This paper analyses election related violence during African elections in the 1990-2009 period to examine the role of international election observations. The paper tests three hypothesis: 1) The presence of credible international observer organisations increases the likelihood of election-related violence in the pre-election period; 2) The presence of credible international observer organisations has no effect on the likelihood of election-related violence on election day; and 3) The presence of credible international observer organisations increases the likelihood of election-related violence in the pre-election period relative to the probability of election-related violence on election day.
A dataset was created that includes all elections held in African countries (including North Africa) for the 1990-2009 period. The data includes 301 elections rounds for the 1990-2009 time period. Data for election violence includes a variety of event types, including violent riots, pro-government violence, and violence by non-state actors related to elections. The data set was used to test the hypotheses.
Key Findings:
- The small but growing literature on electoral violence notes that election day, even when preceded by violence, is often surprisingly peaceful.
- The peacefulness of polling relative to other parts of the electoral process can be linked to the increasing international interest in elections, as exemplified by the rise of international election monitoring.
- International election observers influence the strategic calculation of domestic elites participating in electoral processes.
- The presence of international electoral missions mitigates the potential for election-day violence because domestic actors likely refrain from intimidating opposition candidates or voters before the eyes of international observers.
- However, the presence of international observers creates incentives for political actors to engage in violent manipulation in parts of the electoral process receiving considerably less international attention, such as the pre-election period.
- Violence in the pre-election period is less likely to be detected or, at a minimum, less likely to be criticised as strongly by international actors supervising the electoral process.
- The presence of monitors increases the incidence of violence before elections relative to violence on election day.
Recommendations:
- The effect of international monitors on pre-election violence is weakened for observer missions that pay greater attention to the pre-election period, although this effect holds only for violence committed by non-state actors.
- The findings support the frequent call for greater supervision of all parts of the electoral process, and it is possible that improvements in the quality and numbers of long-term observers will succeed in deterring the use of pre-election intimidation by state actors.
- It is not clear whether the empirical findings apply outside the African context.