1 Introduction

The debate about privatization has become increasingly lively in recent years. One topic which has attracted significant attention is private prisons, understood as correctional facilities that are owned and/or run by non-public organizations, always at the behest of the government and typically for profit.Footnote 1 The emphasis critics set on private prisons is sensible given how common such prisons have become. For example, in the USA, for-profit private prisons house eight percent of the total prison population (Budd & Monazzam 2023). Outsourcing the administration of criminal punishment to private actors also appears to be an extreme instantiation of the trend towards increasing privatization that many countries display.

Critics tend to highlight two types of concerns about private prisons. Some critics worry that private prisons are problematic because of their effects on the quality of imprisonment services or on society more broadly. The view here is that while there is nothing in principle wrong about private prisons, in practice these prisons are likely to produce worse outcomes, such as increased risks of corruption, misuse or waste of public money, rights violation of inmates, and deleterious impacts on local communities (Logan & Rausch 1985; Dolovich 2005; Michaels 2010). Call this the instrumental wrong view.

By contrast, other critics more strongly worry that private prisons are (also) wrong because of features that are inherent to private organizations themselves. More specifically, they argue that even if private prisons did not have any problematic effects on the quality of imprisonment services or society more broadly, they would still be wrong in virtue of failing to meet some of the desiderata which an imprisonment practice is supposed to meet. Call this the intrinsic wrong view.Footnote 2 This second view has been most prominently formulated in terms of two substantively overlapping but analytically distinct arguments. According to the first argument, private prisons are wrong because, qua private organizations, they are bound to act on reasons other than those endorsed by the public in whose name imprisonment is carried out.Footnote 3 Call this argument, which has been most systematically articulated by Chiara Cordelli (2020), the representation argument. According to the second argument, private prisons are wrong because, qua private organizations, they cannot condemn offenders in the name of the state. Call this the condemnation argument. This argument has been most systematically articulated by Avihay Dorfman and Alon Harel (2013; 2016). Worth noting is that while these arguments could conceivably be used to undermine public prisons, too, they are deployed by these authors to speak against private prisons specifically. This means that there is an implicit assumption here that the arguments identify a flaw which is not present, if present at all, when it comes it to public prisons.

This article focuses on and analyzes these two arguments for the intrinsic wrong view. This is because, if successful, the intrinsic wrong view would constitute a robust and possibly conclusive case for resisting or ending private prison practices altogether, such that more careful probing is required before we accept or reject it. Furthermore, compared to the instrumental wrong critique, the intrinsic wrong critique is less hostage to empirical evidence about the performance of private prisons, and, as such, a more suitable target for a normative and conceptual analysis of the sort we offer here.

Our analysis advances two claims which, taken together, undermine the intrinsic wrong view. The first claim is that the arguments for the intrinsic wrong view are not fully persuasive even when applied to for-profit private prisons. This is noteworthy since the authors with whom we engage take for-profit private prisons as paradigmatic for private imprisonment practices. Our second claim is that these arguments are positively implausible when applied to alternative private prison arrangements recently defended in the literature—in particular, when applied to private nonprofit prison organizations (Shelby 2016; 2022). More specifically, we argue that if the arguments are intended to work for private prisons as a general class—and this seems to be the way these arguments should be understood—they fail to establish their conclusion, since there are situations in which private prisons are less problematic than public prisons.

The article is structured as follows. In Sect. 2, we reconstruct and assess the representation argument against private prisons. In Sect. 3, we build on some of the results yielded by this analysis to critically assess a suitably reconstructed version of the condemnation argument. Section 4 offers some observations about what our criticism seems to imply for the wider question of the justifiability of private prison arrangements. We argue that there is some reason to accept a mixed system containing both public and private prisons, especially if the private prisons are non-profit oriented.

2 The Representation Argument

Why would private prisons be intrinsically, rather than merely instrumentally wrong? One prominent answer is:

Representation argument: Prisons should be administered by agents that can adequately represent their principals. It is therefore essential that prisons which administrate sanctions for convicted offenders adequately represent the public. Private prisons fail to adequately represent the public.

The representation argument has been put forward by different theorists, including Dorfman and Harel 2013; 2016) and Cordelli (2020).Footnote 4 Here, we focus on Cordelli, since, to our mind, she has advanced the most systematic and sophisticated version of the argument.

Cordelli’s version of the representation argument is part of a broader account which concludes that, passed a certain threshold, the privatization of the state is a serious political wrong whereby the public illegitimately compromises its capacity to self-rule. The Representation argument offers one route to this conclusion. Its underlying contention is that private actors, “owing to certain constitutive features that differentiate them, qua private actors, from public ones, (…) systematically fail to act in our name” (p. 18; emphasis added).Footnote 5 Importantly, this contention does not deny the possibility that private actors could contingently or coincidentally succeed to act in the public’s name. Rather, the idea is that private agents cannot robustly represent the public and that, because of this, they cannot legitimately engage in practices that ought to be carried out in the public’s name—in particular, in basic practices such as punishment, defense or education (Cordelli 2020; Cordelli 2023).

For an agent to qualify as adequately representative of its principal, Cordelli argues that it is not enough that the agent is authorized by the principal to act in its name. The agent must also satisfy two further conditions:

The “included reasons condition”: the agent acts for reasons that are included, or at least not excluded, by the principal’s authorization.

The “domain condition”: the agent acts for reasons that constitute a reliable interpretation of the reasons that are included, or at least not excluded, by the principal’s authorization.Footnote 6

Cordelli claims that private agents cannot systematically satisfy these two conditions in relation to the public considered as principal. Concerning the included reasons condition, Cordelli (2020) notes that, because private agents in general, and private organizations in particular have “free purposiveness, that is, (…) [an] ability to form and pursue comprehensive ends and organizational goals that are external to the rationale of their public mandates” (p. 170), they are bound to act on reasons that are not included or, worse, are excluded from the reasons which they were authorized to act on by the public. Concerning the domain condition, Cordelli holds that, because private organizations are standardly socialized in interpretive schemes and cultures that differ significantly from those public organizations are conversant in – say, in interpretive cultures that prioritize profit and efficiency – they are bound to often engage in unreliable interpretations of the reasons on which they were authorized to act by the public (pp. 184–193).

Let us provisionally park the worry that these two conditions might be inconsistent since, if private organizations are authentically free to decide the reasons for their actions, it is unclear why they are fated to form or partake in interpretive schemes and cultures that are significantly different from those tracked by public actors. For now, we want to focus on each of these two claims taken in isolation—viz., that private organizations are bound to act on reasons that are excluded from those on which they were authorized to act by the public, and that they are bound to often engage in unreliable interpretations of those same reasons. Specifically, we will examine whether the claims hold true when applied to the case of private prisons.

On the first claim, Cordelli is concerned that the reasons for which private organizations often engage in the administration of basic services such as imprisonment, welfare or education are not the kind of reasons that the public’s authorization would include or, more strongly, that they are outright excluded reasons.Footnote 7

With regard to prisons, Cordelli can argue that publicly recognized penal aims—most notably, deterrence, rehabilitation, retribution, compensation and reparation—plausibly satisfy the included-reasons condition.Footnote 8 This is not only because each of these aims can be reasonably supported, but also because, within the jurisdictions that fall within the scope of Cordelli’s critique, these aims are often the result of a democratically defensible process of reasoning and deliberation.Footnote 9 Since public prisons are structurally bound to track these reasons, one can infer that they are well-placed to represent the public. However, because of their free purposiveness, private prisons are not structurally bound to track these reasons and can therefore bring in and act on reasons many of which will not satisfy the included-reasons condition. Consequently, private prisons fail to represent the public.

Cordelli takes acting for the profit of a private organization to be a paradigmatic example of a reason that does not satisfy the included-reasons condition. This seems true: a private organization’s profit cannot be defended as the public’s reason for imprisoning people—viz., private profit is not a reason of or, for that matter, for the public. The same seems to hold true for other privately held reasons that Cordelli (2020, p. 179) evokes—say, religious reasons that comprehensively guide imprisonment practices. The Quaker prisons run during the nineteenth century were arguably an example of private prisons instituted and administered on religious grounds.Footnote 10 Similar to private profit, redemption or salvation are not reasons of the public, at least not if we are considering non-theocratic polities.

Cordelli’s claim that some of the reasons that private organizations pursue qua freely purposive agents cannot satisfy the included reasons condition is thus convincing. However, this does not rule out that there could be alternative specifications of the kind of reasons that Cordelli identifies which would satisfy the condition. Private profit violates the included reasons condition if the profit is exclusively or largely distributed among the workers or shareholders of the private organizations owning or running the prison. But it is possible to specify private profit considerations such that they do not violate the included reasons condition—for instance, as acting for a private profit that is fairly distributed among the workers and shareholders of the private organization, the imprisoned offenders and, whenever relevant and feasible, members of the communities that are directly affected by imprisonable crimes or the prison industry. This latter reason still contains private profit as a consideration for which imprisonment is imposed but does so in a way that seems consistent with at least some of the reasons that are defensibly public—say, rehabilitation or compensation.Footnote 11

Similarly, it is possible to reformulate certain religious considerations that private prisons could act on such that they do not violate the included-reasons condition. For instance, while considerations such as penance or atonement have a demonstrable theological pedigree, it is possible to formulate them in a secular manner so that they constitute reasons that the public could authorize prisons to act on or, more weakly, cohere with reasons which are already publicly authorized. Rehabilitation or retribution, for example, seem like penal aims which could be aligned with underlying religious ideas of penance and atonement.Footnote 12

Given this, the representation argument for the intrinsic wrong view is not entirely persuasive. We do not deny, of course, that some private organizations that own or run prison facilities could violate the included reasons condition. Our contention is simply that the paradigmatic cases where the representation argument is supposed to apply robustly are amenable to modifications that undermine this argument. Put differently, the reasons pursued by private prisons can be rendered publicly palatable.

Here, supporters of the representation argument might insist that their critique still holds for actual for-profit private prisons. This rejoinder is persuasive, but it significantly reduces the scope of the representation argument to the effect of converting it into an argument about reforming current private prison practices rather than reducing or abolishing them. Moreover, the rejoinder makes it unclear why the intrinsic wrong view uniquely targets private prisons and not prisons in general, especially given evidence that, for the most part, public prisons establish carceral practices that are not particularly sensitive to the public’s reasons about such practices either.Footnote 13

Supporters of the representation argument might also object that we are too optimistic about the possibility of constraining the free purposiveness of private organizations such that their reasons can be rendered publicly palatable. In particular, they might contend that, even if we formulate profit considerations such that they satisfy the included reasons condition and inscribe these considerations in the relevant regulation, private prisons cannot reliably interpret or pursue these and other publicly authorized reasons. This is because, unlike public prisons, private prisons have organizational cultures that prioritize values that that are not publicly legitimate—for instance, a market-oriented or efficiency-based organizational culture.Footnote 14 Put differently, private prisons cannot reliably satisfy the domain condition whereby representative agents—in this case, prisons—have to reliably interpret the reasons of the principal—in this case, the public.

In reply, we argue that some of the values which purportedly structure private organizational cultures are, or can be, present in public organizations, too. This is arguably the case for efficiency, for example, which is certainly a value which is also relevant for public organizations.Footnote 15 More strongly, one could argue that there are private organizations whose organizational culture can reliably interpret and pursue publicly authorized reasons. This is arguably the case of nonprofit private prisons. Nonprofit private prisons have been defended on both instrumental grounds—for instance, for having better effects in terms of reducing recidivism rates (Bayer & Pozen 2005)—and intrinsic grounds—for instance, as a legitimate alternative in contexts where public state institutions have lost their legitimacy to imprison and punish (Shelby 2022; 2016).Footnote 16

The claim is that nonprofit private prisons can be funded through donations given by publicly minded private organizations or through smaller (one-off or ongoing) individual contributions. In short, private prisons would rely on endowment or crowdfunding mechanisms that, if adequately set up, can preclude profit considerations. Given the number of nonprofit private organizations that are already involved in securing important prison services—most notably, food provision, healthcare, and educationFootnote 17—private nonprofit prisons constitute a practically realistic alternative to strictly public prison arrangements.Footnote 18

Tommie Shelby (2022) has recently suggested that nonprofit private prisons might be more legitimate than public prisons in contexts of structural social injustice for which the state bears responsibility. He argues that in contexts where specific groups or communities—for instance, African American, Indigenous or Latinx peoples in the USA—are both economically discriminated against and overpoliced and/or overimprisoned by the state, public police and prison organizations act with “a serious deficit of legitimacy” (Shelby 2022, p. 143). Such contexts create the space for suitably designed private organizations to step in and pursue publicly defensible prison aims. Given this, Shelby thinks that “in the right hands, nonprofit private prison administration and services could be a viable, if temporary, alternative” (Shelby 2022, pp. 145–146). These prison organizations would not have the same legitimacy deficit that state institutions incur in structural injustice contexts. They would also have better prospects to involve and collaborate with socio-economically disadvantaged, overpoliced and overimprisoned communities in running non-state prison organizations.

Building on Shelby’s argument we contend that, while persuasive for many actual for-profit private prisons, the worry that private prisons cannot reliably satisfy the domain condition is implausible if applied to nonprofit arrangements of the kind Shelby outlines. The worry seems particularly misdirected in contexts of structural justice. This is because, if nonprofit private prisons manage to effectively work with private citizens who are disadvantaged and subject to disproportionate policing and imprisonment rates, and since such citizens are arguably in a good position to represent the public’s reasons and reasoning on imprisonment practices, nonprofit private prisons are in a better position to satisfy the domain condition in such contexts—viz., to reliably interpret and act on the public’s reasons for imprisonment.Footnote 19 This significantly weakens the representation argument case against nonprofit private prisons.

Here, one might further worry that, since nonprofit private prison arrangements are idealized institutions, they offer no evidence that private prisons do not violate the domain condition. In reply, it should be noted that nonprofit private prisons actually exist for juvenile offenders and function in ways that do not raise any particular concern for the included reasons or domain condition.Footnote 20 Furthermore, the worry seems disingenuous if advanced by defenders of the representation argument, who often engage in idealization to make the case that public prisons satisfy the included reasons and domain condition—viz., to argue that, if public prisons were to be more responsive to the public’s considered reasons for punishing and imprisonment and if prison bureaucrats would cooperate more with ordinary citizens, public prisons would be in a better position to satisfy the two relevant representation conditions.

More generally, the idealization worry cannot be straightforwardly supported from within the intrinsic wrong view. Remember that the intrinsic wrong view is that private prisons are illegitimate in virtue of their inherent features. Consequently, we should expect these inherent features to persist across a diverse set of private prison arrangements, including hypothetical ones. Given this, supporters of the intrinsic wrong view should welcome idealization as a method for testing the cogency of their claims rather than resisting it, at least as long as idealization remains realistic.

Finally, supporters of the representation argument might contend that, insofar as they are publicly minded, nonprofit private prison organizations are not genuinely private (cf. Dorfman and Harel 2016, p. 412). This contention is problematic on at least two counts. First, it equivocates on the meaning of the concept of private organization. It does so by moving from a legally descriptive definition of private organizations as privately funded and/or administered organizations to a moralized definition whereby private is equated with being non-publicly minded. Second, the contention is inconsistent with the free purposiveness claim that, qua private agents, private organizations have the latitude to choose the aims, purposes, and goals for which they act.

To sum up, our conclusion is that the representation argument is not entirely persuasive when deployed against all for-profit private prison arrangements, and that it is positively unpersuasive when applied to some nonprofit private prisons. Consequently, the argument is insufficiently robust to ground the intrinsic wrong view.

3 The Expressive Argument

So far, we have analyzed the argument that private reasons are intrinsically wrong because they are prone to act on reasons which are not endorsed by the public. A closely related but distinct argument is:

Expressive argument. Imprisonment is a vehicle for a polity to communicate condemnation for criminally wrongful acts. It is therefore essential that the imprisonment speaks in the name of the polity. Private prisons fail to speak in the name of the polity.

The difference between the representation argument and the expressive argument is that the latter is more focused on an ontological point. Its claim is not that private prisons are intrinsically wrong because they go about punishment for inappropriate reasons but rather that private prisons are incapable of delivering the condemnation on behalf of the state because of the kind of entity they are. To anticipate the discussion below, this is because they are insufficiently open to steering from elected representatives. The result is that people in private prisons are condemned by private actors as opposed to the polity.

The expressive argument has been defended by several theorists, including Mary Sigler (2010) and Dorfman and Harel. We focus on Dorfman and Harel here since they have set out the argument most carefully over series of publications.Footnote 21 It is worth noting that while the expressive argument could be deployed to challenge the condemnatory credentials of public prisons too, the debate generally assumes that public prisons can speak on behalf of the polity. This is also the view endorsed by Dorfman and Harel.

The expressive argument relies on several premises. It obviously relies on that punishment is (or should be) condemnatoryFootnote 22 and that the success (or appropriateness) of condemnation depends on the identity of the condemner.Footnote 23 Moreover, it assumes that the polity—i.e., a community governed by formal political institutions—is the appropriate condemner when it comes to criminal wrongdoing. While none of these premises is incontrovertible, they are plausible enough, and so we accept them here. More problematic is the assumption that imprisonment is the same as ongoing condemnation. This assumption is questionable, because one could argue that the condemnation of a crime is completed once the court hands down a sentence, with the ensuing imprisonment being merely an implementation of that condemnation. The possibility that condemnation is complete at sentencing would undercut the expressive argument since it would mean that the private or public nature of a prison has no import for condemnatory success.

We shall accept that imprisonment is ongoing condemnation, however, because we want to focus on the crux of the expressive argument, which is the idea that the polity cannot delegate condemnation to private entities without undermining its genuineness. We approach this question in two steps. First, we consider the more general question of whether those who hold a right to condemn may delegate acts of condemnation to other actors while remaining the source of condemnation. Finding this to be fully possible in some circumstances, we then consider whether there is something specifically problematic about a polity, acting through the state, delegating condemnatory acts to private entities.

3.1 Delegating Condemnation

The first question is whether it is possible for an act of delegated condemnation to count as genuine condemnation. By “genuine,” we mean that the condemnation must be relevantly connected to the actor holding the right to condemn. The idea is that condemnation is the prerogative of the wronged party, and that it would not be possible (or at the very least be inappropriate) for us to condemn someone unless this is somehow connected to the putatively wronged party. In other words, the critical remarks of a complete stranger cannot count as condemnation on our account since condemnation is an agent-relative phenomenon which presupposes some kind of relationship or interaction between the condemner and the one being condemned. In this sense, condemnation differs from moral criticism simpliciter.Footnote 24

To reduce the complexity of the issue, consider an interpersonal example where Condie has been wronged by Wanda and where we assume that Condie has an agent-relative right to condemn Wanda (say, because Wanda violated a friendship norm in her behavior towards Condie). Suppose Condie for some reason does not want to tell Wanda off in person and therefore enlists someone else to “do the talking” for her. It seems clear that there is nothing necessarily problematic, as far as the genuineness of the condemnation goes, about Condie delegating parts of condemning to another agent. Imagine that Condie were to pen a letter in which she blames Wanda but asks someone else to read it out, verbatim, to Wanda. We take it that this third party would then participate in the condemnation of Wanda. However, since the third party would simply read out Condie’s words, and be authorized to do so by Condie, it is unclear why this would render it any less Condie’s condemnation. To forestall the discussion below, we also do not see why it would matter if the third party would take out a fee for reading out the letter.

On the other hand, there are clearly examples where enlisting a third party would undermine the genuineness of the condemnation. Suppose that instead of writing a letter, Condie were to buy a “condemnation service” from a company offering to hold wrongdoers to account. We may imagine that an employee of the company knocks on Wanda’s door and proceeds to lambast her following a generic script developed for the purposes of telling off inconsiderate friends. In this situation, it seems that Wanda can rightly respond that the condemnation is too far removed from its original source to be taken as genuinely condemnatory. She is criticized by the company at least as much as by Condie. This is especially true if Condie did not go over and endorsed the script.

Why is it that the former case allows for genuine condemnation, but the latter does not? The reason, we contend, is fundamentally one of control. When Condie enlists someone to read out her letter, she controls the manner in which Wanda is condemned. The third party is more like a tool that Condie uses than an independent agent. When Condie buys the condemnation service without bothering to check its precise content, by contrast, she is more like a passive participant in a condemnatory process in which someone else is making important decisions. So, while in both cases we are dealing with third parties who are authorized by Condie to condemn Wanda on her behalf, in the latter case, Condie does not sufficiently govern the overall condemnatory act to make it her condemnation.

The emphasis on control suggests that there will be in-between cases where some discretion is exercised by the third party. Suppose, for example, that Condie writes a letter blaming Wanda but is unable to offer instructions about how the third party is to respond to all the various justifications or excuses Wanda might offer upon being told off. It is here unclear which conclusion to draw. It could be argued that Wanda has reason to regard Condie’s letter as a piece of genuine condemnation but disregard the improvised responses by the third party as equivalent to the irrelevant remarks of a stranger. Or one might argue that the third party’s improvised responses are also genuinely condemnatory considering that (a) the third party was authorized to act on Condie’s behalf and (b) the third party responds in a way that lives up to the spirit of Condie’s written words. It is quite clear, however, that the need for improvisation is a challenge to the genuineness of the condemnation. It gives Condie a reason to do the condemning herself since she could then handle the further conversation that may emerge herself.

3.2 Delegating Condemnation to Private Entities

Having noted that there is nothing about delegated condemnation which necessarily makes it less genuine, let us now consider whether a polity may delegate the ongoing condemnation of imprisonment to private entities. The key question is whether private prisons are incapable of participating in genuine acts of condemnation on behalf of a polity. Dorfman and Harel think so. They write:

These privatized activities are not the doings of the state since private entities vested with the formal authority to execute the activities in question cannot speak and act in the name of the state. However, the ability to speak and act in the name of the state is crucial for justifying a violent act (say, that of incarcerating a person or determining the conditions of incarceration, the means of disciplining the inmate, and so on). It is necessary for the punishment to communicate a judgment of the state (concerning the wrongfulness of the act), that is, a judgment made in the name of the political community it embodies. (Dorfman and Harel 2013, pp. 93–94. Emphasis added)

Why is it that private entities cannot, even if they work at the behest of the state in carrying out the punishment, “speak and act in the name of the state”? The answer again turns on control. Dorfman and Harel argue that the staff working in private prisons unavoidably enjoy discretion in carrying out the punishment. The wardens of private prisons in particular have extensive freedom of choice when it comes to things like living conditions, searches, security, and so on.Footnote 25 This means, Dorfman and Harel argue, that the imprisoned will often be subjected to the choices of a private agent. Although the wardens of private prisons have been authorized by the state to carry out the punishment on its behalf, this does not mean that they are simply acting as a vehicle of the state. The situation can be likened to the case where Condie instructs the third party to read out the letter, but the third party will also have to improvise in responding to Wanda’s possible rejoinders. In this case, the third party will not simply be a tool for conveying the condemner’s blame but will in effect subject the wrongdoer to the third party’s beliefs and preferences. This dynamic is what explains, so Dorfman and Harel argue, why private prisons are intrinsically wrong for expressive reasons. When the imprisoned are subjected to coercive choices of a private agents in the context of being condemned for a crime, they are no longer subjected to the entity that is supposed to do the condemning—the legitimate state, acting on behalf of the polity.Footnote 26

We accept Dorfman and Harel’s claim that discretion is unavoidable in running a prison and that it is implausible to suppose that a private prison could be so thoroughly regulated to make it a mere tool for the state (cf. Skarbek 2020). Indeed, meticulous regulation would defeat the very rationale for most private prisons, which is to harness the inventiveness and efficiency of the market (Diller 2002; Cordelli 2020, pp. 193–195). The question we should pose is instead why the discretion which surely exists in managing a prison does not also speak against the ability of public prisons to “act and speak in the name of the state.” Discretion, it seems, is a problem for everyone. Yet Dorfman and Harel want to argue that private prisons is uniquely vulnerable to the expressive argument.

Dorfman and Harel’s answer has to do with the way public prisons are organized institutionally. More specifically, they suggest that public prisons act and speak in the name of the state because they operate as part of a “community of practice” which “integrates the political and the bureaucratic in the execution of the relevant functions” and is “characterized by its principled openness to ongoing political guidance and intervention” (Dorfman and Harel 2016, p. 413). This makes public prisons “deferential” to the public point of view even though prison bureaucrats act with some discretion (Dorfman and Harel 2013, pp. 79–89). The claim is not, then, that public prisons act in the name of the state because they simply are the state, nor is it that public prisons are operated by people who act for public-spirited reasons as opposed to private ones. Instead, Dorfman and Harel’s point is that public prisons, unlike private prisons, act and speak in the name of the state simply because they are open to ongoing political control from elected representatives, much like a third party would be able to act and speak in the name of Condie in the example above if Condie were to continuously instruct the third party on how to go about lambasting Wanda. Dorfman and Harel grant that public prisons may also be staffed and run by people who are more likely to act for public spirited reasons compared to private prisons, but they do not regard this as the decisive factor. Private prisons would be unable to act and speak in the name of the state even if “public and private prisons provide identical conditions for prisoners” and the private prisons were “moved by public concerns” (Dorfman and Harel 2016, p. 408).

While there is something to be said for it, this view is ultimately unpersuasive because it overestimates the importance of the public–private distinction for the ability to condemn on behalf of the state. To see why, suppose public prison bureaucrats, like many professions, develop strong professional norms which go beyond merely implementing the laws. Suppose further that the politicians respect the autonomy of the profession and grant the bureaucrats considerable discretion in the way prisons are managed. Suppose finally that the bureaucrats adopt some norms and practices which are not aligned with the will of the public. It is unclear to us why prisons run by such bureaucrats would be more capable of acting and speaking in the name of the state than a private prison which is in fact highly deferential to the will of the public.Footnote 27

Dorfman and Harel’s answer is that openness to political steering is the important thing, not whether such steering in fact takes place. But this invites several responses. On the one hand, it raises the question why private prisons could not also be regulated so as to be open to “ongoing political guidance and intervention.”Footnote 28 Especially when we are dealing with ex post cases where unwarranted prison practices have occurred, it is unclear why the elected representatives would be in a worse position to intervene and correct when it comes to private prisons compared to public prisons. On the other hand, the mere possibility of steering does not seem a robust way to ground the ability to act and speak in the name of the state. When the imprisoned are subjected to the coercive choices of a public prison bureaucrat who is tracking professional norms rather than being deferential to the will of the public, even though elected representatives could intervene and change the way prisons are run, it seems that the imprisoned can complain that they are not coerced by the polity as much as by a representative of a profession.

To be clear, we do not deny that the possibility of intervening provides an excellent reason for saying that politicians are responsible for what is going on in prisons. The fact that a regulator chooses to take a hands-off approach when it could have steered does not exonerate it from responsibility, for in those cases we might say that the regulator tolerated or acquiesced in the results. Our point is that “acting in the name of” and “being responsible for” are different things. Elected representatives are responsible for coercive choices made in prisons simply because they authorized these institutions to inflict legal punishment, but this does not mean that these institutions speak in the name of the state when they in fact cease to be deferential to the will of the public. Of course, if the state’s acquiescence were sufficient to satisfy the “acting in the name of”-condition, then non-deferential private prisons would be able to act in the name of the state just as much as non-deferential public prisons so long as the state acquiesces in their activities.

Where does this leave the expressive argument? We believe we are entitled to draw two conclusions. The first is that the expressive argument is unlikely to support the claim that private prisons are intrinsically wrong for expressive reasons. Instead, much like in the case of the representation argument, the upshot seems to be that there are a set of contingent factors—such as openness to political steering or public-spirited professional norms—that tend to speak in favor public prisons and against private prisons. Just like we noted in our analysis of the representation argument, there is also a risk that commentators compare an idealized conception of public prisons with real-world examples of private prisons.

The expressive argument does offer, however, a potential explanation for why all private prisons are wrong and not just for-profit ones. Insofar as private prisons are insufficiently controlled by elected representatives to act and speak in the name of the state, any private prison will be vulnerable to the charge that it cannot condemn on the state’s behalf, and this even if it acts on reasons that the state in fact endorses. Now we reject that all private prisons are insufficiently controlled by elected representatives, or at least not insufficiently controlled in a way that is not also true for public prisons, and so we reject this line of reasoning. But it deserves to be stressed that the expressive argument has the advantage that it is less affected by the distinction between profit and non-profit private prisons than the representation argument.

The second conclusion is that the problem seems to lie with discretion rather than with the type of entity which owns a prison. Discretion is probably unavoidable when it comes to running a prison since laws, regulations, and policies will never be able to offer detailed guidance in every situation. But the problem discretion poses when it comes to condemnatory punishment persists whether we are talking about public or private prisons. In either case, it seems that some parts of a punishment are not decided by the actor who holds the right to condemn and thus becomes expressive of someone else’s judgment. A natural thought, therefore, is to assess the condemnatory credentials of prisons based on how well or badly their discretionary choices align with the intentions of the polity.Footnote 29 For the reasons given above, such an approach would not necessarily track the distinction between private and public prisons. It would not support the claim that all private prisons perform badly let alone worse than public prisons.

4 Conclusion

In this paper, we offered a critical analysis of two salient arguments which are supposed to show that private prisons are intrinsically wrong: the representation argument and the condemnation argument. We argued that neither of these arguments is sufficient to oppose all for-profit private prison arrangements, including realistically specified ones, and that the arguments are outright implausible when extended to some nonprofit private prisons. Of course, this does not rule out that many of the currently existing private prisons (many of them for-profit) are morally illegitimate. We fully accept that many private prisons are wrong, including for intrinsic reasons. All we have argued is that the critics have so far been unsuccessful in showing that private prisons are doomed to be wrong in virtue of being private.

While largely negative, our analysis elicits a few positive considerations that can be tentatively adduced in favor of private prisons. First, it suggests that, under non-ideal conditions where public imprisonment practices are wrong in several respects—say, because of mass incarceration or prison overcrowdingFootnote 30—some private prison organizations offer alternatives that could partly and provisionally tackle these wrongs. For instance, structurally linking private prison contracts to recidivism reduction might ultimately reduce mass incarceration and prison overcrowding, and thus somewhat prevent the wrongs which underlie these phenomena.Footnote 31 If so, supporters of the intrinsic view owe us a broader argument about how to balance between different intrinsic wrongs instead of a narrow argument about the intrinsic wrongs of private imprisonment alone.

Second, and relatedly, if we accept the claim that state institutions lose their legitimacy to punish and condemn some of their citizens when they are responsible for grave structural injustice, private prisons arrangements of the kind outlined by Shelby (2022) could provisionally fill in this legitimacy gap and allow the public to punish and condemn at least some of the serious crimes which affect it.Footnote 32

Third, private prisons could contribute to a partial rethinking of the aims of imprisonment—for instance, by adducing new reasons that justify imprisonment or by proposing new interpretations of standard justifications of imprisonment. This is valuable if we believe that, given their organizational commitments and culture, public prisons tend to track a homogenous set of penal reasons and that there is genuine uncertainty about the correct reasons to punish. Put differently, private prisons could be used as platforms for exploring and experimenting with different, and potentially more adequate prison institutional structures and rationales. This seems particularly plausible in a correctional system where the number of private prisons would be capped to allow for sound comparisons between private and public prison set-ups.Footnote 33