Elicited Time Preferences and Behavior in Long-Run Projects

We study whether and how the experimentally elicited risk and time preferences of subjects are associated with their behavior in long-run projects. First, risk and time preferences are elicited from time-dated monetary choices to estimate a general discount and utility function at an individual level, then subjects work on a longitudinal project that requires effort in multiple periods. We find that present bias in the form of a fixed cost or variable cost (quasi-hyperbolic discounting) is not supported by monetary choices. Analyses of allocation patterns of work reveal that the estimated utility and discounting models are not compatible with the observed allocations. We find evidence of both present and future bias, although the former is more prevalent and severe, and subjects exhibit naivete in their choice reversals. Furthermore, discount rate and present bias parameters estimated based on monetary choices have predictive power on how work is allocated in the long-run project.