How Much Can Taxation Alleviate Temptation and Self-Control Problems?

We develop a quantitative dynamic general-equilibrium model where agents have preferences featuring temptation and self-control problems, and we apply it in order to understand to what extent standard investment/savings subsidies can improve welfare. The dynamic model of preferences builds on the Gul-Pesendorfer setting and uses a “quasigeometric” formulation for temptation, the strength of which is modeled with a separate parameter. When this parameter is infinity, consumers always succumb to the temptation, and our formulation reduces to the model of multiple selves studied elsewhere. We embed the consumer in the typical neoclassical growth setting used for macroeconomic analysis and demonstrate how steady states can be analyzed and how equilibria can be solved. We also propose a way of calibrating the model. Our preference-based setting gives unequivocal guidance for policy analysis. In our quantitative application, we show that when preferences are such that self-control is limited and consumers instead mainly succumb to the temptation of overconsumption, policy can play an important role. ∗This paper performs the quantitative analysis formerly studied in a paper entitled “Temptation and Taxation”; that paper, thoroughly revised, now focuses solely on theoretical analysis of optimal taxation with temptation and self-control problems. We thank Daron Acemoglu, Larry Epstein, Xavier Gabaix, Wolfgang Pesendorfer, Iván Werning, Justin Wolfers, and three anonymous referees for important suggestions and comments. Krusell, Kuruşçu, and Smith are at the Institute for International Economic Studies (Stockholm), the University of Texas at Austin, and Yale University, respectively. We thank seminar participants at Arizona State University, Brown University, Carnegie Mellon University, Duke University, Harvard University, Institute for International Economic Studies, Johns Hopkins University, Federal Reserve Bank of Kansas City, Massachusetts Institute of Technology, Université de Montréal, New York University, Princeton University, Stanford University, University of Virginia, Yale University, and the 2001 North American Summer Meetings of the Econometric Society and Society for Economic Dynamics conferences for helpful comments. Krusell and Smith thank the National Science Foundation for financial support.