2 The favorable factors for power sharing
2.2 The composition of segments
2.3 Geographical concentration of segments
2.4 Socioeconomic differences
2.5 Tradition of elite accommodation
2.6 Overarching Loyalties
2.7 External threats
2.8 External actors
3 The concept of favorable factors applied to Bosnia and Macedonia
3.2 The composition of segments
3.3 Geographical concentration of segments
3.4 Socioeconomic differences
3.5 Tradition of elite accommodation
3.6 Overarching loyalties
3.7 External threats
3.8 External actors
With the advent of new media battles and bloodshed have found their way into our daily lives. Traditional and new media alike provide an abundance of images and stories, which can be accessed around the clock. Beyond that, social media platforms like Facebook and Twitter enable individuals to actively partake in the dissemination of information by sharing and dis- cussing their opinions online. The increased participation of various actors in the news pro- cess has generated a flood of information and data. As a result the audience takes less time to review issues in its entirety, like the full-scale process of conflicts and wars including the fol- lowing aftermath. Thus, once a peace agreement is brokered, public interest shifts to "[...] the latest outburst of hostility at another unstable spot across the globe" (Koneska 2014, p. 1).
The current coverage on the Balkans perfectly demonstrates this phenomenon. Whereas, dur- ing the 1990s, news stories about war and riots in the region were at the top of the media agenda, the former Yugoslavian states hardly make headlines nowadays anymore. Coverage is usually limited to the integration process of former Yugoslav countries within the internation- al community, such as membership in the European Union (EU) and North Atlantic Treaty Organization (NATO). Thus, even though the outbreaks of violent protests in June 2013 and February 2014 in Bosnia and Herzegovina attracted worldwide attention, discussion about the region's politics quickly petered out once the protests stopped (Arsenijević 2014a, p. 103). However, the recent turmoil shows that even almost 20 years after the passing of the Dayton Peace Agreement, which put an end to one of the most gruesome wars in recent European history1, Bosnia is far from being entirely democratically consolidated. Similarly, Macedonia has seen recent outbreaks of violence, when thousands of protesters gathered in Skopje past July to demonstrate the conviction of six people charged with terror-related murder motivated by ethnic hatred (Lajmanovska 2014). The severity of the protests in both countries reveals that ethnic tensions are still holding a strong position within the multiethnic society of the country.
Apart from the lacking media coverage on Bosnia and Macedonia, recent events reveal the need for a closer examination of the respective countries. As we are watching the unraveling of former power sharing states like Iraq or Lebanon and the resulting tragedies, it seems ever more relevant to examine the factors that hinder effective democratization, as promoted by consociationalism.
The following study will provide a comparative study between Bosnia and Herzegovina2 and Macedonia, both countries implemented with power sharing structures. In addition they share “[...] structural similarities, such as size and geographic location” (Noel 2004, p. 88). Both former Yugoslav countries have seen violent conflict -Bosnia to a much greater extent than Macedonia- prior to the implementation of power sharing. Furthermore, peace agreements - consisting of the Dayton Peace Agreement (DPA) in Bosnia and the Ohrid Framework Agreement (OFA) in Macedonia- could only be reached due to pressure exerted by external actors. The two countries were equipped with power sharing structures, which were aimed to “maintain the peace after brutal civil wars as well as facilitate the consolidation of democracy after less violent conflicts […]” (Roeder and Rothchild 2005, p. 5). As both countries still face difficulties in their democratization process the objective of this thesis is to assess the factors which favor and obstruct power sharing in Bosnia and Macedonia.
The first part will offer the theoretical framework following Arend Lijphart's Favorable Con ditions3 for Consociational Democracy, which has been subject of constant readjustment over the course of his different studies on consociationalism. Following a brief introduction on the theory of consociationalism, the nine favorable factors of power sharing by Lijphart will be explained in greater detail. External actors will be added as an additional factor up for review. Criticism of Lijphart's factors will conclude the first part of the thesis.
The main part of the thesis is concerned with the actual case studies of Bosnia and Macedonia. The theoretical model will be applied to the countries realities, by dedicating a chapter to each factor. Each chapter will end with a short conclusion on whether the factor can be considered favorable in the respective countries, while also including a brief comparison between Bosnia and Macedonia.
Ultimately, the question “Which factors have been favoring and obstructing power sharing in Bosnia and Macedonia?” will be answered. Moreover, the conclusion will draw a comparison of the effects of the favorable factors in Bosnia and Macedonia and therefore help assessing the relevance of Lijphart’s favorable factors in general.
2 The favorable factors for power sharing
As Philip Roeder and Donald Rothchild (2005, p. 5) state “[…] power sharing has become the international community’s preferred remedy for building peace and democracy after civil wars”. The most common approach of power sharing, used in today’s political science, was formulated by Arend Lijphart in 19684. Lijphart introduced consociationalism in order to ex- plicate political stability in plural societies. Above all, consociational democracy relies on the willingness of elites to cooperate. This requires that the leaders feel at least some commitment to the maintenance of the unity of the country as well as a commitment to democratic practic- es. “They must also have a basic willingness to engage in cooperative efforts with the leaders of other segments5 in a spirit of moderation and compromise” (Lijphart 1977, p. 53). Moreo- ver, his model is based on four basic principles: (1) a broad based or “grand” coalition execu- tive, in which elites form multiethnic coalitions after elections; (2) minority veto rights, given to minorities in order to protect their vital interests; (3) proportionality in the allocation of civil service positions and funds, in order to enable minorities access to power, the right to participate and influence the political life depending on their size; (4) and group autonomy, through either territorial or corporate federalism (Sisk 1999, pp. 36-37).
According to Lijphart (1977, pp. 54-55) [b]oth the explanatory and predictive power of the consociational model can be improved […] by identifying the conditions that are conducive to overarching elite cooperation and stable nonelite support. […] To the extent that these factors contribute to cooperation among segmental leaders and loyal support by the followers in the segment, they are conditions that are helpful not only in establishing consociational democracy in a plural society but also, once it is established, in maintaining and strengthening it.
These factors “[…] have received much attention and are among the most contested elements of consociationalism [...]” (Bogaards, p. 476). Lijphart made his first notion of the favorable factors in The Politics of Accommodation in 1968. He derived the factors from his comparative studies on Switzerland, Austria and Lebanon. During the course of his following publications (1968, 1969, 1977, 1985, 1996, 2008) Lijphart constantly adjusted not only the number, but also the content of his conditions (Bogaards 1998, pp. 476 ff.). Table 1 presents a compostion of Lijphart's different selections over the years.
The most contested element concerning the favorable factors is Lijphart's (1977, p. 55) claim that they “[...] are helpful but neither indispensable nor sufficient in themselves to account for the success of consociational democracy.” This point of criticism will be discussed in greater detail later on.
Ultimately, Lijphart ends up with a selection of nine factors, which will serve as the theoreti- cal basis for this study. The nine factors consist of the smallness of the respective country, absence of a solid majority, small number of segments, segments of roughly equal size and their geographical concentration, the absence of socioeconomic differences, a tradition of elite accommodation, overarching loyalties between the segments and the presence of exter- nal threats (Lijphart 2008, p. 51). For enhancement of the logical structure of this study, the factors concerning the size of segments, including the absence of a solid majority, segments of roughly equal size and small number of segments will be summarized under the chapter com- position of segments. Furthermore, external actors will be used as an additional factor, as they also pose an essential part in the implementation and often maintenance of power sharing in post war countries (see 2.8).
A favorable factor, which has continually been mentioned in Lijphart’s studies, concerns the size of the respective country. He concludes that smaller states are more likely to become consociational democracies in contrast to larger countries. Lijphart bases his assumption on the classification by Val R. Lorwin, who characterizes nine countries with either having a high or medium degree of segmented pluralism. As can be seen from Lorwin's studies, "all of the smaller states have become consociational democracies in contrast to all the larger ones" (Lorwin 1971, p. 148). According to Lijphart, the smallness of a state goes along with various implications. In Democracy in Plural Societies, Lijphart (1977, pp. 65-70) attends to the dif- ferent effects of size on consociationalism. He distinguishes between external and internal, as well as direct and indirect ramifications.
The direct internal effect of smallness increases the likelihood of elites knowing each other personally. Thus they will meet more often and are likely to reach an agreement, because they don't view politics as a zero-sum game.
The direct external effect causes small countries to feel threatened more easily by larger coun- tries, which increases the internal cohesiveness of the country. "The political leaders will tend to draw together and their followers are also more likely to approve of intersegmental cooper- ation in the face of grave external threats" (ibid., p. 66).
The indirect internal effect of small size countries facilitates the governability of the country. This is due to the fact that smaller countries tend to have less complex systems as Lijphart states. As size decreases the number of individuals and the diversity of their interests and atti- tudes, which are represented and mirrored by politicians and decision-makers, declines.
Lastly, the indirect external effect of smallness leads Lijphart to believe that small size coun- tries are less likely to take part in international politics. This reduced decision-making load is attributed to the limited political power they combine in themselves. Therefore they mostly refrain from foreign policy, which may even lead to a policy of neutrality. "If a small country is a plural society, its size and its plural nature, which entails the danger of internal conflict spilling over into the international realm, jointly increase the likelihood of imposed neutrality" (ibid., p. 69).
In his study on South-Africa, Lijphart (1985, p. 480) refers to the population size rather than country size concerning the smallness factor. Moreover, in Lijphart’s (2008, p. 52) latest study on Indian democracy he refers both to population and country size. Therefore, the fol- lowing study will examine both the territorial size and the size of the population of the respec- tive country.
2.2 The composition of segments
In addition to the country’s size, it is important to assess the composition of the present seg- ments.
One of the biggest obstacles for consociational democracy is the existence of a stable majority of one segment. According to Lijphart, for leaders of a majority group the establishment of a majoritan system might be more conducive than consociationalism (Lijphart 2008, p. 51).
Instead, segments should be roughly of equal size, which leads to a balance of power among them, and enhances the probability for cooperation during negotiations (Lijphart 1977, pp. 55 ff.). In addition to the assumption of the importance of size of the groups, it is also critical to take the stability of the groups’ sizes into account, when analyzing a country. As Roeder and Rotchild (2005, pp. 46 ff.) state, unstable demographic composition of groups can lead to in- creased destabilization of the power sharing institutions. If size changes among groups, caused for example by different birth-rates between groups, displaced people or a rise of ref- ugees within the country caused by war in neighboring states, some ethnic groups may demand to readjust their representation in the power sharing institutions in order to reflect the changed demographics of the society. On the other hand, the currently well- over overrepresented group will try to maintain their status quo in the system. Thus, differences between the different groups may arise and hinder the democratization of the country.
Another factor closely connected to this factor is the number of segments that exist within a society. Lijphart states the optimal number for segments between three and five. The mini- mum of three segments derives from the assumption that a dual segmentation is prone to lead to hegemonial power struggles between the segments. "[I]n a society with two segments [...] the leaders of both may hope to win a majority and to achieve their aims by domination rather than cooperation” (Lijphart 1977, p. 55). On the other hand, if there are a larger number of segments, other difficulties may arise. “The reason is that cooperation among groups becomes more difficult as the number participating in negotiations increases” (Lijphart 1977, p. 56).
2.3 Geographical concentration of segments
Moreover, the geographical concentration of segments provides a favorable factor for conso- ciationalism. "[...] [C]lear boundaries between the segments of a plural society have the ad- vantage of limiting mutual contacts and consequently of limiting the ever-present antagonisms to erupt into actual hostility" (ibid., p. 88). In addition, a territorial concentration of groups can lead to an endorsement of group autonomy by applying federalism (Lijphart 2008, p. 52).
2.4 Socioeconomic differences
Another favorable factor of particular importance is the absence of socioeconomic inequality. "If there are large socioeconomic differences among the segments, the poorer segments will likely feel discriminated against and the more prosperous ones may feel threatened” (Lijphart 198, p. 124). Socioeconomic inequality refers to the fair distribution of economic as well as human resources. Those encompass a similar standard of living, level of employment, income as well as equal access to and level of education. If these standards are not met, grave tensions may arise between the different segments and threaten the viability of consociationalism. Lijphart further states that this problem can be mitigated through political action via a fair redistribution of wealth (Lijphart 1985, pp. 124-125).
2.5 Tradition of elite accommodation
In addition to the self-negating prophecy6, which is a conducive factor for elite-cooperation, the prior existence of a tradition of elite cooperation may facilitate collaboration among the elites. Although Lijphart acknowledges the fact that a history of moderation and compromise prior to the establishment of consociational democracy can be a favorable factor, he states that it is not sufficient without the self-negating prophecy. A historical tradition of accommodation and the appearance of a self-negating prophecy complement rather than exclude each other. Therefore he puts an emphasis on the voluntary and rational element of consociationalism opposed to fixed predispositions (Lijphart 1977, pp. 99-103).
2.6 Overarching Loyalties
Another favorable condition Lijphart mentions is the existence of overarching loyalties, when they provide cohesion within the society. He suggests that nationalism may provide such a cohesive force, since a shared national identity fosters solidarity and a sense of belonging to a common polity among the population and elites. In return, this sense of belonging enhances the probability of establishing and consolidating a democratic state. However, if nationalism is only connected to an ethnic group instead of the state, it may divide the society even greater along ethnic lines (Lijphart 1977, pp. 81-83). As Starčević-Srkalović (2010, p. 172) points out, the absence of overarching loyalties is closely connected to the stateness problem. „A stateness problem exists when a significant proportion of the population does not accept the boundaries of the territorial state, whether constituted democratically or not, as a legitimate political unit which they obligate compliance“ (ibid., p.172). In addition, as recent examples show, the prospect of integration within the international community can create a substitute for the lack of a shared national identity (see 3.8).
2.7 External threats
The existence of external threats has already been mentioned in Lijphart’s study on the size of countries (see 2.1 external effect of smallness). Lijphart (1977, p. 67) stresses that this pre- sumption is tied to the premise that "such a threat must be perceived as a common danger by all of the segments in order to have a unifying effect". Therefore, it is the author’s reasoning in this study that external threats are -especially in the cases of Bosnia and Macedonia- closer connected to the existence of overarching loyalties than the smallness factor. If overarching loyalties are missing, external threats may act as a disruptive force, when “[…] a minority feels more loyalty to the ‘external’ homeland, […] and [doesn’t] attach enough loyalty to the state of residence and citizenship” (Slaveski and Atanas 2012, p. 25). In result external conflicts tend to be replicated internally.
2.8 External actors
The role of external actors is not included within the framework of favorable conditions by Lijphart. John McGarry and Brendan O'Leary (2004, p.9) remark this flaw by stating that “[a] related lacuna in traditional consociational theory is that it neglect[s] the possibilities of posi- tive roles for outsiders both in the implementation and in the active operation of power shar- ing settlements”. Since external actors have been playing a vital role in the democratization process of Bosnia as well as Macedonia, it will be added as a further factor in this analysis. “It is difficult to classify [them] as being a favorable or unfavorable factor in the democratization process” (Starčević-Srkalović 2010, p. 175). External actors usually provide the foundation for peace frameworks -classifying them favorable-, but may also hinder or slow the process, once peace is reached, making them an unfavorable factor. The following analysis on the case studies of Bosnia and Macedonia will therefore consider their role from different angles.
As mentioned before, one of the most important criticisms of the favorable factors concerns the non-binding character of the favorable conditions provided by Lijphart. The debate is split up between the ‘orthodox’ and the ‘latitudarians’. The ‘orthodox’, including Pappalardo (1981) and Van Schendelen (1985), stress that the factors should be considered necessary conditions for the implementation and preservation of consociationalism. In contrast, the ‘latitudarians’ including Lijphart himself, emphasizes that the favorable factors are merely helpful, but not indispensable for the establishment and maintenance of consociational democracy (Bogaards 1998, pp. 475 ff.).
Van Schendelen (1984, p. 34) therefore criticizes “Lijphart’s […] conditions may be present and absent, necessary and unnecessary, in short conditions or no conditions at all” and consequently deems them useless. Papallardo (1981, p.384, italics added) follows a similar logic by stating that “[a] condition must be present in all the involved countries”. Otherwise he considers the concept of favorable conditions pointless.
In rebuttal, Lijphart (1985, p. 114) claims that thinking about the favorable conditions to such extremes doesn’t meet the realities of social sciences, since “virtually all social science knowledge is probabilistic in nature [and] [h]ardly any relationships between variables can be stated in absolute terms.” Therefore, favorable factors don’t need to make the claim of being perquisites for consociationalism, but it is sufficient to say “[…] that the presence of [these factors] makes successful consociationalism more probable and its absence makes it less probable” (1985, p. 114).
Another controversial aspect is the assignment of scores for each factor, which Lijphart intro- duced in his study on South Africa (1985). The range reaches from -2 till +2, indicating fac- tors either as very unfavorable (-2), unfavorable (-1), neutral (0), favorable (+1) or very favor- able (+2) according to his judgement. Lijphart (1985, p. 127) states that “[t]he total scores should obviously be taken with a grain of salt, [since] changes in individual ratings [- including the weight, which is attached to each factor-] will affect the totals.“ Even though, he acknowledges the possibility of different scoring outcomes, he doubts “that such changes will […] deviate much from the pattern [he] found” (ibid., p. 127). As Bogaards (2004, p. 485) points out this is a faulty assumption.
John McGarry and S. J. R. Noel (1989), and many with them, predicted a (black) majority segment in a democratic South Africa where Lijphart judged minority parties of more or less equal size the most likely outcome of the first free elections. This makes a difference of seven points and South Africa would in this interpretation end well below zero with a score of -6 […]. In addition McGarry and Noel assign more unfavourable weight to socio-economic inequality (ibid., p. 485).
These results show that the allocation of points does not lead to the hoped increase on the forecasting power provided by the favorable factors. On the contrary, it rather furthers the diffuseness of the conditions. For this reason, the following analysis will not use the five- point scale.
3 The concept of favorable factors applied to Bosnia and Mace- donia
Following, the main part of this study will dedicate one chapter to each factor. First in order Bosnia will be up for review, followed by Macedonia. Each chapter will end with a short conclusion on whether the reviewed factor can be considered favorable or unfavorable for power sharing in the respective country. The chapter conclusion will also provide a brief comparison between Bosnia and Macedonia.
The focus of the study covers the investigation period from 1991 onwards. The phase between 1991-1995 in Bosnia and 1991-2001 in Macedonia encompasses the crucial time prior to the implementation of power sharing. The beginning of the period starts with the declaration of independence of the respective country, which is followed by the subsequent beginning of violent hostilities in Bosnia (1992-1995) and Macedonia (2001). Moreover, the time period after the implementation of the DPA (1995) and OFA (2001) casts light on the current diffi- culties in the development and maintenance of the established power sharing structures.
Bosnia’s territory is comprised of 51,197 square kilometers, which is divided into the Bosni- ak/Croat Federation (about 51 percent of the territory) and the Bosnian Serb- Republika Srpska, short RS (about 49 percent of the territory). The territory is inhabited by 3,871,643 people. Its population is made up of 48 percent Bosniaks7, who are mainly Muslim and speak Bosnian. The second biggest group within Bosnia is comprised of Serbs, who make up 37.1 percent of the population. They are predominantly Orthodox and speak the Serbian language. Moreover, Croats make up 14.3 percent of the population, making them the third biggest group. They are mainly Roman Catholic and speak Croatian. Others, including e.g. Roma, provide with 0.6 percent the rest of the multiethnic society (Central Intelligence Agency 2014a).
Like Bosnia, Macedonia is relatively small in size of its territory as well as in size of its population. Its total area encompasses 25,713 square kilometers inhabited by 2.091.719 people. Macedonia's ethnic population is made up of 64.2 percent Macedonians, who are mainly Macedonian Orthodox and speak the Macedonian language. The second biggest group within Macedonia is comprised of ethnic Albanians, who make up 25.1 percent of the population. They are predominantly Muslim and speak Albanian. Turks (3.5 percent), Roma (2.7 percent), Serbs (1.8 percent) and others (Vlachs and Bosniaks with 2.2 percent) make up for the rest of the multiethnic society (Central Intelligence Agency 2014b).
The factor of smallness, both based on its territory and the size of its population is therefore fulfilled in both countries. In contrary to Lijphart’s assumption, the countries’ small sizes didn’t have immediate, positive effects.
Contrary to Lijphart’s prediction, the direct internal effect has not led to an increase in reached decisions between the different segments. While parties know each and meet on a regular basis, decisions were often blocked by one another. This is shown by the governmen- tal crisis, which happened in Macedonia in 2006. After elections, the largest Albanian party, Democratic Union for Integration (DUI), had obstructed parliament, because the biggest Mac- edonian party, Internal Macedonian Revolutionary Organization - Democratic Par- ty for Macedonian National Unity (VMRO-DPMNE), had refused to form a coalition with it. Nationwide demonstrations were followed by a nearly four month absence of the DUI in par- liament. In Bosnia, the situation is even worse. Key decisions were only reached with the help of external actors and the Office of the High Representative (OHR), an administrative office created by the EU, which is still present within Bosnian political life.8
The direct external effect has been quite contrary to Lijphart’s assumption, who states exter- nal threats may provide cohesiveness. As will be shown later (see 3.7), ultimately external threats have led to war in Bosnia from 1992-1995 and to armed conflict in Macedonia in 2001.
As opposed to Lijphart’s indirect internal effect, which states small countries with easier gov- ernability due their simple governmental structures, Bosnia has a very complex system of government. “With 14 constitutions and governments with legislative powers, the political system of Bosnia is both inherently complex and asymmetrical“ (Bieber 2006, p. 40). Due to this high degree of decentralization, the system is highly costly and is therefore effectively hindering economic development and overall governability of the country9. As Starčević- Srkalović (2009, p. 176) claims, current general government expenditure regularly exceeds the country's Gross Domestic Product (GDP) over 50 percent. However, being not quite as complex as Bosnia’s system, Macedonia’s political system is also highly decentralized due to the power sharing arrangements through the implementation of the OFA. Furthermore, the governability of both countries is aggravated by the ‘brain drain’, which has been happening in both countries. The war in Bosnia created a massive ‘brain drain’, resulting from the vast amount of people who left the country. According to the Anadolu Agency around 150,000 young people have left Bosnia since 1995. The majority of these emigrants is comprised of skilled and highly skilled individuals who are often doctors or hold diplomas in IT or engi- neering. ‘Brain drain’ is also a problem in Macedonia.
1 According to studies, fatalities caused by the Bosnian war are estimated at around 100,000 (Lampe 2015).
2 Hereinafter it will be referred to as Bosnia.
3 The terms conditions and factors are used interchangeably by Lijphart and will therefore also be used interchangeably in this study.
4 The majority of political scientists relies on the consociational model, provided by Lijphart, because it is based on empirical evidence, in contrast to the integrative approach of power sharing formulated by Donald Horowitz (see Horowitz 1985 for further information). While there are numerous examples of consociationalism including Switzerland, the Netherlands and Belgium, there is a lack of empirical examples for the broad approach of inte- grative measures addressing ethnic conflict management. Moreover, both countries analyzed, feature more as- pects of the consociational system. Since this thesis will feature the favorable factors of power sharing, instead of power sharing itself, see Bieber 2006; Friedrich Ebert Stiftung 2008 for further information on the details of the specific consociational elements in the constitutions - including the DPA and the OFA- and governmental pro- cesses of Bosnia and Macedonia .
5 In this study the words “segment”, “ethnic group” and “group” will be used interchangeably.
6 “In deeply divided societies, the realization that further escalation of a conflict will result in mutually damaging outcomes has been referred to by Lijphart as the ‘self-negating prophecy”(Sisk 1996, p. 23).
7 “Bosniak” has replaced Muslim as an ethnic term to avoid confusion with the religious term Muslim. In con- trast, the term “Bosnians” describes the people of Bosnia and Hercegovina, regardless of their ethnic affiliation.
8 The role of external actors and especially the OHR in Bosnia will be discussed in greater detail in 3.8.
9 For further information on the Politics and Governance in Post-War Bosnia see Bieber 2006, p. 40 ff.