Political Trust, Shocks, and Accountability: Quasi-Experimental Evidence from a Rebel Attack

Abstract

How do violent attacks affect political trust and accountability of democratic governments – and ultimately their chances of surviving in office? Using the case of an unanticipated attack by Tuareg rebels on a military garrison in the West African country of Mali in December 2008, we examine how the attack affected presidential approval and presidential trust. We do so by using survey data which – coincidentally – was collected in the days surrounding the Tuareg attack. The chance occurrence of the attack a few days into the survey demarcates respondents into a group surveyed before the attack and a group surveyed after the attack. We exploit the quasi-experimental nature of this design to study how the public’s presidential approval and trust responds to the news of the rebel attack. The credibility of the quasi-experimental design is strengthened by the fact that respondents in the particular region where the attack was carried out were interviewed in the days surrounding the attack, allowing us to mitigate the effects of geographically imbalanced sampling of respondents. Our estimates show that presidential approval and trust decrease significantly following the attack. However, the results also suggest that voters mainly attribute responsibility to the president, and that the effect is strongest in the region of the attack. While parliamentary approval and trust also declines, voters’ approval of and trust in local governments is unaffected. This suggests that in presidential systems, voters do in fact attribute responsibility for political violence such as rebel attacks mainly to the president and to a lesser extent to parliament and local governments.

Keywords: Political violence, rebel attacks, accountability, political trust, quasi-experiment, Mali, Africa