Institutionalism as a Methodology

Institutionalism as a Methodology

By Keith Krehbiel, Daniel Diermeier
2001Working Paper No. 1699

We provide a definition of institutionalism and a schematic account that distinguishes between institutional theories (in which institutions are exogenous) and theories of institutions (in which some, but necessarily not all, institutions are endogenous). Our primary argument is that institutionalism in the contemporary context is better characterized as a method than as a body of substantive work motivated by the so-called chaos problem. Secondary arguments include the following. (1) While it is important to differentiate sharply between institutions and behavior, institutionalism as a method is neither inimical to behavioralism nor devoid of behavior. (2) When making the challenging transition from developing institutional theories to developing theories of institutions, it is essential to hold behavioral axioms fixed and to choose a form of equilibrium that exists for the class of games studied. (3) For most research programs today, a form of Nash Equilibrium has the requisite properties. The core, and structure-induced equilibria (SIE) that rely on the core, often lack the requisite properties.