PDF version

Reducing the Incidence of Vote-Buying in Uganda

Location: Uganda
Sample: 4,200 villages in 103 constituencies
Timeline:
2015 to 2016
Research Initiative: Governance Initiative
Target Group: 
Voters
Outcome of Interest: 
Corruption
Elected official performance
Electoral participation
Transparency and accountability
Voter behavior
AEA RCT Registration Number: 
Partners:

Vote-buying remains a major impediment to full democracy in many low-income countries. Researchers conducted a randomized evaluation to study how a large-scale campaign against vote-buying affected not only citizens’ willingness to sell their votes but also politician and party behavior in the 2016 election in Uganda.  

Policy Issue 

Vote-buying practices are common in many developing countries, undermining democratic elections. Politicians use a variety of practices to buy votes, including rewarding supporters for showing up at polls, driving voters from other constituencies to the polls in competitive areas, or targeting gifts to voters who are likely to reciprocate with their vote. Governments and civil society organizations have tried many strategies for reducing vote-buying, such as encouraging voters to refuse vote-buying arrangements, advising voters to take what they are offered but vote according to their conscience, or providing voters information about the qualifications of candidates. However, there is a need for more evidence on what types of interventions effectively reduce vote-buying. Furthermore, if vote-buying is reduced, what impact does this have on the campaign strategies and characteristics of elected politicians? 

Context of the Evaluation 

Uganda is a semi-authoritarian regime: at the national level, it has a multiparty system dominated by a single ruling party, in power since 1986. In February 2016, Uganda held general elections, including voting for President and Members of Parliament. Vote-buying is pervasive in Ugandan elections. Campaign activists, working for politicians and political parties, often act as key brokers of votes, using a combination of money, gifts, and moral enforcement to bring voters to the polls. As they generally keep a portion of the money they receive from candidates, they are also the main beneficiaries of campaign spending. 

Details of the Intervention 

In partnership with the National Democratic Institute (NDI) and the Anti Corruption Coalition Uganda (ACFIM), researchers conducted a randomized evaluation to study the impact of a large-scale campaign against vote-buying on citizen participation in vote-buying and politician and party behavior. The campaign contained activities designed to dissuade citizens from selling their votes.

The evaluation took place in the months leading up to and following the February 2016 Ugandan general elections and included approximately 4,200 villages across 103 parliamentary constituencies in 48 districts. Constituencies were randomly assigned to one of two groups:

  • Voter Intervention: In these constituencies, NDI used a network of village-level activists to persuade entire villages to commit to not selling their votes, using a combination of information dissemination, community participatory meetings, posters, and robo-calls. Through the intervention, communities produced a collective declaration to refuse offers of gifts or money in exchange for votes, creating “no vote-buying villages.” 
  • Comparison Group: These constituencies did not receive any intervention.

Within the voter intervention, researchers also randomized the share of villages within each constituency that received these activities. This allowed researchers to determine whether the community-based, anti-vote-buying activities had any spillover effects on nearby villages that did not receive the intervention.

Researchers will use official electoral data and a post-election survey to analyze how the interventions affected turnout, vote shares, vote prices, perceptions of vote-buying, and changes in party and politician behavior.

Results and Policy Lessons 

Project ongoing; results forthcoming.