The Yale Law Journal


Unemployment Insurance for the Gig Economy

26 Jan 2022

abstract. For most of its history, U.S. unemployment insurance has excluded workers lying outside the conventional employer/employee binary. That should change. This Essay argues that lawmakers should extend such benefits to workers in the gig economy. To do so, it draws from two widely accepted philosophical frameworks within tax-law scholarship—liberal egalitarianism and utilitarianism—and applies their principles to gig-worker unemployment benefits. It supports, in particular, a gig-worker unemployment-benefit system modeled on existing U.S. unemployment programs. The Essay concludes by discussing solutions to pragmatic challenges that may arise when implementing this reform.


Every year, across all fifty states, millions of laid-off employees claim unemployment-insurance (UI) benefits. Yet many other jobless Americans are not so fortunate. For most of U.S. history, UI has categorically excluded laborers whose work arrangements lie outside the conventional employer/employee binary—including independent contractors and other self-employed workers.1

In recent years, this trend has grown more serious as the numbers of workers joining the gig economy have skyrocketed. In 2017, depending on one’s definition of “gig worker,” as many as 55 million Americans worked in the gig economy (or thirty-four percent of the U.S. labor force).2 Recent projections for 2020 data reach as high as forty-three percent of the workforce.3 Over the long term, these numbers may only rise, with the COVID-19 pandemic expanding the potential for remote work.4

This Essay presents a normative argument for extending UI to gig-economy workers, broadly construed. To do so, it applies two different theories of distributive justice—liberal egalitarianism and utilitarianism—to the question of gig-worker UI. The normative tax-law literature—including that on redistributive-welfare policies5 and unemployment benefits in particular6—generally grounds its reasoning in these traditions. And as this Essay shows, commonly invoked liberal egalitarian and utilitarian principles justify gig-worker UI as much as, if not more than, standard benefits.

The primary contribution of this Essay is normative. Sorting through every implementational challenge involved with instituting gig-worker UI lies beyond its scope. The Essay aims to show, however, that distributive-justice principles support a realistically achievable version of such benefits. To that end, it specifically advances a UI regime that mirrors the funding structure and design of states’ existing unemployment programs, rather than focusing on an ideal UI scheme.7 It also identifies—and briefly sketches potential solutions to—practical problems that officials implementing such a program will likely encounter. The argument here, then, aims to rally real-world policy makers, particularly at the state level, to make current UI more inclusive.

Recent years have seen a surge in legal scholarship on the gig economy. This Essay fills a noticeable gap in that literature. A majority of gig-economy scholarship tackles questions of how to classify various types of gig workers for regulatory purposes.8 Other pieces discuss how the gig economy figures into conventional labor-law topics like collective bargaining,9 arbitration practices,10 and tax compliance.11 Still others analyze questions further afield from employment law, like the antitrust12 and privacy implications13 of gig firms. None advance anything approximating this Essay’s central prescription—extending states’ existing UI programs to gig workers—nor do they defend similar policies on distributive-justice grounds.

This Essay proceeds as follows. Part I offers a brief definition of “gig-economy workers” and an overview of states’ existing UI programs. Part II introduces two widely accepted theories of distributive justice and explains how they support UI for gig and non-gig workers alike. Part III discusses how states finance UI today and argues that gig-worker benefits, if funded similarly, are still supported by distributive justice. Part IV briefly explores implementational challenges and possible avenues for overcoming them.

I. gig workers and ui today

A. Who Are Gig Workers?

The gig economy has no single definition. Commentators use phrases like “gig job” to refer to all sorts of informal, contingent, or otherwise unconventional working arrangements.14 Still, most agree that the gig economy includes those who, pursuant to a labor agreement, provide clients with on-demand services.15 For simplicity, this Essay adopts a broad usage of the term “gig worker,” although implementing policy makers might narrow the definition. This Essay uses the term to cover self-employed, nonfarm workers—a classification which includes independent contractors.16 Not all such laborers provide on-demand services, but a great many do, and most gig-economy definitions include large numbers from this group.17 For tax and regulatory purposes, workers conventionally associated with the “gig economy”—such as Uber drivers and DoorDash deliverers—typically count as independent contractors.18

How large is this group? As of August 2021, according to Bureau of Labor Statistics figures, just over ten million Americans identified as self-employed.19 That estimate falls between the widest gig-economy definitions, which include other categories of “alternative” workers,20 and more conservative tallies of workers at “conventional” gig employers,21 like Uber or Lyft.22 By any metric, though, ten million Americans is a large group. Whether policy makers grant unemployment protections to this many workers, to a smaller subset, or to still more “alternative” laborers, gig-work benefits would affect a huge swath of the workforce.

Although this Essay classifies “gig workers” broadly, it observes a second conceptual distinction among members of this group. The gig economy’s rise has heralded a growing number of internet-based firms23 that, like ridesharing giant Uber or delivery service DoorDash, regularly hire large volumes of contractors for indefinite periods. The literature has classified this subset of firms as the “narrow” definition of the gig economy.24 In practice, firms in this category frequently play roles akin to standard employers for their contractors.25 Like most standard employers, companies like Uber offer structured, long-term jobs that many occupants perform full time.26 As will be seen, this distinction matters particularly for the normative and implementational discussions in Parts III and IV.

B. The UI Landscape

The United States has provided publicly funded UI on a national scale since the Social Security Act of 1935.27 Each state operates its own UI scheme in partnership with the federal government, with the Department of Labor overseeing the entire system.28 Federal regulations set certain minimum rules for states’ programs—including stipulations for the eligibility of seasonal workers and noncitizens29—while states control most of the details.30 States finance the bulk of their programs, with funding for cash benefits coming out of state budgets.31 The federal government reimburses states’ administrative costs.32

Millions of Americans rely on UI. As of 2019, the UI system covers nearly 145 million workers nationwide.33 Total payments that fiscal year surpassed $27 billion, with 5.5 million people newly claiming eligibility.34 Covered individuals may claim benefits only after involuntarily losing their jobs.

The benefits workers receive differ according to state law. Most states’ systems seek to replace roughly half of beneficiaries’ former wages, up to a cap,35 which can vary considerably. The highest-capped state, as of September 2021, is Massachusetts, which limits weekly payments to $823 (or $1,234 for workers with dependents); Alabama’s $275 cap is the lowest.36 Recipients’ eligibility in most states, but not all, lasts for twenty-six weeks, with adjustments for workers with uneven earnings or short work histories.37 States typically impose further restrictions on part-time workers’ eligibility, such that many states effectively disqualify them.38

Gig workers do not receive UI eligibility under any state law.39 State laws also exclude many other categories of alternative or “contingent” workers.40 For the first time ever, though, gig workers across the country gained temporary UI eligibility when Congress passed the Coronavirus Aid, Relief, and Economic Security (CARES) Act in March 2020, near the start of the COVID-19 pandemic.41

The CARES Act expanded UI in three ways. Most relevant for gig workers, the Pandemic Unemployment Assistance (PUA) program paid to extend UI to independent contractors and other self-employed workers (in addition to other categories).42 Additionally, the Federal Pandemic Unemployment Compensation program gave UI recipients a $600 weekly federal benefit, on top of states’ standard payments,43 while the Pandemic Emergency Unemployment Compensation lengthened state eligibility periods by thirteen weeks.44

PUA was set to expire at the end of 2020, but lawmakers extended it as the pandemic wore on. It finally lapsed on September 6, 2021,45 terminating millions of Americans’ benefits. As of the previous July, 5.1 million people were enrolled in UI through PUA, and an estimated 4.2 million remained so by September 6.46 Today, self-employed workers are, once again, categorically ineligible for UI.

II. the normative basis for benefits

This Essay’s normative argument focuses on two distributive-justice theories: utilitarianism and liberal egalitarianism. These two theories dominate the normative scholarship and practice of tax law, with a third tradition, libertarianism, also carrying some influence.47 Utilitarianism has long been the “standard” approach of the field,48 and its principles are orthodoxy of the economics literature on taxation, in particular.49 The liberal egalitarian tradition, by contrast, traces back to John Rawls,50 whose work, A Theory of Justice,expounded a relationship between justice, equality, and personal autonomy.51 Since Rawls, Ronald Dworkin and numerous others have written in this tradition.52

As the rest of this Part argues, UI finds normative backing in both liberal egalitarianism and utilitarianism. Moreover, the logic of both frameworks weighs in favor of extending benefits to gig workers. Given these theories’ predominance, this Part’s normative argument should hold broad-based appeal.

The final significant school of thought in normative tax scholarship—libertarianism—would likely not sanction gig-worker UI. Libertarianism, made “a major theory in political philosophy” by the writing of Robert Nozick,53 typically opposes redistribution beyond a “minimal state” for conflicting with individuals’ free, legitimate private decisions.54 Extreme adherents to this tradition might “condemn[] any form of taxation as evil.”55 Purely libertarian principles, therefore, would likely deny UI in general56—let alone for UI for gig workers—although some scholars argue they support basic social safety nets more than initially apparent.57

This Essay, then, does not take up arguments from the libertarian tradition. But this choice should not blunt the Essay’s normative or practical impact. Libertarianism, despite being the third normative tax-law tradition of significance, is by far the minority position within the economics,58 political-philosophy,59 and legal academic fields60—not to mention the U.S. population.61 The Essay’s case for gig-worker UI thus retains its significance for scholarship and policy making.

A. The Liberal Egalitarian Lens

Liberal egalitarians, similar to those in the libertarian tradition, hold that each individual has equal moral worth and therefore deserves autonomy.62 But liberal egalitarianism splits from libertarianism in that, in order to respect each person’s equal moral worth, it strives to account for differences in individuals’ background distributions.63 As a result, its adherents scrutinize social institutions that shape autonomy and, to aid the least advantaged, seek to correct inequalities that those institutions produce.64

UI, in general, almost certainly aids those disadvantaged in society. By definition, those eligible for UI have involuntarily lost their jobs—often through no fault of their own, but through exogenous economic developments.65 For many unemployed individuals, then, background employment and economic structures have brought income deprivation that leads to severe disadvantages. Jobs, after all, are most people’s largest source of income.66 And while theorists disagree on income’s precise role in individual welfare or autonomy,67 it surely matters heavily for each. Without income, most workers would struggle to meet basic physical needs, rendering nonessential consumption or other pursuits inaccessible. Spending time and resources to find new employment can itself be harder without income security.68 Likely because of job-search costs, research suggests that less generous UI benefits push job hunters into positions that satisfy their preferences relatively poorly.69

Unemployment has other, indirect autonomy-reducing effects as well. Studies reveal, for instance, that unemployment can cause psychological damage, impede skill development, and delay professional advancement70—all limiting workers’ future opportunities. Unemployment also undermines the autonomy of workers’ dependents, who rely on workers’ income71 and suffer psychological harms when their family members lose work.72 By distributing resources to the least socially advantaged individuals, then, UI bears out liberal egalitarian goals.

B. The Utilitarian Lens

UI also finds backing under utilitarian principles.73 Utilitarianism, which takes as its goal maximizing total social welfare, has two main values74 that ground its treatment of UI. The first is redistribution, which entails allocating resources to those who derive the highest marginal utility from them.75 This factor weighs in favor of UI, for the reasons discussed above.76 Because many unemployed citizens have little other income, each UI dollar distributed to them would significantly boost utility.77

Utilitarians also value efficiency, which counsels against distorting individuals’ decision-making. Efficiency dictates, for instance, that policy makers should not incentivize people to partake in activities that they otherwise would not.78 Doing so would undervalue or overvalue particular goods, keeping the market from producing each at utility-maximizing levels.79 At first, UI might seem to distort choices by discouraging the unemployed, once jobless, from seeking work.80 Because UI raises the returns to unemployment, some people might come to prefer joblessness over working.

This outcome is theoretically possible, but empirical studies offer strong reason to doubt it. Most research finds that, at least in the short term, UI does not meaningfully discourage job hunting.81 These findings are unsurprising, perhaps, given that UI benefits are temporary and often low,82 and given unemployment’s indirect harms on workers.83 Even following the COVID-19 pandemic84—when Congress temporarily expanded unemployment payments far beyond normal levels85—early research suggests new benefits had only modest labor-supply impacts.86 Moreover, UI arguably increases efficiency for some who desire work but, without UI, would be too income deprived to spend resources finding jobs effectively.87 Both distributional goals and efficiency, therefore, work to justify UI.

C. Application to Gig Workers

Liberal egalitarian and utilitarian justifications for UI each extend logically to gig workers. As for liberal egalitarianism, those aspects of unemployment that restrict standard workers’ autonomy equally affect many participants in the gig economy. As gig jobs have surged, growing numbers use gig work to replace, not supplement, standard full-time work. Government surveys show, for instance, that nearly three-quarters of workers on app- or internet-based gig platforms rely on gigs as their primary income source.88 By another estimate, roughly one-third of gig workers spend over forty hours per week on gig jobs.89

Moreover, a liberal egalitarian would observe, other factors likely make gig workers worse off, all else equal, than standard employees upon entering unemployment. For one, gig jobs are often low skilled.90 This suggests that the workers filling such jobs, on average, have less earning power.91 Gig workers will therefore tend to begin unemployment with less wealth, meaning that income deprivation restricts their autonomy more than it otherwise would. Additionally, most standard employers give employees health insurance, retirement savings, and other benefits.92 Many gig workers do not receive these. Without these benefits, workers are ill-equipped to weather crises, such as job loss, and build savings “buffers.”93 Studies conducted before the rise of the gig economy, in fact, document low-skilled workers’ tendency to enter “alternative” working arrangements with low pay, sparse benefits, and minimal job security—all of which foster economic vulnerability.94

From a utilitarian standpoint, these same dynamics also suggest that jobless gig workers, like other unemployed individuals, possess high marginal utilities. Distributional criteria therefore support UI benefits for gig workers as much as for other individuals. Efficiency criteria support the same conclusion for another reason. Against the backdrop of modern employment law,95 gig-worker UI would arguably correct inefficient distortions. While UI does not—empirically speaking—significantly distort work/leisure choices,96 it may well distort choices between gig jobs and non-gig jobs. Providing UI for just one type of work may incentivize some workers to switch from one to the other.97 Insuring both sectors equally, therefore, is a more efficient outcome under utilitarian principles.98

III. integrating benefits with costs

The foregoing discussion shows that under accepted distributive-justice principles, providing benefits to unemployed gig-workers is justified. But to determine whether such principles support expanding existing UI programs, we must also account for UI’s costs.99

Theoretically, any government revenue stream could cover these payments.100 But in practice, states fund UI programs with specific taxes on employers.101 At least for the foreseeable future, this practice seems likely to persist. The public, after all, tends to compartmentalize policy views into “silos,” meaning that they evaluate the fairness of redistributive policies on a category-by-category basis.102 This tendency suggests that voters will continue supporting UI systems that (nominally) pin the costs on employers—even if firms pass costs on to workers, or if lower taxes elsewhere offset UI-specific levies. It is “only fair,” the policy-silo logic would run, for employers to shoulder the cost. Employers are the ones who fire workers in the first place.

This Essay aims to justify a UI system that is realistically achievable within American political institutions. Accordingly, this Part explores how gig-worker benefits could be funded within the country’s prevailing UI-program framework, where particular UI taxes provide the entire revenue for benefits. This structure places costs upon different actors than those in alternative arrangements where, for example, a top-one-percent wealth tax or general budget revenue funds UI payments. As this Part argues, however, the ideals of distributive justice still support gig-worker benefits fashioned after the existing U.S. system.

A. Funding Gig-Worker UI

Today, all fifty states tax employers to fund UI payments.103 Under UI-tax regimes, employers’ payments equal a fixed portion of their total wage expenses multiplied by their state’s UI-tax rate.104 The taxes are experience rated, meaning that employers whose former employees claim more benefits pay proportionally more.105 States store tax proceeds in federally managed trust funds, from which they disburse benefits.106

This system would be straightforward to administer for certain gig workers—namely, those who have long-term contracting relationships with major contractors, such as Uber drivers. In these cases, the government could directly tax the contracting corporation. That arrangement would function analogously to normal UI taxes and would therefore introduce no new distributive-justice concerns. The government would likely need a different approach, however, for self-employed workers with successive, short-term labor contracts. Levying a special tax on all parties paying these workers would be administratively costly107—and such an approach could unnecessarily stifle taxpayers’ business operations if it required extensive recordkeeping.108

A more realistic option for most states, therefore, is directly taxing this group of workers. With new legislation, most states could implement this tax straightforwardly via existing mechanisms for reporting and collecting statewide income taxes.109 At first glance, this financing method might introduce distributive-justice problems. From a liberal egalitarian standpoint, the direct tax would amount to forcing gig workers to buy UI. Compelled purchases would interfere with individual autonomy and contravene conventional assumptions that each person knows best how to allocate their own resources.110 From a utilitarian standpoint, the direct tax could cause distortions.111 But, as explained below, neither objection carries much force.

B. The Liberal Egalitarian Lens

Taxing gig workers to pay for UI, a liberal egalitarian might object, would restrict autonomous allocation of resources. This observation is correct, but liberal egalitarians might justify such restrictions for at least two reasons. First, as for the choice to buy UI, gig workers have limited autonomy in the first place. Economists widely believe that—owing to information asymmetries and adverse selection into insurance programs—market-based UI schemes are largely infeasible.112 Absent government-provided UI, then, gig workers who do want insurance cannot buy it on the market. Therefore, tax-funded UI would not deprive gig workers of any real choice that they now enjoy. It would replace one forced outcome with another, potentially superior one.

Second, the preference for autonomy rests in part—at least for many liberal egalitarians—on an implicit assumption that individuals are rational.113 But this assumption likely fails for the decision to purchase UI, which involves saving for (seemingly) remote or unlikely risks. Individuals often behave irrationally in such contexts. Empirical evidence indicates, for instance, that enrolling people automatically in retirement plans increases retirement savings.114 Assuming that manual enrollment is not overly burdensome, and that retirement savings improve lifetime welfare,115 such findings suggest that people do not rationally manage future financial risks.116 This tendency would be even more pronounced with saving for unemployment, an event that, unlike retirement, does not happen for all workers.117

It is also worth noting that taxing gig workers would treat them, in economic terms, roughly the same as normal employees. Ample empirical evidence shows that employers transfer UI-tax incidence118 to employees by cutting wages, particularly in the long run.119 Standard UI thus restricts autonomy as much as a direct tax. Liberal egalitarians who oppose gig-worker UI on direct-taxation grounds, then, likely must also accept the less-intuitive stance that standard UI benefits are normatively wrong.

C. The Utilitarian Lens

Utilitarians might object to direct UI taxation on choice-distortion grounds. Two potential distortions are relevant. Taxing gig workers might distort choices, first, between gig and non-gig jobs.120 However, given that employers pass on UI-tax incidence, as discussed above,121 a gig-worker tax would equalize gig jobs’ treatment with other occupations (where workers pay UI taxes in practice). Just as with extending UI benefits to gig workers,122 forcing gig workers to pay for UI would thus improve efficiency versus today’s baseline. An overall gig-worker UI scheme would yield no distortion for any worker’s job choice—regardless of whether that worker prefers job-insurance protections or higher wages.123

Second, utilitarians might fear intertemporal distortion. Forcing gig workers to pay taxes, in other words, might make them consume less today than they otherwise would (and, presumably, more during periods of unemployment). This distortion is likely inevitable, but it is also easy to justify. As with redistribution from rich to poor, such intertemporal substitution distributes money over the course of one’s life from high-income periods to low-income periods, when marginal utility is higher. Any distortion of this kind, therefore, scores highly by utilitarianism’s distributional criteria.

IV. translating theory to practice

As discussed in the Introduction, this Essay’s central contribution is normative. It shows that principles of justice widely accepted in tax-law scholarship compel gig-worker UI. Articulating the normative case for such a policy—separately from prescribing exactly how to implement it—makes two significant contributions. First, and most fundamentally, it is only through normative inquiry that policy makers, advocates, or anyone else can determine their values and, consequently, which kinds of policies they support. For this reason, as one scholar writes, “[p]hilosophy-free tax theory or practice does not exist.”124 Indeed, the interdisciplinary nature of tax theory, which is “shape[d] . . . at the most basic level” by “moral ideas about justice and human welfare,” is perhaps one reason our real-world tax system is so notoriously complex.125

Second, clearly articulated normative reasoning has direct, measurable effects on policy. Policy makers widely appeal, in explicit terms, to normative theories of justice and their constitutive principles. Utilitarian values dominate contemporary tax-design debates126 and underpin the cost-benefit analyses to which federal agencies subject new regulation.127 Economic efficiency, some joke, is fairly characterized as the “Brookings Religion”128—a nod to the Washington, D.C. policy think tank. The influence of liberal egalitarian ideas is perhaps less keenly felt than utilitarianism, yet these traditions, too, hold noticeable sway over law, lawmaking, and regulation.129

When converting normative theory to practice, lawmakers and regulators will inevitably be left to fill in the details. Although that task exceeds this Essay’s scope, this Part briefly sketches some implementational hurdles policy makers will encounter. In doing so, it aims both to guide future research and to furnish further support for the proposition that gig-worker UI is, in fact, workable.

Crafting a gig-worker UI system requires, first, legal standards for when gig workers are unemployed. Sometimes, this analysis will be straightforward. Many contracting companies already report contractor payments to the Internal Revenue Service;130 agencies can leverage this data and require more frequent reporting. But other cases are more complicated. Would Uber drivers be “unemployed,” for instance, when recessions (or events like pandemics131) allow them to find only a handful of working hours per week? How can lawmakers verify self-employed-worker unemployment in a manner that prevents fraud?

Answering these questions is important for minimizing policy distortions and helping benefits reach recipients proportionally to their need—that is, their low marginal utilities or unearned disadvantages. Agencies, therefore, should develop overlapping sets of legal tests and checks, applicable to different workers. Some self-employed workers, for example, may be asked to document particular year-over-year revenue declines. This requirement would mirror the Small Business Administration’s criteria for Paycheck Protection Program loan eligibility.132 For other groups, quarterly tax filings may suffice—or even deposit statements from bank accounts preregistered with state agencies. States have already accepted this kind of documentation to verify PUA eligibility.133 The precise qualifying criteria, moreover, might be written to vary automatically with economic conditions.134 A stronger “presumption” of unemployment during recessions, for instance, could advance justice by boosting total aid as market-provided job opportunities diminish.

States will have to adjust other UI standards for the gig economy, too. Another important question is how much support jobless gig workers will receive. Current UI systems pay recipients a portion of their former salaries, but gig jobs’ earnings are inconsistent and volatile.135 States will likely need to select discrete time periods, which may vary by industry, over which to calculate “average earnings.” Drawing these kinds of distinctions, again, is a pragmatic way to satisfy the respective proportionality demands of theories of justice.

Further, states will need standards for part-time gig-worker eligibility. Today, UI excludes many part-time non-gig workers because they are not “able and available” for work under state law.136 Short of expanding general part-time benefits, states must develop new formulae for defining “part-time” thresholds for gig workers, which again may vary by industry. Along these lines, states must also determine how to treat those using gig work to supplement traditional employment (on a part-time or full-time basis). Hours worked in gig and non-gig jobs would, ideally, factor equally into regulators’ judgments about individuals’ UI eligibility—as theories of justice would grant that distinction no normative significance. This policy objective may require benefits recipients to make additional disclosures.

Ultimately, the questions above introduce “edge cases” regarding which workers should receive gig-economy UI benefits, and to what extent. Lawmakers and regulators, in all likelihood, can find reasonable ways to draw these lines equitably. But importantly, none of these questions should be sticking points that halt the broader adoption of gig-economy UI. Public concerns about fraud or abuse in the edge cases, in other words, might well preclude extending UI to as wide a group as justice requires (particularly if gig employers continue their prolific political spending137). Regardless, policy makers should not hesitate to grant it to the millions of gig workers whose circumstances are clearer cut. In the short term, they can design benefits underinclusively, if need be, while developing protocols that mitigate public concerns.

As a final issue, gig-worker UI will put significant strain on states’ unemployment-claims processing systems. Many states’ UI infrastructure is outdated, which caused difficulties when Congress expanded unemployment benefits during the COVID-19 pandemic.138 Conceptually, building out claims-processing capacity should not be difficult, but it may require politicians to support upfront public expenditures. In any event, given administrative obstacles’ tendency to inflict the greatest harm on the least advantaged,139 overcoming this hurdle matters for distributive goals.


Gig-worker UI benefits, structured similarly to existing UI programs, can be easily justified under accepted theories of justice—accounting for both the program’s costs and benefits. Justice therefore compels states to extend UI to the gig economy. Of course, implementing any such program, as discussed above, would introduce additional challenges, some of which implicate distributive justice. For instance, states must design unemployment-reporting protocols, boost their own institutional capacities, and develop fair standards for adapting part-time-worker eligibility rules to the gig economy. But surmounting these hurdles serves worthy distributive-justice goals and aligns with the philosophical principles already underpinning our tax system.

Many thanks go to Anne Alstott and Zachary Liscow, for feedback at various stages of drafting, and to the editors of the Yale Law Journal, for exceptionally thoughtful and diligent suggestions.