of constitutional law by explicitly or implicitly assigning values to [an] identified interest[]). Yet in addition to being standardless and difficult to administer, this seems inconsistent with the Court’s failure to include in its analysis a corresponding weight on the other side of the scale. Not once, after all, does the new, four-part constitutional test announced in Mazars ask a court to weigh the burden on Congress if a subpoena is blocked. See Mazars, 140 S. Ct. at 2035–36. The test focuses instead on whether Congress has other ways to achieve its legislative aims (and thus avoid its harm altogether). See id.

      [187].     See Mazars, 140 S. Ct. at 2026–27.

      [188].     Id. at 2035.

      [189].     Id. at 2030.

      [190].     Trump v. Vance, 140 S. Ct. 2412, 2420.

      [191].     See id. at 2431.

      [192].     See id. at 2429 (acknowledging but declining to endorse Justice Thomas’s originalist analysis); see also id. at 2427–29 (grounding rejection of absolute immunity in the existence of other safeguards against presidential stigma and harassment, rather than original meaning); id. at 2423–25 (providing a historical narrative concerning the use of presidential subpoenas in federal criminal proceedings).

      [193].     Id. at 2424.

      [194].     To be more specific, the opinion focuses on the relevant groups, not just the named parties. See id. at 2430 (inquiring into the “public interest” in New York law enforcement); id. (explaining how “presidents,” plural, can protect themselves from harassment moving forward). I discuss this more infra Part II.B.

      [195].     Vance, 140 S. Ct. at 2425.

      [196].     Id. at 2430.

      [197].     Id. at 2429. Technically, the Court ruled against the President on two alternative arguments: one claiming absolute immunity and another seeking a heightened standard for enforcing state criminal process. Both rulings relied on the same harm-avoider arguments, however. See id. at 2428–29 (explaining constitutional limits on subpoenas directed at the Executive Branch); see also id. at 2430 (elaborating on legal options available to the President in response to a subpoena). I thus collapse the two lines of reasoning here for ease of exposition.

      [198].     Id. at 2427.

      [199].     Id. at 2428.

      [200].     Id. at 2430.

      [201].     Id. at 2431.

      [202].     See id. at 2430–31.

      [203].     Id. at 2430 (quoting id. (Alito, J., dissenting)).

      [204].     Id.

      [205].     Id.

      [206].     Id.

      [207].     Id. (emphasis in original).

      [208].     June Med. Servs. L.L.C. v. Russo, 140 S. Ct. 2103, 2112 (2020) (noting that Act 620 is “almost word-for-word identical to Texas’ admitting-privileges law”).

      [209].     See June Med. Servs. L.L.C. v. Gee, 905 F.3d 787 (5th Cir. 2018).

      [210].     See June Med. Servs., 140 S. Ct. at 2134 (Roberts, C.J., concurring) (voting to strike down Act 620 on the basis of stare decisis).

      [211].     See Aleinikoff, supra note 4, at 948. The plurality definitely did not engage in “definitional balancing,” which is to say a ruling that access to abortion always outweighs maternal health, rendering additional balancing unnecessary in subsequent cases. See id. As Chief Justice Roberts wrote in his concurring opinion, “[t]here is no plausible sense in which anyone, let alone this Court, could objectively assign weight to such imponderable values and no meaningful way to compare them if there were.” June Med. Servs., 140 S. Ct. at 2136 (Roberts, C.J., concurring).

      [212].     Id. at 2120 (plurality opinion).

      [213].     See id. at 2129 (explaining that the law would lead to the elimination of Does 1, 2, 3, and 6, leaving only Doe 5 in practice).

      [214].     Id. at 2129–30.

      [215].     Id. at 2132.

      [216].     See id. Together, the plurality and principal dissent cite to Casey and Whole Woman’s Health seventy-four times. See id.

      [217].     See id. at 2122–28 (plurality opinion) (arguing that additional efforts by Louisiana abortion physicians to obtain admitting privileges would have been futile); id. at 2119–24 (arguing that the physicians could have obtained privileges with greater effort).

      [218].     Id. at 2124–25.

      [219].     See id. (explaining that Doe 2 did not have in-hospital experience because abortions are outpatient procedures).

      [220].     Id. at 2162 (Alito, J., dissenting).

      [221].     Id.

      [222].     Id. at 2164.

      [223].     Id. at 2122 (plurality opinion).

      [224].     See id. at 2129.

      [225].     The plurality arguably described its twelve-page analysis of the various doctors’ compliance efforts as an inquiry into whether Act 620 caused the resulting burdens on abortion access. See id. at 2125. But on any theory of causation, it did. Before the law was enacted, five doctors offered abortions in the state. The doctors applied to, and were rejected by, thirteen hospitals. Id. at 2122. So after the law’s enactment, only one provider would remain. Id. at 2129. Whether some doctors might try harder to satisfy Act 620 is thus not a question of causation, but of ex post harm avoidance—much like the question whether a voter might work harder to pay a poll tax she cannot afford.

      [226].     See id. at 2128.

      [227].     Id.

      [228].     Id.

      [229].     Id. at 2129–30.

      [230].     Id.

      [231].     Id. at 2130–31.

      [232].     Id. (citation omitted).

      [233].     There is one additional way to explain the outcome in June Medical: perhaps the case is best seen as a common law constitutionalist ruling, insofar as it relies heavily on a single precedent, Whole Woman’s Health v. Hellerstedt, 136 S. Ct. 2292 (2016). This is surely a reasonable reading, though Justice Alito at least found the two cases distinguishable. See June Med. Servs., 140 S. Ct. at 2158 (Alito, J., dissenting). In any event, Whole Woman’s Health was itself a harm-avoider decision. The Court explicitly held that Texas abortion providers would be unable to avoid the harm of an adverse ruling moving forward simply by obtaining the admitting privileges required by state law. See Whole Woman’s Health, 136 S. Ct. at 2312–13 (discussing inability of an abortion provider named Dr. Lynn to secure admitting privileges). Likewise, many Texas women would be unable to obtain a procedure due to the law’s effect on travel distances. See id. at 2313 (noting that Texas’s admitting privileges requirement increased the number of women who would have to travel more than 200 miles from 10,000 to 290,000). Conversely, Texas had an easy alternative for avoiding its harm: Texas could protect maternal health by resuming enforcement of the law as it stood prior to the admitting privileges requirement. See id. at 2311–12 (noting Texas’s concession that the admitting privilege requirement would not have helped “even one woman obtain better treatment”). Whole Woman’s Health thus bottoms out on a direct comparison of the difficulty of avoidance options available to the competing sides—exactly the kind of harm-avoider comparison at the heart of June Medical.

      [234].     Cf. Richard Posner, Pragmatic Adjudication, in The Revival of Pragmatism: New Essays on Social Thought, Law, and Culture 235, 244 (Morris Dickstein ed., 1998) (arguing that judges should resolve hard cases “in a way that will produce the best results in the circumstances rather than just deciding cases in accordance with rules created by other organs of government”).

      [235].     See infra note 267.

      [236].     See Adrian Vermeule, Connecting Positive and Normative Legal Theory, 10 U. Pa. J. Const. L. 387, 393 (2008) (“[T]he way things are . . . itself serves as a direct source of norms . . . .”).

      [237].     Barry Friedman, The Will of the People: How Public Opinion Has Influenced the Supreme Court and Shaped the Meaning of the Constitution 5 (2009).

      [238].     2 U.S. (2 Dall.) 419 (1793).

      [239].     See Hans v. Louisiana, 134 U.S. 1, 11 (1890).

      [240].     See U.S. Const. art. V (requiring ratification by three-fourths of all of the States).

      [241].     See U.S. Const. art. VI, cl. 2 (“This Constitution, and the laws of the United States which shall be made in Pursuance thereof . . . shall be the supreme Law of the Land . . . any Thing in the Constitution or Laws of any State to the Contrary notwithstanding.”); Marbury v. Madison, 5 U.S. (1 Cranch) 137, 180 (1803) (holding that a federal law that is “repugnant to the constitution is void”).

      [242].     See supra text accompanying notes 152–165.

      [243].     See Act of May 8, 1981, ch. 151, §§ 1–2, 1981 Minn. Laws 457, 458, https://www.revisor.mn.gov/laws/1981/0/Session+Law/Chapter/151/pdf/ [https://perma.cc/WY9N-KGY6].

      [244].     See, e.g., Donald E. DeVore, Defying Jim Crow: African American Community Development and the Struggle for Racial Equality in New Orleans, 1900-1960, at ii (2015) (“[T]he complete loss of political influence by 1900 effectively closed electoral politics to African Americans . . . throughout the South.”).

      [245].     See Tang, supra note 82, at 1801–02 (arguing that courts should be especially deferential to laws burdening politically powerful groups because legislatures are most attuned to the interests of those groups).

      [246].     See Richard H. Fallon, Jr., Strict Judicial Scrutiny, 54 UCLA L. Rev. 1267, 1326 (2007).

      [247].     See supra text accompanying notes 148–150.

      [248].     See Intercultural Dev. Rsch. Ass’n, A History of IDRA Policy Work to Secure Excellent and Equitable Schooling for All Children 3 (2018), https://www.idra.org/wp-content/uploads/2018/07/A-History-of-IDRA-Policy-Work-to-Secure-Excellent-and-Equitable-Schooling-for-All-Children.pdf [https://perma.cc/57GL-H7BL] (showing that the Plyler decision was followed by adoption of HB72).

      [249].     See Glen Hegar, Texas School Finance: Doing the Math on the State’s Biggest Expenditure, Fiscal Notes, Jan. 2019, at 20, https://comptroller.texas.gov/economy/fiscal-notes/2019/jan/history.php [https://perma.cc/K4RS-FDR8]. Technically this avoidance strategy was not the route the Court envisioned. The Plyler Court grounded its conclusion that concerned Texas parents could avoid the harm of educating undocumented children in the trial court record, which found little evidence to believe the quality of education would suffer from the modest infusion of undocumented students. See supra text accompanying notes 149–150.

      [250].     See 457 U.S. 731, 758 (1982).

         [251].     Id. at 757–58. To be sure, congressional oversight (e.g., in the form of conducting hearings to investigate presidential misconduct) and the enactment of articles of impeachment are different than the passage of ordinary legislation. I include them here as examples of public avoidance insofar as they are pathways by which citizens who wish to hold the President accountable—the losing side in Nixon v. Fitzgerald—can do so acting through the legislature.

      [252].     See R.W. Apple Jr., In 2 Years, Watergate Scandal Brought Down President Who Had Wide Mandate, N.Y. Times, Aug. 9, 1974, at 14, https://www.nytimes.com/1974/08/09/archives/in-2-years-watergate-scandal-brought-down-president-who-had-wide.html, [https://perma.cc/TCV6-2CL4] (describing the role that impeachment proceedings and House and Senate committee investigations played in precipitating Nixon’s resignation). Recent events may well cast doubt on the power of impeachment or congressional investigation to rein in the Chief Executive. But that is not the point of the Court’s harm-avoider analysis in Fitzgerald: the point is instead whether the threat of private damages liability was necessary to keep the President from rising above the law. On that score, it’s significant that oversight and the threat of impeachment were sufficient to provoke Nixon’s resignation, without risk of monetary liability.

      [253].     Cruzan v. Dir., Mo. Dep’t of Health, 497 U.S. 261, 286–87 (1990).

      [254].     Id. at 283.

      [255].     See Tamar Lewin, Nancy Cruzan Dies, Outlived by a Debate over the Right to Die, N.Y. Times, Dec. 27, 1990, at A1, A15, https://www.nytimes.com/1990/12/27/us/nancy-cruzan-dies-outlived-by-a-debate-over-the-right-to-die.html [https://perma.cc/9VUQ-DCK2].

         [256].     See id.

      [257].     See id.

      [258].     See Cruzan, 497 U.S. at 289 (O’Connor, J., concurring).

      [259].     See Carolyn Crist, Over One Third of U.S. Adults Have Advanced Medical Directives, Reuters (July 11, 2017), https://www.reuters.com/article/us-health-usa-advance-directives/over-one-third-of-u-s-adults-have-advanced-medical-directives-idUSKBN19W2NO [https://perma.cc/77ZB-6C6H]; see also Strong Public Support for Right to Die, Pew Rsch. Ctr. (Jan. 5, 2006), https://www.people-press.org/2006/01/05/strong-public-support-for-right-to-die/ [https://perma.cc/5EZ2-KHAW] (finding just 12 percent of Americans had a living will in 1990).

      [260].     Ilya Somin, Foot-Voting Nation, in Our American Story: The Search for a Shared National Narrative 123, 124 (Joshua A. Claybourn ed., 2019).

      [261].     See Plyler v. Doe, 457 U.S. 202, 220 (1982).

      [262].     See Matthew Hall & Jacob Hibel, Latino Students and White Migration from School Districts, 1980-2010, 64 Soc. Probs. 457, 464 (2017) (finding that during the 1980s, a “10 percent increase in Latino youth populations correspond[ed] with a reduction in the rate of white populations . . . of about .05 percent”). According to Hall and Hibel, a common challenge of integrating Latinx students in historically White school districts is that White families migrate to other districts, depleting their initial schools of “financial, social, and political resources.” Id. at 471.

      [263].     The canonical example of foot voting is Brown v. Board of Education, 347 U.S. 483 (1954), after which many White families chose to vote with their feet by moving to outlying suburbs with fewer Black families. See Charles T. Clotfelter, After Brown: The Rise and Retreat of School Desegregation 8 (2004) (noting after Brown that “suburban school districts were the most obvious alternative to city school districts with high . . . proportions of minority students”).

      [264].     Likewise, in Mazars, the various House Committees could have issued subpoenas that were more narrowly tailored to their respective legislative purposes before the Court’s decision; they can still do the same after it.

      [265].     The relevance of forward-looking strategies by which losing groups can avoid their harm distinguishes harm-avoider constitutionalism from the conflict-avoidance principle advanced by Professors Barzun and Gilbert. See Barzun & Gilbert, supra note 57, at 16 n.46 (“Our focus is retrospective.”). Consider a case like Cruzan. Both theories would treat as relevant the fact that, before her tragic car crash, Nancy Cruzan could have signed an advanced medical directive appointing a surrogate medical decision-maker, thereby avoiding the harm she would eventually suffer. That retrospective option, however, would do nothing to alleviate the painful consequences of an adverse ruling for Nancy because she obviously could not go back in time to execute such a form. Harm-avoider constitutionalism thus has the virtue of considering an additional prospective option for avoiding Nancy’s harm even after the Court rules against her: the fact that Nancy’s parents could obtain additional evidence demonstrating her intent to withdraw treatment. That, of course, is precisely the forward-looking harm-avoidance option that the Court itself flagged and that Cruzan’s parents pursued successfully, ensuring their daughter’s ability to die with dignity. See supra notes 256–258 and accompanying text.

      [266].     See U.S. Const. art. I, § 3, cl. 1 (“The Senate . . . shall be composed of two Senators from each State . . . .”).

      [267].     In the parlance of modern originalism, the resolution of these cases occurs in the “construction zone.” See Lawrence B. Solum, The Interpretation-Construction Distinction, 27 Const. Comment. 95, 108 (2010); see also Randy E. Barnett, Interpretation and Construction, 34 Harv. J.L. & Pub. Pol’y 65, 69 (2011) (“The text of the Constitution may say a lot, but it does not say everything one needs to know to resolve all possible cases.”); Keith E. Whittington, Constitutional Construction: Divided Powers and Constitutional Meaning 5–9 (1999) (similar). Or in another sense, to say that a case is “hard” is to say a case lacks legal clarity. See Richard M. Re, Clarity Doctrines, 86 U. Chi. L. Rev. 1497, 1498–99 (2019).

      [268].     The Court’s use of a standard list of argumentative modalities is often attributed to work by Professor Phillip Bobbitt. See supra note 5. Other influential accounts of what has come to be known as constitutional pluralism include Richard H. Fallon Jr., A Constructivist Coherence Theory of Constitutional Interpretation, 100 Harv. L. Rev. 1189 (1987); Balkin, supra note 58, at 242–43.

      [269].     See, e.g., Cruzan v. Dir., Mo. Dep’t of Health, 497 U.S. 261, 278 (1990); Nixon v. Fitzgerald, 457 U.S. 731, 749–50 (1982).

      [270].     See, e.g., United States v. Nixon, 418 U.S. 683, 707 (1974) (considering arguments about “the structure of our Government”).

      [271].     See, e.g., Fitzgerald, 457 U.S. at 747–48.

      [272].     See, e.g., Plyler v. Doe, 457 U.S. 202, 221–23 (1982); Cruzan, 497 U.S. at 278–79.

      [273].     Other scholars have, however. To my mind, the best and most thorough work is Gary Lawson’s Evidence of the Law (2017). Lawson argues that the process of answering any legal question entails a standard of proof. Id. at 193. In constitutional law, the Court seems to use a best available alternative standard under which it adopts the constitutional interpretation that is better than all other options, no matter how close or uncertain the evidence. See id. at 202–03; see also John O. McGinnis & Michael B. Rappaport, Originalism and the Good Constitution 116, 142 (2013) (arguing that courts should use the “interpretive methods that the [Constitution’s] enactors would have deemed applicable,” including a general rule in favor of whichever interpretation has “the stronger evidence in its favor”). But one could easily imagine a different standard—such as clear and convincing evidence or beyond reasonable doubt—under which failure to meet the requisite evidentiary burden would trigger the Court’s admission of uncertainty and ability to decide instead on the basis of harm-avoider constitutionalism.

      [274].     Heidi Kitrosser, Interpretive Modesty, 104 Geo. L.J. 459, 466 (2016). But see McGinnis & Rappaport, supra note 273, at 142 (arguing for a lower standard of proof under which the Constitution’s meaning would be settled in favor of whichever original meaning is supported by stronger evidence).

      [275].     Relatedly, it is possible (perhaps even likely) that Justices will suffer from a degree of overconfidence bias, refusing to acknowledge the possibility of cases being hard in light of years of experience resolving close cases. I thank Richard Re for this observation. But one can imagine an institutional response that does not depend on individual Justices candidly admitting difficulty—for example, a rule under which any case that yields a 5-4 vote on the merits is “hard” and thus merits harm-avoider constitutionalism. My gratitude to Dean Andrew Guzman for this suggestion.

      [276].     To be clear, the theory could still function even if only some Justices believed it was an appropriate decision rule (perhaps because the others think it their duty to search out the meaning of “the law,” no matter how uncertain it may be). In that scenario, the Justices who follow harm-avoider constitutionalism would act as tie breakers in the event of disagreement among those who do not. I thank Richard Fallon for flagging this point.

      [277].     See, e.g., Lamps Plus, Inc. v. Varela, 139 S. Ct. 1407, 1417–19 (2019) (disagreeing with dissent’s conclusion that contra proferentem ought to apply to arbitration clause at issue).

      [278].     Cf. Kisor v. Wilkie, 139 S. Ct. 2400, 2408 (2019) (“The [Auer] deference doctrine we describe is potent in its place, but cabined in its scope.”).

      [279].     See text accompanying notes 145–48.

      [280].     See, e.g., Plyler v. Doe, 457 U.S. 202, 220 (1982) (“[Texas’s law] imposes its discriminatory burden on the basis of a legal characteristic over which children can have little control.”) (emphasis added).

      [281].     See id. at 227 (assessing whether the Texas law “furthers an interest in the ‘preservation of the state’s limited resources for the education of its lawful residents’”) (emphasis added). Professors Barzun and Gilbert helpfully refer to this as a “real party in interest” approach. See Barzun & Gilbert, supra note 57, at 33.

      [282].     Plyler, 457 U.S. at 227. Admittedly, the Plyler Court does frame some of its analysis in terms of the State’s interests, rather than the interests of the underlying groups who support the law. See, e.g., id. at 228 (asking whether “the State may seek to protect itself from an influx of illegal immigrants”) (emphasis added). This is likely attributable to the nature of strict scrutiny’s least restrictive alternative inquiry. Under that test, the Court considers whether the State can achieve its interests through other alternatives. But the State is only a stand-in for the interests of groups who support the underlying laws. Harm-avoider constitutionalism thus focuses more properly on the harms (and avoidance strategies available) to those underlying groups, i.e., the real parties in interest.

      [283].     457 U.S. 731, 757 (1982).

      [284].     See id. at 753.

      [285].     Id.

      [286].     Chayes, supra note 65, at 1291.

      [287].     Id.

      [288].     For an in-depth examination of the group identity problem, see Tang, supra note 83, at 1787–88.

      [289].     United States v. Nixon, 418 U.S. 683, 705 (1974).

      [290].     Id. at 708.

      [291].     Id.

      [292].     See, e.g., Cruzan v. Dir., Mo. Dep’t of Health, 497 U.S. 261, 286 (1990) (candidly acknowledging the harms of the Court’s decision to Cruzan’s parents).

      [293].     Consider, for example, the public backlash after the Court effectively minimized the harm suffered by slaves in Dred Scott v. Sandford by holding that they are simply the “property” of White men for purposes of the Constitution. 60 U.S. (19 How.) 393, 452 (1857).

      [294].     Groups may also try to frame their harms at too general a level of abstraction. For instance, supporters of the right to die might argue that a ruling against the Cruzans will destroy individual liberty. The Court should respond by requiring a specific description of the causal pathway—e.g., liberty may be destroyed due to the state’s refusal to permit withdrawal of a patient’s life-saving medical treatment absent a showing of clear and convincing evidence. It is this actual causal mechanism that should serve as the focus of the parties’ arguments over avoidance strategies, since undoing the mechanism will undo the harm. I owe thanks to Richard Fallon for identifying this concern.

      [295].     See, e.g., United States v. Windsor, 570 U.S. 744, 770 (2013) (refusing to uphold laws that have “the purpose and effect of disapproval of [a] class”); Lawrence v. Texas, 539 U.S. 558, 582 (2003) (O’Connor, J., concurring) (“Moral disapproval . . . is an interest that is insufficient to satisfy rational basis review.”).

      [296].     See Matal v. Tam, 137 S. Ct. 1744, 1764 (2017) (“Speech that demeans on the basis of race, ethnicity, gender, religion, age, disability, or any other similar ground is hateful; but the proudest boast of our free speech jurisprudence is that we protect the freedom to express ‘the thought that we hate.’”); R.A.V. v. City of St. Paul, 505 U.S. 377, 380 (1992) (holding that a city ordinance prohibiting “bias-motivated” disorderly conduct was “facially invalid under the First Amendment”).

      [297].     See supra Part I.B (describing cases where the Court compares ease of avoidance, not the severity of the harms themselves). To be sure, in at least one case—Minnesota v. Clover Leaf Creamery Co.—the Court does seem to engage in a direct comparison of the competing harms. 449 U.S. 456 (1981). But that comparison was questionable for reasons aptly offered by Justice Scalia. See supra note 160 and accompanying text.

      [298].     See supra Part I.B (showing how the Court recognizes the important interests on both sides of Cruzan, Fitzgerald, Nixon Tapes, Plyler, Mazars, Vance, and June Medical).

      [299].     See infra Part IV.A.

      [300].     See infra Part IV.A.

      [301].     See Clover Leaf Creamery Co., 449 U.S. at 473 & n.17.

      [302].     See Grace Chen, Public School Boards Demystified: How Parents Can Influence the Board’s Decisions, Pub. Sch. Rev. (Aug. 6, 2018), https://www.publicschoolreview.com/blog/public-school-boards-demystified-how-parents-can-influence-the-boards-decisions [https://perma.cc/E3F5-C9S2]. This inquiry was relevant because while Texas’s law permitted school districts to exclude undocumented children, it was actually the school districts themselves who chose whether to do so. See Plyler v. Doe, 457 U.S. 202, 205 (1982).

      [303].     See Plyler, 457 U.S. at 219–20.

      [304].     See Dred Scott v. Sandford, 60 U.S. (19 How.) 393 (1857); Hammer v. Dagenhart, 247 U.S. 251 (1918).

      [305].     Constitutional errors are far from rare. Using one conservative measure—the Supreme Court’s own confessions of error—the Court has issued 321 mistaken rulings thus far. See Cong. Rsch. Serv., The Constitution of the United States of America: Analysis and Interpretation, S. Doc. No. 112–9, at 2603–15 (2016), https://www.govinfo.gov/content/pkg/GPO-CONAN-REV-2016/pdf/GPO-CONAN-REV-2016.pdf [https://perma.cc/6MB8-CULF]. The list is underinclusive in one sense and overinclusive in another. It is underinclusive because it only includes express overrulings and their “functional equivalent,” leaving out implicitly overruled cases. Id. at 2603. The list is overinclusive because a few of the listed cases involve non-constitutional rulings.

         [306].     See Jonathan Cantarero, The Ethics of Civil Commitment, 16 J. Health & Biomedical L. 105, 115 (2020) (describing Popper’s philosophy of “negative utilitarianism” as advocating “the responsibility to prevent the greatest amount of harm or evil”).

      [307].     Regan, supra note 53, at 302.

      [308].     See Daniel K. Sokol, “First Do No Harm” Revisited, Brit. Med. J., Nov. 2013, at 23.

      [309].     Press Release, Quinnipiac Univ. Poll, U.S Voters Still Say 2-1 Trump Committed Crime 8 (May 2, 2019), https://poll.qu.edu/images/polling/us/us05022019_uwdh82.pdf [https://perma.cc/AE8S-GLJ2].

      [310].     See Eric J. Segall, Supreme Myths: Why the Supreme Court is Not A Court and Its Justices are Not Judges, at ix (2012).

      [311].     See Adam Feldman, Empirical SCOTUS: Differences Between “Obama” and “Trump” Judges, While Sometimes Subtle, Can’t Be Denied, SCOTUSblog (Dec. 4, 2018) (“[P]artisanship, at least in terms of the appointing president, helps to dictate the decisions of federal judges in complex cases moving through the federal judicial hierarchy.”).

      [312].     I do not mean to suggest that harm-avoider constitutionalism will magically clear all partisanship concerns away. Indeed, there is reason to expect that the Justices’ views on who the best harm avoider is in any given case will itself be tainted by partisanship. I respond to this objection infra Part IV.A.

      [313].     See generally John Strate, Timothy Kiska, & Marvin Zalman, Who Favors Legalizing Physician-Assisted Suicide? The Vote on Michigan’s Proposal B, 20 Pol. & Life Sci. 155 (2001) (“Attributes associated with social conservatism—Republican Party identification and frequent church attendance—led to decreased odds of a ‘yes’ vote [on expanding physician-assisted suicide].”).

      [314].     See Cruzan v. Dir., Mo. Dep’t of Health, 497 U.S. 261 (1990).

      [315].     See Nixon v. Fitzgerald, 457 U.S. 731 (1982).

      [316].     See Plyler v. Doe, 457 U.S. 202 (1982).

      [317].     See June Med. Servs. L.L.C. v. Russo, 140 S. Ct. 2103 (2020); United States v. Nixon, 418 U.S. 683 (1974).

      [318].     See supra text accompanying notes 167–208.

      [319].     See supra Part II.A (describing public and private avoidance strategies).

      [320].     Amelia Thomson-DeVeaux & Oliver Roeder, Is The Supreme Court Facing A Legitimacy Crisis?, FiveThirtyEight (Oct. 1, 2018), https://fivethirtyeight.com/features/is-the-supreme-court-facing-a-legitimacy-crisis/ [https://perma.cc/4KZH-H7WS].

      [321].     See supra Part III.B.

      [322].     See Citizens United v. Fed. Elections Comm’n, 558 U.S. 310 (2010).

      [323].     Greg Stohr, Bloomberg Poll: Americans Want Supreme Court to Turn Off Political Spending Spigot, Bloomberg (Sept. 28, 2015), https://www.bloomberg.com/news/articles/2015-09-28/bloomberg-poll-americans-want-supreme-court-to-turn-off-political-spending-spigot [https://perma.cc/JBU6-WDQ5] (finding 78 percent of Americans say Citizens United should be overturned).

      [324].     Adam Liptak, One by One, Protesters Interrupt Supreme Court, N.Y. Times (Jan. 21, 2015), https://www.nytimes.com/2015/01/22/us/one-by-one-protesters-interrupt-supreme-court-over-citizens-united-case.html [https://perma.cc/2SXD-CET9]; Adam Liptak, Rare Supreme Court Outburst Caught on Even Rarer Videos, N.Y. Times (Feb. 27, 2014), https://www.nytimes.com/2014/02/28/us/rare-supreme-court-outburst-caught-on-even-rarer-videos.html [https://perma.cc/8DHP-JFT7].

      [325].     Citizens United, by contrast, is quite consciously not a harm-avoider opinion. The Court instead held that the First Amendment forbids the government to limit corporate campaign expenditures—a plain (though also contestable) first-order based ruling. See Citizens United, 558 U.S. at 365–66.

      [326].     See Lewin, supra note 256; Schiavo Protesters’ Passions Rise, CBS News (Mar. 27, 2005), https://www.cbsnews.com/news/schiavo-protesters-passions-rise/ [https://perma.cc/5NHM-8GPJ].

      [327].     See supra notes 256–258 and accompanying text.

      [328].     See Lewin, supra note 256.

      [329].     See supra Part II.A (describing public and private avoidance options).

      [330].     The same virtues may well reveal themselves in the aftermath of Mazars, Vance, and June Medical, although it is too soon to judge for certain.

      [331].     576 U.S. 644, 719 n.22 (2015) (Scalia, J., dissenting).

      [332].     138 S. Ct. 2448, 2502 (2018) (Kagan, J., dissenting).

      [333].     Id. at 2460 (majority opinion).

      [334].     Consider the tenor of argument in Masterpiece Cakeshop, Ltd. v. Colorado Civil Rights Commission, 138 S. Ct. 1719 (2018), which considered whether a religious baker possessed a First Amendment right to refuse to bake a cake for a gay couple’s wedding notwithstanding a contrary obligation under state anti-discrimination law. The baker asserted that the gay couple had failed even to advance a “plausible reading of the First Amendment.” Reply Brief for Petitioners at 18, Masterpiece Cakeshop, 138 S. Ct. 1719 (No. 16-111). Meanwhile, the couple responded by belittling the baker’s expressive interest in his wedding cakes as “far-fetched” and “misguided.” Brief for Respondents Charlie Craig and David Mullins at 19, Masterpiece Cakeshop, 138 S. Ct. 1719 (No. 16-111).

      [335].     See Pamela S. Karlan, Foreword: Democracy and Disdain, 126 Harv. L. Rev. 1, 68 (2012) (arguing that the Justices display a growing sense of their “superiority at resolving constitutional controversies” that is rooted in constitutional theory’s increasing confidence that “it can deliver ‘right answers’ to even difficult constitutional questions”) (quoting Richard A. Posner, The Rise and Fall of Judicial Self‑Restraint, 100 Calif. L. Rev. 519, 535–36 (2012)).

      [336].     Jamal Greene, Foreword: Rights as Trumps?, 132 Harv. L. Rev. 28, 34 (2018). To be clear, Greene’s target is the Court’s categorical approach to rights, which he criticizes in favor of a proportionality frame. See id. But the underlying concern is the same: when parties are locked into their positions and incentivized to categorically dismiss their opponents’ positions, the result is a process of constitutional argument that exacerbates societal conflict. See id.

      [337].     See supra Part II.A (describing public and private avoidance options). In this respect, harm-avoider constitutionalism may share a virtue with Professor Bill Eskridge’s pluralism-facilitating theory of judicial review: it lowers the stakes of politics. See William N. Eskridge, Jr., Pluralism and Distrust: How Courts Can Support Democracy by Lowering the Stakes of Politics, 114 Yale L.J. 1279, 1303–10 (2005) (arguing that courts should clear out obsolescent laws and refuse to sanction state animus towards out-groups).

      [338].     The Court’s prior ample experience with the kind of factfinding needed for successful harm-avoider reasoning suggests that this reorientation would be feasible. See, e.g., June Med. Servs. L.L.C. v. Russo, 140 S. Ct. 2103, 2122–32 (2020) (discussing trial court findings that formed the basis of harm-avoider conclusions); Whole Woman’s Health v. Hellerstedt, 136 S. Ct. 2292, 2301–02 (2016) (same). One potential wrinkle is that harm may be imposed on groups not technically before the Court as parties. In such cases, a court might assist further factfinding by recognizing party-interveners and relying on amicus briefs. See Fed. R. Civ. P. 24.

      [339].     For example, one might argue that it would be easier for corporations to repeal BCRA’s expenditure ban than it would be for campaign finance supporters to overturn Citizens United via constitutional amendment. See Tang, supra note 83, at 1810, 1810 n.333.

      [340].     See, e.g., Citizens United v. Fed. Elections Comm’n, 558 U.S. 310, 415 (2010) (Stevens, J., concurring in part) (describing the segregated fund alternative in which corporate stockholders and their families can “pool their resources to finance electioneering communications”).

      [341].     See Limits on Contributions Received by the SSF, Fed. Election Comm’n, https://www.fec.gov/help-candidates-and-committees/taking-receipts-ssf/limits-contributions-received-ssf/ [https://perma.cc/48MT-ZU98].

      [342].     See supra Parts III.A–C.

      [343].     See generally Barzun & Gilbert, supra note 57 (arguing that courts should rule against the party that could have most easily prevented a constitutional conflict from materializing in the first place).

      [344].     Notice that the question of abortion’s place as a constitutional right may be difficult on harm-avoider grounds, even if the particular analysis of Louisiana’s admitting privileges law was not. It was only because Louisiana did not seek to overturn Roe and Casey that June Medical Services was decided as a simple harm-avoider decision. See June Med. Servs. L.L.C. v. Russo, 140 S. Ct. 2103, 2135 (2020) (Roberts, J., concurring).

      [345].     See, e.g., Adam Rogers, Abortion Bans Create a Public Health Nightmare, Wired (May 21, 2019), https://www.wired.com/story/abortion-bans-create-a-public-health-nightmare/ [https://perma.cc/2NX5-MQWC] (“Making abortion mostly illegal will kick off an unintentional, vast experiment in public health—one where the outcomes are sick or dying women and children.”).

      [346].     See National Right to Life Mission Statement, Nat’l Right to Life, https://www.nrlc.org/about/mission/ [https://perma.cc/EP5K-RBA5].

      [347].     Indeed, to this point, only 4 percent of all Americans oppose birth control. Very Few Americans See Contraception as Morally Wrong, Pew Rsch. Ctr. (Sept. 28, 2016), https://www.pewforum.org/2016/09/28/4-very-few-americans-see-contraception-as-morally-wrong/ [https://perma.cc/6T3A-G46D].

      [348].     See Christina A. Cassidy, As Abortion Laws Grow More Restrictive, Tennessee Ranks Among Top States Providing Services for Out-of-State Women: AP, Tennessean (Sept. 8, 2019), https://www.tennessean.com/story/news/health/2019/09/08/abortion-laws-us-women-traveling-states/2232711001/ [https://perma.cc/59D6-R6X8].

      [349].     See, e.g., June Med. Servs., 140 S. Ct. at 2130 (describing “disproportionate[]” burdens of an abortion regulation on “poor women, who are least able to absorb them”).

      [350].     See supra notes 1–5; see also Tang, supra note 64 (describing other first- and second-order theories of interpretation).

      [351].     See supra Part I.B.

      [352].     A related workability critique is that identifying the group with the best avoidance strategies will strain the courts’ factfinding capabilities. As noted above, however, the Supreme Court’s successful use of this approach in prior cases suggests this concern may be overblown. See supra note 339. Moreover, courts analyze the question whether a given side’s harm can be easily avoided with surprising frequency: such an analysis is the backbone of the “irreparable harm” prong of both the preliminary injunction standard and the standard for issuance of a stay pending appeal. See, e.g., Winter v. Nat’l Res. Def. Council, Inc., 555 U.S. 7, 22 (2008) (standard for determining irreparable harm in preliminary injunction proceedings); Portia Pedro, Stays, 106 Calif. L. Rev. 869, 870 (2018) (same for stays pending appeal). Indeed, the Court itself sometimes compares the avoidability of each side’s harm in adjudicating the irreparable harm factor. See, e.g., Ross v. Nat’l Urban League, 141 S. Ct. 18, 20–22 (2020) (Sotomayor, J. dissenting) (arguing that whereas the government could easily avoid its harm by “adding resources to accelerate data processing,” the plaintiffs would not be able to remedy the harm of a census undercount “for at least the next 10 years,” when the next census is conducted).

      [353].     See generally Dan M. Kahan, Foreword: Neutral Principles, Motivated Cognition, and Some Problems for Constitutional Law, 125 Harv. L. Rev. 1 (2011) (describing the problem of motivated reasoning in Supreme Court decisionmaking and proposing particular communication techniques to mitigate it).

      [354].     See Cruzan v. Dir., Mo. Dep’t of Health, 497 U.S. 261, 283 (1990) (“An erroneous decision to withdraw life-sustaining treatment, however, is not susceptible of correction.”).

         [355].     See supra Part II.A.2 (describing the private harm-avoidance strategies successfully utilized after Cruzan).

      [356].     See supra Part III.C.

      [357].     See supra Part III.D.

      [358].     Allison Orr Larsen & Neal Devins, The Amicus Machine, 102 Va. L. Rev. 1901, 1952 (2016); see also Henry Paul Monaghan, On Avoiding Avoidance, Agenda Control, and Related Matters, 112 Colum. L. Rev. 665, 668–69 (2012) (arguing that the Court has “embraced . . . the law declaration model” in which the Court emphasizes “the judicial role in saying what the law is”).

      [359].     See Cass R. Sunstein, One Case At A Time: Judicial Minimalism on the Supreme Court 10 (1999).

      [360].     Id. at ix–x. Harm-avoider constitutionalism differs significantly from minimalism in that the latter theory still requires courts to issue difficult and contentious first-order judgments on the Constitution’s meaning—judgments that may have the divisive effect of rejecting some groups’ claims for constitutional recognition.

      [361].     See supra notes 160–164 and accompanying text.

      [362].     See, e.g., W. Lynn Creamery v. Healy, 512 U.S. 186, 200–01 (1994) (recognizing the absence of powerful in-state industries to oppose a challenged law).

      [363].     Another potential critique of harm-avoider constitutionalism’s workability is that the costs of avoidance in constitutional cases (e.g., whether the dairy industry can move for repeal of a milk packaging law) cannot be readily reduced to monetary terms. But this is true of private law harm-avoider cases, too. Courts do not try to quantify the financial cost to rear drivers of leaving more distance to the cars in front of them, nor do they try to compare those costs to any supposed monetary cost that leading drivers would encounter from stopping more suddenly. See supra Part I.A. These private law rules are still workable, though, because ease of avoidance is something courts can compare directly using a measure of common sense. See id. The Court’s own track record in cases like Cruzan, Fitzgerald, and others suggests the same comparative exercise is possible in constitutional law.

      [364].     See James B. Thayer, The Origin and Scope of the American Doctrine of Constitutional Law, 7 Harv. L. Rev. 129, 144 (1893) (“[Courts] can only disregard the Act when those who have the right to make laws have not merely made a mistake, but have made a very clear one, —so clear that it is not open to rational question. That is the standard of duty to which the courts bring legislative acts . . . .”).

      [365].     Plyler v. Doe, 457 U.S. 202, 222 (1982); see also supra Part I.B.1.iii.

      [366].     Plyler, 457 U.S. at 229.

         [367].     See supra note 263 and accompanying text.

      [368].     See supra notes 248–249.

         [369].     Technically, Mazars involved a challenge to a House legislative subpoena rather than fully enacted legislation. Yet the Court still ruled in favor of the challenger on the ground that the House had easier avoidance strategies. See supra Part I.B.2.i.

      [370].     See Trump v. Mazars USA, LLP, 140 S. Ct. 2019, 2035–36 (2020).

      [371].     I have gone so far as to defend it in other work. See Tang, supra note 82 (arguing that laws burdening politically powerful groups warrant special deference).

      [372].     Among other things, such an inquiry is almost impossible to divorce from the question whether voters support the current President—and that is an inquiry that typically divides closely along partisan lines. See Jon R. Bond & Richard Fleisher, The Polls: Partisanship and Presidential Performance Evaluations, 31 Presidential Studs. Q. 529, 529 (2001) (finding “evidence of a more general tendency for partisanship to color assessments of the [P]resident”).

      [373].     For example, 75 percent of Republicans describe illegal immigration as a “very big problem,” compared to only 19 percent of Democrats. Little Partisan Agreement on the Pressing Problems Facing the U.S., Pew Rsch. Ctr. (Oct. 15, 2018), https://www.people-press.org/2018/10/15/little-partisan-agreement-on-the-pressing-problems-facing-the-u-s/ [https://perma.cc/2SKQ-VT6U].

      [374].     Since 1990, between 79 percent and 84 percent of Americans have expressed support for laws permitting patients to decide whether to continue life-saving medical treatment. See Strong Public Support for Right to Die, supra note 260.

      [375].     528 U.S. 119, 124 (2000).

      [376].     I do not mean to minimize the very real possibility that certain persons would feel legitimately in personal danger upon seeing “a four-car caravan” of police officers. Id. But as Justice Sotomayor recently noted, this is a situation when personal safety actually coincides with the best strategy for avoiding a warrantless Terry stop. See Utah v. Strieff, 136 S. Ct. 2056, 2070 (2016) (Sotomayor, J., dissenting) (“For generations, [B]lack and [B]rown parents have given their children ‘the talk’—instructing them never to run down the street; always keep your hands where they can be seen; do not even think of talking back to a stranger—all out of fear of how an officer with a gun will react to them.”).

      [377].     See Ely, supra note 3, at 73–104.

      [378].     See id. at 103 (arguing that courts should intervene aggressively “when the ins are choking off the channels of political change” and when “representatives . . . are systematically disadvantaging some minority out of simple hostility”).

      [379].     Id. at 92 (“[The Constitution is] overwhelmingly dedicated to concerns of process and structure and not to the identification and preservation of specific substantive values.”); id. at 103 (expressing concern for situations where majorities deny minority groups “the protection afforded other groups by a representative system”).

      [380].     See supra Part III.A.

      [381].     Ely, supra note 3, at 152; see also id. at 146 (describing “special scrutiny” of laws relying on suspect classifications). I am setting aside laws burdening political access.

      [382].     See Tang, supra note 83, at 1757 (“Political power’s presence, I want to suggest, is a good reason for judges to defer to democratically enacted laws, even if one thinks its absence is a bad reason to strike laws down.”).

      [383].     Plyler is an example. Even though undocumented immigrant children are not a suspect class, their relative lack of power—their inability to find alternative educational options—still matters for harm-avoider constitutionalism. See supra Part I.B.3.

      [384].     See Ely, supra note 3, at 103 (arguing that court should “intervene[] only when the . . . political market is systematically malfunctioning”) (emphasis added).

      [385].     See Barzun & Gilbert, supra note 57, at 3–4.

      [386].     See id. at 5 & n.15 (identifying harm-avoider constitutionalism as “the clearest example” of a proposal that, like their own, “sound[s] in cost avoidance”). Note that Barzun and Gilbert cite to an earlier, unpublished version of this paper first circulated in the fall of 2019. See id.

         [387].     See id. at 8.

         [388].     See id. at 5.

      [389].     Compare id. at 5 (“[T]he conflict-avoidance principle encourages parties to avoid the sorts of conflicts that produce hard cases.”), with Parts IV.D and IV.E (describing virtues of bolstered institutional legitimacy and a more generative culture of constitutional argument).

      [390].     Cf. Charles L. Barzun, The Positive U-Turn, 69 Stan. L. Rev. 1323, 1323 (2017) (describing the “positive turn” in constitutional law).

      [391].     See supra Part I.B.

      [392].     See Barzun & Gilbert, supra note 57, at 6 (“Why haven’t judges . . . already applied [cost avoidance] to constitutional law?”). Barzun and Gilbert do argue that certain aspects of their theory bear similarities to other kinds of inquiries in constitutional law. See, e.g., id. at 20 (explaining how conflict avoidance “resemble[s] the conditions that federal courts require for parties to have ‘standing’ to sue”).

      [393].     See id. at 26 n.71 (arguing that harm-avoider constitutionalism “would place onerous information demands on courts,” whereas “the information demands of [the conflict-avoidance principle] are lighter”).

      [394].     See supra Part I.B (describing cases in which the Court has already relied on harm-avoider constitutionalism); see also supra Part IV.A (responding to the argument that harm-avoider constitutionalism is unworkable).

      [395].     See Barzun & Gilbert, supra note 57, at 33–34.

      [396].     Id. at 32. In most constitutional litigation, of course, the defendant will be a government entity.

      [397].     573 U.S. 682, 693 (2014).

      [398].     Barzun & Gilbert, supra note 57, at 33.

      [399].     See Brown v. Bd. of Educ., 347 U.S. 483, 495 (1954).

      [400].     Barzun & Gilbert, supra note 57, at 49.

      [401].     See id. (candidly recognizing that under the authors’ proposal, “rapid advances in equal rights for the LGBTQ community . . . might stall”).

      [402].     See supra Part III.D. In this sense, the retrospective focus of Barzun and Gilbert’s principle may be problematic, for it means once a conflict has materialized, there is no longer any reason for the parties to cooperate in pursuit of some mutually agreeable solution. In harm-avoider constitutionalism, the incentive to compromise and render one’s opponent the better harm avoider persists throughout the litigation.

      [403].     See Steven Levitsky & Daniel Ziblatt, How Democracies Die 97–117 (2018) (arguing that the norm of mutual toleration is essential to preserve our democracy).

      [404].     See supra Part I.

      [405].     See supra Part I.

      [406].     See supra note 82 and accompanying text.

      [407].     See supra Part I.B.2.

      [408].     See Baude, supra note 1, at 2351–63.

      [409].     See Strauss, supra note 2, at 33–98.

      [410].     See supra Part I.A.

      [411].     See supra Part III.D.

More in this Issue

The California Act to Save [Black] Lives? Race, Policing, and the Interest-Convergence Dilemma in the State of California

In January 2020, the California Act to Save Lives became law, raising the state’s standard for justifiable police homicide to cover only those police homicides that were “necessary in defense of human life.” Although the Act was introduced in the wake of protests against officer-involved shootings of Black and Latinx people, the Act itself does […]

The Legality of Ranked-Choice Voting

With the rise of extreme polarization, intense political divisiveness, and gridlocked government, many Americans are turning to reforms of the democratic processes that create incentives for candidates and officeholders to appeal to broader coalitions. A centerpiece of these efforts is ranked-choice voting (RCV). RCV allows voters to rank candidates in order of preference: first, second, […]

Democratizing the Supreme Court

Progressives are taking Supreme Court reform seriously for the first time in almost a century. Owing to the rise of the political and academic left following the 2008 financial crisis and the hotly contested appointments of Justices Neil Gorsuch and Brett Kavanaugh, progressives increasingly view the Supreme Court as posing a serious challenge to the […]

Rethinking Disorderly Conduct

Disorderly conduct laws are a combination of common law offenses aimed at protecting the public order, peace, and tranquility. Yet, contrary to common legal conceptions, the criminalization of disorderly conduct is not just about policing behavior that threatens to disrupt public order or even the public’s peace and tranquility. Policing disorderly conduct reflects and reinforces […]