This article presents a field experiment conducted in Benin during the 2001 presidential elections to investigate the impact of clientelism on voting behavior. The study tests whether voters prefer clientelist (personal favors) or public policy messages and examines how ethnic affiliation, gender, and age influence voting behavior. The experiment involved randomly assigning villages to either a clientelist or public policy platform, with a control group. The results show that clientelism is effective for regional and incumbent candidates, while public policy messages are less effective in the North and for regional candidates. Women tend to prefer public goods over clientelist appeals, and younger and older voters have similar policy preferences. The study also finds that clientelist appeals are more credible when delivered by incumbents, as they can demonstrate their ability to provide promised goods before the election. The findings suggest that clientelism is not universally accepted, even among poor voters, and that public goods are preferred by women and other groups. The study highlights the importance of understanding the factors that influence voting behavior in the context of clientelism and public policy. The results contribute to the growing literature on redistributive politics and clientelism, providing empirical evidence on the determinants of voting behavior in African politics. The study also underscores the need for further research to validate the findings and to explore the effects of clientelism in different contexts.This article presents a field experiment conducted in Benin during the 2001 presidential elections to investigate the impact of clientelism on voting behavior. The study tests whether voters prefer clientelist (personal favors) or public policy messages and examines how ethnic affiliation, gender, and age influence voting behavior. The experiment involved randomly assigning villages to either a clientelist or public policy platform, with a control group. The results show that clientelism is effective for regional and incumbent candidates, while public policy messages are less effective in the North and for regional candidates. Women tend to prefer public goods over clientelist appeals, and younger and older voters have similar policy preferences. The study also finds that clientelist appeals are more credible when delivered by incumbents, as they can demonstrate their ability to provide promised goods before the election. The findings suggest that clientelism is not universally accepted, even among poor voters, and that public goods are preferred by women and other groups. The study highlights the importance of understanding the factors that influence voting behavior in the context of clientelism and public policy. The results contribute to the growing literature on redistributive politics and clientelism, providing empirical evidence on the determinants of voting behavior in African politics. The study also underscores the need for further research to validate the findings and to explore the effects of clientelism in different contexts.