What’s (Fundamentally) Wrong with Privatization?
What’s (Fundamentally) Wrong
with Privatization?
Reclaiming the Public. By Avihay Dorfman and Alon Harel.* Cambridge University Press. 2024. Pp. x + 197. $120.00 (hardcover), $39.99 (paperback & e-book).
The Privatized State. By Chiara Cordelli.** Princeton University Press. 2020. Pp. 346. $47.00 (hardcover), $32.00 (paperback & e-book).
Alexander Volokh[1]
Privatization is unpopular in the contemporary academy, bound up with the contested term “neoliberalism” and the supposed death throes of government since Reagan-Thatcher times. This is doubly so when it comes to the “privatization of force”—the prime example being private prisons.
Much of the critique is in the terms of conventional policy analysis: welfare scholars explain the perverse effects of welfare privatization; prison scholars critique prison privatization; international law or national security scholars decry the abuses of Blackwater. Correct or not, these critiques are fact-based and empirical, and they tend to be instrumentalist, treating public or private status as significant only insofar as it affects some other value, like decent treatment of inmates or beneficiaries.
Not everyone is thrilled with this instrumentalist focus. If welfare privatization is bad because a study shows that welfare delivery got worse, what about the next study? Couldn’t we pick a better contractor? If prison privatization is bad because the contract gave the private provider perverse incentives, couldn’t cleverer privatizers have written smarter contracts that alleviated those incentives? If nobody’s monitoring the contractors, couldn’t we invest more in monitoring and terminate the bad guys’ contracts? Of course, contractors and contracts aren’t perfect—but neither is government provision. So, isn’t this just a contingent matter of comparative institutional analysis?
Some would accept this, and even embrace it: Yes, it’s all contingent; let’s learn from our mistakes and do it better next time; “mend it, don’t end it.” Others, while accepting comparativist premises, think the negative empirics are well-enough established, and the reasons for those negative empirics are strong enough, that they’re comfortable opposing privatization, confident that more studies won’t upend their beliefs.
But still others argue that privatization’s problems run deeper. Perhaps, even if narrow “policy considerations” were a wash—supposing, arguendo, that you could construct a private system producing the same (or better) bottom-line results as a public system—the private system would still be unjust or illegitimate.
Hence the quest for a fundamental, noninstrumentalist critique of privatization, a theory of what’s distinctively good about the delivery of certain services by public actors rather than private ones. With such a critique, privatization opponents needn’t fear the next study, the next innovation in contracting, or the next wave of contract monitors.
But good fundamental critiques have been elusive. Scratch a supposedly fundamental critique hard enough, and one often finds that it’s not really a critique of privatization but rather a critique of some other feature that’s merely contingently associated with privatization; or it makes empirical judgments about the nature of privatization that turn out to be contested and possibly inaccurate; or it makes casual generalizations about the supposed essence of the public and private sectors; or it defines the terms “public” and “private” in tendentious ways that don’t match how they’re used in common language or actual privatization debates. In short, many supposedly fundamental critiques of privatization turn out to be either non-fundamental or not really about privatization.
Enter the philosophers. The last couple of years have seen two major scholarly works promising the long-hoped-for fundamental critique. Chiara Cordelli, a political philosopher from the University of Chicago, has published The Privatized State; and Avihay Dorfman and Alon Harel, legal scholars from the faculties of law at (respectively) Tel Aviv University and Hebrew University of Jerusalem, have published Reclaiming the Public. These are, so far, the most interesting and sophisticated efforts toward a noninstrumentalist critique of privatization.
Unfortunately, most legal academics aren’t current with the political philosophy literature. Also, both books are (broadly speaking) in the Kantian tradition, which has advantages and disadvantages. The disadvantage is that U.S. constitutionalism is broadly Lockean, not Kantian, and Kantian analysis tends to be unfamiliar to U.S. legal audiences. The advantage is that a Kantian orientation might be a more promising avenue for developing fundamental objections to privatization.
Why is that? First, let’s understand what the Lockean and Kantian orientations are. Dorfman and Harel explain the broad difference in a footnote (Dorfman & Harel, p. 17 n.3):
While Lockeans believe that the state is contingently desirable to guarantee liberty, they take liberty to exist independently of the state, so that the state is merely an instrument to bring it about. Kantians, by contrast, hold that the state is necessary for the protection of liberty.[2]
Cordelli explains the Lockean instrumental approach more explicitly:
The Lockean . . . assumes that . . . the demands of justice, including respect for independence, can in principle be fulfilled independently of the existence of any shared institution, and that . . . there is no definite proof that public institutions are better means than autonomous private action to fulfill those demands. From this, the Lockean draws the conclusion that . . . we are permitted, compatibly with the requirements of justice, not to support any public institutional arrangement, let alone a full state system (Cordelli, p. 48).
Lockeanism, in other words, endorses the “interchangeability assumption”—the assumption that any “service or function . . . can, in principle, always be performed by either private or public entities and that the choice of an agent to perform the function must be based on addressing the question of who is more capable of performing this function” (Dorfman & Harel, pp. 94–95; see also Cordelli, p. 46).
It should be no surprise, then, that Lockeanism would have trouble explaining why any institutional arrangement (like privatization or in-house provision) is necessarily good or bad from a justice or legitimacy perspective. This is in line with the American constitutional tradition. American constitutional prison litigation, for example, follows a simple approach: because private prison firms perform the “traditionally exclusive public function” of incarceration, courts unanimously agree (under the State Action Doctrine) that their inmates get all the same constitutional rights as they would if the prisons were public.[3] And, the private prison “problem” having been “solved” through this form of constitutional regulation, the contractors’ private status no longer plays any significant role in the analysis.[4] The courts care whether the inmates’ constitutional rights are respected, regardless of who’s doing the respecting. Clearly, this approach is unfriendly to arguments that something’s wrong with private provision as such.
So, if we want fundamental, noninstrumental objections to privatization, jettisoning the interchangeability assumption, Kantian-style, seems like a must.
These considerations don’t often show up in U.S. legal thinking, so there’s a risk that these important contributions will be missed by those who take an interest in privatization.
I’m a legal scholar, not a professional philosopher; I also come from a Lockean, instrumentalist perspective. But having taken an interest in these philosophical issues for the last 20 years, I hope to bridge the gap here. I don’t find either of these books compelling: as I explain in this review, they give short shrift to the possibilities of contractually or statutorily importing accountability mechanisms into privatization arrangements; they take the failures of past instances of privatization to be signs of inherent failure rather than bad implementation; they make unwarranted assumptions about the fundamental nature of private firms or the inherent logic of privatization; and they define “public” and “private” in ways that don’t track how the terms are used in public discourse. In other words—like previous arguments—these arguments are either non-fundamental or not really about privatization. But these authors make their case far better than their predecessors, and privatization scholars should actively discuss their arguments.
I start, in Part II, by describing what a successful noninstrumental critique of privatization should look like. I go on, in Part III, to describe Dorfman and Harel’s book, and, in Part IV, to explain why it fails to provide a compelling critique of privatization. Then, in Part V, I explore whether Cordelli’s book does any better, and provide my critique of her approach as well. I conclude in Part VI.
Let’s think about what a noninstrumentalist anti-privatization argument would look like.
Initially—to state the obvious—an argument about privatization must be truly about privatization. We should be able to agree on what counts and doesn’t count as “privatization” (or “private,” or “public”), or else I’ll point to a supposed good example of privatization and you’ll deny that it’s privatization.
Here’s my definition, which I claim is also the definition that’s in common usage in privatization debates. People are private if they’re outside the formal organization of government; organizations or firms are private if they’re created by private people and operate outside the formal organization of government. We shouldn’t assume that private people or organizations are motivated by profit or any other particular thing—people can be motivated by whatever they like, and they can run organizations for whatever purpose they like. Firms can be for-profit, not-for-profit, religious, secular, or anything else.
If a private person contracts with government without becoming a government employee (under conventional understandings of the employment relationship), or if such a firm or organization contracts with government without becoming a government agency (under the prevailing law of government agencies), we can call that “privatization” or “contracting out.” The concept of privatization depends on conventional understandings of the public-private distinction; in a world without such a distinction, it might not be meaningful.
The privatization debate, then, concerns whether it’s just, legitimate, permissible, or otherwise desirable for the government to provide services by contracting with private persons, firms, or organizations rather than providing those services through its own employees or agencies. You can define “privatization” in some unusual way and argue against that, but then your argument is about your idiosyncratic understanding of the concept; it doesn’t necessarily translate into an argument against privatization as other people understand it, and it won’t necessarily contribute to the privatization debate.
To further state the obvious, noninstrumental arguments against privatization must be noninstrumental.
If someone argued against prison privatization on the ground that it didn’t save money and reduced the quality of confinement,[5] we would recognize this argument as instrumental: privatization is bad because of its effect on something else. Similarly if someone argued against prison privatization on the ground that private prison firms would lobby for more incarceration,[6] or on the ground that private prisons do save money and would therefore wrongly make us lock up more people.[7] Correct or not, these arguments aren’t noninstrumental; they don’t say anything is wrong about public or private status as such.
What if the claim is merely that privatization makes effect X extremely likely, and that X is bad? This is an argument against X, but not a noninstrumental argument against privatization. But what if the association is more than a mere contingency? What if it’s very difficult to eliminate X within the context of privatization—and what if the easiest way to eliminate X is to avoid privatizing?
If the “ifs” in the previous sentence are solid, maybe this comes close enough: there might be noninstrumental reasons to oppose X, and opposition to privatization is “merely” instrumental to eliminating X, but one’s noninstrumental opposition to X basically implies opposition to privatization.
But those “ifs” had better be solid; the association had better be truly robust. Often, though, they are anything but.
- The Essentialist Fallacy
For instance, when the Israeli Supreme Court struck down a prison privatization statute in 2009,[8] it identified as a “critical question” whether the party doing the incarceration is “mainly motivated” by “the public interest” or “a private interest.”[9] Because, it said, private prison firms are “motivated by economic considerations,” that was enough to make the whole enterprise a violation of human rights—even if the inmates were treated identically to public prison inmates.[10]
By contrast, the Israel Prison Service is a “bod[y] that answer[s] to,” “receives its orders from,” “is subordinate to,” “acts through” and “by and on behalf of,” and is a “competent organ[] of” the state or the government or the executive branch—which, in turn, is “the representative of the public.”[11]
But all of this is merely an appeal to the supposed essence of the public vs. private sectors. The public sector is pure and public-interested, even though public actors may fall short of the ideal and care about nothing but their paycheck or their private goals. The private sector, even in its ideal, is about nothing but profit—even if, in reality, a contractor might be run by former corrections officials who care about doing a good job, and even if its employees are in fact less mercenary-minded than their public counterparts. If this is an empirical judgment, it’s not supported by any analysis or argument about whether the empirics are robust. And if it’s not empirical, why are these presumptions about the public and private sectors warranted? Call this the Essentialist Fallacy.
- The Privatization History Fallacy
Or consider the argument from historical practice. Let’s suppose that prison privatization has always led to bad conditions. Would that be enough to call the empirical relationship robust, so that one is justified in opposing privatization based on noninstrumental opposition to bad conditions?
Maybe, maybe not. What if the problem is that the previous privatizers just didn’t care? What if you could do it better if you were in charge? Then historical practice doesn’t support opposition to privatization as such. On the contrary, it invites one to conditionally endorse privatization as long as it’s carried out by a Nice Caring Government that writes better contractual terms and invests in monitoring; if our guys were in power, we’d do it better. Even if our guys are never going to get into power, that doesn’t contradict our conditional endorsement; it just means the condition behind the endorsement is unlikely to be fulfilled. Call this the Privatization History Fallacy.
- The Clever Privatizer Principle
Let’s go a bit further. Whenever privatization critics point to past privatization failures, they should imagine a clever privatization advocate sitting next to them, planning how to solve the problem in future rounds of privatization. One should count on this clever advocate’s being truly clever, and imagining new contractual terms that alleviate past perverse incentives or prohibit past bad practices.
Are these new contractual terms unheard-of? Perhaps, but no matter. Maybe the only reason they’re unheard-of is that no one with any political power has previously insisted on such terms. The absence of such terms in Real Life might be a good reason for an ordinary voter to oppose actual privatization proposals—because you shouldn’t endorse bad proposals based on dreams that they might be better. But they’re not a good reason for philosophers to oppose privatization in principle—because, as before, philosophers are always free to conditionally endorse. Call this the Clever Privatizer Principle.
- The “Regulation By Contract” Principle
One thing a clever privatizer can do is write desired terms into a contract. It’s common for privatization critics to complain that private contractors aren’t subject to the public-law accountability standards that apply to public agencies.[12] So why not just extend public-law norms to private firms—by contract, by statute, or by judicial decision?[13] This objection to privatization isn’t an inherent objection—rather, it’s an objection to certain kinds of privatization, which is another way of saying an insistence on the “right kind” of privatization.
Firms can agree to all sorts of things in their private contracts with other people or firms, and the same is true when their contractual partner is the government. In a capitalist economy, one can find someone who’ll agree to any contractual condition if the price is right; and if the price is wrong, they don’t need to accept the contract.
This point is broader than mere freedom of contract. It takes two to tango. One side may want to negotiate, but the other side is free to insist on non-negotiable terms. The government can demand public-law accountability standards, or any other contractual terms; or the legislature can make such standards a requirement of any government contract for particular functions; or the legislature could simply regulate firms engaging in particular functions.[14]
Perhaps regulating a fully private industry might be thought to reduce the freedom of the participating firms and/or their customers, in which case legitimacy will have some cost—assuming, of course, that such regulations are really necessary to produce legitimacy. But when we’re talking about whether to contract out the provision of a government service, private firms have no preexisting freedom in that area, and so no contractual requirement can be properly considered contrary to the firms’ freedom. Indeed, a firm’s agreement to any contractual condition is an exercise of its contractual freedom, not a violation of it.
Call this the “Regulation by Contract” Principle; and to the extent that firms desire to enter into such contracts, call this the “Freedom of Contract” Principle.
With all this as background, let’s look at the argument of Dorfman and Harel’s book.
In the Introduction, the authors set the stage for their philosophical approach. The value of an institution lies not only in what it does, but also in who does it. Some people can only act for us, but for certain functions, we can’t be free and equal unless we’re truly the authors of the rules that govern us, rather than merely passive beneficiaries or subjects; the actors who generate the rules that govern us must be acting in our name. For those functions, representation is important. Representation requires, in the first place, perspectivism—public officials’ decisions must defer to (i.e., reflect or be consistent with) our perspectives—and, in the second place, attributability—their decisions must be attributable to us, i.e., we can be held responsible for them. No individual can satisfy these conditions, but truly public institutions can.
Chapters 1 and 2 lay out the authors’ basic political theory.
In Chapter 1, “A Public Conception of Political Authority,” the authors give a theory of legitimate political authority.[15] Political authority claims the power to change people’s normative situation, and it can’t be legitimate unless it’s public, in the sense described above—it must represent us, speaking and acting “in our name.” A political authority that speaks “for us” is hierarchical and tyrannical; legitimate political authority is necessarily nonhierarchical. “Representation” doesn’t necessarily require democratic procedures or actual consent; rather, an authority is representative if it endorses its subjects’ comprehensive worldview. Because the authority replicates the citizens’ worldview, its decisions are attributable to the citizens, which means that the citizens are responsible for those decisions. Representativeness and attributability—what the authors call “public”—are necessary conditions for political authority to be legitimate. Legitimate political authority is thus freedom-facilitating rather than freedom-limiting, because the rules that might seem to restrict our freedom are actually our rules, which (in a sense) we ourselves have authored.
Chapter 2, “Law as Standing,” extends this analysis to law. Why does law make a moral difference—why might it matter whether an act is illegal? Law makes a difference to the extent that it’s established by someone with standing to establish binding norms. Free and equal private persons lack standing to dictate norms for others; only someone public, as described above, has standing to dictate norms for us, because such a person is basically us and the norms he establishes are basically authored by us. A legitimate lawmaker must be detached—he must not act in his own name—and must also be representative—he must act in ours. Recall that representativeness, as above, requires that the lawmaker have “a sufficiently tight deferential relation” with the people he’s binding.
Chapter 3, “Speaking in a Different Voice: The Necessity of Institutional Pluralism,” goes into greater detail on the different ways that norms can be established. Dorfman and Harel argue that “different institutions can [establish norms] in different ways; even identical laws may have different meanings and significance depending on their institutional source.” They distinguish between statutory norms, common-law norms, and constitutional norms—even when the norm is the same, the value produced by that norm might depend on which institution provides it.
For instance, if a norm like free speech or marriage equality is merely statutorily protected, that means the right matters because the majority thinks so; but if such a norm is protected constitutionally, that means it matters independently of people’s judgment. And if the norm is created by judge-made law, that means it’s “the outcome of adjudicative deliberation and legal reasoning, resting on values such as reasonableness and coherence with the legal system and its values as a whole.” Protecting a right in the wrong way inflicts a dignitary injury, and so the authors advocate that even someone whose right has been protected (but in the wrong way) should have standing to sue and obtain an institutional remedy, where a court would proclaim that a particular right is, say, constitutional rather than statutory. It also follows that constitutions should be transgenerational and not amendable in ordinary ways.
Chapters 4 and 5, drawing on previous articles by Dorfman and Harel,[16] lay out the core of their anti-privatization arguments.
Chapter 4 introduces the concept of “Inherently Public Goods.” A good that is inherently public can’t be provided by a private party, even in principle: because its value depends on its public provision, a private attempt to provide the good would simply provide a different (and possibly illegitimate) good.
Dorfman and Harel distinguish between two types of fidelity that an agent can exhibit in carrying out tasks: fidelity by reason and fidelity by deference. An agent exhibiting fidelity by reason executes the task with respect to his own point of view; such an agent may execute the task competently, but his acts are his own, not those of the state. An agent exhibiting fidelity by deference, on the other hand, is deferential to the polity’s perspective, which makes his acts those of the state.
But isn’t fidelity by deference impractical for tasks that involve pervasive discretion and the need to use judgment—wouldn’t that require minute-by-minute consultation? Fortunately, constant consultation isn’t necessary. All that’s necessary for fidelity by deference is that the agents have the status of “public officials,” which involves two conditions. First, the agent must defer to a community of practice that he belongs to, “a community that collectively determines what the public interest dictates.” (The agents themselves determine what their own deference requires, but it needs to be a collective practice.) And this practice must also have an integrative form; that is, it must integrate the political and the bureaucratic and have “principled openness to ongoing political guidance and intervention.” Participants in the practice can then be called “public officials” and their practice is “deferential.”
Which goods, though, are inherently public? Dorfman and Harel don’t offer a complete definition, but they give examples. Punishment is inherently public; punishing a wrongdoer is an expressive and communicative act of condemnation—only possible if the condemnation emanates from the appropriate agent. And war—at least, if justified as promoting a legitimate state interest like self-defense—is a quintessential expression of political sovereignty, and every action in war must be attributable to the sovereign. Beyond that, they say, particular goods must be examined individually to see if their value hinges on public provision. As to these inherently public goods, privatization—any provision that’s not “public” in the above sense, involving a community of practice and an integrative form—cuts the political community off from the provision of the good, and therefore provision of the good is conceptually impossible.
Chapter 5, “Against Privatization as Such,” extends the critique even to non-inherently-public goods. The previous chapter argued that public institutions require the possibility for direct involvement of politicians. Because citizens should have a nontrivial degree of influence over politicians’ decisions, this also means that, in public institutions, citizens are indirectly involved in public officials’ decisions. By contrast, when delivery of goods is in private parties’ hands, this cuts out the involvement of politicians and therefore the involvement of citizens—the private providers’ actions aren’t (as they would be in the public case) the actions of the citizens themselves. The polity becomes detached from these decisions; this is a loss of political engagement. And this is at least sometimes undesirable: even for non-inherently-public goods, “privatization reduces the political dimension of responsibility by partially obviating the distinctive role of collective undertaking in discharging the responsibility persons have by virtue of being citizens.” This loss of civic responsibility doesn’t necessarily forbid privatization, but it’s a relevant cost, and these aggregate costs of widespread privatization wouldn’t show up in an activity-by-activity cost-benefit analysis.
Chapters 6 and 7 apply this theory to two particular policy contexts: public property and artificial intelligence.
I’ll focus on three critiques of Dorfman and Harel’s book.[17] The first two, which I discuss in Section A, are about their political theory generally. First, is it even possible for agents to “speak in our name” in the way that they suggest? Second, there are many cases in society where agents seem to not speak in our name, even though inherently public goods seem implicated. Do these challenge Dorfman and Harel’s theory?
The third critique is specific to privatization, so I leave it to Section B. Suppose Dorfman and Harel are right about the desirability and necessity of “speaking in our name.” Why would that rule out privatization—why can’t private parties speak in our name?
In Dorfman and Harel’s theory, the state or officials can’t act in our name unless they’re representative. Assuming one accepts this view, it all hinges on whether this concept of “representation” is attractive or realistic or even feasible.
First, consider one-on-one representation. Dorfman and Harel defend perspectivism: A person (an “authority” or “representative”) can be said to represent another (a “citizen” or “subject” or the “represented”) if they commit themselves to look at the world from the citizen’s perspective (Dorfman & Harel, p. 23), i.e., to defer to that perspective (Dorfman & Harel, p. 24). Deference “involves the willingness to substitute the [representative’s] judgments and/or worldviews and/or essential features with those of the [represented]. Such a deferential stance is a form of recognizing the actual features of the represented person’s self as having a controlling influence on the deliberations of the representative” (Dorfman & Harel, pp. 24–25).
If deference is present, the authority is making decisions not for the citizen but in the citizen’s name. Such decisions can be attributed to the citizen—“in reality, it is the represented who made the decision” (Dorfman & Harel, p. 23).
What does “attributability” mean?
Attributability implies that under the appropriate conditions a citizen can be held accountable for the decision as if it is hers or his, although she or he has not made it. Attributability does not entail that a citizen can be blamed or prosecuted for the acts of her or his government. It does entail, however, that she or he ought to take some responsibility for the decision and that she or he cannot remain indifferent to it (Dorfman & Harel, p. 23).
Representation requires more than just deference; it also “requires adopting a decision-making process capable of accurately identifying and articulating the point of view of the represented person” (Dorfman & Harel, p. 25). (This might or might not involve democracy (Dorfman & Harel, p. 27).) “To the extent that the authority succeeds in representing, the subjects can justifiably claim that they are the genuine authors of the resulting decisions” (Dorfman & Harel, p. 25). The result is thus a “nonhierarchical” relationship between authority and subject.
One could accept the possibility of one-to-one representation but question whether this can be generalized to groups. But is this even plausible in the one-to-one case? If someone buys me a present, we say they’re buying the present for me. But what if they adopt a deferential stance, honestly committing to buying what they think I would like? They’re still buying the present for me. Perhaps we’ve adopted a process that can identify and articulate my point of view—maybe I gave them detailed information about my views (though I couldn’t tell them exactly what to buy because I didn’t know what was available). They might still choose wrongly, because they can’t get inside my head. Unless their choice is exactly what I would have chosen, I wouldn’t say that I was a genuine author of the agent’s decision.
Does this undermine the law of agency? Can I never be held responsible for another’s decisions? Not at all: I’m rightly held responsible for my agents’ decisions (e.g., in tort law), not because they’re my decisions, but because it’s instrumentally useful for the sake of better incentives or better victim compensation.
Now Dorfman and Harel might argue that represent-ativeness is a matter of degree, and perhaps I’m more of an author of the agent’s decision, to the extent that the agent is better able to implement my views. So perhaps some degree of one-to-one representation is feasible in principle, regardless of how likely successful representation might be in practice. But the prospects of representativeness in the one-to-many context seem far worse, simply because the represented class is heterogeneous. Dorfman and Harel recognize the difficulty: “What happens when the relevant features which provide the basis for representation differ among individuals? What if, as is to be expected, the preferences, judgments, and identities of different individuals radically differ?” (Dorfman & Harel, p. 29).
Indeed: people are so heterogeneous that deferring to their perspectives, worldviews, or identities is impossible; representation (in the “in their name” sense) is impossible; and so either legitimate political authority is impossible . . . or we need a different account of legitimate authority. Instead, Dorfman and Harel seek “(at least partially) to bridge the gap” by appealing to a “modest, holistic, and (partially) proceduralist” account (Dorfman & Harel, p. 29). As to modesty, this involves “lowering the bar of what counts as a legitimate authority” so that it’s enough to “reach a certain threshold,” i.e., be “sufficiently representative” (Dorfman & Harel, p. 29). Similarly, as to holism, people might be “adequately represented” even if they’re not represented on particular decisions. And as to proceduralism, people might have a shared perspective on procedures even if they don’t converge on substantive issues.
But the more one retreats from the “near-perfect degree of convergence and precision that a one-to-one representation could achieve” (Dorfman & Harel, p. 30)—already an exaggeration—the less plausible it is to say that deference is possible and thus that citizens are the authors of the authority’s decisions. Representative government is always for, never in their name.
- The Challenge of Non-Integrative Practices
But suppose we accept that deference is both possible and (at least sometimes) required. We might still ask: Can anyone then become a legitimate official—that is, “public”—simply by choosing to adopt a deferential stance and committing to be guided by the polity’s judgments? No:
A person cannot merely approach the performance of the task at stake from the point of view of the state—there is no such ready-made perspective lying out there. The reason that the government cannot turn a willing individual into its agent simply by asking the individual to perform “a task” is that the tasks dictated by the state are typically underspecified, such that they leave broad margins of discretion (Dorfman & Harel, p. 105).
Dorfman and Harel argue that two features are needed for officials to be able to be deferential toward the polity. First, there must be a “community of practice” in which officials “immerse themselves together in formulating, articulating, and shaping a shared perspective” (Dorfman & Harel, p. 107). And second, that practice must have an “integrative” form: it “must be able to integrate the political offices” and “be open to the possibility that politicians change the practice, guide its mode of operation, and reevaluate the norms governing it” (Dorfman & Harel, pp. 108–09).
One challenge, though, is that many political arrangements don’t have an integrative form—so they’re presumably not “public” because they can’t be deferential. (This would imply illegitimacy in the case of inherently public goods discussed in Chapter 4, or at least a limited undesirability in the aggregative sense discussed in Chapter 5.) I discuss four such arrangements below: (1) the private use of force, (2) apolitical agencies (“independent commissions” in U.S. administrative practice), (3) pure adjudication, especially involving juries, and (4) judicial common-law lawmaking and constitutional lawmaking. All of these lack, at the very least, integration of public officials into decisionmaking processes.
One possible response is the radical one—these arrangements are all illegitimate. But Dorfman and Harel don’t reject these arrangements, and even embrace some. Another response is that they’re legitimate because they relate to activities that aren’t inherently public. But they don’t argue that either—rightly so, because some of these arrangements clearly relate to sovereignty. A third response is to claim that they’re legitimate and produce an extension to the model that would cover these cases. This is the approach they take for some of these arrangements—but they don’t adequately explain how to justify them without undermining the rest of their framework.
- Private Use of Force
Dorfman and Harel distinguish “public” from “private lawmaking” (Dorfman & Harel, p. 44):
Contractual parties, private owners (et alia) are vested with normative powers to create rights and duties for themselves as well as for others, including in the case of ownership for nonconsenting others. For instance, a landowner is ordinarily authorized to decide who can enter the land and who cannot. Such instances of private legislation are not relevant to addressing our concern with the standing to call the demands of morality into law. This is because, all else being equal, private legislation is not an instance of private persons purporting to act as sovereigns, by which we mean to speak and act in the name of us all and in respect of an interest common to us all. Our interest is exclusively in the claims of sovereigns (Dorfman & Harel, p. 44).
This appears plausible. Dorfman and Harel’s theory isn’t about all uses of force. It’s primarily about “inherently public goods”—“those goods that cannot be realized unless state institutions provide them” (Dorfman & Harel, p. 114); not all goods are inherently public.
But is it sufficient to note that “private persons [aren’t] purporting to act as sovereigns”? Private persons can determine authoritatively who can and can’t enter property only because the law allows them to. The law allows private persons not only to unilaterally affect others’ rights but also to back this up (e.g., repel trespasses) with force. We have many other coercive rights. We can defend ourselves from attack, sometimes with deadly force—this is criticized by those who believe such force should be reserved to public authorities. We have rights to “imprison” people—carry out citizen’s arrests in ways that would otherwise constitute “false imprisonment,” and merchants have a limited “shopkeeper’s privilege” to detain suspected shoplifters. When people complain about private policing, they’re often unaware that, usually, private police forces are merely efficiently using rights that everyone already has.
We can imagine a system where determining and enforcing these entitlements is exclusively the province of public authorities acting as sovereigns. The whole system of private property and private enforcement is a way of privatizing force. If sovereign acts must be “public” and can’t be privatized, why aren’t these private rights—long recognized by common law or statute—vulnerable?
Perhaps there’s a difference between prevention and punishment; perhaps punishment is inherently public and prevention isn’t, which would open the door to private preventative (though not punitive) force. The trouble is that Dorfman and Harel aren’t completely clear on what is inherently public; they give a high-level definition and some examples. The top examples are criminal punishment (Dorfman & Harel, pp. 115–18) and war (Dorfman & Harel, pp. 119–22), though, even for war, they limit their discussion to “wars that are justified on the grounds that they promote a legitimate state interest, such as the case of waging a war in self-defense”—excluding “wars that are grounded in state-independent ends,” which “can (and, perhaps, must) be fought regardless of the identity of the agent who acts for the sake of these ends—the paradigmatic case being wars justified by reference to the demands of humanitarian intervention” (Dorfman & Harel, pp. 119 n.55).
They “do not argue that the infliction of political violence in furtherance of public purposes exhausts the entire range of activities whose resulting goods are inherently public”—“determining which function falls in the category of inherently public goods depends on the nature of the relevant good and whether its successful provision requires that it be publicly provided, that is, provided by public entities” (Dorfman & Harel, p. 114). In a footnote, they suggest (justifying the doctrine of numerus clausus) that creating new types of property rights is something that only the state can provide (Dorfman & Harel, p. 114 n.44).
But it’s not obvious why letting private parties use force to protect their entitlements isn’t inherently public—which means their critique of privatization might extend further (maybe a lot further) than they intend.
- Independent Agencies
A second issue concerns “apolitical” public agencies. Dorfman and Harel give the example of “an independent election committee in both post-authoritarian and democratic states authorized to enforce campaign finance laws, redraw election districts, and ensure the integrity of the election process more generally” (Dorfman & Harel, p. 139). In the United States, creating apolitical, technocratic agencies that would be shielded from popular pressure was a key progressive goal—we now have multi-member “independent commissions” like the Federal Trade Commission or the Securities and Exchange Commission, whose members are insulated from presidential removal during their terms (though the constitutional status of such agencies is currently shaky). These agencies do the same rulemaking, adjudication, and enforcement as traditional agencies, and clearly they exercise sovereign power. But they lack “integrative form” and thus can’t be said to exercise “deference,” so they violate the definition of “public.” Are they “private” and therefore illegitimate?
On the contrary, Dorfman and Harel firmly commit to their legitimacy, even calling them “public” and their members “public officials” (Dorfman & Harel, p. 140). This is immensely important, so it’s worth quoting their discussion at length:
The insulation of a public institution reflects the polity’s choice to relieve the institution’s agents of the requirement to defer to political officials. The judgment that underlies this choice is that the general interest is sometimes best served not by way of politicians dictating the decisions and actions that participants in the particular practice ought to make, but rather by creating a sufficiently wide arena of permissibility within which the participants can decide, by themselves, what decisions and actions would be best for the polity. Depending on the relevant context, there may be any number of reasons for enlisting the discretionary powers of bureaucrats and experts at the expense of political judgment—for instance, the subject matter of the activity requires special expertise, long-term reasoning, confidentiality, and so on. Furthermore, sometimes the exercise of political judgment may severely undermine the effective pursuit of the general interest up to the point where it would be appropriate to disintegrate the political/bureaucratic nexus. . . .
Granting greater autonomy to bureaucrats may be deemed necessary in cases in which there exist excesses of office politics (in the pejorative sense of the phrase) or populist tendencies (again, in the pejorative sense), making it too difficult for public officials adequately to manage their tasks and serve the public effectively. What is important to note, however, is that the resort to their discretion is sometimes the best, and perhaps the only, proxy for a bureaucrat or an expert to display fidelity of deference to the general interest (Dorfman & Harel, p. 140).
In other words, deference to political officials is fine as long as the polity thinks it’s a good idea; if the polity thinks deference won’t serve the public interest, it can create an agency without it. There may be “any number of reasons” for this—and legitimate reasons for eliminating political control could be nothing more than that such control is dysfunctional.
Deference, then, is a requirement of legitimacy . . . unless it’s not. Dorfman and Harel apparently give no limiting principle that would prevent all agencies from being this way. They suggest technocratic expertise as one of several factors, but, unfortunately, they allow for too many other factors, so that I’m unsure how far this exception applies and whether it swallows up the rule. The exception is potentially so large and so fundamental that it requires a lot more specificity if it’s to be a limited exception, rather than a concession that destroys the theory. Unless tightly limited, this exception conflicts directly with their insistence that publicness is necessary for legitimacy, that deference is necessary for publicness, and that integrative form is necessary for deference. Even if this serves the public interest, it should be seen as “tyrannical” under their view (Dorfman & Harel, p. 15).
What’s more, this concession seems to erect no barrier against privatization. Again, this depends on how far their “apolitical agencies” exception applies, and whether it might overlap with some of the areas where they critique privatization; perhaps in later work Dorfman and Harel might go into more detail on this. But more on that in Section B below.
- Pure Adjudication: Judges and Juries
Another tough case concerns the role of judges and juries in adjudication, whether private-law disputes or public-law disputes between private parties and the government (including criminal cases). Judicial lawmaking is discussed in the next subsection, so here let’s just consider factfinding (e.g., what happened or who’s telling the truth) or applying given norms to particular facts (e.g., whether the parties to a tort case were negligent).
In any lawsuit, some party is demanding relief—money or an injunction or a prison term. Nobody can obtain relief unilaterally; it requires judicially authorized coercive intervention. Ultimately, the court makes the defendant pay money or do something, or lets the defendant go. This is a sovereign act, which seems like it should be considered inherently public. If so, to be legitimate, it should be “public,” i.e., “representative” or “deferential.” But there’s no account here of the legitimacy of adjudication.
First, consider juries. There’s no obvious reason why juries should be excluded from Dorfman and Harel’s theory; the use of juries is a kind of privatization of sovereign power, like the private use of force discussed above. Not all systems have juries, and one might question why certain sovereign decisions should be delegated to private parties who are invited to apply their personal views.
Juries aren’t representative. They aren’t randomly chosen—once the bases for dismissal by the judge or the lawyers, including for personal knowledge or involvement or bias, are taken into account. Even if juries were random, a jury is a small sample, and any jury’s views may diverge substantially from the public’s. More important still, juries evaluating reasonableness are invited to apply their own views, and aren’t required or encouraged to defer to anyone else’s views.
Focusing on the specific factors of community of practice and integrative form, the jury is also lacking. As to community of practice, each juror deliberates with eleven others, but there’s no consistency with juries in other cases. And there’s also no “integrative form”—if politicians, or even the judge (outside of limited contexts), tried to affect jury decisionmaking, that would be “jury tampering.”
Let’s look now at judges, who at least look like public officials, even if not in the executive branch. Judges, by design insulated from popular pressure, don’t seem to defer to the views of the public at large, and such deference (beyond the sense of following the law) is typically considered undesirable. Among judges, there might be a “community of practice”—through common legal education, ongoing judicial training, and appellate review. But there’s no “integrative form,” because it’s considered improper for political officials to change the results of adjudication. (Perhaps the lawmakers’ control over the statutes is sufficient, even though this would only affect future adjudications. Or perhaps the lawmakers here are the judges themselves, so integrative form isn’t necessary for judges. But that still leaves the uncertain status of the jury.)
So at least juries seem to be carrying out sovereign functions without the requirements of publicness that are supposedly necessary for legitimacy and acting “in our name.”
- Lawmaking and Constitution-Making
An obvious tension in Dorfman and Harel’s theory arises in Chapter 3, where they discuss “the necessity of institutional pluralism” (Dorfman & Harel, pp. 64, 65 n.2). They grant that “lawmaking powers are not exhausted by legislative chambers, democratic or otherwise. Courts or other public institutions in some legal systems may qualify for the task” (Dorfman & Harel, p. 44).
The authors discuss how a norm could be established by different institutions—legislatures, common-law courts, or constitutions. The identity of the institution is significant, they say, because a statutory right isn’t identical to the same right protected constitutionally or through judicial lawmaking; statutory rights matter because of majority preferences, constitutional rights matter independently of majority preferences, and judge-made norms matter because of “adjudicative deliberation and legal reasoning” (Dorfman & Harel, pp. 73–74).
This is all reasonable, but can we imagine common-law or constitutional lawmaking being “representative”?
Consider, first, the lawmaking judge—say, a state Supreme Court Justice establishing common-law rights (Dorfman & Harel, pp. 84–85). “[A] common law judge must approach her or his lawmaking responsibility as (in part) an agent of justice, rather than a reflector of public opinion” (Dorfman & Harel, p. 73). Indeed, judicial decisionmaking processes aren’t “public” in Dorfman and Harel’s sense. This seems like a problem—isn’t judicial lawmaking, no less than legislative lawmaking, an inherently public good, so shouldn’t publicness be necessary?
Let’s assume that judges display the necessary “detachment,” identifying with their role and not their personal preferences (Dorfman & Harel, pp. 54–55). So they don’t act in their own name. But do they act in our name? The authors are clear that this doesn’t require democratic representation in a legislative sense (if it did, all federal judges, and many state judges, would fail that test); rather, the judges just need to “make decisions from our perspective,” which means there must be “a sufficiently tight connection between the decisions’ substance and what we want, judge to be just, or who we are” (Dorfman & Harel, p. 37).
Does that look at all like judicial lawmaking? Surely there’s no tight connection with “what we want”; on the contrary, judicial lawmaking “allow[s] for robust insulation from the choices and preferences of the entire population” and “create[s] an institutional space for the use of reason” (Dorfman & Harel, p. 86). And there’s no connection with “who we are”—that’s a reference to the “essentialist” view of representation, under which “the representative makes judgments that accord with the essential or natural features or identity of the represented” (Dorfman & Harel, p. 25). So presumably, if judicial lawmaking is legitimate, it must be because judges decide based on “what we . . . judge to be just.”
But note that I’ve italicized the word “we” in the previous quote. Dorfman and Harel praise judicial lawmaking for establishing “what new rights are dictated by reason, as mediated by legal analysis” (Dorfman & Harel, p. 87). “[J]udges reason from the past, including precedential decisions to the present case” (Dorfman & Harel, p. 87). When they “break new ground,” it’s “by arguing from prior recognition of fundamental principles, canonical cases, and influential dicta” (Dorfman & Harel, p. 87). It’s guided by the need for “coherence across legal domains” (Dorfman & Harel, p. 87).
But none of this is about “what we . . . judge to be just”—none of this can reasonably be called “representative,” and therefore “public”—unless we assume that we are also responsible for (or endorse) all those past precedents, dicta, and fundamental principles. (And even then, judicial lawmaking, even using traditional legal reasoning, has a lot of discretion and policy judgment, even if not as much as legislative lawmaking.)
Constitutional lawmaking seems even more problematic. Constitutions are also independent of popular will, unless we count the will of some subset of the founding generation. Given enough popular will, they can be amended, but some constitutions (especially the U.S. Constitution) are very hard to amend; most constitutional change has happened because courts have revised their views of what the constitution requires, not because the text has changed.
Among constitutional scholars, this poses a challenge. If democracy is so great, why can courts—staffed by unaccountable judges—strike down popular enactments? And even if there are good reasons to strike down popular enactments, why should the enactments be judged against a standard established by long-dead people who weren’t even representative of their own time?
These are the important questions of constitutional theory. The conventional wisdom is that it’s implausible to claim that “We the People” is actually us,[18] so the challenge is to explain why it’s legitimate that this constitution has been imposed on us. This seems to line up with Dorfman and Harel’s concern—that even laws that substantively promote freedom should be thought of as tyrannical (i.e., imposed on us, adopted “for us” rather than “in our name”) if adopted by someone without the necessary standing (i.e., representativeness) (Dorfman & Harel, p. 15). So one might think Dorfman and Harel would reject the idea of being subject
to constitutions adopted long ago and by others, or at least that they’d explain how this fits into their theory.
But they deal with the problem by, on the contrary, celebrating it. They reject Jefferson’s idea of having frequent constitutional change (which would come closer to having the constitution endorsed by the generation subject to it): his view that “it seems unfair to impose the choice of one generation on subsequent ones” is, they write, a “mistake” (Dorfman & Harel, p. 82). Constitutions “must adopt a transgenerational horizon”; “constitutional amendments should not be subject to the simple democratic procedures associated with statutory legislation” (Dorfman & Harel, p. 82), but instead should be subject to amendment processes that are “substantially more demanding” (Dorfman & Harel, p. 83). And this should also be true of exceptional statutes, like “constitutional statutes in the United Kingdom” (Dorfman & Harel, p. 83).
Rather than calling such a regime tyrannical, Dorfman and Harel call it liberty-enhancing, because it “provide[s] the public with participatory liberties—public autonomy, really—that could not have existed in their absence” (Dorfman & Harel, p. 83). “Ironically, such rights may expand our freedom in that they enable us to communicate our convictions concerning the different status and, in particular, the different grounds underlying different legal norms” (Dorfman & Harel, pp. 83–84).
This irony is real and needs explaining. If non-“public” lawmakers impose laws on the public when they have no standing to do so, it’s no defense that the laws are liberty-promoting; standing is an independent requirement. It’s hard to see how a constitution can be justified under their framework, even if it has the beneficial result of promoting “participatory liberties” and “public autonomy.” Even if constitutions can be said to promote not just liberty, but representativeness, in that they allow us to specify not only the content of norms but also the reasons behind them, this still needs to be reconciled with the loss of representativeness that comes from having to be subject to the norms contained in an ancient and hard-to-amend document with questionable democratic credentials.
* * *
In short, Dorfman and Harel’s political theory is problematic. It’s doubtful that the deference required for representativeness is even possible, in which case acting “in our name” is a mirage. But even if it were possible, the theory apparently fails to fit many features of the modern state. In the case of the private use of force, they bracket the issue, though this is a private delegation of coercive power that could be problematic. In the case of independent agencies, they allow for an ill-defined set of exceptions—a concession that threatens to unravel their entire theory. The case of adjudication by judges and juries is left unaddressed, even though this is a clear case of sovereign power where deference seems to be absent. And as to judicial lawmaking and constitutions, Dorfman and Harel celebrate aspects that seem to be completely at odds with representation and deference.
But suppose deference is both feasible and necessary (at least for inherently public goods). Why should this rule out privatization?
Dorfman and Harel’s arguments against privatization suffer from several problems. They adopt a meaning of “public” vs. “private” at odds with how the words are actually used; therefore, their argument isn’t about privatization as we know it. They wrongly dismiss as “fantastic” the possibility that private firms could enter into contracts with strong government control rights. The loophole that they allow for apolitical government agencies seems to allow for privatization as well, but they unconvincingly attempt to distinguish the two cases. And their position that privatization erodes political engagement only makes sense if it’s an empirical statement; and as such, it’s unsupported.
- This Book Isn’t About Privatization
- What Do “Public” and “Private” Mean?
Dorfman and Harel situate their book as a commentary on “[t]he increasing resort to privatization since at least the Reagan-Thatcher era” (Dorfman & Harel, p. 5), and note that “[their] argument against privatization applies to for-profit and not-for-profit private organizations” (Dorfman & Harel, p. 94 n.1). One might expect to find arguments that are relevant to today’s privatization debate. But they’re using a unconventional definition of “public” and “private,” which differs crucially from ordinary usage.
“[P]olitical authority,” they write, “is fundamentally public”—which, as we’ve seen, requires speaking and acting “in the name of the public” (Dorfman & Harel, p. 17); this involves a particular view of representation, which requires deference to the perspective of the represented. This means officials who enforce law for their own benefit and without deference to citizens—say, in some thoroughly corrupt and authoritarian country—can’t be called “public.” Dorfman and Harel’s definition of “public official” or “private employee” is not formal but functional (Dorfman & Harel, pp. 111–12, 139). People don’t become “public officials” unless their practice has the necessary integrative form; “[t]hey are not officials prior to it” (Dorfman & Harel, p. 112). And so the privatization debate in such a corrupt and authoritarian country might seem moot, since everything’s “private” already: there’s no gain or loss of legitimacy from shifting formal organizational modes within this society.
Moreover, in their view, formally private employees could become “public” under appropriate conditions:
[I]n principle, it is possible that private employees of a private firm would be considered, for our purposes, public officials. This may be so in the (fantastic) case in which they satisfy the two conditions we have articulated: that of participation in a practice that takes an integrative form. For such a case to arise, the for- and not-for-profit organizations must turn their backs on the private purposes that provide the grounds for their operations. They must withdraw from their basic commitments to maximize profits or vindicate certain ideals, respectively. They also must display fidelity of deference to the judgment of state officials in all matters pertaining to the execution of the contracted-for task (Dorfman & Harel, p. 112).
This is an extremely important clarification, which shows how at odds Dorfman and Harel’s definition is from common understandings.
Usually, the privatization debate concerns whether formally governmental bodies should contract with, or transfer some of their powers or assets to, persons who aren’t government employees or organizations created under private law and outside the formal structure of government. If I incorporate a company named “Volokh, Inc.” and sign a contract to perform some function previously performed by government employees, in the ordinary world, this is definitionally privatization. (All the more so if Volokh, Inc. were publicly traded and if the CEO could be fired by shareholders.)
But in the Dorfman-Harel world, this may or may not be privatization. It depends on whether my employees belong to a community of practice and whether the practice has an integrative form. Some governmental agencies might be “private”; some private corporations might be “public.” Counterintuitively, if one replaced a properly constituted private corporation with a corrupt public agency, the Dorfman-Harel definition might count this move as “privatization”! As Dorfman and Harel note, their functional definition of “public” vs. “private” “may sometimes be revisionary” (Dorfman & Harel, p. 139).
This is such an important point that I’ll restate it: Dorfman and Harel aren’t arguing against privatization as commonly understood. The public can be “private” and the private can be “public.” They’re arguing against organizations that lack a community of practice and integrative form, and they grant that such organizations could exist in the private sector. From the perspective of a would-be privatizer using common definitions, this is just an argument against certain organizational forms, and is fully consistent with (the right kind of) privatization.
- Why Should We Care?
Why do we care that the authors adopt a special definition of “public” and “private”?
There’s nothing wrong in principle with adopting a functional view of public vs. private; I’m sympathetic to a view that nothing important should hinge on formal designations. Nor is there anything wrong with adopting particular definitions for purposes of argument; defining terms is praiseworthy, and authors can define black as white and white as black if they like—as long as they’re consistent. Nor is there anything wrong with adopting definitions that diverge from popular understandings; perhaps popular understandings are misguided or incoherent, and another definition can be better.
But there are disadvantages in defining common terms in ways that diverge too radically from common understandings.
First, it’s confusing to the public; there’s value in guarding against the easiest misunderstandings of our work, especially ones stemming from unusual definitions.
Second, Dorfman and Harel don’t necessarily use the terminology consistently, as I touched on earlier and explain more fully below. They refer to privatization as a Reagan-Thatcher phenomenon, and several times, they discuss private organizations with the qualifier “for-profit or not-for-profit” (Dorfman & Harel, pp. 94 n.1, 110, 159). This already suggests that they sometimes use the traditional formal definition of “private”; otherwise, they should have allowed for a third category, “apparently governmental but actually private because it lacks the appropriate form” (i.e., the apolitical agencies, which they nonetheless call “public”).
Third, their definitional move loses valuable opportunities to contribute to the privatization debate. Dorfman and Harel’s thesis can be rephrased as a conditional pro-privatization statement. Imagine this statement: “To be legitimate, officials who enforce the law need to act in our name, which means being properly deferential to the polity’s point of view, which means adopting a community of practice and an integrative form. This is in principle achievable within the private sector; as to integrative form, it requires that contracts allow for robust intervention rights of public officials and be easily terminable.” This is still subject to the aggregative concerns of Chapter 5 (more below), but at least adopting such intervention and termination rights should make the system legitimate, i.e., the private contractors would be “public” according to their definition.
This would be useful in actual privatization debates, where this book could represent a nuanced view; some people might favor privatization only to the extent it embodies particular valuable features, like (in this case) integrative form. Some politicians might seek to increase the support for privatization by designing privatization proposals that incorporated the necessary Dorfman-Harelian features.
To be sure, the labeling doesn’t preclude such alliances; I could claim a policy as privatization, while Dorfman and Harel would deny that it was privatization at all; we’d both be happy, though we might want to avoid ruffling feathers by avoiding the word “privatization” and saying “quasi-public contracts” or suchlike. But the labeling needlessly complicates political agreement.
- The Supposedly “Fantastic” Private Option
Dorfman and Harel themselves are skeptical that this could happen: when discussing private firms’ becoming “public” for their purposes by adopting a community of practice and integrative form, they characterize the possibility as “fantastic” (Dorfman & Harel, p. 112)—and not in a good way. “For such a case to arise,” they write, “the for- and not-for-profit organizations must turn their backs on the private purposes that provide the grounds for their operations. They must withdraw from their basic commitments to maximize profits or vindicate certain ideals, respectively” (Dorfman & Harel, p. 112).
But to call this “fantastic” commits the Essentialist Fallacy and ignores the Clever Privatizer Principle and the “Regulation by Contract” Principle. This is capitalism; private firms can agree to whatever they like. The government could demand that it be allowed to intervene in the private firm’s decisions, and that it be able to rescind the contract without damages. People and firms can allow for extensive control rights in their contracts with other people and firms, and they can do the same when their partner is the government. Nothing prevents firms from agreeing to easily rescindable contracts; people do it all the time (e.g., for at-will employment), and governments can do the same with their contractors.
Why would firms agree to this? Would they be
“turn[ing] their backs on [their] private purposes . . . [or] basic commitments”? No. If the government demanded such terms, would-be contractors could agree to them—in exchange for an appropriate payment—because doing so would be profitable. Nonprofits can do the same if they think doing so wouldn’t undermine their mission too much. They might be confident that they’ll do a great job, and that politicians will be impressed and usually won’t intervene. As with any voluntary transaction, a private organization’s agreement to a condition is a sign that the condition is consistent with its purposes. You don’t like the contract (taking everything, including the money, into account)—don’t sign it. Far from turning their backs on their private purposes, organizations that agree to such terms are vindicating their purposes. Outside philosophers might find that strange or even “fantastic,” but that’s a sign that the outsiders misunderstand the organization’s true purposes, perhaps from misplaced essentialism.
Perhaps obtaining a firm’s agreement would require a payment that’s too high for the government’s tastes. But that depends on how burdensome such terms are in practice—and that’s a practical question, not a fundamental concern. The arrangement might still be beneficial from the firm’s perspective without a large extra payment—if it expects the government to be happy with its work. Would thoroughgoing governmental control rights eliminate the benefits of privatization? Maybe, maybe not: that’s an empirical question, and it’s inappropriate for us to speculate prematurely.[19]
- The “Apolitical Agencies” Loophole
Let’s return to the coherence of Dorfman and Harel’s public-private line. As we’ve seen, their treatment of apolitical agencies—“aptly perceived as public” even though lacking an integrative form (Dorfman & Harel, p. 139)—is in tension with their claimed functionalism. They explain that these are permissible because of the benefits of apolitical practice; I’ve questioned above whether such benefits can ever (within their theory) justify the presumed loss of legitimacy stemming from the absence of deference—and whether their policy-based exception has any limiting principle.
The only limiting principle the authors suggest is that “the discretion that is granted to what we call apolitical public practices [must be] qualitatively different from that created by the act of privatization” (Dorfman & Harel, p. 139).
But this distinction is likewise problematic. This is an extremely important point, because if they endorse apolitical agencies but can’t distinguish them from private contractors, their case against privatization collapses. It’s therefore worth quoting them at length:
It may be protested that this sort of “outsourcing” is a form of privatization. We think not. This is because the arena of permissibility granted to apolitical public practices is qualitatively different from the one created by the act of privatization. In contrast to public officials, private actors possess a valid claim right against state interference insofar as they act within the designated arena of permissibility. Instead of being liable to the power of the state to direct their conduct, private agents enjoy a form of immunity on the basis of which they can invoke their right not to follow the demands of the public interest (as viewed from the polity’s point of view). By contrast, agents of apolitical public institutions enjoy no such immunity. Accordingly, they have no valid claim of their own against state intervention whenever the polity determines that the judgments of these agents disserve the general interest.
The fundamental difference is that the discretionary powers, that is, arena of permissibility, in the case of private entities reflects a concession granted to the private entities, and it is designed to allow them to pursue their interests, concerns, and ideals. Therefore, private entities have a right that the arena of permissibility be respected. In contrast, the arena of permissibility given to public officials in apolitical institutions is exclusively designed to promote the public interest. It confers no rights on the public officials even though it does form a genuine obstacle to political intervention. It should be perceived as an exercise of self-constraint on the part of politicians grounded in their judgment that the general interest is better served by apolitical practices.
One implication of this analysis is that sometimes politicians should defer to the decisions made by the private entity insofar as they fall within the arena of permissibility even when these decisions run afoul of the general interest. This is because the private entity has acquired a right to so act. In contrast, public officials enjoy no such right: In principle, when the arena of permissibility granted to officials is invoked in ways that are judged by the polity as being detrimental to the general interest, it (the arena of permissibility granted to them) should be revoked. Of course, making a judgment of this sort raises important concerns—for instance, there must be an appropriate political procedure for making such judgments and for intervening in the decision-making processes of the apolitical institution. . . . [T]he arena of permissibility characteristic of apolitical institutions is qualitatively different from the one granted to private entities and . . . the difference lies in the absence of a valid claim by apolitical institutions to act contrary to the general interest, properly conceived. . . . [W]hereas a public official of an apolitical institution holds a mandate from the polity, a private agent holds a right against the polity (Dorfman & Harel, pp. 140–41).
But the supposed difference between independent agencies’ and private contractors’ arenas of permissibility is illusory.
Imagine an independent agency and a private contractor with equivalent arenas of permissibility. Suppose the government (based on its view of the public interest at that moment) wants to tell them what to do, and they say no (in a context where they’re entitled to do so). What can the government do?
In the case of the private contractor, the government could wait until the contract expires and choose a different contractor (or take the service in-house). (Indeed, the government can make it clear that this will happen, which might make the contractor comply quickly—perhaps even more certainly than in a public agency. But the authors presumably don’t think de facto control is sufficient.) Or the government could breach the contract, which might require paying damages. (Perhaps the authors don’t think this option is sufficient, because the requirement to pay damages would penalize the government for exercising this option.)
In the case of the independent agency, all the government can do is try to get the statute amended. The agency’s insulation from the polity’s contrary views of the public interest is, if anything, more complete than the private firm’s.
Dorfman and Harel write that the private insulation is a “concession” granted to the firm “to allow [it] to pursue [its] interests, concerns, and ideals,” whereas the apolitical agency’s insulation is “exclusively designed to promote the public interest” and “should be perceived as an exercise of self-constraint . . . grounded in [politicians’] judgment that the general interest is better served by apolitical practices.” But there’s no strong difference here. The government isn’t trying to promote the private organization’s “interests, concerns, and ideals” as such, but only agrees to the privatization because it thinks this would promote the public interest.[20]
What if the polity changes its mind and decides that the apolitical agency’s arena of permissibility no longer serves the public interest? Then, indeed, as Dorfman and Harel say, that arena should be revoked, after the “appropriate political procedure” is followed—the legislature should amend the statute. But this is precisely what should be done if the polity decides that the private contractor’s arena of permissibility no longer serves the public interest. There might be some transitional period when the government can’t yet intervene because of contractual expectations—though the contract can be written to be easily revocable without penalty, and if it isn’t, the government can still choose to breach. But this is again the same as with apolitical agencies: passing a statute also takes time, and there’s no reason why rescinding the arena of permissibility is any harder for contractors than for apolitical agencies.
Perhaps Dorfman and Harel are putting some significance on the idea that private contractors are rightsholders, so their claim against the polity takes the form of a “right,” whereas a public agency doesn’t have its own independent purposes and therefore lacks a “right.” But if we’re to make anything depend on that distinction, we’re relapsing into public-private formalism. Functionally, any arena of permissibility is granted by the polity—whether to a contracting firm or to an apolitical agency—because of a belief that doing so serves the public interest; it’s an exercise of restraint on the polity’s part; and whenever the polity thinks this restraint no longer serves the public interest, it should revoke the delegation.
In either case, according to the rest of Dorfman and Harel’s theory, the entity receiving the delegation should be thought of as “public” if it has the requisite deference (i.e., a community of practice and integrative form) and should be thought of as “private” if it lacks that deference. As long as apolitical agencies can be “public,” private firms can be “public” as well, which collapses the case against privatization.
- Privatization as Detachment of Polity
So far, we’ve been talking about Dorfman and Harel’s arguments against privatization for “inherently public goods,” the argument of Chapter 4. In Chapter 5, they give additional reasons to oppose privatization. These reasons apply to any privatization, not just of inherently public goods, and are less absolute—they don’t rule out privatization entirely, but merely identify certain disadvantages. “Privatization cuts off the link between processes of decision-making and the citizens and, therefore, erodes political engagement and its underlying notion of shared responsibility. Consequently, privatization undermines individuals’ public autonomy” (Dorfman & Harel, p. 124).
The reasoning begins with the idea that citizens must have a “meaningful role,” or “some nontrivial measure of control,” over “the making of political decisions” (Dorfman & Harel, p. 142). However, because private contractors have an “arena of responsibility,”
the polity has no direct control over the decisions made by the private contractors and, therefore, no responsibility for these decisions and the actions that follow them. Privatization signifies the detachment of the polity from at least some of the decisions made by the private body. By granting immunity to the decisions made by the private entity, the polity distances itself from the privatized activity or, at least, from those decisions made by the private entity that fall within the scope of the arena of permissibility (Dorfman & Harel, p. 143).
By contrast, public institutions act in the public’s name, which, as we’ve seen, requires an integrative form: “politicians are active participants in the integrative practice of public officials” (Dorfman & Harel, p. 143). Citizens can control their politicians and politicians can control public officials—so citizens ultimately control the public officials.
Dorfman and Harel are making several claims here.
First, because politicians (and therefore citizens) lack direct control over private organizations, they lack responsibility for those organizations’ decisions and are distanced or detached from those decisions. We’ve seen above that politicians can control their private contractors by negotiating contracts with thoroughgoing rights of public control or easy termination. We’ve also seen that Dorfman and Harel are happy with apolitical agencies that lack such control, and the only constraint on such agencies is that the public has to think political insulation serves the public interest. Even if such agencies are legitimate, one would think there’s still the same level of citizen detachment from such agencies’ decisionmaking.
But in addition to these concerns, it’s unclear why we need to solely focus on formal mechanisms of political control. Suppose a private contractor had a very broad arena of permissibility, so that the government had no right to intervene until the end of the contractual term; but the government made it perfectly known to the contractor that if it didn’t do things a particular way, its contract wouldn’t be renewed. This “soft power” can be just as effective as formal political control. Sometimes even more so: government is full of “public” agencies whose internal cultures consistently resist actual reform (e.g., police departments, prison systems, and militaries). Should we focus on formal or de facto control? Sometimes the formalities are important. But if we’re talking about whether the polity is detached or distanced, it seems that the reality should matter.
Suppose the polity hands over prison management to private firms. Rightly or not, the decision is intensely controversial, and citizens and politicians exercise intense scrutiny over prison contractors—greater than they had ever exercised over public prisons. Whenever something goes wrong, there’s vigorous debate over revoking the contract; sometimes, the contract is immediately revoked. In other cases, all contractors become aware that the contract with the offending firm won’t be renewed. Politicians are on the phone with the firms’ CEOs all the time, and the CEOs are constantly testifying before hostile legislative committees, so the consequences of failure are clear. The stock prices of bad-performing firms plummet, and their stockholders replace those firms’ management. As a result, contractors take immense care to avoid problems.
It seems implausible to say that, in this hypothetical, the polity has distanced itself from private firms’ actions. For the distancing thesis to make sense, it should be empirical—and given the reality of unaccountable public agencies and the possibilities of vigorous oversight of private contractors, we can’t say that the empirics necessarily oppose privatization.
Second, Dorfman and Harel say that, because citizens lack direct control, they also lack responsibility. This is an astounding claim. Consider the mass of decisions made in the private sphere, with no direct public control rights at all. Lots of abuses happen in this private sphere: people abuse their family members; businesses rip off their customers; entrepreneurs develop shady products. Other decisions may be public but outside the direct control of our polity: most people live far from Ukraine or Gaza. We’ve left many decisions to the private sphere, and the world is fragmented among many polities. Do we therefore lack responsibility?
Of course not. On the contrary, we have a responsibility to do whatever’s within our reasonable power to fight injustices, within our polities and sometimes abroad. This is why many political issues concern not formal control over the administrative state, but whether to pass or repeal laws (or increase or decrease their enforcement) regulating private activity—abortion, drug use, student loans—or whether to send weapons to Ukraine or Israel.
Dorfman and Harel have an answer to this:
Citizens always have good reason to struggle against injustice simply by virtue of being persons. However, there arises an additional reason to do so when the injustice in question is the doing of public officials. This is because the latter instance of injustice is done in their name—that is, by public officials who act in the name of the polity to which they belong. . . . A citizens’ protest against the injustice committed by a public agency differs from a protest against injustice committed by an individual, private entity, or another state. It is a protest against injustice (or some other grievance) that can be attributed to the citizen who is, thereby, responsible for its occurrence (Dorfman & Harel, p. 144).
Note that, in their view, when an injustice is committed by public officials, that doesn’t affect whether we have a reason to oppose it; it merely gives us an additional reason to oppose it. But what’s the significance of having one sufficient reason rather than two sufficient reasons? In either case, citizens should take action to remedy the injustice. The idea that the polity is more distanced from injustice when it’s not committed by public officials seems like a highly contestable empirical judgment.
What of the argument that, in the case of privatization, the citizen is responsible for the initial delegation? Dorfman and Harel write:
It is, of course, true that the polity and its constituents bear responsibility for making the initial decision to privatize a given activity, selecting the appropriate contractor, and monitoring its conduct. That said, none of these factors could compensate for the lost control over the manner in which the private entity acts (at least insofar as it acts within the arena of permissibility). Even given that the polity had a specific vision when it privatized the activity, it is barred from reconsidering or changing its course and purpose. . . . [W]hat characterizes public officials is the fact that they are constantly liable to the normative power of the pertinent politician (Dorfman & Harel, p. 144).
But everything in the private sector is always liable to politicians’ normative power. If something is wrong, politicians can pass a statute to fix it; this is no different than their power to pass a statute to remove the arena of permissibility from apolitical agencies, which Dorfman and Harel endorse; and if one takes into account the polity’s power to write easily terminable contracts that give strong intervention rights to politicians, controlling private contractors can be even easier.
The idea that privatization erodes political engagement is thus highly contestable. Political engagement arises from many sources, and privatization might merely change the type of political engagement. If the claim is divorced from empirics, the claim is implausible; the claim makes the most sense if it is empirical, but then the authors don’t justify whether it is true.
If Dorfman and Harel’s approach doesn’t provide a truly noninstrumental argument against privatization, can one do better? Let’s consider Cordelli’s book.
Dorfman and Harel’s approach and Cordelli’s approach share important similarities. Both books are within the broad Kantian (as opposed to Lockean) position. Cordelli’s approach, though, is explicitly grounded in Kantian political philosophy, while Dorfman and Harel’s isn’t.[21] Both agree that certain governmental powers must be “not simply authorized by the people but also exercised ‘in their name’ and in a way that carries out their shared will” (Cordelli, p. 8).[22] They agree that the nature of a good can depend on the identity of who provides it (see, e.g., Dorfman & Harel, pp. 11, 115–16; Cordelli, pp. 32–33, 55–58, 65, ch. 5)—so that, say, private incarceration and (properly constituted) public incarceration are actually different goods (and one might be legitimate and the other illegitimate), even if the prisoners are treated identically. In other words, the provider of certain goods must not only do the right thing but also have standing to provide the good (see, e.g., Dorfman & Harel, ch. 2; Cordelli, ch. 5).
The two approaches also differ on a number of points. The clearest difference is in the identification of the activities to which the noninstrumental critique of privatization applies. Cordelli focuses on the broad class of acts that change people’s normative situation, whereas Dorfman and Harel rely on an activity-by-activity identification of “inherently public goods.” Cordelli calls Dorfman and Harel’s argument (as discussed in some of their previous papers) the “most interesting and powerful noninstrumental argument” (aside from, I presume, her own); she calls their argument “essentialist,” because of their claim that inherently public goods can’t be privatized “by their very essence” (Cordelli, p. 39). But she critiques their argument, on the one hand, because it’s “grounded on a very specific (and contested) interpretation of the essence of certain goods as inherently public” (Cordelli, pp. 40–41) (i.e., what if we think punishment is only about rehabilitation or retribution?), and because it’s “only able to condemn a very limited category of cases of privatization” (Cordelli, p. 41).[23]
Cordelli’s account of Kantian political philosophy begins with the Kantian state of nature. The Kantian state of nature is different than its Hobbesian/Lockean counterpart. The Hobbesian state of nature is characterized by the “war of all against all,” and life is “solitary, poor, nasty, brutish, and short.”[24] Life in that state of nature is so miserable that people are willing to escape it by giving up a lot of their natural liberty—even to the extent of accepting an absolutist government.
The Kantian state of nature, though, is subtler. For Kant, submitting to a strong warlord would also violate our freedom, understood as independence from any particular person’s unilateral will. When individuals purport to change our normative situation (i.e., define our rights and duties), their judgment is merely unilateral and thus isn’t morally binding. (Cordelli calls this “provisional,” as opposed to “conclusive.”) There’s a paradox: Our freedom—i.e., our independence—requires that we are able to acquire rights. But in a world of unilateral judgments—the Kantian state of nature—we can’t acquire rights that are conclusive, i.e., rights that impose binding and enforceable obligations on others.
So we have a duty to escape the state of nature (in Kantian terms, a duty to “create a rightful condition”) and establish a political authority that can conclusively determine our rights. A properly constituted state is necessary for individual freedom, because only such a state can act with an omnilateral rather than a unilateral will, thus being able to authoritatively alter our normative situation.
- The Kantian Legitimation of Democracy
What would an omnilateral political authority look like, and how could we create one?
Cordelli says that there should be a strong presumption in favor of representative democracy, and then, more strongly, that democratic procedures are conclusively authoritative.
First, to satisfy “reciprocal nonsubjection” and be “compatible with the fundamental equal normative authority of all” (Cordelli, p. 63), we need a democratically authorized constitution. But could a democratically authorized constitution authorize benevolent dictatorship? No, Cordelli argues, based on the Kantian concept of “rightful honor”: one needs to be “an active agent” rather than “a mere means for the commitments of others”; “our active agency would likely be jeopardized if we were subject to a system of rights and restrictions on our freedom that was simply imposed on us, without our being able to play any part in actively shaping its content” (Cordelli, p. 64).
So the government itself—not just the constitution—should be democratic. And because one’s duties can properly only be discharged by oneself or by someone acting in one’s name, this government should be representative (Cordelli, p. 65). (We could appoint our rulers by lottery, and this would satisfy reciprocal nonsubjection and respect for equal normative authority, but it wouldn’t satisfy active agency and representation (Cordelli, p. 66).)
What’s to prevent democratic government from being a form of unilateral subjection? Cordelli recognizes this possibility: “a minority will be forced to do x . . . just because the majority says so. Yet it would seem that the majority’s will is nothing other than an aggregate of private wills” (Cordelli, p. 66).
Still, she says, there’s “room for hope” (Cordelli, p. 66). First, “[i]ndividuals, in the state of nature, have presumptive nonrefutable reasons, grounded on freedom, to treat democratic procedures as authoritative” (Cordelli, p. 67). But what does it mean to treat something as authoritative? “To treat a procedure as authoritative means to regard oneself has having reasons to comply with the outcome of the procedure just because the procedure selects that outcome, independently of any particular first-order reasons for wanting that outcome or not” (Cordelli, p. 67). If that’s the case, then citizens “do not surrender to any other’s particular will when they do so but rather act on shared, because procedural, reasons that the political process alone generates” (Cordelli, p. 68). Next:
as long as (i) the minority has independent, presumptive reasons to treat the democratic political process as authoritative, and insofar as (ii) treating this process as authoritative means that the minority must act on reasons that are themselves generated by the authority of that process, then (iii) the minority has reasons to comply with the outcomes of the process that are not themselves dependent on the majority’s will (Cordelli, p. 68).
Thus, “the principle of rational independence is satisfied.” This is why “democratic procedures [are] conclusively, and not simply presumptively, authoritative. This, in turn, is why individuals have conclusive, rather than presumptive, nonrefutable reasons to bring about a system of democratically authorized law, as a way of solving the problem of the state of nature” (Cordelli, p. 68).
- The Kantian Legitimation of Agencies
Suppose we’ve legitimized lawmaking; we still need to see whether we can legitimate law enforcement, including (in modern societies) the administrative state. Cordelli argues, convincingly, that “bureaucratic unilateralism” is a problem, because bureaucrats have a lot of discretionary judgment. If we’re concerned about unilateral impositions, then bureaucratic discretion is a serious problem, which might reestablish the state of nature at the implementation stage even if we’ve resolved it at the lawmaking stage.
To legitimate administrative discretion, Cordelli presents three possible models: the “top-down model” of political control, the “fiduciary model” of bureaucratic independence, and the “public participation model” of legitimation from below (Cordelli, pp. 97–98). Top-down control is good because exercises of administrative discretion need to be democratically authorized (Cordelli, p. 99), but excessive majoritarianism is harmful for “respect for the rule of law and support for the effective realization of a rightful condition” (Cordelli, p. 100). The fiduciary model is good for “resisting short-term partisan pressures when these go against the rule of law or other constitutional essentials”; administrators need a “bureaucratic ethos,” which “requires a disposition to exclude private purposes and personal loyalties from consideration, even when the openness of rules leaves wide interpretative discretion in place” (Cordelli, pp. 102–03). But even the fiduciary model might not be enough to avoid reproducing unilateral subjection (Cordelli, p. 108). The public participation model has advantages in limiting bureaucrats’ residual discretion, but suffers from various other problems, such as unequal participation, capture, manipulability, and ignorance (Cordelli, pp. 109–11).
Cordelli suggests a combined model that includes elements of all three of these models. One feature of her model involves a system of “codetermination, coupled with public hearings,” in which
decentralized administrative agencies, as well as local agencies empowered to regulate and oversee the work of street-level bureaucracies within specific issue areas, would be managed by boards of directors including both members of the public and insulated officeholders. The members of the public would be selected by lot, on the model of civic juries, from among those whose rights or duties of citizenship are governed or changed by the proposed regulations (Cordelli, p. 111).
This civic jury would “retain[] a right to veto regulations that, even if perhaps compatible with the intent of the legislature, still fail to take certain relevant interests, or information that has emerged during the process of public consultation, into due consideration” (Cordelli, p. 112).
Having laid this groundwork, Cordelli argues that privatization doesn’t satisfy the necessary legitimacy conditions. The three major conditions of administrative legitimacy are:
- The authorization condition: “democratic mandates, which delegate important legislative or regulatory discretion outside of the legislature, [must] be valid” (Cordelli, p. 114–15).
- The representation condition: “legislative or quasi-legislative discretion [must] be exercised in a representative capacity—in the name of all” (Cordelli, p. 115).
- The domain condition: “the exercise of legislative discretion must carry out (and, in the process, help reconstitute) the shared will of the people throughout the process of administration. This further condition helps ensure not only that bureaucrats do not impose their unilateral judgment on citizens, but also that what bureaucrats end up doing can be reasonably regarded as falling within the scope of their delegated authority” (Cordelli, p. 115).
In her central chapters, Cordelli applies these conditions to privatization. Privatization “compromises the ex ante validity of democratic delegations” (Cordelli, p. 115), which violates the authorization condition. It “compromises the ability of administrators to act ‘in the name of’ the people when exercising relevant forms of discretion by undermining many of the structural features of office, as well as by changing the nature of the bureaucratic ethos” (Cordelli, p. 115), which violates the representation condition. And it “separates, rather than integrates, the bureaucratic and the democratic, thereby preventing the administrative state from carrying out the shared will of the people” (Cordelli, p. 115–16), which violates the domain condition. Because privatization violates the conditions of legitimacy, it doesn’t solve the unilateral subjection problem, so “the privatized state should be understood, normatively, as a state of progressive regression to the state of nature” (Cordelli, p. 116).
- The Problem of Authorization
Cordelli says widespread privatization amounts to an “abdication of the collective right to democratic self-rule” (Cordelli, p. 135), which violates a principle of “[c]ollective [n]onalienation” (Cordelli, p. 134).
How does privatization do this?
This question can be answered only by analyzing a complex set of empirical facts. Building on recent literature on the effects of systematic outsourcing, we can uncover at least three robust causal mechanisms through which privatization distinctively undermines the three fundamental preconditions of self-rule, namely (1) directive control, (2) civic vigilance, and (3) equal opportunities for political influence (Cordelli, p. 142).
- Directive Control
First, consider privatization’s harms to the government’s directive control. The “prima facie case for privatization” is “to improve both efficiency and flexibility while saving costs, and to compensate for a lack of specific capacities to respond to new situations” (Cordelli, p. 142). But this leads to a problem: “if government lacks the capacity to directly perform certain functions or to do so efficiently, it will also likely lack sufficient capacity to coordinate, plan, oversee, and regulate those to whom those functions are delegated, and to do so efficiently” (Cordelli, p. 143).
Moreover, “the more a government privatizes, the more difficult it becomes for it to control its myriad agents” (Cordelli, p. 143). A “brain drain” to the private sector is likely (Cordelli, p. 143). Government then must outsource the monitoring function itself. In addition, “the less government officials perform the relevant functions themselves, the more they lose the ability to actually perform those functions,” so “[t]he chain of delegation then becomes a vicious circle” (Cordelli, p. 144). Government officials, and ultimately the people, “progressively lose both epistemic and practical control” (Cordelli, p. 144).
- Civic Vigilance
Next, consider privatization’s harms to civic vigilance. Civic vigilance, Cordelli writes, has both epistemic and affective components.
Epistemically, privatization makes it harder to detect abuses:
[P]rivate actors, unlike civil servants, tend to have stronger claims (grounded on their preexisting right to freedom of association and organizational autonomy), against intrusive forms of interference and regulation; generally act outside of tight administrative procedures so as to maintain flexibility and efficiency; and, qua private corporations, can operate across multiple jurisdictions (Cordelli, p. 146).
Privatization makes the role of government less visible (making it look like the market at work), which makes citizens unaware of how government works.
And affectively, it’s significant that “private organizations’ symbolic identity visibly differs from the one of public entities” (Cordelli, p. 146); “when people do not see their own government as the main provider of the benefits they receive, they see little reason to care about their government and thus to actively participate in politics” (Cordelli, p. 147).
- Political Influence
Finally, privatization may undermine equal opportunities for political influence. Privatization increases the rate at which economic resources can be converted into political influence, “by providing wealthy private firms and corporations with incentives to direct a large amount of their private resources into politics, thereby also contributing to institutional corruption, understood as a process through which forms of improper, although not necessarily unlawful, influence ultimately render political institutions unable to fulfill their purpose” (Cordelli, p. 148).
Political influence is a problem for three reasons. First, campaign finance regulations can’t fully control the problem. Second, public officials are dependent on private actors because of their reliance “on the private sector for information and resources” (Cordelli, p. 149). Third:
[A]lthough public agencies can obviously also be subject to capture by particular interests, it is a distinctive and constitutive feature of a public office, beyond the lack of free purposiveness, that the officeholder, unlike a private actor, does not own property that he or she can discretionally use or spend. . . . While bureaucrats, in their public capacity, can and do manage state resources, they should not have discretionary control over public property with which to influence the political process (Cordelli, p. 149).
* * *
All of these considerations show, Cordelli argues, that “privatized government . . . poses a distinctive set of threats to the preconditions of democratic self-rule,” even though privatization might perform better than some “widely corrupted or captured nonprivatized governments”: “while the corrosion of self-rule can certainly result from the corruption of public government, this same corrosion is endemic to privatized government and its logic, even in its noncorrupted form” (Cordelli, p. 150). Privatization represents an abdication of the capacity for self-rule; thus, “a democratic people lacks the moral power to decide to privatize” (Cordelli, p. 151). Systematic privatization is not just bad and impermissible, “but also unauthorized” (Cordelli, p. 151). This is an aggregative problem; as for specific instances of privatization, it’s a “case of a problem of many hands” (Cordelli, p. 151). Up to a threshold, individual instances may be fine; but “once the risk of self-rule abdication becomes sufficiently likely, then each instance of privatization may become morally wrong, and morally wrong independently of the particular nature of the privatized function. . . . Stopping privatization is then . . . the morally required solution” (Cordelli, p. 151).
- The Problem of Representation
Next, Cordelli looks at the “[p]roblem of [r]epresentative [a]gency,” or the “problem of representation”: What does it take to “truly act in the name of the people” (Cordelli, p. 156)? This is the question of whether one can in principle be a representative at all, which is prior to the question of whether one is a virtuous representative (Cordelli, p. 157). If a private entity is incapable of acting in the name of the people, then representation is impossible, which immediately rules out legitimacy.
Cordelli rejects various views of representation—including Dorfman and Harel’s view that representation derives from authorization plus deference (Cordelli, p. 159)—before settling on her own “internalist account of representative agency” (Cordelli, p. 169):
[A]n agent (A) does X in a principal (P)’s
name if and only if:
- The authorization condition: P validly granted to A the authority to do X
- The intention condition: A does X intentionally
- The included reasons condition: A does X for reasons that are not excluded in virtue of acting under P’s authorization.
- The domain condition: X falls within the authorized domain of action D, according to a reasonable interpretation of P’s own understanding of the boundaries of D at the time of the authorization, or according to a subsequent review or in-process ratification by P.
Cordelli addresses one immediate objection: If an official’s mental states determine whether their acts are legitimate, aren’t we subject to their goodwill, which is incompatible with independence (Cordelli, pp. 169–70)? Not so, Cordelli says, because “the appropriate intentional orientation required of officeholders and civil servants should be a product of . . . a ‘bureaucratic ethos’” (Cordelli, p. 170), which is institutional and impersonal; you’re thus not being subjected to someone’s unilateral will.
With this framework in mind, Cordelli explains what’s wrong with private actors: They violate (3) the included reasons condition because, due to private organizations’ free purposiveness, they act for reasons that are excluded; and they also violate (4) the domain condition because their organizational ethos is different from the proper bureaucratic ethos. Therefore, private actors’ decisions “often fail to qualify as done in the name of government, and of the political community government represents” (Cordelli, p. 170).
- Included Reasons Condition
To illustrate, Cordelli gives the example of WorkOpts, a firm that contracted with states to deliver welfare services. WorkOpts’s contracts required it to serve 1,200 welfare recipients per year and place at least 10% of them in jobs. Its managers also owed fiduciary obligations to its shareholders, and the parent company required it to make an 8% profit (Cordelli, p. 171); the result was a drastic reduction in caseworkers’ time spent per recipient (Cordelli, p. 172).
In addition to justice-based concerns about which recipients were prioritized, Cordelli notes concerns of representative agency and democratic legitimacy: “On paper, WorkOpts’ managers have followed government’s contractual directives and met contractual standards. Yet can we truly say that they have exercised their decision-making powers and provided welfare in the name of government?” (Cordelli, p. 174). No: Because they intentionally acted to make an 8% profit for their profit and interpreted their contractual obligations from this vantage point, they “intentionally act[ed] for purposes that lack the status of reasons from the perspective of their institutional role as presumptive government’s agents” (Cordelli, p. 179).
- Domain Condition
Cordelli goes on to the “problem of misinterpreted domain”: “private actors who purport to act in the name of government often fail to meet” the domain condition (Cordelli, p. 184). “[T]heir nonpublic organizational culture shapes their interpretive competence, leading them to interpret the purpose of their public mandate in a way that does not align with the principal’s understanding of that same purpose, given the latter’s interpretative framework” (Cordelli, p. 184). WorkOpts’s organizational culture was based on efficiency and profit, which colored how they interpreted their contractual requirements: efficiency, “far from being a means to an end, or a secondary end the importance of which is derivative from more fundamental goals, [became] a primary end” (Cordelli, p. 189).
Isn’t this a problem in public organizations too? Yes, but:
[A] public office, by its own constitution, should include a system of incentives, the purpose of which is to sustain an intentional orientation toward public purposes alone and corresponding reasons for action. By contrast, but for parallel reasons, the competitive market structure within which the managers are situated, and its attendant system of incentives, give rise to an alternative set of cultural lenses through which salience is denied to certain considerations that would be relevant from the perspective of public purposes. We can thus expect the problem of silencing relevant reasons to be inherent to (even properly constituted) private organizations, in a way that it is not to (properly constituted) public offices (Cordelli, p. 191).
- The Problem of Delegated Activity
The final problem that Cordelli identifies is “the problem of delegated activity”: private actors’ “resulting determinations often fail to qualify as acts of lawmaking, for they fail to qualify as acts that the lawmaking community has done together” (Cordelli, p. 197).
Cordelli develops “a moralized account of (legitimate) lawmaking as something that a certain group of people can do only together,” which she calls “the collective action view of legitimate lawmaking” (Cordelli, p. 199). As explained earlier, lawmaking should be omnilateral, not unilateral: “laws and policies that purport to authoritatively change, demarcate, and enforce people’s rights must be regarded as instances of an ‘omnilateral’ will—that is, they must be made in everyone’s name and in a way that carries out the people’s shared will, beyond simply being publicly authorized” (Cordelli, p. 199). The question is then whether private parties are capable of lawmaking in this sense. She argues that they aren’t:
[P]rivate actors, because of their (1) multiplicity of conflicting loyalties and goals, (2) relative lack of a bureaucratic ethos of public service, and (3) lack of integration in a unified procedural structure that links together the bureaucratic and the democratic, often fail both to have the right kind of intentional orientation and to relate to other participants in the appropriate way. They thus fail to act as participants in the collective practice of lawmaking. This, in turn, means that their decisions fail to qualify as acts that the lawmaking community has done together. They thus lack the status of acts of lawmaking that from this community of practice derives” (Cordelli, p. 199).
Cordelli’s view of carrying out the people’s shared will is that, initially, “the substantive content of the end [i.e., the actual law that is enacted] must qualify as a reasonable specification of some of the people’s joint commitments” (Cordelli, p. 205). Moreover, there has to be an “[i]ntentional orientation to joint activity: lawmakers each [must] have an intention, though perhaps motivated by different reasons, to the (joint) activity of reaching an agreement on an end that reasonably specifies the shared will of the people, by further specifying their shared commitments” (Cordelli, p. 206). But there remains the problem that “the people’s commitments, as embedded in constitutional principles, are so general that they will leave ample discretion to legislators as to how to further specify the content of policy ends”—so “how should the legislators proceed in this specification in a way that can be regarded as further carrying out the people’s will?” (Cordelli, p. 206).
Cordelli concludes that, “[w]hile preferences can be, and arguably so, only aggregated, reasons can be shared. A decision-making process oriented toward a consensus on mutually acceptable decisions can, if appropriately deliberative, generate or construct shared reasons” (Cordelli, p. 206–07). This process should be deliberative:
In order to develop concrete proposals, each representative may start by referring to the political views of their constituents, since such views have more specific content than general commitments, but they should do so with an intention to contribute to an overall decision-making process whose aim is to arrive at a reasonable elaboration of the shared commitments of the citizenry as a whole. This, in turn, means that each representative must be willing to revise his or her views in light of what other representatives say; they must take into account the voice of less established views in the legislature, they must be willing to formulate proposals that reflect reasonable compromises among all participants in the process, and they must allow for fair hearings of all. Each representative must thus intend to reach an agreement on an end together with the other participants (as specified in their joint commitment) by way of each acting in accordance with his or her own subplan (voicing and representing the political views of his or her constituencies), in a way that meshes with the subplans of others (by each being willing to listen to others and revise his or her views accordingly), knowing that others will do the same (Cordelli, p. 207).
The legislators should also exhibit “[m]utual responsiveness: each lawmaker [must] attempt[] to be responsive to the intentions and actions of the others, knowing that the others are attempting to be similarly responsive” (Cordelli, p. 208). They should also exhibit a “[c]ommitment to mutual support: each lawmaker [must be] committed to supporting, or at least to not actively undermining, the efforts of other lawmakers to play their roles in the joint activity” (Cordelli, p. 210); this last condition makes lawmaking not only a “shared intentional activity” but also a “shared cooperative activity.”
Once we get to the activity of the administrative state, we similarly need to figure out how administrators can carry out the shared will of the people without exercising bureaucratic unilateralism. Again, we need a shared cooperative activity—as to the relationship between lawmakers and administrators—with a shared intentional orientation, mutual responsiveness, and mutual support (Cordelli, p. 215). In particular, with respect to that last prong: mutual support requires “integrated procedures” (i.e., administrative procedures that “successfully integrate the democratic and the bureaucratic”) and a “bureaucratic ethos” (Cordelli, p. 217).
If all of this is satisfied, then we can say that public officials are carrying out the shared will of the people.
The problem with privatization, though, is that these three conditions themselves require two preconditions: (1) “broad convergence on the end of the joint activity” and (2) “contained alienation” (Cordelli, p. 219). Cordelli adds a third condition: (3) “procedural integration,” as a result of which the actions of the people involved can be attributed to the practice of lawmaking as a whole. And, Cordelli argues, “in the privatized state, private actors systematically fail to meet all the three conditions identified above” (Cordelli, p. 222).
- Broad Convergence on the End of the Joint Activity
Private parties are likely to have a different “object to which their action is intentionally oriented and that guides their decision making” (Cordelli, p. 222). This isn’t a question of motive—“different motives can support the same intention to pursue a shared goal” (Cordelli, pp. 222–23). Rather, it’s because private organizations, “unlike officeholders, exhibit” a “plurality of goals” because of their “free purposiveness” and because they’re “creations of contract rather than of office” (Cordelli, p. 223). Private organizations’ decisionmakers are also “unlikely to have the same level of identification with the goals of governmental institutions as those who choose a career in public service”—not because they’re “naturally more selfish,” but rather because of their different “process[es] of socialization” (Cordelli, p. 224).
- Contained Alienation
“Although the problem of alienation pervades all institutional settings, from public to private organizations, there are reasons to believe that privatization aggravates this problem in distinctive ways” (Cordelli, p. 225). Here, motives can be relevant, insofar as motivation can help support an intentional orientation; here, Cordelli cites some empirical evidence supporting the idea, among others, that public servants have “a higher level of community-service motivation” (Cordelli, p. 226).
“[P]rivate organizations are often positioned outside of the procedural structure that generates tight deliberative relationships between administrators, elected officials, and the public” (Cordelli, p. 226)—a structure that includes the APA and other procedural statutes. And this is bad because “administrative procedures are important mechanisms for containing alienation”; “privatization, by fragmenting the community of policy makers and weakening procedural ties between them, threatens to foster alienation within the structure of government” (Cordelli, p. 227).
- Procedural Integration and Attributability
Finally, there’s an attributability problem: “[F]or an agent’s decision to be attributable to a certain institution, the agent must act within a stable, coherent, and unified institutional structure that, by providing a shared background framework, appropriately connects the practical reasoning and actions of the various participants, and confers unity to their acting” (Cordelli, p. 228). Practically, this requires administrative procedures, but private parties aren’t subject to these.
* * *
Cordelli concludes her section on the failures of privatization by arguing that we have a duty to exit the privatized state.[25]
Cordelli granted that democracy could be just a bunch of people imposing their unilateral will on the dissenting minority, but then said that there was hope (Cordelli, p. 66). How realistic is that hope?
Not all formally democratic structures are necessarily legitimate; democracies must satisfy some other requirements. For instance,
those who participate in giving practical reality to the omnilateral will through democratic procedures should endorse the creation of a rightful condition as the goal of their collective practice. This means, for example, that citizens should refrain from exercising their democratic rights toward purposes that, even if not self-interested, explicitly contradict the substantive aims of the omnilateral will. . . . For example, it is hard to see how voting for a party that proposes to incriminate adults who engage in homosexual sex or that denies basic welfare to its citizens could ever be compatible with a condition of equal freedom (Cordelli, p. 69).
“[D]emocratic procedures” should also have “a strongly deliberative component”:
By providing each other with reasons that are both intelligible and pertinent to the aim of their collective decision, and by allowing the force of the better argument to eventually prevail, citizens treat each other as active participants in the construction of a shared will—the will of no particular group of persons (Cordelli, p. 70).
And nobody should be a “member[] of a permanent minority” in such a democracy, lest they find it unable to “regard themselves . . . as partaking in a shared political will” (Cordelli, p. 70). (See also the discussion of the requirement of “mutual support,” i.e., non-active-undermining, in the discussion of politics as a “shared cooperative activity” (Cordelli, p. 210).)
This is, to put it mildly, ambitious. Does this at all resemble our polarized, dysfunctional democracy? As I write these words at the beginning of Donald Trump’s second term, each half of the country thinks the other half is not just wrong, but evil and acting in bad faith—not open to persuasion, and adopting public-good rhetoric to mask their self-interest. Many people believe the system is run by shady cabals and rigged against particular groups—though they differ on who the cabals are and who’s oppressed.
Indeed, the omnilateral ideal isn’t even an ideal for many. In the standard interest-group model, politics is just about getting your group to win; democratic compromise (or constructing a shared will) isn’t an ideal, but an unfortunate consequence of not getting a majority. Because Cordelli’s conditions don’t come close to being satisfied and most participants don’t even treat it as an ideal, perhaps democracy is just a bunch of unilateral impositions.
Whether Cordelli’s view of legitimate democracy is realistic is doubly significant. First, and primarily, if Real Democracy diverges too significantly from Ideal Democracy, then the presumption in favor of these “presumptive nonrefutable reasons” (Cordelli, p. 67) may be refutable after all.[26] And second—more relevantly here—it undermines her case against privatization, because even if privatization is illegitimate, we’re just choosing between two different kinds of illegitimacy, and her framework doesn’t say how to resolve that question of degree.
Cordelli’s argument consistently assumes, though, that democracy is realistically reformable while privatization is hopeless. “[F]ar from being the solution to, or even an escape from, the vexing problem of bureaucratic unilateralism, privatization makes a potentially tractable problem into an intractable one, thereby reproducing, within the state itself, a condition that is structurally and normatively homologous to the Kantian state of nature” (Cordelli, p. 83 (emphasis added)).[27] And the problem is “intractable” because the private-sector problems are endemic: “[W]hereas public agencies who act for the sake of private ends are conceptually failing in their raison d’être—this is why, after all, we talk of ‘corruption’—the same failure is, by contrast, endemic to private actors, even appropriately constituted ones” (Cordelli, p. 184).
The problem with Cordelli’s justification of democracy carries over to the justification of agencies as well. As if it weren’t hard enough to legitimate democracy itself, modern conditions require hiring enforcers and administrators; but given their unavoidable discretion, their work is also just unilateral imposition unless accompanied by the proper ethos. And even an ethos wouldn’t be enough, so she suggests “codetermination” for bottom-up legitimation of the administrative state. But incorporating members of the public selected by lot seems immediately problematic. Why aren’t their decisions unilateral impositions?[28] After all, the chapter on legitimizing lawmaking teaches us that government by lottery would be “arguably lacking with respect to the condition of active citizenship . . . and of genuine representation” (Cordelli, p. 66). (See also Cordelli’s discussion of the “Taking Turns at Unilateral Subjection” hypo (Cordelli, p. 53).)
Every fix for unilateralism begets more unilateralism. We’re not good enough for Cordelli’s Kantian world; we need a New Kantian Man. But if we’re going to imagine that sort of idealistic, “ideally but realistically conceived” (Cordelli, p. 113) scenario, then we should be similarly imaginative about what’s possible under privatization.
Repeatedly, Cordelli criticizes privatization by pointing out that private contractors usually aren’t subject to the public-law constraints that govern public agencies (see, e.g., Cordelli, pp. 193, 199, 227). Remember that, from a Lockean perspective, none of Cordelli’s desiderata (not the proper orientation or ethos, not shared cooperative activity, not impersonality) are necessarily desirable, though some might be good ideas from an instrumentalist perspective. But let’s assume arguendo that such public-law constraints are desirable. An easy answer is that nothing prevents us from insisting on such constraints as conditions of privatization—this is the Regulation by Contract principle.[29] We have freedom of contract, and though a firm exists prior to its contract and needn’t accept a contract unless it fits with its “interests and goals” (Cordelli, p. 181), the same is true of the government, which is entitled to announce non-negotiable terms and refuse to contract with anyone who doesn’t accept them.
The requirements that could be imposed by contract or statute are infinitely various. They could involve public participation, transparency, fiduciary duties (which is already assumed in some areas, say if the government hires private lawyers for particular jobs) (see Cordelli’s concern at Cordelli, p. 182), and so on. They could limit the “institutional dualism” (Cordelli, p. 175) that happens when a firm’s management imposes profit targets on top of its contractual obligations. Private organizations have “free purposiveness” and can “form and pursue comprehensive ends” (Cordelli, p. 142); but part of one’s free purposiveness involves one’s ability (and, indeed, freedom) to suppress any comprehensive ends except for “Whatever you say, boss” in exchange for money.
But repeatedly, Cordelli makes several moves to undermine the possibility of saving privatization through such “public-ization.” First, doing so would be contrary to the logic of privatization. Second, if this were possible, it would wipe out the benefits of privatization, so firms wouldn’t want the job. Third, if the government did incorporate such conditions and private firms accepted them, that would be bad because it would undermine private firms’ freedom. And finally, if all this were done successfully, the private contractors would essentially have become public, so there would have been no privatization at all.
Let’s look at these moves in turn.[30]
- Does Privatization Have a “Logic”?
A frequent move of Cordelli’s is to argue that particular defects of privatization arrangements are “robust” and can’t simply be “overcome through regulation,” because they’re “dictated by the logic of privatization and by certain constitutive features that distinguish (or should, ideally, distinguish) private from public actors. They cannot therefore be overcome without making privatization a self-defeating project” (Cordelli, p. 116). The “logic and phenomenology of privatization” is that it “aim[s] at once to improve both efficiency and flexibility while saving costs, and to compensate for a lack of specific capacities to respond to new situations” (Cordelli, p. 142). Regulation of privatization arrangements would work against this goal and violate the “prima facie case for privatization” (Cordelli, p. 142; see also Cordelli, pp. 143, 145, 170, 194, 227, 229).
“The logic of privatization,” “the prima facie case for privatization,” or similar phrases—especially with the definite article “the”—show up so often that one might almost be seduced by the “the” and think that privatization has only one rationale, and that the only possible proponent of privatization is some myopic, efficiency-minded accountant. But the reality is more complex, and privatization has no the logic or the justification.
Why privatize? Is it to increase flexibility? To save costs? To compensate for a lack of governmental capacity? To cut red tape? To foster innovation? Any of these is possible, perhaps all of these. None of these goals is necessary. Much privatization may have been done primarily to save costs, and many privatization advocates may have been primarily (perhaps myopically) motivated by cost savings. Indeed, some privatization has been done out of desperation, in response to fiscal crises. Some of this may have been wrong-headed. Let’s assume, arguendo, that all past privatization was done for bad reasons.
But past privatization undertaken for bad reasons doesn’t support a fundamental critique. This is the Privatization History Fallacy. It’s open to us—inspired by Dorfman, Harel, and Cordelli, and even instrumentalist critics—to reject those bad reasons, and contingently endorse privatization only insofar as it’s done for better reasons.
When private firms contract with each other, smart businesspeople know that the best contract isn’t necessarily the one that seizes on the most obvious cost savings. Contracts should incorporate good monitoring and good incentives, lest your partner save costs by sacrificing quality. Some contractual terms are costly, but if someone complains about the cost of this extra term, you can correctly answer that this apparently costly term is necessary to mitigate sloppiness, self-dealing, cheating, etc.; paradoxically, the apparently costly term can increase profit. Indeed, there’s no paradox: it’s just the boring insight that there are tradeoffs in life; an apparently money-saving move might generate unintended consequences that reduce or eliminate the savings; and conversely, an apparently costly protection can alleviate these consequences.
If smart firms know this, why can’t governments? A ruthlessly efficiency-oriented politician can favor privatizing based on hard-headed cost-benefit analysis—and can also recognize that the benefits of privatization require significant investments in monitoring.[31] And this ruthless politician can be joined by Kantian privatization skeptics. The skeptics don’t even need to be a majority: they could just be the swing voters necessary to create a pro-privatization majority, in which case their insistence on accountability can mold the next generation of privatization programs, in which governments will retain epistemic and practical control.
In short, privatization has no single justification, no single internal logic, no single rationale. Privatization is just a policy, with a variety of possible rationales, which can be contingently supported (depending on how it’s structured) by a variety of constituencies. Focusing on just one simplistic rationale—even if adequately grounded in the history of privatization so far—may be a good reason for voting against privatization in a particular political context (perhaps we don’t trust the politicians responsible for the past failures), but it’s not a robust philosophical argument against possibilities for reform.
- Would “Public-ization” Wipe out Privatization’s Benefits?
The second problem with reforming privatization is that—again, because constraining contractors would be so burdensome—firms wouldn’t accept such contracts. There’s some overlap between this point and the previous one, but I’m listing it separately because the “burdens” here include not only the monetary burdens of monitoring and the like but also the burdens on the contractor’s ability to pursue its mission:
[P]rivate associations, unlike public offices, have a weighty interest in organizational autonomy that directly derives from their free purposiveness[, which] sets limits to what contractual offers it may be ex ante rational for a private entity to accept. For example, it may no longer be rational for private organizations to accept a government’s contractual offers if the latter were made conditional on the recipients accepting heavy constraints on their organizational autonomy and ability to express their conceptions of the good, for the independent pursuit of such conceptions of the good is often the very raison d’être (or at least one of them) of those organizations (Cordelli, p. 147).
Also, while the previous point could be interpreted as a conceptual point about the “logic” of privatization, this point is more of an empirical point about firms’ unwillingness to accept the contracts: “[T]he foreseeable prospect of strict monitoring, review, and oversight, as well as intense government pressure and strict regulation through public norms, would likely make it irrational for private actors to accept government contracts” (Cordelli, p. 195; see also Cordelli, p. 227).
But who are we to speculate on what a private organization would accept? Perhaps some would never agree to deliver a service if it meant suppressing their identity or organizational mission, or consenting to costly regulation. But others might. The way we would figure this out is by soliciting bids, making it clear what’s required, and seeing who steps forward. We shouldn’t make that decision for the organizations in advance—not even offering because we think they wouldn’t like it.
Perhaps lots of organizations would accept the burdens, but only for a higher contract price—and perhaps that would erase the gains from privatization. Perhaps—but there, too, we shouldn’t speculate; some organizations may not care about anything except making money by delivering services the way the government wants. This is a good reason to, where possible, maintain a regime of competitive neutrality,[32] where the government compares whatever bids come in against a backstop of in-house provision; then, the government won’t have to accept any bid that isn’t beneficial to both parties.
- Would Impersonality Unacceptably Burden Contractors?
The first problem was that we wouldn’t reform privatization because that would be contrary to its logic. The second was that if we tried to, providers wouldn’t want to participate. The third is that if providers did participate, the necessary reforms—particularly the requirement that providers suppress their organizational missions and act more like public servants—would unacceptably burden those organizations’ “associational autonomy” grounded in their “free purposiveness” (Cordelli, p. 142). In fact, this would undermine “the pluralism of associational life” (Cordelli, p. 147), “would amount to other spheres of society being co-opted by the administrative rationality of the state itself,” and would constitute a “normative colonization of civil society by the state” (Cordelli, p. 194).
But let’s consider the nature of these supposed costs for pluralism. It’s true that private people—and the organizations they create—have freedoms, including associative rights. But one way they can exercise their freedoms is by choosing to contract with others—recall the “Freedom of Contract” Principle that’s a corollary to the “Regulation by Contract” Principle. As part of these contracts, they can insist on giving free rein to their identities and preferences, if that’s important for them. (Some find their missions all-important; others just want to provide the service or make money.) Or, if this is sufficiently important to the other party, and for appropriate compensation, they can agree to wear someone else’s badge. Nobody’s forcing anyone to agree to any contract, so by agreeing to suppress their symbolism and adopt someone else’s, they’re not giving up their freedom or associative rights—in fact, they’re exercising their freedom and associative rights. And if people or organizations can exercise their freedom by making such contracts with other people or organizations, they can do the same by contracting with the government.[33]
These supposed associative costs ring especially hollow when one considers that Cordelli’s alternative is to radically insource these activities—which involves delivering the services using public officials who are required to suppress their individual views in favor of their bureaucratic ethos. Cordelli doesn’t suggest that privatization would increase or decrease the total number of people doing the work, so the number of extra civil servants resulting from insourcing could even be the same as the number of private individuals whose associational freedoms we’re trying to protect. If contracting to be impersonal in the delivery of a service has serious costs for associational freedoms, surely delivering the same service by hiring all the same people to be impersonal as a matter of bureaucratic ethos must be the same sort of imposition.
- Would Adequately Reformed Privatization Be Public?
The final problem with privatization is definitional. Suppose we insisted on regulating private providers, and they accepted the contracts, and they didn’t mind suppressing their private missions. Would they even be private anymore? Or, in attempting to perfect privatization, would we have “blur[red] to the point of disappearance the distinction between public and private entities that makes privatization conceptually possible in the first place” (Cordelli, p. 148)? Cordelli writes:
[I]f private actors were genuinely forced to act for public purposes alone, and were fully embedded within the procedural structure of public administration and within the system of public offices, they would cease to be “private” in the relevant normative sense. They would acquire many of the features that, as we saw . . . , constitutively differentiate public from private actors (although, of course, they may retain some descriptive features that are often taken to characterize private actors). This would make privatization a conceptually empty term (Cordelli, p. 229).
Like Dorfman and Harel, Cordelli operates with a different definition of privatization than the one used in actual privatization debates. If the government wants to have a service delivered, and rather than use public employees, it chooses to solicit bids from private companies (perhaps even for-profit companies that are traded on the stock market), in common usage, this is definitionally privatization. It’s privatization even if the private organization’s employees agree by contract to wear government uniforms, pursue the government’s mission, and adopt fiduciary duties: those are just some of the infinitely diverse things that private people and firms can contract for under capitalism. Dorfman and Harel admit that their definition of “private” is somewhat “revisionary”; Cordelli puts it more strongly, saying that privatization characterized by such impersonality isn’t even “conceptually possible in the first place.” This would be surprising to the officers and stockholders of the private organizations, who thought they were just coming up with a successful form of outsourcing that would finally satisfy the Kantians.
I confess to being more irritated by this move than I should be. What’s wrong with adopting unconventional definitions, as long as you’re clear about what you’re doing? Rather than complaining about a sort of No True Scotsman Fallacy, I can declare victory, and so can Cordelli. I can propose a scheme where the government contracts with a (formally) private firm that commits to everything Cordelli requires; I can rejoice because this is successful privatization; Cordelli can rejoice because it’s consistent with her Kantian program and (therefore) isn’t privatization. Cordelli’s argument against privatization (as she conceptualizes it) can be recharacterized as a contingent argument in favor of privatization (as the term is commonly used). In other words, just like Dorfman and Harel’s book, this book isn’t actually against privatization, as the term is commonly understood.
But, as with Dorfman and Harel, this redefinition of a common term has some costs. First, using common terms with uncommon definitions is confusing to the lay public; I bet I’ll be spending a lot of time telling people “Yes, but pay attention to how they define their terms; those books aren’t actually against privatization as such.” Second, and relatedly, this loses valuable opportunities to contribute to the privatization debate; there could be opportunities for reform that could satisfy both sides, and the differences in labeling could make these sorts of political compromises needlessly complicated.
Third, though, the labeling issue is part of a pattern that recurs throughout Cordelli’s book, and that shows up in Dorfman and Harel’s book too. It’s a pattern of making all the most optimistic assumptions—giving every benefit of the doubt—to the public sector, and being skeptical of the private sector at every turn.[34]
I have a fantasy about my fictional company Volokh, Inc., which I mentioned earlier. Here’s my business plan. The Dorfman-Harel-Cordellians have won the last election, and now their government has an ambitious plan to massively increase the number of civil servants. Unfortunately, there are severe frictions in the civil-service training process. There are only a few civil-service training programs; the schools that offer these programs are unable or unwilling to massively ramp them up; and other schools aren’t stepping up to fill the gap. Fortunately, I have experience with civil-service training, and I have a plan to offer a program that’s substantively identical to the existing ones, but at a massively reduced cost (no fancy campuses and redundant educational bureaucrats for me). I’ll charge the same as the existing programs, and the difference goes into my pockets and those of my shareholders (the equity markets love the scheme, and my IPO was massively successful).
My program is called “Rent-a-Bureaucrat.” The graduates of my program will be my employees, but I’ll provide them to any agency that wants them and they’ll be socialized within the ethos of the agency where they work; they’ll wear their agency’s uniform, and they’ll be instructed to do whatever their agency bosses tell them to do. I commit to not interfere in any way. They’ll stay at their agency unless they choose to leave or are fired—under exactly the same protective tenure rules that apply to ordinary civil servants.
Ordinary civil servants get a salary, but part of their salary always goes to pay off their student loans; in my arrangement, their salary is paid to my firm, and I keep an equivalent “student loan repayment” component and pay them what’s left over. From the students’ financial perspective, this is exactly equivalent to the ordinary system; the only difference is that I make a profit because my costs are so much lower. They might even prefer the arrangement because I’ll even assume the risk that they don’t get hired and pay them regardless; and governments might prefer the arrangement because they’ll have more flexibility in which agencies they want to increase or decrease.
This fantasy arrangement of mine seems to satisfy the Cordellian legitimacy requirements. (For that matter, it also seems to satisfy the Dorfman-Harelian legitimacy requirements; these employees of mine are “public officials,” because they participate in the necessary community of practice; and because their agency “bosses” can always tell them precisely what to do, they’re subject to ongoing political control and thus their practice has an integrative form.)
But isn’t this case “fantastic”? Aren’t I “turn[ing my] back[] on the private purposes that provide the grounds for [my] operations,” and “withdraw[ing] from [my] basic commitment[] to maximize profits” (Dorfman & Harel, p. 112)? Isn’t it contrary to privatization’s “internal logic and dynamic” (Cordelli, p. 142)? Not at all: my private purpose is to make money, and I’m so confident in my business model that this is the best way for me to make money. That’s my basic commitment and my internal logic.
And so, I’m finally told: This isn’t privatization. Because of the community of practice and integrative form, Dorfman and Harel tell me that this is just public provision, under their functional definition which “may sometimes be revisionary” (Dorfman & Harel, p. 139). Cordelli tells me we’ve “cease[d] to be ‘private’ in the relevant normative sense”; we’ve “acquire[d] many of the features that . . . constitutively differentiate public from private actors (although, of course, [we] may retain some descriptive features that are often taken to characterize private actors)”; if I’m going to call this privatization, then privatization is “a conceptually empty term” (Cordelli, p. 229).
Oh, really? That’s news to me, a private entrepreneur who started up this private company to harness the flexibility of the private sector to help the government fulfill its goal. It’s news to my shareholders, who are enjoying the flow of private profits stemming from this innovative and lucrative arrangement. I used my private contractual rights to agree to the whole of list of neo-Kantian requirements—I myself didn’t believe all these requirements were necessary, but whatever, I’ll agree to anything, given a good enough contract price. And the entire arrangement was called privatization in the media and the legislature because everyone recognized this for what it was—a private, profit-making plan to help the government fulfill its goals using means outside the traditional governmental structure. The apparently merely “descriptive features that are often taken to characterize private actors” aren’t trivial; they’re really constitutive of privatization as it’s actually debated politically; doing what the government wants using organizations not owned by government and outside of formal governmental structures is precisely the point of privatization advocates, and a successful arrangement along those lines should be called privatization.
Yes, this is win-win, because different people can support this system for different reasons; I can call it privatization and you can call it public provision. Why does it irritate me? Because the privatization critics—instrumentalist and noninstrumentalist alike—have spent a lot of time documenting everything they think is wrong with privatization; I’ve spent a lot of time explaining that none of those “problems” (whether or not I agree that they’re problems) are inherent to the process; all of these things can be “fixed.” And when I’m done and triumphantly display my finished product, only to be told that my privately owned, profit-making enterprise is really just public, it seems like goalpost-shifting. Saying that everything I’ve agreed to do is contrary to the nature of the private sector seems like a crude anti-private essentialism; it seems like, to some people, “public” just means “everything I like.” I know, I know—nothing prevents me from developing this arrangement anyway and getting it adopted, even though some people will label it using their own idiosyncratic definitions; this is just my own personal irritation, and I should get over it. Still, it rubs me the wrong way.
I opened this Review by stating that most fundamental critiques of privatization have suffered from at least one of two problems: either they’re not really fundamental, or they’re not really about privatization. I suggested that a Kantian approach, as practiced in these two books, had more potential along these lines, because at least Kantian approaches give an inherent role to the state (rather than merely treating it as possibly instrumentally useful to justice) and reject the interchangeability assumption.
Do these books deliver on the promise of providing a workable fundamental critique of privatization? No. Even if one accepts (as I don’t) their basic political theory of why “public” institutions are necessary for legitimacy, the critiques aren’t fundamental, because they hinge on the presence or absence of particular institutional features (like integrative form, public-law constraints, an ethos, etc.) that can be replicated within the private sector. And relatedly, the critiques aren’t really about privatization, because the authors adopt definitions of “public” and “private” that don’t track how the concepts are used in common discourse and in privatization debates.
These books might have been quite convincing if (like many policy analysis books) they had been phrased tentatively and practically: perhaps it would be a good idea to have contracts with greater political control; perhaps the WorkOpts contract was badly structured; perhaps we really need to care about the differences in organizational culture between civil servants and contractors; let’s see if we can take these problems seriously in future rounds of privatization, and if we can’t realistically do so, perhaps the whole enterprise is too risky and we shouldn’t endorse it.
These are reasonable points, though of course too modest (and too instrumental) for our authors’ ambitions. But it’s the authors’ need to make these arguments “fundamental” in a philosophical sense (e.g., Dorfman & Harel, p. 144; Cordelli, p. 9) that leads them into overly broad and implausible generalizations. At every stage, the authors make presumptions against the private sector and in favor of the public sector. Problems with public provision are regrettable departures from an achievable ideal; problems with private provision are part of the essence of the matter. The possibility of thoroughgoing regulation of private providers by contract is “fantastic,” contradicts privatization’s logic and justification (which is assumed to be some sort of simplistic cost minimization or the like), and wipes out any benefits of privatization. Past privatization failures are taken to be robust empirical generalities—rather than just observations about the sorts of people, with myopic concerns, who wrongheadedly chose those bad projects and whose example we ought to reject.
I don’t know what to call this approach, but whatever it is, it sure isn’t fundamental. Nonetheless, public-law scholars should read these books, because (if they’re unfamiliar with Kantian-inspired political theory) it will expose them to an interesting theoretical structure that’s foreign to how we do business in American constitutional law. And privatization scholars should read these books as well—privatization critics so they can add an impressive set of arguments to their toolkit, and privatization advocates so they know the most sophisticated arguments to respond to.
* Professor, Buchmann Faculty of Law, Tel Aviv University; Phillip and Estelle Mizock Chair in Administrative and Criminal Law, Faculty of Law, Hebrew University of Jerusalem.
** Professor, Department of Political Science, University of Chicago.
- Professor, Emory Law School, avolokh@emory.edu. I am grateful to Andrew I. Cohen, Daniel J. D’Amico, Alexander H. Kaufman, and B. Pablo Montagnes for their help and comments.
[2]. As I discuss below, see infra Part V.A, Dorfman and Harel’s approach differs from Kantianism in various ways; Cordelli’s approach is more explicitly grounded in actual Kantian philosophy. Dorfman and Harel’s “approach to determining what counts as an ‘inherently public good’ rejects the Lockean theory of legitimation without thereby necessarily endorsing its Kantian counterpart” (Dorfman & Harel, p. 114 n.45). However, their approach can still appropriately be lumped into the broad Kantian orientation in this limited sense.
[3]. See Alexander Volokh, The Constitutional Possibilities of Prison Vouchers, 72 Ohio State L.J. 983, 1006–10, 1028 (2011); infra note 13.
[4]. Private status still plays a role at the remedy stage: federal inmates don’t get a remedy against private prison firms under the doctrine of Bivens v. Six Unknown Named Agents of Federal Bureau of Narcotics, 403 U.S. 388 (1971), but this is for the explicitly instrumentalist reason that state tort-law remedies are available. See Alexander Volokh, The Modest Effect of Minneci v. Pollard on Inmate Litigants, 46 Akron L. Rev. 287 (2013).
[5]. For a discussion of such matters, see Alexander Volokh, Prison Accountability and Performance Measures, 63 Emory L.J. 339, 347–64 (2013).
[6]. See Alexander Volokh, Privatization and the Law and Economics of Political Advocacy, 60 Stan. L. Rev. 1197 (2008).
[7]. See Alexander Volokh, Privatization and the Elusive Employee-Contractor Distinction, 46 UC Davis L. Rev. 133, 142–43 (2012).
[8]. HCJ 2605/05 Acad. Ctr. of L. & Bus. v. Minister of Fin., PD 47 (2009) (Isr.), available at https://versa.cardozo.yu.edu/opinions/academic-center-law-and-business-v-minister-finance.
[12]. Malcolm Thorburn, Reinventing the Night-Watchman State?, 60 U. Toronto L.J. 425, 441–43 (2010).
[13]. This is, for instance, what we partly do in U.S. constitutional law with private prisons. See supra note 3, accompanying text. Importantly, calling private incarceration “state action” for purposes of the State Action Doctrine doesn’t deny that prison privatization is a type of privatization. It just means we extend to these private actors the same constitutional norms that apply to governmental actors.
[14]. This important point has been made by Jody Freeman, Extending Public Law Norms Through Privatization, 116 Harv. L. Rev. 1285, 1287–88 (2003).
[15]. Dorfman and Harel often say “legitimate political authority” (e.g., Dorfman & Harel, pp. 3, 7, 16, 31–32, 35, 38, 42), suggesting that the bare term “political authority” doesn’t carry any connotation of legitimacy. When they use the bare term “political authority,” sometimes, indeed, it means this general concept that is independent of legitimacy (e.g., Dorfman & Harel, pp. 14, 15). But sometimes they use the bare term “political authority” normatively, to mean “legitimate political authority” (e.g., Dorfman & Harel, pp. 16, 17).
[16]. Avihay Dorfman & Alon Harel, The Case Against Privatization, 41 Phil. & Pub. Aff. 67 (2013); Avihay Dorfman & Alon Harel, Against Privatization as Such, 36 Oxford J. Legal Stud. 400 (2016). I have previously argued against Dorfman and Harel’s views in Volokh, supra note 7, at 159–72; Alexander Volokh, The Moral Neutrality of Privatization as Such, in The Cambridge Handbook of Privatization 117, 125–32 (Avihay Dorfman & Alon Harel eds., 2021).
[17]. I’m omitting some other important critiques. For instance, why must criminal punishment and war necessarily be “public”? In the standard law-and-economics perspective, criminal punishment merely strives to encourage desirable behavior; the choice of public rather than private enforcers is a matter of convenience, having nothing to do with social vs. private communication. As for war, Dorfman and Harel note a “revisionist” view that “there is nothing morally significant about the practice of war that could detach it from the rest of morality, especially ordinary morality,” and that “the moral rules of engaging in war are set by reference to the question of what a private individual ought to do under similar circumstances,” but don’t provide a rebuttal against this view other than to note that it potentially opens the door to military privatization (Dorfman & Harel, p. 121). See also Cordelli, pp. 40–41 (“[Dorfman and Harel’s] argument has nothing to say to those who believe that the core function of punishment is rehabilitative or retributive, rather than communicative.”); Volokh, supra note 16, at 129.
[18]. Dorfman and Harel briefly mention that “striking down laws as unconstitutional is sometimes (or on some views) explained in terms of judges determining what the will of the American people was at the formative period of constitution-making” (Dorfman & Harel, p. 28)—citing Bruce Ackerman, We the People: Foundations (1993)—suggesting a theory according to which past generations can legitimately bind us. But Dorfman and Harel aren’t clear on whether they endorse Ackerman, or why past generations can bind us.
[19]. In any event, in some cases, we can implement a regime of competitive neutrality; in-house provision can compete with competitive private firms, and in-house provision can win if no firm is willing to take the offered contract at a price below some predetermined level. See Volokh, supra note 5, at 370–72.
[20]. This concern may be connected with Cordelli’s discussion of plural goals. According to Cordelli, a provider shouldn’t be able to act based on goals that “could not be reasonably justified to the citizens of a democratic society” (Cordelli, p. 181). Perhaps Dorfman and Harel, similarly, are objecting that the private organization’s “interests, concerns, and ideals” are particularistic whereas public agencies, including apolitical ones, don’t have this problem. But this is subject to the critiques that (1) private organizations’ plural goals can be addressed by contract or statute, and (2) public employees also have their own “interests, concerns, and ideals” that might differ from the ones desired by their agencies and that likewise couldn’t be justified to citizens. See infra note 30.
[21]. By “Kantian” political philosophy, I refer—as does Cordelli—not necessarily to Kant’s work as such, but to “the recent revival of Kantian political philosophy, found in the work of Katrin Flikschuh, Anna Stilz, and Arthur Ripstein, among others” (Cordelli, p. 46). See Katrin Flikschuh, Kant and Modern Political Philosophy (2000); Anna Stilz, Liberal Loyalty: Freedom, Obligation, and the State (2009); Arthur Ripstein, Force and Freedom: Kant’s Political and Legal Philosophy (2009).
[22]. But they disagree on which powers those are. For Dorfman and Harel, those powers relate to “inherently governmental functions,” whereas for Cordelli, it is “the power to make decisions that change the normative situation of citizens” (Cordelli, p. 8).
[23]. The two approaches also differ on a number of other points, such as (1) the nature of political control of the administration, i.e., top-down (Dorfman & Harel, p. 108–12) vs. also bottom-up (Cordelli, p. 111–13), (2) whether democracy is merely instrumentally useful (compare Cordelli, pp. 61–71 with Dorfman & Harel, p. 28; see also Alon Harel, The Kantian Case Against Democracy, 26 Crit. Rev. of Int’l Soc. & Pol. Phil. 243 (2023)), (3) whether representativeness derives from authorization plus deference vs. whether it requires a more internalist account (Cordelli, pp. 159, 169), and (4) whether actions in someone’s name must also be actions of that person (Cordelli, p. 41; Dorfman & Harel, p. 117).
[24]. Thomas Hobbes, Leviathan (1651).
[25]. Cordelli’s book ends with three chapters discussing “how to get there from here”: what philanthropists and service providers should consider their duties to be in the transitional state between privatization and in-sourcing, and how to create a democratic system of public administration. Those chapters are less relevant to my critique of Cordelli’s critique of privatization, and the points made in those chapters that are relevant to this Article have already been made in the previous chapters. Therefore, I don’t spend any extra space discussing those last three chapters.
[26]. Harel makes a similar point in his review of Cordelli’s book, using a hypothetical about cat-lover and dog-lover factions who merely want to “maximize their power and wealth rather than promote the public good” and who oppress each other whenever they are in power. Hence, he writes, “[n]orms that are democratically authorized need not necessarily or even presumptively count as omnilateral” and thus “democracy is not sufficient . . . for establishing Kantian legitimacy.” Alon Harel, The Kantian Case Against Democracy, 26 Critical Rev. of Int’l Soc. & Pol. Phil. 243 (2023). Cordelli responds to Harel in Chiara Cordelli, Kantian Democracy and Administrative Legitimacy: A Reply to My Critics, 26 Critical Rev. of Int’l Soc. & Pol. Phil. 267 (2023).
[27]. Later, at the end of Chapter 6, Cordelli addresses this argument, which she labels “the legitimacy charge”: “Some could argue that the collective action view is too idealistic or even naive . . . [I]f this view is true, no existing government appears to be legitimate, in the sense of being permitted or justified to exercise the lawmaking power it exercises over its citizens” (Cordelli, pp. 229–30). Cordelli argues that, “once properly understood, the demands of the collective action view could be realistically met by many governments.” First, legitimacy isn’t an on-off concept and “comes in degrees,” so some governments are more legitimate than others. People need to be able to assess whether their government is legitimate, so “[a] requirement of publicity is . . . an important, epistemic complement to legitimacy”; because the collective action view requires that people have the appropriate intentional orientation and we can’t know people’s internal states, we need to use “appropriate institutional benchmarks” like data concerning the level of corruption in government. “The collective action view is thus compatible with treating as fully legitimate all governments that, beyond respecting basic human rights and allowing for a fair process of democratic authorization, also meet the above institutional benchmarks—benchmarks that are not impossible to meet for any developed liberal democracy” (Cordelli, p. 231).
But this argument merely shows that governments can achieve some legitimacy. It doesn’t show that any particular government actually achieves sufficient legitimacy to be “treat[ed] as fully legitimate”; it doesn’t show that Cordelli’s actual preconditions are likely to be achieved; and it doesn’t show that privatization, with appropriate safeguards, is likely to be less legitimate than an actual, realistically achievable non-privatized democratic government.
[28]. See Cordelli, supra note 26 (granting that civic juries involved in codetermination don’t need to “achieve the kind of detachment required by the bureaucratic ethos” and can (unlike bureaucrats) “arguably still rely on their partisan affiliations or comprehensive conceptions of the good as both a motive and a reason to support certain policies”).
[30]. These are just the most common recurring problems. For reasons of space, I’m not focusing on various miscellaneous problems.
First, Cordelli assumes that improper political influence is a greater problem when services are privatized than when they are provided in-house (Cordelli, pp. 148–49). Any discussion of this issue should take into account the possibilities of lobbying by government agencies, and the possibilities of self-interested influence by public-employee unions. It should also take seriously the argument that privatization can alleviate improper political influence. For a partial discussion, see Volokh, supra note 6.
Second, Cordelli’s asymmetrical treatment of public and private also extends to her treatment of plural goals. When she discusses the private sector’s plural goals in the “problem of representation” section, she focuses on goals that “could not be reasonably justified to the citizens of a democratic society” (Cordelli, p. 181), like WorkOpts’s need to make 8% profit. By contrast, when she concedes that plural goals exist in the public sector, she only means more innocuous ones—the need to choose between valid goals like giving beneficiaries individualized attention and pursuing cost-effectiveness, both of which “have the same normative source: securing justice” (Cordelli, p. 181). She doesn’t consider that some public servants might not work hard because they prefer to relax at work, or treat the public (or program beneficiaries) badly because they dislike them or because they enjoy the feeling of power that their position gives them—which seems just as unjustifiable to the citizenry as the desire to make an 8% profit.
Third, Cordelli makes an “affective” argument that “when people do not see their own government as the main provider of the benefits they receive, they see little reason to care about their government and thus to actively participate in politics” (Cordelli, p. 147). This argument is broadly similar to Dorfman and Harel’s “privatization as detachment of the polity” argument that I critiqued in supra Part IV.B.3.
[31]. Contrary to what Cordelli argues (Cordelli, p. 143), lack of capacity to perform a function needn’t imply lack of capacity to monitor: I can tell whether house painters have performed well even if I can’t paint houses, and art critics can judge art even if they can’t create it themselves. Monitoring and performance aren’t the same skill.
For an excellent response to Cordelli on this point, see Cordelli, supra note 26, responding to critics of her codetermination proposal:
It is true that citizens must be able to acquire a certain level of specialized knowledge to serve in a civic jury, but the kind of expertise required to exercise veto power within such jury is not the same as the kind of expertise required to develop detailed regulations. To illustrate by means ofanalogy: an informed citizen can have the necessary expertise to judge that a history syllabus is ideologically biased or excessively narrow, while however lacking the knowledge, acquirable through the long-term study of history, to design a balanced and comprehensive history syllabus herself.
Id. at 273.
[33]. There are principles that limit the government’s ability to extract concessions from people who contract with it or receive benefits from it—this is the “unconstitutional conditions doctrine.” See, e.g., Volokh, supra note 3, at 1029–39. But even this doctrine agrees that the government can extract concessions that are relevant to the purposes of the program. See Alexander Volokh, Expressive Discrimination: Universities’ First Amendment Right to Affirmative Action, 77 Fla. L. Rev. 75, 135–40 (2025).
[34]. It may be worthwhile, in this connection, to read Cordelli’s recent article, What is the Wrong of Capitalism?, Am. Pol. Sci. Rev. 119 Am. Pol. Sci. Rev. 1806 (2025).