Procrastination, partial naivete, and behavioral welfare analysis

This paper has a dual purpose. First, I present a new modeling of partial naivete, and apply this to the analysis of procrastination. The decision maker is assumed to have stationary behavior and to be partially naive in the sense of perceiving that his current preferences may persist in the future. The behavioral implications of such partial naivete differ from those of related literature. Second, I suggest a general principle for welfare analysis in multi-self settings through a new application of Pareto-dominance, which is motivated by the existence of time-inconsistency and coincides with dominance relations used elsewhere in game theory. In the case of procrastination, it leads to a clear welfare conclusion: Being partially naive reduces welfare.