Voting and Peer Effects: Experimental Evidence from Mozambique [Jan. 2015]

Voter education campaigns often aim to increase voter participation and political accountability. Randomized interventions were implemented nationwide during the 2009 Mozambican elections using leaflets, text messaging, and a free newspaper. We study the peer effects triggered by the campaign within households and villages. We investigate whether treatment effects are transmitted through social networks and geographical proximity at the village level. For individuals personally targeted by the campaign, we estimate the reinforcement effect of proximity to other targeted individuals. For untargeted individuals, we estimate how the campaign diffuses as a function of proximity to targeted individuals. We find evidence for both effects, similar across treatments and proximity measures. The treatments raise the level of information and interest in the election through networks, in line with the average treatment effect. However, we find a negative network effect of the treatments on voter participation, even though the average effect of the treatments themselves is positive: the effect of treatment on more central individuals is lower and sometimes negative. We interpret this result as a free riding effect, due to the fact that voter participation is costly.