Support Us

Politician Entry, Selection, and Performance in Sierra Leone

Institutions & Governance Sierra Leone

Freetown, Sierra Leone | Photo credit: Adobe Stock Images

Context

In regions which consistently support a single political party, the most important selection of a representative can be at the primary stage. However, in low- and middle-income countries, primaries are frequently dominated by elites and conducted without transparency, decreasing the likelihood that parties select high-quality, representative candidates. Publicly screening candidate debates have previously shown positive effects on voter knowledge, candidate quality, and campaigning effort during general elections, but had neither been tested on a broader scale in a less-controlled environment nor been used to inform citizens’ participation in primaries.

Researchers shed light on these questions by partnering with both traditional political parties in Sierra Leone to experimentally vary the scope for voters to influence parliamentary primaries. In the status quo, party officials choose among potential candidates with no direct participation by voters. The tradeoffs of such a system are unclear; party officials may be better informed about candidate qualifications, but they may also value traits unrelated to performance in office, like party loyalty or willingness to pay for the nomination.

Study Design

For a random selection of races for MP seats, instead of the status quo, the political parties hosted publicly broadcasted conventions where candidates presented their qualifications and policy platforms. The researchers then conducted representative opinion polling of voter preferences over candidates, and shared the results with party officials. While parties were still free to choose a candidate regardless of voter preferences, the intervention significantly increased whether parties chose the voters’ first choice candidate, from 38 to 61 percent. Moreover, when they compare the characteristics of candidates chosen under the new selection method compared to the status quo, they find that the more democratic method selected candidates who had stronger records of previously providing public goods, which is something voters value. Overall, the results suggest that broadening representation benefited voters, as not only were parties more likely to select their preferred candidates, but those candidates had stronger records of public goods provision.

Results and Policy Lessons

Furthermore, the results suggest that a lack of information explains why party officials often failed to select the most popular candidate in the status quo. Survey results suggest that “90 percent of constituency-level party officials presume local voters share their first choice over aspirants, when voters in fact only agree with them about one-half of the time; and in one-third of races, not a single party official accurately guessed which aspirant ranked first among voters” (Casey, Kamara, and Meriggi, AER 2021). Since party officials clearly found the information on voter preferences useful (as they changed their choices of candidates as a result), this raises the question of why political parties were not already collecting this information. The evidence suggests the parties may have lacked affordable ways of polling voters.

Given these results increased representation, both the sitting president and the presidential flagbearer for the opposition party (at the time, now the ruling party) endorsed this project. Highly ranked officials from political parties have appreciated the initiative and requested to extend it to the entire country. Officials from both parties have also expressed interest in using this approach as standard practice in the selection of candidates.

Read more about the impact of this project, and the published paper (Casey, Kamara, Meriggi AER 2021).

Researchers
Partners
  • Search for Common Ground and the Political Parties Registration
Timeline

2017 — ongoing

Share Now

Copyright 2024. All Rights Reserved

Design & Dev by Wonderland Collective