Content uploaded by Benjamin Reilly
Author content
All content in this area was uploaded by Benjamin Reilly on May 10, 2016
Content may be subject to copyright.
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=finp20
Download by: [Murdoch University Library] Date: 09 May 2016, At: 23:43
International Peacekeeping
ISSN: 1353-3312 (Print) 1743-906X (Online) Journal homepage: http://www.tandfonline.com/loi/finp20
Elections in Post-Conflict Scenarios: Constraints
and Dangers
Benjamin Reilly
To cite this article: Benjamin Reilly (2002) Elections in Post-Conflict Scenarios: Constraints and
Dangers, International Peacekeeping, 9:2, 118-139, DOI: 10.1080/714002729
To link to this article: http://dx.doi.org/10.1080/714002729
Published online: 08 Sep 2010.
Submit your article to this journal
Article views: 1167
View related articles
Citing articles: 8 View citing articles
Post-Conflict Elections:
Constraints and Dangers
BENJAMIN REILLY
Elections have become an integral element of many United Nations
peacekeeping missions over the past decade. The reason for this is
clear: the focus of most UN missions has shifted from one of pure
peacebuilding to one of state re-building or, in some cases like East
Timor, state creation. In such cases, elections provide an inescapable
means for jump-starting a new, post-conflict political order; for
stimulating the development of democratic politics; for choosing
representatives; for forming governments; and for conferring
legitimacy upon the new political order. They also provide a clear
signal that legitimate domestic authority has been returned and hence
that the role of the international community may be coming to an end.
For all of these reasons, elections have become a central part of many
UN peacekeeping missions. In addition, electoral assistance outside
peacekeeping missions has become something of a growth industry
since the fall of the Berlin Wall and the third wave of democratization
has led to a threefold increase in the number of putatively democratic
governments around the globe.
Despite this, there has been a considerable variation in the relative
success of elections in meeting the broader goals of democratization
from country to country and case to case. In some cases, such as
Namibia and Mozambique, elections clearly played a vital role in
making a decisive break with the past. In others, such as Angola, flawed
elections created more problems than they solved. In Haiti,
administrative inefficiencies undermined the credibility of the broader
electoral process. By contrast, in Cambodia, technically successful
electoral processes were soon overwhelmed by the realities of power
politics. And in Bosnia, premature elections helped to kick-start the
façade of democratic politics, but also helped nationalist parties cement
an early grip on political power. While this mistake has been avoided
in Kosovo and East Timor, it is still to be seen how elections influence
the broader process of peacebuilding in these two critical cases.
What is clear, however, is that in any UN mission, the holding of
elections forces critical political choices to be made. Elections
92ip08.qxd 25/06/2002 12:46 Page 118
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
represent a key step in a broader process of building political
institutions and legitimate government. Elections influence the extent
to which the internal politics of fragile new states become stabilized,
whether the new political dispensation comes to be viewed as
legitimate, and how the rhythm of peaceful democratic politics can
evolve and become sustainable. Variations in electoral procedures can
also play a key role in determining whether the locus of political
competition evolves along extremist or centrist lines, and hence in
developing moderate and broad-based political parties.
There are three main areas of variation that are crucial influences on
the shape of post-conflict politics in most countries. First, there is the
question of timing: should post-conflict elections be held as early as
possible, so as to fast-track the process of establishing a new regime, or
should they be postponed until peaceful political routines and issues
have been able to come to prominence? Second, there is the mechanics
of elections themselves: who runs the elections; how are voters
enrolled; what electoral formula is used. Third, there is the often
underestimated issue of the effect of the elections on political parties.
Especially in cases of weak civil society, political parties are the key link
between masses and elites, and play an absolutely crucial role in
building a sustainable democratic polity. Hence, the interaction between
parties and the electoral process is itself crucial. Are the political parties
contesting the election narrow, personalized, sectarian or ethnically
exclusive entities, using the political process to pursue their wartime
objectives? Or are they broad, multi-ethnic, programmatic
organizations with real links to the community? How can the former be
discouraged and latter promoted?
More generally, there is the overarching issue of under what
circumstances elections help to build a new democratic order, and
under what circumstances they can undermine democracy and pave the
way for a return to conflict. For example, elections are part of the
broader process of democratization, but ill-timed, badly designed or
poorly run elections can actually undermine the broader process of
democratization. This is the overarching theme of this essay.
Timing
As a starting point, the issue of election timing is a crucial, and under-
appreciated, variable in election planning. Issues of timing also directly
affect administrative choices, electoral system designs and the way
political parties form. For example, in some cases timing demands
particularly the need to hold a quick election have influenced the
119
POST-CONFLICT ELECTIONS
92ip08.qxd 25/06/2002 12:46 Page 119
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
choice of electoral laws, and these have affected not just the party
system but also the broader incentives presented to political actors as
part of the election process. Take the case of Angolas 1992 presidential
elections held under the Bicesse Agreement of May 1991 as part of the
peace process aimed at stopping Angolas long-running civil war. The
major parties contesting the election were the political wings of two
former liberation movements turned armies: the governing MPLA, led
by President Eduardo Dos Santos, and UNITA, led by Jonas Savimbi.
Due to the extraordinary nature of the election (the first ever held in
Angola) and severe timing pressures, a hastily-drafted electoral law was
enacted which included, as part of the presidential election, a run-off
between the top two candidates if no-one gained a majority in the first
round of voting.
This choice of formula had two impacts. First, it precluded any
possibility of power sharing between the two main combatants, as
the election itself could only be won by one candidate. Second, it
provided an escape hatch for parties weakly committed to the process,
who could get an indication of their support levels after the first
round of voting. When Savimbi realized after the first round that he
was unlikely to win the election, he rejected the election and went
back to war. The issues of timing and electoral system choice
thus impacted directly on the overall failure of the peace process in
Angola. Of course, it is possible that this may have occurred anyway.
But the design of the electoral system clearly presented strategic
opportunities for candidates to remove themselves from the contest
an incentive that would have been lower under a different set of
institutional rules.
Such events may suggest that democracy itself is part of the problem
in such highly fraught situations, and that post-conflict situations are
too fragile to be exposed to the competitive pressures of the electoral
process. But this oft-heard critique ignores several factors. First,
elections can be purposively designed to encourage not winner-takes-
all outcomes, but the sharing of power between groups. Indeed, many
would argue that some form of power-sharing is a primary requirement
in post-conflict situations. Second, critics of elections as instruments of
democratization often ignore the real need to construct a legitimate
governing authority in post-conflict circumstances. Not least because
so many of todays conflicts take place within states, the overarching
challenge of many UN missions is to build or re-build a sustainable
democratic state that can function without direct international
involvement. Elections are a crucial element in achieving this. State-
building has been a priority issue in both Kosovo and East Timor, for
120 RECOVERING FROM CIVIL CONFLICT
92ip08.qxd 25/06/2002 12:46 Page 120
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
example, where UN missions were confronted with the challenges of
attempting to build functioning democracies in societies only recently
ravaged by violent conflict and suffering a history of incorporation and
oppression by powerful neighbours.
One valid criticism of elections in post-conflict scenarios, however,
is that if held too early, they can undermine the nascent democratic
order. This has been a fundamental problem of many UN-supervised
elections: they have been held too soon and too quickly after peace has
been restored. In fact, over the last decade, UN peacekeeping missions
appear to have developed a kind of standard operating procedure.
Once a minimum level of peace has been obtained (which does not
necessarily mean a full ceasefire agreement), and a basic level of
infrastructure is in place, the next step is usually to hold some kind of
parliamentary elections often within a year or two of the start of the
mission followed by a rapid hand-over to the newly-elected
authorities, and an even more rapid departure of UN troops and
personnel. Thus in Kosovo there was strong pressure on the OSCE, the
body tasked with organizing elections in the region, to hold elections
as quickly as possible, regardless of whether the social conditions that
exist there are conducive to the cut and thrust of open electoral politics
or not. A similar pressure to hold instant elections was present in East
Timor, where the UN Transitional Administrator, Sergio de Mello,
argued that postponing elections beyond 2001 would be difficult,
arguing that you cant hold back the horses of political development
indefinitely. In both cases, national elections were held in the second
half of 2001.
But if held too early, elections in fragile situations can easily
undermine the longer-term challenge of building a sustainable
democracy. Elections in conflictual situations can act as catalysts for the
development of parties and other organizations which are primarily
(and often solely) vehicles to assist local elites gain access to governing
power. They can promote a focus on regional, rather than national,
issues. They can serve to place in positions of elected authority leaders
committed to exclusionary visions of the country leaders who are, in
many cases, the very same ones who started or fought in the conflict in
the first place. This generals-to-politicians transformation has been a
recurring problem in the Balkans, where nationalist parties and elites
have attempted to use the political process to continue to press their
sectarian aims. Early elections also tend to elicit more extreme
reactions from voters than an election held after a period of state
rebuilding. This is one of the perverse realities of post-conflict
elections: the sine qua non of the democratic process, elections, can
also be its undoing.
121
POST-CONFLICT ELECTIONS
92ip08.qxd 25/06/2002 12:46 Page 121
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
The timing of elections can also impact directly on the shape of the
political party system, and on the degree of coordination between local
and national-level elites. For example, a major goal in building a
sustainable democracy should be the creation of parties which are
broad-based, have strong links to local communities, and campaign on
a national platform. But in post-conflict situations, many political
parties are not broad-based vehicles for presenting competing policy
and ideological platforms, but rather narrowly-focused, personalized
elite cartels. In other cases, political movements are often merely thinly-
disguised variants of the armies which fought in the original conflicts,
as exemplified in Bosnia by the growth of nationalist parties like the
(Croat) HDZ, (Serb) SDS and (Bosniac) SDA, respectively. This problem
also afflicts former liberation movements, such as East Timors Fretilin,
or the Kosovo Liberation Army, as they attempt to transform themselves
into mainstream political organizations. Either way, holding elections
too early in the transition period can have the perverse effect of
blocking the development of more aggregative and programmatic
political parties institutions which are now widely accepted to be
important facilitating agents for successful democratization.
A second issue is the coordination of election timing with sub-
national elections. Some scholars argue that in a new democracy,
holding national elections before regional elections generates incentives
for the creation of national, rather than regional, political parties1 and
hence that the ideal process of election timing is to start at the national
level and work ones way down. Others such as Diamond believe that
simultaneous national and local elections can facilitate the mutual
dependence of regional and national leaders. The more posts that are
filled at the regional and local level the greater the incentive for
regional politicians to coordinate their election activities by developing
an integrated party system.2This was the situation at Indonesias 1999
elections, with identical party-based ballots being presented to voters at
simultaneous elections for national, provincial and local assemblies,
which greatly strengthened the nascent party system. In recent years,
however, UN practice has been the opposite: to start with municipal
elections and work up, as in Kosovo. This approach is particularly
suited to state-building elections, which can help develop party politics
from the ground up. In general, the comparative evidence suggests that
this bottom-up approach to electoral timing is probably the best way to
encourage the development of party politics and to inculcate voters in
the routines of electoral politics.
A more immediate problem often comes not from the domestic
realm but from the approach taken by the international community
122 RECOVERING FROM CIVIL CONFLICT
92ip08.qxd 25/06/2002 12:46 Page 122
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
itself. International policy makers, not least at the UN, have typically
viewed elections as a punctuation point in a peacekeeping mission,
which can usher in not just a new government but also provide a
convenient exit point for international involvement. Thus Cambodias
1993 election, the culmination of the biggest UN peacekeeping mission
to date, was followed by a rapid departure of the UN and other
international forces from Cambodia a departure which did little to
translate the results of an exemplary electoral process into solidifying
a fragile new polity. Soon after, a coup by the second prime minister,
Hun Sen, against the most popular elected party, FUNCINPEC, saw
Cambodia return to its familiar politics of intimidation and
authoritarian rule. Elsewhere, rushed elections (for example, in
Liberia) with little in the way of broader political support, have
undermined the legitimacy of the election process, creating further
problems for future democracy-building efforts.
The wider obsession in the 1990s with elections as a form of
conflict resolution is perhaps the most obvious manifestation of this
quick-fix mentality. The world is littered with elections, often
conducted at the behest of the international community, which only
served to inflame and politicize the root causes of conflict. Given this,
it is not surprising that elections held too early in the process of state
rebuilding (or premature elections, as they have become known)
often have the opposite results to those intended. The December 1991
Algerian elections, which were aborted after the fundamentalist Islamic
Salvation Front won the first round of voting, and which led to the
suspension of the constitution and the strengthening of military rule,
were one case in point. In Burundi, the June 1993 elections
undermined prospects for democracy by fuelling an ethnic
retribalization of party politics, which inflamed pre-existing
TutsiHutu tensions not just in Burundi but in neighbouring Rwanda
as well. In both cases, early or ill-thought-through elections appeared
to undermine the broader path of democratic development.
There are, however, powerful pressures, both domestic and
international, for early elections to occur as part of the process of state
rebuilding in post-conflict societies. For one thing, given the risk-
averse nature of the international community when it comes to
peacekeeping commitments, such elections can (as noted above)
provide a clear exit strategy for international involvement. But
supporting the difficult process of transforming a poor, traumatized
and war-ravaged society into a well-functioning democracy requires
more than the presence of a few hundred UN officials for 18 months,
with an election at the end. It means, quite simply, being prepared to
123
POST-CONFLICT ELECTIONS
92ip08.qxd 25/06/2002 12:46 Page 123
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
invest substantial time and money in an open-ended process of social
and political development. With the exception of the Balkans, which
benefits from its location in Europe (and where observers are talking
about an international presence in the region for decades), there are
few post-conflict societies anywhere in the world where international
actors have the inclination to pursue such an open-ended strategy. In
most cases, the roving eye of the international media and the major
western governments moves on to other, more fashionable, issues.
A second-best alternative to such open-ended commitment is not to
rush into immediate elections following a peace deal, but rather to
encourage local involvement for a few years until some of the basic
elements of a pluralistic party system and a functioning state have been
established. This was the approach taken by the UN in both East Timor
and Kosovo, where local-level democratization and security-building has
taken precedence over the early holding of a national poll, and where
national consultative bodies of local leaders have been introduced
without an electoral process. In East Timor, for example, the UN
developed a National Consultative Council, made up of representatives
of East Timors government-in-waiting, into a form of unelected
legislature, which included representatives of youth, church and
womens groups. In Kosovo, national elections were postponed in
favour of municipal polls, where the stakes are much lower and the
responsibilities of elected officials were focused on service delivery
rather than national issues. In both cases, the evidence suggests that, by
involving local actors in the process of governing while lengthening out
the transition to full-blown national elections, a more mature and
responsible form of party politics has begun to be developed. This
approach has much to recommend it for future operations.
Electoral Mechanics
The mechanics of the electoral process can have a profound and
often profoundly misunderstood impact on the success or failure of
post-conflict democratization. Electoral mechanics can be divided into
two main areas: the electoral system that is, the formula by which
votes are converted into seats, including the way ballot papers are laid
out and the structure of electoral districts, and the electoral
administration such as the electoral management body, the provisions
for voter registration, boundary delimitation and the like. Between
them, these two areas comprise some of the most important variables
influencing the success or failure of post-conflict elections, and indeed
for democratization more generally.
124 RECOVERING FROM CIVIL CONFLICT
92ip08.qxd 25/06/2002 12:46 Page 124
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
While electoral systems have attracted a voluminous academic
literature, issues of electoral administration remain under-studied by
scholars and under-rated in general in terms of their effect on post-
conflict polities. Voter registration, for example, is a perennial area of
concern, not least because nearly all post-conflict elections take place
in an environment where basic census and other records are missing.
The construction of a comprehensive register of voters is thus often a
first step in the bureaucratic process of state-building. It is also often
an enormously time-consuming, logistically-challenging and resource-
intensive process: in Cambodia, for example, the voter registration
period took almost a full year before the election and demanded huge
amounts of time, personnel and money. Because electoral districts and
polling places are often drawn and allocated on the basis of voter
registration records, this process usually impacts on these areas too.
However, probably the most important administrative decision
concerns the composition of the body managing the elections, and
specifically whether the elections are run by the government of the day
or whether some form of Independent Electoral Commission is
established, and whether such a body is composed of political parties
or non-partisan civil servants. The world-wide trend is definitely
towards independent electoral commissions staffed by non-partisan
civil servants; indeed, since the worlds largest democracy, India,
adopted this model at independence it has been widely adopted around
the world. However, the influence of the United States is important
here, as the US form of electoral administration is based around
political appointees and party representatives, and many post-conflict
democracies have also adopted this model. The comparative evidence
suggests that this is a mistake, and that independent commissions run
by apolitical civil servants are definitely to be preferred. Party-based
commissions have an almost inevitable tendency to split along party
lines. In Haiti, for example, the Provisional Electoral Council was
made up of representatives of the political parties, but was also deeply
divided along party lines, and internal mistrust and divisions prevented
it from working efficiently.3In Cambodia, by contrast, a non-partisan
electoral commission was widely seen as one of the outstanding
elements of the entire UN mission. Non-partisan commissions were
also a prominent and successful part of UN missions in Namibia and in
East Timor.
The danger of using party-based electoral administrations was
graphically demonstrated by Indonesias transitional elections in 1998.
Amid the flowering of new political movements that often
accompanies a democratic opening, a requirement that both the
125
POST-CONFLICT ELECTIONS
92ip08.qxd 25/06/2002 12:46 Page 125
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
government and opposition political parties must be represented on
the General Elections Commission (KPU), resulted in a deadlocked and
unwieldy body of no less than 53 persons, most of them party
representatives (including some individuals who were also candidates
for the election). Consequently, during the preparation for one of the
most important transitional elections of the 1990s, the body charged
with running the elections the KPU became almost completely
dysfunctional, being deeply divided along party lines and unable to
take even basic decisions (at one stage, fist-fights broke out between
different members of the commission). After the elections, which were
administratively flawed, the Indonesians moved quickly to discard the
party-based KPU and replace it with a much smaller, non-partisan body
of 11 non-party and non-government representatives, many of them
academics.
Electoral processes also need to be sustainable. While the UN plays
an important vector role in spreading new practices and technologies,
there is a distinction between the ideal electoral technology and the
capacity of a recipient country to handle that technology in a sustainable
manner. A number of internationally-financed and run elections over
the past decade have introduced a level of electoral technology which
was clearly unsustainable for the host country, and could not be
replicated in their second, locally-run elections. Cambodia and
Mozambique both fall into this category. Highly expensive levels of
basic equipment and staffing is a common problem; an over-reliance on
sophisticated information technology more suited to a First World
country than a Third World one is another (a typical example is the use
of computerized electoral rolls in countries where electric power is
unreliable). Building a sustainable electoral administration needs to be
the over-riding aim in such situations, even where this means using
more basic technology or equipment. Similarly, donors need to think
hard about the relative merits of funding expensive one-off
international election observation missions (otherwise known as
electoral tourism) versus the longer-term benefits of directly
supporting the domestic electoral administration and local observer
groups. The latter is less glamorous but usually has a much greater pay-
off in actually assisting the consolidation of a new democracy.
While these and other issues of electoral administration continue to
receive inadequate attention, the design of electoral systems, by
contrast, has long been recognized as one of the most important
institutional choices for any political system. Electoral systems can be
purposively designed to achieve particular outcomes, and serve to
structure the arena of political competition, including the party system.
126 RECOVERING FROM CIVIL CONFLICT
92ip08.qxd 25/06/2002 12:46 Page 126
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
The great potential of electoral system design for influencing political
behaviour is thus that it can reward particular types of behaviour and
place constraints on others. This is why electoral system design has
been seized upon by many scholars as one of the chief levers of
constitutional engineering to be used in mitigating conflict within
divided societies. As Arend Lijphart notes, If one wants to change the
nature of a particular democracy, the electoral system is likely to be the
most suitable and effective instrument for doing so.4As well as their
suitability for engineering, electoral rules also serve to structure the
arena of political competition during election campaigns. This has
important behavioural consequences for both voters and candidates.
Because elections represent a primary arena of political competition in
many new democracies, and different strategies of cooperation or
antagonism between the players can increase or decrease their
prospects for success, the electoral system is also a key mechanism in
shaping wider political practices, and can have an effect far beyond the
elections themselves.
Electoral systems also have a direct impact upon politics in societies
divided along ethnic, religious, ideological or other lines. Donald
Horowitz, for example, argues that the electoral system is by far the
most powerful lever of constitutional engineering for accommodation
and harmony in severely divided societies, as indeed it is a powerful tool
for many other purposes.5Lijphart says that the electoral system has
long been recognized as probably the most powerful instrument for
shaping the political system.6Timothy Sisk writes that electoral systems
play an important role in engineering the results of democratic
voting, and along with other institutional choices can have a profound
impact on the nature of political parties and the general character of
democracy.7Beyond this consensus on the importance of electoral
systems, however, there is profound disagreement among theorists as to
which electoral systems are most appropriate for divided societies.
Two schools of thought predominate. The scholarly orthodoxy has
long been that some form of proportional representation (PR) is all but
essential if democracy is to survive the travails of deep-rooted
divisions. For example, Arthur Lewiss study of the failure of post-
colonial democracy in countries such as Ghana, Nigeria and Sierra
Leone in the late 1950s and 1960s prompted him to argue that divided
societies need PR to give minorities adequate representation,
discourage parochialism, and force moderation on the political
parties.8Such arguments foreshadowed, in part, the electoral
recommendations of consociational approaches to managing ethnic
cleavages in divided societies, which emphasize the need for divided
127
POST-CONFLICT ELECTIONS
92ip08.qxd 25/06/2002 12:46 Page 127
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
societies to develop mechanisms for elite power-sharing if democracy
is to be maintained. In terms of electoral systems, consociationalists
argue that some form of proportional representation is all but essential
for divided societies, as this enables all politically-significant ethnic
groups, including minorities, to form ethnically-based parties. Their
prescriptions for electoral system design often focus on the need for
party list PR, usually in large districts. This is based on the tendency of
PR to produce multi-party systems and hence multi-party parliaments,
in which all significant segments of the population can be represented,
and on the empirical relationship between proportional electoral rules
and oversized or grand coalition governments, which are a
fundamental feature of the power-sharing approach on which
consociationalism is based. The use of large, multimember electoral
districts is particularly favoured, because it maximizes proportionality
and hence the prospects of multiple parties in parliaments, which can
then form the basis of a cross-ethnic government coalition.9PR election
rules are thus important of themselves because they are likely to
facilitate proportional parliamentary representation of all groups and
also an important component of wider consociational prescriptions
that emphasize the need for grand coalitions, group autonomy, and
minority veto powers.
In contrast to this orthodoxy, an alternative approach sometimes
typified as centripetalism maintains that the best way to mitigate the
destructive effects of ethnicity in divided societies is not to simply
replicate existing ethnic divisions in the legislature, but rather to utilize
electoral systems that encourage cooperation and accommodation
between rival groups, and therefore work to break down the salience of
ethnicity rather than foster its representation in parliament.10 Drawing
on theories of bargaining and cooperation, centripetalism advocates
institutional designs that encourage opportunities for dialogue and
negotiation between opposing political forces in the context of electoral
competition. By privileging cooperative campaign strategies with
increased prospects of electoral success, candidates representing
competing (and sometimes violently opposed) interests are presented
with incentives to negotiate for reciprocal support, creating an arena of
bargaining where vote-trading arrangements can be discussed.11
Centripetalist approaches advocate the use of electoral rules which
encourage vote-pooling and preference swapping in order to
encourage inter-ethnic bargaining and promote accommodative
behaviour. At the core of this approach is the need to make politicians
reciprocally dependent on the votes of members of groups other than
their own.12 The most reliable way of achieving this aim, according to
128 RECOVERING FROM CIVIL CONFLICT
92ip08.qxd 25/06/2002 12:46 Page 128
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
proponents of the centripetal approach, is to offer sufficient electoral
incentives for campaigning politicians to court voter support across
ethnic lines. For example, some electoral models such as preferential
systems like the alternative vote (in Fiji) or the single transferable vote
(Northern Ireland) permit (or even require) voters to declare not
only their first choice of candidate on a ballot, but also their second,
third and subsequent choices amongst all candidates standing. Parties
that succeed in negotiating preference-trading agreements for
reciprocal support with other parties will be rewarded, thus
strengthening moderate voices and the political centre. This gives them
strong institutional incentives both to engage in face-to-face dialogue
with their opponents, and to negotiate on broader policy issues than
purely vote-seeking ones. The overall effect is thus to reorient electoral
politics away from a rigid zero-sum game to a more fluid, complex and
potentially positive-sum contest. The success of pro-peace forces at
Northern Irelands breakthrough 1998 election was dependent to a
significant extent on such vote-transfers towards the moderate middle
and away from extremists. Fijis transitional 1999 election also utilized
centripetal procedures, as did the transitional 1990 election in Estonia.
Sri Lanka and Papua New Guinea are other examples of countries in
which centripetal electoral systems have or will be used.
Regardless of whether consociational or centripetal approaches (or
some mixture of the two) are favoured, there is widespread agreement
amongst many scholars that some type of power-sharing government
featuring all significant groups is an essential part of democracy-
building in divided societies. In particular, multi-ethnic coalitions are
favoured by both consociationalist and centripetalists as desirable
institutions for divided societies. This form of the power-sharing model
is most often associated with proportional elections, as PR is the surest
way of guaranteeing fair results and minority representation. Lewis, for
example, argues that one of the advantages of proportional
representation is that it tends to promote coalition government.13 Ye t
the comparative evidence from our cases suggest that power-sharing
has been less stable and less in evidence in post-conflict elections than
many scholars would have predicted. In most cases, moreover,
proportional elections have resulted in majority rule: Namibia,
Mozambique, Liberia are all examples of this. In each case, however,
the largest party would probably have won an even greater majority
had alternative institutional designs been employed.
It is instructive to note that almost all of the major transitional
elections conducted in recent years, including almost all of those held
under UN auspices, have utilized some form of PR. In fact, transitional
129
POST-CONFLICT ELECTIONS
92ip08.qxd 25/06/2002 12:46 Page 129
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
elections in Chile (1989), Namibia (1989), Nicaragua (1990),
Cambodia (1993), South Africa (1994), Mozambique (1994), Bosnia
(1996, 1998, 2000), Kosovo (2001) and East Timor (2001) all used a
form of regional or national list PR for their founding elections.
In fact, party-list PR has become the de facto norm of UN
parliamentary elections. In presidential systems, this has usually been
combined with some form of run-off election for the presidency. Only
Haiti in 1995, which used a run-off system for its parliamentary
elections, has deviated from the PR norm (and there the record of this
system was mixed, to say the least: in Haiti, as in Angola, some losing
candidates trailing after the first round of voting chose to boycott the
second round, thus undermining the legitimacy of the process as a
whole).
As would be expected from their widespread use, PR systems have
many advantages for transitional elections in new democracies: they
are fair, transparent and provide a clear correlation between votes cast
in the election and seats won in parliament. By bringing minorities into
the process and fairly representing all significant political parties in the
new legislature, regardless of the extent or distribution of their support
base, PR is often seen as an integral element for creating an inclusive
and legitimate post-authoritarian regime. But the adoption of such
systems for post-conflict elections has usually been dictated more by
administrative concerns, such as the need to avoid demarcating
individual electoral districts and to produce separate ballot papers for
each district, than these wider political issues. Indeed, in many post-
conflict elections, national PR systems are the only feasible way to hold
an election quickly, as a uniform national ballot can be used, no
electoral districts need be demarcated, and the process of voter
registration, vote counting and the calculation of results is
consequently simplified.
However, national PR systems also have some disadvantages, as
they provide no geographic link between voters and their
representatives, and thus create difficulties in terms of political
accountability and responsiveness between elected politicians and
the electorate. In addition, many new democracies particularly
those in agrarian societies have much higher demands for
constituency service at the local level than they do for representation
of all shades of ideological opinion in the legislature. It has therefore
increasingly been argued in Namibia, South Africa, Cambodia and
elsewhere that the proportional systems used at the first transitional
elections should be modified to also encourage a higher degree of
geographic accountability such as by having members of parliament
130 RECOVERING FROM CIVIL CONFLICT
92ip08.qxd 25/06/2002 12:46 Page 130
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
represent territorially-defined districts and service the needs of a
constituency. A popular choice in recent years has been for mixed
electoral systems, in which part of the legislature is elected on a
national level by proportional representation, and some is elected at
a local level from single-member districts, so that both the
proportionality and accountability are maximized. For example, the
August 2001 elections for East Timors 88-member constituent
assembly the body tasked with drawing up the countrys first
constitution used a mixed system, with 75 of the assemblys seats
elected on a nationwide basis by proportional representation, and
13 seats (one for each of the electoral districts) elected by first-
past-the-post.
There are also variations within PR systems that need to be
considered. For example, the precise kind of PR formula used can
influence the extent to which minor parties are represented, or major
parties are advantaged. For example, the use of a Hare divisor at
the provincial level in Cambodia, rather than using a Largest
Remainder system nationwide, had a major political effect: minor
parties who would have gained seats had one national constituency
been used fell short, while the two major parties the Cambodian
Peoples Party and Funcinpec both gained seat bonuses as a result
of these (apparently minor) system choices. Overall, an additional
ten parties would have gained representation had the election
been held on a national rather than a provincial basis.14 In Namibia,
by contrast, a highly proportional national PR system was
introduced despite concerns voiced by the United Nations Institute
for Namibia that parties should reject any PR system that tends to
fractionalize party representation.15 The final Namibian electoral
system was one of the most proportional in the world: with no legal
thresholdsin place, a party needed less than one per cent of the vote
to gain election.
As such cases suggest, it is impossible to divorce the shape of the
party system, and prospects for post-election power-sharing, from the
design of the electoral system. All three are mutually entwined to a
large extent. For example, different types of electoral formula can
encourage or retard different types of party constellations, and can also
influence the extent to which post-conflict parties are broad-based and
moderate entities, drawing cross-communal support, or whether they
are (as in Bosnia) merely former armies in a new guise wolves in
sheeps clothing. Proportional representation, while fairly representing
all views, can also enable small extremist parties to gain crucial
footholds in power. In support of this contention, some comparative
131
POST-CONFLICT ELECTIONS
92ip08.qxd 25/06/2002 12:46 Page 131
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
studies have found that smaller district magnitude the number of
members elected from each electoral district is the crucial
institutional variable in blocking the rise of fringe or extremist parties
and encouraging the development of a broad-based party system,
suggesting that less proportional systems are to be preferred.16
Other technical considerations can also have major implications.
Take the case of designing list PR systems for ethnically-divided
societies: because such systems can utilize one standard national ballot
paper and do not require electoral districts to be drawn or voter rolls
to be demarcated on a geographical basis, they are by far the simplest
system for electoral administrators and, arguably, voters facing first-
time elections in new democracies. But in Bosnia, the application of PR
has also been seen to undermine the process of democratization by
disengaging politicians from voters and, worse, permitting the
development of hard-line nationalist political parties, who can achieve
electoral success by making narrow, sectarian appeals to their core
ethno-political base. Indeed, early Bosnian elections have served to
emphasize that under such conditions, the surest route to electoral
victory under PR is to play the ethnic card with disastrous
consequences for the longer-term process of democratization.
Because of these concerns, the Bosnian elections of November 2000
therefore utilized an open list PR system, in which voters could
choose not just between parties but also between candidates within
parties, with the expectation that this would encourage greater
identification with and responsiveness from elected politicians. But as
anyone familiar with the use of the same system in the deeply
ethnically-torn country of Sri Lanka could have advised this was a
risky move in a divided society where ethnic affiliation remains the
primary basis of voter choice. In Sri Lanka, parties that have attempted
to field a multi-ethnic candidate list have found that such open lists
can undermine, rather than promote, multi-ethnic government:
Sinhalese voters will, if given the chance, deliberately move Tamil
candidates placed in a winnable position on a party list to a lower
position. This may well have been a problem in South Africa as well,
had not the electoral system used been a closed list, which allowed
major parties such as the ANC and the NP to place ethnic minorities
and women high on their party list. In Bosnia, the 2000 elections saw
a wave of victories for extremist parties and candidates, a wave of
victories that the permissive open list PR electoral system only served
to encourage, as it contained no real incentives for inter-ethnic
cooperation or moderation.
132 RECOVERING FROM CIVIL CONFLICT
92ip08.qxd 25/06/2002 12:46 Page 132
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
Political Parties and Power-sharing
Transitional democracies, particularly those moving from a deep-rooted
conflict situation, typically have a greater need for inclusiveness and a
lower threshold for the robust rhetoric of adversarial politics than their
established counterparts. Similarly, the stable political environments of
most Western countries, where two or three main parties can often
reasonably expect regular periods in office via alternation of power or
shifting governing coalitions, are very different from the type of zero-
sum politics which so often characterize divided societies. This is one of
the reasons that winner take all electoral systems like first-past-the-
post have so often been identified as a contributor to the breakdown of
democracy in the developing world: because such systems tend to lock
out minorities from parliamentary representation they can, in situations
of ethnically-based parties, easily lead to the total dominance of one
group over all others.17 Democracy, under these circumstances, can
quickly become a situation of permanent inclusion and exclusion, a
zero-sum game, with frightening results.
But there are also distinctive elements of political parties in post-
conflict situations that appear to transcend institutional considerations.
Because of the underdeveloped and deeply divided nature of most post-
conflict societies, elections often have the effect of highlighting societal
fault-lines and hence laying bare deep social divisions. In such
circumstances, the easiest way to mobilize voter support at election time
is to appeal to the very same insecurities that generated the original
conflict. This means that parties have a strong incentive to play the
ethnic card or to take hard-line positions on key identity-related issues,
with predictable consequences for the wider process of democratization.
Post-communist elections in Yugoslavia in the early 1990s, for example,
resulted in the victory of extremist nationalist parties, committed to (and
achieving) the break-up of the federation. The 1993 elections in
Burundi, which were supposed to elect a power-sharing government,
instead mobilized population groups along ethnic lines and served as a
catalyst for ethnic genocide a few months later. Similarly, Bosnias 1996
and 1998 elections effectively served as ethnic censuses, with parties
campaigning on ethnic lines and voters reacting to heightened
perceptions of ethnic insecurity by electing hard-line nationalists to
power, greatly undermining the process of democracy-building.
For this reason, scholars and policy makers alike have frequently
identified the need to build broad-based, cross-regional and multi-
ethnic political parties in fragile multi-ethnic states, particularly those
susceptible to separatist appeals. Horowitz, for example, has advocated
133
POST-CONFLICT ELECTIONS
92ip08.qxd 25/06/2002 12:46 Page 133
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
the need for broad multi-ethnic parties or coalitions of parties as a key
facilitating factor in avoiding ethnic conflict.18 Similarly, Huntington
argues that fractionalized and ethnically or regionally exclusive party
systems are extremely damaging for democratic prospects and are,
consequently, found widely in the failed democracies of the Third
World.19 A 26-nation study of democracy in developing countries
concluded that a system of two or a few parties, with broad social and
ideological bases, may be conducive to stable democracy.20 Diamond
sums up the prevailing view of many scholars, arguing that political
parties remain important if not essential instruments for representing
political constituencies and interests, aggregating demands and
preferences, recruiting and socializing new candidates for office,
organizing the electoral competition for power, crafting policy
alternatives, setting the policy-making agenda, forming effective
governments, and integrating groups and individuals into the
democratic process.21 By contrast, under the conditions of polarized
pluralism, featuring competition between extremist movements, the
logic of elections changes from one of convergence on median policy
positions to one of extreme divergence.22 Politics becomes a centrifugal
game. Such fragmented party constellations are empirically much more
likely to experience violence and the breakdown of democracy than
more moderate multipartism based on a few catch-all political
parties.23
For this reason, there is an increasing focus in the policy world
which has yet to be adequately digested by scholars on the need to
build broad-based, programmatic political parties in new democracies,
and to avoid the narrow, personalized and sectarian parties and party
systems that have undermined so many new democracies. Particularly
in societies split along ethnic lines, cross-regional and multi-ethnic
parties that compete for the centre ground appear to be a and
perhaps the crucial determinant of broader democratic consolidation
and peacebuilding. For this reason, new democracies around the globe
have, over the past few years, experimented with an unusual array of
institutional approaches to encourage the development of sustainable
political parties and party systems.
There are several ways of doing this. First, party rules governing the
formation, registration and campaigning of political parties can be
enacted which encourage parties to be cross-regional and cross-ethnic
in their composition. This was the approached used successfully at
Indonesias transitional 1999 elections, where to qualify to compete in
the election, political parties must have established a branch structure
in more than half of Indonesias 27 provinces, and within each of these
134 RECOVERING FROM CIVIL CONFLICT
92ip08.qxd 25/06/2002 12:46 Page 134
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
provinces must also have established branches within over half of all
regions and municipalities. The Indonesian drafters stated clearly that
their aim was to discourage political groups based on ethnicity or
region that could form the basis of secessionist claims, and to
encourage the development of broad-based organizations campaigning
on a national platform.24 The results from the 1999 election were
encouraging for these expectations, as the main electoral contest did
indeed appear to take place between three large cross-regional parties,
and the level of ethnic violence associated with the elections was much
lower than had been feared (although it appears to be rising again in
the post-election period). Variations on this approach have also been
used in several other Asian and West African countries.
Second, electoral systems can be designed to enable voters to rank-
order choices between candidates (preferential voting), a process that
has been shown to help sustain centrist parties. This was the approach
used at Northern Irelands breakthough 1998 Good Friday agreement
elections, which utilized a single transferable vote form of electoral
system that enabled voters to indicate secondary choices on their
ballot. Analyses of these elections have found that the use of a
transferable ballot enabled pro-peace Republican and Unionist voters
to give their first vote to their communal party, but to transfer their
secondary votes to pro-agreement non-sectarian parties (thus
advantaging the moderate middle of non-ethnic parties). Vote
transfers overwhelmingly flowed from sectarian parties on both sides
towards the pro-agreement but non-sectarian middle.25 Pro-agreement
parties on both sides of the sectarian divide benefited from such vote
transfers, which among other things were ultimately crucial in
converting a bare anti-agreement unionist voter majority into a bare
pro-agreement unionist parliamentary majority. Evans and OLeary,
for example, conclude that the principal reason that a workable
assembly emerged from the 1998 elections was the adoption, or re-
adoption, of the single transferable vote voters lower-order
preferences kept the Assembly on-track by reducing the numbers of
seats that the anti-Agreement unionist parties won in the election.26
Third, distribution requirements can be enacted which require
parties or individual candidates to garner specified support levels from
across different regions, rather than just their own. The best-known
example of this type of cross-regional engineering has been in Nigeria.
Nigerias February 1999 presidential elections which swept Olesegun
Obasanjo to power took place under laws which contained a so-called
distribution requirement: instead of the usual majority vote
requirement, successful candidates had to obtain not just a majority of
135
POST-CONFLICT ELECTIONS
92ip08.qxd 25/06/2002 12:46 Page 135
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
the vote, but also not less than one-quarter of the vote cast in at least
two-thirds of the states of the federation. The intention behind this
kind of distribution requirement first introduced in 1979 and since
adopted in two other African countries as well was to ensure that the
winning candidate gained cross-ethnic support across the country
rather than just in one part. Again, a primary aim was to counter the
secessionist tendencies that may have been unleashed by the electoral
process under different rules. From the 1999 presidential election, the
preliminary evidence is encouraging: Obasanjo ran on a cross-ethnic
platform and in fact gained greater votes outside his own region than
within it (precisely because, it appears, he campaigned on a cross-
regional multi-ethnic platform).
Fourth, the rules of the game can be constructed in such a way as
to encourage, or require, parties to put forward multi-ethnic lists of
candidates, thus encouraging multi-ethnicity within parties. In
countries as varied as Lebanon, Singapore and South Africa, the rules
of the game encourage parties to present multi-ethnic candidate lists
to the voters. In Lebanon, for example, election is dependent, at a
practical level, on being part of a mixed list of candidates representing
different religious groups. In most cases, candidates must compete for
election against other members of their own group. Electors choosing
between party lists must thus make their choice on the basis of criteria
other than ethnicity. In Singapore, most MPs are elected from multi-
member districts known as Group Representative Constituencies,
which each return between three and six members from a single list of
party or individual candidates. Of the candidates on each party or
group list, at least one must be a member of the Malay, Indian or some
other minority community. Moving from a compulsory to a voluntary
model of multi-ethnic candidate lists, the closed-list proportional
representation system used in South Africas 1994 elections enabled
the major political parties to voluntarily adopt a multi-ethnic candidate
composition thus enabling the major black party, the ANC, to place
white and coloured members at winnable places on their candidate list.
Finally, external interventions can be used to try to stimulate the
development of a meaningful party system where none exists. In
Kosovo, for example, the OSCE devoted substantial resources to
introducing a network of political party service centres, intended to
support the territorys nascent political groupings by providing them
with logistical and material assistance and, by implication, helping
them move towards becoming functioning, policy-oriented political
parties, rather than the narrow and personalized vehicles for ethnic
extremists that were evident in Bosnia. The party service centres aim
136 RECOVERING FROM CIVIL CONFLICT
92ip08.qxd 25/06/2002 12:46 Page 136
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
to help strengthen the organizational capacity of Kosovos political
parties, to assist them in developing their policy platforms and
preparing for election campaigns. They have a particular focus on
assisting parties that have demonstrated that they are viable and have a
popular mandate. In Papua New Guinea, which has a weak and
fragmented party system that has destabilized executive government, a
new law tries to strengthen the party system by encouraging newly
elected MPs to build stable coalitions in parliament, and by granting
the resulting parliamentary parties monetary and administrative
support. The law also provides for a by-election if an MP votes against
their own party leader in a parliamentary confidence vote. Both the
Kosovo and Papua New Guinea approaches can be seen as top down
inducements to organize and build sustainable parties.
Conclusion
It seems trite to reiterate what has been said many times before:
democratization is a long-term process of social and political
development, rather than a short-term event run by the international
community. But we forget this at our peril. The impact that external
interventions can have on democratization particularly in post-
conflict situations is largely limited to the design and construction of
hardy institutions; the provision of adequate security and
infrastructural conditions; as well as a modest input into the norms and
routines of a first election. Beyond that, democracy is a domestic game,
and its longer-term outcomes are very much the preserve of local actors
and conditions. International interventions are crucial in putting in
place the short-term conditions for a transition to democratic rule, but
their longer-term impacts are necessarily limited.
Given this, the most important contribution that the international
community can make is to help establish coherent and robust political
institutions, rather than to engage in broader attempts at social
engineering. Because institutions structure the routines of behaviour in
which political actors engage, they are crucial elements, over the longer
term, in helping to build a moderate and sustainable political culture,
in which routines of cooperation and accommodation come to be
accepted as the norm rather than the exception. But such routines have
to be allowed to develop organically within a facilitating institutional
framework. The role for the UN and other external actors should
ultimately be to make sure that such a framework is the best and most
appropriate that can be devised. Such a limited focus is necessarily a
modest endeavour but a worthy one nonetheless.
137
POST-CONFLICT ELECTIONS
92ip08.qxd 25/06/2002 12:46 Page 137
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
ACKNOWLEDGEMENTS
This essay draws upon the authors research for a project on the UN and democracy
organized by Roland Rich of the Centre for Democratic Institutions at the Australia
National University and the UNU Peace and Governance Programme. Thanks to Edward
Newman of the United Nations University and Robin Ludwig of the United Nations
Electoral Assistance Division for their helpful comments on an earlier draft of this paper.
NOTES
1. J. Linz and A. Stepan, Problems of Democratic Transition and Consolidation: Southern
Europe, South America, and Post-Communist Europe, Baltimore: Johns Hopkins
University Press, 1996, pp.98107.
2. L. Diamond, Developing Democracy: Towards Consolidation, Baltimore: Johns
Hopkins University Press, 1999, p.158.
3. S. Nelson, Haitian Elections and the Aftermath, in K. Kumar (ed.), Postconflict
Elections, Democratization and International Assistance, Boulder, CO: Lynne Rienner
Publishers, 1998, p.76.
4. A. Lijphart, Electoral Systems, in S.M. Lipset (ed.), The Encyclopedia of Democracy,
Congressional Quarterly Press, Washington DC, 1995, p.412.
5. D.L. Horowitz, A Democratic South Africa? Constitutional Engineering in a Divided
Society, Berkeley: University of California Press, 1991, p.163.
6. A. Lijphart, The Alternative Vote: A Realistic Alternative for South Africa?, Politikon,
Vol.18, No.2, 1991, p.91.
7. T. Sisk, Choosing an Electoral System: South Africa Seeks New Ground Rules, Journal
of Democracy, Vol.4, No.1, 1993, p.79.
8. W.A. Lewis, Politics in West Africa, London: George Allen & Unwin, 1965, p.73.
9. A. Lijphart, Electoral Systems, Party Systems and Conflict Management in Segmented
Societies, in R.A. Schreirer (ed.), Critical Choices for South Africa: An Agenda for the
1990s, Cape Town: Oxford University Press, 1990, pp.1013.
10. T. Sisk, Democratization in South Africa: The Elusive Social Contract, Princeton:
Princeton University Press, 1995; B. Reilly, Democracy in Divided Societies: Electoral
Engineering for Conflict Management, Cambridge: Cambridge University Press, 2001.
11. Ibid.
12. D.L. Horowitz, Ethnic Groups in Conflict, Berkeley: University of California Press,
1985; D.L. Horowitz, A Democratic South Africa? Constitutional Engineering in a
Divided Society, Berkeley: University of California Press, 1991.
13. Lewis (n.8 above), Politics in West Africa, London: George Allen & Unwin, 1965, p.79.
14. My thanks to Michael Maley for the data on this point
15. L. Cliffe, et al., The Transition to Independence in Namibia, Boulder CO: Lynne
Rienner Publishers, 1994, p.116.
16. See J. Willey Institutional Arrangements and the Success of New Parties in Old
Democracies, Political Studies, Vol.46, No.3, 1998, pp.65168 (Special Issue).
17. The classic argument on this remains W.A. Lewis (n.8 above).
18. See Horowitz, Ethnic Groups in Conflict (n.12 above).
19. See Samuel P. Huntington, The Third Wave: Democratization in the Late Twentieth
Century, Norman: University of Oklahoma Press, 1991.
20. See L. Diamond, J. Linz, and S.M. Lipset, Introduction: What Makes for
Democracy?, in Diamond, Linz, and Lipset (eds), Politics in Developing Countries:
Comparing Experiences with Democracy (Second Edition), Boulder: Lynne Rienner,
1995, p.35.
21. L. Diamond, Introduction: In Search of Consolidation, in L. Diamond, M.F. Plattner,
Y. Chu and H. Tien (eds), Consolidating the Third Wave Democracies, Johns Hopkins
University Press, 1997, p.xxiii.
138 RECOVERING FROM CIVIL CONFLICT
92ip08.qxd 25/06/2002 12:46 Page 138
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
22. See Giovanni Sartori, Parties and Party Systems, Cambridge: Cambridge University
Press, 1976.
23. See G. Bingham Powell, Contemporary Democracies: Participation, Stability, and
Violence, Cambridge, MA: Harvard University Press, 1982.
24. John McBeth, Dawn of a New Age, Far Eastern Economic Review, 17 Sept. 1998.
25. See B. Reilly, Democracy in Divided Societies: Electoral Engineering for Conflict
Management, Cambridge: Cambridge University Press, 2001.
26. G. Evans and B. OLeary, Northern Irish Voters and the BritishIrish Agreement:
Foundations of a Stable Consociational Settlement?, paper presented at the American
Political Science Association Annual Meeting, Atlanta GA, Sept, 1999, pp.34.
139
POST-CONFLICT ELECTIONS
92ip08.qxd 25/06/2002 12:46 Page 139
Downloaded by [Murdoch University Library] at 23:43 09 May 2016
