How are Preferences For Commitment Revealed?

A large literature treats take-up of commitment contracts, in the form of choice-set restrictions or penalties, as a smoking gun for time-inconsistency or self-control problems (for short, "present focus"). This paper develops techniques for inspecting this assumption, presents new field-experimental results that challenge it, and provides an alternative approach to measuring present focus. Theoretically, we show that in a standard model of quasi-hyperbolic discounting, demand for commitment is unlikely when there is some uncertainty about the future. In a field experiment with 1292 members of a fitness facility, we test for the presence of imperfect perception of contract value and experimenter demand effects in commitment contract take-up by offering contracts both for going to the gym more and for going to the gym less. We find that there is significant take-up of both types of contracts, and that individuals who take up contracts to go to the gym more are the most likely to take up contracts to go the gym less. These starkly conflicting choices are inconsistent with all standard models of present focus, and suggest that offers of commitment contracts may not be well-targeted policy tools. However, we show that a combination of belief forecasts and elicitations of willingness-to-pay for piece-rate incentives provides more robust identification of present focus and people's awareness of it. We apply the methodology to our experimental results to obtain estimates of the partially sophisticated quasi-hyperbolic discounting model