DEMOCRATISATION AND ELECTORAL ENGINEERING IN POST-ETHNIC CONFLICT BETTINA
SOCIETIES SCHOLDAN
Abstract
The article explores the theory and practice of constitu-tional engineering in the context of ethnically fragmented societies. By looking at the way electoral engineering fits into broader concepts of democratic rights and democratisation processes, the tension between representation and stability is presented as an issue of how to combine democratic legitimacy with efficiency in dealing with post-war situations. Historical and cognitive institutionalism might provide a better understanding of how an ethnic community might evolve into a political community than a rational-choice oriented, positive institutionalism prevalent in conflict management studies. The author suggests three criteria for assessing efforts of constitutional engineering in terms of democratic legitimacy without having to solve the dilemma of whether democratic institutions or democratic culture must come first. By closing with a brief analysis of discussions of the electoral system in Bosnia, the article raises practical questions about the rhetoric of institutionalism and democratic theory in a post-war context.
Bettina Scholdan is a PhD student in the Department for Political Science, University of Vienna, email: scholdan@oefse.at.
o
in rN
o o o
fN
O >
A 3
a
0
LH fN
Introduction: Democracy and Electoral Engineering
For decades, political scientists have discussed the question of how to design institutions for ethnic-conflict prone societies. One of the questions that is dealt with repeatedly is the issue of electoral systems.1 In the traditional political science approach, the question simply boils down to: how can you work out an electoral system which guarantees outcomes that rather contain than intensify already existing fragmentations (or cleavages or divisions) in a society? In 1994, the U. N. expanded its definition of peacekeeping to "election engineering" (Jenkins 1994, 32). Since then, democratisation projects for post-war societies have been major features of any peacekeeping operation. While this has brought along an almost natural involvement of outsiders (i.e., non-citizens) in the drafting of laws and constitution-making processes, the question of the democratic legitimacy of constitutional (or institutional or electoral) engineering rarely features among the concerns of the engineers who are too preoccupied with making it work.2 The outlet proposed by Giovanni Sartori (1994, 201) — to keep in mind the content-neutrality of constitutions — is clearly off the mark; in the context of post-ethnic-conflict democratisation the goal is to devise systems that are "designed to produce determinate outcomes" (ICG Report 1999). The ideal outcome in cases like Bosnia and Herzegovina (B&H) is precisely to remove ethnicity as the (only) content of politics.
The debate about democratisation in transitional societies oscillates between an emphasis on the right to self-government expressed in a democratic culture — that is, a question of "value-rationality", as Claus Offe (1997) has put it — and an emphasis on political stability and the viability of institutions — that is, a question of purpose-rationality. Certainly, stability is an important part of safeguarding human rights, such as security of life and social conditions; this is a point I do not want to be underestimated. Being purpose-oriented — securing satisfactory representation of all parties and groups involved and therefore maintaining peace — makes a lot of sense in post-war societies. Yet Offe also says that the issue is less the introduction of democratic institutions and the rule of law which can be imposed from above or even from outside, but the sustainability of democracy in the political culture and social fabric of a country. Efforts to bring together the recent emphasis on the relationship between human rights and a sustainable democratic culture with the "renewed confidence in our ability to design and deploy institutions to shape our political culture" (Davis 1997, 131) often jar with the instrumentalist language of conflict management literature.
Democracy has long been understood as a prophecy fulfilling itself with the first "free and fair" election. Elected officials have legitimate authority. The perception that democracy is an ongoing process — very much alive in theories of democracy — has only recently shown itself in the theory and practice of conflict management. The engineering of institutions has been very much in the foreground of conflict management theory, whereas the backbone of democratic culture — the habits of the heart, to use a very American expression (Bellah 1996) — has played too little a role in designing societies in which people have a chance to a live in peace and dignity. Or is this just thinking the wrong way — is it the institutions that produce the ways of political life?
In the first part of my paper, I briefly want to introduce different approaches taken towards ethnically proportional representation. This includes not only theo-
ries of conflict management and institutional engineering, but also theories articulated from the point of view of underrepresented minorities in the United States and Canada. The second part will deal with the approach that, to me, seems most likely to be able to combine institutional with cultural aspects and allows a critique of the instrumentalist, rational-choice theory of ethnicity in politics put forward by most of the conflict management literature. After presenting the main points of "cognitive institutionalism," I will apply its arguments to the discussion on the permanent election law for Bosnia and Herzegovina and raise some questions that I think are important when it comes to the democratic legitimacy of projects of institutional engineering.
How Ethnic Are Ethnic Conflicts Anyway?
At the beginning of the 1990s, ethnicity was elevated to a catchall explanation for a myriad of phenomena in various circumstances. Their similarity was either taken for granted or used rhetorically to denounce any minority rights or multicultural politics. "Ethnic conflicts" were proliferating everywhere, counteracting the process of globalisation and supra-nationalism associated with European integration. In the past couple of years, however, it has become increasingly trendy to say that it is not so much a proliferation of ethnic conflicts but of tagging any complex politically volatile situation with the label of "ethnicity". Ethnic conflicts, by this school of thought, are not about ethnicity at all, but about class and distribution of resources (e.g. Allen and Seaton 1999). Much of the criticism of the use of ethnicity as an explanatory category is directed against exactly the talk about "ancient ethnic hatreds."
Quite contrary to the multicultural rhetoric found in Western societies, "ethnic" in the context of democratisation has become synonymous with a "violent" or even "atrocious" character.3 Apart from mere folklore, calling something "ethnic" means actually removing it from the common civic culture of Western liberal democracies, whose institutions should include all citizens regardless of their national, ethnic, cultural origins, not to say anything about religion, social background, or gender. Ethnicity is seen as a disorder, and which, with its primordial character of identification, escapes rational institutions and policy-making (Offe 1997). That it is also an important site of identification situated between the state and the individual, contributes to the difficulty of assessing its proper role in democratisation processes. Many political scientists have opposed this popular understanding of ethnicity. They argue that it is, indeed, very rational for members of the elite to gather people around ethnic identity, and that is it also rational for people to make sense of the world and the events influencing their lives in terms of ethnicity. It does not, however, always make sense to analyse the relationship between groups and events and motives leading to war as ethnic — as long as "ethnic" only means some innate quality that drives members of one particular group to viscerally hate members of another particular group. Writers such as Prunier (1995) for the Great Lakes region or Noel Malcolm (1994) for Bosnia have shown that cultural history and social background are very much part of the makeup of societies that, once in full-fledged conflict, are represented as torn by purely ethnic cleavages.
It is very important to separate criticism against the unreflected, meaningless soundbite of "ancient ethnic hatreds" as a vantage point of analysis from criticism
of the concept of ethnicity itself. Ethnicity has become a stand-in for what anthropologists had called tribe: group membership based on descent, expressed in shared traditions, and cultural as well as political values and institutions (Glickman and Furia 1995, 12n). Cultural anthropologists, such as Clifford Geertz and Mary Douglas, have brought the meaning of ethnicity closer to the constructionist school of culture: culture is a way of life, a way of understanding one's environment. What is one of the most important aspects of the rise of ethnicity, then, is the issue of perception. I use the term, perception, because I want to distance it from the widely held belief that seeing the world in ethnic terms is due to false consciousness. There is no false consciousness because there is no right consciousness. The more sophisticated accounts of ethnicity show its interdependence with so many different cultural and political factors that it becomes hard to understand them as one and the same phenomenon. But as long as ethnic identity remains one of the safest ways to voice political demands, and as long as it is an easy label to tag onto situations that would take too much time and interest to figure out in their complexity, it will not lose its salience.
In an international context, only players credentialized as "nation states" (however improbably) got to gamble at the top tables; in interpersonal terms, various rights to personal identity vested in understandings of history, representations of what language is, relations to place and locality and so forth became derivative of the inalienable personal right to ethnic identity. To lack ethnic identity is thus to be devoid of some degree of normal (late twentieth century) humanity (Fardon 1999, 72).
It is an important reminder that ethnicity is situated at the intersection of nationalism and racism. This is one of the reasons why giving up ethnic identity as a focus of political claim is so hard; on the one hand, the power of ascriptive categories (identity always being produced by boundaries and differentiation from the other) cannot easily be counteracted, especially for a political and cultural minority. On the other hand, getting claims heard and supported on an international level has been too closely tied to the right to self-determination of peoples — whose definition relies on ethnic descent.
While I agree that one should not take the ethnicity label for granted, I want to recall that perceiving the world through an ethnic lens creates reality for many people. This reality is not a false reality; it is the reality on which people base their actions, thus creating patterns of understanding and corresponding institutions. Gérard Prunier put this very clearly for the context of Rwandese society:
With time this cultural mythology became reality, i.e. the social and political actors moved by degrees from their real world into the mythological script which had been written for them (in a way, with their complicity). By the 1940s their lives, their actions, probably their feelings were obeying the logic of the script rather than that of their more complex organic past, which by then was receding into historical unreality (Prunier 1995, xiii).
This could very well be a description of the current political and cultural institutions of Bosnia: how to acknowledge this reality as the grounds for a democratic right to proportional representation and, at the same time, defuse its divisive character, should be the core question in electoral engineering.
Ethnicity is first of all a translation and not an explanation of social phenomena (Fardon 1999, 70). The current language of understanding conflicts is a poor one; trying to get behind the easy labelling is a first step toward envisioning the appropriate institutional accommodation. Understanding ethnicity as an expression of the master-trope of identity (Fardon 1999, 72) requires an equal understanding of the political, social and cultural ways identities are constructed — and might be deconstructed or reconstructed in the process of constitution-making.
Representation vs. Stability: Electoral Systems for Ethnic-Conflict-Ridden Societies
A large part of the literature on designing institutions for fragmented or divided societies consists of devices for electoral engineering. Questions of ethnicity play a prominent role in these considerations. The African democratisation process, for example, is said to have been severely hampered by the fact that elections were held too soon. Democratic elections meant multiparty politics, and the fastest way to gather support had apparently been the articulation of interests in terms of ethnic identities that had either survived the colonial period or were exacerbated, if not produced, by colonialists (Glickman and Furia 1995, 22ff.). The ability of the electoral system to contain and defuse the violent potential of ethnic conflict is thus a very crucial feature in building democratic institutions that can survive.
From the point of view of democracy theory, one can say that there are two principles seeking to be reconciled in the electoral system: the right to representa-tion4 and right to stability — which is certainly very prominently faced by postwar or transitional societies.5 One cannot be obtained without the other: representation cannot be effective and fruitful if there are no institutions sufficiently anchored in order not to give way to divisive factions. Stability will not be present if there is a part of the citizenry too large or violent enough that does not feel a fair representation of their interests. This interdependence has its impact on the legitimacy — an almost non-existent term in the literature of conflict management — of an electoral system and the officials elected accordingly.
How are these goals — which can be described as adequate representation and sufficient stability — be achieved? I do not want to use too much space explaining electoral systems that have been dealt with elsewhere in depth (Ljiphart and Grofman 1986; Horowitz 1991; Jenkins 1994).
Generally spoken, consociational systems which are created by principles of proportional representation at all institutional levels are said to most satisfy the right to representation. As they leave it to the voter for whom to vote, they are less likely to entrench ethnic categories that are ascribed from outside and might prove divisive (Jenkins 1994, 30). A potential problem of this system is the formation of large-party coalitions eventually leaving no margin of influence for small minority parties (Horowitz 1991, 200). Another point of criticism is the fact that proportional representation, even if it does not assume that voters will vote along ethnic lines, puts no obstacles in the way of candidates who play the ethnic card (Jenkins 1994, 25; Glickman and Furia 1995, 22).
Plurality systems in their pure form completely disallow accommodation of ethnic interests. Whoever gains the majority of votes in a district is elected. But once we say "district," we are already in trouble. For the district boundaries are
drawn by the very people who run in elections which has provided many ideas to people who want to stay in power; it has also inspired a literature on "racial gerrymandering" (Guinier 1994; Karlan 1998), which argues that democracy fails as long as electoral districts are created in ways that delegate minorities to political oblivi-ousness. Meeting these and similar critiques, plurality systems have often been modified to accommodate ethnic diversity (Jenkins 1994, 6-24).
What does this tension between representation and stability mean for societies recovering from a violent conflict articulated in ethnic terms? On the one hand, it seems evident that stability has an even larger relevance in post-war societies. The desire to outrightly force political parties and candidates to show moderation and the willingness to compromise apparently outweighs any consideration of ethnic identity as a valid basis for making political claims. On the other hand, people might be unlikely to be able to think in non-ethnic or even cross-ethnic terms when only recently ethnicity might have been a matter of life and death. Trusting somebody from "the other side" or somebody who fraternises with "the other side" might be too much to ask for in an immediate post-war context. In this case, ethnicity might be in the way of democracy in terms of not promoting the stability necessary for democratic institutions to gain confidence.
Faced with an abundance of imagination of how to design electoral systems and the persistence of ethnically articulated conflicts, one must ask the spoilsport's question: can one change the logic of politics by changing the electoral system? If one reads various explanations of why one system is more likely than another to reduce ethnic tension, one can easily believe that engineering the electoral system means not only engineering the political system but society as a whole. A closer look at countries which have implemented one or the other of those systems reveals that electoral systems alone will not do the trick. Neither Northern Ireland nor Sri Lanka or Nigeria, not to speak of the Lebanon, is particularly well known for being pacifist and moderate in dealing with ethnic group conflicts (Horowitz 1991, 173). The problem with functionalist theories is, of course, that much more than with normative theories, their measure is success. The fascination with electoral engineering, in spite of all the functionalist theory, reflects the idealist hope that something can be done about radical politics:
While there is the danger of concerned but passive resignation to seemingly intransigent ethnic strife, a faith in political institutions to manage ethnic divisions can drive policy recommendations, whatever their origin or nature of the divisions (Glickman and Furia, 1995, 11).
Institutional and constitutional engineering represent more complex versions of this faith, which is very well expressed in the following quote:
A concern for constitutional engineering can reverse the thesis of the need for preconditions of democracy: democratic institutions may cause rather than are caused by a trusting civic culture of tolerance and compromise (Glickman and Furia 1995, 23).
"New institutionalism" sheds a light on the circumstances in which institutions can cause a "trusting civic culture" precisely because it does not neglect the fact that institutions are shaped by the cultures surrounding them. It helps to explain and structure many ideas floating in and about theories on democratisation and
electoral systems, and, with its own shortcomings, exposes the flaws in the current wave of electoral engineering.
"New Institutionalism" as a Bridge between Institutional Culture and Everyday Ways of Life
Is it the institutions that can determine the democratic evolution of a society in transition? "New institutionalism" approaches allow analysing the extent to which institutions can form the society in which and for which they are created; one should keep in mind, though, that institutions become the way people are dealing with them. There are no institutions that cannot be ethnicised (or gendered, for that matter).
According to Shaheen Mozaffar, institutions are "configurations of rules" (Mozaffar 1995, 62). In the language of institutionalism, their significance for political life is tantamount to discourse in poststructuralist accounts of power.6 It channels activity, opens and closes doors to possible strategies, and decides about political identities. The rules, once established, permit, require, and prohibit certain strategies (Mozaffar 1995, 44) and offer the framework and guidance for political mobilisation.
Institutions thus derive their political potency from their constitutive role in both defining the very identity of agents as socially relevant persons and embedding them in a structure of constraints and opportunities.... To control institutions is to determine both the political identities of social actors and their incentives and opportunities, hence their life chances (Mozaffar 1995, 45).
The power of institutions is far-reaching. As Mozaffar is quick to assert, institutions are situated at a junction between agency (setting the rules) and structure (the rules once set, as is the case for groups lacking the requisite power to define rules, but also for democratic constitutions, subordinating the power of the lawless to the rule of law). The institution further "structures and animates the strategic relationship between social actors" (Mozaffar 1995, 43). Understanding the way institutions work helps understanding how "institutions shape the conditions and processes that propel the political activation of ethnicity as a recurrent variable in politics" and how "social actors come to select ethnicity as a basis for organising political action" (Mozaffar 1995, 45). The institutional logic of ethnic politics for Mozaffar consists of the necessary move from ethnicity-as-interest to ethnicity-as-identity, where ethnicity is invested with a normative, symbolic, and transcendent quality.
The instrumentality of strategic rationality, even as it structures ongoing interest-based interactions of contingent cooperation among ethnic actors dissolves ethnic interest into ethnic identity. And ethnic identity validates all instruments that promote and protect ethnic interests (Mozaffar 1995, 61).
The logic of institutionalism becomes quite clear: choosing the correct rules in the correct circumstances produces a satisfactory outcome. The only remaining problem is to identify the correct rules (electoral systems) and the nature of the circumstances. According to "positive institutionalists" (Bussière 1999, 8) like Mozaffar, this should not be too difficult a task; he assumes that rational choice is
the underlying motor of human action. Without going into too much detail as to why rational choice is not an accurate description of human behaviour, I, nevertheless want to point out that with the rise of the school of political economy of war, rational choice has definitely ceased to be a safeguard against violent, and even protracted violent conflict.7 But even in such scenarios, rationality remains the guiding force, although it has stopped delivering the salutary results institutional engineers hope for when they set their sophisticated designs in motion. Even a cautious and subtle writer on the issue of constitutional engineering, like Donald Horowitz, is trapped when he thinks that one does not have to bother with concepts as vague and unpredictable as culture and belief, because all politicians certainly want to be elected and re-elected (Horowitz 1991, 261).
Mozaffar does acknowledge that much circumstantial background must be taken into consideration before rendering a definite analysis of why ethnicity can — in one context — be a very peaceful phenomenon, fostering community spirit and civil society without which democracy cannot prosper, and — in another context — add new dimensions to the term "atrocity" (Mozaffar 1995, 61). His framework for a "systematic comparative analysis ... that can productively integrate the generalisations drawn from empirical studies of ethnic politics in Third World countries ... with those drawn from the studies of ethnic politics in Western countries, Eastern Europe, and Russia" (Mozaffar 1995, 34) lacks the insight that a sequence of events does not happen because human actors — and in this case, elite ethnic entrepreneurs — make strategic decisions. Looking for more or less direct correlations between cause and effect is the pitfall of theories based on rational choice, and it deeply affects any effort to institutionally engineer a transition to democracy. Historical and cognitive institutionalism represent another strand of the "new institutionalist" approach by exploring a more complex interaction between institutions, culture, and human agency with regard to specific historical constellations or active interpretations of the role and purpose of institutions.
"Cognitive institutionalism," influenced by social constructionists, might offer a way out of the dilemma. It is flexible enough to accommodate ideas such as contingency and interpretive action. Cognitive institutionalism still places institutions between the poles of agency and constraint (which, according to a discourse theory of power, are mutually reinforcing. Foucault 1984), but allows a space for normative democratic action that exceeds the instrumentality of rational choice theory.
"Politics ... involves ongoing efforts to persuade others via normative arguments that they should think of themselves and their interests differently than they do." Indeed, institutional power in a democratic society flows less from raw power than the legitimacy ultimately conferred by citizens (Rogers Smith quoted in Bussière 1999, 8).
The concept of "institutional mission," borrowed by Elizabeth Bussière from Howard Gillman8 (Bussière 1999, 9), refers to the feeling of actors to work at achieving the same goals within an institution. This is not an intrinsically rational feeling; translated into designing institutions for a post-war society it is the articulation of the aspirational question standing at the beginning of constitutional debates: "Who wants to live in the same polity, on more or less equal terms?" (Horowitz 1991, 278). But it conveys something of what a democratic culture might be and why self-government might be worth achieving. Legitimacy and accountability might
be a rather old-fashioned stand-in for ethnic identification. Can substantially democratic institutions not only be substitutes, but also pathbreakers for democratic elections, as Glickman and Furia suggest?
The key point is that perhaps ethnic strife is part of the process of solving the problem rather than the problem itself, when it comes to the establishment of democratic government. So the key question on elections would be how we can hold elections whose results are not destabilizing. Perhaps elections ought to come relatively late in the process of democratisation, rather than early (Glickman and Furia 199, 23).
What is the consequence of the suggestion that democratic institutions shape a democratic culture and way of life? Is it less the electoral design than the overall institutional pattern that promotes identification with a political rather than an ethnic culture? How can the step be made from an assimilationist civic culture or an exclusionist national culture to a multiethnic political culture? This raises the question of the relationship between citizenship and ethnicity in multicultural societies. I would like to recall the meaning of theories of presence for institutional engineering in this context:
Theories of presence claim that in constructing political institutions, the belonging of representatives is as important as the positions they endorse, the interests they defend, or the voting record they hold. No social group, these theories argue should be excluded from the public sphere (Tamir 1997, 2).
If the political expression of ethnic identity is a human right requiring active protection and promotion from the state, ensuring non-discrimination on the basis of equal citizenship might not be enough. If, as some authors suggest (Kymlicka 1995, Phillips 1995), the feeling of self-government at the heart of democratic citizenship can only be achieved through adequate representation by members of one's own ethnic group (the same goes for gender and sexual orientation), how can democratisation in a multiethnic society take place without, somehow, ensuring the proportionality of ethnic representation?9
Unfortunately, these questions will not be answered within the scope of this article. In the case of societies where violent conflicts had been articulated in ethnic terms, they lead back to the tension between representation and stability that governs efforts to institutionally anticipate and channel ethnic fragmentation. Do ethnic groups in a post-war situation have to waive their right to self-government
—	if we assume for the time being that the link between democratic citizenship and ethnic representation is, indeed, one that cannot and should not be neglected
—	for the sake of stability? As with the question of whether democratic institutions can breed a democratic culture, or whether democratic beliefs and culture are needed to anchor the institutions, democratic rights and stability appear at odds. My point is less about deciding what has to come first; in post-war multiethnic societies constitutional engineering from outside cannot yield identification with the political community necessary for durable democratisation. There is a definite need for the kind of intransigent critique of individuals like Robert Hayden, reminding us that constitutional engineering from outside is always undemocratic and that the current state of Bosnia is absurd (Hayden 1998); yet in our less than perfect world, constitutional engineering from outside very often offers the only
hope for peace. Therefore we must develop criteria for these interventions that satisfy both, the need for efficiency and stability, and the need for aspirational symbolism that forms the core of political identity and culture.
The following criteria might be used in assessing the democratic legitimacy of projects of constitutional and electoral engineering: the inclusive nature of the deliberation process; the democratic legitimacy of individuals passing the respective laws, and the extent to which the institutional arrangements allow for the agency of citizens; how much do their decisions determine the polity as such and society as a whole? The case of Bosnia may show that balancing representation and stability through electoral engineering can be even trickier than textbooks would acknowledge.
Engineering Bosnia
After four years as a laboratory for institutional designers of all sorts and creeds, the common perception is that Bosnia can still not survive on its own (Boyd 1998).10 According to political analysts, part of the blame rests in the way elections were scheduled and administered in B&H. The electoral system served to entrench ethnic divisions, perpetuating the fears and animosities produced in the course of the war, and hindering the evolution of a culture of reconciliation and human rights that would have been the prerequisite for a self-sustaining democracy.
The conditions for the first post-Dayton elections, held in September 1996 were pretty bad, according to the International Crisis Group, whose reports count among the most sophisticated and balanced ones of the post-war Bosnian political culture.11 In May 1996, ICG outlined the "Conditions for Democratic Elections in Bosnia" (ICG Report 1996a) which were declared as not met by August 1996 (ICG Report 1996b). Neither the return of refugees nor freedom of movement, the apprehension of war criminals, or the moderation and fair-play of the political parties, were underway when OSCE declared the conditions for democratic elections as sufficiently fulfilled. (ICG Report 1996b; Udovicki 1997, 284, 307 n. 19 and 20).
A comparison of ICG's somewhat bleak predictions with ensuing developments shows that almost all of the predictions turned out to be correct — to at least some extent. Even as violent incidents have decreased, most of the consequences — cited in the August 1996 report — contributed to the problems encountered in the 1997 and 1998 elections. Voter confusion might seem to be a trivial admonishment, but anyone present in a polling station in the Bosnian countryside will know that it actually turns out to be quite a crucial aspect of the procedural and substantial legitimacy of the election process. Irregularities and disenfranchisement of voters are other and more obvious flaws of the 1997 and 1998 registrations and elections, neither of which have promoted trust in democratic processes or an inclusive sense of equal citizenship.12
Consequences that affect the core of this presentation directly are the "repercussions on the ground": the strengthening of control of the ruling, nationalist parties in their territories of dominance, and the legitimacy and continuation of the "'national projects'" begun with the war. (ICG Report 1996b) One of the reasons for allowing a continuation of this process (Instead of its reversal) were the voting rules: although voting for the municipality of one's residence at the time of the last pre-war census (in April 1991) was supposed to be the rule, voters could also cast
their ballot for the municipality of their post-war residence. In 1996 and 1997 it was also possible to register for a "future municipality," if one could make a credible claim of permanent residence by the time of the elections. Combined, these rules made it easy for local politicians to exacerbate the fear of severe distortions of the election results by absentee voters (to be rightly understood: the victims of "ethnic cleansing"), who were expected to vote for "their" party i.e., the nationalist parties representing the "other" ethnic groups; it allowed ruling parties to play the ethnic card for preserving the gains and recouping the losses. Together, with a system of rigid, if unevenly implemented, tripartite proportional representation (i.e., Bosniac, Serb and Croat candidates, notwithstanding the 7.7 per cent who declared themselves as "others" in the 1991 census), according to the 1991 census, by ethnically designating seats at all levels of public office, the electoral system did nothing to soften the division and calm the fears of the constituency. It did, however, contribute much to the current institutional stalemate that is said to be at the core of the dysfunctionality of democratic institutions in B&H (ICG Report 1998b).
Very much in tune with "new institutionalise approaches, ICG proposed "changing the logic of Bosnian politics" in the spring prior to the 1998 national and regional elections (ICG Report 1998a). The ICG paper does not try to fit Bosnian institutions into the ideal of a multiethnic Bosnia that some have in mind when speaking of pre-war Bosnia. ICG takes into account the distinctiveness of ethnic groups throughout Bosnian history and their dissolution as a recent event (ICG Report 1998a, 1; see also Schoepflin 1993, 186) The logic of the Bosnian electoral system has turned out to do nothing to disperse divisive and potentially volatile structures:
Having been elected on the basis of votes of a single community, an ethnically-based party will only represent the interests, or what they deem to be the interests, of that one people and will feel no obligation to the rest of Bosnia's population. Instead of seeking accommodation in common institutions, these parties view every political issue as a "zero game" in which there will inevitably be a winner and a loser and thus fail at all times to reach compromise. Further, where an ethnically based party has an absolute majority, members of other communities effectively have no political representation and, by extension, no political rights (ICG Report 1998a, 3).
Another reason why taking no responsibility for governance has been such an easy option for the major parties might have been the fact that with the extended exercise of the powers of the High Representative, there was basically no need to reach a compromise. As long as politicians know that someone will pass a law or take necessary action, they can show their ethnic muscles without risking the complete dissolution of the institutions that keep them in power. In terms of cognitive institutionalism — discussed above — the flaw might just lie in not allowing the state of Bosnia and Herzegovina to become an "institutional mission."13 This is more than apparent in the Bosnian Constitution (Annex 4 of the GFAP), which prescribes a split of citizens in terms of voting rights. Article V of the Constitution reads: "The Presidency of Bosnia and Herzegovina shall consist of three Members: one Bosniac and one Croat, each directly elected from the territory of the Federation, and one Serb directly elected from the territory of the Republika Srpska."
As its paper was not heeded before the 1998 elections, ICG did a rerun on electoral reform proposal in March 1999, using slightly more pressing language ("Break-
ing the Mould"); it still acknowledged that "[a]s long as ethnicity is the issue" (ICG Report 1999, 4) exhorting Bosnians to vote on issues rather than ethnicity will lead nowhere in terms of breaking up the current deadlock. By trying to find an overlap between the kind of government that would be best for Bosnia and the kind of government that Bosnians want, ICG produced several formula (of Alternative or Preferential Voting) that are supposed to neutralise radical votes and give a chance to voters for whom ethnicity is not the issue.
Asking citizens about the kind of government they want has been declared a major prerequisite for successful democratic reforms of institutions (Jenkins 1994, 30). Since March 1999, OSCE has been responsible for an information campaign on the election law and assisted in organising a deliberation process between parties, civic associations, and NGOs. It also conducted a public opinion poll among Bosnian citizens (OHR/OSCE 1999a). While the results seem, somehow, overly optimistic about the interests of the general public in the political system, they seem to confirm ICG assumptions about motivations for voting — almost exclusively the ethnicity of a particular candidate (OHR/OSCE 1999a). Simultaneously, the citizens of B&H perceive their elected officials as "unresponsive." Accountability was recognised as a very desirable feature of democratic politics.
The draft for the permanent election law (released on October 21) incorporates several proposals made not only by ICG but also by writers on electoral engineering. Open list proportional representation is incorporated in the elections for all levels of parliament — the system of ethnically slated seats set up in Article IV of the B&H Constitution, however, remains untouched. Regarding the presidential elections, the draft provides for a preferential ranking system, which entices candidates to search not only for the voters whose first preference they are, but also for those who would give them second or third preferences.14 Before being allowed to run for office, the candidates must look for cross-entity support by getting a sufficient number of signatures either from voters with permanent residence in a different entity or support by a simple majority from 2 (independent candidates) or 4 (party candidates) Municipal Council/Assemblies in a different entity.15
The engineering intentions behind the law seem quite clear about the nature and purpose of specific institutions. On the legislative level, the law pays tribute to the right to ethnically proportional representation (and the futility of working against that grain in a society where ethnicity was a question of life and death not so long ago). By reducing the electoral units and requiring candidates to establish permanent residence in their constituency, the draft wants to ensure accountability of candidates. Accountability actually was one of the major issues voiced in reactions during the Permanent Election Law Information Campaign (OHR/OSCE 1999a). The recently established Open-List-System gives voters the chance to vote for particular candidates, thus increasing the influence of voters on who actually gains a seat. On the level of the presidencies, the system hopes for integrative figures, for cooperating individuals who pursue the "institutional mission" of an undivided Bosnia.
But it remains divided under this provision. Signatures of support cannot replace the right to vote for a candidate. Quite understandably, the working group was reluctant to change a constitutional provision as fundamental as the question of who votes for whom.16 One can only hope that this does not make the draft a mere "chimera" — an expression with which one of the international advisers of the working group, Donald Horowitz, denounced any hope for introducing a differ-
ent institutional logic after a new regime is set in place (Horowitz 1991, 244). The question of accepting the draft notwithstanding, it does more than expose the weaknesses and obstacles for democratisation built into the Dayton framework: it shows the limits of engineering ethnic content with the tools of undemocratic citizenship. Whether it will be enough to "engineer" Bosnian society toward the goal of the international community set — a multiethnic Bosnia — will, finally depend on the will of the people not to think like institutions designed in their best interest. Faith in democratic institutions still begins with faith in citizens' agency.
Addendum
Since this paper was submitted for publication in November, the October draft as well as a revised draft dating from January 2000 have been rejected by the parliament of B&H. The deadline for adoption of the law set by the High Representative for 15 February 2000 as a prerequisite for accepting B&H as a member of the Council of Europe has passed. There has been scathing criticism of the proceedings of the OHR and the OSCE, mostly for not contesting the constitution.17 Apart from the question of whether the European Convention on Human Rights overrides all other constitutional provisions, constitutional interpretation or revision should not be the task of a working group on election laws. By exposing the flaws of Dayton, the draft election law has once more opened a discussion on what Bosnia and Herzegovina should look like.18 There may be little to indicate that the current discussion of a revision of the Dayton agreement leads to a true constitutional debate instead of typical, ethnically entrenched party politics. However, trying to force the adoption of the election law by calling on the wrath of the Council of Europe, as the High Representative has recently done, only cuts short a process necessary for democratisation in Bosnia where it has not happened so far.
Notes:
1.	This paper was initially prepared for a conference of the International Political Science Association: Human Rights Research Committee: Human Rights and Democratic Transitions, August 26 - 28, 1999, Ljubljana. I am very grateful to Professor Rizman for giving me the opportunity to contribute to the conference. I also owe thanks to Michael Davis and Birgit Kainz for their very helpful comments and criticism. Jim Rogan of the OSCE Mission to Bosnia and Herzegovina was so kind to provide me with insights into the preparation of the draft for the permanent election law of B&H as well as with the text of the final draft and commentary. Unfortunately, these texts were released only a couple of days before this manuscript was due. The final chapter entitled "Engineering Bosnia" is, therefore, only a very cautious sketch of another article that will discuss its provisions and underlying policy in depth.
2.	Constitutional or institutional engineering refers to designing an entire set or a part of institutions that make up the political system. With constitutional engineering, the additional task of building a new state structure—usually after a war, a secession, or a period of non-democratic rule. Electoral engineering is based on the assumption that the choice of an electoral system influences the entire way of making politics by determining how majorities are created. The kind of election system is thus perceived as crucial to the sustainability of the institutional arrangement set up in the constitution- making process; it is assigned a prominent role in cases like Bosnia where state-building is said to be hampered by political parties counting on the weakness of the identification with state institutions.
3.	See for example, Banks and Wolfe Murray (1999, 149), quoting a Professor of Modern English at the University of Birmingham, who said in an interview with the Guardian (14 July 1994): "If you say someone has ethnic tendencies you are taken to mean that he is a murderer."
4.	Theories of representation in multi-ethnic societies insist on the fact that democratic representation is not achieved simply by putting a ballot in the ballot box on election day. Being enfranchised, having an equally weighted vote, and the right not to completely and invariably lose one's vote in the process of electoral arithmetic, are all part of what is so simply called voting rights. Theories of representation, especially those concerned with the politics of identity, often go further to discuss the appropriate (identity) link between the representative and the represented. Does ethnic identity automatically carry along the same political agenda? See for example Guinier 1994; Philipps 1995; Tamir 1997.
5.Jenkins	(1994) places Ljiphart in the group of those preoccupied with representation and Horowitz among those emphasising stability more.
6.	Unfortunately, I lack the space to explore the discursive distinctions and similarities between discourse theory and institutionalism.
7.	One, if not the most important, representative of this school is David Keen in his article, "Who's It Between? 'Ethnic War' and Rational Violence". In Allen and Seaton 1999, 81-101, he argues that the driving forces behind contemporary wars might be easier understood if one accepts the possibility that it is rational not to win a war (86).
8.	"The Court as an Idea, not a Building (or a Game): Interpretive Institutionalism and the Analysis of Supreme Court Decision-Making," published in C. W. Clayton and H. Gillman (eds.), Supreme Court Decision-Making: New Institutionalist Approaches, 65-87, Chicago: University of Chicago Press.
9.	This appears to me to be even more valid in the case of societies with a high level of ethnic stratification. Where ethnicity is a proxy for assigning social status, the failure to adequately represent an ethnic group translates into more than cultural injustice. Yael Tamir makes this point in her discussion of theories of presence: it is not competence or knowledge, or even role model needs for formerly discriminated and inferiorized groups that is affected by a politics of presence, but the very issue of self-government itself (Tamir 1997).
10.	In a way this perception is even confirmed by Carlos Westendorp's assessment of developments during his tenure (Office of the High Representative, 16 July 1999).
11.There	are two reasons for drawing so heavily on the reports of the International Crisis Group. One is their quite outspoken position on the engineering of democratic institutions in Bosnia. In this case, ICG definitely did not lament the "illegitimacy of the protectorate.".Also, some of the proposals made by ICG concerning electoral reform were taken into consideration for the draft of the Permanent Election Law- albeit after more than eighteen months.
12.	Apart from the exclusion of the famous category, "other," from the tripartite allocation of seats to Bosniacs, Croats, and Serbs in the Dayton constitution, the decision to postpone the right to apply for B&H citizenship until the year 2000 for Serbs from Krajina has excluded about 40,000 people from voting in no less than six consecutive elections on Republika Srpska territory. The persisting problems with voter registration—alhough to some extent nobody's fault—has alienated voters in both entities from the international organizations that want to be seen as teachers of democracy and the rule of law.
13.	It is very much an institutional mission for the international community on the ground. The fervor and passion with which OSCE officials speak about the work on the permanent election law shows that there has been a constitutional moment in the post-war history of B&H. It goes without saying that this won't do.
14.	For the arithmetic of this system see ICG 1999.
15.	Articles 4.4 and 4.8 of the draft for the election law (OHR/OSCE 1999b); In this case, one probably has to pay attention to the extent to which Municipal Councils and Assemblies are constituted by members elected by voters voting for their 1991 residence—which would then support "their" candidate in a cross-entity, but not in a cross-ethnic way.
16.	James Rogan, OSCE, personal communication, 31 August 1999. Not very surprisingly, the critical reactions so far focus on the perpetuation of the inherent inequality of B&H citizenship. E.g. ICG 1999b).
17. Alternative Information Network, 9 February 2000; ICG 1999b.
18. Bosnia Report 1999; Habul 2000.
References:
Alternative Information Network Sarajevo. 2000. Dossier: Election Law in B&H. Alternative Information Network, 9 February, 2000 <www.aimpress.org> [accessed 9 February 2000].
Banks, Marcus and Monica Wolfe Murray. 1999. Ethnicity and Reports of the 1992-95 Bosnian Conflict. In T. Allen and J. Seaton (ed.), The Media of Conflict. War Reporting and Representations of Ethnic Violence, 147-161. London/New York: Zed Books.
Bellah, Robert N. et al. 1996. Habits of the Heart. Individualism and Commitment in American Life. Berkeley: University of California Press (First published 1985).
Bosnia Report. 1999. Should Dayton Be Revised?. Interview with Dr. Zoran Pajic, legal adviser to the International Crisis Group. In Bosnia Report, August - November 1999 <www.bosnia.org> [accessed 9 February 2000].
Boyd, Charles G. 1998. Making Bosnia Work. Foreign Affairs 1, 42-55.
Bussière, Elizabeth. 1999. Where Law and Community Meet: Juries as "Conscience of the Community" in Historical Context. Paper presented at the Law and Society Association Meeting, May 29, 1999, Chicago.
Davis, Michael. 1998. Constitutionalism and Political Culture: The Debate over Human Rights and Asian Values. Harvard Human Rights Journal 11, 109-147.
Election Law Information Campaign: Responses to Discussion Held by the House of
Representatives and House of Peoples of the Parliamentary Assembly of B&H. <http:// www.oscebih.org> [accessed 8 February 2000].
General Framework Agreement of Peace. Annex 4. Constitution of Bosnia and Herzegovina.
Glickman, Harvey and Peter Furia. 1995. Issues in the Analysis of Ethnic Conflict and Democratisation Processes in Africa Today. In H. Glickman (ed.), Ethnic Conflict and Democratisation in Africa, 1-31. Atlanta: The African Studies Association Press.
Guinier, Lani. 1994. The Tyranny of the Majority. Fundamental Fairness in Representative Democracy. New York: The Free Press.
Habul, Emir. 2000. Destiny of Peace Agreement for B&H. Preparations for Amendments or Improvement of Dayton? In Alternative Information Network, January 26, 2000. <www.aimpress.org> [accessed 9 February 2000].
Hayden, Robert M. 1998. Bosnia: The Contradictions of "Democracy" without Consent. East European Constitutional Review 7, 2 <http://www.law.nyu.edu/eecr> [accessed 13 June 1999].
Horowitz, Donald L. 1991. A Democratic South Africa? Constitutional Engineering in a Divided Society. Berkeley: University of California Press.
ICG Report. 1996a. Conditions for Democratic Elections in Bosnia. Sarajevo, 22 May.
ICG Report. 1996b. Why the Bosnian Elections Must Be Postponed. Sarajevo, 14 August.
ICG Report. 1998a. Changing the Logic of Bosnian Politics: Discussion Paper on Electoral Reform. Sarajevo, 10 March.
ICG Report. 1998b. Doing Democracy a Disservice: 1998 Elections in Bosnia and Herzegovina. Sarajevo, 9 September.
ICG Report. 1999a. Breaking the Mould: Electoral Reform in Bosnia and Herzegovina. Sarajevo, 4 March.
ICG Report. 1999b. Is Dayton Failing? Bosnia Four Years After the Peace Agreement. Sarajevo, 28 October.
Jenkins, Laura D. 1994. Ethnic Accommodation through Electoral Systems. A paper prepared for the UNRISD/UNDP project on Ethnic Diversity and Public Policies.
Karlan, Pamela S. 1998. A Bigger Picture. Boston Review, February/March 1998, 19.
Keen, David. 1999. Who's It Between? "Ethnic War" and Rational Violence. In T. Allen and J. Seaton (eds.), The Media of Conflict. War Reporting and Representations of Ethnic Violence, 81-101. London: Zed Books.
Kymlicka, Will. 1995. Multicultural Citizenship. Oxford: Oxford University Press.
Ljiphart, Arend and Bernhard Grofman, eds. 1986. Electoral Laws and Their Political Consequences. New York: Agathon Press.
Malcolm, Noel. 1994. Bosnia. A Short History. London: Papermac.
Mozaffar, Shaheen. 1995. The Institutional Logic of Ethnic Politics: A Prolegomenon. In H. Glickman (ed.), Ethnic Conflict and Democratisation in Africa, 33-69. Atlanta: The African Studies Association Press.
Offe, Claus. 1997. Cultural Aspects of Consolidation: A Note on the Peculiarities of
Postcommunist Transformations. In East European Constitutional Review 6, 4 <http:// www.law.nyu.edu/eecr> [accessed 22 October 1999].
Office of the High Representative. 16 July 1999. Report by the High Representative for
Implementation of the Peace Agreement to the Secretary-General of the United Nations. Annex 1 [Internet] <http://www.ohr.int> [accessed 30 July 1999].
OHR/OSCE. 1999a. B&H Citizens Consider their Elected Officials "Unresponsive," PELIC Survey Reveals. Press release, 5 August, 1999 [Internet] <http://www.oscebih.org> [accessed 13 August 1999].
OHR/OSCE. 1999b. Draft for the Election Law for Bosnia and Herzegovina, released October 21, 1999.
Philipps, Anne. 1995. A Politics of Presence. Oxford: Oxford University Press.
Prunier, Gérard. 1995. The Rwanda Crisis. History of a Genocide. London: C. Hurst & Co.
Sartori, Giovanni. 1994. Comparative Constitutional Engineering. An Inquiry into Structures, Incentives and Outcomes. Basingstoke: Macmillan.
Schoepflin, George. 1993. The Rise and Fall of Yugoslavia. In J. Mc Garry and B. O'Leary (ed.), The Politics of Ethnic Conflict Regulation, 172-203. London: Routledge.
Tamir, Yael. 1997. Multiculturalism and Theories of Representation. Paper presented at the Conference for the Study of Political Thought. Democratisation and Justice, Columbia University, April 4 - 5, 1997.
Udovicki, Jasminka and James Ridgeway, eds. 1997. Burn this House. The Making and Unmaking of Yugoslavia. Durham: Duke University Press.