Abstract. This paper suggests that stabilizing federalism is like solving a public good provision problem. It reviews results in the public good provision literature that are relevant for federalism, and discusses the implications of these results for the institutional design of federalism.
JEL classification: H110, H770, P160Key words: Intergovernmental relations, government performance, institutional designFederalism now reaches to all corners of the globe, yet in political science we still lack a basic understanding of how to design institutions to maximize the chances that a federation will be successful. We write federal constitutions without knowing what might best be included to maximize the possibility of peace and prosperity. We advocate adoption of the constitutional architecture of successful federations, a policy that will meet with diminishing success as federalism moves from the western developed world to cultures and contexts that are quite different. Some adopt federal constitutions despite our relative ignorance, such as Russia, Bosnia, South Africa, and Iraq. Others hesitate, lacking confidence in their ability to engineer a successful federation. For example, Afghanistan would have been a natural case for federation, given its regions' histories of political independence, but Afghani constitutional planners and their consultants rejected a federal solution, not knowing how to build a federation strong enough to resist the peripheralizing forces of aggressive, competitive provinces. And in Argentina, the federal government is pleading for another restructuring of its debt repayment scheme, despite being unable to find a remedy for the root cause of its fiscal misfortune: provinces who have little incentive to balance their books, despite the ill effects for the national economy.What makes a federation successful, robust? Is it historical accident, cultural influences, or the institutional elements in a constitution that affect its success? If institutions matter, have we fully explored their significance? In this paper I advocate a new approach to the study of federalism that describes federalism as a public good provision problem. The implication of this description is that we are able to enlist our understanding of how public good provision problems may be resolved to the design of federalism. However, these results must be modified to fit the particular challenges that federalism presents. This paper focuses on how a network of institutions, complementary in their functional capacity, makes federations productive. In so doing, the project departs from existing studies, which have focused on the effect of the existence of particular institutions-often considered in isolation-on the longevity of the federation or distributional aspects of it.