Who Chooses Commitment? Evidence and Welfare Implications

This paper develops an approach for estimating people's perceived and actual self-control problems in the field, and for investigating whether offers of commitment contracts are well-targeted tools for addressing self-control problems. In an experiment on gym attendance (N=1,248), we estimate a model of partially naive quasi-hyperbolic discounting, and we study (i) how commitment contract take-up relates to structural estimates of people's perceived and actual self-control problems and marginal benefits of behavior change, (ii) if commitment contract take-up is in part due to boundedly-rational understanding of incentives, and (iii) whether commitment contracts increase welfare, and how they compare to linear subsidies or taxes. We find high take-up of commitment contracts to increase gym attendance, which translates to significant increases in exercise. However, contract take-up is not positively related to estimates of actual or perceived self-control problems. A novel test suggests that this may be partly explained by boundedly-rational take-up decisions: many individuals are willing to take up commitment contracts both for higher gym attendance and for lower gym attendance, and there is a significantly positive correlation in people's propensity to take up both types of contracts. Our estimated model of quasi-hyperbolic discounting and gym attendance implies that offering our commitment contracts lowers consumer surplus and is less socially efficient than utilizing linear incentive schemes that achieve the same average change in behavior. We estimate an optimal exercise subsidy of $7.54 per attendance