Abstract

Twenty five years after the conclusion of Dayton Peace Agreement (DPA), Bosnia and Herzegovina (BiH) can still be considered a divided society, structured alongside a foundational ethnic cleavage. The role of the three Constituent People, Bosniaks, Serbs and Croats, is recognised and guaranteed by the Constitution that defines a rigid consociational asset of the Bosnian institutions.  While the crucial influence of the ethnic cleavage on the Bosnian institutions and parties is evident, it might still be worthy analysing how the issue is perceived by the broader Bosnian population, from a bottom-up perspective. Thus, undertaking an empirical analysis is the only way to understand how this ethnic cleavage has been rooted among the people of Bosnia and Herzegovina. Although the results may be affected by the relatively low number of participants and possible selection bias that may limit their generalisation, the findings of this analysis confirm that Bosnians have different perceptions of crucial political issues, including the very same idea of State and citizenship, depending on their ethnic identification. Thus, ethnicity seems to be a stronger factor of identification than national citizenship, confirming the idea of Bosnia and Herzegovina as a divided society.

Key words:Bosnia and Herzegovina, Dayton peace agreement, divided societies, ethnicity

Theoretical Framework

Bosnia and Herzegovina can certainly be defined as a “plural society”. Several definitions elaborated in order to define plural or divided societies do not present substantial differences, therefore we may just adopt the one given by Alvin Rabushka and Kenneth A. Shepsle, (1972, 21) whereby: “the hallmark of the plural society, and the feature that distinguishes it from its pluralistic counterpart, is the practice of politics almost exclusively along ethnic lines. To put the emphasis differently, in the plural society - but not in the pluralistic society - the overwhelming preponderance of political conflicts is perceived in ethnic terms. Permanent ethnic communities acting cohesively on nearly all political issues determine a plural society and distinguish it from a culturally heterogeneous, nonplural society. […] To summarise, a society is plural if it is culturally diverse and if its cultural sections are organized into cohesive political sections.”

The adjustment of the Bosnian political life around an ethnic cleavage has been evident since the very first free competitive elections held in the country in 1990. The electoral outcome sanctioned the victory of ethnic nationalist parties and the definitive collapse of the League of Communist Party of Bosnia and Herzegovina (SKBiH), which was considered to be the most orthodox Communist and conservatives party of all the Republics (Toal, Maksic, 2014, 268) and was already strongly weakened at the time of the election by cases of corruption such as the Agrokomerc scandal (Bieber, 2014, 552). Unfortunately, the dimension of the ethnic conflict in BiH did not remain restricted to the political arena, and after the declaration of independence, in March 1992, it erupted in a military civil conflict that was tearing apart the country for nearly four years. The systematic use of violence against the civil population and the actions of ethnic cleansing radically changed the demography of BiH, adding a new meaning to the term “divided society”. Indeed, in 1995 the division between the ethnic groups was no more merely political, but political and geographical: the deliberate use of violence against civil population reached its aim, leading to the creation of territorial regions whose ethnic composition is homogeneous, or “cleansed”. The territorial segregation represents a huge disruption in Bosnian history, which would always see members of different ethnic groups coexisting in the same areas. True, interethnic violence had already occurred during the second world war, but, nevertheless, in the socialist BiH that emerged from the conflict the different ethnic groups kept living together in the same cities and regions, and BiH preserved the tradition of mixed marriages, to the point that by 1990 around the 40% of Bosnian urban couples were ethnically mixed (Donia, Fine, 1994, 186).

The deep and apparently unbridgeable ethnic cleavage dug by the war has been definitively frozen and consolidated by the Dayton Peace Agreement (DPA). The Bosnian Constitution, which contains the norms concerning the structure of the main Bosnian institutions and the State itself, is indeed part of the DPA, as Annex IV of the Peace Agreement. Cornerstone of the Constitution is the recognition of the three Constituent People of the Bosnian State: the Bosniak, the Serb and the Croat people. The identificationof an individual as member of one of the three peoples is an individual choice that does not require any probation of belonging or acceptance from the group itself, but entirely rely on a system of self-classification[i]. The State structure drawn in the DPA is a “complex” State (Grewe, Riegner, 2011, 26), a hybrid halfway between a confederation[ii] and a federation model, to be analysed on two levels. The top level of analysis is the central State of Bosnia and Herzegovina (BiH), while on the other level there are the two sub-State Entities that Bosnia and Herzegovina is divided into: the Federation of Bosnia and Herzegovina (FBiH) and the Republic of Srpska (RS). Each Entity has its own assembly, president and ministers, and owns large autonomies. An important detail that should be taken into account is the demographic composition of the Entities: RS’s population is composed of the 81, 5% of the Serbs, while the number of the Bosniaks and Croats reaches the 92, 1% of the population of the FBiH[iii]. The borders of the entities basically follow the lines of the battlefront frozen in 1995, and also because of a partial failure of the policy of the war refugees return, the human geography of the country has irreversibly changed: only a small percentage of Croats and Bosniaks remained in the Serb entity, and the same appears to be true for the Serbs in the FBiH (Brubaker, 2013, 1). The demographic composition of the Entities also has a direct effect on the central institutions of the country, characterised by a constitutional architecture that obsessively reproduces consociational mechanisms of power-sharing (classification of Bosnia and Herzegovina as a case of consociational democracy can be found in: Lijphart, 2004, 99). The Presidency of the Republic is a collegial organ composed of three members, one for every Constituent People: the ethnic belonging to the Croat, Serb and Bosniak people is indeed a necessary requirement for the election. The same dynamic occurs in the upper house, the House of People, composed of 15 members, five of whose must be Serbs elected from the territory of the RS, and five Bosniaks and five Croats elected from the territory of the FBiH. The ethnic identity as the requirement for the access to some the highest offices of the State has also been condemned by the European Court of Human Rights, in the case Sjedi-Finci, as discriminatory against those Bosnian citizens who do not declare to be a Serb, Bosniak or Croat[iv]. Furthermore, the lower house, the House of Representatives, is composed of 14 representatives elected from the RS and 28 from the FBiH. The ensemble of these electoral prescriptions reduces by far the necessity for the candidates to look for the votes of the members of ethnic groups different from theirs, encouraging, on the other hand, higher levels of interethnic conflict in the political speech (Dzankic, 2015, 535). It must be noticed that the situation is even more complex than described above, besides the two Entities, the presence of the Brcko District must be taken into consideration. It is located in the crucial point of junction between the two portions of the RS, with the status of condominium subject to the territorial sovereignty of both Entities, but self-ruled by the local institutions[v]. Even though the District has large autonomies, the inhabitants must choice between the citizenship of the RS or the FBiH in order to exercise their political rights (Stjepanovic, 2015, 391), and they are subject to the same rules in the matter of affiliation or not to one of the Constituent People.

Nowadays, the ethnicity can indeed still be considered as the main fracture of the Bosnian society and the foundational cleavage of the Bosnian party system. The presence of two-levels institutions also implies the existence of a two-level party system, with an Entity-level, infraethnic dimension of political confrontation and a State-level, multi-ethnic one, taking place in the consociational central institutions (Sebastian, 2010, 324). The outcomes from the first post-war elections in 1995 to present days show ethnic, nationalist parties winners at the Entity level, following the so-called by Donald L. Horowitz (2000, 359) theory of the ethnic outbidding. Ethnic parties, winners in the local arena, are therefore able to bring their instances to the central level, where they confront with the other ethnic parties. The main Serb party is SNSD, whose leader is Milorad Dodik, which in 2006 overtook the war-time SDS of Radovan Karadzic: since then the SNSD has strongly held its leading role among the RS electorate. In the FBiH the SDA and the HDZ are respectively the main Bosniak and Croat parties. In the Federation a relevant role is also played by the SDP, that incorporates Titoist symbols and arguments in its campaigns, clearly states its multi-ethnicity and pushes an anti-nationalist agenda (Hulsey, 2010, 1134), but that in the spiral of radicalisation that takes place in such a polarised context ends up being identified as a mainly Bosniak party, therefore promoter of Bosniaks’ interest (Kivimäki et al., 2012, 110; Toquet, 2011, 459).

Applying the categorisation of Giovanni Sartori (1976, 116-119), the Bosnian party system could therefore be considered as a polarised pluralism: the consociational constitution, characterised by extensive power-sharing mechanisms and veto instruments at the disposal of the ethnic actors, the so called “ethnic vetoes” (Morrison, 1996, 145-157), gives to multiple parties a significant blackmail power. These powerful actors are promoters of divisive policies and of conflictual visions of the Bosnian state itself. Bosniak parties, carrier of the interest of the majoritarian group, promote the idea of a centralised, unified BiH, and some of the more radical Bosniak exponents even called for the abolition of the RS (Kivimäki et al., 2012, 110), defining it as a “genocidal creation” (Moll, 2013, 915). The counterpart to the Bosniak unitarian speech is the Dodik’s call for a secession referendum for RS: first thrown out by Dodik during the general elections of 2006 (van Willigen, 2013, 148), the threat of a referendum has periodically remerged in occasion of electoral campaigns and in times of higher conflict (Chivvis, 2010, 97-110), as a way to disarm fellow Serb radical opponents and as instrument of bargain with other ethnic groups (Toal 2013, 179). Indeed, even if the actual call for a referendum seems unlikely in the lack of support from Serbia (Kartsonaki, 2016, 506), the sole threat of it can be considered as an instrument to counterbalance the centralist impulse coming from the Bosniak part, and as an instrument to rise internal consensus (International Crisis Group, 2011, 19-20). In such a context the posture of Croat elites assumes a specific importance. In terms of demographic consistence, Croats represent the smallest Constituent People, and as they are not provided with the very same guarantees by the Serbs, in the absence of a Croat entity, the HDZ elites began seeing themselves in a position of disadvantage (Manning, 2004, 62), and at increasing danger of recession, as evidenced by the reactions to the election of Zeljko Komsic and by the “Ljubic case”. Croat elites are marked by the fear of being dominated by Bosniaks, especially in the FBiH, and they have been calling for stronger mechanisms of safeguard: (Sebastian, 2012, 606) this resulted in centrifugal pushes, including calls by Croat representatives for a possible territorial reorganisation of the FBiH[vi] or at least for an electoral reform that would grant the legitimate representation of the Croat constituent people[vii]. Such a political and institutional environment certainly influences the perception that Bosnians citizens have of their fragmented institutions, of the concept of citizenship and of the Bosnian state itself.

Hereby are presented the results of the study with the aim to investigate the long-term consequences of the Dayton order on the identity and ethnic belonging of Bosnian citizens and on the Bosnian idea of citizenship. The question of the prevalence of the ethnic or citizen identity assumes indeed a specific relevance in a divided society, enhanced in the Bosnian case by the existence of the two ethnicity-based Entities.

Study design and data collection

The questionnaire is composed of some identification questions, that will be used as independent variables, and other five main sections. The aim of the first question clarifies which citizenship the individual owns, besides the national citizenship of BiH, which is to be considered as a precondition to be interviewed: the two possible answers are citizenship of the FBiH or the RS. There is the option to insert the third Country citizenship, Croatian or Serbian, in the eventuality that the individual has one of them. Specifically, holding the Croatian citizenship besides the Bosnian one, can often correspond with the identification as a member of the Croat Constituent group, since Croatia adopted an extensive ius sanguinis criteria for the concession of the citizenship, including the great majority of the Croat Bosnians (Sarajlic, 2012, 370). The other important question is about the identification: Bosniak, Serb, Croat or Bosnian. This last answer was given to offer the chance of expressing an identification based on a citizenship criterion rather than an ethnic one, on the model of Yugoslav era questionnaires, that after the “Muslim”, “Serb” and “Croat” answer used to offer a “Yugoslavian” as an answer, often chosen by minorities members (International Crisis Group, 2012, 6-7).

The goal of the first and second sections is to measure the levels of confidence of the interviewed individuals toward national and international institutions. Regarding national institutions, these include the central institutions of BiH, including the national Parliamentary Assembly of BiH and the three members Presidency, the Entity level institutions, including the bicameral Parliament and the Presidency of the FBiH and the Presidency and National Assembly of RS, and eventually the local level of governance that may be the Canton for the inhabitants of the FBiH and the Municipality for the inhabitants of RS. The international institutions are the Office of the High Representative, the EU and NATO: we selected these three subjects because of their deep involvement and of their influence on the politics of BiH. The respondents were allowed to choose a value on a scale going from one to seven, with one corresponding to the lowest level of confidence and seven to the highest. The same scale is employed in the third section, where the respondents were asked to express their positive or negative perception about a list of foreign Countries and Organisations. In the fourth section, through the employment of the same scale from one to seven, the interviewed subjects were asked to express their opinion on five possible options for the future of BiH. The last three, yes-or-no questions close the questionnaire, inquiring the satisfaction regarding the last elections, the inclination of the respondents to vote a candidate from another Constituent People and the willingness to emigrate leaving the country.

A cross-sectional survey was conducted. Study participants were recruited in the FBiH (Sarajevo, Tuzla, Mostar), in the RS (Banja Luka, Bjielijna) and in the district of Brcko, in order to enrol a population sample sufficiently representative of the ethnic composition of the all country. Study participants were recruited through two different modalities: i) direct contact, in parks, universities, and other public locations; ii) internet connection, via mail. In the second case, a snowball technique was used, asking those interviewed to give their name and contacts of friends or acquaintances. The interviews were conducted using the standardised self-administered questionnaire containing ten sections, generally split in more sub-questions. Before filling out the questionnaire, participants were invited to read the instructions and to give their informed consent. The questionnaire was anonymous, and the participants were reassured about the confidentiality of the information. Prior to the study, a pilot phase was carried out in order to assess the quality and the feasibility of the questionnaire, which was translated from English into Serbian and then back-translated into English. It took six months to collect the data, between November 2018 and April 2019.

The following analysis is based on 319 collected questionnaires. Among the interviewed individuals, 142 (44,5%) are male and 177 (55,5%) female, and with regard to their level of instruction, two (0,6%) of them have no formal instruction, two (0,6%) have primary school instruction, 87 (27,3%) have high school diploma, 166 (52%) have a bachelor degree and 59 (18,5%) a master degree, while the remaining three (0,9%) are PhD. The median age of the interviewed individuals is 23 (range: 18-64 years old; interquartile intervals [25^th and 75^th percentile]: 21 and 30 years old). More than a half of the people that were enrolled in the study were born after the conclusion of the Dayton Peace Agreement, in the post-war BiH, and more than three quarters were in very young age at the time, and grew up in a Country already divided in Entities.

As mentioned above, the last two external variables are the citizenship of the Entity and eventual citizenship of third Countries owned by the interviewed subject, and their ethnic or national self-identification. Among them, 138 (43,3%) are citizens of the FBiH, and 96 (30,1%) of the RS, while the others were in possession of a dual citizenship: 50 (15,7%) of them are citizens of the FBiH and Croatia, six (1,9%) of them of the FBiH and Serbia, and 29 (9,1%) of the RS and Serbia. Associated with their identification, 57 (17,9%) identify themselves as Bosnian, 85 (26,6%) as Bosniak, 128 (40,1%) as Serb, and 49 (15,4%) as Croat.

A vast majority of the participants were enrolled in urban areas, and may therefore not be representative of the population of the whole country. Even though the age range taken into consideration is 18-64, three quarters of the people enrolled in the study are no older than 30 years old, with an equal representation of male and female participants. Therefore, it is possible to say that the present study is especially representative of the opinions and perception of the younger Bosnian citizens, who grew up in post-Dayton Bosnia, from urban areas. In the process of data collection, the ethnic composition of the Country was taken into consideration, in order to assure a representation of all ethnic groups, considering the size of their population.

Results

The first step in the analysis consists of generating the main components (Factors) explaining the general structure of the data set from the correlation between original variables (items). The factors were extracted by means of Principal Component Analysis (PCA) and correspond to the eigenvectors of between items correlation. This allows to project the initial multivariate data set into a space spanned by few principal axes, giving account of the correlations among original variables. The resulting Factors are synthetic indexes, mutually independent because of their very construction, so that each Factor represents a latent variable of the system. From the 24 original variables we obtained three main factors capable of explaining the 47% of originally scattered across 24 items. The choice of considering only the first three factors was dictated by the fact that from fourth factor onward (the factors are extracted in decreasing order of variance explained) the presence of noise fuzzified the univocal interpretation of the components. The Factor meaning is interpreted via their loading pattern (Pearson correlation coefficient between original variables and factors). We will consider a correlation as relevant for factor interpretation only if its loading is higher than 0.5 in absolute value (Table 1). The first three factors explain a decreasing amount of the information, thus we have: 25% for Factor 1, 16% for Factor 2 and 8% for Factor 3 respectively. Each factor was assigned a meaningful name on the basis of its loading pattern (Table 1).

Table 1. Factorial patterns

The table reports the Pearson correlation between original variables and factor (loadings). The loadings correspondent to the variables most relevant for factor interpretation are bolded. The factors are each other independent by construction so pointing to separated aspects of the data set, the three factors solution explained collectively the 47% of total information. The three factor solution was selected by Cattell's method, the components of order higher than 3 are indistinguishable from noise. The names attached to the factors stem from the loading pattern.

Descriptive statistics on factor scores (Table 2) allows to detect the strict relation between 'Identification' and relative factor score, whose statistical significance was assessed by Analysis of variance (p < 0.0001 for all the factors). It is worth noting that factor scores have by construction mean zero and unit standard deviation on the entire data set, thus in Table 2 negative values of the mean point to 'lees than average' scores and the opposite is true for positive values.

Table 2. Descriptive statistics

The table reports the descriptive statistics of the factor scores relative to the three Serb, Bosniak, and Croat classes. Keeping in mind that factors have by construction zero mean and unit standard deviation on the entire data set, it is evident how the assignment of Serb, Bosnian and Croat to the first three factors on the basis of their factor patterns is reflected by the score values with the leading class having higher mean in all the correspondent factors. It is worth noting the Croat-Serb opposition (opposite signs with similar absolute values) with Bosniak in the exact intermediate position as for Factor 3 (Croat factor). Factor 1 has a similar structure but reports the Bosnian-Serb opposition, while Factor 2 (Bosnian factor) is a 'unipolar' component opposing Bosnian to the other two groups.

As the second step, through the Chi-square test, we measured the significance of the superposition between “Citizenship”, and “Identification”- that in fact showed a strong correlation Table 3, p < 0.0001). Therefore, it would be reasonable to assume that the significance of “citizenship” is just a mirror of “identification”, and that the determination of “citizenship” is indeed correlated with the “identification” variable.

Once determined the interpretative potential of the “identification” variable toward the above mentioned Factors, we can proceed to establish a relation between these Factors and the specific answers that were given to the question of the “identification” of the citizens of BiH. There were four possible answers to the question: “Bosnian”, “Serb”, “Bosniak” and “Croat”. Three out of four answers describe an identification based on ethnicity, while the first answer, “Bosnian”, describes a self-identification based on the element of the BiH citizenship: not an ethnic identification, but a civic, statehood one. Table 2 reporting descriptive statistics, outlines a predominance of the answers “Bosniak”/“Bosnian” in relation to the Factor 1, of the answer “Serb” in relation to the Factor 2 and “Croat” in relation to the Factor 3. Therefore, we could assign to the Factor 1 the name of “Bosnian/Bosniak Factor”, to the second factor the name “Serb Factor” and to the third “Croat Factor”.

Table 3. Contingency table

Graph 1.

The graphic representation, as the Photo 1, is a useful tool in order to understand how the variable “Identification” is related to Bosnian/Bosniak Factor and the Serb Factor. We can distinguish two clusters in respect to the answers to the Identification question: “Bosniak”, “Bosnian”, and a discrete amount of “Croat” are clustered in the lower, right section of the graphic, while the “Serb” Identification clusters in the upper, left area. As we can see, the two Factors are clearly distinct, identifying different but no opposite answers; in fact, the pattern of answers of the two different components aggregates throughout the same direction.

As mentioned before, we identified three main Factors. Unfortunately, because of the lower explanatory potential of the third Factor, the Croat Factor, we did not consider it in our analysis. Even if the Factor was identified as a stand-alone variable, not enough strong correlations emerged to allow a complete discussion on this Factor. This may probably depend on the lower number of collected interviews and also on the major variability in the answers of the self-identified Croat individuals interviewed in the study, living scattered in towns with different surrounding environments. We will therefore proceed with the analysis of the Bosnian/Bosniak Factor and with the Serb one.

Bosnian/Bosniak Factor

The Bosnian/Bosniak Factor presents a strong correlation with positive levels of trust toward the international institutions: the highest confidence scores are the ones referred to the European Union and NATO. It must be noticed that the confidence scores that the interviewed subjects assigned to the international institutions, in particular EU and NATO, are higher than the ones referred to the local, Entity level and National institutions. Considering this high confidence score of the European Union, the strong level of support toward the process of integration of BiH in the European Union should not surprise at all. Referring to the opinion that the answering subjects have of foreign Countries, a strong feeling of sympathy toward Austria and Germany is evident, followed by Turkey and Arab Countries. Lastly, in the opinion expressed about the possible options for the future of BiH, the Bosniak/Bosnian Factors shows a significant relation with high levels of enthusiasm toward the process of centralisation of the Country. Slightly less marked, but still significant, is the level of appreciation toward the possible overcoming of the division in Entity. This second option about the future of BiH, apparently similar to the first one, is probably less significant because it is more radical: this represents a step further than a vaguer “process of centralisation” and it is probably perceived as harder, if not impossible to reach, considering the current political divisions. What is also evident, in relation to the Bosniak/Bosnian Factor, is that the future of BiH is desirable to be in the European Union. Therefore, the expectation for the future of BiH is the one of a more centralized, unified State, member of the European Union.

Serb Factor

In relation to the Serb Factor, we can easily identify a strong level of confidence toward the Entity level institutions of the RS: with a correlation coefficient of 0.808, this represents the most significant correlation among all the factorial patterns. The amount of correlation decreases when shifting to the level of trust toward the local and State authorities. These results outline a clear preference toward the authorities of the RS rather than anyone else. Therefore, confronting the confidence levels toward national and international authorities emerged in relation with the Bosnian/Bosniak Factor and the Serb Factor it is possible to identify two distinct attitudes. The Bosnian Factor shows more significant levels of trust toward the international authorities, with no indications of special sympathy for local authorities or for the ones of the FBiH, while the Serb Factor shows a clear preference for the authorities of the RS. Regarding the opinion about other Countries, the Serb Factor shows an easily predictable correlation with a positive opinion of Russia and Serbia. On the other hand, even if the Bosnian/Bosniak Factor and Serb Factor show a positive correlation with different Countries, it is relevant that the positive opinion in a Factor toward specific Countries is not matched by a negative opinion toward the ones appreciated in the other Factor. Finally, with respect to the possible future of BiH, there is a clear positive correlation with the hypothesis of a referendum for the secession of RS: an outcome that we could define the opposite of what found about the Bosnian/Bosniak Factor, whose preference went to the perspective of a more centralised BiH.

Discussion and Conclusions

The results of the research prove the existence of two radical opposite views of BiH and of the very same idea of citizenship of the Country. On one side we have the Bosnian/Bosniak Factor, that seems to be related to the idea of a centralized, united BiH, while the Serb Factor shows a widespread feeling among the Serb community of belonging to the Entity of the RS more than to BiH itself.

First of all, it is necessary to analyse the so called Bosnian/Bosniak Factor: as we can say from the very same name given to the Factor, it implies the overlapping of two different elements. Indeed, the Bosnian/Bosniak Factor includes two different typologies of identification: an identification based on the idea of citizenship and civic belonging to a State, the so called “Bosnian” element, and an identification based on ethnic criteria, the “Bosniak” element, that indicates the belonging to one of the three Constituent People of BiH. The dual nature of the Bosnian/Bosniak Factor occurs due to the fact that the answers given by the people who identified themselves as Bosnian and Bosniak do overlap. In the first place, when explaining this phenomenon, we must take into consideration that there are relevant chances that the individuals who identified themselves as “Bosnians” must be part of the Bosniak group or, less likely, Croats identifying themselves on a citizenship rather than ethnic criteria. Indeed, only the 1% of the individuals in possession of the citizenship of the RS, where the 92,13% of the Serbs of BiH live, identified themselves as “Bosnian”, leaving little, marginal chances for Serbs identifying themselves as “Bosnians”. The substantial overlapping seen in the answers given by self-identified Bosnians and Bosniaks can be explained by the fact that the Bosniak politicians are the main bearers of the idea of a unitarian BiH, while the Serb ones tend to defend the current status quo (Touquet, 2011, 459). The Bosniak Constituent People represents indeed the largest ethnic group in BiH, since Bosniaks amount to the 50,11% of the total population of the Country. Bosniaks elites are well aware of the fact that a unitarian country would advantage their ethnic group, that is the largest one, and they are therefore the main promoter of the idea of a centralized Country (Sarajlic, 2012, 375-376).

On the opposite side of the BiH political spectrum stand the Serb elites. The RS, with its ensemble of safeguards and institutions, represents a solid stronghold for the Serb community against any attempt of changing the status quo in favour of the central institutions. As already stated, Serbs represent the 81,5% of the population of the RS, but only the 30,78% of the total population of BiH. This is how the calls for referendum by the Serb leadership must be interpreted, and the results of the research seem to confirm the appeal that this political claims have on the Serb population. According to these results, the perspective of a referendum for the secession of the Entity from the rest of BiH seems to be, among the interviewed Serbs, a more appealing option than maintaining the status quo, and even more appealing than a possible path of integration of BiH into the European Union.

The differences between the Bosniak/Bosnian Factor and the Serb Factor are extended also to the scope of foreign politics, as seen in reference to the judgment that the interviewed people expressed about international Institutions and third Countries. Of course, the Factor is strongly correlated with a positive opinion of the EU and NATO, and among the three international institutions, they have been asked about, the OHR seems to be the one with lowest levels of trust, probably because the Institution, after reaching its peak of interventions in Bosnian politics under the Ashdown mandate in 2002-06 (Domm, 2007, 163), has gradually lost relevance in the Bosnian political framework from the Schwarz-Schilling mandate on (Belloni, 2009, 364-365). In regard to the other Countries, as mentioned above, the Bosnian/Bosniak Factor shows an especially strong positive correlation with Austria and Germany: this can be partially based on to the association of these two Countries to the European Union, but mostly because of the special relation existing between Austria and Germany and Bosnia and Herzegovina. During the Yugoslavian Wars, Austria and Germany hosted, respectively, up to 320.000 and 86.500 Bosnian refugees (Barslund et al., 2016, 4), and to this day the two Countries have represented two of the most important destinations for the Bosnian diaspora. The explanation of the positive scores obtained by Turkey and Arab Countries requires to go far back in time: the historical and emotional bond between the Bosnian Muslim population and Turkey, indeed, dates back to the Ottoman domination of the Balkan peninsula. The XX century proved to have been especially troubled, and witnessed the emerging and retreat of religious and ethnic pushes, but at the end of the century the connection with the Muslim world remerged once again with the formation and rapid growth, in 1990 - the first one to be created among all the Bosnian ethnic parties- of the openly Muslim SDA, that clearly catered only to Muslims (Bieber, 2014, 551).

On the other hand, the Serb Factor shows a clear, positive correlation with Russia and Serbia. The reasons for the bond between the RS and Serbia are pretty obvious: the Republic of Srpska was formed at the end of a conflict perceived by the Bosnian Serb as a defensive war against Croat and Bosniak attempting to separate them from the Serbian Motherland (International Crisis Group, 2011, 11) and even if Serbia does not support any secessionist attempt from BiH, the position of the Serbian presidency in favour of the preservation of the status quo and of the Entity of the RS has been clearly stated[viii]. Then, the relation with Russia seems to be strong, if not in real economic and political terms, at least with regard to the collective narrative that stresses the bond between the Russian people and the Serbian people: the historical narrative of the two Orthodox and Slav brother Countries, that met an especially wide spread during the Kosovo crisis (Miklóssi, Korhonen, 2010, 60-66), is still strong nowadays. Historical and popular narrative aside, the element must also be confronted with a more sceptical attitude of the Serb population and of the élites of the RS toward the European Union: it is no coincidence that the second official visit of Milorad Dodik abroad, just after Serbia, was to the Russian Federation, where he met Vladimir Putin. While the institutions of the Federation of Bosnia and Herzegovina seem more eager to approach the European Union, the RS is more hesitating, looking Eastward whenever possible, to Russia or even to the Far East, as the new highway built thanks to Chinese investments shows.

Before drawing firm conclusion, some limits and biases of the study should be mentioned. First of all, the sample was relatively small; for this reason, a two-steps statistical analysis capable to maximize the information gathered from the study was applied. Secondly, the participants were mostly enrolled in urban areas, thus they are more likely to have higher educational level and access to services that may to some extent influence the perception of the presence and the value of national and international institutions; for this reason, the opinion of the participants might not be representative of the all country, leading to recruitment bias. Furthermore, only a proportion of those who were contacted accepted to be interviewed, leading to possible selection bias. This problem is more likely to occur with the street interview than with the snowball method. Whether this may have affected the generalisation of the results remains undefined.Even though two different methods were used to collect the information, in both cases the questionnaire was self-compiled, and the comparison between the results obtained using the two different methods did not reveal significant differences (data not shown).

The results of the analysis outline how people gave different answers on the majority of the issues taken into consideration in the analysis on the base of their ethnic identification. Therefore, the ethnic identification seems to be the main cleavage in the Bosnian society and the individuals seem likely to assume different positions on political issues on the base of their ethnic belonging, as in the definition of “divided society” that was given at the beginning of the article. BiH citizens identifying themselves as members of different ethnic groups showed different opinions on the idea of the Bosnian statehood and citizenship, on the national and international authorities, on the future of the State of BiH and even on other Countries. The evidence emerged from this research confirms the theory that the ethnic identification is a stronger factor of identification than citizenship: being a member of a Constituent People comes before being a citizen of the State of Bosnia and Herzegovina. Therefore, according to the results showed through this analysis, Bosnia and Herzegovina seems to be, still nowadays, 25 years after the signature of the Dayton agreement, a divided society.