post-accord elections, power sharing and conflict
DESCRIPTION
Election. Power sharing. conflictTRANSCRIPT
Post-Accord Elections, Power Sharing and Conflict
© Anna K. Jarstad
Department of Peace and Conflict Research, Uppsala University
Draft date 19 March 2008
Comments are welcome!
AbstractHow does power sharing influence conflict after post-accord elections? In order to enable ananalysis of the relationship between power-sharing agreements, elections and conflict, Ipresent a new data collection, PAE (Post-Accord Elections). Initial analysis of the data showsthat power sharing plays an important role in relation to the first post-accord election.Parliamentary elections that are preceded by political power sharing as stipulated in a peaceaccord are followed by peace to a higher extent than elections not preceded by power sharing.Elections that take place after power sharing are also considered ‘free and fair’ to a higherextent than elections conducted without previous power sharing. In addition, this paper showsthat a period of power sharing does not mean that that former rebels automatically gain votesin the future. On the contrary, there are cases were a regime change has taken place after aperiod of power sharing. The research presented in this paper is in line with the hypothesisthat power sharing reduces the uncertainties involved in a peace process by ensuring includedparties a share of political power and thereby reducing the predicaments of post-accordelections.
Paper prepared for the 49th Annual Convention of the International Studies Association in
San Francisco 26-29 March 2008.
2
IntroductionThe first election after a peace agreement can start off the process of democratization.
However, there is an increasing appreciation of the obstacles involved in democratization in
war-torn societies. Elections are often postponed for security reasons. When elections are
held, they are often fraudulent, and marred by violence and boycotts. Former warring parties
frequently play a key role in war-shattered societies and the very issues that caused the
conflict are often reflected in the electoral contestation. An increasing body of literature on
war-torn societies has concluded that elections can even increase the risk of conflict.1
Power sharing has been put forward as a method to overcome the initial obstacles of
governance after a violent conflict. South Africa is a case in point where transitional power
sharing paved the way for majoritarian democracy (Sisk and Stefes 2005). However, there is
scant quantitative research on elections in war-torn societies and we therefore do not know
how common it is that elections are held following peace accords, and whether or not power
sharing is important for the holding, conduct and outcome of such elections. Furthermore,
there have been few attempts to provide a comprehensive theoretical explanation for why
power sharing would facilitate democratization. On the contrary, anecdotal evidence often
suggests that power sharing can have detrimental effects on democratization (Jarstad 2008;
2005; Spears 2000). This means that despite the fact that a great deal has been written on
power sharing and elections respectively, the relationship between the two is understudied.
In this paper, I begin to address this shortcoming in the literature by analyzing new data
on elections and conflict after power-sharing processes in civil wars. This paper serves two
purposes: 1) to present the key variables in a new dataset, PAE (Post-Accord Elections), and
1 See, for instance, (Baldwin 2006; Höglund; Höglund 2008a; Höglund 2008b; Höglund, Jarstad, and Kovacs;
Mansfield and Snyder 1995; Mansfield and Snyder 2001; Mansfield 2005; Maoz and Russett 1993; Oneal,Russett, and Berbaum 2003; Paris 2004; Rosato 2003; Snyder 2000; Thompson and Richard 1997).
3
2) to conduct an initial analysis of these variables. The research is driven by the following
research query: How does power sharing affect conflict after post-accord elections? The PAE
data collection enables an analysis of the relationship between power-sharing agreements,
elections and conflict. Initial analysis of the data shows that power sharing plays an important
role in relation to the first post-accord election. Parliamentary elections that are preceded by
political power sharing as stipulated in a peace accord are followed by peace to a higher
extent than elections not preceded by power sharing. Elections that take place after power
sharing are also considered ‘free and fair’ to a higher extent than elections conducted without
previous power sharing. In addition, this paper shows that a period of power sharing does not
mean that that former rebels automatically gain votes in the future. On the contrary, there are
cases were a regime change has taken place after a period of power sharing. The research
presented in this paper is in line with the hypothesis that power sharing reduces the
uncertainties involved in a peace process by ensuring included parties a share of political
power and thereby reducing the predicaments of post-accord elections. However, more
research is needed to explore the relationship between power sharing and post-accord
elections.
The paper begins by outlining a few theoretical points of departure, including research
questions, definitions and operationalizations. Secondly, the new dataset PAE is briefly
described. Thirdly, I conduct a tentative analysis of the relationship between post-accord
elections and conflict, with a special focus on political power sharing, judgments on ‘free and
fair’ conduct of elections and the fate of rebels in elections. Finally, I summarize the findings.
Facilitating Elections by Pact-MakingSince the 1990s democratization became an integral part of peacebuilding. A large number of
elections have been held each year with the specific purpose of promoting peace. However, in
4
war-torn societies, there is often an increase of violence in relation to elections and in some
cases elections are even followed by armed conflict. This raises questions concerning how the
risks involved in elections can be reduced.
Power sharing has been described as a conflict management device to end conflicts, but
has at the same time been seen as an alternative to, or a constraint on democracy. Recent
events in contemporary conflict locations such as Sudan, DR Congo, and Kosovo have again
raised questions regarding the usefulness of power sharing as an instrument of ending civil
wars and paving the way for democracy. On the other hand, the case of Kenya shows that
peace accords with the explicit purpose of ending conflict through power sharing continues to
be a preferred remedy in internationally mediated conflict resolution. While power sharing
continues to be applied in a variety of contexts, we have yet not developed a sufficient
understanding of the processes involved to determine when power sharing can work.
Advocates of power sharing suggest that it can be of relevance in three different
contexts: 1) in transitions to democracy, 2) in divided societies, and 3) in war-torn countries.
To begin with, elite-negotiated pacts have been suggested as an important explanation for
successful transition to democracy in countries such as Spain, Poland, Hungary and
Czechoslovakia. A transition towards democracy means the opening up of political space and
competitive politics. Typically, political parties are formed and they seek popular support in
elections. Under such conditions, both the old political elite and the opposition leaders risk
ending up as a minority political party, and even become excluded from government. A
political pact is an agreement among a select group of elites which seek to define, or redefine,
rules regarding “the exercise of power on the basis of mutual guarantees for the ‘vital
interests’ of those entering into it” (O'Donnell and Schmitter 1986, 37). In this way, political
pacts guarantee inclusion, regardless of election results or as an alternative to elections.
5
Hence, pact-making is expected to create a stable environment and limit the uncertainty
during the transition.
Political pacts have also been suggested to be of importance in divided societies. In
Arend Lijphart’s theory on consociationalism, political pacts are the key to make democracy
function in societies divided along ethnic lines. Such guaranteed inclusion of the elites of all
major rival groups is deemed necessary for democracy to prevail. Lijphart suggests that under
conventional democracy, the minority risks being permanently excluded. Where people vote
along ethnic lines, political parties representing ethnic minorities have no chance of ever
forming a majority, and shifting majorities in parliament are therefore unlikely. Majority rule
in divided societies then spells “majority dictatorship” (Lijphart 1985, 102). Also under PR
(Proportional Representation), the prospects for ethnically defined minorities to gain political
power are limited. The only possibility for minority parties to gain access to positions in the
executive, is to form a coalition with larger parties. Although this is difficult even under
normal circumstances, in divided societies cross-ethnic parties or coalitions tend to receive
little popular support (Reilly 2001).
There is an increasing body of quantitative studies on power sharing and civil war. This
strand of research suggests that political power sharing defined as guaranteed inclusion in
government is less effective than other solutions, such as territorial and military power
sharing, in promoting durable peace. Several studies have analyzed the effects of promises of
power sharing in negotiated settlements on durable peace (e.g.,Hartzell and Hoddie 2003;
Hartzell 1999; Hoddie and Hartzell 2005; Walter 2002). Few studies have analyzed whether
provisions for political power sharing are indeed carried out. A recent study, however,
demonstrates that while implementation of military and territorial provisions are associated
with peace, the implementation of political pacts does not increase the prospects for peace
6
(Jarstad and Nilsson forthcoming 2008). The present paper focuses on another issue, namely
the first elections conducted after peace agreements. It gives rise to a set of questions: How
common is it that elections are held after a peace agreement has been signed? Are such
elections frequently followed by conflict? Is there a higher risk that conflict resumes if the
former rebels lose the elections? Does power sharing affect the likelihood of conflict
following elections?
The PAE data collectionA new dataset will enable analysis of the questions raised above. In order to study elections
held after civil conflict, I introduce a new dataset – PAE (Post Accord Elections) – that
contains unique information on elections conducted after peace agreements struck during the
post-Cold War period. This dataset includes information on post-accord national elections
(presidential and legislative) as well as information on local elections held in new entities
following a settlement on territorial devolution. To date, the dataset covers elections
following all 82 full or partial peace agreements signed during 1989–2004.2
The dataset
includes detailed information on the first election (of each of the three categories included)
held after the peace agreement, but also records any subsequent elections. However, this
paper only focuses on the first post-accord parliamentary elections. PAE also includes data on
conflicts from UCDP (the Uppsala Conflict Data Program) until the end of 2007, and on
power-sharing provisions from two previous data collections TOPAD (Terms of Peace
Agreements Data) and IMPACT (Implementation of Pacts).3
2 Information on the peace agreements is based on the original texts available from the UCDP:
http://www.pcr.uu.se/research/UCDP/3 TOPAD includes information on political, military and territorial pacts in all peace agreements signed in the
period 1989–2004 (Nilsson, Svensson, and Sundberg 2006). IMPACT contains unique information on theimplementation of power-sharing pacts in peace accords in the post-Cold War period (Jarstad, Nilsson, andSundberg 2006). Data has been coded by Ralph Sundberg.
7
Table 1. Post-Accord Elections 1989–2006
Type Number of elections held
Presidential elections 31
Parliamentary elections 39
Local elections in new entity 16
Source: PAE dataset, version 11 March 2008.
Key VariablesThe dataset includes several variables related to post-accord elections. All cases include the
experience of armed conflict and at least one peace agreement. However, not all of these
cases are post-conflict cases, since the signing of a peace agreement does not always mean
that violent conflict ends. On the contrary, there are cases where armed conflict continues
throughout the studied period. Elections Held captures whether or not parliamentary elections
were held within five years after the signing of the agreement. There are two dependent
variables that include information on post-election conflict during the period 1989–2007. The
dataset only includes so-called ‘state-based’ conflicts, i.e. conflicts in which at least one
warring party is the government. Hence, the dependent variables do not capture organized
one-sided violence (violence directed by the government or a rebel group against civilians) or
non-state conflicts (where two organized groups, neither one being the state, engage in
violence against each other). The first dependent variable, Post-Election Conflict, records if
any party in the country engages in violent armed conflict that reaches at least 25 battle-
related deaths during a period after the election. Conflict is reported from they day after the
election and the remaining period the same calendar year as the holding of the elections and
the calendar year following on the elections. This variable includes signatories and non-
signatories to a peace deal, and also the conflict behavior of warring parties in a different
8
conflict in the same country. The dependent variable Signatory Post-Election Conflict records
if armed conflict between the signatories to a peace agreement resumes (or simply continues),
or whether peace prevails, following national elections. Conflict is considered to take place
when the government and one or more of the rebel groups that have signed the peace
settlement, or a splinter faction of a signatory group, engage in armed conflict that reaches at
least 25 battle-related deaths from the day after the election until the end of the calendar year
following the election.
In this study, I begin to study factors that potentially influence conflict after post-accord
elections: the type of peace agreement (provisions for political power sharing), the conduct of
elections (‘free and fair’), and the outcome of elections (rebel inclusion or rebel exclusion). A
peace agreement is a formal document, signed by the government and one or more rebel
group, which addresses the contested issue (incompatibility) by settling all or part of it.4
Power Sharing denotes peace agreement provisions for political power sharing, in line with
Barbara Walter’s definition of political pacts which have “offered the combatants guaranteed
positions in the new government at the level of cabinet or above, or a specific quota of
political power in at least one of the main branches of government” (Walter 2002). In
addition, power sharing is here only considered to be in place if the stipulation indeed has
been implemented, according to all of the following three criteria: (1) all named groups and
individuals have claimed their seats in government, (2) the assembly has been inaugurated,
and (3) the institution of joint governance has begun to function (Jarstad, Nilsson, and
Sundberg 2006). This means that, in contrast to most quantitative studies on the topic, power
sharing here only refers to cases which have had the experience of joint rule.
4 This study does not include so called process agreements, i.e., agreements that merely outline a process for
how to resolve the incompatibility, or agreements that are to be confirmed in a forthcoming agreement.
9
Free and Fair refers to whether the election was considered free and fair by external
observers. This coding is based on a variety of sources, ranging from Freedomhouse.org to
electoral observation missions for specific elections. It is coded as 1 if the elections were
declared free and fair, and coded as 0 if they were not. Some elections are coded as 2,
denoting that the elections were certified but were deemed by observers to have serious
irregularities or flaws.
The outcome of the elections, particularly the fate of the rebels – Rebel Outcome –
could also be important for peace. In relation to parliamentary elections, it is recorded
whether rebels, or a political party associated with a former rebel group, win seats in
parliament (Rebel Inclusion) or if the rebels are defeated or did not participate in the elections
(Rebel Exclusion).5
Patterns of Post-Accord ElectionsBetween 1989 and 2004 a total of 82 peace agreements were signed in 37 different countries.
The data shows that 55 out of these peace agreements stipulated that elections were to be held
on the national level (president and/or parliament). 16 peace agreements included provisions
for local elections as part of a territorial power sharing deal, i.e. stipulated that elections were
to be held to a new entity of devolved governance. This means that the vast majority of peace
agreements provide for elections. A first question is then whether elections have a positive
effect on peace.
5
Defining who is a rebel or rebel-associated party is a problematic, see Söderberg Kovacs (Söderberg Kovacs2007). In the code book for the PAE dataset the following examples are given: “Sometimes rebel groupstransform into political parties after a conflict has ended. If this is the case in a specific conflict the codingdenotes the performance of this political party. Examples of such parties are the Revolutionary United Front’sRUFP (Revolutionary United Front Party), and the National Patriotic Front of Liberia’s NPP (National PatrioticParty). At times rebel groups are also closely linked to political parties that act within the democratic sphere of acountry. If this is the case the coding denotes the performance of this political party. An example of such arelationship can be found in Congo, where the different militias have clear links to ‘normal’ political parties,such as the UPADS party’s support of the Cocoyes militia” (Jarstad, Nilsson, and Sundberg 2006).
10
In some countries these agreements introduce democracy for the first time. However,
even when stipulated in peace agreements, elections are not always held. In other war-torn
societies, elections are held without a provision in a peace accord. Procedures of democratic
governance sometimes continue throughout the conflict and the following peace process. This
was the case in for example Chad and Mexico. In order to cover all of the first elections held
after a peace agreement, we thus also need to look into cases where peace agreements have
been signed but where these do not include explicit stipulations of elections.
Despite the predicaments involved in holding elections in war-torn societies, elections
were held in a vast majority of all countries where a peace agreement has been signed. In 33
countries parliamentary elections were held within five years after the signing of the peace
agreement. The analysis includes a total of 39 parliamentary post-accord elections conducted
in these 33 countries.6 Almost half of the elections – 17 out of 39 – were followed by conflict.
This could indicate that elections in war-torn societies trigger conflict. A fairly short period
has been chosen for the analysis of conflict following elections in order to limit the risk of
including other triggering events that could lead to conflict. Nevertheless, the coding also
captures conflicts that are on-going before the elections, and territorially based conflicts that
might be less affected by national elections than conflicts over government. Therefore this
data should be interpreted with caution. At least, it seems reasonable to conclude that
elections do not in general prove to be effective as a way to end or prevent new conflicts in
countries already shattered by conflict.
However, after 31 out of 39 parliamentary elections there was no conflict between the
signatories to the peace agreement following the election during the period from the day after
the election until the end of the following calender year. This suggests that peace agreements
6 These 39 elections denote the first parliamentary election held after a peace agreement. In addition, it has been
recorded whether subsequent elections were held.
11
are indeed important for partial peace following elections.7 Signatories were engaged in
conflict after 8 elections. This means that 20,5 percent of the elections were followed by post-
election conflict among the signatories.
It should also be noted that there has been armed conflict in most of the few cases
where elections have not been held during the analyzed period. In Cote D´Ivoire, there has
been one-sided violence, but no armed conflict as defined by UCDP is recorded. As the
dataset only covers cases in which peace agreement were signed 1989–2004, the 2004 peace
agreement concerning Cote D´Ivoire is the latest one included in the dataset. Elections have
yet not been held (but five years have not yet passed). In Afghanistan there has been armed
conflict throughout the period included in the dataset. For this conflict, three agreements are
included, two signed in 1993 and one in 1996. No elections where held within the five-year
period.8 Also after the agreements in Rwanda 1993 and Somalia 1993 and 1997 the
signatories were engaged in conflict and no elections were held within the five-year period.
This means that refraining from elections does not in general provide for peace. A summary
of the overview of elections, peace agreements and conflict is presented below.
Table 2. Post-Accord Elections and Conflict
Peace agreements: 82
Agreements that stipulate national elections: 55
Agreements that stipulate local election in a new entity: 16
Countries with peace agreements: 37
Countries holding post-accord parliamentary elections: 33
Post-accord parliamentary elections: 39
Conflict following the first post-accord parliamentary election: 17
Signatory conflict following the first post-accord parliamentary election: 8
7 This is in line with research conducted by Desirée Nilsson. Her research focuses on peace agreements and does
include elections (Nilsson 2006; Nilsson forthcoming 2008).8 The Bonn agreement is not considered as a peace agreement by UCDP. Therefore the 2005 elections is not
included as post-accord elections in the datset.
12
Power Sharing and Post-Accord ElectionsIt has been suggested that power sharing can ease the transition from war to democracy.
While there is a large academic discourse on power sharing and elections, thus far no global
quantitative study has analyzed the relationship between the two. I begin to address this topic
by mapping out the general pattern of parliamentary elections following peace agreements
that include provisions for political power sharing: what role does power sharing play in
regard to elections?
To recall, 39 parliamentary post-accord elections are included in the analysis. In 13
cases the provisions were implemented ahead of the elections.9 Among the total cases of
parliamentary elections, 8 were followed by signatory conflict (20,5 percent). When
comparing elections that have been preceded by political power sharing (13) with elections
that were held without such power sharing (26), I find that only one election preceded by
power sharing was followed by signatory conflict (8 percent), whereas 7 out of the 26
elections without power sharing were followed by signatory conflict (27 percent). This
indicates that power sharing indeed plays an important role for warring parties that have
signed a peace agreement to stick to peace, even after elections.
Even more interestingly, power sharing might be important for overall peace. There was
no conflict the year after elections in 9 out of 13 power sharing cases (69 percent), whereas 22
out of 39 (56 percent) of the post-accord parliamentary elections that were not preceded by
power sharing were followed by conflict. This means that not only do the signatories stick to
9 15 of these elections were preceded by peace agreements stipulating some form of political power sharing
agreement. In 13 out of these 15 cases political power sharing was implemented (Sierre Leone is here consideredto be a case of implemented power sharing, although the latest power sharing agreement was not implemented.No conflict is recorded within the analyzed period.) The two cases where political power sharing was neverimplemented are Mexico and Uganda. Whereas no conflict followed the elections in Mexico, in Uganda, LFAcontinued fighting during 2007. In total, there are nine parliamentary elections preceded by implemented powersharing were peace holds.
13
peace to a higher extent in power sharing processes, but there are also fewer cases where
other actors resort to conflict following elections preceded by power sharing. More research is
needed to find out if there is any causal relationship between power sharing and overall peace,
and if so which are the causal mechanisms at play.
Table 3 . Post-accord parliamentary electionsElections held ’on time’ Signatory Conflict Any Conflict
Elections preceded bypolitical power sharing
13 1 4
Elections without politicalpower sharing
26 7 13
Total 39 8 17
Both elections and political power sharing are tools that can be used to regulate conflicts over
governance. It is plausible that these measures are less effective in managing territorial
conflicts. It can therefore be of interest to focus on the cases that include conflicts over
government and exclude territorial conflicts. In total 25 parliamentary post-accord elections
have been conducted after peace agreements signed in government conflicts during the period
studied.10 13 of these were preceded by power sharing, whereas 12 were not. It is more
common that elections are held when the peace agreement does not provide for political
power sharing.11 However, elections are more frequently followed by conflict when they are
not preceded by powersharing: Among those elections that indeed are held there is only one
case of signatory conflict after elections following a period of power sharing, compared to
four cases of signatory conflict following elections without power sharing. How should this
10
Government conflict here includes all government conflicts according to the definition of UCDP, and alladditional conflicts in which peace agreements providing for political power sharing has been struck.11
There are 10 government conflicts which include peace agreements without provisions for power sharing. Inall of these elections have been held. 15 government conflicts include peace agreements stipulating politicalpower sharing, but in 2 of these cases the provisions were not implemented at all. There are 3 countries were
14
be interpreted? It is conceivable that power sharing can provide opportunities for all included
groups to influence the peace process, including the design of the new rules of governance
and also the date for elections and its conduct. It could therefore be the case that power
sharing ensures that elections only take place when the warring parties have reached a
minimal level of cooperation and commitment to the peace process.
Does this also mean that power sharing has a positive influence on the conduct of
elections? The data shows that elections conducted after power sharing are deemed free and
fair to a higher extent than other elections.12 It is possible that the administration of the
elections and conflict managing bodies are more inclusive during power sharing governments,
but this remains to be studied. Another possible explanation could be that there is a greater
international involvement in power sharing processes than in other war-torn societies. This
could have a positive effect on the security situation in relation to elections. Another
possibility is that the classification ‘free and fair’ is a result of political expediency rather than
objective criteria.13 Hence, this issue warrants more research.
Table 4. Post-accord parliamentary elections in conflicts over governmentElections held ’on time’ Free and fair Signatory Conflict
Elections preceded bypolitical power sharing
13 11 1
Elections without politicalpower sharing
12 7 4
Total 25 18 5
political power sharing was implemented, but where election have not been held within the five-year period afterthe signing of the latest peace agreement struck the period 1989–2004.12
The number of free and fair first post-accord elections not preceded by political power sharing after all typesof conflicts is 17 out of 26.13
This critique was recently raised by the Human Rights Watch in its release of the World Report 2008, seehttp://hrw.org/englishwr2k8/docs/2008/01/31/usint17940.htm.
15
In most cases, elections are about winning or losing political power. However, political power
sharing guarantees inclusion in the national government. Does this also mean that parties to a
power sharing deal make sure that they will not be completely defeated in an election? The
data shows that this is not the case. On the contrary, there are more cases of rebel exclusion as
a result of post-accord parliamentary elections following power sharing (5) compared to rebel
exclusion after elections without previous power sharing (3).14 In the only case of post-
election conflict, Cambodia, the rebels were included as a result of the post-accord elections.
Somewhat surprisingly, there is no power-sharing case where the excluded rebels resort to
conflict. On the contrary, there are four cases where rebels are defeated in the elections, but
where there is nevertheless no post-election armed conflict among the signatories during the
period after the elections. The relevant elections are Bangladesh 2001, Guinea-Bissau 1999,
Liberia 2005, Niger 1997 and Sierra Leone 2002.15 Possible reasons for the rebels’ acceptance
of an electoral defeat could be that they have disarmed, or lack other sufficient support to
restart a conflict, but further research is needed on these cases in order to understand the
causal mechanisms involved. The fact that elections are held after post-accord power sharing,
and that they are seldom followed by conflict, suggests that power sharing might play a role
in mitigating the risk of post-accord elections.
Concluding remarksSeveral findings have been made in the initial analysis of post-accord elections presented in
this paper. A first set of findings regard the prevalence of post-accord elections. It is common
that the issue of democratic elections is included in the processes of promoting peace and
establishing a new political order: 55 out of 82 peace agreements stipulate national elections.
14
The number of rebel exclusion as a result of the first post-accord elections not preceded by political powersharing after all types of conflicts is 16 out of 26.
16
Elections were also carried out in the vast majority of countries – 33 out of 37 – within five
years after a peace agreement had been signed. A second finding is that post-accord elections
do not in general provide for peace: 17 out of the 39 parliamentary elections were followed by
armed conflict. Thirdly, power sharing seems to affect the risk of conflict following elections.
While most of the post-accord parliamentary elections where held without previous political
power sharing, there was only one case of conflict between the signatories following national
elections among the 13 cases where power sharing had been stipulated in a peace agreement
and where it had also been implemented. Furthermore, elections preceded by power sharing
were consider free and fair to a higher extent than elections without previous power sharing.
At this stage, it is not possible to draw any conclusions from this particular finding. Finally,
exclusion of rebels as a result of electoral defeat or abstention was found to be more common
among elections preceded by power sharing, compared to cases without power sharing. Also
this issue warrants more research.
In sum, this study suggests that there is a co-variation between power sharing and the
absence of armed conflict after post-accord elections. How are we to understand these results
in relation to previous research on this topic? At first glance, these results seem to contradict
previous studies, which have found that peace agreement provisions for political power
sharing does not significantly affect the durability of peace (Hoddie and Hartzell 2005). A
recent study shows that such provisions do not provide for peace even when they are
implemented (Jarstad and Nilsson forthcoming 2008; Jarstad and Sundberg 2007). However,
while previous studies analyze the impact of peace agreements, this study focuses on
elections. A closer look at the cases shows that in most instances several attempts at conflict
resolution have been made before elections eventually are held. It is also common that several
15 In Guinea-Bissau, there was a coup attempt that did not reach the level of 25 deaths. In Niger there are otherconflicts after the election in 1997, but not involving the signatories. Uppsala Conflict Data Program (Date of
17
peace agreements are reached before peace finally holds. However, deeper case studies are
needed to found out if there is any causal relationship between any of the factors related to
power sharing, elections and peace.
retrieval: 2008/03/19) UCDP Database: www.ucdp.uu.se/database, Uppsala University.
18
ReferencesBaldwin, Clive. 2006. Minority Rights in Kosovo under International Rule: Minority Rights
Group International.Hartzell, C., and M. Hoddie. 2003. Institutionalizing Peace: Power Sharing and Post-Civil
War Conflict Management. American Journal of Political Science 47 (2):318-332.Hartzell, Caroline A. 1999. Explaining the Stability of Negotiated Settlements to Intrastate
Wars. Journal of Conflict Resolution 43 (1, February):3-22.Hoddie, Matthew, and Caroline A. Hartzell. 2005. Power Sharing in Peace Settlements:
Initiating the Transition from Civil War. In Sustainable Peace. Power and Democracyafter Civil War, edited by P. G. Roeder and D. Rothchild. Ithaca and London: CornellUniversity Press.
Höglund, Kristine. Electoral Violence: Causes, Concepts and Consequences. unpublishedpaper.
Höglund, Kristine. 2008a. Peacemaking in the Shadow of Violence. Leiden: Martinus NijhoffPublishers/Brill Academic Publisher.
Höglund, Kristine. 2008b. Violence in War-to-Democracy Transitions. In War-to-DemocracyTransitions: Dilemmas of Peacebuilding, edited by A. K. Jarstad and T. D. Sisk.Cambridge: Cambridge University Press.
Höglund, Kristine, Anna Jarstad, and Mimmi Söderberg Kovacs. The Predicaments ofElections in War-Torn Societies. unpublished paper.
Jarstad, Anna K. 2008. Power Sharing: Former Enemies in Joint Government. In War-to-Democracy Transitions: Dilemmas of Peacebuilding, edited by A. K. Jarstad and T. D.Sisk. Cambridge: Cambridge University Press.
Jarstad, Anna, and Desirée Nilsson. forthcoming 2008. From Words to Deeds: TheImplementation of Power-Sharing Pacts in Peace Accords. Conflict Management andPeace Science 25 (3).
Jarstad, Anna, Desirée Nilsson, and Ralph Sundberg. 2006. The IMPACT (Implementation ofPacts) Dataset Codebook. Uppsala: Department of Peace and Conflict Research,Uppsala University.
Jarstad, Anna, and Ralph Sundberg. 2007. Peace by Pact: Data On the Implementation ofPeace Agreements. In Globalization and Challenges of Building Peace, edited by A.Swain, R. Amer and J. Öjendal. London: Anthem Press.
Lijphart, Arend. 1985. Power-Sharing in South Africa. Berkeley: Institute of InternationalStudies, University of California.
Mansfield, E.D., and J. Snyder. 1995. Democratization and the Danger of War. InternationalSecurity 20 (1):5-38.
Mansfield, E.D., and J. Snyder. 2001. Democratic Transitions and War. From Napoleon to theMillennium's End. In Turbulent Peace. The Challenges of Managing InternationalConflict, edited by C. A. Crocker, F. O. Hampson and P. Aall. Washington D.C.:United States Institute of Peace Press.
Mansfield, Edward.D. 2005. Electing to Fight. Why Emerging Democracies Go to War.Cambridge, Massachusetts
London, England: MIT.Maoz, Zeev, and Bruce M. Russett. 1993. Normative and Structural Causes of Democratic
Peace 1946–1986. American Political Science Review 87 (3):624–638.Nilsson, Desirée. 2006. In the Shadow of Settlement. Multiple Rebel Groups and Precarious
Peace, Department of Peace and Conflict Research, Uppsala University, Uppsala.
19
Nilsson, Desirée. forthcoming 2008. Partial Peace: Rebel Groups Inside and Outside of CivilWar Settlements. Journal of Peace Research 45 (4).
Nilsson, Desirée, Isak Svensson, and Ralph Sundberg. 2006. The Terms of Peace AgreementsData (TOPAD) Codebook: Department of Peace and Conflict Research, UppsalaUniversity.
O'Donnell, Guillermo, and Philippe Schmitter, eds. 1986. Transitions from AuthoritarianRule: Tentative Conclusions about Uncertain Democracies, Transitions fromAuthoritarian Rule. Baltimore: Johns Hopkins University Press.
Oneal, John R., Bruce M. Russett, and Michael L. Berbaum. 2003. Causes of Peace:Democracy, Interdependence, and International Organizations, 1885-1992.International Studies Quarterly 47 (3):371-393.
Paris, Roland. 2004. At War's End: Building Peace After Civil Conflict. Cambridge:Cambridge University Press.
Reilly, Benjamin. 2001. Democracy in Divided Societies. Electoral Engineering for ConflictManagement. Cambridge: Cambridge University Press.
Roeder, Philip G., and Donald Rothchild, eds. 2005. Sustainable Peace. Power andDemocracy after Civil Wars. Ithaca and London: Cornell University Press.
Rosato, Sebastian. 2003. The Flawed Logic of Democratic Peace Theory. American PoliticalScience Review 97 (4):585-602.
Sisk, Timothy D., and Christoph Stefes. 2005. Power Sharing as an Interim Step in Peacebuilding: Lessons from South Africa. In Sustainable Peace. Power and DemocracyAfter Civil Wars, edited by P. G. Roeder and R. Donald. Ithaca and London: CornellUniversity Press.
Snyder, J. 2000. From Voting to Violence. Democratization and Nationalist Conflict. NewYork, London: W. W. Norton & Company.
Söderberg Kovacs, Mimmi. 2007. From Rebellion to Politics. The Transformation of RebelGroups to Political Parties in Civil War Peace Processes. Uppsala: Department ofPeace and Conflict Research, Uppsala University.
Spears, Ian S. 2000. Understanding inclusive peace agreements in Africa: the problems ofsharing power. Third World Quarterly 21 (1):105-118.
Thompson, William R., and Tucker Richard. 1997. A Tale of Two Democratic PeaceCritiques. Journal of Conflict Resolution 41 (3):428-454.
Walter, Barbara F. 2002. Committing to Peace. The Successful Settlement of Civil Wars.Princeton: Princeton University Press.