Designing Public Organizations and Institutions: Essays on Coordination and Incentives
Abstract: This thesis consists of an introduction and four self-contained chapters that address issues of how organizations and institutions should be designed in order to provide adequate incentives and enable coordination. The first essay uses a multi-task principal-agent model to examine two reasons why coordination problems are common when public sector organizations share responsibilities: the incentives to coordinate resource allocation and the difficulties of measuring performance. When targets are set individually, the resulting incentives may induce inefficient resource allocations, even if agents' motivation is fully in line with the principal's interests, and performance measures are perfect. Introducing shared targets may improve the incentives to coordinate, especially when agents are highly motivated by activities for which responsibility is not shared, performance measures overestimate the value of these activities, and the imprecision of performance measures is low. Simulations indicate that the one-shot model is a good approximation in the short run when one parameter is initially unknown, but imprecise measures yield substantial deviations from equilibrium when three parameters have to be learned. The second essay develops a model of costly communication with the weakest-link game as a basis, and boundedly rational agents that choose myopic best replies, have limited information processing capabilities, and may occasionally experiment or make mistakes. Communication may increase efficiency, but as agents face a trade-off between lowering the strategic uncertainty for the group and the costs of communication, the least efficient state is often the unique stochastically stable state. Making communication mandatory on the other hand induces efficient coordination, whereas letting a team leader handle the communication increases efficiency when the leader expects others to follow and has enough authority over the group. Simulations show that the stochastically stable state is also overrepresented in the short run, especially if groups are large. The third essay examines theoretically and empirically the relationship between budget institutions and fiscal performance in local governments. The model suggests that conflicts of interest over fiscal discipline between the central and the local levels imply that the central level needs institutions that curb the bargaining power and align the incentives of the local levels. A survey covering 265 out of 290 Swedish municipalities is used to collect information on conflicts of interests and budget institutions. The regression results indicate that interactions between institutions and conflicts of interest are important, as the estimated correlations depend on the strength of conflicts. Centralization of the budget process, a credible threat of replacement of managers following systematic deficits, and surplus carry-over rules all appear beneficial to net revenues, but only in municipalities that report substantial conflicts of interest. For municipalities where the conflict is small, a deficit carry-over rule is positively correlated to net revenues. The fourth essay examines a case in which the Swedish central government provided conditional grants to 36 financially troubled municipalities. The synthetic control method is used to identify suitable comparison units for each of the 36 municipalities. The estimates of fixed effects regressions on the resulting sample indicate that for most of the admitted municipalities, costs seem to be largely unaffected by the program. However, a non-negligible share is able to hold back costs more than expected, and the development of net revenues is favorable for the group as a whole. Thus, participation in a conditional bailout program need not erode fiscal discipline, and may even induce a greater concern for fiscal discipline.
CLICK HERE TO DOWNLOAD THE WHOLE DISSERTATION. (in PDF format)