Abstract
Dynamic consistency is a key behavioral property in dynamic economic models, making it possible to use tractable dynamic programming techniques. However, when combined with the separation of time and risk preferences, it can lead to unrealistic predictions that contradict empirical evidence. This paper demonstrates that dynamic consistency can be relaxed to hold over a much smaller domain of consumption programs while maintaining sufficient richness for practical applications and allowing the separation of risk aversion from intertemporal substitution. As an application, I introduce a new model of dynamic preferences, the Epstein–Zin–Selden–Stux preferences. These preferences are recursive only within a restricted domain. Recent experimental results by Meissner and Pfeiffer (J Econ Theory 200:105379, 2022), which Epstein–Zin preferences cannot rationalize, find a natural explanation through this new model. Finally, I consider an application of this new model to a consumption-investment problem.
Supplementary notes can be added here, including code and math.
Lorenzo Maria Stanca
Assistant Professor of Economics
Greetings! I hold concurrent appointments as an Assistant Professor at Collegio Carlo Alberto and within the Department of Economics, Social Studies, Applied Mathematics and Statistics (ESOMAS) at the University of Turin. My academic focus is centered on economic theory, with a particular emphasis on decision theory.