Why Do Voters Support Corrupt Politicians? Experimental Evidence from South Africa

Democratic elections are supposed to prevent corrupt politicians from winning office. In practice, however, voters frequently vote for corrupt politicians. In this paper, we examine why voters sometimes support corrupt candidates. We interrogate this seeming paradox from the perspective of explanations highlighting that voters support corrupt candidates because of lack of information; because of clientelist exchanges of material benefits in return for votes; or because of party loyalty. We test these explanations through an embedded experiment in a new nationwide survey in South Africa – a country where issues of corruption are highly salient. We find that voters express strong willingness to punish corrupt candidates across all treatment conditions. However, voters are more lenient towards corrupt politicians when they are offered material benefits in return for their vote as part of a clientelist exchange. This suggests that clientelism serves to reproduce corruption, and have important implications for the fight against corruption.