Anodyne Privatization

: Privatization of state services has been a flashpoint for political conflict over the past several decades. The goal of this paper is to explain why someone who is a supporter of the welfare state might also support the privatization of certain state services, in certain cases. Recent philosophical literature has focused on the most problematic privatization initiatives, especially the introduction of private prisons and military contractors. As a counterpoint, this paper describes a set of anodyne privat-izations, understood as privatizations that no reasonable person could object to. The key step in this analysis involves showing that privatization is not a unitary phenomenon. There are different types of privatization, different degrees of privatization, and also different motives for privatization. There are also important normative differences between these initiatives, which might lead a reasonable person to support some but not others.


I. INTRODUCTION
Public debate over the privatization of state services often exhibits a Manichean tone, as though the boundary between the public and private sectors was the site of an epochal struggle between good and evil, with the state on one side, representing the forces of justice, and the private sector on the other, representing the greed of capital.Discussion of the issue among political philosophers has been somewhat more nuanced, yet there is still a tendency to treat privatization as a unitary phenomenon, the desirability of which can be determined in the abstract, on the basis of fundamental principles of justice. 1 On top of this, the assumption remains widespread that all those who support the welfare state, along with its central mission of attenuating the injustices generated by the market, should be opposed to privatization in all forms.Thus a great deal of energy has been invested in the task of producing increasingly clever arguments against privatization (with particular emphasis on considerations that rule it out on deontological grounds), without much attention to the administrative difficulties that arose within the public sector that provided the primary motivation for it (Dorfman andHarel 2013, 2016;Cordelli 2021).
My central ambition in this paper will be to explain why someone who is a supporter of the welfare state might also support the privatization of certain state services, in certain cases.A great deal of the philosophical literature has focused on the most problematic privatization initiatives, especially the introduction of private prisons and military contractors (DiLulio 1988;Pattison 2010).As a counterpoint to this, I would like to describe a set of anodyne privatizations, understood as privatizations that no reasonable person could object to.This provides a useful framework for assessing the acceptability of any particular privatization proposal, because most can be situated between these two extremes.The first step in developing this analysis involves showing that privatization is not a unitary phenomenon.Most importantly, there are different types of privatization, and confusingly, different degrees of privatization.There are also importantly different reasons for privatization.Once it is made clear that a half-dozen quite different phenomena are routinely lumped together under the heading of 'privatization', it becomes far less surprising to discover that there are important normative differences between these various initiatives, which might lead a reasonable person to support some, but not others.Thus the normative assessment of privatization must be a great deal more nuanced than it has been, or than the current tenor of the philosophical literature encourages.

II. PRELIMINARY CLARIFICATIONS
As a prelude to any detailed discussion of the merits of privatization, it is necessary to issue a few clarification and qualifications.First of all, it is essential to distinguish core functions of the state from what are often described as its 'economic' activities (Stiglitz 1989).It is the latter that have most often been the focus of privatization efforts.There are a variety of ways of drawing the core-periphery distinction, some more tendentious than others.Perhaps the least controversial is to begin with the Weberian definition of the state as the institution that exercises a monopoly on the legitimate use of force in a territory (Weber 2004).The core components of the state can then be defined as those that are involved in constituting the institution with this power.This will include not only the agencies involved in securing and preserving that monopoly of coercive force de facto (that is, the military, police forces, the penal system), but also those involved in establishing the conditions of its legitimate exercise (that is, the legislature and judiciary).Although there has been some refinement of technique, these components of the modern state have been in place for centuries.In more recent years, however, it was discovered that an institution possessing a monopoly of force could also be an important economic actor, and indeed, that it could accomplish certain tasks that no private actor seemed prepared to take on.The result has been a massive expansion of state economic activity, particularly in wealthy industrialized democracies, where state spending as a fraction of GDP typically rose from less than 2% at the beginning of the 20th century to roughly 40% at the end (Lindert 2009).Areas in which the state has become a particularly prominent actor include banking, education, health, infrastructure, mass transportation, communications, and insurance (including pensions) (Moss 2002;Barr 2020). 2ost of the political debate over privatization has been focused on these economic functions, since this is where the bulk of state spending is concentrated.As a result of privatization, most Western states have withdrawn from ownership of automobile manufacturing (Jaguar, Renault), airlines (British Airways, Lufthansa, Air Canada), retail banking (BNP Paribas, Credit Lyonnais, Thomas Cook), railroads (British Rail, Canadian National), petroleum (British Petroleum, Elf, PetroCanada), mining (Charbonnages de France, British Coal, PotashCorp), and telecoms (Telstra, France Télécom, Deutsche Bundespost Telekom).There were particularly acrimonious debates in the 1980s and 90s over privatization of the water supply and railroads in the U.K., telecoms in France, electricity systems in the U.S., as well as the state petroleum company in Canada (Donahue 1989).The current philosophical debate, by contrast, has tended to focus on privatization initiatives that impinge upon the core coercive apparatus of the state, such as private prisons or the use of military contractors.Here it would not be surprising to discover that there are strong objections to privatization, stemming from basic philosophical principles.
It is generally agreed that the state does not arise spontaneously, through a mere confluence of individual interests or out of voluntary contractual relations.After a coercive public power has been established, then individuals can accomplish a great deal through private contracting, but these contractual relations are only binding because of the background guarantee offered by the public authority; they cannot serve to establish that power.As a result, the core coercive apparatus of the state operates on a different logic from that of any private institution.It is, therefore, courting disaster to begin replacing elements of that apparatus with private organizations.Most obviously, because of the dangers of corruption, there is considerable merit in compensation systems that neutralize pecuniary concerns for state employees who are directly involved in the exercise of coercive power.It is, of course, important to avoid the fallacy of thinking that those who work in the public sector automatically acquire a commitment to the public interest (or that individuals in the private sector can be expected to act in a self-interested fashion, merely because the organizations in which they work are committed to maximizing profit) (LeGrand 2011, 88).This involves a confusion of individual motivations with organizational objectives.It is, however, important to observe that public-sector organizations are in a position to institute management models that significantly reduce the risk of corruption or abuse of power in a variety of ways: by having strong employment security, relatively dull financial incentives (for example, seniority-based compensation), strong shared values (including patriotism, the rule of law, and equal citizenship), and a culture of deference to elected officials.There is such a thing as a 'public sector ethos', particularly at higher ranks of the bureaucracy, even though it stops short of making public officials purely disinterested (Pratchett and Wingfield 1996, 639-656).
In part for these reasons, most advanced welfare states have not been especially keen to privatize core state functions.There is also the simple fact that in most countries these functions are not particularly expensive, and so privatization does not have much upside from a fiscal point of view.Two prominent exceptions are the U.S., which maintains an unusually large prison population of over 2 million persons, with another 4 million on probation or parole (Harding 2001), and Israel, which maintains what amounts to a permanent military occupation over a subject population of more than 3 million (Seidman 2010).It is perhaps not surprising that, in these countries, the temptation to privatize elements of the coercive state apparatus is much greater than elsewhere, and that the controversies generated have occupied the attention of academics (Medina 2010).In most other countries, however, these issues are not particularly pressing.And so, in the discussion that follows, I will focus my attention only on the privatization of economic functions of the state.Apart from the fact that this is where most privatization has occurred, this is also where one is likely to find the more uncontroversial instances of it.
The second major qualification worth registering is that, in a complex mixed economy, the public sector is almost never public 'all the way down'.In other words, if one examines a particular public agency, engaged in the provision of some public good or service (for simplicity I will refer to both as 'goods'), there will inevitably be some elements that are 'contracted out' or purchased from private suppliers. 3The U.S. Postal Service, for example, operates over 300 vehicle maintenance facilities and employs over 4,000 mechanics, in order to keep its fleet of 150,000 delivery vehicles in a state of good repair.At the same time, it does not actually construct the vehicles-the 'Grumman LLV' that they use was built by Grumman and General Motors.This is fairly typical-if one digs deep enough, one will eventually come upon inputs to government activity that are supplied by the private sector.This boundary is often easy to discern, because most states are required to go through a public tendering process in order to acquire these inputs.The set of rules governing this acquisition process is extremely complex and forms the cornerstone of the arcane but enormously important domain of public administration known as 'procurement' (Thai 2001).
This observation is philosophically significant because it serves as a reminder that the distinction between public and private is often not either-or, but rather a question of fixing the boundary at which the public stops and the private begins.Failure to attend to this leaves some arguments against privatization, which impose an overly categorical prohibition, vulnerable to a reductio, because they wind up prohibiting any sort of procurement from the private sector.For example, if one appeals to distributive justice considerations to claim that the state should hire its own cafeteria workers, rather than contract out food service to a private company, on the grounds that the state pays higher wages, then it follows that the state should not be purchasing semi-prepared frozen foods either, but instead should be doing all food preparation in-house, because the state can pay its prep workers a higher wage, or that the state should not be purchasing food from private farmers, but rather should be operating its own agricultural facilities, because the state can pay its growers and pickers a higher wage, and so on.Anywhere that one attempts to draw the boundary between public and private, there will be an obligation to extend the public further, and so the argument amounts to a requirement of economy-wide socialism.
The other thing worth noting about this issue is that, at a certain level of granularity, the question of where the boundary should fall between public and private seems unlikely to be one that can be settled through appeal to first principles (Stark 2009).In practice, it typically gets made on the basis of simple considerations of cost, quality, and convenience (or 'net value') (Moore 1995).In this respect, public managers engage in deliberations that are strikingly similar to those engaged in by private managers confronting the 'make or buy' decision, the analysis of which has figured centrally in the transaction-cost theory of the firm.On this view, the boundary of the firm is determined by the scope of the employment relation.Various firm inputs, including labor, can be obtained through market contracting.The employment relation is also contractual, but differs in one key respect, which is that its terms are incomplete.Rather than agreeing to perform a precisely delineated task in return for some consideration, the way that a contractor does, an employee instead agrees to a relatively open-ended job description, but under the authority of management.In other words, the employee agrees to perform tasks as assigned.
The choice between public and private, at the boundary, typically involves the exact same question about the capacity of the state to achieve its objectives through contracting, rather than through its own employees.For example, while Canada Post also operates maintenance facilities for its delivery vehicles (mainly Grumman LLVs), more extensive refurbishment is contracted out to a private firm (Jasper Engines and Transmissions).While there may be some good reason, grounded in basic principles of political philosophy, for the postal service to be in the public sector, it seems rather unlikely that these same principles can resolve the question of how extensive a repair program must be before it makes sense to contract it out to a firm that specializes in engine rebuilding.The latter question is often a very difficult one, which involves primarily comparing the state's 'agency costs' (those involved in getting its own employees to do it) to its 'contracting costs' (those involved in getting an outside party to do it).These costs include not only the direct budgetary costs, including both salaries, payments, and the cost of supervision, but also indirect costs, such as delay, inflexibility, and opportunism, as well as hidden costs, such as failures of performance, quality defects, and shifts in riskbearing.
Public managers are, in my experience, neither stupid nor particularly ideological, and so typically are in a position to recognize when they are able to draw up contracts and oversee them effectively and when they are not.There is a major body of work in public administration-somewhat absurdly named-of 'deliverology' dedicated to the question of how to deliver results in government, in which questions about contracting, and the contractual specification of 'deliverables', looms large (Barber 2015).Contrary to the narrative of rampant neoliberalism, which registers only a one-way process of selling off state assets, there is actually a great deal of both out-sourcing and in-housing at the boundary of the public sector, along with significant movement back and forth (that is, not just privatization, but also reversals of privatization) (Hefetz and Warner 2004). 4  There is also a great deal of renegotiation of these contracts over time (Perard 2009, 202).So it is important to recognize that a great deal of privatization is not driven by political ideology or rival conceptions of justice, but by pragmatic managerial decision-making that merely adjusts the inevitable boundary between public and private.

III. THE TRADITIONAL PUBLIC-SECTOR MODEL
Much of the recent philosophical literature on privatization has not been informed by a particularly fine-grained appreciation of how state operations are carried out, or of how much variation in institutional practice exists from one jurisdiction to the next.In particular, when speaking of the 'public sector', many philosophical discussions appear to presuppose what Evan Davis refers to as the 'traditional public sector model' (TPSM) as the standard for service delivery (Davis 1998).The key characteristic of this model is that it involves the provision of some good by employees of the state, funded through general taxation, such that it is free to citizens at the point of service.A paradigmatic example of this is the way that policing is funded and delivered, or the way that primary education has traditionally been provided.The crucial feature of this model is the fact that the state stands on both sides of the transaction-as both the purchaser (that is, the source of funding) and the provider.
One can see the power that this model exercises over the philosophical imagination in the way that debates over public health care, or 'socialized medicine' have been conducted.Here the image of the National Health Service (NHS) in the U.K. looms large.What is striking about the NHS is not just that patients are not charged any fees at the point of service, but that doctors and nurses are salaried employees of the state.While granting that there is something to be said for this arrangement, it is important to observe that it is extremely uncommon as a model of socialized medicine.The more typical arrangement, such as one finds in Germany or the Netherlands, is one in which the state intervenes quite heavily in the health insurance sector, in order to guarantee universal coverage, but then leaves a great deal of the provision of health care services to the private sector.5 Furthermore, doctors are often not salaried, but rather are compensated on a fee-for-service basis.The absence of pointof-service fees (or co-payments) is also not universal-Sweden for instance charges user fees for physician appointments, hospital stays, and emergency room visits (Glenngård and Borg 2019).
More generally, it is extremely common to find arrangements in which the state acts as the sole provider of a good, but where citizens must purchase it in the same way that they would any other commodity.For example, even when the state acted as the sole provider of home electricity, consumption was always metered, and individuals had to pay for the amount that they used.(The old state-owned British Gas even installed coin-operated gas meters in many homes, so that residents would have to keep a supply of shillings on hand, in order to feed the meter when it began to run low.)The state often tacitly subsidized consumption by keeping rates low, but there was never any suggestion that consumption should be entirely funded by general taxation, and thus be free to individuals at the point of service.More generally, while the U.K. Labour Party was once committed to state ownership of the means of production, the goal was never to make the products of companies such as British Steel or British Leyland, much less Rolls-Royce, freely available to all.
There are also many cases in which the state acts as the sole (or primary) purchaser of a commodity (that is, a monopsonist), but where supply is completely private.Road construction is perhaps the most common example of this.The state is practically the only purchaser of large-scale roads and highways-which it typically then makes available 'for free' at the point of use to all citizens-but the actual construction of these roads is almost always done by private contractors.Construction of government buildings and other infrastructure, such as bridges, is the same, not to mention smaller items, such as signage, garbage cans, tree planters, and public art.These areas all feature a 'purchaser-provider split', where the state acts either as the purchaser, or the provider, but not both.The central feature of the traditional public sector model, by contrast, is that it involves the state acting as both purchaser and provider.
Given the ubiquity of purchaser-provider splits in the public sector, on might wonder why the TPSM has occupied so much attention in the theoretical literature.One reason is no doubt the influence of Paul Samuelson's definition of 'public goods' as those that are perfectly non-rival and non-excludable (Samuelson 1954).This has given rise to a great deal of confusion, since practically no goods provided by the public sector possess either quality.If one thinks of insurance, for example, and insurance-like arrangements, which constitute an enormous fraction of the economic activities of the welfare state, not only are the payments perfectly rival, but it is in most cases easy to exclude individuals from the benefits.But nevertheless, for goods that do happen to be relatively diffuse and optimally shared by many, the TPSM typically represents the most efficient mode of provision.
A non-excludable good is just the limit case of a good that generates positive externalities.If the entire value of the good takes the form of an externality, then it may make sense for the state to produce the good and make it freely available to all (think, for example, of direct air capture carbon scrubbing).But if the good generates some mixture of private value to individuals and positive externality for others (as with various forms of green energy), it may make more sense for the state simply to subsidize production of the good, or make it available to individuals at a below-market price, in order to promote optimal production of the externality.In other words, the appropriate price to charge for such a good is dictated by the fraction of the good that constitutes private value (that is, appropriated by the individual), and the fraction that takes the form of a positive externality.Making goods free at the point of service is appropriate for Samuelson public goods only because these are goods whose value would be impossible to monetize if an attempt were made to provide them privately.
The second major reason for an overemphasis on the TPSM is normative, stemming from the conviction that an important function of the welfare state is to achieve a 'decommodification' of certain goods.Michael Walzer is no doubt the most influential philosophical proponent of this view, with his claim that certain goods, such as health care, belong in their own 'sphere of justice', in which they are distributed in accordance with principles that are fundamentally different from those governing the distribution of market commodities.(His discussion echoes the NHS constitution, which states as its second fundamental principle that access to NHS services should be "based on clinical need, not an individual's ability to pay"-DHSC 2021.)For this reason, Walzer's vision of a public health care system is one in which "ownership is abolished" and "practitioners are effectively conscripted or, at least, 'signed up' in the public service.They serve for the sake of the social need and not, or not simply for their own sakes" (Walzer 1983, 89).In order to achieve full decommodification, on Walzer's view, the state must act as both purchaser and provider, because anything less than this will result in the transaction being mediated through a money price (with the state either taking bids from rival suppliers, or else providing goods to those willing and able to pay).
But again, it is important to observe how seldom the welfare state actually achieves, or has even aspired to achieve, this sort of decommodification.Even with the NHS, the suggestion that health care is distributed in accordance with need is something of an illusion.It is actually distributed in accordance with a physician's determination of the patient's needs.Unlike the U.S., in which patients can directly access specialist services, in the U.K. general physicians act as gatekeepers, limiting access to the system so that more expensive services are available only by referral.(This is a general feature of most 'needs-based' systems, such as elder care, in which access is always determined by a bureaucratic assessment of need-see Evans 2016.)But more importantly, complete decommodification is not necessary in order to institutionalize this access principle.The Canadian single-payer health system, in which the state acts only as the purchaser, leaving provision to the private sector, has an access structure almost identical to that of the NHS.Health care is distributed in accordance with need because it is paid for by insurance, and all insurance (whether public or private) offers indemnification based on the need generated by the insured event (Horne and Heath 2022).
Nevertheless, there is a widespread belief that the welfare state imposes limits on the market, by taking certain goods and both producing and distributing them in accordance with non-market principles.This is often approved of, not just because of a distaste for the commodity form, but also because of its egalitarian distributive consequences.Since the income tax system is progressive, it will typically be the case that the burden of paying for goods that are financed through general taxation will fall more heavily on the rich.This distributive argument typically underdetermines the case, however, because the same objective could be achieved by redistributing income, or by providing individuals with vouchers to access certain goods (Daniels and Trebilcock 2005).And indeed, this is precisely what the state does in many domains (ranging from food security to higher education access).Thus the egalitarian perspective offers no explanation for why the state needs to be more deeply involved on either the purchaser or the provider side of the transaction.
As a result, there has been a certain amount of puzzlement among political philosophers, generated by the fact that money is still changing hands in so many areas of welfare-state intervention, and a great deal of provision is being carried out-and always has been carried out-by the private sector (Radin 1989).These goods (such as public roads) are sometimes referred to as 'partially decommodified'.But this obviously invites the question, whether the objective was ever to decommodify these goods in the first place, or whether some other normative principle provides a more perspicuous reconstruction of the reasoning that led to them being in the public sector.The practice of partial decommodification is so ubiquitous that it tends to raise doubts about the plausibility of the decommodification standard.It is also worth noting that, unlike the postal service, which enjoys a state-imposed monopoly that prevents private firms from offering competing services, many other goods provided through the TSPM can be 'topped up' by private purchase.Typically, the state provides these goods at a very basic level for free, but allows individuals to purchase additional quantities, or higher quality versions, from the private sector.This is unlike goods such as transplant organs, or surrogacy services, in which the state prohibits all commercial transactions (Sandel 2012).By contrast, there is an active market for private health care services and health insurance in the U.K. Similarly, with garbage disposal the state acts as the primary purchaser, but businesses are typically obliged to contract directly with private firms to secure disposal.The police provide basic reactive security, leaving individuals free to purchase additional security (especially proactive security) from the private sector (Claassen 2011;Schafer and Fehling 2021).Thus these goods are not really being 'decommodified' through state involvement.
My own view, which I will mention briefly here, is that the best way to understand all of the primary domains of economic welfare state activity is to see them as responding to one or another form of market failure (Heath 2011). 6In some cases, such as education, state intervention is motivated by the presence of significant positive externalities (and the tendency of markets to underproduce them).In other cases, such as health insurance, it has to do with information asymmetries (and the fact that the state is not vulnerable to adverse selection).In other cases, it has to do with market power (and the tendency for natural monopolies to form in the private sector).The important point is that market failures can occur on either the demand side, or the supply side, or on both sides of a market.It is only in the latter case that the TPSM is necessary as a response; in other cases state intervention is only required on the purchaser (demand) side or the provider (supply) side, depending on where the market failure is occurring.

IV. TYPES OF PRIVATIZATION
Given the prevalence of purchaser-provider splits in the welfare state, it seems unlikely that there could be any principled objection to such arrangements (certainly there can be none that appeal to precedent and best practice).Otherwise put, the insistence that all economic activities of the welfare state conform to the TPSM is excessively rigid and rules out a number of highly successful welfare-state arrangements (such as the Canadian health care system, or the U.S. Social Security system).Furthermore, it runs into serious difficulty, and perhaps conceptual incoherence, once the boundary problem is recognized (that is, how far 'down' does the TPSM have to go?).Does anyone care that mail delivery vehicles are not being built by government employees?But then does it matter whether they are maintained by government employees?
It is useful therefore to distinguish the TPSM from a purchaser-provider split model (PPSM), in which the state acts either as the primary (or sole) purchaser of a good, or as the primary (or sole) provider of a good.The former can be described as the demand-side public sector model (DPSM), the latter as the supply-side public sector model (SPSM).This then generates a variety of different forms of privatization (labelled 1 through 5 in Figure 1).In principle one could move from the TPSM to a fully private market (FPM), characterized by private provision and funding through direct purchase, although this is relatively uncommon.Most privatizations have involved movement either from the TPSM to a PPSM (that is, the introduction of a purchaser-provider split), or from a PPSM to a FPM (cf.Starr 1989, 24).Many philosophical critics of privatization, it should be noted, such as Alon Harel, Avihay Dorfman, and Chiara Cordelli, appear to be concerned only with shifts from state to private provision, and so raise no principled objections to type 1 and 4 privatization, focusing only on types 2, 3, and 5. 7 Public debate, by contrast, has been sparked by all 7 See Dorfman and Harel (2013) and Cordelli (2020).My central objection to these views is not merely that they ignore certain forms of privatization, but that, in order to rule out the forms of privatization that they do object to, they impose normative conditions on agents of the state so strict that they rule out large segments of the administrative five, and so it is worth considering what is at stake in each of these decisions about the role of the state.
Perhaps the closest thing to a type 5 privatization, or a direct movement from the TPSM to a FPM, was the (abidingly unpopular) privatization of residential water services in England and Wales.Historically, residents had been charged a flat rate for water and sewage (that is, regardless of usage), calculated as a fraction of municipal property taxes.The distribution network and supply facilities were owned and operated by the state (as they are still in Scotland and Northern Ireland).Privatization (carried out in 1989 by the Thatcher government) involved not only the creation and sale of shares of these state-owned firms, but also authorization (and encouragement) to institute new billing arrangements.The latter was aimed at promoting the installation of water meters, which allow utilities to charge consumers for the exact amount of water used.This was motivated in part by the fact that the U.K. was lagging most European countries in the transition to metered water supply (although it should be noted that in these other countries there was no privatization of water provision, and so the installation of meters in those countries involved only a type 1 transition from the TPSM to a SPSM).
The more common form of privatization is type 3, a shift from the SPSM to an FPM.Highly publicized cases of this sort involved state-owned enterprises (SOEs), such as airlines, automobile manufacturers, or oil companies, which already functioned very much like private firms in their relations with customers, being sold off to the private sector.Somewhat less common is type 4 privatization, or a transition from a tax-financed to a user-pay model, where the supply is already in the private sector.The most familiar cases of this involve situations in which the provision of some good is subsidized from general tax revenues, which the state then withdraws.The elimination of gasoline and other fuel subsidies are perhaps the best-known instances; the 'delisting' of certain services from public health insurance plans would be another example.
What gives rise to a great deal of confusion, however, is the fact that the term privatization is used not only to refer to these transitions to a fully private market, but also to describe many transitions from a TPSM to either a SPSM or DPSM.These forms of privatization do not result in the formation of a fully private market, because the state still retains an important role on one side of the transaction.Thus the shift away from taxpayer funding toward tuition-based payment is described as the privatization of higher education (type 1), even when ownership remains public.Or the shift away from having city employees collect residential garbage, to the contracting out of this service to a waste management firm, is described as privatization (type 2), even when it remains financed by taxation and free at the point of service.This is certainly the area in which one is likely to find the most anodyne instances of privatization, simply because the state retains the ability to exercise considerable control over the market by virtue of dominating one side of it.8At very least, there can be no categorical rejection of such arrangements, simply because the purchaser-provider splits that are created through privatization are already ubiquitous in the public sector.
The major question with type 1 and 2 privatizations is, given that the state is still going to be maintaining an important presence in the sector, why not just stick to the TPSM?In particular, the traditional model is often favored for reasons of distributive justice.On the supply side, government employees are often better paid and enjoy better working conditions, including more extensive unionization, than private-sector employees. 9On the demand side, funding through general taxation is progressive with respect to income, whereas user payments are not.For this reason, opponents often allege that privatization is favored by the enemies of equality precisely because the TPSM has these progressive features (see, for example, ITPI 2016).More careful analysis, however, reveals a somewhat complicated situation.Some of these distributive justice intuitions opposed to privatization turn out to be invalid upon closer examination, while others, which turn out to be valid, can still be achieved under a PPSM, using different policy instruments.Thus the distributive justice issue often just muddies the water, making it more difficult to focus on the issues that are actually driving the decision whether or not to privatize.
In the remainder of this section, I will describe some of the factors that have played a central role in these deliberations.I will preface these remarks by setting aside some of the 'free lunch' arguments that are sometime made, both by critics and defenders of the welfare state, which suggest that one or the other ownership structure is inherently superior to the other.For example, it is commonly claimed by proponents of privatization that government is necessarily less efficient or less innovative than private enterprise. 10This may be a valid generalization, but it is not a universal law (Estrin and Pérotin 1991).It is certainly the case that highly competitive markets tend to promote firms with efficient internal organizational structures, just as markets with low barriers to entry are very effective at rewarding innovation.But it is not the private ownership structure of firms that is generating these beneficial effects, it is the external environment in which they are operating.Some of the greatest disappointments in the history of privatization have occurred when states sold off assets to the private sector, but failed to create a competitive market for those firms to operate it, leading to the grim discovery that private monopolies can be just as inefficient as public monopolies, and often more infuriating to deal with.Thus the promise of efficiencies arising merely from the change in ownership structure winds up being a free lunch (that is, an illusory benefit).
Opponents of privatization often make an opposite, but equally invalid claim, by suggesting that the public sector is always more efficient because it is not burdened by the expense of paying out 'profits' to investors.This view remains quite persistent, despite the fact that it is an obvious free lunch.Since every welfare state in the world today runs a substantial budget deficit, every marginal increase in expenditure is financed by debt.And so either way, providers of capital get paid, whether in the form of interest on government bonds to finance public-sector investment, or in the form of profits on shares to finance private-sector investment.Explicit borrowing costs are somewhat lower in the public sector, since the state need not pay the 'equity premium' that is enjoyed by shareholders in private firms.But that is because the state is bearing all of the risk in the case of public financing, which means that lenders are being compensated in the form of a guaranteed return on their investment.
Either way, the providers of capital get paid an equivalent return.The alternative, of financing expenditure through increased taxation, is obviously not free either, although the cost is much more difficult to calculate (involving both transaction costs and deadweight losses).Nevertheless, it is difficult to find any estimate of the cost of raising an additional dollar of tax revenue that finds it to be lower than the rate of return on equity investments.
Thus there are no global arguments for or against privatization, each case must be evaluated on its merits.When thinking about these merits, it is helpful to distinguish the type 1 from type 2 cases.

IV.I. Motives for Type 1 Privatization
The easiest way to understand the motivation for type 1 privatization, which involves shifting from tax-financing to beneficiary-financing of state-provided goods, is to consider the consequences of the fact that most goods provided by the welfare state are not actually public goods in the Samuelson sense.Because of this, they are not non-rival in consumption, but rather are subject to congestion, whereby one person's use either reduces the quantity available for others or erodes the quality.Obviously burning natural gas, drinking water, using electricity, or living in an apartment makes these goods unavailable to others.Less self-evidently, driving on a road creates congestion that diminishes the value of the road to others; attending an educational institution increases class size, which may reduce the quality of the learning experience; showing up at the emergency room creates delay that may reduce the quality of care for others; and so on.There may be an excellent case for having these goods provided by the state, but when they are free at the point of service it has the potential to generate collective action problems.Because citizens have no external incentive to limit their consumption, they will tend to ignore the negative effects their decisions have on others, resulting in either shortages (and queues) or quality degradation.This can have negative downstream consequences as well.One of the major factors contributing to urban sprawl, for instance, is that not only are most roads unpriced at the point of use, but municipalities often use general tax revenue to pay for expansion of water, sewage, electrical, and transit systems, as a result of which homeowners do not pay anything close to the true cost of lowdensity exurban development.Under the TPSM, the government, which is supposed to be in the business of solving collective action problems, may find itself creating them instead.
An important motive for type 1 privatization is therefore to resolve collective action problems that have been created by public financing of a rival, or partially rival good.The installation of water meters is perhaps the least controversial instance of this, in part because of the vast amounts of water that are wasted when it is unmetered (which apart from being expensive, can have very negative environmental consequences) (Ornaghi and Tonin 2021).People growing lawns in arid or semi-desert conditions, or hosing down the sidewalk outside their front door instead of sweeping it, are both easily observable consequences of financing water provision through a flat tax on households.Restaurants in California have on occasion been forced to stop offering water to their customers, while farmers nearby waste literally billions of gallons growing crops like rice and almonds in the desert, all because of the government's failure to charge users for the amount of water that they consume (and a prohibition on reselling it) (Fox 2015).
Along similar lines, there is a fairly compelling argument to be made for road pricing as an antidote to congestion.Drivers in many jurisdictions do actually finance road construction and maintenance through fuel taxes.The problem is that they pay nothing for roads at the point of use, which leads them to ignore the effects that their decisions about where and when to drive have on other users.Back when tolls had to be collected by hand the imposition of a pricing system would slow down traffic, but with electronic billing this obstacle has been removed.And yet unlike most public transit users, who are forced to pay per trip, the payments made by drivers are still funneled through the tax system, in a way that creates a fairly serious collective action problem.Making them pay directly for road use is a way of reorganizing these incentives.(In opposition to this suggestion, one is likely to encounter the egalitarian objection, that usage charges are equivalent to 'regressive' taxes, because they cost the poor a larger fraction of their income.This is a misleading characterization, however, insofar as a price is not a tax, and every price in the economy is, in a sense, regressive, since it costs the poor a larger fraction of their income to buy a television, or a slice of pizza, than it costs the rich.The solution, however, is not to give everyone free roads, televisions, or pizza, and simply tolerate the collective action problems that result, but rather to give income assistance to the poor, and charge everyone the true social cost of their consumption.) The second major argument for type 1 privatization is that the use of general tax revenue to fund provision of a good may result in unwanted forms of cross-subsidization between different classes of citizens, including some that are highly perverse.For example, it will generally be the case that the lack of user fees, or the presence of subsidies, in utilities and infrastructure will leave those who seek to minimize their environmental footprint subsidizing the lifestyle of those who do not.There is also the concern that if consumption of the good is skewed toward the wealthy, the net effect of public provision on the TPSM may be regressive, even though the tax system as a whole is progressive.The best example of this is university education, which despite generating positive externalities, also produces non-trivial private benefits.The introduction of variable fees for students was intended to eliminate unjustifiable forms of cross-subsidization between different categories of student, as well as between taxpayers and students who derived significant private benefits from their education (Barr and Crawford 2005). 11 Finally, it is important to recognize that fiscal discipline may be very difficult to achieve under the TPSM because of the 'soft budget constraint'.Under the very old-fashioned model, public-sector economic activities were simply part of the operations of government departments, often with no clear demarcation between the budget of different components.Thus the Department of Transportation would spend money on highways, rail transit, airports, and seaports, and earn revenue from various sources, such as a gas tax, carriage fees, ticket sales, and docking charges, all of which would simply be absorbed into the department's budget, then perhaps supplemented with a subsidy from general tax revenue.This made it very difficult to get a clear picture of the fiscal situation with respect to any particular operation, such as a rail service.In order to avoid the rather serious accountability problems caused by this arrangement, the moderately old-fashioned model saw particular operations being consolidated into distinct 'crown corporations' or SOEs, which would be assigned a budget and then held accountable for their own expenses and revenues.While in principle this promised to improve fiscal responsibility, in practice it achieved much less than initially hoped.The central problem was that the state had difficulty imposing credible budgetary discipline on these SOEs.If revenue fell short of projections, or cost overruns led to additional expenses on a capital project, the central state had little choice but to cover the shortfall.So while SOE managers were instructed not to run deficits, there was not much that could be done to deter them from doing so (other than threatening to fire senior managers, which was also for various reasons seldom a credible threat).This created the problem that János Kornai referred to as the soft budget constraint (Kornai 1986).
Because states have improved their capacity to manage these budgetary difficulties, it is easy to lose sight of how serious a problem they became during the early period of welfare state growth.During the 1960s and 70s, SOEs lost extraordinary amounts of money, enough to undermine the stability of several European currencies, as well as threatening the fiscal integrity of the state.In 1975, for instance, the total budgetary deficits of SOEs in the U.K. was equal to 35% of their contribution to GDP (Vickers and Yarrow 1991, 122).Books such as James O'Connor's, The Fiscal Crisis of the State, published in 1973, reflected the widespread perception that the welfare state was financially unsustainable.A great deal of this red ink was due to simple mismanagement.For example, public sector operations were often overcapitalized in comparison to the private sector: airlines had too many airplanes, flying too many empty seats; railroads had too much rolling stock, and kept open too many branch lines, and so on.But as long as SOEs had the capacity to tap into general tax revenue to cover their deficits, it proved extraordinarily difficult to get them to stop losing money.More generally, the public sector has considerable difficulty penalizing failure (for example, closing things down), in part because those who are adversely affected have political avenues of redress.
The primary mechanism through which these problems were addressed initially was 'corporatization', whereby SOEs were put in increasingly arms-length relations to traditional line departments, and given their own budgets, allowed to enter legal contracts, assume commercial liabilities, and so on.Many, of course, continued to be funded by the state, but increasingly their budgets became subject to explicit contracting (for example, through 'pay for performance' arrangements, in which budgets would be increased only in return for the achievement of specific objectives).The major reason that these arrangements were introduced within the public sector was to enhance financial accountability, and to supplement traditional lines of ministerial authority with more binding directives.Once this was accomplished, however, budgetary discipline could be further enhanced by requiring these SOEs to finance their own operations by charging for their services, an arrangement that in many cases involved type 1 privatization.Such an arrangement can be used, for example, to force SOEs to finance their own capital expenditures from the revenue stream they can be expected to generate (something that officials acting in an oversight capacity, looking at an expenditure request, are often at pains to assess).Thus type 1 privatization wound up being used often as a way of imposing, if not a hard budget constraint, then at least a firmer constraint on SOEs.
It is important to emphasize that there are better and worse ways of doing all of this, and there are many examples of privatizations motivated by one or another of these concerns that worked out very poorly.What I have described are simply three problems associated with the financing of state activities through general taxation: it may generate collective action problems (which manifest themselves in the form of wasteful consumption, congestion, quality degradation, and queuing); it may produce unwanted cross-subsidization among different categories of citizens; and it may give rise to a soft budget constraint that encourages wasteful public expenditure.In certain cases, it may be possible to attenuate or resolve these problems by imposing payment at the point of consumption, or otherwise transferring the cost of providing the good to its beneficiaries.As a general principle, one can say that with state programs that are primarily motivated by the goal of correcting market failure, if it is not overly costly to identify the individuals who are benefiting from the program, if the magnitude of that benefit can be estimated with some precision, and if there are no overriding reasons of fairness not to do so, then the state should charge those individuals for the benefit that they receive.12These are the cases where type 1 privatization is justifiable.

IV.I. Motives for Type 2 Privatization
The second, somewhat more controversial form of privatization involves cases in which the state continues to fund the provision of some good through tax revenue, but instead of having the actual provision done by state employees, it contracts with either private individuals or organizations to do the provision.Type 2 privatization therefore involves a shift in the public-private boundary, so that goods that had once been made internally are now bought from external sources.The most important thing to recognize, when it comes to the 'make or buy' choice, is that the state is often faced with very unattractive options on both sides.The problems with state employees will be discussed below; for now I would like to draw attention to the fact that governments are in many ways handicapped compared to private corporations when it comes to contracting out, because of the extremely onerous constraints that are imposed on procurement.Keeping in mind that practically nothing is public all the way down, every government department and SOE purchases a large number of inputs from the private sector, in the same way that private corporations maintain a large number of supply relationships.The state, however, has enormous difficulty getting good value for its money from these relationships, in part because it is prevented by procurement rules to developing strong relationships with suppliers (Kelman 2002).
Although technically the relationship between a firm and its suppliers is a market exchange, in practice private corporations often cultivate highly cooperative relationships with their long-term suppliers.A supplier may know, for example, that if it overperforms with respect to a contract (for example, delivering goods early or at lower cost, suggesting a more technologically efficient solution, etc.) it will be rewarded through preferential access to future contracts.Government, however, is usually constrained by the requirement that it take the lowest bid on any tendered contract.As a result, the state may find itself obliged to accept a bid from a firm that it considers exceedingly unlikely to fulfill the contract, over another firm that has performed exceptionally well in the past.For the same reason, firms have no incentive to overperform, because they know that it will gain them no consideration in the future.As a result, they tend to behave very legalistically when it comes to performance, providing no more than the letter of the contract obliges them to.The famed capacity of the private sector for innovation and creative thinking is often applied to the task of cynically exploiting government contracts.
Because of this, the state often finds itself in a situation in which it would like to contract something out, but is unable to do so because it cannot secure good value.If the contract is insufficiently complete, contractors will exploit ambiguities, but making it complete would impose excessive rigidity.For example, the public sector tends not to have very sophisticated information technology competence in-house, in part because it cannot compete with the private sector for talent (Kelman 1990).As a result, when it wants to computerize some process for example, switch to electronic record-keeping), it must put out a contract.The problem is that procurement regulations often do not permit the state to specify its needs in outcome-based terms ('we would like to solve problem x'), it must instead specify in considerable detail how it would like the problem solved, and precisely how much it is willing to pay.Unfortunately, it often lacks the technical expertise to know how the problem should best be solved.Thus it may find itself unable to tender a contract, because it cannot articulate its needs at a level of specificity that will both satisfy the rules governing the procurement process and protect itself from exploitation by contractors.Ironically, this creates an incentive to privatize government agencies that purchase high-value inputs, precisely because this will free them from government procurement regulations, allowing them to get better value out of their supply relationships.
Attentive conservatives have in fact noticed that procurement regulations are a major barrier to their project of achieving meaningful reductions in the size of the state.Some have argued that instead of engaging in top-down privatization of traditional public services-like selling off the local water treatment plant-which tends to evoke political resistance, conservative governments should instead focus on improving the efficiency of procurement practices, which would in turn promote a bottomup privatization of state operations, as public managers increasingly find themselves able to secure better value by contracting out more services.The problem with this strategy is that it is practically impossible to improve procurement practices without increasing the discretion of state officials, which many conservatives are also opposed to, in part because they are suspicious of the motives of public officials.As a result (and contrary to the narrative of rampant neoliberalism), conservative parties have not achieved much success at solving the basic conundrum of government procurement, and have therefore done little to incentivize bottom-up privatization.
When one turns to the other option, of doing in-house provision, the choices may also be unattractive, but for quite different reasons.Although partisans of the welfare state have been extremely reluctant to admit it, public sector unions are the source of an enormous number of difficulties that afflict in-house provision. 13Unions were developed as a response to the adversarial labor relations of market contracting in private corporations, and so the importation of those adversarial practices into the public sector has been a poor fit in a number of ways.Superficially, the major problem is that, unlike in the private sector, in which unions are forced to moderate their demands in order to ensure the ongoing viability of their employer, public sector unions need not worry about the state going bankrupt.As a result, they are able to engage in much more aggressive rent-seeking, which results in many countries experiencing more labor militancy, and a great deal more work disruption, in the public sector than in the private.The only real constraint on this is the threat of back-to-work legislation, or revocation of the right to strike (for instance, through designation of a class of workers as 'essential').This has a number of peculiar effects, including the fact that it may make unions more obstreperous in dealing with left-wing governments, whom they expect to be more reluctant to enact such legislation.The most unfortunate effects, however, arise at the municipal level, since these governments seldom have the power to enact back-to-work legislation, and so must appeal to a higher order of government to resolve recalcitrant labor problems.In countries such as the U.S., where municipal politics is structured by the same party divide as state and federal politics, this assistance may not be forthcoming for a variety of reasons, include baser forms of partisan politics.Thus municipal governments often have the greatest incentive to contract out various forms of labor, because they are in such a weak position when dealing with public-sector unions.
Sentimental attachment to the labor movement has made both leftwing activists and academics extraordinarily averse to criticizing publicsector unions, or to pointing out the various ways that their actions have contributed to the decline of the socialist project.Despite widespread mobilization against police violence in the U.S., for instance, it took progressives a very long time to acknowledge that police unions were an enormous part of the problem (Levin 2020).Similarly, considerable apprehension has been expressed about the role that corporations running private prisons might play in pressuring states to expand incarceration, with total disregard for the role that (far more powerful) prison guard unions have played in lobbying for the exact same goals (Pfaff 2017).Teachers' unions in the U.S. also served as a major obstacle to school re-opening in the wake of the COVID-19 pandemic, leading to large-scale flight to private schools (which remained open) in many regions (Marianno et al. 2022).And while there has been an extraordinary amount of attention paid to the influence of corporate interests in the political process, the problem of capture of social-democratic parties by public-sector unions has been relatively neglected (Barber 2015).
With respect to privatization, perhaps the most dramatic example of public sector unions undermining the case for state ownership is the degradation that has occurred in public housing projects in North America.The issue here is not just wages but also the extremely rigid work rules that unions favor.Keeping older buildings in a state of good repair requires a complex combination of skills.Private landlords will typically employ a 'handyman', who is able to carry out most small maintenance and construction tasks, including plumbing, electrical work, carpentry, and painting.Unions, by contrast, will usually insist on a strict separation of trades.The result, as one official at the Boston Housing Authority described it, is that even small repair jobs may require appointments with several different trades ("if you had a leak under the sink, a laborer had to go and break open the wall.A plumber would have to get in there and fix the pipe.A carpenter had to come back up and frame the hole back in.The plasterer would have to plaster.And the painter would have to paint"- Moore 1995, 199). 14These onerous work rules are a major reason that public housing systems across North America have a multi-billion dollar backlog of repairs.Budgetary authorities, however, have no confidence that sinking a few more billions of dollars into these operations will actually result in many repairs getting done.As a result, the housing stock is simply allowed to degrade until eventually it becomes uninhabitable.Under such circumstances, there is a strong argument to be made for privatization (combined, as appropriate, with a voucher scheme allowing low-income residents access to rental housing on comparable terms).If the practical consequence of public provision of housing is the degradation and eventual destruction of entire neighborhoods, it is unreasonable not to explore other ways of securing access to those goods.
The example of public housing is somewhat anomalous, in that there were not really significant market failures in this sector.While private landlords exercise some market power over tenants, because of the costs involved in moving, the primary motivation for the development of public housing was always to improve affordability, which is an essentially redistributive objective.The challenge with redistributive programs is then to justify in-kind provision, since if the market is reasonably efficient, the egalitarian objectives can be achieved through monetary transfers.Thus certain type 2 privatizations are motivated by the realization that, in order to ensure that all citizens have access to certain goods, it is not necessary for state employees to be providing them directly.The Supplemental Nutrition Assistance Program in the U.S., for example, is run by the Department of Agriculture because in the distant past the U.S. government used to take surplus agricultural product and transfer it directly to the poor.Food stamps were introduced in 1939 in order to give recipients greater freedom of choice, while still focusing on the disposal of agricultural surplus (as part of a broader price-management scheme).This evolved into the contemporary program, which provides recipients with an electronic card that can be used to purchase most food items, and in some cases inexpensive restaurant meals.Most countries take the easier route of simply providing the poor with cash payments, with the expectation that they will use a certain fraction of it to purchase food.In each case, Western states have essentially gotten out of the business of providing food directly to the poor; food security is achieved by providing a means-tested transfer, leaving the provision of food to the private sector.
Housing projects in the U.S. are also peculiar in that they are administered municipally, but derive significant funding from the federal government (from the U.S. Department of Housing and Urban Development).This illustrates another important point, which is that theorists often refer to 'the state' as though it were a unified organization, ignoring the fact that authority is typically fragmented between different levels of government.This is particularly important in federal states, where regional governments often wield enormous power, and yet act in an entirely strategic manner toward other regions in the confederation, as well as the federal government.Thus the idea that government officials act in 'the public interest' harbors a significant ambiguity, since the public interest at lower levels of government is typically identified with a regional interest.Many state or provincial politicians do not even attempt to sustain the pretense that they care about the interests of the country as a whole, because of the electoral advantages that can be obtained from adopting an antagonistic stance toward outsiders.
This can become a serious problem when SOEs belong to these lower orders of government.For example, elected representatives from coalproducing regions of the U.S. can be counted on to oppose any effort at climate change mitigation.This is typically taken to be a consequence of successful lobbying and political donations by the coal industry.But now imagine if state governments owned the coal companies!This would, if anything, make the problem worse, since public managers have internal access channels (and are immune to lobbying restrictions imposed on private firms) that can make them an incredibly powerful force within government.One can see this in Canada, where provincial governments have jurisdiction over natural resources, which gives them both control over resource royalties and the freedom to develop SOEs in the mining sector.This has made provincial governments extremely aggressive in defending the interests of the mining industry, with a number of consequences, including considerable success in opposing federal environmental regulation.The influence of the Quebec government, for example, in defending the interests of the Société nationale de l'amiante is one of the reasons that asbestos mining continued in Canada for so much longer than in Europe and the U.S. (Shields 2012).
All of these are examples of the way that the accumulation of significant resources and economic power within the state can be, in effect, corrupting, benefiting constituencies whose interests are poorly aligned with the general public interest.This is a reflection of the fact that public management-which has as its central priority maintaining that alignmentis extraordinarily difficult, being subject to a great many more constraints than private management.As a result, cases may arise in which it is easier to maintain that alignment through external contracting than through internal supervisory arrangements.Under such circumstances, there may be a compelling case to be made for type 2 privatization.

V. DEGREES OF PRIVATIZATION
In addition to the five types of privatization described in the previous section, it is important to observe that privatization need not be an allor-nothing affair.It is possible to partially privatize government operations on the supply and the demand sides of a transaction, or both.Consider, for example, type 5 privatization, which involves a shift from the TPSM to a fully private market.As already noted, states only rarely enforce a monopoly in the provision of a good, and so private markets typically co-exist alongside state provision.In the standard run of cases, the state provides the good at a relatively low level, but makes it universally available.Individuals who want more can then purchase additional quantities through private contracting.In such markets, type 5 privatization sometimes occurs by attrition, as the state simply reduces its funding for provision of the good, so that the private sector takes on an increasing share of the burden.For example, there was clearly a partial privatization of policing (or security) in the U.S. with the formation of gated communities, since the security guards hired by homeowner associations were not just supplementing the police, but rather were replacing them in the performance of certain tasks (Jones and Newburn 2006;McKenzie 1994;Stark 2009, 29-42). 15imilarly, there are many cases in which the state subsidizes the provision of some good, but where consumers must still pay a fraction at the point of delivery.Many public services have user fees (for example, public swimming pools and sports facilities, summer camps and after-school programs, university tuition), which do not cover the full cost of providing the relevant good.It is possible then for the state to shift the publicprivate balance by drawing down the subsidy, so that a larger fraction must be financed by these fees.This can be seen as partial privatization of the type 1 or 4 variety.
There are many cases in which this sort of privatization is unjustifiable or even pernicious.In cases where the state is the sole provider of a good, underfunding will typically generate complaints about quality degradation or increased queuing.The availability of a private alternative, by contrast, typically insulates wealthier individuals from the effects of declining quality or availability in the public sector (they are more likely, in other words, to exercise an exit option rather than voice).This leaves only relatively powerless citizens consuming the public version of the good, people whose complaints can more easily be ignored.Thus partial privatization may constitute a form of 'privatization by stealth', which winds up being the course of least resistance, politically, only because the costs are borne disproportionately by the disadvantaged.
At the same time, it is important to recognize that there are other cases in which is it perfectly appropriate for the state to adjust the publicprivate ratio in the provision or subsidization of certain goods.On the subsidization side, circumstances that warrant it include those in which severe congestion or queuing has developed in the public sector, or when users have figured out a way of appropriating for their private benefit a larger fraction of the value produced by the good.For example, with the increased demand for 'quants' in investment banking, university mathematics programs found themselves teaching more and more students who were preparing for a career in finance, rather than science.Introducing a separate finance stream, and charging students higher tuition for those degrees, struck many as a reasonable response.On the provision side, partial privatization is often used to introduce competition, which puts pressure on public-sector organizations to improve their performance.Michael Barber, for instance, estimates that contracting out 10-15% of operations in some domain is often enough to generate noticeable improvement in the segment that remains in the public sector (Barber 2015).This is partly because the availability of a private-sector alternative generates benchmarks, which allow for the assessment public-sector performance, and partly for its incentive effects, as public-sector organizations improve their performance (and scale back labor militancy) in order to avoid being privatized.
What is less often realized is that when the state sells off certain assets to the private sector, it may not do so entirely, but will retain some fraction of the ownership, often enough to continue exercising control over the enterprise.In many cases the state will also create for itself a special class of shares (that is, a 'golden share' arrangement) that allow it to exercise control regardless of what happens with the shares held by private investors (Bortolotti and Faccio 2009).For example, the sale of Canada's national oil company, Petro-Canada, in 1990 was by far the most controversial privatization undertaken by the government of Prime Minister Brian Mulroney.And yet the Canadian state actually retained a 19% ownership stake in the firm, and placed conditions on the traded shares that prohibited any individual or organization from acquiring more than 10% ownership.(Similarly, despite being routinely classified as an SOE, British Petroleum was never entirely owned by the British Government.When the first round of 'privatization' occurred in 1979, the Thatcher government merely sold 5% of the company's shares, drawing down state ownership from 51% to 46%.) States also often direct public pension funds or banks (such as the public investment bank Bpifrance) to invest in the newly privatized corporations, so that private ownership is actually something of an illusion.
This type of mixed ownership is possible because of the prior process of corporatization, which transformed the relationship between many SOEs and traditional line departments.These reforms were undertaken with the primary objective of improving accountability within government.One of its effects, however, was to radically reduce the difference in governance structure between public and private enterprises, which in turn made mixed ownership feasible.Most obviously, corporatization disembedded economic operations from traditional departments, where they were often under direct ministerial authority, and placed them (typically) in a special-purpose statutory corporation, under the authority of a more-or-less independent Board of Directors.And instead of being given a complex mandate, with multiple benchmarks of success, managers were instructed to operate under standard business principles, including in some cases the goal of maximizing profit (especially when demand was fully private).The pursuit of 'social objectives' was secured, not by assigning additional objectives to managers, but rather by contracting with the SOE, in effect paying it additional sums to pursue goals such as regional development, hiring disadvantage workers, or keeping the cost of certain goods artificially low. 16he latter point merits brief comment, since critics often observe that after privatization states are still forced to engage in complex contracting and regulatory intervention in order to control the private enterprises that are supplying the goods that were formerly provided by state employees.This is particularly noticeable in cases in which states privatize operations but fail to create a competitive market, and so are forced to regulate, and in some cases supervise, the newly-created private monopoly.This has given rise to some notoriously opaque and ill-conceived contracts.Critics were certainly correct in observing that the more enthusiastic proponents of New Public Management reforms vastly overestimated the capacity of the state to formulate complete contracts with private suppliers, mainly because they ignored the information asymmetries involved in the provision of many traditional state services (Heath and Norman 2004). 17All of this goes to show how easy it is to underestimate the agency problems involved in contracting with private firms.At the same time, it is also easy to underestimate the agency costs involved in traditional state employer-employee relationships.Critics of privatization sometimes write as though, under the TPSM, elected officials simply gave orders, which were then faithfully carried out by a cadre of loyal civil servants (Farrell 2019).In fact, these internal channels of authority are often ineffective or non-existent, as a result of which the state already engages in a significant amount of regulation and contracting in dealing with its own branches and operations (Hood et al., 2003).For example, the Minister of the Environment cannot simply order public utilities to reduce their pollution output.Emissions reduction from public utilities is achieved in the same way that it is with private utilities, through regulation.Thus the difference in ownership between SOEs and private corporations does not make as much difference as one might think, when it comes to asserting control over these organizations.
As a result, when the state sells off some fraction of an SOE, this does not necessarily change the level of control that it can exercise over the firm, especially when the state retains a controlling stake and so is in a position to make all appointments to the Board. 18Partial privatization, in this case, is often just a way of financing capital investment by drawing in equity instead of issuing debt.(Similarly, many public-private partnership arrangements are designed to keep large capital investments, especially in infrastructure projects, off the state's balance sheet.)In some of these cases privatization more closely resembles an accounting shenanigan than an actual sale.Consider the situation of the British Water Authorities prior to privatization.The British state had, over the course of decades, failed to make appropriate investments in upgrading its treatment facilities and distribution network, as a result of which it was facing the need to make massive capital investments in its water services.Apart from political ideology, it was this need for capital that served as a major motive for privatization.The standard way for the state to finance this sort of investment would have been by selling bonds.The problem for the government of the day was that issuing bonds shows up as government debt, and the Thatcher government could not tolerate the optics of increasing the annual deficit.Equity, by contrast, can be booked as revenue, not debt, and so by attracting equity investment the government was able turn requires making trade-offs, where finding the optimal balance requires judgment, the outcome of which cannot be specified contractually.
to create the impression of having reduced its deficit.Of course, in the case of water the British state sold off the assets entirely, but had it merely sold off a fraction, sufficient to cover the capital cost of the new investment required, it would have managed to borrow millions of pounds, while at the same time reducing the state's official debt, and without actually relinquishing any effective control over the water service.(A cynical observer might suggest that the government of France has on occasion done precisely this, using partial privatization as a way of borrowing money from the pension fund of public-sector employees without having to book the loan as debt.)

VI. TECHNOLOGICAL CHANGE
The most highly valued economic activities of the welfare state occur in areas in which the private alternative is compromised by some form of market failure.The concept of market failure, however, is not structural, but is always relative to the current state of technology.What typically causes the market to fail is not the impossibility of solving a particular organizational problem privately, but rather the cost of doing so.For example, it is commonly claimed that negative externalities arise because of incompleteness in the system of property rights.It is, however, seldom impossible to establish property rights in areas in which they currently do not exist, such as over the open seas or the atmosphere, it is just prohibitively expensive to do so.Similarly, insurance markets are often undermined by asymmetric information, but there is usually some price at which the desired information could be obtained.The market fails simply because the cost of organizing the transaction exceeds its economic value, and so private parties do not bother undertaking it.Or in the case of partial market failure, the cost results in inefficiently low levels of production of the good.
Just as technological development can change the profile of transaction costs faced by firms, creating an incentive for them to readjust their boundaries, technological development can resolve old and create new market failures.Most obviously, it can allow competitive markets to develop in sectors that had previously been natural monopolies.Telephone service, for example, had once been a natural monopoly, because it was necessary to connect each building up to a physical network of transmission lines.With the development of wireless service, the cost of developing a competitive market (that is, redundant geographical coverage) declined significantly, resulting in most states opting for a regulated competitive market over monopoly state provision.Even in France, where the state went to considerable effort to ensure that France Télécom, reborn and rebranded as Orange, would retain a dominant market position, no attempt was made to prevent the emergence of a competitive private market.(Orange was also subsequently semi-privatized, with the French state maintaining a controlling stake.) More subtly, the development of information technology has radically reducing the cost of private-sector contracting.The decline of the mid-20th century conglomerate firm is largely due to a decline in the cost of market contracting relative to managerial organization.Similar forces have affected the public sector.Although the cost of collecting taxes has declined, the cost of private transactions has declined a great deal more, which has eroded the comparative advantage of the public sector.In particular, keeping track of people's consumption and then billing them has become dramatically less expensive than it once was.This was the major factor making it possible to split distribution from provision in physical networks such as the electricity grid or natural gas pipelines.As a result, it became possible to create a competitive market for electricity generation, while keeping electricity distribution a publicly owned monopoly.The development of water meters, or time-of-use pricing for electricity, has also become feasible through a dramatic decrease in the cost of collecting data from locally installed devices.And finally, computerization has generated a significant decrease in the transaction costs involved in running auctions, which has in turn enhanced the state's ability to generate competition among private providers in the absence of conventional markets (such as with electricity supply) (Maurer and Barroso 2011).
In other cases, the general cheapening of commodities has resolved problems involving the underproduction of positive externalities (or nonexcludable goods), and thus reduced the need for state involvement.During the first half of the 20th century, for example, the city of Montreal made a major investment in the construction of public baths, aimed primarily at ameliorating the quality of life and social integration of poorer immigrant and working-class communities, who had no hot water in their homes (Labonne 1996).With improvements in indoor plumbing, however, most of these baths fell into disuse (and municipal bylaws no longer permit the construction or rental of living units without hot water and private bathing facilities).As a result, the baths were either sold off or transformed into swimming pools.Although there was some mobilization aimed at saving the buildings-some of which were quite beautiful-no one opposed the actual privatization, simply because there was no longer any case to be made for public baths.One can imagine something similar happening with public libraries and books, once distribution becomes fully electronic. 19 These examples of technological change resolving market failures should not lead one to assume that change always runs in the direction of greater privatization.In some cases, technology has created market failures where previously there had been none.For example, physical limitations often made it possible to bundle together products in a way that facilitated private provision of certain positive externalities.Newspapers are perhaps the best example.A traditional newspaper derived significant revenue from its classified advertising section, which it then used to subsidize other aspects of its operations, most notably the news section.The development of the internet made this sort of bundling infeasible, and paid classifieds were soon replaced by online platforms offering the service without charge.This represented an obvious gain for consumers, except that it removed the subsidy to news production, which has produced an increasingly conspicuous market failure in the production of 'public interest' journalism, particularly at the municipal level where many local newspapers have disappeared.
Technology has also allowed the formation of new natural monopolies, particularly through the network externality effects that are highly visible on social media.In many cases, the solution lies in the development of open-source versions (such as Linux in the domain of operating systems, or Apache in web servers) and so state ownership and provision are not required.The point is that conditions are dynamically unstable, and so it is important to maintain ideological flexibility when considering the overall merits of state ownership in particular sectors.There are, as they say, many ways to skin a cat, and so when circumstances change the best way to achieve certain state objectives-whether best articulated in the language of social welfare or of social justice-may also change.In certain cases, privatization is justifiable because state ownership or financing is simply not required any longer to achieve those results. 19And so, while I love libraries almost as much as Martin O'Neill, I am not persuaded by his claim that they are public because "there is something transcendent, something deeply valuable, indeed almost miraculous, about a shared public space devoted to a role as a repository for human learning and memory" (O'Neill 2020, 187).I can imagine a similar paean having been written 80 years ago about the solidarity-enhancing qualities of communal bathing, and yet no one would think, in the present day and age, that this constitutes a strong argument for public baths.

VII. CONCLUSION
My objective in this discussion has not been to defend privatization as a general panacea for the various frustrations that citizens experience when dealing with government.The narrow objective was merely to present a counterpoint to the extremely negative portrayal of privatization that is common in the philosophical literature, which arises primarily from a focus on worst-case scenarios.This is not to deny that there have been many terrible, ill-considered, or even flagrantly corrupt privatizations.My idea, however, was to present a few best-case scenarios, in order to show that privatization of the economic functions of the state is not intrinsically objectionable, even though it may be, for various contingent reasons, highly undesirable.Examples of anodyne privatization would include the selling of SOEs in sectors that were once natural monopolies, but where technological change has made it possible to organize appropriately competitive markets; the imposition of user charges for publicly provided goods that are experiencing quality degradation from open access (such as electronic tolls on highways that suffer from severe congestion); and the contracting out of public services in cases where relatively complete and enforceable contracts can be written (such as municipal garbage collection).Thinking about these cases alongside the undesirable ones allows one to form a conceptual framework, or a grid of sorts, that will make it easier to situate and to develop a more nuanced analysis of any proposed privatization.
A great deal of the political rhetoric surrounding privatization is a hold-over from an earlier era, in which social-democratic parties were still committed to the gradual nationalization of the entire economy.It is important to recall, for instance, that during the Thatcher years the British Labour Party still had the controversial 'clause IV' in its constitution, which expressed a desire for "common ownership of the means of production, distribution, and exchange" (Gani 2015).If this is the objective, then of course any privatization whatsoever constitutes a setback that must be reversed at the earliest opportunity.But since that time, most progressives have come to accept that total state ownership and control of the economy is neither feasible nor desirable.Indeed, most now accept some version of the provocatively ambiguous German Social Democratic Party slogan: 'As much market as possible, as much state as necessary' (Crouch 2021).
But if the state is no longer aiming to take over the entire economy, the most pressing question for the democratic welfare state becomes one of determining which parts should still be owned and/or controlled.My own view, which I have defended at length elsewhere, is that state involvement in the economy should be guided primarily by the norm of efficiency, which is to say, the objective of correcting market failure, and in cases in which it cannot be corrected, the substitution of state purchasing and/or provision for market contracting (Heath 2020).This is a view that is commonly held by economists, and much less commonly held by philosophers and political theorists.The received view among that latter has been that state involvement in the economy is primarily motivated by the goal of promoting greater equality.It follows from this that a certain amount of inefficiency in state provision is tolerable, if the arrangement is still beneficial from the standpoint of distributive justice.(To give a concrete example, proponents of such a view may oppose road tolls, even though the failure to price vehicle travel generates serious collective action problems, because it nevertheless makes roads accessible to both rich and poor on equal terms.)The inefficiency, from this perspective, is a cost worth bearing for achieving the more egalitarian outcome.
It is this last conclusion that I would like to cast doubt upon, or at least problematize, in my concluding remarks, by drawing attention to some hidden costs associated with state involvement in economic activities in which it does not enjoy a comparative advantage over the private sector.It is helpful to begin by identifying areas in which the state is an essential actor, in the sense that if it does not provide a good, no one else is going to step forward and do so.(This is the case, for example, with various elements of the social safety net, where if the state does not provide these forms of security to citizens, no private insurance company is going to start selling policies that offer indemnity against those risks.)There are other areas in which state involvement is obviously optional, in the sense that withdrawal from the sector would create a vacuum that private corporations would rush to fill.In between are areas in which state involvement is merely desirable, in the sense that it provides an improvement over what the private sector would have provided.
The problem with maintaining state involvement in sectors in which its activities are purely optional, or do not have a strong efficiency rationale, is that there are often private-sector alternatives available, and it is in many cases easy for citizens to compare the performance of the state to that of private firms.This is a comparison that seldom works out favorably for the state.There are complex reasons for this, but one of the most important is that providing polite and responsive customer service is an extraordinarily difficult managerial challenge, which many highlymotivated private firms fail to do well, but which public sector organizations, facing only dull pecuniary incentives, not to mention an entirely unionized workforce that cannot easily be penalized for mere rudeness, unhelpfulness, or obstructionism, almost inevitably do badly.Unfortunately, most citizens formulate their opinions about the qualities of the state as an institution on the basis of these interactions with street-level bureaucrats (Rothstein 2011).They also tend not to distinguish essential from optional state activities, and so to the extent that they develop a negative attitude toward state involvement in the economy, tend to apply it to all its activities.As a result, negative experiences with a state-run airline or hotel can easily translate into opposition to state pensions or health care, becoming in turn easy fodder for right-wing politicians with an ideological hostility to government activity in all domains. 20As a result, when the state does a bad job in areas in which its involvement is optional, it may erode the legitimacy it requires to carry out tasks in which its involvement is essential (Rothstein 2011, 93).
As the experiences of the 20th century have shown, successful government programs, such as Social Security in the U.S., the National Health Service in the U.K., or Medicare in Canada, which both create significant value and promote greater equality, are positively loved by citizens, and may become an important locus of solidarity and national identity.By contrast, poorly run programs, which frustrate citizens and provide undeserved economic rents to insiders, are detested, and have the capacity to generate blanket hostility toward government.Thus it is important that supporters of the welfare state maintain a flexible and intelligent approach to thinking about the role of the public sector, and encourage it to focus its efforts on areas in which it has a demonstrated capacity to create value.Within these domains, there may be many opportunities to finance or distribute these benefits in a way that will reduce overall inequality.The important point is that these egalitarian ambitions be pursued without sacrificing the positive sum character of the overall program.
Finally, it would be helpful if supporters of the welfare state, instead of simply opposing privatization politically, were to pay more attention to improving public administration in ways that would make privatization less attractive organizationally (mirroring the way that clever opponents of the welfare state have focused on improving procurement, in order to make contracting out more attractive).Many supporters of the welfare state have been too quick to accept the refrain that problems in the public sector can all be blamed on miserly governments imposing budgetary restrictions.As Michael Lipsky has observed, budgetary shortfalls are a structural feature of most public programs, because demand typically expands to absorb any increase in supply. 21Thus it is important to be judicious in evaluating claims about underfunding.While it is true that some problems can be solved by spending more money, questions of privatization often turn on the relative efficiency of different organizational forms, which means that spending more money on public operations often just bolsters the case for privatization.The fact that public-sector organizations can simply demand more money from the public purse is one of the major problems with public-sector organizations, that makes them so difficult to manage.The arrangement we should be aiming for is one in which the state chooses carefully the tasks that it is to perform, in such a way as to ensure that what it does, it does well.In order to achieve this, the state must obviously be prepared to stop doing-which is to say, to privatize-that which it does badly, or that others have learned to do better.