Measures of Trust and Participation in Collective Actions: Evidence from Rural Senegal

Abstract

This paper investigates the extent to which different measures of trust are connected to a community's ability to undertake collective action. The objective is to clarify what trust measurement method is better at explaining how heterogeneous individuals manage to engage in joint activities, despite the associated problem of free-riding. For this purpose, we use a version of the Trust Game in which subjects are matched with a random anonymous member of their own community. Results are compared to survey trust questions: the World Value Survey trust question and specific questions regarding trust in neighbors and trust in strangers. We find that trust, as measured by survey questions, has poor predictive power, while the results from a simple trust game are much better predictors of local participation in collective activities. More particularly, a one-unit change in the amount sent increases the probability of participation in Tontine, Credit Association and Animation respectively by 6%, and 3%. Furthermore, we find that individuals who consider that most people can be trusted have 17.58% more chance to participate in Animation activities.