Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
Abstract If the future is uncertain, optimal intertemporal decisions rely on anticipating one's own optimal future behavior as is typical in dynamic programming. Our aim is to detect experimentally stylized facts about intertemporal decision making in a rich stochastic environment. Compared to previous experimental studies our experimental design is more complex since the time horizon is uncertain and termination probabilities have to be updated. In particular the decision task is non-stationary as in real life which seriously complicates the task of diagnosing behavioral regularities. In this study we give some illustrative results and provide some general perspectives. Our main result is that subjects'reaction to information about termination probablilities are qualitatively correct.