Democratization and Electoral Violence in Sub-Saharan Africa, 1990-2007

Abstract

This paper investigates two principal questions that are underexplored in the existing literature on democratization and African politics. First, how frequent has electoral violence been in Sub-Saharan Africa since the continent’s grand democratic experiment began in 1990? Second, what explains the variation in why some elections are violent but others are not? To answer the first question, we disaggregate the concept of “electoral violence” into pre- or post-vote violence, incumbent or challenger violence, and level of violence, and we construct an original dataset measuring these characteristics. We generate several findings; most importantly, we find that significant electoral violence occurs in 19% of elections, that most violence occurs before the vote and is perpetrated by incumbents, and that if violence breaks out after an election it tends to be more severe and involve challengers. To answer the second question, we map out causal mechanisms linking elections to violence and derive some preliminary hypotheses. Our quantitative analysis using descriptive statistics does not yield any strong patterns, but suggests that margin of victory, regime type, cleavage structure, income level, and pathway to power matter. However, many of the hypothesized causal mechanisms are not captured in the quantitative cross-national data, and the data themselves are limited. To compensate, we turn to qualitative analysis, examining structurally similar cases that vary in whether or not significant electoral violence occurs. We find a consistent pattern that access to vital resources, especially land, is the central issue in cases of high electoral violence; that elections trigger claims of indigeneity in states with significant migration; and that high violence is more likely where local rival groups fighting over access to vital resources are essential to the winning electoral strategy of national political parties in close elections.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 146 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.