Strategic Analysis of Petty Corruption, I: Entrepreneurs and Bureaucrats (Research Seminar, April 8th, 2004) Roy Radner New York University Abstract This is the first of two papers whose objective is to analyze a family of game-theoretic models of "petty corruption" by government officials. Such corruption is widespread, especially (but not only) in developing and transition economies. The models go beyond the previously published studies in the way they describe the structure of bureaucratic "tracks," intermediaries, and supervisors, and the various information flows among the participants. In the basic model of the present paper, entrepreneurs may apply, in sequence, to a "track" of two or more bureaucrats for approval of their projects. Each entrepreneur has a project that has a specific (expected present) value that would be realized if the project were approved. This value is known to the entrepreneur but not to the bureaucrats. The entrepreneur must apply to each bureaucrat in the track in a prescribed order, and her project is approved if and only if every bureaucrat in the track approves it. In principle, approval of a project requires that it be qualified to meet certain requirements, and this qualification imposes a cost on the entrepreneur. Each bureaucrat may demand a bribe as a condition of approval a qualified project (extortion), and a larger bribe may be demanded for the approval of an unqualified project (capture). Each participant learns the previous history of play with a fixed delay. We show that the corresponding sequential game has a continuum of "trigger-strategy equilibria," and we show how this family of equilibria depends on the parameters of the game. In particular, we show that as the discount rate of the bureaucrats approach zero, there is a sequence of trigger-strategy equilibria for which the social loss approaches zero (a "Folk Theorem). We also characterize in more detail the equilibria in this class that minimize the social loss due to the system of bribes, and those that maximize the expected total bribe income of the bureaucrats. Some policy relevant implications of the analysis are that: (1) rotation of the bureaucrats among different offices is likely to increase the minimum social loss by increasing the effective rate at which bureaucrats discount the future; (2) replacing a track of N bureaucrats by a single "window" of one bureaucrat can decrease the minimum social loss but increase the maximum social loss. The second paper will study the effects of introducing intermediaries in the system. Joint work with Ariane Lambert-Mogiliansky (CERAS, Paris) and Mukul K. Majumdar (Cornell University). A paper will be available from Julia Mills early next week.