Political tensions in Kenya are high in the run-up to elections scheduled for next year.
Opposition leaders have led angry protests against the country’s electoral commission and have accused the government of everything from rampant corruption and ethnic favoritism to involvement in the murder of a prominent businessman.
Meanwhile, security forces have forcefully broken up opposition protests, and government leaders accuse the opposition of disregarding Kenya’s constitution and the political process in favor of rioting in the streets.
Given the country’s history of election-related violence, it’s no surprise that many observers believe the risk of violence breaking out around next year’s elections is rising.
At the same time, political science research has tended to assume that, if politicians employ violence, it is because they benefit from doing so. Thus, the literature has largely focused on the ways in which violence helps politicians win office, with much less attention paid to how violence may hurt politicians’ ability to win votes.
Yet evaluating both the benefits and the costs of a violent electoral strategy is crucial to understanding how violence shapes voting behavior and the election outcomes.
How significant are these costs, and how do they compare to the potential benefits laid out in existing research? How do politicians perceive the effect of violence on their ability to win voter support, and are these perceptions accurate? Is violence as effective an electoral tactic as scholars and the politicians they study tend to assume?
My research in Kenya suggests that the electoral backlash against politicians who employ violence is far greater than most analysts—and Kenyan politicians themselves—tend to believe. In particular, while politicians believe that violence is at worst irrelevant or at best useful for consolidating coethnic support and winning elections, the reality is that it sharply reduces support among voters, including coethnics.
Furthermore, analyses of the relationship between violence and election outcomes suggests that the voter backlash may be great enough to offset any advantages that violence may bring.
What do Kenyan politicians believe about the effect of violence on voting? Evidence from interviews and survey experiments conducted with these politicians suggests that they see violence—and the heated ethnic rhetoric that often leads to it—as at worst irrelevant, and at best useful, in their efforts to consolidate coethnic support and win elections.
Describing why voters would support a coethnic candidate despite (or because of) a history of violence, politicians claimed that “People will always rally behind their own” or that “[h]e is considered a fighter for their interests. He cares for them. He’s willing to eliminate the intruders as they see them.”
This contrasts sharply with the results of survey experiments with Kenyan voters, which showed that violence and heated ethnic rhetoric sharply reduced voter support compared to otherwise identical candidates, even when the candidate was a coethnic.
The backlash was so strong that voters would vote for a nonviolent, non-coethnic candidate over a violent coethnic one, despite coethnicity being a very strong predictor of vote choice in Kenya.
The large and broad-based voter backlash against violent campaign tactics suggests that violence may be less effective for winning elections than most Kenyan politicians and scholars assume.
In fact, my analysis of constituency-level elections in the 1990s, when what violence occurred was committed almost exclusively on behalf of the ruling KANU (Kenyan African National Union) party, shows that violence had no significant effect on KANU presidential or parliamentary vote share, or on the likelihood of a KANU MP win.
And in the 2013 general elections that followed the large outbreak of violence around the 2007 elections, those politicians allegedly involved in organizing the violence were reelected at a much lower rate than those that weren’t—just 30 percent, compared to a reelection rate of 60 percent for incumbent MPs overall.
The evidence from Kenya therefore suggests that the backlash against violent electoral tactics is much larger than the literature—and Kenyan politicians themselves—tend to assume, and that this backlash may be large enough to offset the electoral benefits that violent coercion provides.
In many ways, these findings are highly encouraging. By showing that violence can backfire, they suggest that providing political elites with more and better information about the true effects of violent electoral tactics can help reduce their incidence, simply by appealing to politicians desire to win office. Such an approach may be more effective than normative appeals since it is aligned with the electoral incentives that often drive politicians’ behavior.
Could an information-based intervention actually work? Evaluating such an intervention would be a highly promising avenue for future research.