Life-cycle Models, Economic Puzzles and Temptation Preferences

This paper focuses on the difficulty of standard life-cycle models to predict the behavior observed in the reality, most noticeably the excess sensitivity of consumption to income, the under-provision for old-age consumption, the limited participation in the financial market, and the lack of asset decumulation after retirement. It shows that allowing for preference reversals, as it is the case in the “quasi-hyperbolic discounting” and “temptation” models, may contribute to explain jointly these economic puzzles. A life-cycle model based on temptation preferences, in particular, is attractive as it preserves time consistency and can be solved with standard dynamic programming techniques.