Showing papers by "Jean Tirole published in 1989"

Chapter 2 The theory of the firm

Abstract: Publisher Summary The theory of the firm has long posed a problem for economists. This chapter discusses the analytical models of the firm that go beyond the black-box conception of a production function. The firm is seen as a contract among a multitude of parties. The main hypothesis is that contractual designs, both implicit and explicit, are created to minimize transaction costs between specialized factors of production. This follows Coase's original hypothesis that institutions serve the purpose of facilitating exchange and can best be understood as optimal accommodations to contractual constraints rather than production constraints. There are three problems that need attention. A first step is to develop and apply techniques that deal with nonstandard problems, such as incomplete contracts, bounded rationality, and multi-lateral contracting. The second step ought to integrate observations from neighboring fields, such as sociology and psychology, in a consistent way into the theoretical apparatus. The third step will be to increase the evidence/theory ratio, which is currently very low in this field.

Noncooperative game theory for industrial organization : an introduction and overview

Abstract: Publisher Summary Noncooperative game theory is a way of modelling and analyzing situations in which each player's optimal decisions depend on his beliefs or expectations about the play of his opponents. Game-theoretic methodology has caused deep and wide-reaching changes in the way that practitioners think about key issues in oligopoly theory, much as the idea of rational expectations has revolutionized the study of macroeconomics. This chapter reviews those aspects of the noncooperative game theory that are most commonly used by industrial organization economists and describes the most important applications of the theory. There are two equivalent ways of formulating a game. The first is the extensive form. An extensive form specifies (1) the order of play, (2) the choices available to a player whenever it is his turn to move, (3) the information a player have at each of these turns, (4) the payoffs to each player as a function of the moves selected, and (5) the probability distributions for moves by “Nature.”

Provision of quality and power of incentive schemes in regulated industries

Abstract: We study the incentives of a regulated monopoly to supply quality. For an experience good, the current level of sales yields no information about quality and the cost reimbursement rule is the only instrument to achieve the conflicting goals of provision of quality and cost reduction. A high concern for quality moves optimal contracts toward cost-plus contracts; and an increase in the discount factor raises the power of incentive schemes. In contrast, for a search good, direct sales incentives can be provided to supply quality; whether a high quality concern drives optimal contracts toward cost-plus or fixed-price contracts then depends on whether quantity and quality are net substitutes or net complements. 1 . Introduction . An unregulated monopolist may have two incentives to provide quality: the "sales incentive" and the "reputation incentive." When quality is observed by consumers before purchasing (search good) , a reduction in quality reduces sales, and thus revenue as the monopoly price exceeds marginal cost. In contrast, when quality is observed by consumers only after purchasing (experience good) , the monopolist has no incentive to supply quality unless consumers may repeat their purchase in the future. The provision of quality is then linked with the monopolist's desire to keep its reputation and preserve future profits. In this paper, we investigate whether similar incentives to provide quality exist in a regulated environment. Before doing so, it is useful to distinguish between observable and verifiable quality. Quality is usually observable by consumers either before or after consumption. Quality is furthermore verifiable if its level can be (costlessly) described ex-ante in a contract and ascertained ex-post by a court. When quality is verifiable, the regulator can impose a quality target to the regulated firm or more generally reward or punish the firm directly as a function of the level of quality. For instance a regulatory commission may dictate the heating value of gas or may punish an electric utility on the basis of the number and intensity of outages. Formally, the regulation of verifiable quality is analogous to the regulation of a multiproduct firm, as the level of quality on a given product may be treated as the quantity of another, fictitious product. This paper will be concerned with observable but unverifiable quality. The effectiveness of a new weapons system, the quality of broadcasting by a regulated television station, the level of services enjoyed by a railroad See Sappington [1983] and Laffont-Tirole [1988] for information-based theories of the regulation of a multiproduct firm with and without cost regulation. passenger or the probability of a core melt-down at a nuclear plant are hard to quantify and include in a formal contract. As Kahn [1988, p. 22] argues: But it is far more true of quality of service than of price that the primary responsibility remains with the supplying company instead of with the regulatory agency, and that the agencies, in turn, have devoted much more attention to the latter than to the former. The reasons for this are fairly clear. Service standards are often much more difficult to specify by the promulgation of rules. When quality is unverifiable , the regulator must recreate the incentives of an unregulated firm to provide quality without throwing away the benefits of regulation. First, it must reward the regulated firm on the basis of sales. Second, the threat of nonrenewal of the regulatory license, of second sourcing or of deregulation makes the regulated firm concerned about its reputation as supplier of quality. The focus of our analysis is the relationship between quality concern and power of optimal incentive schemes. An incentive scheme is high(low-) powered if the firm bears a high (low) fraction of its realized costs. Thus a fixed-price contract is very high-powered, and a cost-plus contract is very low-powered. The link between quality and the power of incentive schemes has been much discussed. For instance, there has been a concern that "incentive regulation" (understand: high-powered incentive schemes) conflicts with the safe operation of nuclear power plants by forcing management to hurry work, take shortcuts and delay safety investments. There have been accounts that the switch to a high-powered incentive scheme for British Telecom (price caps) after its privatization produced a poor record on the quality front (Vickers -Yarrow [1988, p. 228]). 2 Similarly, Kahn [1988, I, p. 24] contends 2 It is not surprising that the dissatisfaction with the quality performance subsequently led to the costly development and monitoring of quality indices to be included in the incentive schemes. that, under cost-of -service regulation (a very low-powered Incentive scheme), In the matter of quality "far more than in the matter of price, the interest of the monopolist on the one hand and the consumer on the other are more nearly coincident than in conflict." Kahn's intuition is that the regulated monopolist does not suffer from incurring monetary costs to enhance quality because these costs are paid by consumers through direct charges. This intuition is incomplete. First, some components of quality Involve non monetary costs. Second, and more importantly, under pure cost-of -service regulation, the regulated firm does not gain from providing costly services either so that a low perceived cost of supplying quality does not imply a high 3 incentive to supply quality. Last, in the context of military procurement, Scherer [1964, pp. 165-166] has suggested that There is reason to believe that the use of fixed-price contracts would not greatly reduce the emphasis placed on quality in weapons development projects, although it might affect certain marginal tradeoff decisions with only a minor expected impact on future sales. To give formal arguments to assess the relevance of these perceptions, we introduce two related natural monopoly models of an experience and of a search good. Whether the power of regulatory contracts decreases when quality becomes more desirable depends crucially on whether contractual incentives can be based on sales (on top of cost) or not, i.e., on whether the regulated firm 3 In practice, one does not observe pure cost-of-service regulation. Due to the regulatory lag, the regulated firm is, like an unregulated monopolist, the residual claimant for the revenue it generates and costs It incurs between rate reviews (the differences being that the prices are fixed and that the regulated monopolist is concerned about the ratchet effect) ; thus actual cost-of-service incentive schemes are not as low-powered as one might believe. We will not try to study (variants of) cost-of-service regulation, but will rather focus on optimal regulation. See Joskow and Rose [1987] for empirical evidence on the level of services under cost-of-service regulation. supplies a search or an experience good. In our two-period model of an experience good, the regulator purchases a fixed amount from the regulated firm. Because quality is ex-ante unverifiable , the regulator has no alternative than to accept the product. The supplier's incentive to provide quality is then the reputation incentive, i.e., the possibility of losing future sales. In contrast, our static search good model has the firm sell to consumers who observe quality before purchasing. The former model is best thought of as a procurement model, and the latter as a regulation model, although other interpretations are possible (in particular, some regulated products are experience goods) . To separate issues in this paper we choose cost functions for which the incentive-pricing dichotomy (Laffont-Tirole [1988]) holds. Pricing is not used to extract the rent due to asymmetric information. In the case of an experience pood , we argue that incentives to supply quality and those to reduce cost are inherently in conflict. The regulator has a single instrument -the cost reimbursement rule -to provide both types of incentives. High-powered incentives schemes induce cost reduction but increase the firm's perceived cost of providing quality. This crowding-out effect implies that the more important quality is, the lower the power of an optimal incentive scheme. We also show that when the firm becomes more concerned about the future, its perceived cost of supplying quality decreases, which induces the regulator to offer more powerful incentive schemes. We thus find Scherer's suggestion quite perceptive. In the case of a search good , the crowding-out effect is latent but has no influence on the power of incentive schemes. In our model, the regulator can separate the two incentive problems because it has two instruments: cost reimbursement rule and sales incentives. The incentive to provide quality is provided through a reward based on a quality index, which is the level of sales corrected by the price charged by the firm. As in the case of an experience good, the cost-reducing activity is encouraged through the cost reimbursement rule, which is now freed from the concern of providing the right quality incentives. This dichotomy does not, however, imply that an increase in the desirability of quality has no effect on the power of incentive schemes; it has an indirect effect because higher services may increase or decrease the optimal level of output, which in turn changes the value of reducing marginal cost and thus affects the regulator's arbitrage between incentives and rent extraction. While there exists a vast literature on the provision of quality by an 4 unregulated monopoly, surprisingly little theoretical research has been devoted to this issue in a regulated environment. Besanko et al . [1987] assume that the monopolist offers a range of verifiable qualities to discriminate among consumers with different tastes for quality (a la Mussa-Rosen [1978]) and investigate the effect of imposing minimum qualit