Privatization and Antitrust (and other aspects of Competition Policy) — Part 2

This is the second in a series of posts serial-blogging a draft chapter I’ve written on privatization and competition policy (including antitrust) for a forthcoming Stanford University Press book on Competition and the Role of the State. The first part was yesterday. What follows represents the current draft, though sans footnotes for the blog format. Sorry, I don’t have it posted on SSRN yet, so I can’t link to the full document, but drop me a line if you’d like the Word document.

Last time, I discussed theoretical reasons to think that whether a service is privatized may not be relevant. Now, I go on to see why these theoretical reasons might not hold up, and why privatization might be relevant anyway.

Relevance?

Sappington and Stiglitz didn’t present a theory of government behavior that would allow a comparison of government and private ownership; that’s the story of the post-1987 theoretical research. I summarize a few of these models below, to give a flavor of the sorts of contractual incompleteness that could support a difference between different modes of ownership.

One possible difference between public and private agents, suggested by Jean-Jacques Laffont and Jean Tirole, is that the private manager has to answer to two principals, the government and the shareholders, while the public manager only has to answer to one. The disadvantage is that cost-reducing effort is suboptimal: “The multiprincipal situation dilutes incentives and yields low-powered managerial incentive schemes.” The advantage is that the private manager has greater investment incentives, because the shareholders are less likely to expropriate the manager’s investment to secure greater social benefits. So privatization trades off these two types of effort and investment.

Another difference, suggested by Klaus Schmidt, is that the government might have better access to inside information about the agent’s finances when the agent is public. Knowing the agent’s costs can lead to bad incentives. If costs turn out to be high, it will be ex-post optimal to subsidize the enterprise, but this leads to low ex-ante incentives to reduce costs. A government ignorant of costs won’t subsidize as much and will inefficiently shut down some enterprises, but at least cost-reducing incentives will be improved. The tradeoff is between productive and allocative efficiency.

Another strand of the literature, spawned by Hart, Andrei Shleifer, and Robert Vishny (HSV), stresses the importance of residual control rights—that is, the right to do whatever’s not prohibited by the contract. A contract to run a government program—say, a prison—only specifies a basic service, but the agent can invest in thinking up various innovations to the service. Some innovations cut costs (and, by assumption, reduce quality); private managers spend too much effort thinking of these, because they keep every dollar saved. Other innovations improve quality (and, by assumption, cost more). Private managers will be able to appropriate some of the net benefit by renegotiating the contract with the government; their incentives to think up such innovations are suboptimal, but at least better than those of public managers, who have a more precarious bargaining position. Privatization trades off these two types of innovation.

The HSV model has been quite influential. Hart himself used it in later work on public-private partnerships (PPPs), where he discussed whether the contractor who builds the facility should be the same as the one that runs it. Patrick Schmitz also used an HSV-style model to discuss the advantages of joint ownership, and Timothy Besley and Maitreesh Ghatak used it to discuss privatization in the presence of altruistic providers.

Most models assume a benevolent government. (“Benevolence,” in the lingo of economists, is invariably taken to mean social welfare maximization in a utilitarian framework, or something more or less like it.) In some models, non-benevolence makes privatization more desirable, as private ownership limits the scope of politicians’ private agendas. But of course a non-benevolent government can also transfer resources to its supporters by non-benevolently regulating private firms, so the superiority of privatization doesn’t necessarily follow: “with full corruption the allocation of control rights and cash flow rights between managers and politicians does not affect either the efficiency of the firm or the transfers it receives.” Moreover, if the government continues to regulate firms heavily, privatization (the transfer of cash-flow rights) “may actually make things worse. Politicians continue to use their control of regulated firms to pursue political objectives, but it is now less costly for them to do so.” Transferring control rights—not just cash-flow rights—to managers, a process that Shleifer and Vishny call “commercialization,” may be necessary for privatization to work. So the costs and benefits of privatization depend on the precise form of non-benevolence.

Empirics and the Uses of Theory for Policymakers

No serious economic model unambiguously predicts benefits from privatization, so empirics are a useful check. Unfortunately, empirical studies of privatization are less useful than they may appear because the choice of enterprises to privatize is non-random. Profitable firms may be privatized first, whether because these firms are the most likely to restructure, because such restructuring will generate less unemployment and be more politically palatable, because buyers are easier to find for profitable firms, or because the sale of profitable firms generates more money for the government. Alternatively, sometimes it’s the unprofitable firms that may be privatized, for instance through liquidation. Either way, naïve estimates are misleading.

Thus, the best empirical studies are those that use instrumental-variables approaches or other methods that deal with selection bias. A recent review of the literature by Saul Estrin and co-authors, which takes selection bias into account, concludes that in Central and Eastern Europe, privatization has mostly increased productivity; the effect is mostly much smaller in the countries of the Commonwealth of Independent States (which now includes all former Soviet republics except the Baltic states and Georgia). Privatization also has a non-negative effect on profitability in Central and Eastern Europe, the former Soviet countries, and China, though the size and significance of the effect depends on the type of ownership: concentrated domestic private ownership, managerial ownership, and foreign ownership are good.

An earlier review by William Megginson and Jeffry Netter concluded that, for non-transition economies, privatization increases “output, efficiency, profitability, and capital investment spending,” and decreases leverage; the studies “are far less unanimous regarding the impact of privatization on employment levels in privatized firms.” To prevent the political fallout from investors who lost money after buying shares in privatized firms, governments engaging in their first large share-issue privatizations typically “establish (or augment)” an SEC-like agency, establish regulatory bodies for natural monopolies, improve information disclosure rules, and otherwise modernize their corporate governance systems.

It’s immediately apparent from these empirical results that privatization doesn’t exist in a vacuum. The results are different in Russia than in the Czech Republic, or when ownership is foreign vs. domestic, because institutional details matter—many more than are apparent from reading the theoretical models. Those who are familiar with the sociology and reward structure of the economics profession won’t find it surprising that the theoretical models focus on a small number of factors. There’s no glory or prestigious publication in finding that privatization is better or worse depending on the competence or public-spiritedness of public servants. On the one hand, this point is obvious (and a paper making this point will be accused of baking the results into the pie with the assumptions); on the other hand, making a judgment about the competence or public-spiritedness of public servants is politically charged. Nor is there glory or prestigious publication in finding that the desirability of privatization depends on the quality of corporate governance or antitrust law in the private sector, or on the presence of civil service protections, public-employee unions, APA/FOIA-type statutes, or public procurement regulation in the public sector. In principle, any problems along those lines can be fixed by statute (or in some cases by constitutional amendment or by a change in judicial interpretations), so if those are contractual incompletenesses, surely they aren’t interesting ones.

Hence the articles locating the key contractual incompleteness in supposedly more “fundamental” differences between public and private, like the allocation of residual control rights, or the government’s loss of accounting information about private entities. This is of course true and interesting, and the tradeoffs identified are real. But all the boring differences between sectors are relevant as well, even if they’re not “core.” Government employees might indeed be less motivated, or they might be more public-spirited. Corporate governance or antitrust or civil service law or public employee unions might be hard to change, or the optimal fix for those policies might be elusive. Unlike theoretical economists, policymakers live in a world where not every policy is politically on the table.

The same is true of various other perhaps-“accidental” features of the public or private sectors. Private prison firms and public-sector corrections employee unions may have different abilities and incentives to lobby for greater incarceration. Prison privatization may make juries less deferential toward prison officials, since juries are often tougher on corporations than on governments. Soldiers can be imprisoned for disobeying orders; not so for employees of private military companies. And so on.

There’s no shame, therefore, in discussing what isn’t covered in the theoretical models, and I’ll do so often in the rest of the chapter. More generally, it’s clear that privatization causes institutional change and is also affected by preexisting institutions, notably corporate governance and competition policy. All this is the subject of the rest of the paper.

Privatization as Catalyst

In some ways, privatization can be a catalyst for other forms of social change: some have argued that privatization can promote democracy or broaden the distribution of wealth or stock ownership (thus limiting the scope of action for future left-wing governments); privatization can also help develop a country’s financial markets. For all the talk of the optimal sequencing of privatization (antitrust before privatization? corporate governance before privatization?), the feasibility of the program depends on political support. Privatization may in some cases be necessary to produce the dispersed class of owners that will then be a political force for greater institutional reform.

For instance, it hasn’t necessarily been all negative that Russian privatization ended up with ownership concentrated in the hands of a relatively small number of tycoons, colloquially called “oligarchs.” Concentrated ownership is often positive for productivity, since it helps overcome the free-rider problem in the monitoring of management. And the oligarchs have often been a positive force not only for enterprise restructuring and productivity but also for institutional reform in Russia—to say nothing of constituting a big-business lobby and funding source that may have played a role in preventing the return of Communists to power. Of course, the experience has had its strong negatives as well—the oligarchs haven’t been so keen on competitiveness or free trade, the privatization was far from transparent and far from egalitarian, and the political constituency for privatization in Russia has suffered as a result. The important point, though, is that the development of transparent economic institutions in Russia should be judged by what’s feasible, not by the optimal textbook approach.

Next time: Privatization and corporate governance, and the beginning of privatization and competition policy.