Three Factors Limiting the Power of Incentives

We have already observed that asymmetric information allows regulated firms to extract large rents under high-powered incentive schemes. This possibility in turn motivates the use of cost- or earnings-sharing schemes. This section investigates three factors that make sharing by consumers or by the government particularly desirable.

Quality Concerns

A well-known drawback of high-powered schemes is that they make it very costly to firms to supply quality. The provision of quality raises cost and is therefore borne entirely by the firm if the latter is residual claimant for its cost performance. The firm may, therefore, decide to skimp on quality if quality is not minutely specified in the regulatory contract. In contrast, low-powered schemes, by failing to make the firm accountable for its cost performance, make it very cheap to supply quality.

This point is well known. Indeed, the U.S. Department of Defense has often invoked it to motivate the use of cost-plus contracts in contexts in which quality is a sensitive issue and its specifications are hard to pin down exactly in the initial procurement contract with the defense contractor. Similarly, the argument has been made several times that the introduction of incentive regulation for power companies conflicts with the safe operation of nuclear power plants. As a last illustration, quality started deteriorating shortly after British Telecom's 1984 privatization and design of more powerful incentives in the form of a price cap, and quality standards and verification mechanisms had to be set up as a consequence.

There is a natural complementarity between powerful incentive schemes and a close monitoring of quality standards. If the regulator is unable to specify the standards or to monitor compliance with these standards, then there may be no other choice than the adoption of low-powered incentives.

Regulatory Commitment

Regulatory contracts are generally much shorter than the relationship between the regulatory authority and the regulated firm. They do not exceed five years for price cap regulation and are usually shorter under other forms of regulation.

Let us for the moment assume that the official length of the contract is respected, so that the regulatory review will occur at the expiration of the current regulatory contract. And consider a high-powered incentive scheme covering the length of the regulatory contract. It is clear that even though the firm is formally residual claimant for its cost savings, an effort to reduce cost by $1 is not rewarded $1 overall. A lower cost will convince the regulatory authority of a higher efficiency and will make it more demanding in the regulatory contract designed for the firm at the next regulatory review. So while a $1 cost reduction yields $1 to the firm in the short run, it also entails a long-term penalty in the form of higher performance requirements at the next review. This is the well-known ratchet effect. The ratchet effect imposes a bound on the incentives that can be provided even by formally high-powered incentives.

In practice, the actual length of the regulatory contract may be shorter than its formal length. That is, the contract may be renegotiated before the next regulatory review. On the one hand, the regulator, usually under political pressure, may be tempted to force the firm to renegotiate before the end of the contract when the latter makes substantial profits. Such renegotiation exacerbates the ratchet effect and makes the firm even more cautious about taking full advantage of (formally) high-powered incentives to reduce cost. On the other hand, the firm may force the regulator to renegotiate midway and offer more favorable terms if the initial contract proves unprofitable and makes credible the threat of bankruptcy or at least that of forgoing investments that the regulator deems necessary. The firm is then said to face a "soft budget constraint," since it is rescued by the regulator despite a commitment not to intervene before the next regulatory review. A high-powered incentive scheme raises the likelihood of the occurrence of either situation (very high or very low profits) and therefore of contract renegotiation. Because in both cases contract renegotiation ex post rewards the firm's inefficiency (punishes its cost-reducing effort), early regulatory reviews further decrease the real power of formally high-powered incentive schemes.

Regulatory Capture

The debate on incentive regulation focuses on the agency problem between the regulator and the regulated firm. There is in practice a second agency problem, namely, that between the general public and the regulator. And, as we will see, the two agency problems may interact in an interesting way.

Understanding the second agency problem requires asking why there is a regulator in the first place. Quite clearly, the regulator is an informational intermediary. The members of the collectivity face a collective action problem. They individually have no incentive to make the very heavy investment required for a good understanding of both the technology and the economics of telecommunications. This free-riding problem is somewhat alleviated by the election of political representatives, who themselves are informational intermediaries. But the free-riding problem subsists, in an attenuated but still substantial way, at the congressional level. It is resolved only (and still imperfectly) by the creation of specialized congressional committees and bureaucratic agencies.

In a nutshell, congressional committees and regulatory agencies are "informational intermediaries," "delegated monitors," or "supervisors." Their role is to fill some of the informational gap between the collectivity and the industry. But this informational expertise is precisely what provides the delegated monitors with discretionary power and what creates the second agency problem. This second agency problem can be divided into two parts. First, the informational intermediaries may not have sufficient incentives to collect the information about the technology and the economics of the industry, in the same way the regulated firms may not have enough incentives to control their cost. Second, when endowed with a given amount of information, the informational intermediaries may not make use of this information in the direction that would benefit the collectivity. That is, they may abuse their discretion.

Here, we are interested in this potential for abuse, and in particular in the possibility that the informational intermediaries are captured by interest groups. Political scientists (Montesquieu, the American Federalists, Marx, Bernstein, . . . ) and economists (the Chicago school with Stigler, Posner, Becker, and Peltzman, and the Virginia school with Buchanan and Tullock) have long recognized the threat posed by interest-group politics for the efficiency of economic regulation.

For an economist interested in regulatory design and reform, the interesting question, though, is not the existence of a threat of capture, but rather what is to be done about it. To approach this question, we must, as we discussed, first explain why regulators have discretion and so may collude with interest groups. Their role as informational intermediaries seems crucial in this respect. It is precisely because their principal (the full Congress or the collectivity as a whole, depending on the interpretation) is uncertain as to the ranking of alternative policies that informational intermediaries can get away with policies that favor specific interest groups to the detriment of the collectivity. Furthermore, a delegated monitoring approach not only explains why there is scope for regulatory capture, but also suggests policies that will reduce the likelihood of capture, as we now illustrate in the context of incentive regulation.

Recall that high-powered incentive schemes are associated with high rents, and therefore with high stakes for the regulated firm. High-powered incentives thus generate for the regulated firms a large benefit from capturing their regulators. Using a procurement example, and in the presence of uncertainty about the firm's cost, the regulator has a lot of discretion when designing a fixed-price contract, in that the level of the fixed price is highly subjective and affects the firm's welfare substantially. In particular, when the regulator obtains information that the firm has a low cost, a piece of information that if disclosed would permit the elimination of the firm's potential rent, the regulator may be lenient with the firm and "forget" the information he acquired. In contrast, a cost-plus contract is more mechanical, and regulatory decisions have a much lower impact on the firm's welfare. A cost-plus contract is therefore less sensitive than a fixed-price contract to the risk of regulatory capture by the regulated firm (at least when good accounting procedures are in place).

The reader may wonder whether the influence of the regulated firm might not be offset by that of interest groups with opposite interests. Indeed, taxpayers in a procurement context and consumers in a regulatory context are hurt when the firm enjoys a rent, since they then have to pay higher taxes and prices for the services, respectively. It can be shown, however, that, even if they solve their collective action problem and are properly organized, taxpayers or consumers do not form an effective counterpower to the industry when they lack the information about policy rankings. But this lack of information is precisely what motivated the use of an informational intermediary in the first place. Put differently, collusion may occur among members of a "nexus of information," that is, between economic agents sharing information; it is for this reason that collusion is an important matter in situations of delegated monitoring.

The delegated-monitoring view of regulatory agencies, therefore, leads to the following implication for incentive regulation: low-powered incentive schemes, because they are less discretionary (make less use of the regulator's private information), are more robust to regulatory capture; it is therefore advisable to lower the power of incentive schemes when the threat of capture by the regulated firm is serious. Or put differently, the adoption of high-powered schemes must go hand-in-hand with the existence of political and bureaucratic institutions that alleviate the capture problem.

More generally, the delegated monitoring view identifies four types of policies with a potential to curb regulatory capture. The first two aim at reducing the gains from collusion; the last two make it more difficult for the parties to collude to reap those gains.

Reduction of Regulatory Discretion The use of low-powered incentive schemes is an illustration of a more general principle: Substantial regulatory discretionthat is, a high sensitivity of regulatory decisions to the regulatory agency's assessment of the regulated firm's cost and demand environmentcreates high stakes for the interest groups and therefore a concern about regulatory capture. The potential for regulatory capture, therefore, reduces the use that is made of the regulators' private information and thereby creates a more bureaucratic environment. We will later offer other examples of reduced stakes and less discretionary policies.

Making Regulators Accountable by Offering Them a Stake Another way of fighting capture is to make regulators internalize at least partly the welfare of other groups so as to induce them not to favor a specific interest group. Let us here provide some analogy. A lead investment bank, leveraged buyout specialist, or venture capitalist brings investors to buy a new debt or equity issue; it thereby acts as a delegated monitor. It specializes in acquiring information about the issuer not held by the other investors. But, of course, it may also collude with the issuer, possibly getting some kickback on the side. To reduce the risk of collusion (as well as to encourage the collection of information), investors require the lead investment bank, leveraged buyout specialist, or venture capitalist to take a stake in the issue. The incentives of the delegated monitor and the principal (the investors) are then better aligned. Stake taking by delegated monitors is actually a widespread institution in situations in which the principal's welfare is easily measurable.

Alas, what works well in financial environments is less effective in a regulatory context. Stake taking by the regulator would mean that the regulator's salary would be made contingent on the amount of money paid by taxpayers (in a procurement context), on the level and structure of consumer prices (in a regulatory context), and possibly on the quality of service (in both contexts). It is clear that such performance measures either are hard to specify exactly ex ante, or when they can be easily described, as in the case of the taxpayers' bill, they lack a natural benchmark (in contrast, the benchmark can be the market rate of return in the financial investment analogy). So, in general, it is difficult to prevent capture by providing regulators with monetary incentives that would align their interests with those of taxpayers or consumers.

Making Collusion More Difficult We have not yet discussed the process of capture. In general, the regulator may be willing to do a favor to an interest group if a quid pro quo is available. This quid pro quo in practice takes many forms: direct monetary transfer (as in most cases of corruption), campaign contributions for politicians, lack of complaint about the regulatory activity, future employment as a top executive or consultant (the revolving door), friendship, entertainment expenditures, and so forth. The nature of the bribe varies greatly with the situation and with the personality of the parties involved in the capture process.

One may try to reduce capture by making these quid pro quos more costly to the parties. But such policies have also a cost. For example, the prohibition of the revolving door may make it difficult to find qualified regulators if industry executives envision that they will not be able to return to the private sector when they quit the regulatory agency. It is only recently that economists have started developing formal models of the impact of regulatory institutions on the ease with which transfers can respond to stakes and facilitate capture.

Reduction of the Asymmetry of Information between Regulators and Their Principal To the extent that capture is related to the regulators' informational superiority over their principal (full Congress, collectivity), a reduction in this informational asymmetry reduces the scope for capture. Obviously, one should not expect too much from such policies, since delegated monitoring is the heart of the problem. But some steps can be undertaken that reduce the manipulation of information by regulators. In particular, and following the Anglo-Saxon tradition of transparency, one can force regulators to adopt a very open process: open regulatory hearings, use of consultative documents, independent appeal procedures, written and detailed explanation of decisions, and so forth. Transparency gives the regulators' principal access to other sources of opinion (interest groups such as customers and competitors), to data, and to the regulators' reasoning. It alleviates, although it does not eliminate, informational asymmetries between the delegated monitor and its principal.

Economic Principles: Performance-Based Regulation

The Rationale for Performance-Based Regulation

One of the motivations for the design of new regulatory schemes is the perception of a poor cost performance of regulated firms. Depending on the situation, this poor performance may take several forms: insufficient internal control, labor hoarding, incompetent management, undue perks, lack of innovation, imprudent investments, and so forth. The suboptimal performance is generally attributed to a lack of incentives. Deregulated firms have traditionally operated under a regime in which cost increases were automatically passed through to consumers in the form of higher charges for the operators' services. Recent reforms have been trying to make firms accountable for a substantial fraction of their costs.

1.1 A Procurement Example

To get a good grasp at the incentive issues, let us abstract from pricing considerations in a first step. Suppose that a government wants to procure some (fixed amount of) public good. Only one firm has the know-how or capacity to supply this public good. Unfortunately, the government has imperfect information about the cost that will be incurred by the firm. Its expected cost depends on both exogenous and endogenous variables.

Exogenous variables refer to the technological requirements that the firm will be facing in producing the public good and to the firm's opportunity costthat is, the profit forgone by the firm when producing the public good instead of undertaking an alternative project for another principal. To the extent that the firm is better informed about these exogenous factors when contracting with the government, as is likely to be the case, the government faces an adverse selection problem. Roughly speaking, the government does not know whether a low payment will suffice to convince the firm to undertake the project or whether it will have to "pay the high price." Needless to say, the firm will not be eager to reveal that its production cost is low even if this is the case, since it is in its interest to persuade the government that only a high price will do.

Endogenous variables refer to those postcontractual decisions taken by the firm that, together with the exogenous variables (and some ex post uncertainty), determine the firm's final production cost and that cannot be contracted upon because they are not verifiable by a court of law. These discretionary choices by the firm create the scope for poor performance that we alluded to earlier. That is, the government faces moral hazard on the firm's side. Only if the firm is made accountable for a large fraction of its realized cost will it not abuse this discretion.

1.2 The Basic Trade-Off between Incentives and Rent Extraction

Suppose that the government wants to buy the public good but would like to pay as little as possible. Its instrument is the payment to the firm, which can be made contingent on the firm's ex post realized cost (provided, of course, that the firm is willing to provide the public good under the conditions offered by the government). That is, the government offers a cost reimbursement rule specifying the payment that will be made to the firm for each realization of the cost.

A key concept is that of the power of the incentive scheme. A high-powered incentive scheme is one in which the firm bears a high fraction of its cost at the margin. That is, when the firm raises its cost by $1, its net payment (that is, the payment it receives over and beyond the reimbursement by the government of the realized cost) is reduced by an amount close to $1, or, equivalently, its gross payment (that is, the payment it receives if, by accounting convention, it pays the cost itself) hardly moves with the realized cost at the margin. In a procurement context, a fixed-price contract, in which the contractor receives a fixed gross payment, is the prototypical high-powered incentive scheme, since the firm is made fully accountable for its cost savings. In contrast, a low-powered incentive scheme is one in which a $1 increase in the firm's realized cost translates into about a $1 payment by the government, and so hardly affects the firm's profit. In particular, in a cost-plus contract, the firm's cost is reimbursed, and so the firm is not made accountable for its cost savings or overruns.

If the government were fully informed about the firm's cost, that is, if there were no adverse selection, the government's design of a contract for the firm would be a simple matter: by offering a fixed-price contract, the government would give perfect incentives to the firm. The latter would fully internalize its cost savings and therefore would exert the socially optimal level of effort to reduce costs. Besides, the government would not have to worry about paying too high a price. Being well-informed about the firm's technology, it could perfectly choose the fixed price so as to leave the firm with no rent; that is, it could select the lowest price that is consistent with the firm's being willing to agree to produce the public good. We conclude that, in the absence of adverse selection, the government would offer a high-powered incentive scheme.

In the presence of incomplete information about the firm's technology or opportunity cost, though, the government faces a trade-off between giving good incentives to the firm and capturing its potential rent. Recall that proper incentives for effort are created by a fixed-price contract (or more generally by a high-powered incentive scheme). But a contract that yields $1 to the firm each time the firm endogenously reduces its cost by $1 also gives it $1 whenever its cost is lower by $1 for exogenous reasons; that is, a firm is residual claimant also for cost factors that are outside its control. This fact generates substantial rents. In contrast, a cost-plus contract (or, more generally, a low-powered incentive scheme), while providing poor incentives to keep cost down, is efficient at capturing the firm's potential rent. Indeed, the firm does not benefit when it is lucky and its cost is exogenously reduced by $1, since the cost is fully borne by the government.

To illustrate this adverse selection problem and the impact of the power of the incentive scheme, suppose that there is no moral hazard problemthat is, that the firm's cost is exogenously determined. This cost can be either 5 or 10. If the government is constrained to offering a fixed-price contract, and if the public good is socially sufficiently valuable so that the government must supply it, then the government has no choice but offering 10 to the firm. While this offer ensures that the firm is willing to produce the public good, it also leaves a rent equal to 5 if the firm has a low cost. In contrast, a cost-plus contract pays only what is needed to let the firm break even. To be certain, the realized cost is then, say, 8 or 13 (depending on the firm's intrinsic efficiency), but the payment matches the cost.

We thus conclude that there is a basic trade-off between incentives, which call for a high-powered incentive scheme, and rent extraction, which requires, in the presence of adverse selection, low-powered incentives.
This simple implication of the theory is too often forgotten. High-powered incentive schemes (price caps in a regulatory contract) have repeatedly been hailed as a breakthrough in the economics of regulation. While they indeed deliver a good cost performance, they are also likely to leave substantial profits to the firms' owners. There is no magic cure. Those who support or just accept the use of high-powered incentive schemes should be ready to refrain from forcing contract renegotiation when they observe large profits. Experience (the 1995 early review of the U.K. regional electricity companies is a case in point) shows that this point is not always understood.

1.3 One Size Does Not Fit All

A second insight, "One size does not fit all," can be gleaned from our procurement example. The same contract in general does not fit all types of supplier. The point is more subtle than the incentives-rent-extraction trade-off, but it is equally important. In the presence of asymmetric information between the firm and the government, the key issue for the latter is how to screen the firm's type through its contractual choice.

Recall that the difficulty encountered in extracting the firm's rent resides in the firm's ability to pretend to face a high cost when it actually has a low cost. The government can, however, reduce the gain enjoyed by a low-cost firm by making this strategy unappealing. To build a simple example, suppose that the firm's intrinsic cost can be "high" or "low" (that is, the exogenous cost parameter can take one of two values). And suppose that the government offers a menu of two contracts, a fixed-price and a cost-plus contract, from which the firm selects one (or none). The fixed-price contract is designed so as to let the firm just break even (that is, enjoy no rent) when it has a low cost. Yet the firm when it has a low cost is not tempted to choose the cost-plus contract, since its cost efficiency yields no gain under a contract that reimburses all costs anyway. And it is easy to see that the firm strictly prefers the cost-plus contract when its cost is high because it would lose money under the fixed contract, which, as we have seen, is designed so as to let the more efficient type just break even.

This menu of a fixed-price contract for a low-cost type and of a cost-plus contract for a high-cost type allows the government to ensure participation by the high-cost type while not giving rents away to the low-cost type. This "perfect screening" is of course purchased at a high expense: The high-cost type faces no incentive whatsoever to control cost. But it illustrates nicely the scope for screening