Elections and Collective Action: Evidence from Changes in Traditional Institutions in Liberia

Numerous recent field and lab experiments have found that elections cause higher subsequent levels of collective action within groups. This article questions whether effects observed in these environments apply when traditional institutions are democratized. We test the external validity of the experimental findings by examining the effects of introducing elections in an indigenous institution in Liberia. We use a break in the process of selecting clan chiefs in Liberia at the end of the country’s civil war to identify the effects of elections on collective action within communities. Drawing on survey data and outcomes from behavioral games, we find that the introduction of elections for clan chiefs has little effect on community-level and national-level political participation but it increases contentious collective action and it lowers levels of contributions to public goods. These findings provide an important counterpoint to the experimental literature, suggesting elections have less salutary effects on collective action when they replace customary practices. How do community-level institutions affect collective action? Political scientists have long argued that elections can create senses of community, inculcate participation, and legitimize leaders in a way that facilitates subsequent collective action.1 Furthermore, a number of recent lab and field experiments have empirically demonstrated that the introduction of elections can facilitate collective action and civic participation.2 Yet many communities that do not select their leaders by elections have also proved themselves capable of overcoming collective action dilemmas. In particular, the reliance on “traditional” leaders to mobilize communities for collective action is common in developing countries. Traditional leaders are typically not elected, yet they are argued to be effective in mobilizing collective action because of their ability to appeal to custom and long-established norms as a source of legitimacy.3 In the context of customary institutions for local governance, would elections for leaders improve collective action? We believe this question is still open, as the findings from recent lab and field experiments may not apply in cases in which well-established indigenous institutions are democratized. On the one hand, lab experiments ask respondents to make decisions in novel environments in which non-electoral institutions may be relatively ineffective because they do not draw on established norms and expectations. On the other hand, in most field experiments conducted to date, elections have been introduced into new institutions that have paralleled rather than supplanted traditional governance structures. As a result, these studies may not capture the full effects of replacing customary methods with elections in existing traditional institutions. It is important to complement these analyses with studies in which customary methods of selecting leaders are replaced by electoral methods. The main challenge to studying the effects of introducing elections in “traditional” institutions is developing an identification strategy. Clearly, few entrenched local leaders would volunteer to be part of an experiment that randomly introduced elections into their communities. Instead, this article takes advantage of a break in the process of selecting local leaders in Liberia following the country’s civil war. At the end of the war, some local clan chiefs were appointed by small groups of elite, while others were elected by their communities; however, virtually all clan chiefs who left office after the end of the war were replaced by chiefs selected through elections. This break permits us to identify the effect of introducing elections in traditional institutions.