Uncompensated Takings: Insurance, Efficiency, and Relational Justice

Article - - Issue 5


The basic principles governing the power of eminent domain might seem so well established by now as to be beyond controversy: the government may confiscate private property for public use but must compensate the owner for the value of what the government took. Those rules have been part of the U.S. Constitution since the enactment of the Fifth Amendment’s “Takings Clause,” which declares “nor shall private property be taken for public use without just compensation,” and many state constitutions contain similar provisions.[1] A long line of U.S. Supreme Court cases has emphasized that the Takings Clause’s “just compensation” requirement mandates that the government pay “a full and exact equivalent” in value for the seized property.[2]

Yet in recent decades prominent scholars have argued that the government should not be obligated to compensate for taken property at all.[3] Building upon an idea, common in economic analyses of eminent domain, that takings compensation is a form of government-provided insurance, these scholars argue that leaving property owners to purchase that insurance from private companies would benefit society by reducing both “moral hazard”—distortions in landowners’ incentives to improve their property—and administrative costs. Today the case against compensation—for convenience, one might call it the “anti-compensation thesis”—is a staple of many economically influenced analyses of takings law, and it routinely appears in leading casebooks.[4]

The influence of these arguments may, at first glance, seem surprising, since the government’s explicit constitutional obligation to pay compensation for taken property seems unlikely to be repealed.[5] In fact, however, the issues raised by the anti-compensation thesis are pivotal for understanding eminent domain for two reasons.

The first reason is practical. Although the Takings Clause requires payment of “just compensation,” the scope of the Takings Clause’s application is still very much contested.[6] The Clause itself does not indicate exactly what sorts of government actions qualify as “takings” and therefore are bound by the Clause’s “public use” and “just compensation” requirements.[7] Thus, courts continue to confront questions about whether certain physical interferences with possession are takings, as well as whether certain regulations are takings.[8] Nor have courts established a canonical definition of “property” for constitutional purposes, and even the U.S. Supreme Court has, in different cases, defined property in ways that are in tension with each other.[9] As a result, the set of circumstances requiring “just compensation” can expand or contract.[10]

If the anti-compensation thesis is correct that ideally the government would never pay for property that it takes, then courts’ adopting very restrictive criteria for what constitutes “property” or a “taking”—and thus for when “just compensation” is owed—would be reasonable, indeed advisable. Alternatively, if those arguments are not plausible, then the theoretical availability of private insurance for takings losses offers no reason to adopt a narrow view of the Takings Clause’s scope.

Likewise, the Takings Clause does not specify the amount of compensation that constitutes “just compensation.” Although the U.S. Supreme Court understands “just compensation” to mean fair-market-value compensation, that interpretation establishes merely a constitutional minimum.[11] Congress and individual states are free to provide a higher level of compensation, and in many cases they do.[12] If, however, the anti-compensation thesis is correct, then the optimal amount of compensation is zero, and, consequently, laws that increase takings compensation are unnecessary at best, and harmful at worst.

Beyond these practical considerations, the anti-compensation thesis also has fundamental implications for the basic principles of takings law. The challenge that the thesis poses to standard justifications for requiring government compensation reveals a critical gap in contemporary understandings of eminent domain’s normative foundations. This Article aims to remedy that gap.

This Article has three major contentions: First, although the anti-compensation thesis may seem startling, the standard justifications for requiring takings compensation struggle to meet its challenge. If those established justifications were the only relevant considerations, then the anti-compensation thesis would actually be quite plausible.

Second, the thesis nevertheless is false. Contrary to what economic analyses of eminent domain commonly assume, takings compensation and insurance are fundamentally distinct because they involve importantly different relationships between the party suffering a loss and the party paying “compensation” to alleviate that loss. When property is taken, the relationship among the property’s owner, the community that took the property, and the entity that pays compensation involves a specific type of justice, a type that this Article calls “relational justice.” This sort of justice is distinct from other forms of justice common in legal scholarship—such as distributive justice, procedural justice, and corrective justice—and it has been overlooked in existing literature on eminent domain. This Article argues that recognizing the relevance of relational justice, and what this Article terms the “duties of reasonable accommodation” that it imposes, shows how the government’s paying compensation for taken property has an irreplaceable role in legitimizing exercises of eminent domain.

Third, there is no need to choose between relational justice and economic efficiency, because the economic-efficiency concerns that motivate the anti-compensation thesis are illusory. Although proponents of the anti-compensation thesis have carefully argued that government compensation for takings creates inefficiencies, they have paid remarkably little attention to assessing the size of those alleged inefficiencies. This Article remedies that crucial gap in the existing literature, arguing that any such inefficiencies are likely to be negligible and that their net total may even be negative. Thus, economic efficiency considerations offer no reason to disregard the requirements of relational justice.

The Article proceeds as follows. Part I describes the two standard arguments in favor of replacing government compensation with private insurance and then examines how they do, or could, respond to standard justifications that scholarship has identified for requiring government compensation. Part II develops the argument that relational justice requires government compensation. The discussion begins by identifying a critical oversight in the assumption that takings compensation is merely a form of government-provided insurance. It then analyzes the common characterization of takings as “forced sales,” bringing to light the indispensable role that compensation plays in making a transfer of ownership legitimate. Part II next addresses the implications of the fact that even though exercises of eminent domain are coercive, they are not wrongful. It then explains why relational justice does not require that the government compensate for every burden that it imposes—that is, for every cost of “legal transitions”—and why taxation and takings are fundamentally distinct. Part III reexamines the economic case for the anti-compensation thesis, arguing that the supposed efficiency gains from replacing government compensation with private insurance are likely to be negligible at best, and thus they provide no reason to disregard the requirements of relational justice. Part IV concludes.

I. The Case Against Compensation

The first step toward evaluating the anti-compensation thesis is understanding its foundations. Historically, the case against requiring compensation for taken property has rested on two concerns, one involving “moral hazard,” the other focusing on administrative costs. These arguments run squarely contrary to the standard justifications that existing literature has offered for requiring government compensation. This Part will survey both competing sets of arguments and examine how the anti-compensation thesis’s defenders respond, or could respond, to the standard justifications for requiring government compensation. The responses available to the anti-compensation thesis will, perhaps surprisingly, prove largely effective against the standard justifications for requiring compensation. If those justifications were the only relevant considerations, then the anti-compensation thesis would have at least prima facie plausibility. This conclusion then sets the stage for Parts II and III, which will show why that superficial plausibility does not survive closer examination.

A. The Core Case

The roots of the anti-compensation thesis lie in Lawrence Blume and Daniel Rubinfeld’s seminal article providing “an economic analysis of compensation as a form of insurance.”[13] The basic structure of this analysis is simple. Claims for takings compensation arise from the losses that owners of taken property suffer because their property has become more valuable to society as a component of some public project than it would have been in the owners’ hands. In this respect, losing one’s property to an exercise of eminent domain might seem similar to losing it to any other event that is beyond the owner’s control, such as a fire or a natural disaster.[14] Meanwhile, the taken property’s owner has already been paying to the government, in the form of higher taxes, the sums needed to make the general system of takings compensation possible.[15] On this analysis, these tax payments are analogous to the premiums paid to purchase private insurance against a loss, and the government’s paying compensation is analogous to payment of an insurance claim for that loss.[16] Treating the “just compensation” requirement in this way as a form of government-provided insurance is now common even among scholars who do not go on to advocate replacing government compensation with private insurance.[17]

This insurance-based conception of takings compensation sometimes takes the form of assertions that such compensation is merely one species of a broader general category of “transition relief”—that is, compensation from the government for costs that individuals suffer as a result of some legal change.[18] The fact that legal change is pervasive and routinely results in individuals’ suffering losses for which they receive no compensation then raises the question of why takings losses should be singled out to receive public compensation when other legal-transition losses do not.[19] Some commentators go even further, arguing that takings losses are fundamentally no different from losses caused by any of the myriad factors that affect property’s value, including factors that have nothing to do with government action, none of which are accompanied by government compensation.[20]

Having thus conceived of takings compensation as merely a form of government-provided insurance against one particular type of risk among many similar types of risk, the canonical case for the anti-compensation thesis then argues that providing this insurance would optimally be left to the private market.[21] This conclusion rests on two grounds: a concern that the availability of government compensation biases property owners’ incentives in a socially inefficient direction, and a concern about unnecessarily large administrative costs. Each concern can be briefly summarized.

1. Avoiding Moral Hazard.—If takings compensation is a form of insurance, then it is natural to worry that its availability may create a “moral hazard.” That is, it may encourage the insured to act in ways that increase the risk of socially undesirable results.[22] The specific hazard attributed to takings compensation is an increased risk that landowners will make excessive improvements to their property, improvements that will go to waste if the government takes the property for a public project that has no use for that improvement.[23]

A simple hypothetical example can make the problem clear. Suppose that an owner is considering improving her property and that the improvement will cost $4,000 to construct. Assuming that she is economically rational, she will make the improvement only if the payoff for doing so is greater than $4,000.[24] Further suppose that she concludes that the improvement would offer a payoff of $5,000 over its lifespan.[25] Ordinarily, one would expect her to make the investment (unless an even better investment opportunity presents itself) and that doing so would increase the total amount of wealth in society by $1,000. In other words, her making the investment would be socially efficient.

But now add to this picture the possibility that the government might exercise its power of eminent domain over her property. Suppose that there is a 50% chance that the government will take her property to build an airport and that the government’s doing so would itself substantially increase total social wealth. For the sake of simplicity, assume that if the taking occurs, it will occur immediately after the owner has finished constructing the improvement.

In this revised scenario, the owner’s building the improvement would no longer be socially efficient. The mathematically expected value of the improvement’s payoff would now be only $2,500 because there is now a 50% chance that the improvement will actually return zero rather than the $5,000 that is hoped.[26] Meanwhile, constructing the improvement would still cost $4,000. Thus, the net expected return from building the improvement would be negative: –$1,500.

If the government pays no compensation for the property that it takes, then a rational owner will make this same calculation herself, realize that the mathematically expected outcome of making the improvement is a personal loss of $1,500, and choose to invest her $4,000 somewhere more profitable. The course of action that is personally best for her and the socially efficient course of action are the same.

However, the argument continues, this synchronization of personal advantage and social efficiency ceases if the government fully compensates owners for the value of taken property.[27] For if the government reimburses owners for the value of any taken improvements, then an owner who is deciding whether to invest in an improvement will, quite rationally, pay no attention to the probability of the property’s being taken. In this example, if the property is not taken, then the owner would receive $5,000 in value from enjoying the effect of the improvement on her property, and if the property is taken, then she would receive $5,000 in compensation from the government. Either way, she receives $5,000 in return from a $4,000 investment. Hence, she will choose to make the investment, even though, as noted a moment ago, doing so imposes a $1,500 expected loss on society. The decision that is most advantageous to her now differs from the socially efficient decision. Therefore, this argument concludes, the government’s paying compensation for taken property produces inefficient incentives to overinvest in improvements that will go to waste if the improved property is taken.[28]

By contrast, leaving a property owner to rely on private insurance for reimbursement of the value of taken property would avoid this hazard because the price charged by private companies to insure against takings would depend on both the value of the insured property and the risk of a loss. As a result, insurance would become more expensive if an improvement increased the insured property’s value or if a substantial risk arose that the property might be taken.[29] This increase in insurance premiums would encourage owners to take account of the risk of a taking—i.e., to “internalize” that risk—and thus to make the socially optimal decision when deciding whether to improve their property.[30] The ultimate effect, the argument concludes, would be to avoid the “moral hazard” that leads to a socially inefficient result when the government pays compensation.

2. Administrative Cost Savings.—The second canonical argument for the anti-compensation thesis springs from the observation that funding government compensation for taken property requires increasing tax revenue (assuming that the government does not cut spending elsewhere).[31] Thus, if government compensation for takings is a system of government-provided insurance, those higher taxes are effectively “premiums” paid for this insurance.[32]

As a result, the argument continues, replacing government compensation with private insurance would not increase owners’ net expenses, since the money spent to purchase private insurance would be offset by the money saved in correspondingly lower taxes.[33] Indeed, owners would actually save money because private insurance is likely to have lower administrative costs than government compensation schemes do.[34] Thus, private insurers could simultaneously offer the same amount of coverage as the government does and do so at a lower price. Moreover, the reduction in taxes would also benefit non-owners, who have no need for takings insurance. Everyone, the argument concludes, would be economically better off.[35]

B. Insurance and the Standard Justifications for Compensation

These arguments for the anti-compensation thesis run squarely contrary to a standard set of justifications for requiring government-paid compensation for taken property. Hence, the thesis’s plausibility will depend on its ability to answer those justifications. As will soon become evident, the standard justifications score no clear victory over the private-insurance alternative.

Categorizations of these justifications commonly distinguish between those that focus on incentives for socially efficient behavior and those that focus on “fairness.”[36] For convenience, each set of justifications can be discussed separately.

1. Efficiency.—One common rationale for the “just compensation” requirement involves a trio of closely related arguments focused on socially efficient decision-making. A brief survey of these arguments can make their difficulties evident.

The first of the three arguments asserts that relieving governments of an obligation to compensate for losses inflicted by their takings would subject governments to a “fiscal illusion” about the size of their projects’ social costs.[37] In economics terminology, removing the compensation obligation would allow governments to avoid internalizing all of the costs of their takings decisions.[38] The ultimate result would be governments’ taking private property more often than is socially optimal.

The “fiscal illusion” justification is now well-established, and it has an equally well-established counter-argument: government officials’ primary motivations are political rather than fiscal, and therefore reducing the fiscal costs of takings is unlikely to have much influence on officials’ decisions about what projects to pursue.[39] This response is itself somewhat controversial. For example, it may be that local governments are more sensitive to fiscal incentives than state and federal governments are, and thus are relatively more susceptible to “fiscal illusion,” at least with respect to projects that are not receiving state or federal funding.[40] At the same time, there is empirical evidence that governments do respond more to political incentives than they do to fiscal concerns.[41]

Nevertheless, even if the fiscal-illusion concern has some prima facie plausibility in some circumstances, the presence of a private-insurance alternative to government compensation would reduce or eliminate much of that concern’s force. Whether or not governments are responsive to fiscal concerns, no one disputes that they respond to political pressure, and private companies that provide takings insurance would have a natural incentive to exert political power against projects that required a taking. There is no obvious reason to think that the political effect of this lobbying would be less effective in forcing governments to recognize the costs of their takings than the fiscal effect of requiring compensation would be.

A second efficiency-based justification for requiring government compensation focuses on the risk of “rent-seeking” by opportunistic, politically influential private actors hoping to induce the government to take property for public projects that will benefit them personally even when society as a whole would be better off if the project did not occur.[42] This second justification posits that a government’s susceptibility to that sort of influence depends, at least in part, on what the government would have to pay to undertake those socially inefficient projects. Increasing those costs by requiring the government to pay compensation for taken property might then limit that susceptibility, thereby creating an obstacle to opportunism.[43]

Once again, however, this justification loses force if government compensation is replaced by private insurance. As noted a moment ago, the companies providing that insurance would have a natural incentive to oppose such takings (since they would have a natural incentive to oppose all takings), and there is no reason to think that their political influence would be any less effective than fiscal constraints are at thwarting such rent-seeking.

The third justification is a mirror image of the second, focusing on impediments to efficient takings rather than on the risk of inefficient takings: requiring the government to pay compensation serves to “buy off” politically influential property owners who otherwise would have a natural incentive to use their influence to stop socially beneficial takings of their own property.[44] This justification is, in effect, an observation that compensation might create a socially beneficial form of “moral hazard” in the takings context, decreasing owners’ incentives to “take precautions” by lobbying to avoid the government’s taking property.[45]

Unlike the other two efficiency-based justifications, this third justification’s force is not undermined by the availability of private insurance. Even if there is no need to “buy off” individual owners, since they are insured against takings losses, there could still be a need to “buy off” the companies that provide this insurance and thus have a strong incentive to oppose governments’ exercising their power of eminent domain.

However, the justification itself seems, at best, markedly incomplete. First, it rests on an assumption that no better way to address harmful political pressure exists than to, in effect, bribe it away. Moreover, since the government does not generally attempt to purchase the acquiescence of policies’ opponents by paying them money, this justification creates a puzzle: Why should eminent domain be a pronounced exception to the government’s typical practices when undertaking policies that do not enjoy universal approval?[46] There is no obvious way for this justification to answer that question.

2. “Fairness.”—Even if one sets aside the difficulties just noted, efficiency-based justifications for requiring government compensation all face a fundamental problem. What the Constitution requires, and courts have repeatedly demanded, is not payment of “efficient” compensation but rather payment of “just” compensation.[47] In the oft-quoted words of Justice Black in Armstrong v. United States,[48] the Fifth Amendment’s “just compensation” requirement “was designed to bar Government from forcing some people alone to bear public burdens which, in all fairness and justice, should be borne by the public as a whole.”[49] Hence, no account of eminent domain’s normative foundations will be adequate unless it includes a plausible account of how, and under what circumstances, takings are just and not merely efficient.[50] Even if a concern for efficiency is a part of justice, it is not the whole of justice.[51]

The second set of standard justifications for requiring government compensation springs from this concern about what justice requires. However, as will soon be clear, these “fairness”-based justifications are also vulnerable to the anti-compensation thesis’s challenge. When faced with the possibility of private insurance as an alternative to government compensation, the standard justifications struggle to explain why “just compensation” is not zero compensation.

One of the most influential “fairness”-based justifications rests on Frank Michelman’s famous assertion that utilitarian analyses of takings should include assessments of “demoralization costs”—i.e., the value of unhappy feelings that property owners would experience if their property were taken without compensation and that other owners would experience from worrying about whether their property too would suffer such a fate.[52] Michelman argued that when demoralization costs from denying compensation are higher than the “settlement costs” of administering the compensation-payment system, then society would benefit from requiring the government to compensate.[53] On this account, then, there is no inherent problem with uncompensated takings, only a problem that may arise if those takings happen to affect the feelings (and thus the “utility”) experienced by various people in certain ways.[54]

The availability of private insurance might go far to alleviate this unhappiness, since insurance payments would diminish owners’ monetary losses.[55] Because reducing owners’ net losses would leave owners with less cause for regret, demoralization costs would accordingly shrink, and the case for requiring government compensation would also diminish.[56]

Moreover, the demoralization-costs argument faces a fundamental objection arising out of its dependence on utilitarianism.[57] On such accounts there is nothing inherently wrong with the government’s forcibly taking property and not paying for it. Uncompensated takings are condemned only if they produce negative psychological reactions. There may, however, be more than one way to avoid those reactions. Thus, this theory would offer no objection to the government’s taking property without paying for it, so long as those takings are accompanied by a suitably effective plan to avoid upsetting the populace, perhaps by deploying pervasive propaganda or imposing strict secrecy measures to ensure that the government’s actions did not become widely known. Michelman himself suggested that, at least in some cases, the government “could reasonably count on holding down the demoralization costs of a failure to compensate either by preventing the matter from becoming widely known or by not revealing the general implications of this particular decision,” even when “[f]airness rather clearly requires compensation . . . .”[58] The difficulty now is evident: suggesting that justice is indifferent between compensating people and deceiving them, except insofar as one approach or the other is more effective at producing positive feelings, calls into question whether this account really rests on considerations of justice at all.[59]

A second “fairness”-based account—one more closely related to the ordinary idea of justice—rests on a particular concern about equal treatment, frequently associated with the Armstrong Court’s assertion that the purpose of requiring governments to compensate for taken property is to avoid “forcing some people alone to bear public burdens which, in all fairness and justice, should be borne by the public as a whole.”[60] The equal-treatment account interprets this principle as concerned about the inequity of burdening one person with a taking when other people, perhaps very similarly situated, are not burdened.[61]

Although concerns about equal treatment may have considerable intuitive appeal, this particular argument provides little reason to favor government compensation for takings over a system of private insurance. Under a system of private insurance, all property owners are burdened by having to purchase insurance, and each owner of taken property receives “compensation” payments from the relevant insurer. Thus, each property owner equally bears the cost of takings. Although the burden of purchasing insurance falls only on those who own property, unequal treatment of that sort is common and not generally considered unjust. Property owners routinely have burdens that are considered legitimate despite not being shared by everyone—for example, an obligation to pay annual taxes on the value of their property.[62] Thus, there is no obvious reason to conclude that the particular sort of unequal treatment that remained after replacing government compensation with private insurance would be unjustified.[63]

Moreover, even among property owners the burdens imposed by the law are frequently unequal and uncompensated. As Thomas Merrill has noted, the equal-treatment argument “fails to explain why some forms of ‘unequal’ treatment are compensated while others are not. . . . Clearly, some limiting principle other than equal treatment is needed to explain why compensation should be forthcoming in some cases but not in others.”[64] This objection is, in effect, a specific application of a general argument that any adequate justification for requiring compensation for takings will need to explain why takings should receive compensation when many other burdensome “legal transitions” do not.[65]

Hence, neither established efficiency-based justifications nor established “fairness”-based justifications offer a convincing reason to reject the anti-compensation thesis and require government compensation for takings.[66] If these justifications were the only relevant considerations, then leaving owners to rely on private insurance to protect against takings losses could seem a quite plausible alternative to government compensation. As the next Part will show, however, the considerations behind these standard justifications are incomplete. They fail to account for an additional important but often overlooked consideration—the requirements of what this Article terms “relational justice.”

II. Takings and Relational Justice

The discussion to this point has argued that the anti-compensation thesis’s challenge to standard justifications for the “just compensation” requirement cannot easily be dismissed. However, as this Part will explain, neither the considerations that produce those standard justifications nor the considerations that motivate the anti-compensation thesis exhaust the concerns of justice. In particular, established arguments have failed to notice that justice has an important “relational” dimension in the eminent domain context. Recognizing that dimension both reveals the fundamental error in the economic approach of considering takings compensation to be merely a form of government-provided insurance and explains why private insurance is an inadequate substitute for government compensation.

A. Justice’s Intrinsic and Relational Aspects

Many of the arguments discussed so far—both for and against the “just compensation” requirement—have shared two basic limitations, both of which are common in contemporary analyses of takings law. First, they are wholly instrumental. That is, they justify requiring compensation, or not requiring compensation, by appealing to some other presumed good that compensation policy is asserted to promote—such as socially efficient decision-making or avoiding feelings of demoralization. This fact is not surprising in light of the widespread influence of economic analyses of law, since economic reasoning is paradigmatically instrumental.[67] Justice, however, is not a purely instrumental good. Rather, it is important, at least in part, for its own sake.[68] Thus, one would expect that the most compelling account of the “just compensation” requirement would offer, at least in part, an identification of compensation’s intrinsic importance, not merely its instrumental usefulness.

Second, both the standard justifications for requiring compensation and the arguments for replacing public compensation with private insurance have largely been indifferent to the relationships that exist among the various people and institutions who are involved when the state exercises its power of eminent domain. The importance of relationships among specific parties is again easy to overlook from a purely economic perspective because the quintessentially economic concerns of efficiency—i.e., the total amount of wealth in society—and “fairness”—understood as the distribution of wealth among broad social classes—both focus on aggregates of people and institutions.[69] The aggregate point of view necessarily abstracts away from relationships among individuals, except to the extent that those relationships, taken together, instrumentally affect the total amount of wealth or its general distribution. Thus, to the extent that such relationships have normative importance—and this Part will argue that they do—even the most plausible accounts discussed so far are necessarily incomplete.

The one exception to this characterization is the “equal treatment” account, which differs from the other accounts in two significant ways. First, at least in some forms, it does regard equal treatment as intrinsically important, not merely as instrumentally desirable for attaining some other end. Second, it does recognize the significance of one specific type of relationship among individuals: the quantitative relationship of being more burdened by a taking than other people are.[70]

However, the relationship upon which the equal-treatment account focuses is quite minimal. It is not based on any interaction or connection between the related parties, but only on a comparison between them, based on whether a given owner’s burden is greater than, less than, or equal to other people’s burdens. This sort of merely comparative relationship exists among every person, and even between persons and inanimate things. It is the same type of relationship that everyone has to the Queen of England or to Mount Everest—i.e., being younger (or older) than her or smaller than it.

These highly attenuated sorts of relationships may be of interest to metaphysicians, but they are quite remote from the interpersonal relationships that play a central role in human life.[71] Some of those interpersonal relationships are largely involuntary, such as being in a particular family or political community, while others are chosen, such as being a creditor or a member of a club. Both types of relationship, however, are normatively fertile in that they can give rise to obligations and entitlements, whether it is a moral duty to care for an aging parent or a moral and legal duty to repay a loan from a friend. As a result, justice inevitably will have something to say about them.

A brief terminological note may be useful. As will soon be evident, the issues of justice that arise out of these relationships—what, for convenience, one might call “relational justice” issues—are distinct from oft-discussed concerns about the overall distribution of wealth in society—i.e., from the “distributive justice” that today is frequently equated with “fairness.”[72] Although there is no canonical taxonomy of types of justice, the closest traditional category into which relational justice would fit is perhaps “commutative justice.” However, even classic sources disagree about the sorts of considerations that fall within that category. For example, in commenting on Aristotle’s analysis of justice, Aquinas applied the term “commutative justice” both to justice that involved voluntary transactions and to justice that involved rectifying crimes and other wrongs.[73] By contrast, Hobbes asserted that “[t]o speak properly, commutative justice, is the justice of a contractor . . . .”[74]

Moreover, neither definition quite captures the specific type of situation that arises in the context of eminent domain. Justice in consensual transactions is an imperfect fit because takings are inherently non-consensual. Rectification of wrongs (commonly associated with “corrective justice”) is not quite apt either because the state does not commit a wrong when it exercises its power of eminent domain.[75] Ultimately, however, taxonomies are useful only to the extent that they clarify the nature of the things that they categorize. Because pigeonholing the sorts of considerations raised by takings compensation into one philosophical category or another is unlikely to offer any additional clarity, this taxonomical question can safely be put aside, and this Article will simply use the term “relational justice” to refer to the general category of justice at issue here.[76]

For present purposes, two specific kinds of relationship will turn out to be essential for understanding takings law’s “just compensation” requirement—both why it exists and why private insurance would not be an adequate substitute. One of these relationships is that which exists between property owners and the community to which each belongs.[77] The second is the relationship that arises when one person benefits by imposing a burden on another person.

As the discussion in this Part will explain, one of the immediate benefits of considering these relationships is that doing so brings to light the error in assuming that takings compensation is merely a form of government-provided insurance. Recognizing that error in turn will clear the way to see how compensation’s role in eminent domain is not principally to mitigate losses that may happen to befall a property owner, but instead to make the exercise of eminent domain be legitimate. It will then follow that compensation must come from the community that benefited by imposing the burden created by the taking, not from a third party whom the burdened owner had paid to protect it from that burden. That is, the government must compensate for the property that it takes, and the reason it must do so is to make that taking be relationally just.

B. Reconsidering the Insurance Analogy

As noted earlier, the anti-compensation thesis rests on a fundamental assumption that takings compensation is a form of government-provided insurance against takings losses.[78] That assumption then raised the question of whether private insurance might serve this function just as well as government insurance does, or perhaps even better.

At first glance, the assumption that takings compensation is merely a form of insurance might seem plausible. From one perspective, there may seem to be no difference between, for example, an increase in wildfire risk’s causing an owner to lose her house when it is consumed in a wildfire and the increase in wildfire risk’s causing the owner to lose her house when the government takes it to build an additional fire station to help control wildfires. In both cases the owner loses the property, and in both cases the ultimate cause is the increase in wildfire risk. Since the former loss is the sort of loss for which private insurance is routinely relied upon, why shouldn’t the latter be as well?

However, two fundamental structural differences between insurance and takings compensation become evident once one considers the relationships between the people paying the compensation and the owners who receive it.[79] Treating takings compensation as akin to insurance rests on a mistake about those relationships.

The first important structural difference arises from the fact that government-paid compensation is funded by tax revenue. Since property taxes are only one type of tax, the people who pay “premiums” for this “insurance” and those who are “covered” by it are not the same.[80] The set of “premium” payers is the set of all taxpayers—including those who pay income tax or sales tax—but the people “covered” are only those taxpayers who are also property owners.[81] This mismatch implies that what non-owner taxpayers are paying for is not “insurance.” Taxpayers who do not own real property, and therefore are at zero risk of a taking, obviously are not buying insurance against a loss to which they are inherently immune.[82]

A second important structural difference springs from the fact that the community, acting through the government, does not merely compensate for the taking but also benefits from the taking. In this crucial respect, the losses imposed by takings are fundamentally different from losses that are typical in ordinary property and casualty insurance. If a burst water pipe damages an office, a wildfire consumes a house, or a collision damages an automobile, those losses do not benefit anyone, except incidentally those who profit from repairing such damage. Unlike in eminent domain cases, the damage suffered in these cases is pure loss. Because no one benefits from these sorts of losses, it is natural to conclude that the burden of such losses must be borne by whoever was unfortunate enough to suffer them, and thus that the victim must rely on private insurance for relief.[83] No one else bears a relevant relationship to the burden.[84]

By contrast, an essential fact about takings losses is that the burden placed on taken property’s owners is related to producing a benefit to others, and that relationship is not merely incidental. The burden is not merely a byproduct of the government’s producing the public benefit but rather a means of the government’s doing so.[85] The benefit results from imposing the burden, such as by using the taken property as the site for a park, a school, or a fire station.[86]

As a result, in takings cases, but not in the case of accidents or natural disasters, a particular relationship exists between the cause of the loss and the party that suffers the loss: the former has benefited by imposing a loss on the latter.[87] Recognizing this relationship suggests that what connects taxpayers who are paying for takings compensation is not that they all are protected by a policy of compensating for takings—as noted earlier, they are not—but rather that they all belong to the community that benefits from the taking.[88]

This recognition also casts a different light on the Armstrong principle. When the Armstrong Court referred to “public burdens which, in all fairness and justice, should be borne by the public as a whole,” it was silent about why those burdens should be borne by the entire public.[89] As noted earlier, a common assumption is that the court thought that the public should bear those burdens because otherwise, unequal treatment would result.[90] However, an equally possible interpretation is that the court meant that the public should bear those burdens because the public imposed them for its own benefit.[91] In fact, in context, the latter interpretation is actually more plausible than the former. In the same paragraph in which the Court stated the principle, it emphasized that the government executed the taking “for its own advantage,” but the Court made no mention of how the owner’s burdens compared to other people’s.[92]

The existence of this particular relationship between an owner burdened by a taking and the public that benefits by imposing that burden suggests that when the public, acting through the government, pays compensation for takings, it is acting as a purchaser of a good rather than as an insurer against a loss (that it itself caused). The next subpart will examine that suggestion, and its implications, in more detail.

C. Relationships and Justification

1. Eminent Domain as a “Forced Sale.”—Because the losses suffered when property is taken directly benefit the community that takes that property, it is not surprising that courts and commentators have often found it natural to refer to the government’s condemnation of private property as a “forced sale.”[93] For example, Justice Cardozo declared that “‘[c]ondemnation’ is an enforced sale, and the state stands toward the owner as buyer toward seller. On that basis the rights and duties of each must be determined.”[94] A closely related strain of thought treats the state’s obligation to pay compensation for takings as equitable, involving an implied contract.[95] Reinforcing the idea that takings fundamentally are compelled versions of voluntary sales, many state statutes require that the government first attempt to purchase property in a negotiated transaction before exercising its power of eminent domain over that property.[96]

Once one thinks of the two parties in an instance of eminent domain—the state and the owner—as a buyer and a seller, then it becomes easy to recognize why it is important that takings compensation be paid by the taker rather than by private insurance purchased by the taken property’s owner. Consider the following simple hypothetical: someone goes to a grocery store and walks out with a carton of milk without paying for it, but a bystander takes pity on the shopkeeper and pays her a sum equivalent to the price of the milk. In this case, the shopkeeper has suffered no loss, but the unpaid taking of the milk nevertheless is wrongful. A person who takes milk from a store avoids acting wrongly only if he pays for the milk that he takes.[97] Payment of the stated price makes a customer’s acquisition of the milk legitimate—payment is what makes him a “customer” rather than a “thief”—but only if the person who pays the price is the same as the person who took the milk. Otherwise, the person who takes the milk has misappropriated it, and the shopkeeper has merely been fortunate to benefit from the bystander’s charity.

Changing the scenario slightly so that the money that the shopkeeper receives is an insurance-claim payment from an anti-theft policy that she had purchased does not affect the injustice of the customer’s action. The payment of money equal to the price of the milk (“compensation”) absolves the customer from misappropriation only if the customer himself makes the payment. The relationship among the customer, the shopkeeper, and the payment is inseparable from the moral status of the customer’s acquisition of the milk.

The implication for eminent domain is straightforward. To the extent that acquisition of property through takings is analogous to acquisition of property by sale, the acquisition will be legitimate only if the entity that acquires the property also pays compensation to the owner from whom it was acquired.[98] Justice is not satisfied by leaving the burdened owner to receive “compensation” only from third parties, whether they be motivated by charity or by the requirements of an insurance contract.

Of course, there is a crucial disanalogy between purchases and takings: takings are not voluntary exchanges, since the property’s transfer is compelled by the state. A “forced sale” is not the same as a “sale.” As will soon be clear, however, the fact that eminent domain is coercive does not alter the conclusion that the entity taking the property is the entity required to pay for it.

At an intuitive level, the fact that the state (the “buyer”) has coerced an owner into giving up property, rather than acquiring it with the owner’s permission, does not plausibly reduce the state’s obligation to compensate the burdened owner, nor does the state’s acting coercively somehow make it more appropriate for the state to require that the owner seek help from others to mitigate the loss. If anything, the opposite would seem to be true: the presence of coercion would only increase the grounds for insisting that the coercer be the party who pays compensation.

However, this intuitive argument is merely suggestive rather than conclusive because most ordinary cases of coerced transfers—e.g., thefts—have the additional feature of being wrongful. Because one person’s having wronged another naturally suggests that the first person must compensate the second, there might be a natural inclination to think that compensation for coerced transfers is required, but only for tacit reasons that are not relevant to transfers that are coerced yet not wrongful.

This possibility requires attention, because, as the next section will discuss, a signal feature of eminent domain is that the state does not commit a wrong when it exercises its power of eminent domain, even though taking possession of property without the owner’s consent is ordinarily a wrong. Since ordinary judgments about the wrongfulness of coercive transfers therefore do not apply to takings, perhaps ordinary judgments about compensation also do not apply, and for similar reasons. However, as will soon be evident, that inference does not follow. Examining the reason why takings are not wrongs in fact confirms the conclusion that the government owes compensation for property that it takes.

2. Necessity and Reasonable Accommodation.—The fact that the state does not wrong owners when it exercises its power of eminent domain is evident in how the law treats those exercises, most obviously in providing for them at all. If exercising that power were wrongful, the appropriate response would simply be to prohibit that exercise—to ban the government from taking property or require the government to return the property that it already took, not to permit the government to take property and keep it so long as it merely pays the property’s market value in return. Prohibition is in fact how the law treats certain types of takings, namely those that fail to satisfy the Constitution’s “public use” requirement.[99] Because takings purely for private use are wrongs, owners can stop them altogether.[100] By contrast, takings for public use are routinely permitted. Indeed, it has long been said that the power of eminent domain is an inherent attribute of any government’s sovereignty.[101]

Given therefore that the state does not act wrongly when it exercises its power of eminent domain, it must be the case that the state has a right to take property, at least under certain conditions. Equivalently, restated from the perspective of the taken property’s owner, it must be the case that, under certain conditions, owners have a duty to yield their property to the government that acts on that community’s behalf.[102] That observation naturally prompts a further question: Why does this particular duty exist?

Since there is no obvious reason to think that eminent domain is sui generis—a unique legal authority somehow separate from the rest of the law—consulting the rest of property law can provide some illumination. The universe of duties that community members may have toward each other is potentially broad and varied.[103] Familiar among them, however, is a general responsibility not to inflict certain types of harm on other members of the community.[104] In property law, for example, these duties are evident in laws prohibiting trespass and nuisance.[105] This duty, however, is not absolute. Thus, for example, courts will not enjoin reasonable uses of property as nuisances, nor will they deem a trespass to have occurred when someone fleeing a mortal threat seeks shelter on private property without permission.[106]

These doctrines recognize that community membership requires some accommodation of the interests of other members of the community—that is, the existence of what one might call a duty of reasonable accommodation.[107] Recognizing that duty makes it possible to restate why the government does no wrong in taking, on behalf of the community, private property needed for some beneficial public project: an owner has a duty of reasonable accommodation toward his or her community, and the community’s taking the owner’s property, when doing so is necessary for some public project, does not exceed the community’s corresponding right to receive that reasonable accommodation.

However, a qualification embedded in that explanation is important to make explicit: The absence of wrongdoing depends upon the taking’s necessity. If the project could proceed just as well without taking property against its owners’ wishes, impeding the state’s acquisition of the property would not harm the community, and therefore a duty of reasonable accommodation would not extend to yielding that property to the government.[108] Only if the taking is necessary will the duty apply.[109]

Now, in any system based more on notions of civic equality than on caste privileges, the following proposition is likely to be uncontroversial: While each property owner has certain duties to the political community to which he or she belongs, merely by virtue of membership in that community, by the same token, every other member also has duties toward the rest of that community.[110] The duties are reciprocal.[111] This proposition might not receive universal assent, of course. For example, anarchists would deny the existence of any duties at all toward a community.[112] And some people might reject the premise that civic equality does or should exist. Nevertheless, the minimal observation that civic duties do exist is a nearly inescapable premise of any account of government power and obligation, even if philosophers have long labored to identify the correct account of exactly why those duties exist.[113]

Recognizing both necessity’s essential role in justifying a taking and the reciprocity of the duty of reasonable accommodation now makes evident why the community must pay compensation for taken property.[114] While one member of a political community may have a duty not to impede something that is necessary for attaining some public end, the other members of that community equally have a duty not to impose more of a loss on that owner than is necessary to attain that end. Each has duties of reasonable accommodation toward the other, and inflicting unnecessary harm is not reasonable.[115]

Refusing to compensate owners of taken property ordinarily is not necessary to effectuate a project.[116] For example, although building a highway connecting two towns is impossible without using property located somewhere between them, it is perfectly possible both to build a highway and to pay for the property used to do so.[117] (If any proof were needed of this fact, one could simply consult any road map and observe that highways in fact do exist, despite the fact that the law has long required compensation for property taken through eminent domain.) The necessity of taking certain property for some public purpose depends on questions of geography, geology, civil engineering, and the like. Those questions are entirely distinct from the question of who pays for that property.

Thus, even when owners’ duty of reasonable accommodation compels them not to prevent necessary public use of their property, that duty ordinarily does not extend to ceding their property without compensation.[118] If the community fails to compensate the owner of the taken property, it would violate the very same duty of reasonable accommodation that had served to justify its taking the property without permission in the first place. Thus, the community’s paying compensation is a necessary condition for exercises of eminent domain not to be wrongful.[119] The fact that the state does not act wrongly when it takes property without the owner’s permission rests on an assumption that the state is paying compensation for what it takes.

Crucially here, the key assumption is not merely that someone—such as an insurance company—happened to pay the owner a sum equal to the compensation owed, thus mitigating the owner’s loss, but rather that the political community that benefited from the taking has done so.[120] The community will not have fulfilled its duty if someone else happens to step in to do what the community was obligated to do. The relationship between the party paying compensation and the party receiving it thus is essential to the justification of a taking.

3. Taxation and Takings.—An additional implication of this account is worth noting. Commentators have observed that any adequate theory of takings compensation will need to be able to explain why taxation is permissible even though uncompensated takings are not.[121] The distinction just noted between the necessity of imposing a harm and the necessity not to compensate for that harm provides the required explanation. Governments need to raise money through taxation, and doing so obviously is impossible if they compensate each taxpayer for the value of the taxes paid.[122] (The net effect of such a compensation policy would be to raise exactly zero revenue in taxes. Every dollar paid would immediately be refunded.) Thus, in the special case of taxation, the need for the government to act and the need for the government not to pay compensation for the burdens that act imposes are inseparable, and the latter necessity legitimates non-compensation.[123] However, necessity’s ability to legitimate what otherwise would be impermissible extends only to what actually is necessary, and in ordinary cases of takings only one of those two types of necessity is present—the necessity to acquire the property in question.[124] Since governments ordinarily have the capacity to pay for property that necessity compels them to take, they are not exempt from paying that compensation. Taxation of property and takings of property through eminent domain therefore are fundamentally distinct.

It should now be clear that takings’ not being wrongful despite being coercive does not justify the government’s refusing to compensate for what it takes. To the contrary, the duty that explains why the coercion inherent in eminent domain is not wrongful, combined with the reciprocal applicability of that duty, together imply the opposite: compensation in fact is required. Because of the relationship between the burdened owner and the community that benefits from imposing that burden, justice inherently requires that the community pay compensation, independent of any economic or other instrumental considerations. The “just compensation” requirement reflects the demands of relational justice.

D. Why Not Compensate for Every Cost of Legal Transitions?

One remaining question is how this relational justice account can answer a challenge, commonly posed in the literature on just compensation, to explain why the law should treat takings differently from the vast array of other legal changes that inflict losses on individual members of a community but routinely do not provide compensation.[125]

The first step in seeing how this account can meet that challenge is to note that the existence of a prima facie reason to require compensation from those who benefit by imposing burdens on others is not the same as a conclusion that, all things considered, those who benefit should pay such compensation. The latter implication will follow only if no countervailing reasons exist that would outweigh the prima facie reason.

In this case, even if one assumes that the relational justice account provides a prima facie reason to require compensation for a wide range of government-imposed burdens, not just takings, a relevant countervailing reason is immediately obvious: there is no practical way that the government could possibly compensate for every loss that its actions impose. Legal change happens every day, and changes that leave absolutely everyone better off are rare at best.[126] As a practical matter, therefore, compensating for every burden imposed by the state is simply impossible.[127] The costs of a system of universal compensation for such burdens would be prohibitively high.[128] Necessity therefore compels the state not to pay compensation for every loss that it imposes, and as noted earlier, the law has long held that necessity can legitimate what otherwise would be impermissible.[129]

As a result, the existence of reciprocal duties of reasonable accommodation does not imply that the community, acting through the government, has an all-things-considered duty to compensate for every burden that it imposes. An impossible accommodation is not reasonably demanded. However, the fact that compensating for every loss is impossible does not imply that communities need not compensate for any loss. Even if limited resources preclude doing the best thing all the time, that limitation does not justify failing to do the best thing to the extent possible. The question then is why, if some burdens must receive compensation while others necessarily do not, eminent domain falls in the former category rather than the latter. I.e., Why do takings losses have priority over other losses that follow from government action?

Although a general theory of losses from government action lies outside the scope of this Article, it is easy to identify one clear distinction among two broad categories of losses, based on the relationship between those losses and the benefit that the community receives from imposing them: some losses are incidental to the benefit, while other losses contribute to the benefit.

For example, a decision to build a multilane expressway between two cities may ultimately reduce the profitability of small businesses that are located along a smaller, pre-existing highway.[130] Such losses, although real, do not themselves create the public benefits that creating the expressway generates. Those benefits relate to the speed and ease of travel between the two cities. The fact that businesses along the formerly popular route have lost customers does not cause travel along the expressway to be any faster or easier. Thus, these businesses’ losses are merely an incidental side-effect of the new expressway’s construction.

By contrast, a law that requires every citizen to serve in the military imposes a burden—mandatory military service—that cannot be separated from the benefits of having a larger military. In this case, the burden of serving in the military actually produces the benefit of a larger military. Unlike the losses suffered by downtown businesses when a freeway bypass opens, the losses suffered by unwilling conscripts contribute to the benefits of universal military service.

All else being equal, a community’s obligation to pay compensation when it benefits by imposing a loss on one of its members intuitively seems stronger than its obligation to compensate people who suffer losses that are merely incidental side-effects of socially beneficial actions. In the latter case, the loss is regrettable, and the community would have preferred to avoid it; in the former case, the loss actually provides the benefit.

Although developing a philosophical foundation for that intuition lies outside the scope of this Article, the intuition itself has been widely held. One sign of its appeal is the enduring philosophical relevance of the “Doctrine of Double Effect.” Originally formulated in the thirteenth century, the Doctrine is an analytical tool for assessing the moral permissibility of actions that have multiple effects, some of which would ordinarily be morally impermissible. One of the Doctrine’s key criteria is whether an action’s morally problematic effects play an instrumental role in achieving the action’s goal—in which case the action is deemed morally impermissible—or instead are merely unintended side-effects, in which case the action may be morally permissible (if other criteria, not relevant here, are also satisfied).[131] Although the Doctrine is not universally accepted among philosophers, the seriousness with which it is still taken in formal moral thought is testimony to the intuitive appeal of its core distinction.[132]

In the specific context of takings, the basic implications of the distinction between incidental and contributory losses are straightforward. Exercises of eminent domain fall squarely within the realm of losses imposed in order to produce a benefit, while many other losses from legal transitions are merely incidental to the benefits that those legal changes obtain. Therefore, in a world of limited resources, losses from eminent domain would inherently have a stronger claim for compensation than losses from those other transitions would, all else being equal.[133]

That fact, however, cannot be the whole story. As a unanimous U.S. Supreme Court noted in Connolly v. Pension Benefit Guaranty Corp.:[134]

In the course of regulating commercial and other human affairs, Congress routinely creates burdens for some that directly benefit others. For example, Congress may set minimum wages, control prices, or create causes of action that did not previously exist. Given the propriety of the governmental power to regulate, it cannot be said that the Taking Clause is violated whenever legislation requires one person to use his or her assets for the benefit of another.[135]

The Court’s observation highlights the fact that a state-imposed burden’s priority for receiving compensation may depend on multiple considerations, not just on the relationship between the burdened party and the benefited party.[136] In particular in the takings context, a critical additional consideration is whether the burden in question involves interference with rights to property, as opposed to some other rights. The Fifth Amendment’s requirement that just compensation be paid applies only when “private property” has been taken. Regulatory measures such as price controls and minimum wages—examples that the Connolly court offered—do benefit some by means of burdening others, but in neither case are the imposed burdens interferences with property rights. A general interest in being able to charge more for what one sells or to pay less for what one purchases is simply not property. Thus, the fact that the government is not required to compensate those who are burdened by regulations abridging these lesser, non-property interests does not show that compensation is not owed when property is taken.

In practice, determining which rights qualify as property rights thus will be critical for determining whether a particular government-imposed burden requires compensation or not. Likewise, a fully comprehensive theoretical account of compensation requirements ideally would explain why property is or should be given the special status that it has in the law. However, creating a complete general theory of property that would resolve those issues is not necessary for present purposes. The account offered in this Article focuses only on the core case of takings—coercive acquisitions of tangible property—and it is now clear why those acquisitions are justifiably treated differently from other losses that the government may impose. What light this account might shed on more remote applications of takings principles, such as regulatory takings, is a question that can be left to another day.

III. Assessing the Economic Case

If the discussion thus far has been convincing, then the intrinsic importance of the government’s paying compensation for taken property is clear. Independent of any concerns about economic efficiency, relational justice requires that the community that benefited by taking property compensate the owner who was burdened by that taking.

So far, however, this argument has shown only that such compensation is intrinsically important. It has not yet addressed whether that importance is sufficient to justify requiring compensation even in light of the efficiency gains supposedly offered by replacing government compensation with private insurance.[137]

How to address conflicts between intrinsic concerns of justice and instrumental concerns about efficiency is a perennially thorny question.[138] Fortunately, in the present case, that question can safely be set aside because it is relevant only when such a conflict exists—that is, only when adopting the intrinsically just policy would appreciably diminish social efficiency. As will become evident, in the case of takings compensation there is no actual conflict.

Existing discussions of the anti-compensation thesis pay surprisingly little attention to the question of how much social efficiency is actually at stake. They argue that the just-compensation requirement creates a moral hazard and inflates administrative costs, but they offer no estimate of the size of those effects.[139] As this Part will explain, the supposed moral hazard is likely to arise only rarely, if ever, and any administrative cost savings would likely be minimal. Meanwhile, switching from government compensation to private insurance would actually reduce efficiency by increasing information costs and distorting incentives. Hence, any net efficiency gains from replacing government compensation with private insurance would likely be small or nonexistent, and thus would provide no reason to set aside the requirements of relational justice.

A. How Much Is at Stake?

1. How Frequently Would the Moral Hazard Arise?—As noted in Part I, one central pillar of the anti-compensation thesis is the argument that government compensation for taken property creates incentives for owners to improve their properties even when doing so is socially inefficient.[140] If this moral hazard were a significant problem, one would expect to find widespread evidence of actual costly improvements that have been made despite owners’ awareness that those improvements likely would be taken and destroyed.[141] Thus, it is striking that the anti-compensation literature offers few concrete examples drawn from actual exercises of eminent domain.[142] The scarcity of actual examples becomes unsurprising, however, once one realizes that this particular moral hazard rarely would produce inefficient results.

One reason that the moral hazard is likely to be irrelevant in practice is that the number of improvements that even could be inefficient in this way is apt to be quite small. The hazard would arise and be costly only when all three of the following are true: (1) owners decided to make expensive improvements to their property; (2) the expected value of those improvements, given the probability of a taking, was negative; and (3) the government actually did take that property, making the owners’ investments go to waste. In reality, however, many owners never improve their property beyond modest maintenance; very little property is ever taken by the government; and even owners who improved property that the government ultimately took may, by luck, have made those improvements when (unknown to those owners) the improvements’ expected value was in fact positive. Hence, the set of possibly inefficient “excess” improvements is likely to be quite small.[143]

Moreover, owners’ disregarding the probability of a taking will reduce efficiency only if the opposite course of action—i.e., attempting to consider that probability—would not have been futile. If owners are unable to assess that probability with sufficient precision and accuracy to make optimal decisions, then discouraging them from considering it will have no practical effect other than avoiding wasted effort.[144] Unfortunately, there is little reason to believe that owners ordinarily have this ability, since it is not clear how even moderately well-informed owners could assign anything other than a vague probability to the risk that their property will be taken.[145]

As a result, many scenarios in which considering the risk of a taking theoretically would change a landowner’s investment decisions never would arise in practice. For example, recall the hypothetical example described in Part I: an improvement project that costs $4,000 and will return $5,000 if the government does not take the property but zero dollars if the government does take it. The expected value of undertaking that project would then depend on the probability of the government’s taking the property. If that probability is 15%, then the project’s net expected value would be positive and society would benefit from the project, but if that probability is 25%, then the net expected value would be negative, and the project is socially inefficient.[146] The main point of the moral hazard argument is that the availability of government-provided compensation will lead the owner to ignore the risk of a taking, and thus to build the improvement irrespective of whether the probability of a taking is 25% or 15%.

But the blame for this undesirable effect cannot be placed on the availability of government compensation unless, in the absence of such compensation, the owner actually would have been sensitive to that particular difference in probabilities. And, in reality, there is little reason to expect owners to be able to estimate the risk of expropriation with sufficient precision and accuracy to distinguish situations where the chance of condemnation is 15% from situations where that chance is 25%.[147] At best, owners are likely to be able to make only very rough assessments of condemnation risk—e.g., that it is “negligible,” “low,” or “high.”[148] Hence, much of the time, the efficient decision may simply be unknown to the decision maker, and thus any moral hazard will have no practical effect.[149]

Moreover, even when the necessary calculations are possible, one might expect owners naturally to avoid building improvements that have a substantial risk of being taken, since an improver’s goal is not to break even by having the improvement taken and reimbursed, but rather to enjoy the improvement, including by making plans for its use that would come to fruition only after the time of the likely taking.[150] Presented with a substantial likelihood that any improvement would be taken, such an owner would already have good reason to avoid building the improvement there, and instead to sell the property, buy land elsewhere, and build where he could count on enjoying the value that he had purchased.[151]

The incentive simply to move and build elsewhere is especially strong if the improvement is for commercial purposes because construction takes time. An owner who builds a revenue-producing improvement, such as an office building, even though that property is likely to be taken, will have to build the improvement twice—once on the current parcel and another time on the new parcel after the first parcel is taken. Since the improvement cannot generate profit until it has been built, this second course of action would be more costly than moving immediately to another parcel and building there.[152] For example, if creating the improvement takes one year, then by building on the soon-to-be condemned parcel, the owner will miss out on two years of potential profits rather than just one—profits from the year spent building the first building, and then profits from the second year spent building a new building after the first building was taken. Since the government generally does not compensate for takings’ consequential losses, such as lost profits, the owner would have a strong economic incentive to avoid that unnecessary loss.[153]

Thus, in practice, the frequency with which the purported moral hazard would actually lead to socially inefficient investment seems close to zero.[154] Those owners who were oblivious or merely lacked unusually precise information about the probability that their property would be taken would not be able to make the optimal decision anyway (except by chance), and those who did have that information would likely prefer simply to move and build elsewhere.[155] As a result, any efficiency gains from eliminating that hazard likely would be negligible.

2. The Hope of Administrative Cost Savings.—The second pillar of the anti-compensation thesis is the argument that replacing government compensation with private insurance would reduce administrative costs, thus saving money for all taxpayers, including property owners. In fact, however, any net savings in administrative costs are likely to be meager, at best.

Hypothetically, competition among private-insurance providers might encourage diligence in eliminating administrative waste, and that encouragement might in turn lead to lower administrative costs.[156] However, there is no obvious reason to believe that the current administrative costs of eminent domain compensation are excessive. Hence, the efficiency gains that are available even in theory may be few, and if so, then any improvement caused by competition would only be minor. Outside of simple faith in the power of private markets and the irremediable inefficiency of government, there is no reason to think that the potential administrative-cost savings here are substantial at all.

One way that replacing government compensation with private insurance might nevertheless appear to reduce administrative costs is if that private alternative were to offer fewer procedural protections for owners—e.g., fewer opportunities to challenge the accuracy of the value placed on the taken property.[157] However, while removal of procedural safeguards—just like the removal of safeguards in general—might reduce the monetary costs of operating a system, it would do so only at the expense of increasing the risk of the kind of harm that the safeguard was intended to prevent (such as erroneous valuations). Since the optimal level of procedural protections cannot be known a priori, one cannot assume that any reduction of administrative costs by reducing procedural protections really would be socially efficient rather than merely a false economy.

Moreover, switching to a system of competing insurance providers would introduce other costs. One arises from a reduction in risk-spreading, the other from an increase in solvency risk.

In general, the economic motivation for insurance rests upon the benefits obtainable by spreading across many insured parties risks that would otherwise be concentrated on a few people.[158] Because risk-spreading is the key to effective insurance, the larger the pool of insured parties over which the risks are spread, the better, all else being equal.[159] If the government provides compensation for takings losses, then the relevant pool is effectively all taxpayers—in other words, almost everyone. But if private insurance is used instead, then each competing insurance company will necessarily have a markedly smaller pool over which to spread its risks and thus will be able to provide insurance only at a higher cost.[160]

A second, related cost springs from the fact that private-insurance companies face solvency risk—the possibility that, through misfortune or mismanagement, a company may not be able to pay all of the claims that it receives. Even when that risk is low, it will inevitably still exceed the risk of the government’s becoming insolvent and unable to pay its debts, because only the government has the power to raise money through taxation.[161] Thus, switching from government compensation to private insurance imposes additional risk on property owners, and that risk in turn gives property owners reason to monitor the solvency of the firms that insure them against takings losses. Such monitoring, when possible at all, requires an additional investment of resources.

How large these additional costs would be relative to the supposed reduction in costs from competitive pressures or attenuated procedures is an empirical question. Unfortunately, answering that question would be difficult or impossible, since doing so would require measuring the costs of an alternative compensation system that does not exist. What can be said is that even if switching to private insurance did reduce some administrative costs, there is no reason to think that this reduction would be large, and since such a switch would simultaneously increase other costs, the change’s net economic effect is unlikely to be significantly positive and might even be negative.

B. Additional Costs of Private Insurance

Switching to private insurance also would incur two additional sets of costs that the anti-compensation argument overlooks. One set involves information costs, and the other involves distorted incentives.[162]

1. Information Costs.—Eliminating government compensation for taken property would require that owners who are considering improving their property, or companies who would insure those improvements against takings, estimate the probability that the government will take that improvement before it exhausts its useful life. That calculation is, of course, what the anti-compensation thesis relies upon to avoid the moral hazard that government compensation allegedly creates.[163] As noted earlier, owners likely will often be unable to obtain these estimates at all, at least with the accuracy and precision needed to make those estimates.[164]

However, even if those estimates were feasible, acquiring the information needed to make them would be costly. At best, obtaining the relevant publicly available information would require constant monitoring and assessment of the public statements of government officials, agencies, and legislatures. And, quite likely, an accurate estimate would require access to a wide range of data that are not readily available to private entities at all, such as the content of government deliberations and the government’s internal estimates of the social value of various public project alternatives. Insurance companies might be able to acquire that information, either licitly or illicitly, despite these obstacles, but only at additional expense.

Compounding the problem, the probability of a taking might change over time. Thus, owners who undertook projects requiring significant time to complete could not rely on just one estimate, made at the project’s outset, of the probability of a taking but rather would need continually to reassess the probability of a taking, in order to determine whether they should continue the project or instead abandon it.

Of course, removing the government’s compensation obligation would decrease the government’s own information costs because the government would no longer need to determine the precise market value of any taken improvements in order to pay the correct amount in compensation. However, this reduction in information costs would occur only in those few cases in which the government actually takes property—or at least seriously plans to take property—while owners’ information costs would increase every time owners consider making an improvement to their property. In any given year, the number of property improvements undertaken or contemplated presumably dwarfs the number of improvements that the government takes.[165] Thus, the net effect of removing the government’s compensation obligation likely would be an increase in total information costs.

2. Private Companies and Fiscal Incentives.—Additional social costs may arise from the creation of inefficient incentives. If governments respond to political pressure, then replacing government compensation with private insurance might result in the government’s taking property less often than it should.[166] Since every profit-oriented business has an incentive to reduce its expenses, private companies who provide insurance against takings would have a strong incentive to lobby governments to reduce the use of eminent domain.[167] Insurance companies are exactly the sort of concentrated interest endowed with ample economic resources that are commonly thought to have considerable political influence.[168] Hence, this self-interested lobbying might be especially effective in shaping governments’ decisions about which projects to undertake or to forego.[169]

Moreover, private companies often directly benefit from the exercise of eminent domain, and their decisions about property obviously are not made independently of fiscal considerations, even if the government’s decisions are. (Indeed, the responsiveness of private actors to fiscal considerations is the very foundation of the “moral hazard” argument.) Historically, one prominent use of eminent domain was to enable large private businesses to accumulate land that was needed for their operations. Thus, in the nineteenth century, states often delegated their eminent domain power to railroads and canal companies to use in building transportation networks.[170] Similarly, in the twentieth and twenty-first centuries, states have delegated their eminent domain power to public utilities.[171] And the U.S. Supreme Court’s landmark decision in Kelo v. New London[172] upheld the constitutionality of using eminent domain to transfer property to private businesses as part of a government plan to promote economic development.[173]

In all of these cases, the companies that benefited from the exercise of eminent domain were, of course, obligated to pay compensation for the property that they obtained by using this delegated power.[174] However, if the government’s obligation to pay compensation for taken property were removed, leaving burdened owners to rely on private insurance to alleviate their losses, then socially inefficient ventures would become much more attractive to companies that did not have to pay for the property required to undertake them.[175] Growing businesses that have been delegated the power of eminent domain would then have a powerful incentive to exercise that power to acquire property for ventures that otherwise would not have been sufficiently profitable to be worth undertaking.[176] As a result, removing the obligation to compensate for taken property would predictably increase the frequency of inefficient land-use decisions by those private businesses.[177]

Furthermore, businesses would also have strong incentives to lobby governments to undertake “economic development” projects—even if not socially optimal—in which those businesses could participate, thereby benefiting from the opportunity to acquire property through eminent domain at no expense. Of course, since the incentive effects of pro-taking lobbying by private beneficiaries of eminent domain and anti-taking lobbying by insurance companies pull in opposite directions, these two distortions of incentives might conceivably cancel each other out. However, such a result would merely be a lucky coincidence. There is no obvious reason to think that those distortions would in fact negate each other, much less that they would reliably do so over time. And even if, by chance, they did happen to cancel out, the resources spent in lobbying to reach that stalemate would still be a direct social cost.[178]

IV. Conclusion

The argument in favor of replacing government compensation for taken property with private insurance thus rests on an illusory economic foundation. Considering all of the effects of such a change, and—crucially—the likely size of each effect, shows that any efficiency gains from eliminating government compensation for takings would most likely be negligible at best, and might even be negative. But the anti-compensation thesis has an even more fundamental flaw: it misunderstands the basic nature of takings compensation. “Just” compensation is not reducible to the instrumental promotion of other goods, such as economically efficient maximization of social wealth. Instead, the government’s compensating for property that it takes has an intrinsic value that is inextricably tied to the legitimacy of the power of eminent domain, a connection that springs from the particular relationships that exist between property owners and the rest of the community. Requiring the government to pay compensation for what it takes is thus fundamentally different from the provision of insurance against takings losses, and attempting to substitute private insurance for government compensation would neither improve social efficiency nor satisfy the requirements of relational justice.

  1. .U.S. Const. amend. V. For examples of state constitution provisions, see Cal. Const. art. 1, § 19, cl. a (“Private property may be taken or damaged for public use only when just compensation, ascertained by a jury unless waived, has first been paid to, or into court for, the owner.”); Mass. Const. art. 10 (“And whenever the public exigencies require that the property of any individual should be appropriated to public uses, he shall receive a reasonable compensation therefor.”); N.Y. Const. art. 1, § 7, cl. a (“Private property shall not be taken for public use without just compensation.”).
  2. .See, e.g., Monongahela Navigation Co. v. United States, 148 U.S. 312, 326 (1893) (reading the Fifth Amendment as “a declaration, that no private property shall be appropriated to public uses unless a full and exact equivalent for it be returned to the owner”); Olson v. United States, 292 U.S. 246, 254–55 (1934) (endorsing the Monongahela Navigation Co. court’s statement).
  3. .See infra Part I.
  4. .See, e.g., Steven Shavell, Foundations of Economic Analysis of Law 127–34 (2004) (discussing arguments for and against governmental compensation and concluding that the argument in favor of compensation “is significantly qualified . . . by questions concerning whether the actual incentives of the state to take are excessive and by related issues”); see also Daryl J. Levinson, Making Government Pay: Markets, Politics, and the Allocation of Constitutional Costs, 67 U. Chi. L. Rev. 345, 393 (2000) (“From an efficiency perspective, a rule of no compensation seems at least as plausible as a rule of full or partial government compensation.”). For examples of casebook discussions, see, for example, Jesse Dukeminier et al., Property 1064 (7th ed. 2010) (noting the debate about whether private insurance is a preferable alternative to public compensation and commenting that it, “so far as we know, remains unresolved”); Robert C. Ellickson & Vicki L. Been, Land Use Controls: Cases and Materials 175–78 (2d ed. 2000) (excerpting a seminal article on the topic and discussing the issue more generally); Thomas W. Merrill & Henry E. Smith, Property: Principles and Policies 1169 (3d ed. 2017) (noting literature identifying potential obstacles to the availability of private insurance against takings losses but commenting that “[i]n theory . . . if these problems could somehow be overcome, private insurance would do the trick just fine”); Richard A. Posner, Economic Analysis of Law 73 (8th ed. 2011) (discussing the question: “Why not just let property owners insure the market value of their property against the risk of its being taken by eminent domain?”).
  5. .If the taken property is owned by a foreign national, expropriating it without paying compensation might also violate international law. See, e.g., G.C. Christie, What Constitutes a Taking of Property Under International Law?, 38 Brit. Y.B. Int’l L. 307, 307 (1962) (noting that there exists a “widely recognized rule of international law that the property of aliens cannot normally be taken, whether for public purposes or not, without adequate compensation”).
  6. .For a survey of the key arguments and issues, see generally David A. Dana & Thomas W. Merrill, Property: Takings 58–168 (2002).
  7. .See Thomas W. Merrill, The Landscape of Constitutional Property, 86 Va. L. Rev. 885, 891 & n.17 (2000) (noting that “there is an enormous literature about what it means to ‘take’ property” and collecting sources).
  8. .For a prominent recent dispute over whether a particular physical interference constitutes a taking, see Ark. Game & Fish Comm’n v. United States, 133 S. Ct. 511, 515 (2012) (holding that repeated temporary flooding of private property is “not categorically exempt from Takings Clause liability”). For an overview of regulatory takings, see Dana & Merrill, supra note , at 121–68.
  9. .See Merrill, supra note 7, at 889–90 (discussing four inconsistent Supreme Court cases and noting that “not one of the four decisions makes any reference to any of the others, or makes any effort to integrate its innovations . . . into the preexisting fabric of the law”).
  10. .For a description of the evolution of the meaning of “property,” see 2 Nichols on Eminent Domain § 5.08 (2016). See also Maureen E. Brady, Property’s Ceiling: State Courts and the Expansion of Takings Clause Property, 102 Va. L. Rev. 1167, 1174 (2016) (describing courts’ expansion of the definition of “property” to provide a right of compensation in “street grade” cases).
  11. .See United States v. Miller, 317 U.S. 369, 374 (1943) (determining that a property owner is entitled to the “fair market value” of the property taken).
  12. .See Nicole Stelle Garnett, The Neglected Political Economy of Eminent Domain, 105 Mich. L. Rev. 101, 121 (2006) (noting that, as a result of various statutory provisions, “[f]requently, owners are legally entitled to substantially more than the fair market value of their property”).
  13. .Lawrence Blume & Daniel L. Rubinfeld, Compensation for Takings: An Economic Analysis, 72 Calif. L. Rev. 569, 571 (1984).
  14. .See, e.g., Louis Kaplow, An Economic Analysis of Legal Transitions, 99 Harv. L. Rev. 509, 534–35 (1986) (“A private actor should be indifferent as to whether a given probability of loss will result from the action of competitors, an act of government, or an act of God, except to the extent that the source of the risk will affect the likelihood of compensation or other relief.”).
  15. .See, e.g., Shavell, supra note 4, at 128 (noting that “through payment of higher taxes to finance compensation for takings, individuals must implicitly pay exactly the premium they would be charged for private insurance coverage against takings”).
  16. .See, e.g., id. (“[O]ther things being equal, there is an equivalence between the state paying compensation in the event of takings and individuals purchasing insurance coverage against uncompensated takings.”); Blume & Rubinfeld, supra note , at 590–92 (arguing that “compensation, provided ex post, acts as a rudimentary form of insurance”).
  17. .See, e.g., Abraham Bell & Gideon Parchomovsky, The Hidden Function of Takings Compensation, 96 Va. L. Rev. 1673, 1711 (2010) (“In essence, by paying compensation to owners any time it takes their property, the government grants all property owners a publicly provided insurance policy against the risk of future takings of their property.”); Daniel A. Farber, Public Choice and Just Compensation, 9 Const. Comment. 279, 282 (1992) (“In effect, [the citizen] is buying insurance against a taking . . . .”).
  18. .See, e.g., Lee Anne Fennell, Unbundling Risk, 60 Duke L.J. 1285, 1316 (2011) (“Transition relief—such as grandfathering, recognition of vested rights, or compensation for governmental takings—amounts to embedded insurance against legal change.”).
  19. .See, e.g., Kaplow, supra note , at 563 (asserting that determining “what constitutes a taking for constitutional purposes . . . has long been recognized as a central difficulty in takings doctrine, which attempts to limit the constitutional requirement of just compensation to a small subset of the diminutions in value that can result from government action” (footnote omitted)).
  20. .See, e.g., id. at 534 (asserting that “none of the distinctions [that commentators] offer for treating government and market risks differently withstands scrutiny”).
  21. .Two articles were seminal in creating the canonical case: Lawrence Blume, Daniel L. Rubinfeld & Perry Shapiro, The Taking of Land: When Should Compensation Be Paid?, 99 Q.J. Econ. 71 (1984); and Louis Kaplow, An Economic Analysis of Legal Transitions, 99 Harv. L. Rev. 509 (1986).
  22. .See Thomas J. Miceli, The Economic Theory of Eminent Domain: Private Property, Public Use 86 (2011) (noting that “[f]rom a purely economic perspective” the anti-compensation argument “is not a particularly surprising one considering that it is a simple application of well-known results from the economics of insurance,” specifically moral hazard). For a succinct explanation of moral hazard as a general problem in insurance, see Robert Cooter & Thomas Ulen, Law & Economics 238 (6th ed. 2012).
  23. .See Blume, Rubinfeld & Shapiro, supra note 21, at 81 (arguing that private investors “do[] not take into account the loss to society if the project is undertaken and the capital invested in . . . land is lost”); Kaplow, supra note 14, at 539–40 (arguing that full compensation for takings leads to overinvestment by landowners). Blume, Rubinfeld, and Shapiro’s paper appeared a decade after William Baxter and Lillian Altree had published a related argument, in a somewhat different context, about distorted incentives for improving property when conflicts between land uses might arise. William F. Baxter & Lillian R. Altree, Legal Aspects of Airport Noise, 15 J.L. & Econ. 1, 5–6 (1972). For a succinct technical exposition of the moral hazard concern, see Daniel Klerman, Comment, Takings, Fiscal Illusion, and the Median Voter, 173 J. Institutional & Theoretical Econ. 71, 72–73 (2017).
  24. .Strictly speaking, she will do so only if the payoff is sufficiently above $4,000 that no alternative opportunity would offer her an even greater payoff for her $4,000 investment. That minor complication does not affect the substance of the example.
  25. .Again, speaking strictly, the example supposes that $5,000 is the present value of all future payoffs from the investment. For a succinct discussion of present-value calculations, see Cooter & Ulen, supra note , at 37.
  26. .The expected value of an event with an uncertain outcome is determined by multiplying the probability of each possible outcome by the payoff that would result if that outcome becomes actual, and then summing all of those products. Cooter & Ulen, supra note , at 43. Thus, in this example, the expected value of building the improvement would be (0.5)($5,000) + (0.5)($0) = $2,500.
  27. .See Shavell, supra note 4, at 131 (arguing that full compensation may “excessive[ly] incentivize [individuals] to invest in improving their property” and lead to “socially undesirable” results); Blume, Rubinfeld & Shapiro, supra note 21, at 81, 88 (arguing that full compensation leaves land owners with little incentive to account for the risk of a taking and that in some instances no compensation may be more desirable).
  28. .See Kaplow, supra note 14, at 541 (arguing that “[f]ull compensation is . . . undesirable whenever the market would not have provided full protection”).
  29. .For a brief discussion of how insurance firms determine how much to charge for insurance, see Robert W. Klein, A Regulator’s Introduction to the Insurance Industry 7 (2d ed. 2005).
  30. .In the hypothetical example above, requiring the owner to purchase private insurance for protection against takings losses would add the cost of insurance premiums to the contemplated improvement’s $4,000 cost of construction, and that extra cost would increase as the chance of the property’s being taken increased. If there is a 50% chance of the government’s taking the property, then an insurance company would charge at least $2,500 to insure against the loss of an improvement worth $5,000. See id. (explaining that individuals with higher risks pay higher insurance premiums). As a result, the owner’s total cost of making the improvement would rise to $6,500 ($4,000 in construction expenses plus $2,500 in insurance expenses), and the owner’s personal net expected payoff from building the improvement would consequently become negative (–$1,500), matching the negative net social payoff of making the improvement. An economically rational owner therefore would refrain from making the improvement, and consequently would end up making the socially efficient choice.
  31. .See, e.g., Farber, supra note 17, at 282 (noting that owners protected by government compensation will, as a result, have to pay higher taxes that “[o]n average . . . just balance the possible expectation of compensation”).
  32. .See, e.g., id. at 283 (“Functionally . . . the taxes are equivalent to insurance premiums for risk-averse taxpayers.”).
  33. .See, e.g., Shavell, supra note , at 128 (“[T]hrough payment of higher taxes to finance compensation for takings, individuals must implicitly pay exactly the premium they would be charged for private insurance coverage against takings.”).
  34. .See, e.g., id. at 129 (observing that government-based compensation systems might incur higher administrative costs than private-insurance systems in determining amounts of takings compensation); Steve P. Calandrillo, Eminent Domain Economics: Should “Just Compensation” Be Abolished, and Would “Takings Insurance” Work Instead?, 64 Ohio St. L.J. 451, 507 (2003) (reasoning that “the overt motives of profit-maximization and the pressures of market competition” minimize administrative costs for private insurance firms). Kaplow endorsed the idea that private insurance might have lower administrative costs, but he acknowledged that the infrequency of the government’s taking property complicates the picture. Kaplow, supra note 14, at 547.
  35. .In economic language, the argument is that replacing government compensation with private insurance would be Pareto superior. For a discussion of Pareto superiority and its significance, see Jules L. Coleman, The Grounds of Welfare, 112 Yale L.J. 1511, 1515–20 (2003) (reviewing Louis Kaplow & Steven Shavell, Fairness Versus Welfare (2002)).
  36. .The specific terms used and the exact groupings can vary. In particular, some discussions treat “public choice” arguments as distinct from “efficiency” arguments. Nevertheless, the basic catalog of standard arguments is widely established. See, e.g., Abraham Bell & Gideon Parchomovsky, Taking Compensation Private, 59 Stan. L. Rev. 871, 877–85 (2007) (articulating fairness-based, efficiency-based, and political-based justifications); Michael H. Schill, Intergovernmental Takings and Just Compensation: A Question of Federalism, 137 U. Pa. L. Rev. 829, 841–65 (1989) (addressing fairness and efficiency arguments for compensation); Christopher Serkin, Passive Takings: The State’s Affirmative Duty to Protect Property, 113 Mich. L. Rev. 345, 360–71 (2014) (addressing utilitarian-based and fairness-based takings justifications).
  37. .See, e.g., Dana & Merrill, supra note , at 41–46 (discussing the fiscal illusion argument and concluding that “[i]f we do not require the government to pay compensation for takings . . . government officials may suffer from the ‘fiscal illusion’ that the resources they take have no opportunity cost”).
  38. .See Cooter & Ulen, supra note 22, at 181 (suggesting that “the compensability of takings” acts as a check upon overregulation by forcing the government to “internalize” the costs of its takings).
  39. .See, e.g., Shavell, supra note 4, at 130 (“[T]he individuals who make decisions whether or not to take property may themselves not be much affected by the state’s compensatory disbursements.”); Kaplow, supra note 14, at 569 (noting that “requiring compensation would not necessarily counteract fiscal illusion” because of “political complications . . . and the frequent division of responsibility between those who decide issues of taxation and those who make other decisions, such as choices of government projects”).
  40. .See generally Christopher Serkin, Big Differences for Small Governments: Local Governments and the Takings Clause, 81 N.Y.U. L. Rev. 1624 (2006) (arguing that local governments are sensitive to fiscal incentives in takings decisions, even if those incentives have less effect at the state and federal levels).
  41. .See Ronit Levine-Schnur & Gideon Parchomovsky, Is the Government Fiscally Blind? An Empirical Examination of the Effect of the Compensation Requirement on Eminent-Domain Exercises, 45 J. Legal Stud. 437, 463 (2016) (offering empirical support for the idea that political, rather than fiscal, concerns motivate government officials). Lawrence Rosenthal has offered an argument that bridges the gap between these two competing accounts. Rosenthal argues that because fiscal constraints limit government officials’ ability to direct resources in politically advantageous ways, government officials’ sensitivity to political incentives does not imply an insensitivity to fiscal incentives, but rather the opposite. Lawrence Rosenthal, A Theory of Governmental Damages Liability: Torts, Constitutional Torts, and Takings, 9 U. Pa. J. Const. L. 797, 832–37 (2007).
  42. .See Schill, supra note 36, at 861 (noting that just compensation reduces the incentive for politically powerful groups to use uncompensated takings as a means to effect self-interested ends).
  43. .See id. (“Requiring the federal government to compensate property owners when it takes their property reduces the incentives for this type of ‘rent-seeking,’ by spreading the costs of such behavior to all citizens, including those in power.” (footnote omitted)).
  44. .See Dana & Merrill, supra note 6, at 46–47 (explaining the “buyoff” theory and noting that “compensation [may] function[] as an important element in overcoming the opposition of intense minorities to projects that are in the interest of the diffuse majority”); Farber, supra note 17, at 290 (“The effect of the compensation requirement is to buy off the landowners and shift the cost of the [government] project to other groups.”).
  45. .For a general description of moral hazard, see Cooter & Ulen, supra note , at 238.
  46. .Subpart II(D) infra will return to this general question, showing how the account offered in this Article can answer it.
  47. .U.S. Const. amend. V.
  48. .364 U.S. 40 (1960).
  49. .Id. at 49. For a history of “the Armstrong principle” and its widespread influence, see William Michael Treanor, The Armstrong Principle, the Narratives of Takings, and Compensation Statutes, 38 Wm. & Mary L. Rev. 1151, 1153–54 & nn.17–22 (1997).
  50. .Thomas Miceli notes that when the Blume, Rubinfeld, and Shapiro article arguing for the anti-compensation thesis was published, it “caused something of a stir, principally because it seemed grossly unfair and also flew in the face of the constitutional requirement of just compensation.” Miceli, supra note , at 86.
  51. .Even many law-and-economics scholars acknowledge the relevance of at least some concerns other than efficiency, commonly including the distribution of wealth in society. See, e.g., Louis Kaplow & Steven Shavell, Why the Legal System Is Less Efficient than the Income Tax in Redistributing Income, 23 J. Legal Stud. 667, 674–75 (1994) (acknowledging concerns about wealth distribution but arguing that they are best addressed not by changing the allocation of legal rights and duties in specific corners of the law, but instead by increasing taxes on the wealthy and government spending on the poor). For a recent textbook statement of a similar view, see Cooter & Ulen, supra note , at 8. See also A. Mitchell Polinsky, An Introduction to Law and Economics 153–62 (4th ed. 2011) (asserting that while “the efficiency analysis should be of principal importance,” other factors like the redistribution of wealth are relevant). For criticisms of these economic approaches to distributive concerns, see, for example, Hanoch Dagan, Takings and Distributive Justice, 85 Va. L. Rev. 741, 761, 789 (1999) (positing that distributive considerations should play a prominent role in determining the amount of compensation paid for each taking). See also Cooter & Ulen, supra note , at 4 (describing economic analyses of law as including attention to “how laws affect the distribution of income across classes and groups”); Blume & Rubinfeld, supra note , at 606–07 (suggesting that only the poor—as the “most risk averse”—should receive takings compensation because the rich can afford to purchase private insurance); Lee Anne Fennell & Richard H. McAdams, The Distributive Deficit in Law and Economics, 100 Minn. L. Rev. 1051, 1056 (2016) (arguing, inter alia, that adjusting legal rules is “not axiomatically inferior to tax-and-transfer as a means of achieving or maintaining desired distributive results”).
  52. .Frank I. Michelman, Property, Utility, and Fairness: Comments on the Ethical Foundations of “Just Compensation” Law, 80 Harv. L. Rev. 1165, 1214 (1967). Michelman thought that the utilitarian theory implied a further assumption that the intensity of those negative feelings was increased by recognition that takings losses are deliberately inflicted by the government. Id. at 1216. Dana and Merrill have criticized that assumption on the grounds that takings losses also are for the public good and therefore may be less demoralizing than losses that seem to serve no purpose. Dana & Merrill, supra note , at 39. But see James Geoffrey Durham, Efficient Just Compensation as a Limit on Eminent Domain, 69 Minn. L. Rev. 1277, 1307 (1985) (arguing that because “[e]minent domain is one of the few situations in which citizens see the direct consequences of governmental actions that negatively affect them,” demoralization costs will increase if the project for which the taken property is used “is not to the owner’s liking”).
  53. .Michelman, supra note , at 1215.
  54. .See id. at 1213 (“[W]e must remember that the utilitarian’s solicitude for security is instrumental and subordinate to his goal of maximizing the output of satisfactions. Security of expectation is cherished, not for its own sake, but only as a shield for morale.”). Nestor Davidson has argued that arguments based on “demoralization costs” caused by the risk of legal change should also recognize that the possibility of legal change might also have some positive psychological effects on property owners. Thus, the net psychological effect of the possibility of legal change may be less than often assumed. See Nestor M. Davidson, Property’s Morale, 110 Mich. L. Rev. 437, 477 (2011) (“Legal transitions can support [positive psychological effects] as easily as they can signal instability.”).
  55. .Not everyone would agree. William Fischel and Perry Shapiro argue that insurance would have no effect on demoralization costs because “[a]n insurance payment simply spreads this loss over time. . . .” William A. Fischel & Perry Shapiro, Takings, Insurance, and Michelman: Comments on Economic Interpretations of “Just Compensation” Law, 17 J. Legal Stud. 269, 287 (1988). Hence, even with insurance, property owners would still experience the same losses and thus would suffer the same demoralization. Id. at 286–87. This argument seems incomplete. Insurance does not spread losses merely over time but also over a broad group of people, most of whom will not themselves directly experience a loss. (For example, most houses never catch fire, even though homeowners routinely carry fire insurance.) Thus, the payment received from insurance in the event of a loss is ordinarily greater than the premiums that the recipient had paid to purchase insurance. That difference, in fact, is a principal motivation for buying insurance in the first place, and it is also the source of insurance’s “moral hazard” problem. (There would be no moral hazard if insurance had no effect on the experienced costs of damaging events.) Thus, even if insurance did not eliminate the burdens of takings losses, it ordinarily would sharply reduce them, and thus would predictably reduce demoralization costs suffered by those whose property was taken. Of course, everyone who pays premiums for that insurance might feel some “demoralization” from the need to make those payments, and those feelings would need to be included in any calculation of insurance’s net effect on demoralization. Whether the total amount of demoralization caused by everyone’s paying for insurance would exceed the total demoralization caused by uninsured losses is an empirical question that a priori theorizing cannot definitively resolve. However, since most people seem to be risk averse, it seems likely that insurance would tend to reduce overall demoralization costs. See, e.g., Daniel Kahneman, Thinking, Fast and Slow 318 (2011) (“People are willing to pay much more for insurance than expected value—which is how insurance companies cover their costs and make their profits. Here again, people buy more than protection against an unlikely disaster; they eliminate a worry and purchase peace of mind.”). Moreover, there is empirical evidence that risk aversion varies across demographic groups, such as age and gender. See J. François Outreville, Risk Aversion, Risk Behavior, and Demand for Insurance: A Survey, 37 J. Ins. Issues 158, 166–70 (2014) (supporting the conclusion that “[c]haracteristics such as gender, age, race, and religion clearly affect one’s level of risk aversion”). Hence, the amount of demoralization from a lack of insurance may also vary across those groups, raising the possibility that even if insurance were to have no effect on the total amount of demoralization in society, it still might affect the distribution of that demoralization and thus be desirable on distributional grounds.
  56. .Cf. Dana & Merrill, supra note , at 38–39 (arguing that Michelman’s account needs to explain why takings should receive government compensation when protection against many disappointing events is left to private insurance).
  57. .The literature on utilitarianism is now vast. For a classic discussion, see generally J.J.C. Smart & Bernard Williams, Utilitarianism: For and Against (1973).
  58. .Michelman, supra note , at 1224.
  59. .Although Michelman’s discussion of the utilitarian argument for compensation is the best-known part of his 1967 paper, the second half of that paper sought to avoid the weaknesses of the utilitarian account by offering an alternative account based on several philosophical papers by John Rawls. Id., at 1219–45. Space does not permit an analysis of this alternative account, other than to note that it focuses on the distribution of wealth and thus fundamentally differs from the relational-justice account that will be developed in Part II.
  60. .Armstrong v. United States, 364 U.S. 40, 49 (1960).
  61. .Some courts, even before Armstrong, had expressed a similar idea. See, e.g., Monongahela Navigation Co. v. United States, 148 U.S. 312, 325 (1893):

    And in this there is a natural equity which . . . prevents the public from loading upon one individual more than his just share of the burdens of government, and says that when he surrenders to the public something more and different from that which is exacted from other members of the public, a full and just equivalent shall be returned to him.

    The idea also has historical antecedents in scholarly commentary. In a prize-winning student essay, James Bradley Thayer argued:

    [P]roperty which is taken [by eminent domain], is something . . . above or aside from the owner’s regular share in the common expenses; . . . but no man can be supposed to have agreed to bear more than his share of the common burdens; there is no principle upon which such an unequal distribution could be based; and therefore, if a man’s property be taken by the State, he is entitled to have this, his excessive contribution, made up to him by compensation.

    J.B. Thayer, Essay, The Right of Eminent Domain, 19 Monthly L. Rep. 241, 250 (1856) (footnotes omitted).

  62. .For an overview of the property tax system, see generally Ronald C. Fisher, What Policy Makers Should Know About Property Taxes, Land Lines, Jan. 2009, at 8. Larissa Katz has argued that, in fact, property is “an office through which the state assigns burdens” to property owners. Larissa Katz, Governing Through Owners: How and Why Formal Private Property Rights Enhance State Power, 160 U. Pa. L. Rev. 2029, 2034 (2012). If Katz’s argument is correct, it would follow that this sort of unequal treatment is an inherent feature of property ownership and thus presumably would not automatically justify demands for compensation.
  63. .A related open question is whether the proper perspective from which to assess the equality of treatment is ex ante or ex post. Although ex post the treatment of the owner of taken property differs from the treatment of owners whose property was not taken, ex ante they were all subject to the same risk that their property might be taken, and thus they were treated equally at that point. The losses that followed, then, were merely the actualization of those equally distributed risks. The equal-treatment principle by itself does not settle which of these two perspectives is the morally appropriate one for purposes of evaluating the equality of treatment.
  64. .Thomas W. Merrill, Rent Seeking and the Compensation Principle, 80 Nw. U. L. Rev. 1561, 1579–80 (1986) (reviewing Richard A. Epstein, Takings: Private Property and the Power of Eminent Domain (1985)).
  65. .Subpart II(D) infra will discuss that argument in more detail.
  66. .Some commentators identify a third category of justification for requiring the government to pay compensation, one framed in terms of concerns about flaws in the process by which governments decide whether to take property. For example, Saul Levmore has offered a prominent account of takings compensation based on concerns about “singling out” individual property owners. Saul Levmore, Takings, Torts, and Special Interests, 77 Va. L. Rev 1333, 1344–48 (1991) (restating the account and emphasizing the phrase “singling out”). Levmore argues that takings compensation is required to protect small, politically uninfluential groups, such as individual property owners, from potential exploitation by a politically powerful majority. Saul Levmore, Just Compensation and Just Politics, 22 Conn. L. Rev. 285, 291–93, 305–11 (1990). Although at first glance Levmore’s account focuses on concerns about failures of the political process, the account is not fundamentally different from the two types of accounts already discussed, because the reasons that Levmore offers for concern about these failures ultimately rest on the risk of inefficient takings or of unequal treatment. Id. at 308–11. Thus, Levmore’s account amounts to a variant of those two theories, and rejoinders that the anti-compensation argument could offer to each of them would apply to Levmore’s account as well. See Kaplow, supra note , at 605 (“Concerns about abuse of power are potentially far more important in the context of takings than in most other transition contexts precisely because takings often single out individuals or groups. Again, the increased likelihood of private insurance in the absence of a compensation requirement . . . would do much to alleviate this problem.”); Levinson, supra note , at 394–95 (discussing and criticizing the “singling out” account).
  67. .See, for example, Lionel Robbins’s classic definition: “Economics is the science which studies human behaviour as a relationship between ends and scarce means which have alternative uses.” Lionel Robbins, An Essay on the Nature and Significance of Economic Science 15 (1932); see also Max Weber, Economy and Society 63 (Guenter Roth & Claus Wittich eds., 1978) (“‘Rational economic action’ requires instrumental rationality . . . .”). Strictly speaking, Robbins’s definition is of economics as understood in neoclassical theory. See Daniel M. Hausman, Introduction (arguing that Robbins is “attempting to define economics as neoclassical theory”), in The Philosophy of Economics: An Anthology 1, 38 (Daniel M. Hausman ed., 2d ed. 1994). However, since the economics that inspires law-and-economics scholarship and the anti-compensation arguments at issue here is predominantly neoclassical, that limitation is immaterial for present purposes. See, e.g., Christine Jolls, Cass R. Sunstein & Richard Thaler, A Behavioral Approach to Law and Economics, 50 Stan. L. Rev. 1471, 1545 (1998) (“Traditional law and economics is largely based on the standard assumptions of neoclassical economics.”); see also Herbert Hovenkamp, Rationality in Law and Economics, 60 Geo. Wash. L. Rev. 293, 293 (1992) (noting that law-and-economics scholarship is built upon a foundation in which “[e]conomists model and attempt to predict individual behavior on the assumption that people act rationally with the goal of maximizing their expected utility”); Ejan Mackaay, History of Law and Economics (describing the neoclassical influences on the seminal first decades of law-and-economics scholarship but also noting the emergence of competing approaches in later years), in Encyclopedia of Law and Economics Volume I: The History and Methodology of Law and Economics 65, 65–67 (Boudewijn Bouckaert & Gerrit De Geest eds., 2000).
  68. .For a brief discussion of the early roots, stretching back to Plato’s Republic, of the idea that justice is valuable for its own sake, see 1 Terence Irwin, The Development of Ethics: A Historical and Critical Study 100–01 (2007).
  69. .See, e.g., Cooter & Ulen, supra note , at 4 (describing economic analyses of law as being concerned with efficiency—“a comprehensive measure of public benefits”—and “how laws affect the distribution of income across classes and groups”). Cooter and Ulen note that some scholars argue that efficiency alone is the proper focus of economic analyses of law, thus setting distributional concerns entirely aside. Id.
  70. .See supra section I(B)(2).
  71. .The same can be said of the common claim that property as an institution is inherently “relational,” in the sense of the textbook principle that “[f]or lawyers, if not lay people, property is an abstraction. It refers not to things, material or otherwise, but to rights or relationships among people with respect to things.” Dukeminier et al., supra note , at 83. The type of relationship that this sort of assertion typically has in mind is purely formal, inspired by Hohfeld’s classic discussion of the relationship between various correlated legal statuses, as when one person’s right against another (e.g., a right to exclude) corresponds to that second person’s duty toward the first (e.g., a duty not to trespass). See, e.g., Stephen R. Munzer, A Theory of Property 17 (1990) (“If the law views property as relations among persons with respect to things, which relations are involved? A start on an answer lies in Hohfeld’s Fundamental Legal Conceptions.”). For Hohfeld’s analysis, see generally Wesley Newcomb Hohfeld, Some Fundamental Legal Conceptions as Applied in Judicial Reasoning, 23 Yale L.J. 16 (1913).
  72. .For an example of this common use of the term “distributive justice,” see Michael J. Sandel, Justice: What’s the Right Thing to Do? 192 (2009) (“When we discuss distributive justice these days, we are concerned mainly with the distribution of income, wealth, and opportunities.”).
  73. .Thomas Aquinas, Commentary on the Nicomachean Ethics 400–03 (C.I. Litzinger trans., 1964) (commenting on Book V, Chapter 4 of Aristotle’s Nicomachean Ethics). Adam Smith used a similar definition. Adam Smith, The Theory of Moral Sentiments 318 (Knud Haakonssen ed., Cambridge Univ. Press 2002) (1759) (“The first sense of the word [‘justice’] coincides with what Aristotle and the Schoolmen call commutative justice . . . which consists in abstaining from what is another’s, and in doing voluntarily whatever we can with propriety be forced to do.”).
  74. .Thomas Hobbes, Leviathan 100 (J.C.A. Gaskin ed., Oxford Univ. Press 1998) (1651).
  75. .For discussion of the non-wrongfulness of takings, see infra section II(C)(2). For a representative characterization of corrective justice as involving rectifying wrongs, see Jules L. Coleman, Risks and Wrongs 320 (1998) (“[A] natural understanding of the principle of corrective justice is the following: Corrective justice imposes the duty to repair the wrongs one does.”). Coleman’s own view is slightly different: “The view I want to defend is that the duty of wrongdoers in corrective justice is to repair the wrongful losses for which they are responsible.” Id. at 324. Ernest Weinrib’s theory of “corrective justice” encompasses not only rectification of wrongfully inflicted losses but also restitution for certain lawful acts that enrich one party but harm another. Weinrib’s example is restitution for damage caused by one party’s entry onto another’s property to seek shelter from an imminent grave peril, entry that was made without the owner’s consent but nevertheless is permitted by law. (Weinrib offers the famous case of Vincent v. Lake Erie Transp. Co., 109 Minn. 456 (1910), as a specific example.) Ernest J. Weinrib, The Idea of Private Law 196–200 (1995). Some historical views of the nature of takings compensation thus might fit fairly easily within Weinrib’s category of “corrective justice.” On those views, whether the state is permitted to take private property for public use and whether the state is obligated to compensate for taken property are two separate questions, with answers that derive from two separate sets of reasons. As a result, although the state would be wrong not to compensate for taken property, failure to compensate would not make the taking itself wrongful. See, e.g., Philip Nichols, The Law of Eminent Domain § 204, at 623–24 (1917) (providing historical support for the idea that “the obligation to make compensation is a condition imposed upon the exercise of the power in all well[-]ordered communities” rather than “an essential part of the power of eminent domain”). Similarly, Weinrib’s discussion treats the question of the permissibility of unconsented entry as distinct from the question of whether compensation is owed for losses caused by the entry. See, e.g., Weinrib, supra, at 198 (arguing that in cases such as Vincent v. Lake Erie Transportation Co. “[t]he fact that the use of the dock was justified does not mean that the defendant should retain the benefit of that use by avoiding its costs”). However, a significantly different competing view—a view for which this Article is arguing—is that payment of compensation is (at least normally) a necessary condition for an exercise of eminent domain to be legitimate at all. See infra note 97 and accompanying text. Such a view seems to fall outside Weinrib’s category of “corrective justice.”
  76. .Although identifying this category’s outer contours lies beyond the scope of this Article, it seems likely that “relational justice” is a general category that includes both “corrective justice” and purely contractual “commutative justice” (of the sort that Hobbes had in mind), as well as at least one other subcategory of justice—namely the subcategory that includes the justice required by the relationships that exist in the takings context. Since coining yet another term for this particular subcategory would not enhance the clarity of this Article’s discussion, and the multiplication of jargon can confuse more than it enlightens, this Article will simply use the term “relational justice,” with a tacit qualification that the type of relational justice relevant in the takings context is only one of several possible types of relational justice.
  77. .Cf. Joseph William Singer, Democratic Estates: Property Law in a Free and Democratic Society, 94 Cornell L. Rev. 1009, 1047–48 (2009) (discussing implications of the fact that “property law concerns . . . relations among persons”).
  78. .See supra subpart I(A).
  79. .Some prominent commentators have expressed moral qualms about pervasive substitution of commercial relationships for other forms of relationship. These critics might consider the replacement of public compensation with private insurance purchased in the open market to be another instance of the “commodification” that they warn against. However, the arguments advanced in the present Article rest on other grounds and do not presuppose any general aversion to (or endorsement of) the spread of market-based relationships. For examples of these concerns, see Elizabeth Anderson, Value in Ethics and Economics 166 (1993) (“In treating human relations as indifferently substitutable means for acquiring goods, welfare economics blinds itself to the ways markets undermine certain expressive relations with others.”); Margaret Jane Radin, Contested Commodities xii (1996) (asserting that “there can be coexistent commodified and noncommodified understandings of various aspects of social life” but also that it is necessary to ask “whether that coexistence is unstable, threatening to decay into a monolithic structure of commodification”); see also Michael J. Sandel, What Money Can’t Buy: The Moral Limits of Markets 51 (2012) (prefacing a critical discussion of using monetary incentives as a policy tool with the observation that, “[t]o a remarkable degree, the last few decades have witnessed the remaking of social relations in the image of market relations”).
  80. .See Fisher, supra note , at 9 (reporting the revenue that different levels of government receive from property taxes, relative to other revenue sources).
  81. .Cf. Schill, supra note , at 854 n.98 (“A major difference between compensation and insurance is that for compensation, there is no necessary correlation between the amount of taxes paid by a citizen and the expected loss to that citizen from condemnation. With regard to insurance, however, the insured’s premium includes a payment which is approximately equal to the expected loss from the event insured against.”).
  82. .For the same reason, salaries paid to soldiers conscripted into the military in times of war are not analogous to insurance claims payments made to those conscripted citizens to alleviate the burdens of military service, payments that might have been replaced by requiring all military-aged males to buy private insurance against conscription. (Since conscription is done by lottery, the similarity between money paid to conscripts and private insurance might be even greater than in the case of takings. Hence, the implausibility of thinking of military pay as insurance payments further suggests that skepticism of that analogy in the takings context is appropriate.)
  83. .One might, of course, have a view that the costs of all such losses should be spread across everyone—“socialized”—and thus that the government should compensate each victim who suffered a loss. Such a view would obviously provide an additional reason to require the government to pay compensation for the property that it takes, but trivially so, since the government would be paying compensation for all losses.
  84. .There is one notable exception to that general characterization of casualty insurance: When the loss is the result of wrongdoing by someone other than the victim, this additional party—the wrongdoer—does have a relevant relationship to the loss. The existence of a robust private insurance market against such losses—for example, theft insurance—does not suggest that private-insurance is an appropriate substitute for compensation by the wrongdoer who caused those losses. Victims of theft, negligent harms, and the like need to insure against those losses because determining the identities of the relevant wrongdoers may be difficult—for example, some thieves are never caught—and even if their identities are discovered, suing them for compensation can be expensive and may ultimately be fruitless if the defendants lack the resources to provide the required compensation. See generally, e.g., S. Shavell, The Judgment Proof Problem, 6 Int’l Rev. L. & Econ. 45 (1986) (offering an economic analysis of “judgment proof” defendants). Hence, compensation through private insurance is merely a second-best solution, one that does not relieve the wrongdoer of an obligation to pay compensation. These reasons for settling for a second-best solution do not apply in eminent domain cases, since the taker’s identity is obvious and governments’ power to tax ensures that governments rarely will lack sufficient funds to pay compensation. Moreover, in these sorts of casualty-insurance cases, subrogation typically allows the insurer who pays compensation to seek reimbursement from the perpetrator. See 25 C.J.S. Damages § 190 (2012) (“[T]he insurer is entitled to be subrogated to the rights of the insured as against the tortfeasor or . . . may recover back from the insured the amount of the recovery.”). So even if takings compensation were analogous to casualty insurance, and thus private insurance could displace the government in its role as insurer against losses from takings, that fact would not itself displace the government’s role as perpetrator of the taking, and it therefore would leave the government still liable to pay compensation (this time, however, to the insurance companies).
  85. .Cf. United States v. Causby, 328 U.S. 256, 262 (1946) (distinguishing between “a definite exercise of complete dominion and control over the surface of the land” and “a case of incidental damages arising from a legalized nuisance”). This characterization of eminent domain’s burdens does require some qualification, since taking property typically imposes more than one type of burden. Some of those burdens are merely incidental to the benefit received—for example, the transaction costs of finding and acquiring suitable property to replace the property that was taken and the psychological pain that owners may suffer from leaving their property. The value lost in suffering those burdens is not transferred to the parties that benefit from the taking. Instead, it simply is lost. See Lee Anne Fennell, Taking Eminent Domain Apart, 2004 Mich. St. L. Rev. 957, 964 (2004) (“Because [the subjective premium] is personal to the individual landowner, its confiscation in the course of eminent domain necessarily means its outright destruction rather than its transfer to someone else.”). Nevertheless, the chief burden that a taking imposes is the loss of the property itself, and that loss is essential in producing the gain that the public receives as a result of the public project that made the taking necessary. (If the loss were not essential to produce the gain, then the taking would not be necessary and therefore would not be permitted at all. See 6 Nichols on Eminent Domain § 24.11[2][a][iii] (2018) (“Before exercising the power of eminent domain there must be a determination that the taking is necessary to advance a legitimate public purpose.”).)
  86. .See, e.g., Shoemaker v. United States, 147 U.S. 282, 297, 321–22 (1893) (addressing condemnation of land for the creation of Rock Creek Park in Washington, D.C.); Kohl v. United States, 91 U.S. 367, 374–75 (1875) (addressing condemnation of land in Cincinnati so that a post office and other public buildings could be built there).
  87. .Kaplow acknowledged that losses from takings differed from losses that the government did not cause, but he did not recognize that the relevant difference lies not only in who caused the loss but also in who benefits by imposing the loss. As a result, he saw no reason to have different approaches to compensating for those losses:

    With fire insurance or market risks, one expects to be self-reliant in securing protection; when the risk is directly linked to the government, one is more inclined to look to the government for protection. But this distinction does not indicate what different values, if any, are implicated by the origin of unequal burdens, or that any such difference in values would call for a governmental response that diverges from what investors would find worthwhile when responding to market risks.

    Kaplow, supra note , at 578. Recognizing the importance of the community’s having benefited by imposing the loss explains why the inclination to look to the community (acting through the government) for compensation makes sense for certain types of policy change, independent of any economic considerations of the sort that Kaplow’s account focuses on.

  88. .Although each member of the community likely does not personally benefit from every particular exercise of eminent domain—for example, an aged childless widow does not personally benefit from the construction of a public school on the other side of town—that fact is immaterial. What matters is that such people are still members of the community that does benefit from the exercise. Doctrinally, this fact is reflected in an acknowledgment that a taking can validly be for “public use” even if not every member of the public personally benefits from it. See, e.g., Fallbrook Irrigation Dist. v. Bradley, 164 U.S. 112, 161–62 (1896) (“It is not essential that the entire community or even any considerable portion thereof should directly enjoy or participate in an improvement in order to constitute a public use.”).
  89. .Armstrong v. United States, 364 U.S. 40, 49 (1960).
  90. .See supra section I(B)(2).
  91. .Thomas Merrill’s discussion in Rent Seeking and the Compensation Principle reflects this ambiguity. He says: “The justification for compensation that emerges most clearly from the opinions of the Supreme Court is that of equal treatment. Simply put, it is said to be ‘unfair’ to make a few pay for the good of the many.” Merrill, supra note , at 1579. The latter sentence seems implicitly to acknowledge the normative importance of the relationship between the people who are burdened and the people who are benefited by means of imposing those burdens. However, Merrill’s subsequent elaboration of this point instead focuses exclusively on the relationship between people whose property is taken and people whose property is not taken—i.e., on the relationship between those who are burdened and those who are not burdened. Id. at 1579–80.
  92. .Armstrong, 364 U.S. at 48.
  93. .See, e.g., Madden v. Comm’r, 514 F.2d 1149, 1151 (9th Cir. 1975) (referring to exercise of eminent domain as a “forced sale”); Law Offices of Vincent Vitale v. Tabbytite, 942 P.2d 1141, 1147 (Alaska 1997) (asserting that takings compensation “can readily be categorized as money accruing from a sale of the land in question”); People ex rel. Dep’t of Pub. Works v. Church, 136 P.2d 139, 145 (Cal. App. Dep’t. Super. Ct. 1943) (“The relations between a condemnee and a condemnor have often been likened in decisions to those between a seller and a purchaser.”); Pearson v. Johnson, 54 Miss. 259, 263 (1876) (“The proceeding has the elements of an enforced sale.”); see also 1 Nichols on Eminent Domain § 1.142[6] (2016) (asserting that acceptance of the theory that eminent domain is a compulsory sale “seems almost inevitable” in jurisdictions that require payment of compensation in advance for taken property); 1 William Blackstone, Commentaries *139 (“The public is now considered as an individual, treating with an individual for an exchange. All that the legislature does is to oblige the owner to alienate his possessions for a reasonable price . . . .”). The element of compulsion, however, keeps eminent domain largely outside the realm of contract law, except by implication if the state fails to pay compensation. See, e.g., In re Parking Place in Hempstead, 140 N.Y.S.2d 341, 343 (N.Y. Sup. Ct. 1955) (“It seems clear that acquisition of real property by condemnation possesses no contractual attributes. Its very nature is opposed to the idea of contractual relationship since it is a power to take property in spite of the owner’s objection so long as just compensation is made.”). Doctrinally, the analogy to a sale thus has its strongest legal force if compensation is not paid. See, e.g., Carman Fitz Randolph, The Law of Eminent Domain in the United States 209–10 (1894) (“The right to compensation[,] whether it be deemed inherent or constitutional, is not usually considered as founded on contract. . . . But where land is condemned and compensation withheld, the transaction is generally treated as an ordinary purchase, and the owner holds a lien for the compensation.”).
  94. .Jackson v. State, 106 N.E. 758, 758 (N.Y. 1914) (Cardozo, J.).
  95. .See, e.g., Alabama v. United States, 282 U.S. 502, 503 (1931) (“There is a strong analogy in the cases requiring the United States, under an implied contract, to pay for private property which it takes for public use.”).
  96. .See, e.g., 7 Nichols on Eminent Domain § G2.05[3] (2018) (stating that “a good faith attempt to negotiate a purchase” is one of the “more common” conditions precedent that must be satisfied before a government is entitled to take property).
  97. .The payment of money is a necessary constituent of a legitimate act in such circumstances. For a related discussion of this point, see Brian Angelo Lee, Emergency Takings, 114 Mich. L. Rev. 391, 402 (2015) (describing the role of “constitutive compensation” in legitimating actions). Jules Coleman has offered a similar example to make a similar point. See Coleman, supra note , at 291–92 (distinguishing among three different roles that compensation may play and offering an example of eating a salad at a Korean produce market, where “[m]y eating the salad they offer is justified provided I pay them what they (fairly) charge me for it”).
  98. .This conclusion is in tension with some older descriptions of the compensation obligation’s nature. For example, in 1879, the U.S. Supreme Court asserted that “[t]he clause found in the Constitutions of the several states providing for just compensation for property taken is a mere limitation upon the exercise of the right [of a state to take property through eminent domain].” Boom Co. v. Patterson, 98 U.S. 403, 406 (1879). The implication was that payment of compensation was not inherently a necessary condition for an exercise of eminent domain to be legitimate, although a constitution might choose to add such a requirement. See Thayer, supra note 61, at 251 (“[T]he right of the State to take . . . has no condition of compensation annexed to it, either precedent or subsequent.”). Other older cases disagreed. See, e.g., Gardner v. Trs. of Newburgh, 2 Johns. Ch. 162, 166 (N.Y. Ch. 1816) (Kent, C.) (asserting that provision of “fair compensation” is a “necessary qualification accompanying the exercise of legislative power, in taking private property for public uses,” one that “is admitted by the soundest authorities, and is adopted by all temperate and civilized governments, from a deep and universal sense of its justice”). The argument offered in this Article casts doubt upon the soundness of the Boom Co. Court’s approach.
  99. .See, e.g., Kelo v. City of New London, 545 U.S. 469, 477–78 (2005) (stating the unconstitutionality of a taking to confer “a private benefit on a particular private party” or “under the mere pretext of a public purpose”); Mo. Pac. Ry. Co. v. Nebraska, 164 U.S. 403, 417 (1896) (holding that a taking for private, rather than public, use violated the Fourteenth Amendment).
  100. .See, e.g., Haw. Hous. Auth. v. Midkiff, 467 U.S. 229, 245 (1984) (“A purely private taking could not withstand the scrutiny of the public use requirement; it would serve no legitimate purpose of government and would thus be void.”).
  101. .See, e.g., Georgia v. City of Chattanooga, 264 U.S. 472, 480 (1924) (“The power of eminent domain is an attribute of sovereignty, and inheres in every independent State.”); Boom Co., 98 U.S. at 406 (“The right of eminent domain, that is, the right to take private property for public uses, appertains to every independent government. It requires no constitutional recognition; it is an attribute of sovereignty.”); Prop. Reserve, Inc. v. Superior Court, 375 P.3d 887, 905 n.12 (Cal. 2016) (“California and federal authorities establish that eminent domain . . . is ‘an inherent attribute of sovereignty’ . . . .”); H.W. Halleck, Elements of International Law and Laws of War 74 (1866) (“It is a right which, from its very nature, is inseparable from the sovereignty, and is necessarily transferred with the sovereignty.”).
  102. .Rights and duties ordinarily go together. For a classic discussion of the correlation between rights and duties in legal theory, see Hohfeld, supra note 71, at 30–32. See also David Lyons, The Correlativity of Rights and Duties, 4 Noûs 45, 45 & n.4 (1970) (collecting sources on the correlation in the context of moral philosophy).
  103. .Although the idea that property owners have duties toward their communities is unlikely to provoke vigorous opposition, some recent scholars have argued that those duties deserve more attention than current property scholarship typically offers. See, e.g., Joseph William Singer, Entitlement: The Paradoxes of Property 16 (2000) (criticizing theories that “obscure or deny something important: owners have obligations as well as rights”); Gregory S. Alexander, The Social-Obligation Norm in American Property Law, 94 Cornell L. Rev. 745, 748 (2009) (arguing that “the responsibility dimension of private ownership has been sorely under-theorized in American law”); Nestor M. Davidson, Property and Relative Status, 107 Mich. L. Rev. 757, 771 (2009) (examining an “enduring communitarian perspective” that offers “a functional vision of property as the realm of deeply embedded relationships and community, with a normative focus on the obligations that arise from these interconnections”); David Lametti, The Concept of Property: Relations Through Objects of Social Wealth, 53 U. Toronto L.J. 325, 326 (2003) (proposing “a new metaphor for understanding private property” that “allows specific objects of property to carry with them duties of stewardship or obligations to use in a certain manner”); see also Sandel, supra note 72, at 240–41 (criticizing “moral individualism” for being unable to account for “familiar features of our moral and political experience,” including “the special responsibilities of family members, and of fellow citizens, for one another”).
  104. .This requirement has been established for so long that its standard formulation is in Latin: sic utere tuo ut alienum non laedas, commonly translated, “Use your own property so as not to harm that of another.” See, e.g., 57A Am. Jur. 2d Negligence § 89 (2018) (interpreting the common law maxim).
  105. .For an overview of trespass law, see Joseph William Singer, Property 23–44 (3d ed. 2010). For an overview of nuisance law, see id. at 97–127.
  106. .See, e.g., Ploof v. Putnam, 71 A. 188, 189 (Vt. 1908) (concerning a boat that moored to a dock without permission to seek safety in a storm); 1 Dan B. Dobbs, Paul T. Hayden & Ellen M. Bublick, The Law of Torts § 117 (2d ed.) (2011) (discussing “private necessity”); 2 id. § 399 (distinguishing nuisance from trespass on the basis of nuisance’s requirement that the relevant interference be unreasonable).
  107. .Cf. Commonwealth v. Tewksbury, 52 Mass. 55, 57 (1846) (Shaw, C.J.) (“All property is acquired and held under the tacit condition that it shall not be so used as to injure the equal rights of others, or to destroy or greatly impair the public rights and interests of the community; under the maxim of the common law, sic utere tuo ut alienum non lædas.”). Attempting to establish the ultimate philosophical foundations for the existence of this duty would be far more ambitious than needed to address the specific questions at issue in this Article. This Article’s argument, by design, assumes no commitment to any particular theory of the sources of political obligation. For present purposes, it will be sufficient if the characterization of this duty is at least roughly plausible and is broadly consistent with what can be observed in the law.
  108. .For a distinct but related idea, see Larissa Katz, Spite and Extortion: A Jurisdictional Principle of Abuse of Property Right, 122 Yale L.J. 1444, 1482 (2013) (identifying a category of property doctrines that prevent “abuse of right”).
  109. .See, e.g., 6 Nichols on Eminent Domain § 24.11[2][a][iii] (2018) (noting that “[b]efore exercising the power of eminent domain there must be a determination that the taking is necessary to advance a legitimate public purpose” and collecting cases). The necessity in question here is disjunctive necessity. I.e., it is necessary to take some property within a given set of properties, and the specific property that is taken belongs to that set. Hence, taking Blackacre for a public project will qualify as necessary even if the project could have proceeded just as well by taking Whiteacre instead, provided merely that the project’s success required taking one of the two. See, e.g., Fitz Randolph, supra note , at 49 (“The absolute necessity of a particular location is not in any case a prerequisite to the exercise of the eminent domain . . . .”).
  110. .That the existence of such duties depends on the existence of a specific type of relationship between the person and the relevant community is evident in traditional rules limiting a country’s legal authority solely to particular geographic regions or groups of people. See, e.g., Antonio Cassesse, International Law 49–50 (2d ed. 2005) (discussing international law principles governing the scope of states’ sovereignty). Eminent domain is no exception to this general restriction on governments’ authority. Thus, the United States has no power of eminent domain over property owned by foreigners in foreign countries. See, e.g., Joseph Story, Commentaries on the Conflict of Laws § 428, at 717 (3d ed. 1846) (“[R]eal estate, or immovable property, is exclusively subject to the laws of the government, within whose territory it is situate.”). Likewise, within the United States, an individual state’s power of eminent domain extends only to property located within its borders. See, e.g., Crosby v. Hanover, 36 N.H. 404, 423 (1858) (“Our official powers are confined to the limits of our own State, and the court in this case cannot require or authorize the town to go into Vermont to take the property of the corporation there, and the commissioners had no power to condemn property situated in that State.”).
  111. .The notion that members of political society have reciprocal duties toward each other, merely by virtue of their shared membership in that society, may seem obvious. Nevertheless, various theorists have taken care to emphasize the point. See, e.g., Dagan, supra note 51, at 772:

    The mere fact of belonging or membership entails special responsibilities. Hence, land ownership—like ownership at large—is perceived not merely as a bundle of rights, but also as a social institution that creates bonds of commitment among landowners and between landowners and others who live, work, or are otherwise affected by the landowners’ properties.

    The historian James Kloppenberg identifies “the ethic of reciprocity” as one of the three “premises that . . . lie beneath modern democracy.” James T. Kloppenberg, Toward Democracy: The Struggle for Self-Rule in European and American Thought 9–10 (2016); see also Christopher Serkin, Affirmative Constitutional Commitments: The State’s Obligations to Property Owners, 2 Brigham-Kanner Prop. Rts. Conf. J. 109, 110–11 (2013) (arguing that the demands of reciprocity both in Dagan’s account and in progressive “human flourishing” theories of property imply that the government sometimes has affirmative obligations to protect private property).

  112. .See, e.g., Robert Paul Wolff, In Defense of Anarchism 72 (1998) (concluding, somewhat reluctantly, that “[t]here would appear to be no alternative but to embrace the doctrine of anarchism and categorically deny any claim to legitimate authority by one man over another”).
  113. .That civic duties exist is obviously an essential presupposition of the entire legal system, for without it, there would be no law at all. For a succinct survey of the contending accounts through history of why those duties exist, see Richard Dagger & David Lefkowitz, Political Obligation, in Stanford Encyclopedia of Philosophy (Edward N. Zalta, ed., Fall 2014), https://plato.stanford.edu/entries/political-obligation/ [https://perma.cc/Z86H-XBEW].
  114. .For an argument that reciprocal obligations may sometimes compel the government to protect property, see Serkin, supra note , at 130–32.
  115. .See, e.g., E.L. Strobin, Annotation, Right to Condemn Property in Excess of Needs for a Particular Public Purpose, 6 A.L.R. 3d § 2[a], at 297 (1966) (noting “the widely recognized principle that the power of condemnation may not be used to condemn property in excess of that needed for public purposes”).
  116. .James Bradley Thayer noted, more than a century ago, the distinction between the need to take property and the need to take without paying compensation. See Thayer, supra note , at
    250–51 (“There is a necessity for the taking, but none for taking without compensation, or without a just compensation, and such as shall put the party affected, so far as may be, on a level with the rest of the community.”). However, Thayer seems to have overlooked this distinction’s implications. He went on to argue that the government would not have “exercised its power wrongfully” if it failed to pay compensation for property that it took because “the right of the State to take springs from . . . a necessity of government,” while owners’ rights to compensation for taken property spring from a different source—“the natural rights of the individual.” Id. at 251. Whether a taking is legitimate and whether compensation was owed were, Thayer thought, two independent questions. The flaw in Thayer’s argument arises from the distinction that Thayer himself had noted. Even if taking property is necessary, it does not follow that taking the property without compensation is necessary, and thus it does not follow that “necessity” gives the State the right to engage in uncompensated takings.
  117. .In rare cases, circumstances may arise in which the government genuinely cannot pay for the property that is needed for a public project. It may be that this distinct type of necessity (a need not to pay compensation) might excuse or justify the government’s not paying compensation or paying only partial compensation. For a discussion of actual examples of this possibility, see Lee, supra note , at 404–07. However, instances of non-compensation necessity are likely to be rare because governments have the power to tax and thus generally can acquire the money needed to pay compensation, even if political considerations would make them prefer not to.
  118. .The same principle applies to trespasses that are compelled by private necessity. Such incursions are not deemed to be wrongs, but nevertheless, compensation may be owed. See Restatement (Second) of Torts § 197 (Am. Law Inst. 1965) (indicating that a private-necessity privilege exists for trespass but noting that where “entry is for the benefit of the actor or a third person,” the actor “is subject to liability for any harm done in the exercise of the privilege”).
  119. .The role of compensation in takings is not to remedy a wrong but rather to prevent a wrong from occurring in the first place. See Lee, supra note , at 401–04 (distinguishing among types of compensation and referring to the role of compensation in eminent domain as “constitutive compensation”).
  120. .In practice, governments often borrow money to fund their expenses, including to pay compensation. As a result, the people ultimately paying compensation may not be the same people who decided to exercise the power of eminent domain or who initially benefited from the project. This wrinkle does not affect the analysis here, because even though the identities of the individuals who make up a community inevitably change as generations are born and pass away, the communities themselves typically remain, and the rights, privileges, and obligations of membership in that particular community are not determined solely by what happened during the time that each current member has been alive and living within that community. (Over long enough periods of time, of course, nearly every community will eventually disappear, just as the Roman Empire today exists only in history books and the imagination. However, a community’s lifetime is typically substantially longer than the span of time over which the costs and benefits of any given exercise of eminent domain are likely to be felt.)
  121. .See Eric Kades, Drawing the Line Between Taxes and Takings: The Continuous Burdens Principle, and Its Broader Application, 97 Nw. U. L. Rev. 189, 191 (2002) (“Perhaps the most surprising observation about recent commentary on drawing the line between taxation and takings is its paucity. This is surprising since the issue is so fundamental.”); Levmore, supra note , at 292 (“[E]very theory of takings law should explain or at least struggle with the question of why the power to tax—without compensation, of course—is not fundamentally inconsistent with the constitutional obligation to compensate condemnees.”). A survey of the relevant literature appears in Kades, supra, at 191–208.
  122. .See, e.g., Kaplow, supra note , at 519 (“Compensation for taxation would nullify the tax itself and, along with it, any prospect for raising revenue or implementing distributional policy.”). Taxation is not the only governmental policy with this feature. For example, Thomas Merrill has noted:

    [T]he losses caused by governmental action are often deliberately inflicted to provide incentives for private parties to take or avoid certain kinds of actions. A typical pollution control regulation, for example, is imposed at least in part to discourage polluting activities, through cost internalization. If the government compensated for such losses, this incentive effect would be eliminated.

    Merrill, supra note , at 1582.

  123. .Note that this account avoids any assumption that every taxpayer receives “in-kind” compensation that equals or exceeds his or her tax burden. For a discussion of accounts that do rely on such an assumption, see Kades, supra note , at 200–02. One problem with that assumption is that it seems unlikely to be true. See, e.g., Levmore, supra note , at 292 (“Broadly based taxes that are used to finance a range of projects . . . do not pretend to strike fair bargains with all taxpayers, but, instead, benefit some citizens and burden others.”). For a generally skeptical assessment of arguments based on appeals to in-kind compensation, see generally Brian Angelo Lee, Average Reciprocity of Advantage, in Philosophical Foundations of Property Law 99 (James Penner & Henry E. Smith eds., 2013).
  124. .Space does not permit examining whether the existence of reciprocal obligations not to inflict unnecessary harm might also argue in favor of certain types of tax policies, such as utility-equalizing “progressive” taxation.
  125. .See, e.g., Bell & Parchomovsky, supra note , at 1677 (“But the fact is that most government actions that harm citizens—from school closings to tax changes—do not entitle the affected citizens to compensation.”); Michael Heller & Rick Hills, Land Assembly Districts, 121 Harv. L. Rev. 1465, 1476 n.31 (2008) (“The traditional academic criticism of the obligation [to compensate] is that government does not routinely compensate persons for losses incurred as a result of other legal transitions—for instance, when government repeals a tax exemption or prohibits a previously legal activity.”). For a broader version of this challenge, based not only on losses caused by the government but also on losses caused by other factors, see Kaplow, supra note , at
    533–34 (noting earlier accounts that observed similarities between market risks and the risks of losses from government action, and suggesting that “none of the distinctions they offer for treating government and market risks differently withstands scrutiny”). Cf. Merrill, supra note , at
    1581–82 (asserting that “[t]he insurance theory [that compensation serves to maximize wealth by pooling risk for the benefit of risk-averse members of society] fails to explain why we willingly compensate in some cases while refusing to compensate in others”).
  126. .In fact, even changes that make some people better off while leaving everyone else’s situations unchanged (and thus no worse off)—i.e., “Pareto superior” changes—are rare in practice. See, e.g., Daniel A. Farber, What (If Anything) Can Economics Say About Equity?, 101 Mich. L. Rev. 1791, 1795 (2003) (reviewing Louis Kaplow & Steven Shavell, Fairness Versus Welfare (2002)) (“In reality, however, Pareto improvements are often hard to find. Because unanimous consent is unlikely to exist for changes in legal rules, law and economics practitioners often fall back to a broader but less compelling standard, Kaldor-Hicks efficiency.”).
  127. .Cf. Michelman, supra note 52, at 1178–79:

    [T]o insist on full compensation to every interest which is disproportionately burdened by a social measure dictated by efficiency would be to call a halt to the collective pursuit of efficiency. It would require a tracing of all impacts, no matter how remote, speculative, or arguable, and a valuation of all burdens, no matter how idiosyncratic or imponderable. If satisfactory performance of such an obligation is not absolutely impossible, at least it is clear that in many situations its costs would be prohibitive.

  128. .Cf. Blume & Rubinfeld, supra note , at 599–600:

    [M]arket failure provides a rationale for considering the payment of compensation not only when land is taken by eminent domain, but also when zoning changes or other governmental regulations affect land values. But this does not mean that compensation should be provided in all cases of insurance market failure. There are a number of costs associated with compensation, and we can only conclude that compensation ought to be paid when the benefits outweigh the costs.

  129. .See Dobbs, Hayden & Bublick, supra note 106, § 117, at 362 (discussing how “private necessity” may “protect defendants whose acts in emergencies would otherwise count as trespass to land or chattels or as conversion”).
  130. .This example is not purely hypothetical. See, e.g., Kaid Benfield, The Death—and Life—of Small Downtown America, Citylab (Sept. 7, 2012), http://www.citylab.com/work/2012/09

    /how-main-street-can-be-saved/3200/ [https://perma.cc/SH4Y-AKTM] (describing how the creation of interstate highway bypasses hurt small businesses that were located in downtown shopping areas).

  131. .The Doctrine of Double Effect’s origin is commonly attributed to Thomas Aquinas, 2 Summa Theologica Pt. II, question 64, art. 7 (Benzinger Brothers, Inc. 1947) (13th c.). One of the seminal sources of the Doctrine’s present-day scholarly interest is Philippa Foot, The Problem of Abortion and the Doctrine of the Double Effect, 5 Oxford Rev. 5 (1967).
  132. .The philosophical literature on the doctrine is now voluminous. For a recent overview, see Alison McIntyre, Doctrine of Double Effect, in Stanford Encyclopedia of Philosophy (Edward N. Zalta ed., Winter 2014), https://plato.stanford.edu/entries/double-effect/ [https://perma.cc/GVT5-T38A].
  133. .Jed Rubenfeld has argued that the “just compensation” requirement should apply solely to instances of the government taking possession of property and using it. Jed Rubenfeld, Usings, 102 Yale L.J. 1077, 1149 (1993). Although the analysis in the present Article agrees with Rubenfeld in treating transfers of possession as the quintessential type of taking, it does not imply that such “usings” are the only type of action that qualifies as a “taking” and thus requires compensation.
  134. .475 U.S. 211 (1986).
  135. .Id. at 223.
  136. .Among the other relevant considerations may be the extent to which the burdens imposed are unequal. This Article does not contend that the “equal treatment” account discussed in subpart I(B) is entirely wrong, but rather that, at best, it is only part of the story.
  137. .The discussion that follows will simply grant, for the purposes of argument, the anti-compensation thesis’s assumption that private insurance against takings losses would be available to purchase if the government were to stop compensating for takings. That assumption is controversial but ultimately speculative and therefore impossible to assess conclusively a priori. See, e.g., Blume & Rubinfeld, supra note 13, at 593–96 (attributing the absence of a private market for takings insurance to moral hazard and adverse-selection problems); Jonathan S. Masur & Jonathan Remy Nash, The Institutional Dynamics of Transition Relief, 85 N.Y.U. L. Rev. 391,
    421–26 (2010) (discussing the difficulties of pricing insurance against the risks of legal transitions); see also Farber, supra note 17, at 285 (noting the “element of speculation” in this debate and asserting that “conjecture” about whether a private market would arise “is not a satisfying basis for either a positive or a normative theory”).
  138. .This question is merely a specific instance of a longstanding general philosophical problem of how to decide what moral judgments should result when incommensurable values clash. See, e.g., Bernard Williams, Conflicts of Values (noting the existence of “a plurality of values which can conflict with one another, and which are not reducible to one another . . .”), in Moral Luck: Philosophical Papers 1973–1980, at 71 (1981).
  139. .For an impassioned assertion of the importance of assessing the sizes of alleged economic effects, rather than merely establishing that an effect exists or may exist, see Deirdre McCloskey, The Trouble with Mathematics and Statistics in Economics, Hist. Econ. Ideas, 2005, at 85, 88–90.
  140. .See supra section I(A)(1).
  141. .In general, cases in which improvements are made immediately before condemnation are not by themselves evidence of moral hazard unless the owner both recognized and discounted the risk of condemnation. The mere fact of a short gap in time between the improvement and the taking implies nothing by itself.
  142. .One recent article asserted:

    Numerous cases have come before our courts where improvements were made to land immediately prior to its condemnation . . . . These examples of eminent domain resulted in the demolition of millions of dollars worth of buildings and improvements that might never have been constructed at the outset if the landowners involved were properly incentivized to account for the risk that their property could be taken.

    Calandrillo, supra note , at 506–07. To its credit, that article cites four actual cases to support its assertion, a helpful improvement on the typical practice of resting the moral hazard concern on purely hypothetical examples. However, the four cases cited provide little evidence of any significant moral hazard. The strongest cited example is a Maryland case that does indeed involve an owner’s improving property despite having good reason to anticipate the government’s taking it, but, even in this case, the government admitted that the owner’s improvements might have been compatible with the use that the government would make of the property. Hence, the social cost of the improvement here may have been negligible. J.L. Matthews, Inc. v. Md.-Nat’l Capital Park & Planning Comm’n, 792 A.2d 288, 305–06 (Md. 2002). In another cited case, there was no evidence that the improvement’s builder anticipated losing the property, and in fact the government’s decision to take was a reaction to the improvement’s commencement. Div. of Bond Fin. of Dep’t of Gen. Servs. v. Rainey, 275 So. 2d 551, 553 (Fla. 1973). In a third case, the property increased in value before the taking, but there was no indication that the increase was the result of any improvements rather than merely a rise in the real estate market, and the court explicitly said that there was “no indication that the increase in value of defendant’s property was due to bad faith improvements by defendant . . . .” Saratoga Fire Prot. Dist. v. Hackett, 97 Cal. App. 4th 895, 906 (2002), as modified on denial of reh’g. Finally, in the remaining case, the “improvements” were crops that had been planted in the ordinary course of farming—raspberries that had been planted two years before the taking was authorized, Christmas trees that had been planted seven years before the taking was authorized, and hay that had been planted recently but that the government was able to reap and sell. Town of Newington v. Estate of Young, 777 A.2d 219, 238–41 (Conn. Super. Ct. 2000). It seems unlikely that eliminating whatever moral hazard existed in this case would have avoided significant costs, if any at all.

  143. .Shavell, supra note , at 132–33 concedes that the infrequency of takings may limit the efficiency gains from eliminating compensation’s moral hazard but appears not to see that fact as raising doubts about the desirability of eliminating the compensation requirement.
  144. .In a related vein, Thomas Miceli notes that “an important shortcoming of the existing literature on efficient compensation rules” is that the economic models it uses “simply assume that the court has the necessary information without worrying about how it acquired it.” Miceli, supra note , at 112. Miceli ultimately simply resigns himself to accepting the difficulty: “In the end, all we can say is that some residual inefficiencies are the unavoidable cost of relying on non-market allocation mechanisms that require information that the government or the courts cannot easily obtain.” Id. He seems not to have noticed that private owners face the same information problem, nor the implications of that fact for the moral hazard argument.
  145. .Abraham Bell has suggested that when considering “the efficiency of property development, it may not be necessary to quantify the ‘takings risk . . . .’” Abraham Bell, Not Just Compensation, 13 J. Contemp. Legal Issues 29, 56 (2003). Bell’s suggestion springs from a general idea that if “the market in which the care is to be exercised is considered roughly efficient”—by which Bell seems to mean that the widespread, established practices in the relevant field of action produce socially efficient outcomes—then an individual could identify the socially optimal course of action in a given instance simply by conforming to those practices. Id. Thus, in the takings context, “we might determine the reasonable standard of property development by reference to similar properties in similar locations.” Id. This suggestion has two fundamental problems. One basic problem is that markets and social practices are not inherently efficient; they become efficient as a result of the informed decisions of the many participants in those markets or practices. Hence, if it is impossible for Person A to identify the socially efficient development decision in the face of potential takings—e.g., because the information necessary to make the required expected-value calculations is unavailable—then it will do no good to tell Person A just to do what Persons B, C, D, E, and F have done, since all of those people made their decisions burdened by the same information constraints that affect Person A’s decision. Under certain circumstances, aggregating predictions made by basically knowledgeable people can harness what James Surowiecki popularized as the “wisdom of crowds,” and thus arrive at a more accurate prediction than any of those individuals would have made alone. James Surowiecki, The Wisdom of Crowds, at xiv (2004). However, this aggregation procedure can only refine the knowledge that was implicit in each individual prediction; it cannot make something out of nothing. Thus, for example, in one of Surowiecki’s famous examples, the location of a vanished submarine was determined with remarkable accuracy by aggregating the guesses of “a team of men with a wide range of knowledge, including mathematicians, submarine specialists, and salvage men.” Id. at xx–xxi. However, had the searchers not consulted experts but instead aggregated the guesses of ordinary people with no relevant technical expertise, the resulting prediction might have been slightly less bad than any individual layperson’s prediction, but it still would have been useless. One cannot bootstrap oneself from ignorance to knowledge by conforming one’s views to those of other people who are equally ignorant. See, e.g., Cass R. Sunstein, Infotopia: How Many Minds Produce Knowledge 25 (2006) (asserting that “[t]he accuracy of judgments of statistical groups is best explained by reference to the Condorcet Jury Theorem,” which rests on an assumption that “each person [in the group] is more likely than not to be correct”); Dan Cassino, The “Wisdom of the Crowd” Has a Pretty Bad Track Record at Predicting Jobs Reports, Harv. Bus. Rev. (July 8, 2016), https://hbr.org/2016/07/the-wisdom-of-the-crowd-has-a-pretty-bad-track-record-at-predicting-jobs-reports [https://perma.cc/M9HZ-PFM7] (noting that “aggregation doesn’t get rid of any errors that are shared across the people being aggregated”). Moreover, even if by some fortunate accident, other property owners’ decisions in the face of potential takings happened to be efficient, that happy situation would not help an owner who lacked adequate information about the probability that his or her property would be taken. To follow Bell’s proposal, such an owner would need to identify the decisions that others had made in situations similar to the owner’s—i.e., made when the probability of a taking was similar in size to the probability that this owner faces. But the ability to identify properties that faced a similar probability of being taken depends on already being able to determine both the probability that one’s own property will be taken and the probabilities that those other properties would have been taken. Since the ability to adequately assess the probability of a parcel’s being taken is exactly what the owner lacks, and is exactly what Bell’s suggestion aspired to make unnecessary, Bell’s proposal provides no way to avoid the problem of inadequate information.
  146. .If the risk of condemnation is 15%, then the project’s net expected value would be [(0.85)*(5000) + (0.15)*(0)] – 4000 = $250. If the risk is 25%, then the net expected value would be –$250. The “break-even” point occurs when the probability of a taking is 20%. In that case, the expected value of the project would be (0.8)*(5000) + (0.2)*(0) = $4000, which exactly equals the project’s $4000 cost, producing a net expected value of $0.
  147. .If the owner has purchased private insurance against takings, then the premiums charged by the insurance company to insure the contemplated improvement should reflect the risk that the improvement will be taken by the government. This price signal might enable the owner to know when the project would be socially efficient and when it would not. But even if insurance companies would have better information about takings than individual owners do, there is no obvious reason to think that they could reliably distinguish between situations with a 15% risk of a taking and situations with a 25% risk. Takings occur infrequently enough and in a wide enough diversity of circumstances that insurers would lack sufficient data to generate such precise estimates.
  148. .Steven Shavell’s hypothetical illustration of the moral hazard problem involves an improvement that would have a positive expected value only if the probability of the government’s taking the property is less than 20% but that the owner chose to undertake despite the probability of a taking’s being 40%. Shavell, supra note , at 131. Shavell’s example thus relies on owners’ being able accurately to distinguish situations in which the probability of a taking is 20% from situations in which that probability is 40%. In practice, it seems much more likely that in both cases owners would be able to say only that a taking seems unlikely but possible, and thus they would be unable to make the calculation upon which Shavell’s argument rests.
  149. .This objection applies equally to arguments that justify the “just compensation” requirement on the grounds that without such compensation risk-averse owners may make fewer investments than is socially optimal. For an example of this sort of argument, see Dagan, supra note 51, at 749–50 (asserting that “no-compensation regime[s] may also generate inefficiencies” by producing under-investment in assets potentially subject to takings).
  150. .Moreover, the owner cannot always be certain of breaking even, because there is always a risk that the government’s estimate of the property’s market value will be inaccurate. For example, Yun-Chien Chang has argued that in practice the value of improvements to property is often not included in the calculation of that property’s fair market value when takings compensation is calculated. Yun-chien Chang, Private Property and Takings Compensation 27–30 (2013). Risk-averse owners would prefer to avoid that sort of hazard. Owners also might naturally prefer to avoid the frustration of seeing what they built be demolished.
  151. .Adding to that incentive is the risk that widespread knowledge of the possibility that property in that area might be taken will, over time, cause a substantial drop in that property’s market value and thus in the amount of compensation received. See Gideon Kanner, Condemnation Blight: Just How Just Is Just Compensation?, 48 Notre Dame Law. 765, 767–69 (1973) (discussing how, because government projects take time, “notice that a taking is imminent becomes widespread, which in turn promotes a wholesale departure of tenants, reluctance on the part of owners in the affected area to invest in improvements and maintenance, and distortion of the real estate market”). Note that, for the purposes of compensation, taken property’s fair market value is determined as of the date of the taking, not as of an earlier date when the possibility of a taking began to be anticipated. See Horne v. Dep’t of Agric., 135 S. Ct. 2419, 2432 (2015) (“The Court has repeatedly held that just compensation normally is to be measured by the market value of the property at the time of the taking.” (internal quotations and citations omitted)).
  152. .For a related argument, see Kanner, supra note , at 768 (“What businessman in his right mind would buy or lease under such circumstances? Why should he remodel, install trade fixtures, buy stock-in-trade, and develop goodwill for his business, only to have it all confiscated when the threatened condemnation comes?”).
  153. .See, e.g., Mitchell v. United States, 267 U.S. 341, 345 (1925) (denying any obligation to compensate owners of taken property for consequential losses to their business as a result of the taking); 4 Nichols on Eminent Domain § 12B.09 (2018) (collecting sources excluding evidence of past and future profits from consideration).
  154. .The moral hazard argument for the anti-compensation principle faces the additional critical difficulty that, even if the hazard were large, it would not imply that the government should pay zero compensation. Instead, it would leave the correct level of compensation indeterminate. As Blume, Rubinfeld, and Shapiro noted, paying any fixed amount of compensation would eliminate the moral hazard in question. Blume, Rubinfeld & Shapiro, supra note 21, at 78. In fact, any compensation scheme that disconnects the amount of compensation received from the value of the taken improvements would work, including randomly determining the amount of compensation to be paid or paying a sum equal to the average value that improvements on similarly sized property have. More complicated, but also potentially effective, is Fischel and Shapiro’s suggestion that the government pay property owners for options to take their property without compensating for any improvements built on it. See Fischel & Shapiro, supra note 55, at 274 (“One way of dealing with the moral hazard problem is for the government to purchase in advance of the landowner decisions an option to take the property without paying for lost capital.”). Such an approach, in effect, simply amounts to making a partial taking with compensation.
  155. .Bad-faith improvements made merely to “punish” the government by increasing its cost of taking the property already are excluded from the compensation that the government must pay. See 4 Nichols on Eminent Domain § 13.12[3] (2018) (collecting cases).
  156. .See, e.g., Cooter & Ulen, supra note 22, at 178 (“The discipline of competition causes a higher level of administrative efficiency in private insurance funds than in state insurance funds.”); Calandrillo, supra note 34, at 507–08 (explaining incentives for private insurers to minimize administrative costs).
  157. .Both Kaplow and Shavell seem to have had this potential source of savings in mind. See Shavell, supra note 4, at 129 (suggesting that “insurance-related administrative costs would be lower than” state-compensation-related administrative costs because “the process by which the state determines the amount of compensation is likely to be more cumbersome than insurers’ procedures”); Kaplow, supra note 14, at 547 (noting that state-compensation-related administrative costs may be higher because “government institutions . . . rely upon review procedures that typically are more costly than those used by private parties in similar contexts”).
  158. .See Klein, supra note 29, at 6 (explaining the concept of diversification of risk).
  159. .See id. at 7 (“Risk and uncertainty is reduced through . . . the greater predictability of losses achieved by increasing the number of members of the pool.”).
  160. .For a similar argument, see Cooter & Ulen, supra note , at 178 (“The state can spread the risk of takings through the base of all taxpayers, which is broader than the base of all policy holders in any insurance company. So, risk-spreading argues for public insurance.”).
  161. .Reinsurance can potentially reduce this extra risk, but only at an extra expense, and cannot eliminate it. See generally 44A Am. Jur. 2d Insurance § 1812 (2016) (describing the concept of reinsurance).
  162. .Abraham Bell and Gideon Parchomovsky warn that creating a situation in which the government has nonpublic information (about what property will be taken) that is economically significant both to owners and insurance companies would create socially costly incentives for rent-seeking. Bell & Parchomovsky, supra note , at 1711–12. That potential consequence of limited information is independent of the information-costs concern raised here.
  163. .See supra section I(A)(1).
  164. .See supra section III(A)(1).
  165. .Unfortunately that presumption is necessarily speculative. Surprisingly, no tally of the frequency of eminent domain seizures appears to exist. See U.S. Gov’t Accountability Office, GAO-07-28, Eminent Domain: Information About Its Uses and Effect on Property Owners and Communities Is Limited 8 (2006), http://www.gao.gov/new.items/d0728.pdf [https://perma.cc/GD77-MJYX] (“[W]e were unable to determine the number of times and the purposes for which eminent domain has been used across the nation because of a lack of centralized or aggregate data.”).
  166. .Shavell suggests that governments might generally be inclined to take property less often than is optimal, but he advances that suggestion as an argument against the “fiscal illusion” concern about excessive takings. His discussion does not consider the possibility that removing compensation might amplify any such tendency to take too infrequently. Shavell, supra note , at 129–30.
  167. .Calandrillo notes this possibility in passing but seems to discount its significance for unstated reasons. Calandrillo, supra note , at 515 n.331; see also Farber, supra note 17, at 295 (“The same is true of insurance companies, which could be expected to lobby hard against the losses covered by their policies. If anything, these groups might be even more effective as lobbyists than the individual property owners, since they are repeat players.”).
  168. .See, e.g., Farber, supra note 17, at 289 (“If public choice has any one key finding, it is that small groups with high stakes have a disproportionately great influence on the political process.”); Saul Levmore, The Public Choice Threat, 67 U. Chi. L. Rev. 941, 948 (2000) (reviewing Robert D. Cooter, The Strategic Constitution (2000) and Dennis C. Mueller, Constitutional Democracy (1996)) (explaining “public choice” and “special interest” models in relation to legislation); see also Dagan, supra note 51, at 754 (suggesting that “strong potentially injured parties exert more pressure on the public authority than people who belong to marginal groups or to the nonorganized public”).
  169. .See, e.g., Farber, supra note 17, at 284 (arguing that “[p]rivate insurers might in fact be more capable of taking precautions against takings than individual citizens” because they are better able to “overcome free-rider and other organizational problems”). Lobbying might not be the limit of insurance companies’ actions. Abraham Bell and Gideon Parchomovsky suggest that if private insurance replaced public compensation for takings, then “[i]nsurance companies who stand to lose from certain takings might exert improper influence on governmental decisions, by means of political contributions or bribes, in order to increase their revenues or minimize their losses.” Abraham Bell & Gideon Parchomovsky, Takings Reassessed, 87 Va. L. Rev. 277, 310 (2001).
  170. .See James W. Ely, Jr., The Guardian of Every Other Right: A Constitutional History of Property Rights 76 (2d ed. 1998) (noting that “state governments . . . aggressively us[ed] eminent domain power to promote transportation projects” in the nineteenth century). These delegations occurred even before the Constitution’s enactment. See id. (“As early as 1786, South Carolina conferred the power of eminent domain on the Santee Canal Company to obtain land and materials for the construction of the canal.”); see also Boom Co. v. Patterson, 98 U.S. 403, 406 (1879) (“The property may be appropriated by an act of the legislature, or the power of appropriating it may be delegated to private corporations, to be exercised by them in the execution of works in which the public is interested.”).
  171. .See, e.g., 1A Nichols on Eminent Domain § 3.03[3][b][i] (2018) (detailing the delegation of eminent domain power to “public service corporations”).
  172. .545 U.S. 469 (2005).
  173. .Id. at 475, 488. For an overview of the many private enterprises for which eminent domain may be used, see 2A Nichols on Eminent Domain § 7.07 (2018).
  174. .See Robert Brauneis, The First Constitutional Tort: The Remedial Revolution in Nineteenth-Century State Just Compensation Law, 52 Vand. L. Rev. 57, 107 (1999) (reporting that a “large portion of nineteenth-century just compensation litigation . . . involved corporate defendants. Everyone expected that turnpike, canal, and railroad corporations, and usually municipal corporations as well, would bear the costs of property acquired for their benefit”).
  175. .Eliminating a business’s obligation to compensate for property that it takes would decrease the cost to the business of the project’s inputs, and thus would increase the project’s accounting profitability. However, that elimination would have no effect on the total amount of wealth created by the project—only its distribution—and thus would not affect its economic profitability (i.e., its social efficiency). For a textbook discussion of the distinction between accounting profit and economic profit, see Paul Heyne et al., The Economic Way of Thinking 166–68 (11th ed. 2006).
  176. .More precisely, they would have an even greater incentive than they already have to profit from the use of eminent domain as a substitute for negotiated purchases. Even when compensation is owed, that incentive may exist. See Daniel B. Kelly, Strategic Spillovers, 111 Colum. L. Rev. 1641, 1717 (2011) (noting that “it is often less expensive for an assembler to convince a local government to exercise eminent domain on its behalf than to purchase the parcels in the real estate market”).
  177. .One might try to avoid this result by eliminating the compensation requirement only for takings that the government does directly, rather than takings by private entities to whom the government has delegated its taking power. However, neither the moral hazard argument nor the administrative-costs argument for eliminating the compensation requirement draw any distinction between direct takings and delegated takings. Both types of taking have the same effect on incentives to improve property and both incur the same administrative costs. Hence, this response would require substantial revision of the anti-compensation thesis.
  178. .See Richard L. Hasen, Lobbying, Rent-Seeking, and the Constitution, 64 Stan. L. Rev. 191, 228–34 (2012) (summarizing the social costs of lobbying).