This paper examines clientelism in distributive politics of welfare benefits in light of the theoretical framework and its empirical application to the case of the Thailand Village Fund. The Fund is analyzed as a form of targetable welfare transfers that have become a popular electoral campaign strategy to directly target votes in many new democracies, especially in those rooted in clientelism. The Dixit-Londregan (1996) model of redistributive politics has been modified to capture clientelist exchanges where groups of voters are identified by their clientelist ties. The findings obtained from household survey data in 2004 and 2010 show that clientelist voters who are connected to their party patrons are targeted for transfers because the benefits can be distributed more efficiently to whom the patrons know well. The empirical evidence suggests that, after correcting for the selection bias, particularistic transfers existed in the form of an additional amount of transfers, rather than the likelihood of receiving them. Effects particularism in the patron-client form seem less powerful in 2010.
