Brief 55: Candidate Participation in Electoral Debates
EGAP researcher: Horacio Larreguy
Other authors: Jeremy Bowles
Geographical region: Africa
Research question: What are the electoral consequences of candidate selection into debates that disseminate programmatic information?
Preparer: Catlan Reardon
The study takes place in Liberia, a low-income West African country with high levels of clientelism and a weak media sector. The intervention was implemented, in partnership with USAID and the NGO Internews, during the House of Representatives election of October 2017 across 73 electoral districts. Representatives in the House are tasked with making laws, overseeing access to development funds, and allocating and implementing public goods.
The partner, Internews, ran 129 standardized debates with the goal of disseminating programmatic policy promises of participating candidates. By randomly varying the intensity of the invitation to the debate, the authors are able to assess the effects of variation in incumbent and challenger participation in the debates on political knowledge and behavior of voters.
The partner NGO organized 129 standardized debates to elicit the policy promises of candidates for the House. The platforms were then rebroadcast by community radio stations. The debates were held across all 73 districts from mid-August to mid-September prior to the October 2017 elections for the House of Representatives. In order to induce variation in debate participation, the authors randomly varied the intensity of the invitations to all candidates in each district.
The control condition involved an invitation by the relevant Liberian journalist association along with basic logistical information. The treatment condition added a set of official letter invitations, personal phone calls from prominent journalists, and SMS reminders to candidates. The intervention was designed to reduce uncertainty about the debates and allay concerns regarding participation that differentially affected the decision of leading candidates. The control (low-invitation effort) and treatment condition (high-invitation effort) were randomly allocated across all 73 districts. Randomization was blocked on a set of pre-treatment covariates, resulting in 19 blocks, with 3 to 4 districts per block. The authors, then, randomly assigned 38 districts to the treatment condition and 35 to the control.
The authors employed various data collection methods, including a nationwide panel survey of over 4,000 citizens, a survey of over 600 candidates, a survey of more than 50 radio stations, transcripts from debates, 20 focus groups, and polling station-level electoral results.
Overall, 59% of candidates participated in the debates, with a higher proportion of opposition candidates involved (60% compared to 48% of incumbents). In the treatment condition, incumbents and their main challengers were more likely to attend their debates compared to those candidates in the control condition. The treated districts show a 21.2 percentage point increase in the share of incumbent and challenger candidates attending the debates, where incumbents were 76% more likely to participate and challengers 43%. Debate exposure – whether citizens heard the debate and how often – was also higher in treated districts. The intervention also increased political engagement, in the form of seeking out political information, of citizens in the treatment condition.
Receiving the treatment also led to candidates making heterogeneous decisions regarding their other campaigning efforts. Challengers in treated districts responded by decreasing on-the-ground campaigning efforts. Meanwhile, incumbents increased their use of radio campaigning.
The authors found that debate exposure decreased uncertainty among citizens about the competence and policy priorities of participating candidates. Yet, this updating of beliefs tended to only concentrate in, and benefit, incumbent candidates. This is due to the fact that participating incumbents were judged to have better qualifications and policy positions than opposition candidates. In turn, voters were more likely to vote for incumbents in treatment districts, particularly when they matched voter policy preferences. The authors suggest the ability of more experienced incumbents to correctly evaluate the returns to selecting into debate participation explain these effects.
The results reveal the challenge of increasing accountability in contexts with high levels of clientelism and a weak media sector. While most models of electoral accountability focus on the provision of information, the reality in many of these settings is that many candidates do not face incentives to disseminate programmatic information. Inducing participation in debates that provide policy information can be one solution to the pervasive use of non-programmatic tactics in developing low-information contexts. However, the results show that future interventions should focus on leveling the playing field for less experienced challenger candidates, as more experienced incumbents are better able to self-select into providing programmatic information.