I. Introduction

Professor Bagenstos presents us with a puzzle: Given that public benefits programs tend to be extremely sticky, why were COVID-era benefits so slippery? Why, indeed, does the retrenchment of COVID benefits seem to have given rise to a broader retrenchment of entitlement programs, like Medicaid and the Supplemental Nutrition Assistance Program (SNAP)?

I have a hypothesis—a theoretical answer to this puzzle—that is quite different from the ones Professor Bagenstos offers in his article. My hypothesis is grounded in John Stuart Mill and Harriet Taylor Mill’s so-called “harm principle,”[1] which is (as I have argued at length elsewhere) better understood as an “uncertainty principle.”[2]

In my brief remarks here, I will tell you what the Mills’ uncertainty principle is and how it functions, and then I will use the uncertainty principle to explain the slipperiness of COVID benefits. Specifically, I will highlight some important differences between 2020 on the one hand and 1929 and 1963 on the other: differences that can help explain why New Deal and Great Society benefit programs were so much stickier than their COVID-era counterparts.

II. The Uncertainty Principle

In the introduction to On Liberty, John Stuart Mill and Harriet Taylor Mill announced a “very simple” thesis that the only legitimate purpose of social coercion is to prevent “harm to others.”[3] For nearly a century and a half, scholars have taken the Mills at their word, attributing to their treatise a “harm principle” that permits social intervention only to prevent interpersonal injuries. On this understanding, the primary purpose of the Mills’ principle was to prohibit paternalistic regulation: regulation intended to prevent only self-harms.[4]

As the Mills elaborated their principle throughout the body of their treatise, however, “harm to others” became, for the Mills, a mere shorthand for “harms that are observable from the outside and are thus reasonably certain to cause net welfare losses in the absence of intervention.”[5] That is, the distinction that mattered most to the Mills was not a distinction between “harm to others” and “harm to self”; it was a distinction between observable and unobservable net harm.

This interpretation of “harm” in the Mills’ treatise is easiest to see in the specific examples they offered of regulable and nonregulable conduct.

First, several of the Mills’ examples of regulable conduct, such as voluntary slavery contracts,[6] involve only self-harm, but those examples all involve harms that are obvious and observable from the outside. Everyone knows that slavery is harmful to the enslaved person, and when an enslaver whips an enslaved person for disobedience, everyone can observe the resulting injuries.[7]

On the other side, most of the Mills’ examples of nonregulable conduct, such as dancing among Puritans,[8] involve interpersonal harms, as the Mills explicitly recognized.[9] The heathens’ dancing scares the Puritans. Those harms, however, are not observable using ordinary sense perception; they are entirely internal harms, existing only in the minds of the harmed individuals. The Puritans’ fear that God will punish them for permitting others to dance is a harm that no one outside of the fearful Puritan himself can observe.[10] Unlike the injuries that arise from whipping, chronic fear is imperceptible from without.[11]

Throughout the Mills’ treatise, then, the factor that consistently justified regulation was certainty (not “harm to others”), and the factor that consistently demanded liberty was uncertainty (not “self-harm”).

Indeed, the Mills were explicit that their ideal principle was not about harm at all. Instead, their ideal principle centered explicitly on uncertainty. As the Mills put it: “It is easy for any one to imagine an ideal public which leaves the freedom and choice of individuals in all uncertain matters undisturbed, and only requires them to abstain from modes of conduct which universal experience has condemned.”[12]

The only reason the Mills articulated their principle in terms of harm rather than certainty, they said, was that humans tend to “assume that their certainty is the same thing as absolute certainty.”[13] The Mills worried that a rule forbidding regulation under conditions of uncertainty would provide no meaningful constraint on regulation—and no meaningful scope for individual liberty—because human regulators rarely recognize their own uncertainties.[14]

These days, however, the opposite might be true. Modern scientists—who became an important influence in politics with the rise of technocratic governance during the New Deal Era[15]—sometimes recognize too much uncertainty.

During the Modern Era, in which the Mills were writing, science was still deductive, as it had been since the time of Aristotle.[16] Before the positivist turn of the nineteenth century, scientists asserted first principles dogmatically, and they derived laws of nature mechanically and syllogistically from those principles. Early modern science thus denied any role for uncertainty—which is why Galileo was imprisoned for heresy. No one was willing to admit that the theistic “first principle” of geocentrism might be wrong.[17]

As the Victorian Era faded into the early twentieth century, science took a hard positivist turn,[18] becoming increasingly observational, inductive, and quantitative.[19] In the wake of that turn, uncertainty began to predominate. The introduction of Knightian uncertainty in 1921[20] and Heisenberg’s uncertainty principle in 1927[21] transformed social and physical sciences from practices of logical determination into practices of probabilistic estimation. Today, no scientific model can claim coherence without acknowledging an error term: a variable that captures the residual uncertainty inherent in all human observations.

This change in the nature and pervasiveness of uncertainty, relative to the Mills’ Victorian Era, might undermine the usefulness of an “uncertainty principle” in the opposite direction from the one the Mills feared. An uncertainty principle cannot usefully guide human governance under either the Victorian or the modern notion of uncertainty. If uncertainty is either everywhere or nowhere, then an “uncertainty principle” will not provide a useful limit on regulation. A Victorian uncertainty principle would lead to tyranny—because Victorians are never uncertain—but a modern uncertainty principle, if taken to encompass all kinds of uncertainty, would lead to anarchy—because modern scientists are always at least a little bit uncertain.

The Mills, however, had a more specific understanding of uncertainty than either the Victorians’ arrogance or the moderns’ humility. And, once decoded, the Mills’ notion of “uncertainty” does a remarkably good job, I think, of describing governmental behavior in a liberal democracy, not only with respect to individual liberty but also more generally in the regulation of human life.

To the Mills, “certainty” was about observability, and “observability” was about either sensory perception or widespread consensus.

When the core costs and benefits of a regulatory question are observable—either to our physical senses or by way of near-unanimous subjective agreement—then society is competent to intervene to advantage. Of course, regulators can never obtain perfect information, even when regulatory variables are visible and quantifiable; humans are not omniscient. But when costs and benefits are observable, regulators can attain “the amount of certainty attainable by a fallible being.”[22] As the Mills put it, “There is no such thing as absolute certainty, but there is assurance sufficient for the purposes of human life.”[23]

When, however, the core costs and benefits of a regulatory project are entirely unobservable—because they are wholly internal and/or subject to deep dissensus—then society should abstain from intervening. In such cases, intervention will be guaranteed to generate costs—because regulation always incurs, at minimum, an opportunity cost—but there will be no way to determine whether those costs have generated any benefits at all, much less enough benefits to outweigh the inevitable costs of intervening.[24]

The easiest way to illustrate this more-limited meaning of “uncertainty” is through two concrete examples.

First, regulatory regimes that center primarily on monetary transactions provide an easy example of Millian certainty.[25] Money is a quantified medium that facilitates observation of individuals’ expected costs and benefits.[26] When trade is voluntary and informed, my willingness to pay $5 for a widget is a strong indication that the widget will, in my estimation, provide me at least $5-worth of benefit. Likewise, the seller’s willingness to accept $5 is a strong indication that she will, in her estimation, suffer less than $5-worth of harm from giving it up. Furthermore, if an observer knows that the widget-manufacturer spent more than $5 to produce the widget, the observer can conclude that the manufacturer sold the widget at a loss (incurring a real harm), and if other widgets of comparable quality regularly sell for $20, the observer can presume that I obtained a larger-than-usual consumer surplus from the acquisition. That is, the observer can presume that the benefits I gained from the widget far outstripped the break-even point of the $5 I spent.

Both the monetization of a single exchange and the presence of a broader monetized market thus facilitate observation of a social interaction’s net effects for each of the individuals involved. Prices do not provide perfect or complete information—and sometimes money does not capture any of what regulators care about, such as when someone pays for sex[27]—but prices illuminate at least a portion of both parties’ costs and benefits.

The second example that helps to illustrate the Mills’ principle—the clearest example of a regulatory question suffused with Millian uncertainty—is religion.[28] Sensory perception cannot determine the objectively correct theory of divinity. We cannot meet any God or gods that might exist, except perhaps after death, and we cannot interview dead people to learn about their experiences.[29] Religion is a matter of faith, not observation. As a result, a regulator could not coherently claim to have made the world objectively better off by increasing the percentage of, say, Hindus in the world.

Furthermore, shifting the question from objective reality to subjective preference does not help in the context of religion. There is deep and widespread dissensus on the question of divine meaning. Importantly, this dissensus centers entirely on internal feelings and beliefs that cannot be observed from the outside. A prohibition on practicing Christianity would cause emotional pains for Christians, but those pains would be neither perceptible nor measurable. As a result, the pains would be impossible to weigh against the (equally unobservable and nonquantifiable) satisfaction that a regulatory majority might experience from prohibiting the religion. That is, a Christian’s pain at being forced to practice Islam probably outweighs a Muslim’s pleasure at seeing a Christian pray to Allah, but the actual cost-benefit calculation is impossible to perform, even in the rough “attainable certainty” sense, given that both the pleasure and the pain are unobservable from the outside.[30]

Finally, even if we knew it to be true that everyone would be objectively better off if, say, they accepted Jesus as their Lord and savior, the effects of regulation would be impossible to assess. Whether an erstwhile Buddhist has truly, in her heart-of-hearts, accepted Jesus is a question that, in the absence of mind-reading machines, cannot be answered by anyone outside of the regulated individual. (You can learn to pronounce “Shibboleth” without true faith![31]) Even assuming, then, that maximizing the presence of Christians was the objectively best result, regulatory intervention to accomplish that goal would be unable to claim demonstrable benefits that could outweigh its costs.

When regulatory costs, benefits, and effects are all uncertain in this sense, regulation will, by necessity, be wholly arbitrary. There will be no way to estimate the optimal regulatory approach ahead of time nor will there be any way to assess regulatory successes or failures after regulations go into effect. Governmental intervention, under such conditions, can never be anything but a set of “arbitrary impositions and purposeless restraints.”[32]

With this idea of certainty and uncertainty in mind, one general rule of regulability emerges, which will remain important as my remarks continue. Under the Mills’ principle, interpersonal welfare transfers—especially when they are unidirectional—are easier to regulate than are intrapersonal welfare tradeoffs.

If I punch you in the face to steal your money, I have inflicted physical and financial harm on you without giving you anything in return. I have transferred welfare entirely from you to me, without any offsetting transfer from me to you. The welfare effects are both interpersonal and unidirectional. Under such circumstances, regulators can observe the harm I have inflicted on you, and they can confidently conclude that you would never have chosen to endure that harm if you had had the power to avoid it. Not only the effects but also the net effects, for each of the individuals involved, are easy to observe. My net effect is wholly positive; yours is wholly negative.

If, however, you and I agree to compete in a boxing match, regulators will have no way of knowing whether the net effects of the punches are positive or negative for the individuals involved. Regulators will be able to see the physical harms each of us endures—the blackened eyes and bloodied noses—but they will not be able to determine whether the thrill of the contest and the expectation of victory are sufficient, for either of us, to outweigh those harms. Each of us will have made an intrapersonal welfare tradeoff—the pain against the thrill—that regulators will be unable to observe. Intervention, in such circumstances, will not be able to claim cost-benefit legitimacy. Saving us from the pain of black eyes might make us worse off, on net, by depriving us of the thrill of the contest.

In general, then, intrapersonal welfare tradeoffs involve more Millian uncertainty than do interpersonal welfare transfers.[33] To justify regulation of intrapersonal tradeoffs, a Millian regulator would need to satisfy the high bar of showing that “universal experience has condemned” the individual’s choice.[34]

III. Economic Insecurity

Social and economic guarantees[35]—which are the subjects of most entitlement programs and thus the primary subjects of Bagenstos’s puzzle—present an intermediate case under the uncertainty principle. When the government addresses economic insecurity, its ends are clear and agreed-upon, providing some dimensions of certainty. But there is widespread dissensus as to means.

I will focus here only on those entitlements that Bagenstos mentioned in his article: income assistance, unemployment insurance, health insurance, and (although Bagenstos mentions it only tangentially) food.[36]

To start, there is no doubt that everyone is better off when financially secure. An economically insecure life is one that “universal experience has condemned.”[37] No one believes—or even argues—that life is better for those who cannot afford to pay for the food, clothes, or healthcare they need. And no one believes or argues that any human being can survive without food, shelter, or healthcare.

Furthermore, the harms of starvation, exposure, and illness are easy to observe from the outside, using ordinary sense perception. I can see the difference between a starving body and a well-fed one; I can see sunburns, shivers, and frostbites that result from exposure; I can hear people’s coughs, sneezes, and sniffles; and I can use a thermometer to measure people’s temperatures. I can tell the difference between a person who is alive and a person who is dead. All the costs of economic insecurity—all the costs associated with lack of food, shelter, and healthcare—are easy to observe from the outside. In one sense, then, there is no Millian uncertainty involved in the question of economic insecurity. Superficially, it seems that society is competent to regulate economic welfare, and public benefits programs should gain widespread support.

There is, however, deep disagreement as to what causes economic insecurity. That disagreement, in turn, translates into a widespread dissensus as to the best regulatory means of addressing the problem: a dissensus that importantly tracks the logic of the uncertainty principle. Specifically, there is a long-running debate among political and economic theorists as to whether insecurity arises primarily from interpersonal transfers or from intrapersonal tradeoffs.[38]

Bagenstos gives you some insight into the “transfers” side of that debate, treating economic insecurity as a symptom of capitalism.[39] On the theory that Bagenstos adopts, capitalists take from their workers in a unidirectional transfer of welfare. Owners and employers demand that their employees labor for the profit of the employers, not for the well-being of the laborers, giving rise to an interpersonal transfer that “society is competent to restrain.”[40]

This theory builds on a deeper jurisprudential argument that economic insecurity—and the need, among employees, to labor for wages—arise from the government’s protection of property rights.[41] If the government didn’t protect privately-owned homes against trespass, then unhoused people could sleep in your bedroom instead of dying of exposure. If the government didn’t protect grocery stores’ property interests in their stock, then food-insecure people could waltz into the nearest Whole Foods to take what they needed. And if the government didn’t protect capitalists’ ownership of their real and monetary properties, then unemployed people could simply take the capitalists’ money to support themselves rather than being “forced,” by property rights, to work for a living. The harms of economic insecurity, on this view, are side effects of other regulatory interventions that explicitly transfer welfare (in the form of “ownership”) from one person to another and that condition such transfers on observable monetary payments[42] that not everyone can afford.[43]

On these property-based theories, economic insecurity is obviously a fit subject for regulatory correction under the uncertainty principle. The causes of such insecurity are interpersonal transfers that center primarily on money.

On the competing theory, however, individuals become economically insecure through a series of their own decisions.[44] A cascade of intrapersonal welfare tradeoffs—a decision to quit a stable job to start one’s own business, a decision to discontinue one’s mental health medications, or a decision to start using heroin, for example—leads to foreseeable insecurity that the individual effectively chose for herself. On this theory, the cost-benefit tradeoff is impossible for regulators to assess because the benefits are internal. I cannot observe the pleasures that other people derive from heroin, so I cannot definitively conclude that life is objectively better when lived sober and housed rather than high and unhoused. Perhaps some people derive so much pleasure from heroin that the accompanying costs of homelessness are, for them, worth bearing.[45] (I doubt it! But I cannot use my sense perception to prove that my doubts are well-founded.)

On this theory, economic insecurity is not a fit subject for regulatory intervention because the net effects of the intrapersonal tradeoffs are impossible to observe. A government program that “helped” people find houses and jobs might actually make them worse off.

Consistent with the uncertainty principle, those who believe that economic insecurity arises from interpersonal transfers tend to favor public benefits programs while those who believe that economic insecurity arises from intrapersonal tradeoffs tend to oppose such programs. These competing theories thus give rise to widespread dissensus surrounding the appropriate means of addressing economic insecurity: whether regulation is a proper tool to address the problem or whether voluntary private charity is the better approach.

All of that said, this debate has been raging since the dawn of time—or at least since the dawn of democracy. The dissensus surrounding appropriate means of addressing economic insecurity thus cannot explain the difference between 2020 on the one hand and 1929 and 1963 on the other. Indeed, debates between laissez-faire and socialism were equally as widespread in 1928 and 1962 as they were in 2019.

IV. Exogenous Shocks

According to Professor Bagenstos, the Biden Administration believed that the “big, exogenous shock[]”[46] of the COVID pandemic would create an opportunity to enact long-lasting expansions of public benefits. Indeed, Bagenstos tells us, exogenous shocks often give rise to “critical junctures” in politics, opening “opportunities for major institutional reforms.”[47]

As a descriptive matter, it is hard to argue with Bagenstos’s understanding here, both because he was working in the Biden Administration at the relevant time[48] (and would therefore know better than I!) and because exogenous shocks did, indeed, precede every moment of major institutional reform in American history. The Great Depression preceded the New Deal; the Kennedy assassination preceded the Great Society; and COVID preceded the current phase of retrenchment.[49]

As a theoretical matter, however, it is not clear why exogenous shocks should open opportunities for long-lasting reforms. Institutional reform might be necessary to address the shock itself, but it is not at all obvious that those reforms should outlast the immediate effects of the shocks. Indeed, when constitutional doctrines flex to accommodate emergencies, everyone expects them to return to normal after the emergency ends.[50] The idea that shocks should create opportunities for major institutional reforms, then, is a theory that begs to be cabined in some way.

Applying the uncertainty principle to the concept of exogenous shocks sheds useful theoretical light on (a) why exogenous shocks generally create space for regulatory interventions that were previously impossible and (b) why the COVID shock, unlike the Great Depression and the Kennedy assassination, has caused retrenchment rather than persistence of entitlement programs.

A. Shocks and Certainty

Under the uncertainty principle, the general relevance of exogenous shocks is that they give rise to observable causes or effects of regulatory costs and benefits that are, during ordinary times, hard to observe. Shocks shine light on otherwise unobservable regulatory variables.

1. Unemployment and Depression

Consider unemployment first. During ordinary times, it is impossible to determine why any given person is unemployed. As noted above, some people believe that unemployment is a systemic problem while others believe that it is an individual choice.[51] When looking at any single unemployed person, however, it is impossible to differentiate between the two theories. A libertarian can always trace an unemployed person’s situation to choices she made, and a socialist can always identify systemic factors that limited and shaped those choices as well as ongoing systemic barriers to her reemployment. Determining whether the systemic or individual variables predominate in any one person’s case is almost always impossible.

During a shock, however, systemic variables obviously predominate for most people. During the Great Depression, the global collapse of capital markets caused unemployment rates to skyrocket.[52] The idea that any single person, at that time, was choosing to remain unemployed would be so improbable as to be laughable (even though there were, at the time, undoubtedly some people who were unemployed by choice). Likewise, during COVID, mandatory quarantines forced many businesses to shutter, and widespread fear of engaging in public life caused a sustained economic downturn that was beyond any individual’s control.[53] As in the Great Depression, many of the people who were unemployed during COVID were obviously not “personally responsible” for their economic insecurity.

2. Healthcare and Pandemic

The story for healthcare is slightly more complicated but has similar features. During ordinary times, the cost of uninsurance is hard to measure. The core problem is that both the ontology and the etiology of symptoms are often impossible to observe.[54] Most viral infections, for example, do not have reliable diagnostic tests, and the clinical presentation of respiratory infection, especially for someone who does not have a fever, is similar to the clinical presentation of anxiety (fatigue and shortness of breath, for example).[55] Pain—which is the dominant symptom of many medical conditions—is impossible to observe from without. I cannot literally “feel your pain.” As a result, doctors frequently have a hard time differentiating among illness, hypochondria, and malingering.[56] Whether an individual really requires medical care is sometimes much less observable than one might hope.

During the COVID pandemic, however, the probability that a person with respiratory symptoms was carrying a virus, rather than merely malingering or experiencing anxiety, significantly increased.[57] Furthermore, the social cost of uninsurance became so widespread that the cost became obvious even to a skeptic. At the height of the pandemic, few Americans would have argued that their uninsured and COVID-infected roommate should be turned away from the hospital; they would have feared infection and death more than they feared increased taxation.

3. Assassination and Government

The Kennedy assassination did not have similarly obvious impacts on individual Americans’ lives, but it did shine a harsh light on the unrest that had been stirring in the decade leading up to that November day.

There were two major movements ongoing at the time of JFK’s death: the Second Red Scare and the fight for civil rights.[58] McCarthyism and the House Un-American Activities Committee (HUAC) were hunting communists and suppressing speech throughout the 1950s and 1960s.[59] While McCarthy had been discredited almost a decade before Kennedy’s death,[60] HUAC’s work continued until the committee was repurposed in 1969,[61] and the ripple effects of HUAC’s “blacklist” continued to spread even after the committee reconstituted itself.[62] Meanwhile, the Supreme Court ordered an end to racial segregation in 1954,[63] giving rise to immediate and intense backlash throughout the South.[64] Then, in 1955, Emmett Till was lynched, and his murderers were wrongly acquitted.[65] In the months immediately preceding JFK’s assassination, Martin Luther King, Jr., was imprisoned in Birmingham,[66] and Medgar Evers was assassinated.[67]

Throughout this time, northern White anti-communists could have felt relatively removed from the turmoil. Indeed, they could have taken comfort in uncertainty, choosing to believe that the government’s actions in all of the controversial cases might have been justified. A northern “Karen” might have thought that the Hollywood Ten and blacklisted writers really were communists (unlike her); that ideological communists really did represent a threat to Americans’ safety such that the mere communist ideology (unlike her own) deserved to be eradicated; that Emmett Till’s murderers really did deserve to be acquitted (unlike someone who might murder her); that MLK was a dangerous criminal who (unlike her) deserved to be imprisoned; and that Byron De La Beckwith deserved to walk free after shooting Medgar Evers.

When Kennedy was killed, however, Karen could no longer feel safely removed from the dangers of her time. JFK was, after all, a White man from Boston who had expressed no sympathy for Soviet Russia or communism and who had committed no crime. His death, unlike the many tragedies that preceded it, could be neither justified nor dismissed as “other.”

Furthermore, uncertainty surrounding the circumstances of and motive for Kennedy’s assassination, rather than providing comfort and distance, became a source of profound skepticism toward prevailing institutions. Immediately following Kennedy’s assassination, a Gallup poll found that only 29% of Americans believed the government’s assertion that Lee Harvey Oswald had acted alone.[68]

All told, then, Kennedy’s assassination shook Americans’ faith in the existing system.[69] Systemic harms that, before 1963, northern Americans could have dismissed as plausibly cost benefit justified became much more obviously unjustified.

Under the uncertainty principle, all of these shifts in observability and belief created new justifications for regulatory interventions, transforming previously uncertain and thus nonregulable questions into reasonably certain and thus regulable ones.

When everyone can see that unemployment is a systemic rather than an individual problem, government should intervene to fix it. When everyone can see that uninsurance imposes social costs that are worth mitigating, government should expand insurance coverage to encompass as many people as possible.

And, in general, when an exogenous shock shines light on the failures of existing institutions, government is justified in reforming or replacing those institutions.

B. Entrenchment?

The core of Bagenstos’s puzzle, however, is not why reforms happen during shocks, but rather why they last after the shocks have ended. Entitlement programs that were enacted during a shock often become entrenched, persisting far beyond the life of the shock itself. In the midst of the Great Depression, everyone recognized the need for Social Security, but the Social Security program has long outlived its period of obvious and observable necessity.[70]

The traditional reason given for entrenchment—and the one that Bagenstos offers—is that entitlement programs’ beneficiaries become a large constituency in favor of the programs’ preservation.[71] People become attached to, or even dependent upon, their government handouts, and they refuse to vote for anyone who suggests taking them away. The widespread, supermajoritarian preference for maintaining entitlements becomes an insurmountable obstacle to their repeal.

This theory is quite plausible, and it provides a compelling description of American experiences, especially during and after the New Deal. But, as Bagenstos notes, the COVID experience does not fit the standard story.

Now, I think there is an explanation for COVID’s divergence that is even simpler than the ones Bagenstos and I are offering. The COVID shock was extraordinarily short-lived, relative to the Great Depression and the Civil Rights Movement. Thanks to the “[w]arp [s]peed” at which scientists developed a working COVID vaccine,[72] many people returned to normalcy less than a year after the pandemic’s outbreak.[73] As a result, voters had far less time to grow attached to or dependent upon their government handouts during the pandemic than they did during the Great Depression.

That said, however, the concept of regulatory uncertainty and observability sheds an interesting alternative light on the COVID experience: a light that is worth considering in addition to the other explanations on hand.

1. The Role of Law

The Great Depression and the Kennedy assassination revealed systemic failures of American law. The Great Depression revealed weaknesses in the laissez-faire approach; the Red Scare revealed the impossibility of regulating people’s beliefs, showing that any attempt to prohibit feelings leads inevitably to arbitrary tyranny; the Civil Rights Movement shed light on the sheer injustice of legal segregation; and the murders of Till and Evers demonstrated the legal system’s inability to curb the brutality of individual segregationists.[74]

The harms that became observable in the harsh light of the 1930s and 1960s were harms that arose from bad laws and that could be corrected by good ones.[75]

The harms that became observable during the COVID pandemic, by contrast, did not arise from bad laws. They arose from a virus. The coronavirus organism—unlike the “viruses” of laissez-faire and segregation—could not be legislated out of existence. While legislation could mitigate some of the pandemic’s downstream harms by providing financial assistance for unemployment and accelerating vaccine development,[76] for example, no law could eliminate the pandemic’s etiology. As a result, the pandemic did not, in and of itself, undermine public faith in preexisting institutions in the same way that the Great Depression and Kennedy assassination did. At the end of the pandemic, there could be no fear that, by repealing COVID entitlements, the government might cause the virus to return.[77]

Indeed, the COVID-era government was attacking a foe that was largely beyond its control. Pandemics are complex systems.[78] Their outcomes are observable, but those outcomes are impossible to trace to discrete antecedent causes.[79] When a government tries to alter the trajectory of a complex system like a pandemic, regulatory design is extraordinarily difficult.[80] As a result, some of the government’s interventions are bound to cause more harm than good, and all of the government’s interventions will have a hard time demonstrating that they were net beneficial.

Consider a few concrete examples from COVID. Supermajorities of Americans throughout the country favored reopening public schools toward the end of the pandemic,[81] but many northern schools nevertheless remained closed.[82] In hindsight, it appears that school closures did mitigate the virus’s spread, but they also resulted in easy-to-measure learning losses and mental health harms.[83] The mitigation of the virus’s spread is an unobservable benefit; there is no way to measure avoided harms: harms that never materialized. But the learning losses and anxiety are easy to see. The cost-benefit tradeoff, therefore, was presumably negative from the perspective of the uncertainty principle. School closures traded concrete harms for unobservable benefits.

Similarly, scientists are confident that cloth masks are less effective at slowing COVID transmission than are their N95 counterparts,[84] especially given that cloth masks are often worn incorrectly. COVID-era mask mandates, however, did not require people to wear cloth masks correctly or to wear N95 masks instead. As a result, whether mask mandates accomplished anything at all, as used in the real world, is impossible to assess. The benefits of cloth masks were established exclusively in lab settings; once the masks entered the complex system of human use, their benefits were impossible to measure. Mask mandates thus angered many people, giving rise to concrete harms, without providing demonstrable benefits.[85]

For similar reasons, the government’s messaging during a pandemic is unlikely to inspire public confidence. When tackling a complex system, discrete causes are impossible to observe, and trajectories are difficult to predict.[86] A government’s best efforts at providing information and guidance will inevitably go astray, at least some of the time. For example, the debate between the “lab leak” theory and the zoonotic transmission theory will never be resolved with perfect clarity, but the debate over the competing theories further undermined public trust in government throughout and beyond the pandemic.[87]

Notably, all of these factors are relevant under the uncertainty principle, and all of them, according to Millian logic, could be explanatory variables in the retrenchment of COVID benefits. The mix of observability and uncertainty that inheres in complex systems makes legitimate intervention both necessary and impossible. Intervention is necessary because harms are real and observable, but the government will be wholly unable to calibrate its interventions precisely to the underlying causes of those harms. Even regulatory effects will be hard to observe. Some regulation will therefore appear arbitrary even when the government is doing the best technocratic job possible under the circumstances. These problems—the unavoidable presence of uncertainty and arbitrariness—will make intervention less legitimate under the uncertainty principle, eroding public support for regulation generally and making the idea of retrenchment more politically palatable than it otherwise would be.

From this perspective, it may not be coincidental that COVID preceded a phase of retrenchment and the Spanish Flu preceded the Roaring Twenties: the apex of laissez-faire.[88] Because there is a tremendous amount of uncertainty involved in global pandemics, such pandemics inevitably make governments appear feeble, error-prone, and illegitimate. Under the uncertainty principle, pandemics inevitably undermine government’s legitimacy.

2. Policy Effects

The final point I want to make is that the observable effects of COVID monetary policies also undermined the entrenchment of those policies. Remember that purely monetary policy is far easier to observe than other kinds of interventions, so monetary interventions must produce concrete benefits quickly if they want to survive. COVID interventions had some easy-to-observe negative side effects, which emerged immediately after their enactment.

First, COVID stimulus checks injected money into the demand side of the economy without first rebuilding the supply side, contributing measurably to COVID-era inflation.[89] The COVID experiment with universal basic income, then—although confounded by lingering effects of quarantines and pandemic fear—correlated strongly with immediate, observable effects that argued against the program’s entrenchment.

Second, COVID income assistance and unemployment programs exacerbated rather than clarified the dissensus between libertarians and socialists in their competing theories of unemployment. With the “Great Resignation,” many workers revealed that they would, under the right conditions, choose not to work. Indeed, they showed that they would quit even if, by doing so, they would forgo significant financial benefits, like health insurance, that the government was not providing for them.[90] Although many of the quitters’ choices were undoubtedly influenced by systemic factors like wage stagnation, the decision to become unemployed was, nevertheless, for many individuals, a real choice. That is, the Great Resignation does not describe systemic unemployment; it describes individual decisions to become unemployed.[91] Of course, the COVID pandemic did involve some systemic unemployment, too, as noted above. Many people lost their jobs due to the interpersonal transfers of quarantines.[92] But the pandemic also involved widespread and observable individual unemployment, resulting from the intrapersonal tradeoff of choosing not to work. Under the uncertainty principle, the Great Resignation undermined the Millian legitimacy of COVID entitlement programs—both because intrapersonal tradeoffs are impossible to regulate technocratically and because the Great Resignation exacerbated dissensus around the appropriate means of addressing unemployment.

The lesson that Bagenstos takes from the Great Resignation is that it provoked the business lobby to attack COVID entitlement programs,[93] which is undoubtedly true. But those same lobbying groups have also attacked Social Security (with no success so far) and Obamacare (with only mixed success so far).[94] In general, lobbying will not succeed, no matter how well-funded, unless the lobbyists’ underlying claims are politically plausible. If Elon Musk started lobbying for the decriminalization of murder, his efforts would go nowhere despite the vast resources he could devote to the project. The idea of decriminalizing murder is not at all politically plausible. The business lobby’s attacks, then, are not a sufficient explanation for the retrenchment of COVID entitlements. Under the uncertainty principle, however, the lobbyists’ substantive objections to COVID entitlements were, in fact, theoretically plausible.

Under the uncertainty principle, both the exacerbation of uncertainty, in the widespread choice to quit working, and the revelation of new data, in the observable economic effects of universal basic income, supported the decision to repeal COVID-era entitlements.

Importantly, these factors also provide real distinctions between COVID programs on the one hand and New Deal and Great Society programs on the other. Social Security did not cause large swaths of people to opt out of the workforce; it supported people who had aged out of the workforce or who had suffered an injury that removed them from the workforce. Medicare and Medicaid may have, over time, contributed to healthcare inflation, but the effects of those programs are extremely difficult to trace causally. In the meantime, Medicare and Medicaid have also provided concrete and observable benefits to covered individuals and to society at large.[95]

V. Conclusion

When the government enacts policies we don’t like, it is always tempting to attribute the outcome to the crass power plays of our opponents. It is much harder to believe that the outcomes of liberal democratic processes—especially when they appear wrong-headed to us—might be rational under utilitarian principles. As I hope is clear from my wording throughout these remarks, I do not claim to know or to be able to prove that the retrenchment of COVID-era benefits was, in fact, cost-benefit justified due to the elements of uncertainty that pervaded pandemic-era regulation. Perhaps I am simply naïve (an Ivory Tower maven to the death!) to believe that there could be an immanent rationality in modern American politics, especially during the MAGA era. But examining the puzzle of COVID retrenchment through the lens of the Mills’ uncertainty principle at least sheds some interesting light on the question. And it makes me feel better about being on the losing side.

  1. See John Stuart Mill [& Harriet Taylor Mill], On Liberty 13, 52 (Batoche Books 2001) (1859). For the propriety of crediting Harriet Taylor Mill as a coauthor, see John Stuart Mill, Autobiography 251–52 (3d ed. 1874) (“The ‘Liberty’ [i.e. On Liberty] was more directly and literally our joint production than anything else which bears my name, for there was not a sentence of it which was not several times gone through by us together, turned over in many ways, and carefully weeded of any faults, either in thought or expression, that we detected in it.”); Christoph Schmidt-Petri, Michael Schefczyk & Lilly Osburg, Who Authored On Liberty? Stylometric Evidence on Harriet Taylor Mill’s Contribution, 34 Utilitas 120, 135–36 (2022) (using stylometric analyses to demonstrate Harriet Taylor Mill’s distinctive voice in the treatise).

  2. See Abigail R. Moncrieff, Constitutional Technocracy 111–12 (June 9, 2025) (unpublished manuscript) (on file with author) [hereinafter Moncrieff, Constitutional Technocracy]; Abigail R. Moncrieff, The Uncertainty Principle: A Reconceptualization of American Substantive Rights 3, 16 (Jan. 7, 2026) (unpublished manuscript) (on file with author) [hereinafter Moncrieff, The Uncertainty Principle].

  3. See Mill [& Mill], supra note 1, at 13 (“The object of this Essay is to assert one very simple principle, as entitled to govern absolutely the dealings of society with the individual in the way of compulsion and control, whether the means used be physical force in the form of legal penalties, or the moral coercion of public opinion. . . . [T]he only purpose for which power can be rightfully exercised over any member of a civili[z]ed community, against his will, is to prevent harm to others. His own good, either physical or moral, is not a sufficient warrant.”).

  4. For a few examples of how anti-paternalism has become a dominant theme in understanding the harm principle, see generally Piers Norris Turner, “Harm” and Mill’s Harm Principle, 124 Ethics 299 (2014) (proposing that the harm principle is “merely an anitpaternalism principle, concerned with allocating decisional authority between society and the individual on the basis of what sorts of reasons are in play”); Ben Saunders, Reformulating Mill’s Harm Principle, 125 Mind 1005 (2016) (discussing the permissibility of soft paternalism under Mill’s harm principle); Sven Ove Hansson, Mill’s Circle(s) of Liberty, 41 Soc. Theory & Prac. 734 (2015) (discussing how the principles of harm, free trade, and individual liberty interact with each other to form Mill’s account of liberty and how they connect with his anti-paternalistic view). For a more nuanced (and more accurate) discussion of the Mills’ attitude toward paternalism, see generally Raphael Cohen-Almagor, Between Autonomy and State Regulation: J.S. Mill’s Elastic Paternalism, 87 Philosophy 557 (2012) (defining Mill’s view of paternalism as “elastic” due to his favoring of both soft and hard paternalism depending on the situation at hand).

  5. See Moncrieff, The Uncertainty Principle, supra note 2 (manuscript at 16).

  6. . Mill [& Mill], supra note 1, at 94.

  7. See generally Solomon Northrup, Twelve Years a Slave (David Wilson ed., Univ. of N.C. at Chapel Hill Libr. 2011) (1853) (giving a firsthand account of the horrors of slavery in the American South).

  8. Mill [& Mill], supra note 1, at 77–80.

  9. Id. at 80, 83–84 (recognizing Puritans’ fear “that God not only abominates the act of the misbeliever [who works on Sunday or who dances], but will not hold us guiltless if we leave him unmolested”).

  10. See id. at 80, 84.

  11. Acute fear is often perceptible; you can see someone cowering or shaking. But the chronic fears that arise from general policies—such as modern fears that tolerating abortion will lead to eternal damnation—do not give rise to perceptible symptoms. You cannot tell, when you meet a fearful pro-life person, that he is scared of damnation.

  12. . Mill [& Mill], supra note 1, at 78.

  13. Id. at 19 (emphasis added).

  14. Mill [& Mill], supra note 1, at 19.

  15. See Moncrieff, Constitutional Technocracy, supra note 2 (manuscript at 4, 45).

  16. See id. (manuscript at 178–79, 188 n.j); see also Michael A. Williams, Francis Bacon’s Inductive Science and its Victorian Consequences, Victorian Web (Jan. 8, 2018), https://victorianweb.org/science/francisbacon.html [https://perma.cc/GG9P-ZKDQ] (explaining how Bacon’s inductive science influenced late nineteenth century scientists).

  17. See generally Maurice A. Finocchiaro, The Church and Galileo, 94 Cath. Hist. Rev. 260 (2008) (providing an overview of the historical and intellectual relationship between the Catholic Church and Galileo, including the doctrinal and institutional context of the Galileo affair).

  18. See generally Dylan Riley, The Paradox of Positivism, 31 Soc. Sci. Hist. 115 (2007) (describing the development, domination, and ongoing resilience of “methodological positivism,” an approach to science that seeks ontologically real and observable laws that exist independently of the sciences that are trying to describe them).

  19. For a critique of the modern tendency to seek quantification, see generally Sally Engle Merry, The Seductions of Quantification: Measuring Human Rights, Gender Violence, and Sex Trafficking (2016) (arguing that overreliance on quantifiable variables causes social scientists to miss or misunderstand many dimensions of profound social problems, especially including gender violence and sex trafficking).

  20. See generally Frank H. Knight, Risk, Uncertainty and Profit (1921) (articulating a distinction between risk and uncertainty and then arguing that the pervasiveness of uncertainty prevents “perfect competition” from functioning as theorized, leaving economic markets with only “imperfect competition” as their best possible standard).

  21. See generally W. Heisenberg, Über den anschaulichen Inhalt der quanten-theoretischen Kinematik und Mechanik, 43 Zeitschrift für Physik 172 (1927) (Ger.), translated in Nat’l Aeronautics & Space Admin, NASA TM-77379, The Actual Content of Quantum Theoretical Kinematics and Mechanics (1983), https://ntrs.nasa.gov/citations/19840008978 [https://perma.cc/WU5P-V737] (arguing that observation always perturbs the system under observation, making it impossible to measure two conjugate variables (like position and momentum of a quantum particle) with simultaneous precision).

  22. . See Mill [& Mill], supra note 1, at 23, 41.

  23. Id. at 21.

  24. See Moncrieff, Constitutional Technocracy, supra note 2 (manuscript at 139–40).

  25. Indeed, the Mills were explicit that their principle could not justify the “doctrine of Free Trade,” even though voluntary trade is definitionally harmless. Mill [& Mill], supra note 1, at 87–88.

  26. See generally F.A. Hayek, The Use of Knowledge in Society, 35 Am. Econ. Rev. 519 (1945) (putting forward Hayek’s information theory of prices).

  27. See Moncrieff, The Uncertainty Principle, supra note 2 (manuscript at 32–33).

  28. Id. (manuscript at 27).

  29. There are many people who claim to have a spiritual connection with dead people—and even people who claim that they can speak to dead people—but there is no way, empirically or observationally, to test the validity of their claims. See generally Gillian Bennett & Kate Mary Bennett, The Presence of the Dead: An Empirical Study, 5 Mortality 139 (2000) (offering both skeptical and non-skeptical interpretations of the self-reported experience of feeling dead souls’ presence and noting that there is no way to prove or disprove the reality of the experience).

  30. Restricted individuals can report their feelings, but observers have no way to verify or falsify the truth of the subjective reports. Lie detector tests don’t work, and mind-reading machines don’t exist.

  31. See Judges 12:6 (King James).

  32. See Poe v. Ullman, 367 U.S. 497, 543, 550 (1961) (Harlan, J., dissenting).

  33. Notice, here, that the distinction I’m drawing is tantalizing close to the distinction between “harm to others” and “self-harm.” Indeed, the phrase “intrapersonal welfare tradeoff” might sound like an ivory tower evening jacket for the Mills’ much simpler concept of “self-harm.” But, of course, a boxing match does involve “harm to others.” When I punch my opponent, I’m harming her, not myself, and vice versa. Nevertheless, I would guess that the Mills would deem boxing nonregulable because the net harms to each individual are impossible to observe. Likewise, there are many interpersonal welfare transfers that do not involve “harm to others” that the Mills nevertheless deemed regulable under their principle, such as voluntary trade. As noted, “harm to others” was merely a shorthand for observability. The distinction between interpersonal transfers and intrapersonal tradeoffs is a more precise distinction for what the Mills were trying to capture.

  34. . Mill [& Mill], supra note 1, at 151.

  35. See Courtney Jung, Ran Hirschl & Evan Rosevear, Economic and Social Rights in National Constitutions, 62 Am. J. Compar. L. 1043, 1053, 1055 (2014) (finding significant variance, globally, in constitutions’ protections for social and economic rights). See generally Cass R. Sunstein, The Second Bill of Rights: FDR’s Unfinished Revolution and Why We Need It More Than Ever (2004) (arguing that the U.S. Constitution should include more social and economic guarantees); Cass R. Sunstein, Why Does the American Constitution Lack Social and Economic Guarantees?, 56 Syracuse L. Rev. 1 (2005) (considering four possible reasons that the American constitution, unlike many of its counterparts globally, fails to protect social and economic rights).

  36. Samuel R. Bagenstos, COVID and the Great Retrenchment, 63 Hou. L. Rev. 791, 802, 806, 808 (2026).

  37. Mill [& Mill], supra note 1, at 151.

  38. See Isaiah Berlin, Two Concepts of Liberty, in Liberty 172, 173 (Henry Hardy ed., 1995).

  39. See Bagenstos, supra note 36, at 817, 823. See generally Thomas Piketty, Capital in the Twenty-First Century (Arthur Goldhammer trans., Belknap Press of Harvard Univ. Press 2014) (2013) (arguing that capitalism, in the absence of intentional redistribution, will always lead inexorably to gross wealth inequality and economic insecurity among the lowest classes).

  40. . Mill [& Mill], supra note 1, at 170.

  41. See generally Robert L. Hale, Coercion and Distribution in a Supposedly Non-Coercive State, 38 Pol. Sci. Q. 470 (1923) (pointing out that economic insecurity is shaped not only by capitalism but also by common law rules of property and contract, characterizing distribution as a creature of regulation rather than a natural state of affairs); Barbara H. Fried, The Progressive Assault on Laissez Faire: Robert Hale and the First Law and Economics Movement (1998) (tracing the history and broader social context of Hale’s critique).

  42. See, e.g., U.S. Const. amend. V (requiring the government to pay money, in the form of “just compensation,” when it uses property for the general public rather than preserving an individual owner’s legally enforced monopoly over that property).

  43. See generally Thorstein Veblen, The Engineers and the Price System (Batoche Books 2001) (1921) (arguing that monetary prices do not reflect human need and that the price system therefore ought to be replaced by a system that better captures real need).

  44. For a description and critique of this theory, see generally Andrew Koppelman, Burning Down the House: How Libertarian Philosophy Was Corrupted by Delusion and Greed (2022) (showing how modern libertarianism has become increasingly tough-minded and blame-oriented, turning “personal responsibility” into a complete disclaimer of other-regardingness). For an older version of the choice-based theory, see Herbert Spencer, Social Statics: Or, the Conditions Essential to Happiness Specified, and the First of Them Developed 353 (1851).

  45. Notably, this debate centers on the philosophical debate between “negative” and “positive” conceptions of liberty (which is quite different from the lawyerly debates between negative and positive rights). A philosophical positivist would argue that the preference for heroin is clearly irrational, such that a regulator increases freedom by empowering people to escape their addiction. A negativist would take people’s preferences where she found them, refusing to question the rationality of a strong preference for heroin. The debate is long running in the philosophical literature, dating back to Kant and Hegel (positivists) and their objections to Locke and Bentham (negativists). See generally Berlin, supra note 38 (articulating this distinction between negative and positive liberty). See also Moncrieff, Constitutional Technocracy, supra note 2 (manuscript at 114, 119–20) (arguing that the Mills were not believers in negative liberty, Berlin’s treatment of them as such notwithstanding).

  46. . Bagenstos, supra note 36, at 830–31 (quoting Paul Pierson, Politics in Time: History, Institutions, and Social Analysis 135 (2004)).

  47. Id.

  48. See Samuel R. Bagenstos, Mich. L., https://michigan.law.umich.edu/faculty-and-scholarship/our-faculty/samuel-r-bagenstos [https://perma.cc/Y86S-3HZB] (last visited Jan. 21, 2026).

  49. There are, of course, other instances of this pattern in American history that I am not including here. The Civil War and Reconstruction preceded the laissez-faire era, for example, and the September 11 attacks preceded the rise of the surveillance state. See Burton W. Folsom Jr., The Fall and Rise of Laissez-Faire in the United States, 1789–1900, 23 Indep. Rev. 357, 365 (2019); Don Bell, We Built a Surveillance State. What Now?, Project on Gov’t Oversight (Aug. 20, 2024), https://www.pogo.org/analyses/we-built-a-surveillance-state-what-now [https://perma.cc/J3FX-7GZP]. For the sake of brevity, however, I am focusing only on the examples that most clearly relate to entitlement programs, which are, after all, our current topic of interest.

  50. See generally Richard A. Posner, Not a Suicide Pact: The Constitution in a Time of National Emergency (2006) (arguing that constitutional constraints must flex during times of emergency but must then be restored when the emergency ends).

  51. See supra Part III.

  52. Efraim Benmelech, Carola Frydman & Dimitris Papanikolaou, Financial Frictions and Employment During the Great Depression 1–2, 8 (Nat’l Bureau of Econ. Rsch., Working Paper No. 23216, 2017), https://www.nber.org/system/files/working_papers/w23216/w23216.pdf [https://perma.cc/6PM4-YKAW].

  53. Austan Goolsbee & Chad Syverson, Fear, Lockdown, and Diversion: Comparing Drivers of Pandemic Economic Decline 2020, J. Pub. Econ., Nov. 25, 2020, at 1, 6–7.

  54. See J. Michael McWilliams, Health Consequences of Uninsurance Among Adults in the United States: Recent Evidence and Implications, 87 Milbank Q. 443, 444–46 (2009), https://pmc.ncbi.nlm.nih.gov/articles/PMC2881446/pdf/milq0087-0443.pdf [https://perma.cc/SJF5-LUZD].

  55. See Anxiety Flu Symptoms: 7 Clues It’s Not a Viral Infection, Tenn. Behav. Health (Mar. 30, 2025), https://tennesseebehavioralhealth.com/anxiety/anxiety-flu-like-symptoms/ [https://perma.cc/R6V4-GZA6].

  56. See Phillip K. Martin & Ryan W. Schroeder, Challenges in Assessing and Managing Malingering, Factitious Disorder, and Related Somatic Disorders, Psychiatric Times (Oct. 20, 2015), https://www.psychiatrictimes.com/view/challenges-assessing-and-managing-malingering-factitious-disorder-and-related-somatic-disorders [https://perma.cc/K5TY-QYFZ].

  57. See Ruth Plackett & Ella Ferris, The Prevalence and Determinants of Health Anxiety During the Covid-19 Pandemic: A Systematic Review and Meta-Analysis, PLOS Mental Health, Dec. 30, 2024, at 2, 14. Of course, anxiety and hypochondria also rose during COVID. Id. at 14. Nevertheless, the probability of respiratory infection was demonstrably higher during the pandemic than it was during ordinary times, and the harm arising from lack of healthcare was observably higher, too.

  58. I’m setting aside second-wave feminism here, even though I suspect that second-wave feminism was important to the inclusion of “sex” in the Civil Rights Act of 1964 and to the enactment of Medicaid in 1965. But because causation is impossible to prove in these kinds of histories and because the timeline is not as suggestive with respect to feminism, it is harder to draw out feminism’s influence. Betty Friedan’s The Feminist Mystique, which is often treated as the spark that set second-wave feminism alight, was published in February of 1963, very shortly before the Kennedy assassination. Jacob Muñoz, The Powerful, Complicated Legacy of Betty Friedan’s ‘The Feminine Mystique’, Smithsonian Mag. (Feb. 4, 2021), https://www.smithsonianmag.com/smithsonian-institution/powerful-complicated-legacy-betty-friedans-feminine-mystique-180976931/?utm [https://perma.cc/97HL-ZFMT]. While Kennedy was, in fact, quite a strong ally of women’s rights, most voters probably did not think of him as such given that his movements in that direction were often subtle and not part of a broader feminist movement at the time. See generally Cynthia E. Harrison, A “New Frontier” For Women: The Public Policy of the Kennedy Administration, 67 J. Am. Hist. 630 (1980) (demonstrating Kennedy’s feminist credentials).

  59. See generally Landon R.Y. Storrs, The Second Red Scare and the Unmaking of the New Deal Left (2013) (showing, through newly declassified documents, how the Second Red Scare drove left-leaning thinkers out of government, based on faulty accusations that they were communists).

  60. See generally Nelson W. Polsby, Towards an Explanation of McCarthyism, 8 Pol. Stud. 250 (1960) (providing a postmortem on McCarthyism, published the same year Kennedy was elected).

  61. See Historic Committee Names, Hist. Art & Archives: U.S. House of Representatives, https://history.house.gov/Records-and-Research/FAQs/Committee-Names/ [https://perma.cc/94FL-VZ99] (last visited Mar. 5, 2026).

  62. See generally Lester Cole, Hollywood Red: The Autobiography of Lester Cole (1981) (noting that the effects of HUAC blacklisting, for many Hollywood actors and writers, lingered well beyond Dalton Trumbo’s return to film in 1960).

  63. See generally Brown v. Bd. of Educ., 347 U.S. 483 (1954) (holding that racially segregated schools violate the Fourteenth Amendment because they can never be “equal”).

  64. See Charles M. Payne, “The Whole United States is Southern!”: Brown v. Board and the Mystification of Race, 91 J. Am. Hist. 83, 83–84 (2004) (noting that the biggest demonstrable impact of the decision in Brown was “the mobilization of white southern resistance to racial change”).

  65. See Hugh Stephen Whitaker, A Case Study in Southern Justice: The Murder and Trial of Emmett Till, 8 Rhetoric & Pub. Aff. 189, 195, 210–11 (2005) (recounting the tragic history of Emmett Till’s murder and the sham trial that followed).

  66. See Martin Luther King, Jr., Letter from Birmingham Jail 1 (Penguin Books 2018) (1964).

  67. See generally Michael Vinson Williams, Medgar Evers: Mississippi Martyr (2011) (biography of Medgar Evers).

  68. See Lydia Saad, Gallup Vault: Few in 1963 Thought Oswald Acted Alone, Gallup (Oct. 26, 2017), https://news.gallup.com/vault/221048/gallup-vault-few-1963-thought-oswald-acted-alone.aspx [https://perma.cc/3QTK-KPFE].

  69. See generally Michael L. Kurtz, The Assassination of John F. Kennedy: A Historical Perspective, 45 Historian 1 (1982) (examining the many ways that the assassination and its aftermath undermined public faith in government).

  70. See Social Security Act (1935), Nat’l Archives (Feb. 8, 2022), https://www.archives.gov/milestone-documents/social-security-act?utm [https://perma.cc/9AHR-V98G].

  71. Abbe R. Gluck & Thomas Scott-Railton, Affordable Care Act Entrenchment, 108 Geo. L.J. 495, 557–58 (2020).

  72. See David E. Sanger, Trump Seeks Push to Speed Vaccine, Despite Safety Concerns, N.Y. Times (Apr. 29, 2020), https://www.nytimes.com/2020/04/29/us/politics/trump-coronavirus-vaccine-operation-warp-speed.html [https://perma.cc/8YKP-U8A7].

  73. See Katie Thomas et al., F.D.A. Clears Pfizer Vaccine, and Millions of Doses Will Be Shipped Right Away, N.Y. Times (Sep. 17, 2021), https://www.nytimes.com/2020/12/11/health/pfizer-vaccine-authorized.html [https://perma.cc/7MLP-PBT6].

  74. See supra Section IV.A.

  75. I do not mean to argue, here, that Kennedy’s assassination caused the War on Poverty. There is some evidence that Congress adopted the Civil Rights Act in Kennedy’s honor (and might not have had he not died), but the broader War on Poverty was not a direct reaction to the assassination. See President Lyndon B. Johnson, Annual Message to the Congress on the State of the Union (Jan. 8, 1964), https://www.presidency.ucsb.edu/documents/annual-message-the-congress-the-state-the-union-25 [https://perma.cc/4SAK-X6UB]. Instead, what I mean to argue here is that the Kennedy assassination caused a soft vibe shift in American consciousness that made the War on Poverty more politically palatable than it would have been in the absence of that shift. This claim is impossible to prove, of course, but it seems plausible.

  76. Benjamin A. Mandel & Mark S. Ludwick, Bureau of Econ. Analysis, COVID-19 Pandemic: Federal Recovery Legislation and the NIPAs 4–5 (2021), https://www.bea.gov/sites/default/files/2020-06/COVID-19 Pandemic-Federal Recovery Legislation and the NIPAs.pdf [https://perma.cc/6QW3-58J7].

  77. Vaccine policies are a possible exception to this statement. The current CDC’s efforts to limit access to COVID vaccines might, in fact, give rise to a COVID resurgence. See, e.g., Joshua P. Cohen, New Limits on Who’s Eligible for COVID-19 Vaccines Raise Access Questions, Forbes (Sep. 2, 2025, at 17:28 ET), https://www.forbes.com/sites/joshuacohen/2025/09/01/new-limits-on-whos-eligible-for-covid-19-vaccines-raise-access-questions/ [https://perma.cc/WHJ8-QJH9].

  78. See generally M. Mitchell Waldrop, Complexity: The Emerging Science at the Edge of Order and Chaos (1992) (describing complex systems theory).

  79. See generally P.W. Anderson, More Is Different, 177 Science 393 (1972) (explaining the impossibility, within complex systems, of predicting macrolevel phenomena by observing the behavior of microlevel constituents).

  80. See generally My Villius Zetterholm & Päivi Jokela, Addressing Complexity in the Pandemic Context: How Systems Thinking Can Facilitate Understanding of Design Aspects for Preventive Technologies, Informatics, Jan. 11, 2023, at 1, 5, 11 (applying complexity theory to pandemic response in order to suggest some changes to the design of preventive technologies).

  81. See Megan Brenan, Amid Pandemic, 79% of K-12 Parents Support In-Person School, Gallup (Mar. 11, 2021), https://news.gallup.com/poll/336173/amid-pandemic-parents-support-person-school.aspx [https://perma.cc/3883-H4CJ].

  82. See Jeremy Singer, School Reopening Decisions During the COVID-19 Pandemic: What Can We Learn from the Emerging Literature? 2 (Brown Univ. Annenberg, Working Paper No. 22-617, 2022).

  83. See Samuel Hume, Samuel Robert Brown & Kamal Ram Mahtani, School Closures During COVID-19: An Overview of Systematic Reviews, 28 BMJ Evidence-Based Med. 164, 172 (2023) (reviewing empirical data on the effects of school closures).

  84. See generally Abrar A. Chughtai, Holly Seale & C. Raina Macintyre, Effectiveness of Cloth Masks for Protection Against Severe Acute Respiratory Syndrome Coronavirus 2, 26 Emerging Infectious Diseases, no. 10, 2020, at e1 (finding that cloth masks are less effective than N95 masks).

  85. Masking might have slowed the virus’s spread—or it might not have. Again, in a complex system, discrete causes are extraordinarily difficult to trace. The claim in text is only that any benefits could not be observed or demonstrated, but the costs were concrete and observable.

  86. See Anderson, supra note 79, at 393, 396.

  87. See generally Philip Elliott, How Distrust of Donald Trump Muddled the COVID-19 ‘Lab Leak’ Debate, Time (May 26, 2021, at 14:37 CT), https://time.com/6051414/donald-trump-wuhan-laboratory-leak/ [https://perma.cc/7ZQL-P5WN] (describing the politicization of the debate over origins of the COVID-19 pandemic).

  88. See generally Roger W. Garrison, The Roaring Twenties and the Bullish Eighties: The Role of Government in Boom and Bust, 7 Critical Rev. 259 (1993) (comparing the laissez-faire fiscal policies of the 1920s to those of the 1980s).

  89. See Julian di Giovanni, et. al, Quantifying the Inflationary Impact of Fiscal Stimulus Under Supply Constraints, Fed. Rsrv. Bank of N.Y., Staff Rep. No. 1050, at 1, 5 (2023), https://www.newyorkfed.org/medialibrary/media/research/staff_reports/sr1050.pdf [https://perma.cc/56VK-B8WK].

  90. See generally Miriam F. Weismann, How the “Great Resignation” and COVID Unemployment Have Eroded the Employer Sponsored Insurance Model and Access to Healthcare, 49 Am. J.L. & Med. 415 (2023) (noting that many people who quit their jobs were also choosing to give up their health insurance).

  91. See id. at 421–23.

  92. See supra Section IV.A.1.

  93. See Bagenstos, supra note 36, at 834.

  94. For current versions of the Social Security attacks, see, e.g., Lorie Konish, As Social Security Faces an Uncertain Future, the Debate on Privatization Heats Up Again, CNBC (Mar. 14, 2025, at 14:07 ET), https://www.cnbc.com/2025/03/13/as-social-security-faces-an-uncertain-future-some-say-it-should-be-privatized.html [https://perma.cc/WU95-8JMZ]; Social Security Reform: Options to Adjust Benefits, Peter G. Peterson Found. (Apr. 24, 2025), https://www.pgpf.org/article/social-security-reform-options-to-adjust-benefits/ [https://perma.cc/6WZY-LKA2]. For the Obamacare attack’s mixed success, see generally Nat’l Fed’n of Indep. Bus. v. Sebelius, 567 U.S. 519 (2012) (upholding the individual mandate as a tax but not as a regulation and holding that the Medicaid expansion cannot be obligatory for states).

  95. See generally David W. Brown, Amanda E. Kowalski & Ithai Z. Lurie, Long-Term Impacts of Childhood Medicaid Expansions on Outcomes in Adulthood, 87 Rev. Econ. Stud. 792 (2020) (finding that Medicaid beneficiaries return to the federal government $0.58 of every dollar spent, through increased tax revenue, while also benefitting physically, financially, and mentally from the coverage).