Can we escape privatisation dilemmas? Reflections on Cordelli’s The Privatized State

In the last decades, a broad privatisation movement has swept many nations, leading to formerly public services being outsourced to a variety of private parties. While lawyers, economists and social scientists have studied this development, political philosophers have had relatively little to say on the matter. However, privatisation raises important normative questions and concerns. Fortunately, with Chiara Cordelli’s The Privatized State, we now have a wonderful, closely argued, and much-needed in-depth philosophical treatment of privatisation. Privatisation refers to the outsourcing, delegating or contracting out of public functions (24). Many of these functions, while being ‘required by justice’, do not have to be provided by the state itself, in-kind. Instead, provision can often be delegated to private third parties. How to evaluate such delegations from a normative point of view? In chapter 1, Cordelli starts with a road map of existing philosophical positions on this topic. She divides them into two camps. Most liberal egalitarians are ‘instrumentalists’, since they follow economists in thinking about the rightfulness of privatisation as a matter of efficient outcomes (31). Against this, Cordelli maintains that we need richer normative frameworks, allowing for considerations beyond outcomes, pertaining to the process itself and its legitimacy. Others have provided non-instrumental arguments. One example is the commodification theorists who argue that privatisation corrupts the social meanings of goods (37). Against them, Cordelli restates the familiar criticism that such social meanings are often controversial. Liberal societies cannot therefore legitimately base their decisions on these meanings. Can we find a third way, transcending the instrumental/non-instrumental dichotomy? (43) Cordelli introduces her Kantian approach in chapter 2. Her aim is to show that there is ‘a fundamental and morally necessary connection between justice and public action’ (46). We can only live in a just society where there is a public authority guaranteeing justice. Such an authority is legitimised through an omnilateral will, ensuring nobody is subject to the arbitrary, private wills of others. This requires the familiar institutions of representative democracy (62–68). Cordelli leaves open the theoretical possibility that all of this may be compatible with a ‘privatization of governance’, as long as public authorities retain the monopoly on lawmaking and adjudication. However, in a second step, she wants to show that there are tight limits to this compatibility, for

privatization generates, at the lower level of governance, the very same problem that the justification of political authority is meant to solve to begin with. It does so by reproducing within the civil state the very same normative conditions that characterize and define the precivil statethe state of natureas a state of merely provisional justice. (46) This passage contains the essentials of Cordelli's argumentative strategy. Privatisation is, in Kantian terms, a fallback into the state of nature, where we (as citizens) are ruled not by an omnilateral will, but by the unilateral will of private providers. Hence privatisation is akin to a form of feudal patrimonialism (10,116).
By taking this route, Cordelli's book provides the authoritative statement of what a Kantian strategy would say about privatisation. In this it follows the lead of others who have recently developed approaches to contemporary problems inspired by Kant's political (in contrast to: moral) philosophy, such as Katrin Flikshuh, Anna Stilz, and Arthur Ripstein. This turns our attention away from distributive justice (in the post-Rawlsian mode) and efficiency (as economists do), and towards political authority and legitimacy. Do we still live under an omnilateral will if public functions are outsourced to an important extent?
One of the virtues of Cordelli's approach is that she doesn't attack the question of privatisation directly, but via the route of another problem: that of administrative discretion. In chapter 3, she argues that there is an inherent tension in the administrative state itself. The Kantian requirement that citizens are subjected only to an omnilateral will encounters the real-life predicament that administrators often need to have a measure of discretion to make judgments in practice. This however threatens to usurp legislative and administrative powers that should remain public. Cordelli solves this 'problem of bureaucratic unilateralism' (83), by positing a range of conditions to make this problem tractable. The result is her 'integrative model' (84), which integrates a set of conditions that competing theories of administrative legitimation bring forward as unique solutions (97). The three conditions of the integrative model refer to (i) a valid authorisation of administrative decisions through democratic channels (worked out in chapter 4), (ii) a representation of the will of the people by administrators (chapter 5), and (iii) an integration of the bureaucratic and the democratic elements in the exercise of discretion (chapter 6).
Privatisation comes in only after this analysis of administrative discretion; as a violation of the conditions of the integrative model. The advantage of this roundabout strategy is that it acknowledges there are no easy choices, since the problem is simply inherent in large bureaucratic states in which citizens rule indirectly, through an apparatus of countless civil servants. Still, privatisation is suspect for Cordelli, since it 'makes a potentially tractable problem into an intractable one' (83). My critical comments in the following are meant to inspire some doubt about the starkness of this claim.

The privatisation dilemma (1): for-profits
Cordelli sets out the authorisation condition in detail. She argues that privatisation decisions are invalid authorisations, since they are an 'abdication of the collective right to democratic self-rule, through the erosion of this right's minimal preconditions' (135). When people authorise private parties to fulfil public functions, they undermine the conditions of their own collective self-rule in the future. How does this work?
The argument relies on 'three robust causal mechanisms' (143), which Cordelli argues accompany privatisation and undermine collective self-rule. I focus on just one of them, what Cordelli calls 'directive control'. Cordelli argues that the case for privatisation is often made in terms of efficiency (cost-saving) and flexibility, in the face of a state's lacking capacities to deal with specific problems. However, this rationale undercuts the ability of the state to adequately regulate and monitor the industries it privatises. This necessitates subsequent rounds of delegation (of the task to regulate and monitor itself), with a further loss of 'practical' and 'epistemic' control from the state (143-145). This problem leads to what I will here call The Privatization Dilemma. One can always spend more money and resources on regulation and monitoring, to overcome an initial loss of control. However, by doing so the efficiency gains motivating privatisation from the start get lost. Privatisation becomes a 'self-defeating political project ' (194, also 227).
The question is whether the dilemma is as stark as posited here. The structure of Cordelli's argument seems to be zero-sum: whatever we gain in efficiency we lose in terms of democratic control. But this seems to me an overgeneralisation. Certain forms of privatisation in the past have been done with satisfactory results (telecommunication services come to mind, also air and rail transportation and energy infrastructure, when handled well), increasing efficiency and stimulating innovation, while still maintaining an adequate level of regulatory control. Or at least that's the kind of argumentwell-known from the economic literaturethat Cordelli's account needs to say more about. The problem comes back in chapter 5, in the context of the discussion of the representation condition.
Here Cordelli presents a case study of WorkOpts, a firm that contracted with the state for providing unemployment persons with job placements (171). The firm faced an 8% profit target, and as a consequence engaged in cream-skimming, focusing on the easier cases and neglecting the harder ones. This kind of cost-cutting strategy in the face of commercial pressure to make profits is familiar. But the case raises many questions that go unanswered. Were the goals expected of the company's service sufficiently specified by the state? Was public funding adequate to the task? Was there enough competition? These are all design challenges, when creating regulated public markets (or 'quasi-market'), i.e., markets for goods which fulfill a public function and hence are to be heavily regulated. 1 My point is not that such markets are easy to implement or maintain. Neither is my point to deny that there are trade-offs between efficiency and state control. But I would stress that in the context of public services, efficiency itself is a public value. For public resources are scarce, hence governments need to take into account how to get most (public!) value for their dollars. 2 And it just isn't clear to me that, as an intermediary form between completely unregulated markets on the one hand, and completely publicly provided services on the other hand, regulated public markets are necessarily doing worse in providing a satisfactory balance between these two values, for all types of public goods/services under consideration. To say privatisation is a priori illegitimate because it provides insufficient (or alternatively: only extremely costly) opportunities for democratic control, requires more argument. Regulated public markets need to be tested empirically and comparatively against purely private markets and purely public provision, and the details of regulation matter heavily. 3 My impression is that the optimum balance of both values may lie anywhere on the spectrum, depending on empirical circumstances.
Finally, wherever the argument goes, it does seem to make Cordelli's argument turn heavily on the kinds of efficiency considerations she rejected as insufficient at the start of her project. I appreciate how these considerations are here integrated into a political philosophical theory of democratic legitimacy. Nonetheless, despite the heavy philosophical armoury, the efficiency considerations do become the ultimate arbiter of where that theory of legitimacy goes.
3. The privatisation dilemma (2): non-profits I now turn to Cordelli's treatment of non-profit organisations. Cordelli's book runs the arguments against privatisation in a generic fashion, and she draws on both non-profit from for-profit examples while developing her arguments. The dynamics and problems are very different, though, and a bit more separation of these two contexts could have been helpful. Cordelli's argument with respect to non-profit privatisation also relies on a privatisation dilemma, albeit of a different kind. This dilemma too is perhaps less stark than she claims it to be.
Cordelli argues that government authorisation to have civil society organisations provide public services requires them to 'assume a public impersonal face' (147). This is because of the government's own imperative (familiar from liberal theory) not to express or endorse any particular comprehensive conception of the good life. This, however, clashes with these organisations' wish to express such comprehensive conceptions. These organisations are characterised by a 'free purposiveness', which is severely constrained when they become an arm of the state bureaucracy. In terms of a privatisation dilemma, the first horn is when a non-profit organisation can rule itself according to its comprehensive doctrine. People subject to its services are then ruled arbitrarily, subjected to the unilateral will of the non-profit. This is a fatal loss of democratic legitimacy (imagine a state-sponsored Catholic charity handing out foot to poor, Buddhist clients, on condition they convert to Catholicism). On the second horn, the civil society organisation is prohibited from expressing its own comprehensive identity, while providing services with an impersonal face, as the state would do when it would provide the service publicly. Then, however it is basically integrated into the bureaucratic apparatus, having lost the organisational independence which is required to make civil society a valuable sphere for free, associational activity (279-281).
Here too I think there are good reasons to reject the dilemma, and be less skeptical about whether mixed forms are always unworkable and dysfunctional. Many European countries have a tradition of schools, hospitals and other non-profit organisations providing education, health care etc. (i.e., public functions). In doing so, they are regulated and inspected by governments on many different aspects (from safety and hygiene to content of curriculum, etc.). Still, they maintain distinct identities. Citizens as private persons value having a choice between provider of different (religious, ideological etc.) orientations. In their public capacity, citizens legislate common public demands, which may differ between countries and times. Such systems do lead to tensions, when these values clash (e.g., teaching creationism in biology class). Of course these tensions would disappear in a purely public system. But that would lead to a complete loss in citizens' opportunities to choose between providers of different orientations. I would maintain citizen choice (like efficiency in the for-profit context) is a public value. 4 And a satisfactory balance between both values may be struck in a mixed system. It is at least not a priori clear to me that the balance is better in a purely public system of provisioning.
This point not only relates to organisations with religious identities. For example, academics in a university context also have (often strong) beliefs about what constitutes 'good research/teaching'. In countries where universities are publicly funded, academics also have to obey many government directives in terms of standards and regulations (say, on making publications open access, or installing curriculum committees). Depending on how many factors play out in the context, students may be just as well, or better off in such a mixed system, where a creative tension is in-built between the internally developed ethos of academics, and the democratically legislated demands of the larger public. A similar argument can be made as applied to publicly funded broadcasting. In conclusion: if we do not create such opportunities for private initiative in such important sectors as media and universities, schools and health care, but make them all 'purely public', tightly held to the bosom of the national state apparatus, then we relegate the energies of civil society to the sphere of our hobbies. In the end, I think this reduces the space for associational life more than Cordelli fears that the embrace of state funding does (279). 5

Back to the public functions question
With these points in mind, I would like to close by considering the question of which goods are to be considered fulfilling a public function, so that they fall under Cordelli's Kantian injunction not to privatise. The criterion of justice won't help us here, becauseas Cordelli has argued herselfmany important goods that everyone should have access to in a just society can be distributed through the market (think of food). Also, many goods governments use in the administration of justice can be bought on a market (e.g., the office supplies and computer hardware). I take it that there are two basic ways to look at this task of delineating public functions.
One way is to look at the extent to which the Kantian conditions (of authorisation, representation, and integration, see above) are fulfilled, at the level of individual goods. Once so and so much directive control is lost, for example, authorisation becomes illegitimate. Cordelli's take on the matter is different. She adopts an alternative analysis, where the matter needs to be judged at the level of all goods taken together. The wrong of privatisation is a matter of an accumulation of privatisations of individual goods; at some threshold point, we can no longer say that there is collective self-rule (150-154). On the analogy with the accumulated harm of carbon emissions, Cordelli argues that it is 'systematic privatization' (151) which erodes our democratic ability to rule ourselves. This may make it seem as if the specific types of goods do not matter. 6 There may be something to this holistic assessment of the matter. Still, we are faced with the necessity to make decisions between different baskets of goods, each of which could satisfy such a threshold condition. In Cordelli's argument, public functions closely related lo law-making, administration and adjudication figure prominently. She often mentions prisons and the military as examples. This would make sense, given how Kantian theory relates the omnilateral will to the three classical branches of the state. However, this would make it seem that only the functions of a nightwatchman state would fall under the category of public functions which should not be susceptible to the erosion of privatisation. Is this the best interpretation of where a Kantian argument would lead? This is arguably not Cordelli's intention, since she also often mentions welfare services as examples. But this example makes us enter the vast terrain of the public sector (that is: publicly financed sector, in my Dutch/European context) at large, with publicly financed but privately provided schools, hospitals, universities, broadcasting agencies, etc. Where to stop?
I have no suggestions on how to decide this question, but it seems that the systemic, holistic assessment of privatisation needs to say more about which public functions are the subject of its holistic assessment (similar to the fact that we need to know which gasses we need to count in, when assessing the accumulated harm of climate change).

Conclusion
Cordelli has written a very important book, which should spur intense discussion about privatisation. In taking an uncompromising stance towards privatisation, she challenges all who have taken more pragmatic, efficiency-oriented perspectives. In putting a Kantian account of political authority centre stage, she provides a strong antidote to all those political and public discussions which step over these fundamental considerations of democratic legitimacy. It contains many thought-provoking detailed discussions of legal, economic and philosophical theories about administration, authorisation and representation, which I have not been able to convey in this short piece. Above all, her book provides a timely warning signal about the dangers of privatisation, whichdespite the reservations above -I fully share.

Disclosure statement
No potential conflict of interest was reported by the author(s).