On Sunday, Lebanese citizens will vote in national elections for the first time since 2009. These are the first elections since the passage in June 2017 of a new electoral law and the first since the 2016 Beirut municipal elections, when a grass-roots campaign won almost 40 percent of votes, challenging Lebanon's long-standing patronage-based sectarian parties.
Will that challenge actually change the voting behavior of Lebanese citizens on the national level? To test the relative influence of service provision, programmatic platforms and religious identity on citizens' political behavior, we conducted a survey experiment in October and November 2017 in which 2,400 respondents were asked to choose between two hypothetical candidates whose profiles varied randomly on a range of attributes. We found that while specific types of clientelism and policy issues have some limited influence, voter preference remains most influenced by ethnicity and religion.
Clientelism is not just about short-term exchanges
Academic literature on politics in developing countries -- including those with politicized ethnic and religious divides, like Lebanon -- often highlights the role of vote-buying during elections. However, our study shows that different forms of goods have different effects on electoral behavior. While low-value benefits such as cash and food baskets do not move voters, people are on average more likely to choose a candidate who provides medical treatment and employment opportunities -- by 2 percent and 4 percent, respectively -- to family members.
This suggests that disbursing higher-value clientelist goods, which require continuous relationships between patrons and clients and greater investment, is more effective than one-shot transactions.
Issues are important
Much scholarship on political behavior in the Middle East and other developing regions indicates that programmatic considerations, or the details of a candidate or party's platform, do not drive voter behavior. Where parties are weak and legislatures have little influence on policymaking, policy platforms are less effective. Our study shows otherwise.
Programmatic appeals matter, but not in the way we expected. Rather than simply having a well-articulated platform, what matters is the issue itself. Having a detailed plan does not increase support for a candidate. Neither does a programmatic emphasis on waste management, which is surprising given the ongoing "garbage crisis" in Lebanon. But respondents were 4 percent more likely to vote for candidates who raise the issue of unemployment.
Ethnicity and religion matter, not just as a cover for clientelism
Our survey respondents were about 10 percent more likely to select a co-ethnic, or a candidate from the same religious community, which is the strongest predictor of support for candidates. If religion trumps clientelist distribution and policy concerns about pressing issues such as garbage collection and unemployment, what does it mean to citizens? Why do they seem to care more about a politician's religious identity than the ability to deliver clientelist handouts or policies that they favor?
Our survey allowed us to probe whether a candidate's religion serves as a cue for potential access to clientelist benefits, a guarantee to focus on protecting the community from outside threats, an indicator of shared policy preferences among members of the same ethnic or religious group, or a preference for people from the same community -- what social psychologists refer to as "in-group love."
Although we find some evidence that clientelist benefits offered by a co-ethnic have more appeal, none of the other potential explanations is supported by the data, with the exception of in-group love. The apparent favoritism for co-ethnic candidates that respondents expressed may then reflect a human tendency to distinguish between in- vs. out-groups, fulfilling a need for belonging and facilitating social cooperation. This penchant for communal group attachment may be all the more compelling in the context of weak state institutions, which fail to induce an attachment to a broader national political community.
Lebanon's national elections are a useful case study from which we can derive a better understanding of what drives people's electoral behavior in ethnically diverse patronage democracies. The results of Sunday's national elections may show what some already suspect -- specifically, that in terms of garnering votes, independent groups often focus quite heavily on programmatic platforms and, to their own detriment, ignore communal affinity. Such a conclusion would have major implications for any independent and policy-centered campaign both in Lebanon and the region.
Melani Cammett is Clarence Dillon Professor of International Affairs in the Government Department at Harvard University.
Dominika Kruszewska is a PhD candidate in the Government Department at Harvard University.