392 Cohen Hall
Manoj Mohanan, Assistant Professor in the Sanford School of Public Policy at Duke University
Abstract: In settings where citizens participate in collective action towards a social goal, information interventions are commonly employed to encourage participation. A key concern in collective action is that of free riding behavior of individuals. In contrast, the presence of common knowledge within social networks – where individuals are informed of their neighbors’ awareness and vice versa – might be able to mitigate such free riding and instead lead to individual participation increasing as a function of participation within her own networks. However, disentangling endogeneity to estimate the causal effect of participation within networks on one’s own participation has been extremely difficult. We attempt to estimate this relationship by relying on an experiment conducted in the context of social accountability interventions in Uttar Pradesh (India), where village members were invited to participate in meetings to review performance of health care service delivery in their villages. Based on data on social networks of all individuals in the village, we randomly selected individuals to receive reminders about upcoming meetings and also receive small financial incentives to attend the meeting. We conducted the experiment in a setting without common knowledge, and then repeated the experiment with creation of common knowledge. In an instrumental variables framework, using predicted attendance within networks as a function of randomized incentives, we find evidence suggesting that introduction of common knowledge leads to individuals being more likely to participate as a function of higher participation within their own networks. We also observe evidence that is consistent with free riding when there is no common knowledge - individuals’ participation reduced with the share of village households that received the information.