Volume 166


On January 30, 2017, President Donald J. Trump signed Executive Order 13,771, which directs each agency to repeal at least two existing regulations before issuing a new regulation (referred to as the “one‐in‐two‐out” requirement) and imposes a regulatory budget that sets a cap on total incremental regulatory costs (set at zero for fiscal year 2017). The regulatory budget concept has been kicked around for decades, while the one‐in‐two‐out requirement is more recent and has been implemented in Canada, the United Kingdom, and Australia to various extents. Legal scholars and commentators have been quick to opine on the Order, with some pointing out ways in which it is irrational or impractical and others defending aspects of the Order such as the imposition of a regulatory budget.

In this Essay, we take a somewhat different approach by evaluating how well the Order is likely to achieve its purpose of helping agencies “be prudent and financially responsible in the expenditure of funds.” Although vaguely laudable, this purpose is illaudably vague. We will therefore ground our analysis by defining the goal (or goals) of the Order according to priorities that have been adopted by prior administrations or promoted by scholars, commentators, or interest groups. For purposes of this analysis, we take no normative position on whether these end goals are desirable.

The three potential goals that we evaluate the Order against are: (1) increasing the net benefits of regulations, (2) decreasing regulatory burdens, and (3) increasing presidential control over agencies. We then compare the Order against regulatory reform efforts in other countries. We conclude that the Order is unlikely to sensibly (much less prudently or responsibly) achieve any of these goals without significant changes.

The efficiency gap has recently been touted as a general partisan fairness measure with the ability to “neatly slice the Gordian knot the Court has tied for itself, explicitly replying to the Court’s ‘unanswerable question’ of ‘[h]ow much political . . . effect is too much.’” The measure was endorsed by the district court in Whitford v. Gill, and is currently on appeal to the Supreme Court. The plaintiffs in Whitford, based on a forty‐two‐year analysis of state legislative elections, have proposed a “conservative” 7% threshold on the efficiency gap as a standard for judging the constitutionality of partisan gerrymandering. If the Court adopts the efficiency gap as a partisan fairness measure, this decision could have far‐reaching implications for redistricting practices and litigation.

We examine the properties of the efficiency gap as a measure of partisan unfairness. With the right to vote, the Court has consistently sought to guard against sophisticated as well as simple‐minded modes of discrimination. We explore whether the efficiency gap is up to this task by first considering the mathematical properties of the efficiency gap. Does it provide a consistent and stable interpretation across electoral maps (for different states, for different types of elections, e.g. Congress or state houses, and for different time periods), allowing us to compare the value of partisan fairness from one map to another? We then evaluate whether the measure is impervious to the data that is used to compute it. We finally explore its ability to tap the concepts of responsiveness and bias, separate and important facets of partisan fairness.

On June 23, 2017, the Supreme Court’s 5–3 decision in Murr v. Wisconsin provided some long‐awaited guidance for courts to use in defining the relevant unit of property to perform regulatory takings analyses. The Court announced a new three‐part test, tasking lower courts with two different multifactor inquiries to determine whether a federal regulatory taking has occurred. First, the court must figure out the right unit of property, and second, it must determine whether just compensation is owed using the three‐pronged test articulated forty years ago in Penn Central Transportation Co. v. City of New York. Despite its resort to multiple factors to determine the relevant unit of property, Murr is most striking because of an important one it minimizes: state‐specific positive law. Takings doctrine has long valued, incorporated, and awarded protection to varied unique forms of state property. This constitutional property federalism is widely perceived as desirable, encouraging beneficial competition and innovation in the forms and content of property rights in different jurisdictions. This Essay argues that the Murr decision threatens constitutional property federalism by inviting courts to define the scope of protected property interests using law and regulation across jurisdictional boundaries. As lower courts and scholars struggle to give meaning to Murr, its effects on the federalist structure of property should not be neglected.

In a perverse twist, the Supreme Court’s 2017 Star Athletica copyright decision might have more significant effects on design patent protection and the larger intellectual property landscape than it will on copyright law’s useful article doctrine. As Professor Jane Ginsburg foretold, the works at issue in Star Athletica provided a poor vehicle for addressing the thorny problems bedeviling the useful article separability doctrine. Unfortunately, in their zeal to nudge the Court toward tighter limits on copyright eligibility for useful articles, several intellectual property scholars and advocates mistakenly argued that design patent protection extended to functional features of articles of manufacture as a means to persuade the Court that it should tighten copyright law’s useful article separability doctrine. The gambit backfired.

According to the majority in Star Athletica v. Varsity Brands, features that are incorporated into a useful article are separable, and thus potentially copyrightable, if they “(1) can be perceived as a two‐ or three‐dimensional work of art separate from the useful article and (2) would qualify as a protectable pictorial, graphic, or sculptural work—either on [their] own or fixed in some other tangible medium of expression—if [they] were imagined separately from the useful article.”

This formulation has not been well received, primarily because it seems to allow copyright protection for many more features of useful articles than was permitted by pre‐Star Athletica doctrine, notwithstanding the wide range of tests previously used and the doctrine’s well‐recognized conceptual incoherence. I share the concern about over‐protection, even though the Supreme Court’s test bears some similarity to the one my co‐authors and I proposed in an amicus brief in the case. As we demonstrated in our brief, Congress intended the useful articles doctrine to distinguish applied art from industrial design, making the former eligible for copyright protection but excluding the latter. Courts, we argued, should operationalize that distinction by deeming features separable only when they “(1) can be extracted from the article, even if doing so would destroy the article, and (2) once extracted, would stand alone as an original pictorial, graphic, or sculptural work (PGS work) that is not a useful article.”

In Star Athletica, LLC v. Varsity Brands, Inc., the U.S. Supreme Court had an opportunity to clarify copyright law’s treatment of product designs that incorporate functionality. Its opinion failed to do so in a host of different ways. In this Essay, we address just one of the opinion’s shortcomings: its failure to adequately define and distinguish between a design’s functional and expressive features. Not only does the Court’s neglect produce uncertainty for creators, litigants, and judges, the opinion makes it substantially easier for copyright claimants to obtain protection for utilitarian aspects of designs, contrary to copyright statute and policy.

The U.S. Supreme Court granted certiorari in Star Athletica, LLC v. Varsity Brands, Inc. to decide “[w]hat is the appropriate test to determine when a feature of a useful article is protectable under Section 101 of the Copyright Act?” The Supreme Court likely granted certiorari because the Sixth Circuit identified no fewer than nine different tests that courts and commentators had proposed for resolving the issue, and the Sixth Circuit adopted a tenth. The Supreme Court then issued its own opinion on March 22, 2017, setting forth a two‐part test grounded in the text of the statute while rejecting several proposed alternatives.

But there is an antecedent question that must be answered before the Supreme Court’s two‐part test in Star Athletica may be invoked; namely, whether the pictorial, graphic, or sculptural work at issue is the design of a “useful article” in the first place. On this question, the Supreme Court’s opinion is disappointingly silent.

The parties in Star Athletica and the Court both implicitly assumed, without analysis, that the cheerleading uniforms at issue were “useful articles.” That is likely the correct conclusion; however, the “useful articles” doctrine has always been an uncomfortable “fit” for clothing, costumes, and other garments, and there will undoubtedly be future cases involving other types of allegedly useful articles in which the antecedent question is even less certain. This essay aims to patch that gap, by providing guidance on the antecedent question of “what is a useful article?”

In Star Athletica v. Varsity Brands, the Supreme Court granted certiorari to resolve confusion in the lower courts regarding the “separability” predicate to copyright protection of decorative features of useful articles. Adopting the Gordian imagery evoked by other appellate courts, the Sixth Circuit in Varsity Brands lamented “[c]ourts have twisted themselves into knots trying to create a test to effectively ascertain whether the artistic aspects of a useful article can be identified separately from and exist independently of the article’s utilitarian function.” Star Athletica involved the “surface decorations” of stripes, chevrons, and color blocks applied to cheerleader uniforms. While the Supreme Court clarified the meaning and application of the “separability” standard for decorative elements of the kind at issue in that case, the decision leaves the knots as tangled as ever when a claim of copyright concerns the entire form of a useful article.

Speaking of the Copyright Act of 1909, noted copyright scholar Benjamin Kaplan had this to say about the role of judges therein:

“[T]he statute, like its predecessors, leaves the development of fundamentals to the judges. Indeed the courts have had to be consulted at nearly every point, for the text of the statute has a maddeningly casual prolixity and imprecision throughout. . . .

Judges, however, who in recent times have inclined against brutality, have run the risk of appearing slightly ridiculous in their tortuous interpretations.”

The Copyright Act of 1976 was designed to avoid this imprecision and overt reliance on judicial creativity. Supposedly comprehensive in its coverage, and the result of delicate compromises, its provisions were crafted to avoid the problems that had brought its predecessors into disrepute.

Nevertheless, an examination of the Supreme Court’s recent copyright jurisprudence—most of which has involved its interpreting the statute’s directives—leaves one with the impression that the Act has had little success in achieving its objectives of textual clarity and precision. And the Court for its part, has played an important role in supplementing the statute with its decisions, in an effort to offer litigants and lower courts operating in the area sufficient direction.

All the same, the Court’s copyright jurisprudence, self‐styled for the most part as an interpretive exercise, is not without its own significant shortcomings. In an effort to interpret the statute (and its accompanying legislative intent), the Court’s holdings are often themselves the source of immense ambiguity, which has only served to further compound the underlying problem of textual imprecision. Leaving aside the theoretical problems with the ambiguities in the Court’s jurisprudence, these decisions have done little to produce a nationally “uniform” copyright jurisprudence, one of the motivating goals of the 1976 copyright regime. Laudable as the Court’s foray into squelching the imprecision of the statute may have been, its copyright jurisprudence has unfortunately begun to entrench an additional layer of interpretive ambiguity and confusion. And in the last two decades, only once has the Court directly revisited the ambiguity generated by its own decision, meaning that much of the uncertainty has fallen to lower courts and litigants to resolve in their everyday application and invocation of copyright law.

The Court’s foray into the protectability criteria for “designs of useful articles” under the Act represents its most recent effort in this direction. Called upon to examine the circumstances under which a design embodies “pictorial, graphic, or sculptural features” that are separable from the useful aspects of the design, the Court entered into murky waters characterized by nine plausible alternatives to choose from(!).The Court’s solution was to offer its own reading of the statute, and with it an altogether new test for the question—one that none of the lower courts had previously adopted.

While prior efforts to understand the test of separability had self‐consciously engaged the policy goals of the copyright regime and related factors, the Court’s majority opinion in Star Athletica took an altogether different tack. Announcing that the matter had to be resolved without a “free‐ranging search for the best copyright policy,” it made clear that the issue was “solely” a matter of “statutory interpretation,” where the Court had to give effect to “the clear meaning of [the] statute[] as written.” And from there, Justice Thomas, the author of the majority opinion, proceeded to derive the Court’s test as though it flowed seamlessly from the words of the statute in self‐evident terms: first, “look at the useful article and spot some two‐ or three‐dimensional element that appears to have pictorial, graphic, or sculptural qualities,” and second, “determine that the separately identified feature has the capacity to exist apart from the utilitarian aspects of the article.” Simple enough. With its new test, the Court also jettisoned the distinction between “physical” and “conceptual” separability, a distinction that most lower courts had used, relying on the legislative history accompanying the statute. To the Court, since the “text of the statute” made no reference to this distinction, it was altogether irrelevant regardless of what the legislative history said.

Thus, in one fell swoop the Court undid the jurisprudence of separability that courts around the country had developed in common law fashion over the last four decades, all in the name of interpreting the clear and plain textual meaning of the statute. Of course, the Court didn’t stop there, but proceeded to apply its own test to the rather uncomplicated set of facts in the case before it, only to conclude that the test had indeed been satisfied. While the Court’s seemingly straightforward test was tailor‐made for the unduly simple set of facts that had motivated the case before it, a point that the concurring opinion of Justice Ginsburg emphasized, the majority opinion spent little time thinking through how its test might apply to future—and more complicated—scenarios. It therefore did scantily little to guide courts and litigants on future applications and extensions of its clear meaning approach to separability. Much like its other copyright opinions, the Court’s ruling is therefore likely to put significant pressure on lower courts to understand, apply, extend, and “interpret” its holding for future circumstances.

It is with this recognition that the University of Pennsylvania Law Review Online invited a renowned group of copyright and intellectual property scholars to reflect on the Court’s reasoning in the opinion. The objective at hand was not to provide an academic/theoretical commentary on the Court’s test and holding, but instead to assess the practical import of the opinion across a range of questions and dimensions, unique to each author’s interests and expertise. Put simply, the goal here was to translate, guide, and clarify the content of the Court’s opinion in Star Athletica for courts, and future litigants.

This Online Symposium is the first in a series of similar efforts that will endeavor to capitalize on the Law Review’s online presence to produce timely and relatively short commentary on prominent Court opinions and legal topics in order to aid the legal profession. As such, it represents an intervention in the domain of what Judge Harry Edwards termed “practical scholarship,” an idea that others have further developed. The effort endeavors to provide timely scholarship that will be of “particular help” and “interest to the members of the practice of the bar or judges.”

In this Essay, we assert that, despite what is written in the collective bargaining agreement (CBA), NBA players should—and in fact do—maintain their legal rights under employment discrimination statutes. At first blush, allowing players to sue the team or the NBA itself may seem to impose an unwieldy burden on professional sports. However, current disability discrimination laws like the Americans with Disabilities Act (ADA) already strike the proper balance between the employer’s interest in maintaining a safe workplace and the employee’s interest in being free from discrimination. As we will demonstrate, many ADA claims will only be cognizable when the team chooses to ignore a Fitness‐to‐Play decision in favor of the player.

This Essay explores the rights of current NBA players regarding their Fitness‐to‐Play requirements. In Part I, we outline recent changes to the CBA, highlighting how the newly effective agreement governs Fitness‐to‐Play determinations. Part II turns to players’ rights as employees under disability discrimination statutes, using the ADA as a case study. Lastly, Part III explores the interplay between the policies and procedures articulated in the CBA and NBA players’ statutorily protected antidiscrimination rights. We conclude that, at a minimum, the NBA and the NBPA should eliminate the CBA provision that requires players to waive their antidiscrimination rights. Generally speaking, sophisticated parties should be left to privately agree as to how to resolve issues that arise between them. By contrast, because parties cannot prospectively waive their statutorily guaranteed rights, a blanket exemption from employment discrimination laws for professional sports is a question for legislatures, not for the collective bargaining process. With that said, we propose that a more meaningful and reliable alternative to the current regime would be to make the Fitness‐to‐Play proceedings—however defined—binding not only on the players, but also on the team and the NBA.

This Essay highlights emerging governance problems presented by persistent Unicorns. It argues that recent market trends and deregulatory reforms have weakened or eliminated the principal mechanisms that imposed discipline on start‐up company founders. Recent scandals at prominent Unicorns suggest that investors have erred in placing blind faith in the honesty and capabilities of start‐up founders. Policymakers should learn from these disasters and close regulatory loopholes that allow Unicorns to persist in limbo between private and public status for extended periods of time.

Part I provides an overview of how the IPO has shifted from the preferred exit strategy in the eyes of entrepreneurs to a regulatory morass to be shunned. It traces developments in the market for start‐up company shares, and regulatory reforms that facilitated the proliferation of Unicorns. Part II highlights unique governance risks posed by Unicorns, addressing both societal and investor protection concerns. Part III offers suggestions on how to address Unicorn risks, and raises fundamental questions about the future of Unicorns in our economy.

Imagine three men—Alex, Brian, and Chris—who rob and kill a young woman in the parking lot of a Pennsylvania shopping mall. Once a subsequent investigation begins to tighten, Alex (the lookout) turns himself in, helping police arrest Brian (the driver) and Chris (the shooter). After an interrogation mimicking the classic Prisoner’s Dilemma, Brian partially confesses, exposing himself and Chris to criminal liability. At arraignment, Alex pleads guilty; Brian and Chris proceed to a joint trial in state court, where both decline to testify.

This is standard fare for courts, prosecutors, and defense attorneys. The Supreme Court has expressed a preference for joint trials, praising the resultant benefits for judicial economy and for avoiding inconsistent verdicts. And although somewhat dated, a study from the 1990s suggests nearly half of felony defendants invoke the privilege against self‐incrimination.

Yet these dueling propensities for joint trials and for not taking the stand put Chris on the wrong side of a collision with another constitutional provision: the Sixth Amendment’s Confrontation Clause. The right to confrontation means defendants must have the opportunity to crossexamine the witnesses against them. But what should the court do in this situation? Since Brian’s confession is admissible under the hearsay exception for a statement by an opposing party, the jury will hear evidence that incidentally incriminates Chris. And with Brian not testifying, Chris is unable to challenge the incrimination.

This constitutional collision iterates daily in state and federal courts across the country. The solution—developed over three Supreme Court decisions: Bruton v. United States, Richardson v. Marsh, and Gray v. Maryland—is to redact the confession so it remains fair and useful evidence against the declarant (Brian) without unfairly prejudicing the nonconfessor (Chris). But jurisdictions vary considerably in applying this doctrine, and the fractured approach has left trial courts and prosecutors with a Gordian knot.

In particular, this doctrinal confusion has caused a circuit split over neutral term substitution—a specific type of redaction that discretely replaces references to codefendants’ names with a nonspecific term or adjective to prevent incidental incrimination. And while the federal courts disfavor circuit splits generally, this split’s particular harm amplifies when federal courts review state judgments under 28 U.S.C. Section 2254. Suppose three other men—David, Ed, and Frank—rob and kill a woman in New Jersey, independent of Alex, Brian, and Chris’s crime in Pennsylvania. Although both states of course follow Bruton, Pennsylvania courts take a slightly more flexible tack: Pennsylvania allows neutral term substitution in all cases; New Jersey only allows it if there are many criminal actors, requiring more thorough redaction in other cases. After the respective prosecutors redact each confession according to their state’s standard, both sets of codefendants are convicted; the state intermediate appellate courts affirm; the state supreme courts deny review; and the U.S. Supreme Court denies certiorari. So far, so good—similar defendants similarly committed similar crimes, so they deserve similar results. But if the defendants file habeas petitions in the Third Circuit, the similarity turns on its head: the New Jersey petitioners are denied collateral relief, yet their Pennsylvania doppelgangers prevail (unless the state can show harmless error or a procedural bar), solely because the Third Circuit disfavors Pennsylvania’s approach to neutral term substitution. To be sure, this outcome raises deference questions under the Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA), but it also reveals a thorny upshot to the debate over Bruton: intracircuit splits.

This Comment seeks to resolve these jurisprudential conflicts. Part I retraces the rule’s evolution to reveal the foundation laid by the Supreme Court and its relationship to the current divide over neutral term substitution. Part II examines ways to resolve this disagreement by comparing Bruton’s functionalist approach to confrontation to the more recent formalist understanding exemplified by Crawford v. Washington. In the end, this Comment calls for the Supreme Court to adopt a rebuttable presumption allowing neutral term substitution.

Shortly after Donald Trump’s victory in the 2016 Presidential election, but before his inauguration, a group of concerned scholars organized in cities and college campuses across the United States, starting with the University of Pennsylvania, to prevent climate change data from disappearing from government websites. The move was led by Michelle Murphy, a scholar who had previously observed the destruction of climate change data and muzzling of government employees in Canadian Prime Minister Stephen Harper’s administration. The "guerrilla archiving" project soon swept the nation, drawing media attention as its volunteers scraped and preserved terabytes of climate change and other environmental data and materials from .gov websites. The archiving project felt urgent and necessary, as the federal government is the largest collector and archive of U.S. environmental data and information.

As it progressed, the guerrilla archiving movement became more defined: two organizations developed, the DataRefuge at the University of Pennsylvania, and the Environmental Data & Governance Initiative (EDGI), which was a national collection of academics and non-profits. These groups co-hosted data gathering sessions called DataRescue events. I joined EDGI to help members work through administrative law concepts and file Freedom of Information Act (FOIA) requests. The day-long archiving events were immensely popular and widely covered by media outlets. Each weekend, hundreds of volunteers would gather to participate in DataRescue events in U.S. cities. I helped organize the New York DataRescue event, which was held less than a month after the initial event in Pennsylvania. We had to turn people away as hundreds of local volunteers lined up to help and dozens more arrived in buses and cars, exceeding the space constraints of NYU’s cavernous MakerSpace engineering facility. Despite the popularity of the project, however, DataRescue’s goals seemed far-fetched: how could thousands of private citizens learn the contours of multitudes of federal environmental information warehouses, gather the data from all of them, and then re-post the materials in a publicly accessible format?

Federal records laws and policies in the U.S., like the FRA, the Presidential Records Act (PRA) of 1978 and the Freedom of Information Act (FOIA), were created to prevent government administrators from engaging in censorship, lack of transparency, and the destruction of relevant government records. The DataRescue project proves that these laws, in their current iterations, are insufficient in the digital information age, and must be reformed to ensure proper records preservation and access. This Essay describes DataRescue’s efforts and limitations, and examines the FRA, prescribing statutory updates that would safeguard online access to government records, allowing volunteers across the nation to rest easy knowing that climate change data and other information will be preserved through changing presidential administrations.

In early January, the U.S. Supreme Court heard oral argument in Husted v. A. Philip Randolph Institute. The case presented a question of pure statutory interpretation: whether Ohio’s procedure for updating its voter registration rolls violates § 8(b)(2) of the National Voter Registration Act (NVRA). At oral argument, Justice Sonia Sotomayor raised the possibility of resolving the case according to a new, potentially revolutionary canon of statutory construction: interpreting federal laws, where reasonably possible, to avoid, minimize, or prohibit racially disparate impacts. Adopting this rule would have important ramifications across numerous fields of law, raise serious constitutional questions, and represent a tremendous victory for the critical race theory movement, which has long emphasized the law’s role in perpetuating structural racial inequality. This potential canon also implicates compelling questions—questions that arose during Justice Sotomayor’s confirmation proceedings—concerning the proper role of judges.

In PHH Corporation v. Consumer Financial Protection Bureau, the D.C. Circuit upheld the constitutionality of the Consumer Financial Protection Bureau’s “novel” structure over vociferous dissent. Yet the divisive constitutional and statutory challenges to the Bureau mirror those levied against its predecessor and partner agency, the Federal Trade Commission, beginning nearly a century ago. The anxiety produced by consumer protection’s infringement on freedom of contract is not “novel.” In this Essay, we explain how the trans‐substantive constitutional and administrative law principles at play in PHH should not distract from the substantive debates about consumer protection underlying the controversy. In short, PHH should be recognized as much as a referendum on “Consumer Financial Protection” as it is on the “Bureau.”

A new hire signs an employment agreement. The agreement contains an arbitration provision. Embedded in the arbitration provision is a one‐sentence delegation clause granting the arbitrator exclusive authority to decide threshold issues of arbitrability, including whether the arbitration provision is valid and enforceable. The agreement has a general choice‐of‐law clause providing that California law applies to the entire agreement, including the arbitration provision and delegation clause. An employee based outside California sues the employer in court and argues that the arbitration provision and delegation clause are unenforceable. Should the court apply California law or the law of the employee’s home state to the enforceability analysis? Several courts have confronted and wrongly answered that question in recent litigation involving Uber. Why they got it wrong—their misapplication of the Supreme Court’s severability rule—is the subject of this essay.

The tragedy of the Parkland, Florida high school shooting shocked the nation and sent thousands of student protestors out of the classrooms and into the streets. Sadly, the nation once again found itself asking the increasingly familiar question of how such senseless tragedies can be prevented. As the search for an answer to this question continues, several avenues of response are being explored. Some have focused on a failure of the “system” and take federal and state authorities to task for not heeding the warning signs. Others are considering how society can deal more effectively with the problem of mental illness. Still others are calling for more restrictive gun laws to address this problem. These calls for action are familiar and the likely federal response is equally familiar: nothing. Federal legislative action that puts further significant limitations on gun ownership is unlikely in the short term. As a result of legislative inaction, we are now seeing a grassroots response not only from concerned individuals, but from corporations willing to take actions that they hope will lessen the likelihood of another act of gun violence by someone under the age of twenty‐one. To accomplish this, retailers such as Walmart, Dick’s Sporting Goods (DSG), Kroger, and L.L. Bean have modified store policies across the United States and will no longer sell long guns (shotguns and rifles) or ammunition to those under twenty‐one. DSG was one of the first to implement this restriction on February 28, 2018. These actions have already been met with resistance from consumers in Oregon and Michigan who allege that such policies violate state public accommodation laws. While the scope of the public accommodation laws’ protections varies among states, Oregon and Michigan are among nineteen jurisdictions that consider age to be a protected class. The first lawsuits in the nation were filed in Oregon against DSG and Walmart by Tyler Watson, a twenty‐year‐old Oregon resident who was unable to purchase a rifle due to the retailers’ newly enacted age restrictions, alleging violation of the public accommodation laws. This Essay explores the merits of this claim using Mr. Watson’s case against DSG as the illustration since it is the furthest along procedurally. After explaining why Mr. Watson is likely to prevail in court, this Essay then concludes with a discussion of the implications of this case for other jurisdictions.


In Legislating in the Shadows, Professor Walker has deftly presented one side of the drafting conversation that occurs between agency officials and congressional staff. Looking through the window of technical drafting assistance—assistance typically provided through conversations responding to congressional staff requests without Office of Management and Budget (OMB) review or approval—Professor Walker sets forth and develops wide‐ranging and nuanced findings from agency participants in these conversations. His findings describe how agency officials understand their active and nonpublic role in the formulation of statutory text from its early stages. He then cogently analyzes certain implications from the standpoint of how agencies apply their interpretive authority and power.

There is much to admire in Professor Walker’s approach. Empirically, he identifies and unpacks an underappreciated aspect of contemporary federal lawmaking. Analytically, Walker offers a rigorous assessment of doctrinal implications from a statutory interpretation standpoint. He situates both the purposivist possibilities and the deference‐related risks in the context of other scholarly efforts, including several quite recent contributions.

I approach the drafting conversation described by Professor Walker from the opposite side—that of congressional staff. In doing so, I suggest that this conversation is complex and variable in ways that go beyond what might typically be understood by agency officials. Building off of this complexity, I explore certain implications of Professor Walker’s analysis, by examining the four principal reasons that agency officials offered for responding to every congressional request for technical assistance. While I share Professor Walker’s view that agencies gain special purposivist insights from their privileged participatory position in the drafting process, I am not convinced that this privileged position is due in any meaningful way to agencies’ role as shadow legislative drafters. Consequently, I remain skeptical that the shadow drafting role creates special risks of self‐aggrandizing interpretive powers, or that it materially contributes to existing concerns regarding the amount of deference agency interpretations should receive.

In their recent article, Keith Whittington and Jason Iuliano marshal considerable evidence for the proposition that the nondelegation doctrine is little more than a myth. The authors review some two thousand U.S. federal and state cases, and acknowledge that “American courts have long recognized a basic constitutional principle that legislative powers cannot be delegated to other political actors. Yet the authors also show that this principle has had remarkably little impact in practice: delegations of lawmaking authority have been upheld at high rates throughout American history by federal and state courts alike. At the Supreme Court level, “[a] review of the Court’s treatment of challenges to federal and state statutes on the grounds that they had impermissibly delegated legislative power to nonlegislative actors does not provide much basis for thinking that there was ever a seriously confining nondelegation doctrine as part of the effective constitutional order.” In sum, notwithstanding its supposed basis in the American structural commitment to separation of powers among coequal branches of government, nondelegation doctrine has never been of much practical significance in the context of U.S. constitutional law.

But in this Response, I argue that nondelegation principles may prove to be of far greater practical significance in a different context—namely, in U.K. constitutional law. In doing so, I demonstrate that the origins of the nondelegation doctrine long predate the structural constitutional provisions and pre‐New Deal cases highlighted by Whittington and Iuliano. Indeed, questions about legislative delegations of power extend back at least as far as the reign of King Henry VIII. The Tudor monarch’s shadow is so long that parliamentary delegations of authority to amend primary legislation continue to be known as “Henry VIII clauses,” and the prospect of including such delegations in Prime Minister Theresa May’s Great Repeal Bill may well complicate Britain’s ongoing exit from the European Union. Thus, while the nondelegation doctrine may merely be a myth in the United States, it is poised to play a key role in one of the most important constitutional debates in recent memory in the United Kingdom.

As reliance on algorithmic decisionmaking expands, concerns are growing about the potential for arbitrary, unfair, or discriminatory outcomes in areas such as employment, credit markets, and criminal justice. Legal scholars have lamented the lack of accountability of these automated decision processes and called for greater transparency. They argue that the way to avoid unfair or discriminatory algorithms is to demand greater disclosure of how they operate. Accountable Algorithms resists this call for transparency, calling it “a naive solution.” Instead, it argues that technology offers tools—“a new technological toolkit”—that can better assure accountability.

One of the examples that Kroll et al. rely on to illustrate their argument is the goal of ensuring that algorithms do not discriminate. Many commentators have pointed out the risk that automated decision processes may produce biased outcomes, and in prior work, I have argued that serious policy concerns are raised when these algorithms exacerbate historic inequality or disadvantage along the lines of race, sex, or other protected characteristics—what I’ve referred to as “classification bias.” Recognizing that the precise meaning of discrimination is uncertain and contested, Kroll et al. do not try to resolve debates over the meaning of discrimination. Instead, without choosing among the competing definitions, they simply survey the available technical tools, suggesting that these tools will be more effective at ensuring nondiscrimination than calls for transparency.

Transparency involves outside scrutiny of a decision process, for example, by allowing third parties to examine the computer code or the decision criteria it implements. Auditing is another method for promoting transparency. When the goal is nondiscrimination, auditing could involve techniques to ensure that an algorithm follows a specified rule—for example, sorting must not occur based on race or sex. Alternatively, auditing for discrimination could take the form of examining inputs and outputs to detect when a decision process systematically disadvantages particular groups. The latter form of auditing does not involve direct examination of the decision process, but is useful in detecting patterns. This type of auditing, in the form of field experiments, is well established in the social science literature as a technique for testing for discrimination in decisions such as employment and consumer transactions. Auditing the effects of decisionmaking algorithms similarly offers a method of detecting when they may be biased against particular groups. Kroll et al., however, express skepticism about auditing as a strategy, arguing that it is not only technically limited, but also likely restricted by law. More specifically, they suggest that when an algorithm is found to have a disparate impact, the Supreme Court’s decision in Ricci v. DeStefano may prevent correcting for that bias.

This Essay responds to Kroll et al., arguing that, despite its limitations, auditing for discrimination should remain an important part of the strategy for detecting and responding to biased algorithms. Technical tools alone cannot reliably prevent discriminatory outcomes because the causes of bias often lie not in the code, but in broader social processes. Therefore, implementing the best available technical tools can never guarantee that algorithms are unbiased. Avoiding discriminatory outcomes will require awareness of the actual impact of automated decision processes, namely, through auditing.

Fortunately, the law permits the use of auditing to detect and correct for discriminatory bias. To the extent that Kroll et al. suggest otherwise, their conclusion rests on a misreading of the Supreme Court’s decision in Ricci. That case narrowly addressed a situation in which an employer took an adverse action against identifiable individuals based on race, while still permitting the revision of algorithms prospectively to remove bias. Such an approach is entirely consistent with the law’s clear preference for voluntary efforts to comply with nondiscrimination goals.

(Visited 881 times, 1 visits today)