Can the endowment effect be used to increase the power of health incentives?

Firms, insurers, healthcare providers, and health-policy researchers are increasingly using financial incentives to target improvements in health behaviors.  Given limited resources, it is important to establish what types of incentive structures generate the strongest behavioral responses. We test whether the endowment effect can be used to improve the effectiveness of financial incentives for health behaviors.  The endowment effect, a consequence of loss aversion, is the phenomenon that people appear to value objects much more if they feel a sense of ownership for them.   

          In partnership with a commercial fitness facility, we enrolled newly-enrolled members for a randomized control trial. We randomize individuals into 1 of 4 groups: a group that choses a gift from Amazon (the endowment group) and receives the gift if they go to the gym 9 times during the first 6 weeks of membership, a group that receives a $30 Amazon gift card if they go to the gym 9 times during that same period, a group that receives a $60 Amazon gift card if they go to the gym 9 times during that same period, and a group that receives a $30 Amazon gift card unconditional on attendance. The idea of this intervention is to increase the power of a given financial reward by transforming an abstract monetary prize to specific objects for which the individual may feel a sense of ownership. We contrast the gym attendance rates of each group.

     This research will be significant in a number of ways.  A small literature has begun to show that incentives that leverage loss aversion can be more powerful, but it is often difficult to generate sensations of losses in incentive programs in a reliable way.  Leveraging the power of loss aversion through attachment to selected objects (i.e., endowment effect) is a novel approach.  Second, by randomizing the size of incentives for those receiving the gift-card (i.e., standard) incentive approach, we will be able to quantify the importance of the endowment effect in dollar terms.  Early research on loss aversion suggests the power of an incentive experienced as a potential loss could be equivalent to a gain-framed incentive of twice the size, but there is little evidence in field trials.  Understanding the dollar value of this treatment will help establish whether the additional benefits cover its (modestly) higher administrative costs.  Finally, this study will target newly enrolled members of a gym whose characteristics suggest they are unlikely to establish a lasting gym-use habit.  As such, this study will provide some of the first evidence of the potential benefits of early interventions to establish habit formation with those attempting to establish a new exercise routine.