Combating vote-selling: A field experiment in the Philippines

By Dean Yang

Dean Yang

Vote-buying and vote-selling have by and large been detrimental to democratic consolidation, yet they are unfortunately pervasive phenomena in many developing democracies. Because of their inimical effects, governments, NGOs, and international donors have directed significant attention and resources toward combating vote-buying and selling through grassroots-level voter education campaigns. To-date, however, little is known about how well these campaigns work in terms of changing voter perception and behavior.

U-M faculty Dean Yang (Ford School), Allen Hicken (Political Science), and Stephen Leider (Ross Business School), and Ford School doctoral candidate Nico Ravanilla teamed up over the winter and spring of 2013 to conduct a field experiment in the Philippines to assess the efficacy of anti-vote-selling interventions. In particular, they sought to evaluate how pairing a standard educational appeal to voters not to sell their votes with a) a simple (and unenforceable) promise not to accept money or b) to accept money, but to vote according to conscience, can impact voters’ perceptions and behavior at the polls. Recent evidence from behavioral psychology and economics has shown that promises and other informal agreements can substantially change behavior and lead to more socially efficient outcomes by changing the social norms. They brought these insights to the field and tested the hypothesis that promising not to sell one’s vote will reduce voter susceptibility to vote buying.

An IPC Research Grant allowed Ravanilla to travel to the Philippines in advance of the May 2013 midterm elections. Ravanilla set the groundwork running and managed the field implementation of the baseline and end-line surveys as well as the promise interventions. The study focused on the elections for mayor, vice mayor, and city council in Sorsogon City, Philippines. The study team conducted two treatments, asking subjects to either promise not to accept money from a candidate, or promise to vote their conscience even if they accept money. In comparison to a control treatment that included only an informational appeal, the promise not to accept money led to a 10 percent decrease in vote selling in the city council election (where monetary offers were small). By contrast, the promise to vote one’s conscience led to a 10 percent increase in vote selling in the mayoral races (where monetary offers were larger). Both offers increased awareness of and sensitivity to vote-buying behavior.