Networks and Information: Evidence on Political Participation in Mozambique

Political accountability requires electoral participation and informed voters. Both have been lagging in Mozambique. We designed and implemented a field experiment during the Mozambican municipal elections of 2013. We study the impact on political participation of social influence through text messages (SMSs), and of information provided through a free newspaper. We randomly assign peers to experimental subjects in order to test for homophily. Measurement of political participation comes from behavioral measures related to the sending of SMSs by subjects, from several individual measures of voter turnout based on survey information, and from the replication of the voting procedure. We follow a pre-analysis plan when showing our results. We find positive average effects of the SMSs on political participation, but no clear evidence of homophily. We do not find effects of the distribution of the newspaper, but positive interaction effects between SMSs by peers and newspaper distribution.