An incentive-compatibility approach to the problem of monitoring a bureau. (includes appendix)

Although Tullock (1965) and Downs (1967) provide two early insightful looks into the nature of bureaucracies, the first systematic attempt to analyze bureaucracies within a public choice framework is that of Niskanen (1971). In his seminal work Bureaucracies and Representative Government, Niskanen presents and develops a model in which a bureau is able to obtain budgets greater than those desired by its sponsor.

Most works that follow in the tradition of Niskanen analyze bilateral, one-period games between the bureau and its sponsor in a partial equilibrium setting. Such efforts include articles by Miller (1977) and by Moene (1986). Miller views the conflict over the size of the bureau's budget as a bargaining game between the supplier of bureaucratic output (the bureau) and the demander of bureaucratic output (the sponsor); Moene considers the possible outcomes associated with whether the bureau or the sponsor moves first when decisions are made sequentially under conditions of uncertainty. More recently, Chan and Mestelman (1988) analyze the strategic behavior between a bureau and its sponsor in a general equilibrium setting, whereas Carlsen and Haugen (1994) evaluate bureaucratic interaction in multiperiod games where the sponsor and the bureau move alternately. Wintrobe (1997) provides a good review of this literature.

In nearly all such analyses that follow in the tradition of Niskanen, a bureau's ability to obtain budgets that are larger than the sponsor desires stems from three important assumptions regarding the bargaining environment. First, the bureau is a monopolist supplier of the particular public good it produces. Second, due to the nonmarket nature of a bureau's output, only the bureau knows its true cost schedule. Third, the bureau is permitted to make take-it-or-leave-it budget proposals to its sponsor. However, as Mueller (1989, 255) notes, a bureau's ability to present take-it-or-leave-it budget proposals to its sponsor "gives the bureau an extremely strong agenda-setting role, a fact that presumably occurs to the sponsor. The sponsor might reasonably request that the bureau state the costs of a range of outputs from which the sponsor then chooses."

Drawing from the analyses of Breton and Wintrobe (1975) and Bendor, Taylor, and Van Gaalen (1985), Mueller (1989) develops a model in which the sponsor chooses a level of output based on a bureau's announced price per unit of output. In Mueller's setup, the bureau maximizes its budget subject to the constraint that its total budget must be at least as large as its total cost. So, the budget-maximizing bureau presents the sponsor with a price per unit at which it will supply bureaucratic output. Based on the bureau's reported price, the sponsor selects a level of output for the bureau in accordance with the sponsor's demand schedule for the bureau's output. However, since the sponsor's demand schedule is known to the bureau and to the sponsor, the bureau will overstate its price report whenever doing so will result in a larger budget. Hence, Mueller concludes that in this new bargaining environment, the bureau will be able to expand its budget beyond its sponsor's preferred level only when the sponsor's demand for the bureau's output is own-price inelastic (p. 256).

Following Mueller (1989), this article replaces the assumption that the bureau is permitted to make take-it-or-leave-it budget proposals to the sponsor. Beyond Mueller, however, I permit the sponsor of the monopoly bureau to regulate the bureau as a monopolist. That is, in addition to selecting the level of output, the sponsor selects a price per unit of output that it will pay the bureau based on a report that the bureau makes concerning its marginal cost per unit of output. To analyze such a situation, …