How Different Can Strategic Models Be

Abstract This paper shows that the dynamic implications of strategic behavior can differ remarkably from those obtained under first-best or "competitive" solutions. We employ the framework of common-property resource models, whose first-best version is the familiar neoclassical growth paradigm. As is well known, Markovian first-best solutions always exist here and imply very regular dynamic behavior (viz., state trajectories are monotone, and converge to the "golden rule"). In the corresponding dynamic game, however, we find that the analog of these solutions, Markov Perfect Equilibria, need not exist; when they do exist, their trajectories could be very complex; and, even if monotone trajectories result, their limit points could be very non-intuitive. Journal of Economic Literature Classification Numbers: C72, C73.