The lotteries of elections and violence:
Why actors in post-conflict countries decide to organise
electoral violence
Hannah Smidt
School of Public Policy
Department of Political Science
University College London
Presentation prepared for the Peace Science Conference
12-14 November 2015, Oxford Ms
1 / 18
Puzzle
Why are some post-conflict election more violent
than others?
N=287, Latin America and Africa
2 / 18
Explaining Côte d’Ivoire
Large-scale violence 2010
Electoral peace 2015
3 / 18
Different perspectives on electoral violence
1
Studies on electoral violence
(Fisher 2002 to Norris 2015)
√
• Motives resulting from electoral dynamics
2
War-to-democracy research
(Rationalist explanations for electoral violence à la Fearon 1995)
√
• Inclusive to spoiler violence
√
• Opportunities
3
Democratization literature
(Rustow 1970 to Przeworski 2014)
√
• Motives: Losses from democratic transition
4
Conflict research
(Gurr 1970 to Cederman et al. 2013)
√
• Motives: Resentment of rivals’ electoral victory
4 / 18
My contribution
•
Novel framework drawing on lottery analogy
and synthesizing different perspectives
•
First large-n cross-national study in
post-conflict countries specifically
5 / 18
Lotteries of elections and violence
Lottery of elections
• Anticipated losses (large?)
• Probability of losing elections (certain?)
versus
Lottery of violence
• Anticipated losses (severity of punishment?)
• Probability of succeeding with violence (opportunity?)
6 / 18
When the lottery of elections is unattractive
Loss aversion:
H1:
Losses in the lottery of elections
+
−→
electoral violence
• relative losses vis-à-vis other actors
• relative losses vis-à-vis position to which actors feel entitled
Certainty effect:
H2:
Certainty of electoral defeat
+
−→
electoral violence
7 / 18
When the lottery of violence is attractive
Punishment:
H3:
Losses in the lottery of elections
−
−→
electoral violence.
Opportunity:
H4:
Probability of winning by force
+
−→
electoral violence.
8 / 18
Testing observable implications
Concept
Losses (elections)
Certainty of defeat
Losses (violence)
Probability (violence)
Indicator
Experience with democracy
Plurality
Conflict duration
First PC elections
Negative power shifts
Information about fraud
Unfavourable polling (Gov.)
Independent judiciary
GDP per capita
Peace spell before elections
Press freedom
Effect
–
+
+
+
+
+
+
+
–
–
–
9 / 18
Research design
•
Cross-national comparison
• 287 national-level elections
• 47 post-conflict countries in Africa and Latin
America
• 1990-2010
•
Negative binomial regression
• DV = Count of violent events (SCAD)
• related to elections
• +/-90 days of election day
•
IVs = Measures of the parameters in two lotteries
10 / 18
Results
If actors have more to lose and are certain of defeat in the
lottery of elections, then electoral violence increases.
Losses elections
Certainty defeat
Democratic experience
Plurality
Conflict duration
First post-conflict elections
Negative power shift
–
+
n.s.
+
+
– 0.98 ev
+ 0.46 ev
Information about fraud
+
+ 0.49 ev
+ 0.32 ev
+ 1.37 ev
11 / 18
Results
The probability of winning by force and losses in violence
cannot predict variation in electoral violence.
Losses/punishment
Judicial independence
n.s.
Probability force
GDP per capita
Free press
Peace spell
n.s.
n.s.
n.s.
12 / 18
Robustness
Robust for...
√
no issue restriction
√
pre-electoral and post-electoral violence separately
√
only elections held immediately after conflict (N=116)
√
excluding outliers (N=280)
13 / 18
Robust for government violence
Differences to general model:
Conflict duration: –
⇒ Longer conflicts make peaceful governments
Negative polling: n.s.
Fraud:
n.s.
⇒ Fraud assures governments of victory
GDP per capita: –
⇒ Opportunity (cost) matters for governments
14 / 18
Robust for opposition violence
Differences to general model:
Conflict duration:
⇒ Hardened identities and larger losses
First post-conflict elections / power shifts:
⇒ Losses for incumbents only
+
n.s.
Judicial independence:
⇒ Better indicator for fairness?
–
GDP per capita:
⇒ due to Mexico
+
15 / 18
Conclusion and next steps
Implications of findings:
√
Large losses and unfair contests matter...
f but parameters of lottery of violence do not.
⇒ Motives instead of opportunistic behaviour
⇒ Policy implication: Electoral and institution-building support
may be more effective than deterrence.
Next step:
• World-wide sample
• Analyse impact of 3rd generation UN peace-building
16 / 18
Bibliography of presentation
Cederman, Lars-Erik, Gleditsch, Kristian S., & Halvard Buhaug
(2013) Inequality, Grievances, and Civil War. Cambridge, UK:
Cambridge University Press.
Fearon, James (1995) "Rationalist Explanations for War."
International Organization 49(3): 379-414.
Fischer, Jeff (2002) "Electoral conflict and violence." IFES White
Papers - Making Democracy Work.
Gurr, Ted R. (1970) Why men rebel. Princeton University Press.
Norris, Pippa, Richard W. Frank & Ferran Martinez i Coma (2015)
Contentious Elections: From Ballots to Barricades. New York:
Routledge.
17 / 18
Bibliography for presentation
Przeworski, Adam (2014). Acquiring the Habit of Changing
Governments Through Elections. Comparative Political Studies
00(00): 1-29.
Rustow, Dankwart (1970). Transitions to democracy: Toward a
dynamic model. Comparative Politics 2(3): 337-363.
18 / 18