Clientelism and Preferences for Redistribution: Insights from a Survey Experiment in South Africa

Clientelistic practices are generally thought to stand in the way of redistribution via public policies because poorer voters, in particular, give up their policy preferences in exchange for a particularistic good. The underlying assumption is that poorer voters would ultimately support a redistributive political agenda but collective action problems or short-term credit constraints make them prefer an immediate transfer. In this paper, we study the relationship of clientelism and preferences for redistribution in one of the most unequal countries in the world, South Africa. In particular, we use a survey experiment to study the effect of a mobilizational message against inequality on redistributive preferences. We study the effect separately for individuals with a favorable view of clientelistic practices vs. individuals who find it unacceptable. Preliminary results from our pilot study with 300 respondents in townships around Cape Town indicate that in the control group, individuals favorable to clientelistic practices do indeed have lower preferences for redistribution than individuals rejecting these practices. More importantly, the treatment effect of our mobilization message is bigger for individuals favorable to clientelistic practices. For example, such individuals appear to be more inclined to increasing top taxes, the minimum wage and are more likely to say that the rich don't deserve their incomes. This transformation of clientelistically inclined respondents from most anti-redistributive to the most “left-wing” in our sample lends support to the idea that demand for clientelistic and for public policy redistribution are substitutes. The paper will be based on the full survey with 3000 individuals that will be carried out from mid-January to mid-February 2014.