Should a public activity be outsourced to private firms? The answer to this complex question often comes with a simple and somewhat redundant answer: governments must take care of public activities. A more informative response, of course, requires a precise definition of what “public” means.

Scholars initially proposed that governments should guarantee the provision of public goods. The economist Paul Samuelson (1915–2009) originally talked about goods that an individual can consume without subtracting from the amount of goods available to other people. Subsequent definitions further specified that individuals cannot be excluded from accessing public goods and enjoying their benefits. Clean air, shared knowledge, and safe streets are all illustrations of public goods.¹

Yet not all public activities involve pure public goods. For instance, public schools can target certain local communities (i.e., families outside the community are excluded) and some schools, especially the best ones, may not be able to accommodate all applicants. In other words, they can potentially exclude some beneficiaries, even if they are free of charge. However, they have a sense of “publicness” because they are accessible to a broad population, at least to a higher extent than expensive private schools targeting handpicked students and wealthy families.

It is easy to see how these goods create formidable challenges to private provision. In the example of private schools, accepting and sponsoring disadvantaged students implies that the schools’ owners will have to forgo profits that they would reap if they focused on affluent customers. In the case of pure public goods, the situation is even more complicated, as, by definition, nobody is supposed to be excluded.

Interestingly, groups at opposite ideological sides similarly consider that public goods require public delivery. Proponents of free markets tend to argue that the state should only care about providing public goods and leave everything else to private initiative. Critics of privatization, in turn, argue that it is inappropriate to have social activities managed by actors with “business” motives – in other words, these activities are public and should be delivered by the state.

As it turns out, both positions oversimplify the debate and fail to consider recent developments in the analysis of how public activities can be properly organized.

1.1 Public goods do not need public management

Enter Ronald Coase (1910–2013), an eclectic economist interested in the benefits and costs of markets and with a unique approach to research. Rejecting the economic profession’s infatuation with clean and abstract models, he was interested in the messy, complex, confusing world of real markets and organizations. In 1974, he published a paper focused on a very curious subject: the organization of maritime lighthouses in England.²

Coase noted that economists, Paul Samuelson included, had long used maritime lighthouses as examples of public goods: any person can see the beam of light without affecting the ability of other people to also benefit from it. As such, the argument goes, lighthouses could hardly be provided by private enterprises. Picture a private owner trying to charge passing ships a fee to recover the investment in the lighthouse and make a profit. In a free market, sailors could agree or refuse to pay for the service. But how could the owner of the lighthouse exclude nonpaying sailors from seeing the light?

Here is the dilemma: if some sailors refuse to pay and safely navigate with the help of the rotating beam, then the paying sailors will soon question why they are paying for the service if others can benefit from it at no charge. For this reason, economists predicted that this market would soon unravel, and lighthouses would fail to attract private investment. Governments would need to step in, possibly levying taxes to fund and operate state-owned lighthouses.

Coase showed that, contrary to this prediction, England did experience cases of private enterprises building, maintaining, and charging for lighthouse services. Although today’s ship owners can use GPS devices to navigate, lighthouses were an essential service in the past. In 1514, King Henry VIII authorized a private organization, Trinity House, to offer navigation and lighthouse services. Trinity started as a kind of guild representing seamen and evolved to become a charity supporting their families as well as the widows and orphans of deceased mariners (it still exists today under a royal charter).

Several private lighthouses were subsequently built with patents granted by the Crown and remained in private hands until 1836. To support their investment and maintenance expenses, private operators were authorized to charge fees varying according to the size and number of voyages of ships using the lighthouses. No mystery here, Coase claimed: lighthouses operated like any other service provided by private firms and paid for by their clients.

Perhaps not surprisingly, the lighthouse article was celebrated by defendants of free markets and critics of government intervention. The logo of the Independent Institute, a think tank, pictures a lighthouse in honor of Coase’s paper – “as a symbol of courage, enlightenment and independence.”³ The mere existence of public goods could not be a valid justification for public provision. Suddenly, private entrepreneurs seemed to have no limit on what they could do for society.

Subsequent work, however, reinterpreted some of Coase’s conclusions and provided additional details on how private lighthouses operated. Coase was right to argue against the received wisdom that lighthouses were public goods and as such required public support. Yet the government was far from absent. Private owners of lighthouses were granted licenses to operate for a prescribed time and were authorized by the government to enforce the payment of fees, which were also regulated. Some ship owners were coerced to pay fees even against their will. And, during the course of the nineteenth century, the funding of lighthouses became a system supported by public money.

Although Coase did report these latter events in his article, critics argued that he largely underestimated the role of the government in regulating and sponsoring public services. Definitely, the system was not a free market, where firms and users voluntarily negotiate service terms and prices.⁴

However, both admirers and detractors of Coase’s analysis missed the point: he never meant to use the lighthouse case as proof that private enterprises can successfully replace governments in the provision of public goods. On the contrary, he was very clear that we should avoid generalizations without a deep understanding of how either model, with more or less private participation, could work in the real world. Such an approach, Coase argued, would help us understand “the richness of the social alternatives between which we can choose.”⁵

This statement is pure Coase: before his Reference Coase1974 paper on lighthouses, he had already written fundamental papers arguing for a comparative approach to economic organization. Put simply, given any social or private objective, the Coasean tradition calls for a detailed examination of the merits and costs of alternative organizational models, free from preconceived positions.⁶

1.2 Do private motives undermine public gains? The cost–quality trade-off

For a long time we did not have a clear theory of what key factors would affect the benefits and costs of private versus public delivery. A substantial advance came with a paper by Oliver Hart, Andrei Shleifer, and Robert Vishny, published in 1997. I will henceforth refer to this paper as HSV.

Hart was already a leading scholar studying the boundaries of the firm: that is, whether firms should integrate or outsource a given activity – a line of research inspired by Coase’s early writings on comparative economic organization. A natural extension would then be to examine the boundaries of government: factors that may stimulate or prevent the outsourcing of public activities to private actors. Shleifer and Vishny, in turn, were studying the pros and cons of government intervention, following privatization reforms across the globe.⁷

A key assumption of the HSV model – as in much of the literature on the boundaries of the firm more generally – is that contracts are incomplete. In the context of prisons, which HSV use to illustrate their model, the key difficulty is how to specify contractual terms that guarantee high “quality” of service.

In examining the private provision of prison services, HSV call attention to two important issues. First, will prison managers use excessive force and violence to avoid escapes and deal with internal conflict? Second, will the firm running the prison hire properly qualified personnel to guarantee adequate treatment and provide a broad array of complex services (including health and legal assistance to inmates)?

We can also consider other relevant dimensions such as the extent to which the firm will try to promote resocialization via within-prison training programs, psychological support, and other activities that eventually reduce recidivism (i.e., repeated engagement in criminal activity after inmates are released).

Notice, therefore, that “quality” refers not only to service attributes that generate direct benefits to inmates (e.g., health support or training) but also to a host of “external effects” on society. Recidivism, for instance, causes societal loss due to recurring criminal activity and the need to maintain costly police activity on the streets. These are negative externalities emanating from low-quality services within prisons. Therefore, a well-run prison has public good characteristics: it not only promotes resocialization but also improves the security of citizens.⁸

Why are these quality dimensions difficult to contract for? Imagine trying to specify a desirable “level of violence” in a contract with a private prison operator. Even if the government tries to specify a list of banned behaviors, deviations occurring within the prison will be observed by the inmates and prison managers, not by an external judge assessing whether the private firm is guilty of violating contractual terms. That is, some quality dimensions may not be verifiable by third parties.

Complicating things, the very process of drafting and monitoring contracting terms is costly, let alone all expected judicial costs in the case of disputes. All these difficulties have to do with what Coase and several subsequent economists referred to as transaction costs. Indeed, contractual terms in privately outsourced prisons are generally left unspecified, especially for more complex dimensions of performance.⁹

Because these quality dimensions are hardly contractable, a crucial aspect is who will define and monitor what are the appropriate levels of violence, services provided to prisoners, resocialization initiatives, and all other service attributes that not only benefit the inmates but also generate positive externalities. If contractual terms are vague, then the private manager has the right to decide whether to invest or not in the provision of these attributes.¹⁰ Yet providing high quality is costly, and private operators may be tempted to focus on activities that directly affect the bottom line. For instance, whether resocialization practices make former prisoners commit fewer crimes outside the prison is immaterial to the management of operations inside the facility.

The HSV model provides a structured way to think about the costs and benefits of private engagement. On the one hand, the management of activities by private actors will create incentives for cost reduction. This is not necessarily bad: we regularly complain about the high costs of various services. In the case of prisons, firms profit from payments that come from government contracts. More generally, the fact that private actors appropriate financial gains incentivizes better operational performance and perhaps even innovations that increase productivity – for instance, surveillance technologies and improved process management practices.

On the other hand, the profit-based incentives of private operators may lead to the neglect of desirable quality dimensions that are difficult to measure and specify in contracts. This temptation will not be as strong in state-run prisons, as public managers will not directly gain from cost reductions or any other action that improves productivity. They usually receive flat wages and often do not have a clear mandate to pursue aggressive savings – in the jargon of organization economics, their incentives are “low-powered.”

Yet it is precisely public managers’ lower inclination to perform on economic dimensions that also reduces their temptation to cut costs at the expense of quality. Therefore, the decision to privatize involves a cost–quality trade-off: private management is likely to improve productivity and reduce costs, but it might also lead to savings on quality attributes that are not specified in contracts.¹¹

As discussed in the Introduction to this book, another way to frame this trade-off is to note that the performance of a given public activity involves the ratio of social benefits to costs. An effective prison will avoid excessive violence, promote resocialization, and take care of all relevant service dimensions that are valued by society. The ratio of these benefits to all relevant operation costs is an effectiveness–cost indicator of service performance or, to simplify the term, a measure of its effectiveness.¹² We cannot generally say whether public or private management will be more effective because they differently affect this ratio. Private firms may reduce the denominator (costs) but their cost-cutting effort may reduce the numerator as well (benefits). However, if the hazard of neglecting important but noncontractable attributes escalates, privatization becomes less and less appealing from a societal standpoint.

Therefore, a general prediction of the HSV model is that privatization can be particularly problematic in cases where private operators can substantially profit from inattention to social or public dimensions that are difficult to specify in contracts.

Empirical studies have generally supported this prediction. Services whose quality attributes can be easily measured and enforced via contracts – such as street cleaning and maintenance of public facilities – are more likely to be privatized and to yield improved public outcomes under for-profit private management than services with difficult contracting attributes – such as complex health programs, crime prevention, and prisons.¹³

Yet, as is usual in the social sciences, straightforward predictions may not hold when some crucial contextual conditions change. Although useful and generally aligned with empirical evidence, the HSV model does not fully accommodate the various organizational arrangements that are found in practice and that can affect the cost–quality trade-off.

To continue with the example of prisons: first, a prison may not be fully “privatized” but instead be subject to close public supervision. As mentioned in the Introduction, Sandro Cabral, Paulo Furquim, and I studied prisons in a Brazilian state (Paraná) where the operations of the prison were outsourced but the state kept public agents (wardens) working on site.¹⁴

In this hybrid model, while contractual terms were vague and incomplete, the private firm also did not have full control of decisions that were not contractually specified. For instance, any attempt by the private firm to cut service costs (such as reducing the extent of health and legal assistance for prisoners) would need to be sanctioned by the public warden. Supervision helped to mitigate contractual incompleteness.

To be sure, this model entails a big risk. What if the private firm bribes the public warden to approve unjustified cost reductions? These bribes are also difficult to observe and verify by external parties.

We unveiled, however, a host of factors that helped control this type of corruption. The public wardens appointed to the private prisons were handpicked by the government and received additional temporary compensation (in some cases close to 40 percent higher than they would receive in similar public-sector jobs). Even though some of them could not be completely fired from the public sector, they would lose this extra salary if caught deviating. This mechanism is what economists refer to as an efficiency wage: an extra payment devised to incentivize higher performance.¹⁵

But how could these deviations be detected anyway? Another important factor was the presence of intense external monitoring by civil society – nonprofits specializing in human rights, religious organizations, and the media. Although corrupt deals can be hardly observed, allowing the private operator to pursue excessive savings would be likely to lead to quality deterioration and complaints – which can be directly observed by those external actors who are in touch with inmates and which generate public commotion when images of violence and internal rebellion appear in the news.

As such images tarnish the popularity of the politicians in charge, pressure to replace the public warden would probably follow suit. No bribe might compensate for the loss of the extra salary (and good reputation) that supervisors could otherwise obtain by running an effective prison. (See Appendix 1 for a formalized explanation of this mechanism.)

Second, the cost–quality trade-off assumes that private firms will be profit maximizers and will probably pursue short-term gains by economizing on quality. What if private firms focus on the long term and, like the public warden in the above example, try to preserve their reputation in future exchanges with the government?

A study by Eyub Yegen found that, as HSV predicts, privatized prisons underperform their public counterparts in terms of difficult-to-contract dimensions (measured in Yegen’s study as the incidence of prisoner suicides). However, this negative effect of privatization was not necessarily observed in the case of prisons run by private firms whose owners have a longer-term orientation (such as pension funds without a strong emphasis on immediate profitability).¹⁶

Third, governments may try to “complete the contract” by finding ways to measure relevant social outcomes and compensate private operators (and their investors) according to these metrics. A highly publicized initiative along these lines was the launch of the first social impact bond in the United Kingdom in 2010, a joint effort by Social Finance (a nonprofit organization) and the UK’s Department of Justice.

The project involved a prison in the city of Peterborough, and the chosen metric was prisoner recidivism – measured as the percentage of released inmates committing new crimes, compared to the percentage among inmates coming from other prisons. Investors funded a bundle of activities, including job training and family support, which were found to reduce recidivism by 9 percentage points.¹⁷

All these actions require governments to contract with private firms and establish evaluation and monitoring mechanisms. A more fundamental question, however, is: Are governments even necessary to such arrangements? Can private actors alternatively self-organize to achieve collective benefits?

1.3 Covenants without a sword

The philosopher Thomas Hobbes (1588–1679) famously coined the term “Leviathan” to describe the coercive power of the state – in his view, necessary to “bridle men’s ambition, avarice, anger, and other passions.” The cost–quality trade-off discussed earlier is a manifestation of this more general problem: private actors may try to increase their profits at the expense of relevant service attributes. Could we ask them to make promises (written or spoken) to act with a more publicly oriented mindset? Not according to Hobbes: “Covenants, without the sword, are but words.”¹⁸

This allegory was used by political scientist Elinor Ostrom (1933–2012) and colleagues to examine the possibility of collective gains with and without governments.¹⁹ The “sword” is an “external enforcer.” If a contract can be written, then the judicial system can apply sanctions in case of verified deviations. However, as a critical assumption of the cost–quality trade-off, contracts are incomplete and hence some critical service attributes may be left unspecified. Leviathan may have to retain control rights to supervise private actors or directly command what needs to be done.

Contrary to this view, Ostrom has forcefully argued that covenants can work without a sword. Her focal problem was also how to solve dilemmas in the provision of goods that are valued by society. There are public goods, which are broadly available (nobody is excluded) and whose consumption by one person does not subtract from the amount of goods available to other people. We have also discussed cases of goods that have a “public” aspect but are not totally inclusive – for example, education in excellent public schools, which are limited in supply. And, of course, there are private goods whose access is restricted to those who can afford them.

Ostrom highlighted another important type of good: common-pool resources. Typical examples are natural resources: forests, lakes, mines, and fertile land. Although in some cases these goods are privately owned (such as company-owned mines), in other important instances they are in the public domain and accessible by either legal or informal means. Distinct from public goods, however, common-pool resources are subtractable: each additional fish caught in a lake reduces the stock of fish available to other people.

The ecologist Garrett Hardin (1915–2003) coined the term “tragedy of the commons” to describe the dilemmas of exploiting common-pool resources.²⁰ Imagine two farmers whose crops border a forested environmental protection area on land that they own. They can agree to not cultivate in that area, because they realize that preserving it will provide common benefits – for instance, the forest will prevent excessive soil erosion in case of heavy rains and will help preserve river sources and natural predators of crop pests.

Yet money is on the table. A farmer can marginally expand their own crop area and still benefit from the (partial) forest kept by the neighbor. But why would the latter keep the area if there are substantial gains from farming in the remaining forest area? If these economic gains are high enough, the farmers will soon arrive at a situation where they both farm in the protected land and the forested area disappears. Picture thousands of farmers facing the same situation, and Hardin’s tragedy can be easily seen.

Leviathan can step in and control the access to common-pool resources via ownership (the protection area can be expropriated and become publicly owned) and/or impose restrictions on the use of the area and heavy fines to those who deviate. All these remedies are imperfect. Not only are contracts incomplete: so are regulatory systems. The government or regulator may have a hard time monitoring and enforcing protection rules in large areas; farmers may bribe public auditors and, even if this fails, they can contest their fines in court.

Could farmers alternatively organize themselves to preserve the forest, even without Leviathan? Ostrom develops a theory of self-governance based on the creation of voluntary, mutually agreed rules. Going back to the previous example, imagine that there are many farmers and that they decide to create a cooperative to jointly sell their products. They realize mutual gains from preserving the green areas: this will allow more sustainable production and enable the farmers to target conscious customers who are willing to pay a premium price for environmentally friendly products.

Instead of a centralized Leviathan, we would have a localized nexus of voluntary agreements to manage common-pool resources. Essentially, farmers would need to define the boundaries of their group and the roles of their members (who is in, who is out, who does what). They would also need to define rules of action indicating what can be done and what is forbidden in their agricultural areas.

They would need to create communication and deliberation mechanisms to define these rules and adapt them to changing conditions. Finally, and crucially, they will have to define how the surplus from the whole operation can be split among members and how to proceed when some farmers deviate from the original rules.²¹

At a more fundamental level, self-governance is based on a relational mechanism: parties define and self-enforce their desired outcomes. Relationships essentially emerge from the possibility of repeated interactions. The “shadow of the future” creates an incentive to abide by the rules because potential deviations can trigger sanctions and therefore result in forgone future gains. For instance, deviating farmers may be expelled from the cooperative and lose the opportunity to sell their products at higher prices. The “shadow of the past,” in turn, results from a history of interactions and allows the members to learn from one another, communicate adjustments that need to be made, and develop norms of behavior.

This relational mechanism is, however, more easily described than created. In order for it to work, two key conditions must be present. First, there must be clarity on what mutual gains can be achieved. Some farmers, for instance, may not fully understand the potential benefits of preserving the area and may think that the rules are constraining their freedom to profit from their land. Broad and frequent communication is essential.

Second, the rules must be credible: reneging on the overall agreement must trigger sanctions. If influential farmers are able to persuade other cooperative managers to keep them in the group even after substantial deviations, this will reduce the credibility of the agreement and undermine incentives to follow the preagreed rules.²²

If groups overcome these difficulties, they will be able to solve social dilemmas without Leviathan. An irony, however, is that the same relationships that support self-governance can also improve private interactions where Leviathan is a partner instead of an external enforcer.

To be sure, mentioning “relationships” with the government can immediately raise eyebrows, given the risks of corruption and preferential treatment of well-connected private actors. For this very reason, it is common to govern public–private interactions via arms-length public auctions where several firms regularly compete for a given contract. At first glance, such an impersonal approach to contracting may appear to be antithetical to relational governance. In reality, relationships are pervasive in public–private interactions.

Ricard Gil and Justin Marion studied more than 5,000 procurement auctions coordinated by the California Department of Transportation (Caltrans) between 1996 and 2005. Private firms had to choose price bids, and the lowest bid would win each individual auction. However, there were multiple auctions and the same firm could participate in sequential contracts. The repeated nature of the auctions created a shadow of the future: a firm would consider not only what it could gain in any given procurement contract being auctioned off, but also potential revenues from future contracts.

The shadow of the past was also relevant: each private bidder nurtured a stock of relationships with subcontractors based on past interactions. As it turns out, private firms lowered their bids when they had more past relations with subcontractors and expected more future business with the government. Caltrans may have not realized it, but the department benefited from evolving relationships with recurring private partners.²³

1.4 Government failure and public bads

Public goods, externalities, and the tragedy of the commons are all seen as potential sources of market failure: individual, decentralized decisions may fail to generate collective benefits. Although it is always tempting to recommend government action as a solution to market failure, we have just discussed how relational governance may support voluntary arrangements where repeated interactions supplant atomized exchanges. In other words, echoing Coase, market failure is not a necessary condition for invoking Leviathan.

Some push this argument even further: government action may itself be problematic. More precisely, attempts to solve market failure may create another form of inefficiency: government failure.²⁴ The remedy may end up being worse than the disease.

First, Leviathan may be a bad owner, as is easily explained by the well-known principal–agent problem. In private firms, there is a group of owners or “principals” who delegate key tasks to managers or “agents.” To ensure that agents adequately perform the assigned tasks, principals must set performance standards and monitor organizational outcomes.

But who are these principals in public bureaucracies? Ultimately, money to fund state-owned operations comes from taxes levied on citizens. Yet citizens can hardly check whether public managers are doing a good job, except perhaps in the case of services that they more directly demand. Even in these cases, citizens may not have a well-defined way to complain or to request improvements in service delivery.

In democracies, society ultimately delegates the monitoring function to elected politicians who, with some exceptions, generally have scant incentives to push public managers toward higher performance standards. In authoritarian regimes, the dictators in charge will do whatever they want in ways that may be largely dissociated from citizen’s interests.

Second, as well as being a bad owner Leviathan may be a bad manager. Our previous discussion on the cost–quality trade-off noted how public managers are less inclined to be cost-effective. Unlike the wages of managers in private firms, who may receive bonuses or performance-contingent payments based on indicators of profitability and productivity, the wages of public managers are more responsive to technical descriptions of the job and tenure. Eliciting higher productivity via career development is even more challenging, and in certain contexts it is difficult, if not impossible, to fire public agents who consistently underperform.²⁵

As if these problems were not enough, Leviathan is not a single actor but instead a web of political interests. Consider the case of state-owned enterprises created to manage natural resources or provide essential services (such as water or energy). While private owners will tend to choose corporate investments depending on their expected profitability and to hire new personnel based on their skills and merit, politicians may try to influence investments in regions where they can reap higher political dividends and may appoint their cronies to high managerial ranks.

Consistent with this discussion, several studies have detected operational performance gains in the transition from state to private ownership. William Megginson, a specialist in privatization, surveyed more than 100 studies of state ownership in several countries and became convinced that “‘state capitalism’ is an essentially failed model.” As always, there are exceptions and the final effect is highly context-dependent – for instance, operational gains seem to increase when there are well-defined private owners monitoring the performance of the privatized companies as well as well-designed privatization processes to prevent corruption in the passage from state to public ownership.²⁶

These arguments indicate that public organizations are expected to act in a way that is dissociated with the overall objective of solving market failure and that may even deliberately create government failure.

The escalation of ineffective public action was vividly discussed by William Niskanen (1933–2011), an economist who contributed to the market-oriented programs put in place by US president Ronald Reagan in the 1980s. Reagan declared in his inaugural speech that “government is not the solution to our problem, government is the problem.” Accordingly, Niskanen argued that bureaucrats will tend to “overproduce” their services and will eagerly fight for public resources. They will maximize their own budgets and may eventually generate societal loss if excessive taxation to fund public activities fails to generate corresponding public benefits.

The work of Peter Klein, Joseph Mahoney, Anita McGahan, and Christos Pitelis, who studied the performance implications of public capabilities, suggests that Leviathan may promote public goods but may also become increasingly competent at providing public bads.²⁷ Accordingly, a frequent theme in this book is that the performance of state-run operations will crucially depend on the ability to keep government failure in check – or at least to keep it at levels that do not create large downstream distortions.

1.5 Is privatization “just”?

Most studies comparing private and public management focus on whether they are more or less effective in generating societal gains. Are public schools or hospitals better than similar privately run units? Do they generate benefits that outweigh their costs? Such analysis examines the supply side.

However, there is a related problem on the demand side. Even if private firms perform better than public units, can they profitably operate in more vulnerable segments such as remote and deprived areas or populations with severe income constraints? And if not, should governments tax more affluent citizens to support services for the least advantaged – either via direct public action or via private management with public sponsorship? Are these taxes even justified in the first place?

A heated debate in the 1970s between two scholars at Harvard University’s Department of Philosophy shed light on these questions. One of these philosophers, John Rawls (1921–2002), published in 1971 a book that became a seminal reference in the study of how societies should address problems of liberty and equality. That book, A Theory of Justice, follows a liberal tradition: in principle, individuals are free to choose the rules of interaction that best promote cooperation and progress.²⁸

In a thought experiment, Rawls asks what general principles of justice free individuals would choose if they were asked to forget who they are: their wealth, their genealogy, their position in society, their whole history. Rawls refers to this condition as the veil of ignorance.

Rawls suggests that people under the veil of ignorance would arrive at two general principles. The first principle states that all people should enjoy general liberties that include freedom to work, speak, think, and vote, as well as protection from theft and arbitrary expropriation of their possessions. The second principle, in turn, starts with the idea that people must have equal opportunity to assume relevant positions in society, such as competing for a public office or becoming a successful entrepreneur.

However, not all individuals will have the necessary skills and resources to be selected to those positions. Rawls thus proposes an additional requirement: every existing inequality must be for the benefit of the worse-off. A contemporary term for these objectives is inclusion: granting access to all, especially the most disadvantaged populations in society.

Who are those disadvantaged groups? Most discussions on the matter focus on income disparities. Thus, the worse-off are individuals and families constrained in their ability to purchase products and services with adequate quality standards. Yet, as economist Amartya Sen forcibly argued, income deprivation is only one of the multiple challenges of the disenfranchised.²⁹

For instance, even when individuals can afford basic goods, they may have a host of cognitive or physical disabilities that limit their potential to learn fundamental skills, find a profession, and fully enjoy their lives – related not only to their income but also to a broader set of adverse conditions brought by impaired physical capacity, lack of family support, societal violence, susceptibility to natural disasters, and many other threats.

Broadly speaking, while the arguments we have discussed emphasize the relative effectiveness of public and private management, we can also assess the two types of management in terms of their social justice. How can we guarantee that the least advantaged are well served? Rawls gives an example of a concrete policy: “the government tries to insure equal chances of education and culture for persons similarly endowed and motivated either by subsidizing private schools or by establishing a public school system.”

In other words, according to Rawls, public management is a possibility, but it is not necessary to achieve social justice. In Coasean fashion, he suggests that “the question then becomes one of comparison of possible alternatives.”³⁰ However, for Rawls, society must prioritize the needs of the least advantaged – meaning, for instance, that governments must guarantee that good education is not available only to those who can afford premium private schools.

In the same department at Harvard, Rawls’s colleague Robert Nozick (1938–2002) disagreed with his formulation. Three years after A Theory of Justice, he published another seminal book: Anarchy, State, and Utopia. Even if inequalities exist, Nozick argues, can anyone – including governments – have legitimate discretion to do something about them? For instance, is it acceptable to put limits on how much money individuals can make or require that they redistribute their wealth via compulsory policies?³¹

If Nozick’s principles apply, will the least advantaged be poorly assisted and condemned to their underprivileged status? Not necessarily, argues Nozick. For instance, to combat low wages or a poor provision of valuable goods, Nozick asks why workers cannot organize themselves in a voluntary entrepreneurial activity. A possible strategy could be to form a worker association or cooperative, in which all members have equal voting rights. There are also cooperatives designed to provide critical services, such as education or credit.

It is a not a coincidence that this suggestion resembles Elinor Ostrom’s relational mechanism discussed in the previous section. Nozick also considers that individuals do not necessarily need Leviathan to solve collective dilemmas. In this case, however, the cooperative (or any other form of collective agreement) is designed not only to address potentially failed markets, but also to repair failures of justice.³²

Following this argument, we also cannot rule out the possibility that private corporations will see opportunities to target vulnerable population segments. A thriving literature in business strategy examined how companies may not only target the least advantaged but also do so profitably. The idea is that firms can leverage their distinct competencies to create and deliver better products for high-growth, low-income markets that are poorly served by existing players.

C. K. Prahalad and Alan Hammond estimated a “poverty premium” due to this lack of sufficient supply. For instance, the residents of Dharavi, a slum in Mumbai, India, paid 10 times more for medication and 53 times more for credit services compared to prices in more affluent areas.³³ Given their large size in many developing and emerging countries, private companies could potentially exploit these neglected markets with better products and productive operations.

More generally, management thinking has increasingly considered that private companies need to move beyond an owner-centered perspective and accommodate the interests of a broad set of relevant actors affected by their business model – the so-called stakeholders.

A variant of the stakeholder-based view of the corporation is more pragmatic: its proponents argue that managers simply cannot create wealth for their owners if they do not take care of their multiple stakeholders. For instance, treating and compensating employees well may eventually lead to better products and more satisfied customers. The company then increases sales and profits.

This mechanism is the basis of the popular “doing well by doing good” tenet and follows a more instrumental perspective of stakeholder relations. In this view, whether companies are genuinely interested in the common good is immaterial, since they may have a self-interested incentive to create and share economic gains.³⁴ Nothing, in this case, is coercing firms to benefit their stakeholders. They do so because it is in their own interest to engage stakeholders in the creation of a unique business model. Corporate managers and owners would be totally free to move in an alternative direction, at their own will and risk. Nozick would be happy.

We can push the argument even further by arguing that the owners of private companies may follow objectives other than profitability. An influential report by J. P. Morgan in 2010 explained the emergence of impact investors seeking companies that blend profitability and socio-environmental dimensions of performance – for instance, startups providing low-cost learning technologies or affordable health clinics.³⁵ Specialized impact-investing funds, such as Bridges Ventures in England or Vox Capital in Brazil, were created to promote such entrepreneurial ventures using resources from wealthy families and conscious investors.

In fact, the old notion that companies should focus on profit maximization is being increasingly challenged in both theory and practice. We can again invoke Oliver Hart (who proposed the cost–quality trade-off), this time in an influential paper written with Luigi Zingales. They proposed that corporations should pay attention to the preferences of shareholders, not necessarily the market value of their shares.³⁶

The presence of investors who willingly seek to generate public benefits – and even sacrifice part of their profits to achieve these – opens the possibility of successfully engaging private firms in activities that were traditionally managed or funded by the state. And this potential, at least in principle, potentially reconciles Rawls’s emphasis on the disadvantaged with Nozick’s opposition to involuntary transfers.

However, there are also important limits to private arrangements intended to promote inclusion. Consider the following story. (I have personally come across variations of this dilemma on several occasions.) An urban development company proposes to revitalize a poor neighborhood. In the beginning, the idea is to improve the local infrastructure and build homes for families that live in the region.

However, the project is so successful that many people outside the area are interested in buying or renting homes. Prices soar, and the firm considers further upgrading the infrastructure to lure higher-income residents. Internal studies indicate that adjusting the original plan to accept the more affluent customers would generate higher profits in the long term.

This adjustment can be easily executed, and the company is free of any government regulation requiring that it should necessarily accommodate low-income families. However, adjusting the plan would imply deviating from the initial focus on the more disadvantaged residents, which is inconsistent with Rawls’s emphasis on the wellbeing of the worse-off. Low-income families do not have sufficient resources to afford housing at the higher prices. They could possibly self-organize and create a form of credit cooperative, as Ostrom and Nozick would suggest, but such an investment would be relatively large and require additional funds that might be difficult to access in credit markets.

The company is free to choose either course of action: stick with the original plan, or adjust it to boost its profits by attracting more affluent clients. By adjusting the plan, the company does well financially but fails to give priority to the poorer families. By keeping the original plan, the company benefits these families but underperforms from a financial standpoint. Doing well by doing good is not possible here. The company needs to make a choice, and a very fundamental one: prioritizing profits or the wellbeing of their low-income clients.

Suppose that company managers decide to stick with the original plan to support income-constrained families. A key question will be how to fund their acquisition or rental of units. An option would be to structure internal cross-subsidies: the more expensive units, paid for by the more affluent residents, could support lower prices for the constrained families. In this arrangement, the owners of the site will be asked to sacrifice part of their profits. They will be tempted to simply set market-based prices, attract higher-income clients, and devote only a tiny fraction of the area, or even none at all, to the lower-income populations who were part of the original plan.

How can this hazard be overcome? A possibility is to attract impact investors with a more balanced pursuit of economic and social objectives. They can sit on the board of the company, closely interact with managers, and persuade the remaining owners that they should commit to the original plan. Even wealthy investors, however, have limits on the number of beneficiaries and projects they can support. What if the initiative is so successful that other cities and even countries would like to replicate it? Perhaps this could be part of a larger public policy program to support affordable housing.

Chile, for instance, implemented a system of subsidies given to low-income families to purchase residential units built and sold by private firms. In principle, this is a market-like mechanism whereby constrained individuals are provided with monetary resources to freely choose their preferred options. If the subsidies were insufficient to cover the whole price of the unit, then families could borrow money from private banks. However, because these private banks considered some families extremely risky given their poverty condition, a state-owned bank, BancoEstado, had to step in and guarantee lending to this vulnerable segment.³⁷

In this example, voluntary private action without government-imposed transfers was not possible. Leviathan was called upon to ameliorate demand-side constraints and also to mitigate the reluctance of private lenders to serve very poor segments. As in the case of effectiveness, achieving inclusion also requires an informed analysis of the merits and costs of alternative delivery forms – whether public, private, or a mix of those two. We cannot simply assume that either private or public solutions will automatically and single-handedly address critical problems of access.

1.6 From public goods to public values: the legitimacy of privatization

The previous discussion converged into two main outcomes to be pursued in the decision to privatize: effectiveness and inclusion. Yet this choice is constrained by societal perceptions of how to properly organize a critical activity. The sociologist Max Weber (1864–1920) declared that the state has a monopoly in the use of force in a given territory.³⁸ Although private enforcement organizations do exist, the police and the military providing security at a large scale are all state actors. Could these activities be outsourced to private firms?

As frequently seen in dystopian stories, the presence of corporations in public security activities raises fears of abuse for private gain and dominance. More generally, citizens may understand that they have rights – in this example, peace, freedom, and life – and that the government has the inalienable prerogative to defend these rights. In democracies, the creation and funding of public organizations should be approved by legislatures composed of elected officials. Policies engaging public organizations should derive from broad discussion, and the missions of these organizations should be explicitly tied to societal demands.

Public management scholars refer to these principles influencing what governments should or should not do, and how to do it, as public values. In our context, these values, as well as considerations of effectiveness and inclusion, may constrain the decision to engage private actors in public initiatives. A hypothetically benign private military organization might be quite effective and inclusive (in the sense that all citizens are equally protected), but society may consider that these services should require public deliberation and public involvement.

In other words, privatization is not a “technical” decision defined by a selected group of policy makers and their economic advisers. The whole process must be perceived as legitimate by society.³⁹

Legitimacy crucially depends on the perceived intentions of those in charge of public activities. Oliver Williamson (1932–2020), a key contributor to the comparative analysis of organizational forms in the Coasean tradition, argued that probity is a key determinant of decisions at the public–private boundary. In his view, “probity implies a high standard of integrity, to include professional excellence, in the organizational unit to which a task has been assigned.”⁴⁰

To illustrate, Williamson gives the example of how to manage a country’s foreign affairs. What would happen if these activities were outsourced to a private corporation? Suppose that the country faces an unexpected external threat and requires additional costs that are not specified in the original agreement. The political leader in charge may also require the support of costly and specialized personnel to tackle diplomatic bickering.

Perhaps the government could impose rules and regulations indicating how the private firm would have to respond to shifting public demands. However, Williamson argues, “the private bureaucracy will view these through the conflicted lenses of foreign policy efficacy and the bottom line, which is to say there are trade-offs.”⁴¹ The solution, then, is not to outsource these activities and instead to rely on public servants loyal to the state mission and buffered from financial pressure.

This problem is, in its essence, related to the aforementioned cost–quality trade-off. Any contract between state leaders and a private firm managing foreign affairs will be substantially incomplete, and this will potentially lead to the neglect of relevant noncontractable attributes. Echoing HSV, Williamson offers a prediction in the case of prisons: “Ownership of the specialized physical asset (the prison) by the government and franchising out the operation of the prison is a possibility for prisons, but quality considerations – herein reflected as probity – signal precaution.”⁴²

Legitimacy concerns are also highly relevant in the implementation of privatization programs involving the sale of state-owned enterprises providing services such as transport and utilities. The neoliberal reforms of the late twentieth century were met with criticism that public policies previously executed by those enterprises could end up in the hands of private capitalists. Will they care about whether local populations have adequate access to the services? Are they just trying to make profits out of large customer markets or are they alternatively motivated to promote local development? It is no wonder that several privatizations were partial, with governments retaining minority shares in the newly privatized companies.⁴³

Perceived legitimacy, however, may vary across organizations. Consider the case of institutes, foundations, or civil society organizations participating in the delivery of public services. Their not-for-profit orientation helps mitigate the cost–quality trade-off by curbing excessive incentives to cut costs or increase operational surplus at the expense of noncontractable service attributes.

They also tend to be mission-driven and specialized. Unlike for-profit firms, their objectives may be more naturally aligned with well-defined groups of beneficiaries –as in the case of nonprofit hospitals and schools targeting low-income populations.⁴⁴ All these factors may increase their perceived legitimacy and attenuate criticism that privatization programs are designed with the purpose of creating profit opportunities out of public activities.

As always, there are trade-offs. Because managers of nonprofits cannot legally appropriate their surpluses, their incentives to perform, as in public organizations, are low-powered. Their need to continuously fund and expand their operation may also lead them to cater to potential donors instead of beneficiaries. And their limited resources may constrain the scale of their potential impact.

In other words, even if the engagement of nonprofits in public services is perceived as more legitimate than the engagement of for-profit companies, it is not totally clear whether these organizations will be more effective or inclusive than the alternatives (including state-run operations). Legitimacy cannot be seen in isolation; it also depends on organizational capacity to deliver results at a scale.

1.7 One issue, many arguments

The plethora of arguments in favor of and against private engagement in public initiatives (summarized in Table 1.1) may cause an initial impression that the overall debate is fickle and inconclusive. Their apparent disparity, in reality, is due to their emphasis on distinct focal issues. While the perspectives from the points of view of public good, cost–quality trade-offs, relational contracting, and government failure focus on effectiveness, the justice perspective centers on whether public or private actors will promote inclusion or will instead fail to address the need of vulnerable population segments.

In addition, most of these perspectives generate arguments that both support and discourage private engagement. For instance, the cost–quality trade-off predicts that for-profit companies may cut costs at the expense of relevant service outcomes. Yet lower costs are also a positive attribute of public services; they reduce public expenses and the need of substantial taxation. Could governments intervene to guarantee adequate service quality and still benefit from these potential savings?

Instead of selectively picking a theoretical argument and generically defending public or private engagement, the appropriate exercise is therefore to scrutinize the trade-offs involved in each form and propose ways to deal with their inherent dilemmas. We thus need a consolidated decision framework that shows when and in which conditions a certain delivery form is preferable to others.

To advance this analysis, I will proceed in the next chapter with a more detailed discussion of the broad outcomes that privatization decisions may pursue: effectiveness and inclusion. In the sequence, I will propose a consolidated decision framework indicating in which conditions public, private, or hybrid (public–private) forms of delivery may or may not help achieve those intended outcomes, conditional on their perceived legitimacy – so as to avoid recommending unrealistic “solutions” that eventually face strong public opposition and never see the light of day.