| In democratic societies, the one political certainty, more or less, is change. With such turnover, one might assume, public policies will be volatile, changing as often as political regimes do. But policies do not always change as regimes do in representative democracies. In this dissertation, a collection of three essays, I address the issue of when political officials will be able and, perhaps more importantly, when they will want to make durable public policy.;The first essay is a theoretical critique of the argument that agency structures that contain protective, insulating mechanisms to prevent future sabotage, will be used in uncertain electoral environments. Analyzing two repeated games, I derive two results. First, in parliamentary systems, cooperation is most sustainable when uncertainty is maximal; second, in presidential systems in which statutory insulation mechanisms are easier to utilize, it is electorally weak groups that will insulate their policies when given the opportunity.;The second essay tests this theory on the case of a particular budgetary mechanism: adoption of line-item veto amendments to state constitutions. Analyzing data from the fifty American states, I find that the most convincing explanation as to why legislative majorities would cede institutional power to the governor is insulation. In particular, conservative majorities who feel they are likely to be out of power are the most likely to propose these mechanisms.;The final essay is a more microscopic study of a mechanism public officials can utilize to turn previous public policy to their own political ends: the assignment of policy jurisdiction to agencies. Using a multi-task agency model, I derive the conditions under which tasks will be assigned to separate agencies and when they will be centralized. Perhaps the most interesting case is that when faced with an existing task that an official does not want implemented and one that she does, if the costs are not prohibitive and high-powered incentives can be provided for the task she prefers, the official will centralize. The reason is that two purposes can be achieved in this case: implementation of the preferred task and distraction of attention away from the task that she dislikes. |
