The importance of preaching to the converted : The strategic use of campaign rallies, campaign promises, clientelism, and violence in African elections

This dissertation addresses one of the main puzzles concerning elections in young democracies: Why do undecided voters receive so little attention in parties' election campaigns? While most theoretical models would expect parties to focus their campaign efforts on voters who do not have strong preferences for any party, this is not what the empirical evidence from young democracies shows. Rather, parties spend much time and money on campaigning among their own supporters, which is what the standard models would simply regard as a waste of valuable resources. I provide three key answers for this puzzle. First, parties do not waste resources on courting their supporters who are certain to turn out, at the expense of campaigning among swing voters. They rather mobilize those supporters who would otherwise not go and vote. Second, in contexts where campaign promises have little credibility, the organizers concentrate their energy on voters who are likely to trust them the most, namely their core supporters. Third, in situations in which parties use electoral violence to affect election outcomes, they concentrate their intimidation strategies on citizens who would be difficult to win over, so as to disenfranchise them. This frees up resources to offer benefits to their supporters to mobilize them to turn out on Election Day. The arguments are subjected to various empirical tests analyzing a range of campaign strategies used by presidential candidates in young democracies, and potentially in more established democracies as well. These strategies include: visits by presidential candidates to electoral constituencies to hold campaign rallies; promises of local club goods to constituencies as opposed to national programmatic promises; and attempts at winning votes using voter bribery as opposed to exercising violence. I test predictions of my argument applying a range of methodological approaches and using various original data sources. During fieldwork in Ghana, I collected event data on the journeys of presidential candidates across the country on the basis of content-analysis of two daily newspapers during the campaigning period. In addition, I compiled audio recordings from these campaign rallies and conduct content-analysis of the campaign speeches these candidates held. Furthermore, I conducted qualitative interviews with campaign managers at the national, regional and constituency level as well as focus group interviews with voters. In addition to these observational data, I carried out a survey experiment on the credibility of campaign promises. To establish external validity for the findings from campaigns in Ghana, I compile individual and regional-level data on the use of clientelism and violence in a total of 10 African countries, combining Afrobarometer survey data with regional-level election data. The overall results show that candidates use campaign rallies, campaign promises and clientelistic benefits largely to mobilize turnout among their potential supporters. The findings further show that candidates concentrate promises of local club goods in contexts in which they enjoy a comparatively high level of credibility. In line with my expectation, the incumbent makes many of such local promises in constituencies where his partisans are concentrated. Furthermore, it is nearly exclusively the incumbent who promises local club goods in the first place. As incumbents already exercise discretion over the use of public resources at the time of the campaign and can thus make costly investments, their promises are more credible than those of opposition candidates. The results from the survey experiment conducted in Ghana's capital Accra confirm that the incumbent is regarded as more credible with his promises than the challenger. They further support my argument that partisans evaluate campaign promises made by the candidate they support as much more credible than if the same promises are attributed to a different candidate. Finally, in line with the hypothesis that parties concentrate bribes or promises of redistribution on their partisans, because they can disenfranchise voters that do not support them, I find that independent voters and those living in contested regions are most at risk of being subjected to violent intimidation. This dissertation thus provides key answers for the puzzle of why parties in young democracies court their own supporters so intensely. The findings also have important implications for the study of election campaigning in young democracies, beyond this puzzle. The evidence presented shows that the prevalence of mobilization as a campaigning strategy has been seriously under-estimated by past research. This informs an important debate in the literature and speaks in favor of turnout-buying rather than vote-buying, and mobilization rather than persuasion. The dissertation also advances recent efforts to integrate the use of clientelism and violence as repertoires of campaign strategies. The findings suggest that these two strategies are used among different types of voters and with different goals