Multicultural Vanguard? Sarajevo’s Interethnic Young Adults between Ethnic Categorisation and International Spaces

Abstract This article analyses how Sarajevo’s young adults from a middle class, interethnic background deal with the rigid ethnic categorisation enforced by state institutions and society. Their strategies (exit, reframing, and partial separation) appear to be unsatisfactory to the actors themselves, and wield generally no influence on the institutions they wish to change. Three factors have been setting into motion this dynamic: first, the difficulty of escaping ethnic group thinking when attempting to reframe ethnic categories; second, the rationality of avoiding open defiance to ethnic categorisation; and third, the young adults’ tendency to centre their life on interethnic and international spaces. As a ‘project elite’, Sarajevo’s young adults are rather separated from society, both discursively and socio-economically.

. Sarajevo is known for its many paradoxes, and has therefore been referred to as 'Europe's jerusalem' (Karahasan 2010, p. 15) or as the place where 'East meets West' (Novakovich 2013). Across the city centre it is not rare to walk by many mosques, Catholic churches, and Orthodox churches standing side by side. Markowitz describes it in the following way: Valuing the plurality of experience in a world where Catholic and Orthodox Christianity, Islam, judaism, and socialist modernity all manifest themselves in the landscape, in family rituals and in public celebrations, in nuances of language, and in people's belongings is what many Sarajevans describe when they invoke the spirit of their city or the soul of Bosnia. (Markowitz 2010, p. 144) Donia (2006, p. 355) noted that, in today's Sarajevo, common life often crosses ethnonationalistic boundaries. Many observers have however remarked that Sarajevo's tolerant character is a merely superficial layer that has little resemblance with the city's daily reality (Pollock et al. 2002, p. 3;Matejčić 2011). The 1992-1995 war is generally described as an 'ethnic war', which has been caused by 'ancient hatred' (Kaplan 1993). This image is ultimately questioned by Hayden, who remarked that Sarajevo's people intermingled for a very long time (Hayden 2002, p. 173). Either way, the war can be said to have negatively influenced the tolerant character of the city. The Dayton Agreement and the ethno-nationalistic politics that it strengthened (Mujkić 2008;Maček 2011, p. 674) created a separation between the three so-called 'ethnic cultures' (Mujkić 2008, p. 42). These groups, however, have always been living 'intermixed but not intermingled', preserving in this sense the salience of community boundaries (Hayden 2007, p. 111). Bosnia's strengthened ethnic separatism has not left Sarajevo untouched. The Inter Ethnic Boundary Line (IEBL) between the Bosnian-Croat Federation and the Serbian Republika Srpska literally runs through the suburbs of Sarajevo (Armakolas 2007).
In this article, we investigate the ethnic relations and the effects of ethno-politics in the daily Sarajevan life, by zooming in on highly educated young adults with an ethnically mixed background. These Sarajevans fall in between the ranks of the city's recognised ethnic categories. Their parents used to identify with the category 'Yugoslavs', and they themselves are (along with many other young adults) critical of ethnic categorisation and of the ethnonationalistic political parties that currently rule the country. This proposition hence questions whether these young, well-educated people raised in interethnic marriages are agents of change vis-à-vis the establishment of more hybrid/open approaches to ethnicity. This article is therefore centred on the following question: how do these young adults deal with specific instances when ethnic background affects their personal life in a country divided according to ethnicity ? We start by describing the way the categorisation system works. What is the institutionalised pressure on Sarajevo's young adults? What are the moments in which their ethnic identity needs to be expressed? After describing the categorisation system, we address the way the informants have dealt with this institutionalised pressure, and their agency in this process. We then illustrate the contradictions apparent in these strategies. We investigate whether their strategies regarding the categorisation keep the system in place, or contribute towards changing the categorisation system. We explore this nexus by looking at how the interviewees assessed the changes they brought about, as well as by looking at the potential impact of their strategies on the wider society. The latter is tentatively assessed by investigating the range of various social spaces they participate in, and attempt to subsequently influence. 1 Building on these insights, we ultimately ask which image of Sarajevo is the most accurate: the one depicting a 'multi-multi city', or one that is basically characterised by separation, in which tolerance represents an out-of-date feature and an outwardly public relations tool. A few criteria led us to classify Sarajevo as a multicultural city. Namely, that, beyond the fact of ethnic and religious groups living in the same city (and diverse ethnic symbols being present in public life), there would also be intermixing (people from different ethnicities living together in the same neighbourhoods), as well as intermingling (social contacts between the ethnic groups). While this ought not to be seen as an either/or question, it ultimately facilitates an exploration of the intermixing and intermingling in the Sarajevan context, while capturing the views that young adults have expressed in relation to this dichotomy.
First, however, we will elaborate upon theories and debates that are central to the scope and purposes of this article. This segment will be followed by a section delving into our methodological approach. In the theoretical section, we present insights from the literature on intermarriages, citizenship, ethnic categorisation, affirmative action, political economy, and 'project elites'. This wide range of sources was crucial to support the analysis of data, which we collected through interviews, participant observation, and web research.

Theoretical framework
The majority of our informants are children from ethnically 'mixed' marriages. As members of a potentially progressive and open segment of society (which also includes journalists and young NGO workers who are not from mixed marriages, for instance), they are regarded here as an interesting group to study changes in categorisation and citizenship. The issue of intermarriages itself does not sit at the core of the article. However, it is important to note that many authors (Bringa 1993;Anderson 1999;Renn 2000;Lindenberg 2005;Song 2010) have argued that the identity of children from interethnic marriages is treated as relatively fluid and changeable. We will briefly address theories of citizenship, ethnic categorisation, affirmative action, as well as the literature on the impacts of political economy and class aspects on ethnicity and nationalism. Finally, we will introduce the concept of 'project elites' to analyse the economic and class aspects of our target group: Sarajevo's interethnic urban young adults.

Citizenship and ethnic categorisation
We make a distinction between externally and internally defined citizenship (jenkins 1994, pp. 218-19). Externally defined citizenship contains a categorisation created by government regulations or, alternatively, social processes. Internally defined citizenship, on the other hand, relates to individual sense(s) of belonging. Ethnic identity, as argued by jenkins, is a social identity that is greatly impacted by externally located processes of social categorisation (jenkins 1994, p. 197). As we will show later, this study is ultimately the outcome of a clash between internal and external definitions of identity (Spaskovska 2012). With regard to Bosnia & Hercegovina, externally defined citizenship entails what Spaskovska called 'ethnizenship'-'where citizens realise their rights, duties and participation in the public and political sphere solely as members of ethno-national or religious communities' (Spaskovska 2012, p. 385). 2 According to Sarajlić (2012), dual citizenship is a particularly salient characteristic of the regional citizenship constellation in the former Yugoslavia, mainly in relation to the citizenship arrangements with neighbouring states (Serbia and Croatia in primis) of which many Bosnians hold citizenship. The Dayton Agreement created a 'two-tiered citizenship', with both statebased and entity-based levels of legal belonging (Sarajlić 2012, p. 371). 3 Sarajlić emphasises: 'Bosnian citizenship needs to be seen in the context of broader, overlapping and competing visions of citizenship, state-building and nationhood rather than as a self-sustaining linear path to political identity' (Sarajlić 2012, p. 368). This implies, for example, that one can hold a Croatian citizenship while living in Bosnia & Hercegovina. In this context, the terms ethnicity and citizenship overlapped in various situations.
Affirmative action-intended here as the minimisation of existing inequalities on the basis of group identity-represents a specific form of externally defined citizenship. Middleton and Shneiderman (2008) showed that state-sponsored initiatives to favour marginalised groups via affirmative action created identity problems and conflicts in regions featuring relative interethnic unity. In their study, the politics of positive discrimination triggered a sort of competition between different ethnic groups, which all wanted to be recognised as being part of the marginalised group, to be able to share in the resources. Middleton and Shneiderman also noted: 'the divisions between groups continue to deepen as political allegiances increasingly map onto cultural differences' (Middleton & Shneiderman 2008, p. 41). In De Zwart's views (2005) this represents a 'dilemma of recognition', which usually emerge vis-à-vis the failure of anti-inequality policies.
Illustrating the results of a study focused on Mostar (Bosnia & Hercegovina), Vetters remarked that 'on the one hand, newly created administrative hierarchies of provision can discursively be articulated as an opposition between original inhabitants and displaced newcomers … [but] can also become a powerful incentive for collective action, and an avenue for imagining and enacting local forms of citizenship that transcend ethno-national boundaries' (Vetters 2007, p. 203). Vetters argues in this sense that aspects of externally defined citizenship in Bosnia & Hercegovina-specifically the administrative categorisation practices that are not based on ethnicity (such as, for example, the categorisation of Internally Displaced Peoples)can create a shared experience and a 'source for imagining a multi-ethnic, civic community' (Vetters 2007, p. 187). In other words, externally defined citizenship can generate a sense of belonging, and ultimately an internally defined form of citizenship, which goes beyond ethnic distinctions.

Responses to external categorisation
Zahra's theory of 'national indifference' outlines one peculiar way of reacting to external categorisation. Here, Zahra argues that we should 'analyse inaction, evasion, and indifference as potential forms of political agency' (Zahra 2010, p. 113), as 'indifference to politics has rarely been entirely apolitical and has instead carried multiple possible political meanings' (Zahra 2010, p. 114). Fenton argues that national membership 'does not guarantee enthusiasm for the "nation" and it cannot be taken as a signal of nationalism' although in much of the literature it has been treated as such (Fenton 2007, p. 321). Billig, in very similar terms, remarks that nationalism is 'not a flag which is being consciously waved with fervent passion, it is the flag hanging unnoticed on the public building' (Billig 1995, p. 8). Not belonging to one of the recognised ethnic groups makes some of the informants indifferent to and, most of all, disapproving of Bosnia's (nationalistic) politics. However, returning to Zahra's (2010) point, this does not imply that they are apolitical. Conversely, we would argue that this is an explicitly political act, which comes to encompass other categories, including naming a child, as illustrated by the choice of many informants to pick a 'neutral' name for their child. We will elaborate further upon this later in the article.

Political economy and 'project elites'
The 'reduction of Bosnian realities to their ethnic dimensions' has to be seen, according to Bougarel, Helms, and Duijzings, as a flaw common to numerous analyses. These analyses tend to reduce the impact of the war to the spatial separation of the country along ethno-national lines, whereas the authors make a plea for a more diverse approach. The use of ethnic statistics and maps by organisations intending to 'undo ethnic cleansing' risks reinforcing what these organisations are ultimately designed to fight (Bougarel et al. 2007a, p. 13). This article is focused on strategies of the management of ethnic categorisation; its core focus, however, is not narrowly placed on ethnicity, which ultimately downplays other factors such as social status. To this end, we have taken into account the impact of the informants' economic and political circumstances, by situating the informants and their ethnic-nationalist orientations in terms of class/social status. Our analysis will therefore address critical studies on Bosnia's transformation from a political economy perspective (Donais 2003;Pugh 2006). More broadly, we will draw on studies on the effects of neoliberal policies on the local population in (South) Eastern Europe (Visser & Kalb 2010;Dale 2011;Kalb 2011;Vetta 2011;Music 2013). In addition, we will draw on the literature that linked nationalism and ethno-politics with neoliberal policies and their socio-economic effects (Ost 2005;Kalb 2011). As our study is predominantly focused on the middle class of urban interethnic young adults, we will draw extensively on the concept of 'project elites', in order to gain insight into the economic and class-based aspects of their strategies for dealing with ethnic categorisation.
In the post-transition era, a new kind of elites has reportedly emerged across the Balkan region. One of these elites is that which Sampson calls the Euro-, or project elite. Members of the project elite work in foreign-dominated private firms or organisations, have cosmopolitan attitudes, consumption patterns, and lifestyles, and are primarily employed in the aid business (Sampson 2002, p. 300). Sampson also notes: 'They are conscious of their transnational aspects, of being at once both within and without; they may live in a country but are not of it. They may speak the local language but they often act within a completely different code' (Sampson 2002, p. 301). These project elites strongly support Western ideas, speak English, and are externally focused. Their strategy is 'intense participation in global civil society networks' (Sampson 2002, p. 310).
Closely linked to the notion of a project elite are the debates on the dichotomy between the supposed 'progressive urban' and the 'primitive village' spaces. In the urban versus rural discussion, Sarajevo would be characterised as more multi-ethnic than the rural areas, which are therefore seen as more ethnically segregated. Armakolas argues that Sarajevans distinguish between the old inhabitants of the city and people who came in 'from the villages' (Armakolas 2007, pp. 89-90): 4 the former, in this context, tend to accuse the newcomers of having negatively influenced the cosmopolitan, multicultural sphere of the city.
Sampson's project elite members have many similarities with the subjects of this article. By displaying their opinions, behaviour, and views on the ethnic structure they have to deal with in daily life, we hope to determine whether they are agents of change trying to help solve domestic ethnic issues or, conversely, a group of more or less isolated individuals focused on transnational organisations and linkages.

Methodological approach
A qualitative, ethnographic methodological approach was followed during data collection and analysis. One of our aims was to address the complexities and subtleties of how Sarajevans cope with ethnic classification, including the limitations set by formal institutional categorisations, and materialised in censuses or job applications. A qualitative approach based on open, in-depth interviews was ultimately needed, in order to allow interviewees to freely express their attitudes, including subtleties, paradoxes, and contradictions.
The data were gathered through ethnographic research, which included semi-structured and open interviews, and a set of participant observations that we conducted in Sarajevo between 2 February and 2 May 2011. Additional data were collected from media and web-based sources between 2012 and 2014. The authors completed a series of more updated media and web searches throughout 2014, in order to include information on that year's 'Bosnian spring' riots. Information from census statistics and policy reports was also used. Interviews were held with 12 young adults of interethnic background, with representatives of three interreligious dialogue organisations (including a choir, the interreligious council, and the Franciscan multireligious and inter-cultural centre), five journalists, and three professors conducting research on interethnic relations. Informal conversations were conducted with some 20 Bosnian young adults and foreigners who had been living in Sarajevo for several years. The young adults interviewed in the 12 semi-structured interviews had an average age of 27 years, with the youngest being 20 years old and the oldest 35 years old. The informants' personal details are anonymised; the names used in this article are fictional. The research focused primarily on well-educated interethnic young adults who can be classified as middle class in the context of Bosnia, holding relatively well-paid office jobs, often for foreign companies and NGOs. The interviewees all held higher education degrees (or were still studying). 5 All the interviewees spoke English (and were interviewed in English), and most of them had a connection with 'the international community'. Four of them worked in NGOs, four in the private sector (two in finance, one in ICT), two had their own company (one in trade, and one in tourism), and two worked in education (one as a high school teacher, the other as a university lecturer). At a time when over half of the young adults in Bosnia were estimated to be unemployed, all but two interviewees had a job, with the latter two still studying for their university degrees. About half of the interviewees owned or rented their own apartments and were in a financial position to travel abroad either within or outside the Balkan region, while the other half of the interviewees lived with their parents.
The range of interviewees, in this sense, was certainly not representative of young adults in Sarajevo, let alone Bosnia. However, the selection bias (highly educated, middle class) is consistent with the article's purposes, insofar as it focuses on that segment of Bosnian society that is potentially most open, progressive vis-à-vis ethnic issues. Participant observation was conducted to observe the visibility of ethnic and religious markers in Sarajevo's public sphere, to build trust with the informants, and to gain information through informal conversations.
In the analysis presented here, we also followed an inductive, ethnographic approach (Cerwonka & Malkki 2007). We started our fieldwork with an open, flexible research design, using open research questions in order to capture the informants' points as fully as possible. Thus, at the start of the data collection no hypotheses were formulated, and there was no predetermined framework for analysing the responses to ethnic classification. The various strategies of ethnic issues/classification described in this article emerged out of the fieldwork and the ensuing data analysis. 6

Demography, census, and entities
The Dayton Agreement ended the war in 1995 but institutionalised, at the same time, ethnic separation in the part of Yugoslavia that then officially became Bosnia & Hercegovina (Mujkić 2008). Bosnia & Hercegovina was divided in two ethnicity-based entities: the Serbian Republika Srpska, and the Bosniak-Croatian Federation of Bosnia & Hercegovina (see Figure 1). 7 The two entities are politically, administratively, and fiscally autonomous, and have their own constitutions. They carry out most government functions independently, and also have their own armies (Bisogno & Chong 2002a).
Elements of Billig's (1995) 'banal nationalism' are present. Through this concept, Billig attempted to convey that, in their daily lives, people are constantly reminded of nationality through certain symbols. For example, the Republika Srpska has its own flag, educational curriculum, and even a short-lived currency (Armakolas 2007, p. 82). 8 These symbols accentuate the differences with other parts of the country. However, '[S]igns of social separation are obvious throughout the country' (Hayden 2007, p. 110), from traffic signs exclusively in Cyrillic or Latin, newly-built churches and mosques, divergent advertisements for restaurant names and brands of beer (Croatian or Serbian beer in some areas, Sarajevo beer around Sarajevo), to monuments of certain war heroes or historical figures. 9 The Dayton Agreement recognised three ethnic groups; Bosniaks (previously referred to as Bosnian Muslims), Bosnian Serbs, and Bosnian Croats (Mujkić 2008, p. 113). As our analysis will show, this arrangement leaves very little room for the establishment of a pan-Bosnian identity. Before the 1992-1995 war, a 'pan-ethnic, citizenship based category of Yugoslavs' had existed (Markowitz 2010, p. 79). The existence of the term Bosniak points to inter-linkages in Bosnia & Hercegovina spanning religion, nationality, and ethnicity, as well as to the changes in perception of ethnic identity that took place throughout the last decades (Bringa 1993;Markowitz 2010). Sell describes the ratio between the different ethnic groups in Sarajevo as follows: Before the war, Sarajevo was one of the most cosmopolitan cities in Yugoslavia. According to the 1991 census, of Sarajevo's total population of 527,000, 49% were Muslim, 30% Serb, 7% Croat, and 14% other, primarily Yugoslavs who were often the children of mixed marriages. (Sell 1999, p. 181) Figure 2 shows the country's ethnic majorities at the time of the 1991 census. It provides an insight into the distribution of the different ethnicities throughout the country, while indicating that Sarajevo was surrounded both by areas with a majority ethnic group, and areas where none of the different ethnicities constituted a majority.
After the war, Sarajevo's population decreased to 350,000, with 90% of inhabitants being Muslim and only 5% Serbs (Sell 1999). This demographic shift was directly caused by the war-time displacement of a significant amount of people (Bougarel et al. 2007a). Bisogno and Chong argued, on the basis of World Bank data, that displacement also had an economic aspect: as it is more difficult to find a job as a member of an ethnic minority, there is an (extra) impetus for people to move out of areas where they are not part of a majority (Bisogno & Chong 2002b). To capture the shift in Sarajevo's 'ethnic composition', Table 1 compares the results of the two censuses held across the city in 1991 and in 2002. As we will argue in greater detail below, changing ethnic composition was not exclusively stimulated by instance migration but was also brought to the fore by wider changes in classification. In particular, the decline in the 'other' category (from 13% to below 5%)-to which people from mixed marriages tended to ascribe-appears to (also) result from respondents being classified (or self-classifying) as belonging to one of the three ethnic categories. Hayden argues that pre-war interethnic sphere in rural Bosnia consisted of 'respect, reciprocity, but also separation' (Hayden 2007, p. 111). In his views, local Muslims, with the help of the international community, created the image of 'Bosnia the good' (Mahmutćehajić 2000); 10 a place where 'unity in diversity' had always existed. Numerous foreign intellectuals contributed to reinforcing this image (Hayden 2007, p. 112). Coexistence alone, however, 'did not mean that the peoples of Bosnia considered themselves to be one nation, a collective body with common interests' (Hayden 2007, p. 113). Since the onset of the Ottoman era in the fifteenth century, the country has never been ruled by an independent government led by Bosnians, as it has always been included in empires (Ottoman or Austro-Hungarian) or formed part of a larger state (Yugoslavia) (Hayden 2007, p. 111). In 1974, constitutional changes 'established (con)federal relations between Yugoslav republics as nation states of their respective ethnic majority' (Sarajlić 2012, p. 368), paving the way for the creation of tighter conceptual and practical links between ethnicity and citizenship. Sarajlić explains: A strong political identification across all groups with the Bosnian republic was lacking, mainly not only because of the distinctive federal setting and internal ethnic complexity, but also because of the existence of two national homeland republics adjacent to Bosnia and Herzegovina, which in turn considered Bosnian Serbs and Croats as elements of their ethnic stock. (Sarajlić 2012, p. 369) Pre-war censuses, which were held across Yugoslavia every ten years, represented a most appropriate avenue for people to register their ethnic belonging. In former Yugoslavia, ethnicity was a relatively fixed concept, 'taking the categories Serb, Croat, and so on, as literal, bounded entities' (Denich 1993, p. 49). In this discourse, children from intermarriages had an ambiguous status, as they belonged to more than one category, and at the same time to none of them completely. During the existence of Yugoslavia, the government register included many different ethnic categories, such as 'Yugoslav', 'Croatian', and 'Slovenian'. The last official nationwide census took place in 1991, and in 2002, a census was held in the Federation, the Bosniak-Croatian half of the country. Informants never referred to the 2002 census. The 1991 census adopted 25 categories to count and categorise the population on 10 See also Hayden (2007, p. 112 (Markowitz 2010, p. 82).

External classification-the political structure
The Dayton Agreement strengthened the role that ethnicity occupies vis-à-vis several aspects of Bosnia's daily life. The most obvious one is politics. The informants demonstrated a rather negative view of both national politics and state institutions. For instance, Ana said: 'Our biggest problem is the government', 11 while Edin mentioned that '[P]olitics is corrupted at every level '. 12 In Bosnia, most political parties tend to feature a clear ethnic identity. The general opinion seems to be that politicians are bad, corrupt people who strengthen ethnic partition, while the inhabitants of Sarajevo in general are regular, decent people who live together with their neighbours of different ethnic backgrounds (Bringa 1993;Bougarel et al. 2007a, p. 24;Spaskovska 2012). Emir, for instance, argued that the nationalist politicians were 'not a good reflection of general people in Bosnia'. 13 The report 'The Silent Majority Speaks', which was based on a nationally-representative survey of 3,580 respondents, did nevertheless reveal that only a small majority of the people in Bosnia is willing to abandon ethnic politics (UNDP 2007). The ethnic question does not seem to dominate the lives of ordinary people, at least until the start of electoral campaigns (Mujkić 2008, p. 146). However, as Donia argues in his biography on Sarajevo, most Sarajevans 'reject nationalist exclusivity in principle, yet they have repeatedly opted to put nationalist political leaders in office' (Donia 2006, pp. 351-52).
In representation of the three narodi, Bosnia & Hercegovina has three presidents, who switch place every eight months. The presidential voting system restricts people to vote only for the representatives of their canton. People living in the Federation can therefore only vote for the Croatian and Bosniak delegate and not for the Serbian one, as the latter is exclusively chosen by the people who live in the Republika Srpska (Sarajlić 2012). Damir reflected about this situation as follows: 'Maybe some Croat would like to vote for a Serb, because it is a better man. Why would I have to vote for a Serb because he is a Serb? Stupid. So I think that politicians are putting all people into boxes'. 14 Another unintended consequence of this presidential system relates to the exclusion of children with interethnic background from running for the presidency, especially if they do not identify with one of the three main ethnic groups. 15 The population's discontent with the strongly ethnically divided government, which is widely seen as inefficient and corrupt, did not, however, translate in a collectively organised form of protest, at least not until 2014.

Collective response: the 2014 riots
In the spring of 2014, a series of demonstrations and riots erupted across the country, and soon came to be known as the 'Bosnian Spring'. In Sarajevo, thousands of people took to the streets and set government buildings on fire, with hundreds of people being injured in clashes with the police (judah 2014). Remarkably, those protests were not expressed along ethnic lines, as they mostly focused on socio-economic concerns. At the time of the protests, our field research and data analysis work had been completed; the study's main conclusions, in this sense, had already been drawn. These very recent developments (which would themselves require a separate study) did not impact as a consequence on the focus of this article. A brief description and some analytical reflections on the 2014 protests can however offer useful additional insights to the argument outlined here.
In the course of the 2000s, the Bosnian economy saw some recovery from the very dire situation that crystallised in the early post-war years. From 2002 to 2008, Bosnia's GDP grew on average by 5% annually (World Bank 2015a), fuelled as it was by aid money and growing exports. Unemployment, while still very high, was declining from its 32% rate in 2006 to a somehow more manageable 24% in 2008 (World Bank 2015c). Between 2006 and 2008, youth employment declined from 63% to 48% (World Bank 2015b). However, the benefits of this recovery were mostly, if not exclusively, reaped by the elites and, to a lesser extent, the middle class. The income of the richest 10% and 20% of the population respectively rose from 22% and 37% in 2001 to 27% and 43% in 2007. In the same timeframe, the income of the lowest 20% fell from 9% to 6.6% (Pula 2014).
This recovery stalled when Europe was hit by the 2008 global financial crisis. Bosnia's GDP contracted by nearly 3% in 2009 alone. Unemployment increased again, from 24% in 2008 to 28% in 2012 (World Bank 2015a), 16 while Bosnia's youth unemployment rate, which in 2012 was the highest in the world, shot up from 48% to 62% across the same timeframe (Pula 2014;World Bank 2015b). A worsening economic situation, growing inequality, and the popular perception of a widely corrupt elite represented the major causes of the growing discontent that offered fertile ground to the outburst of protests in Bosnia and other parts of former Yugoslavia (Musić 2013).
The Bosnian protests started in Tuzla, an industrial town in the north of the Bosnian-Croat Federation, to then spread to some 20 urban centres across the Federation, including Sarajevo, Mostar (with a predominant Croat ethnic population), and Banja Luka (the capital of the Serbian part of the country). The protests brought together people from different backgrounds and different ethnicities in a demand for justice, jobs, a chance for a decent life, and an end to corruption (Žižek 2014).
The economic hardship that triggered the protests was partly the result of the protracted deadlock between the three main ethnic groups, since the government structure, with its rotating presidents and ethnic structured parties, made governing and decision making a difficult and convoluted process. Dzidic argues that Bosnia's population remains 'deeply skeptical of a political class widely believed to be ruling in the interest of the elite, not the people' (Dzidic 2014). Politicians have instilled ethnic thinking by keeping fear of the 'other', thereby diverting attention from the state's poor economic performance. The 2014 protests uncontrovertibly signalled that the population's fundamental concerns remain economic and not ethnic (Dzidic 2014;Žižek 2014). In Tuzla, for instance, protests were sparked by the closure of five newly privatised factories, leading to unpaid redundancies for thousands of workers (Pula 2014). Not only were these protests framed in economic instead of ethnic terms but nationalist tendencies were being criticised and solidarity was expressed across ethnic boundaries. On the walls of government buildings in Tuzla, protestors wrote, 'Everybody to the streets. Death to nationalism!'. 17 Others wrote: 'We are hungry in three languages' (Pula 2014), referring to the three constitutive languages of the country. At the demonstration in Banja Luka, one of the organisers of the rally, Aleksander Zolja, told the crowd: 'We gathered to support the protests in Banja Luka where people are fighting for their rights' (Cerkez 2014), 'we are all citizens of Bosnia and we all have the same difficult lives here'. 18 Several local governments resigned as a result of the demonstrations and riots. 19 Some even started a dialogue with the protest movements, which established, in several towns including Tuzla and Sarajevo, plenums and people's assemblies. Notably in Tuzla, the local authorities started an active dialogue with the people's assembly. It might be still too early to tell whether large-scale protests and ensuing assemblies will stimulate a move away from ethnic politics, 20 and whether this will in turn stimulate people to move away from ethnic self-categorisation in the long term. The outcome of the national elections suggests that relatively little has changed, as nationalist parties have once more gained victories in the elections of October 2014 . However, at the municipal level, more change might be taking place. Overall, it seems that continued pressure from the population is necessary to bring about political change. 21

Responses to external classification
This section shifts our attention away from collective, explicitly political responses and places it onto the less visible but by no means irrelevant 'micro-politics' (Brubaker 2002, p. 170): the way people are dealing with the category that is imposed on them in daily life. We will first look at our informants' individual responses to external ethnic classifications, both those imposed by the state (institutional classification or categorisation) and those reinforced by other people (social classification or categorisation). To this end, we will discuss how institutional classification affects social classification, and also analyse the role of instances of social classification based on appearances, language, and names. 20 To what extent the interethnic youngsters have played an active role in these protests remains a question for future research. The limited information available suggests that interethnic youngsters and intellectuals worked actively together with workers and trade unions in the assemblies that emerged or continued their activities in the wake of the protests. 21 However, it is unclear to what extent the protest movement can develop into a longer-term movement with continued support from the population. By mid-April 2014, the number of participants of the Sarajevan daily protests had already diminished to around 30 people (De Noni 2014). Furthermore, there was a threat that the protest would be captured by ethnic and nationalist forces. Ethnic and economic lines might become blurred, as was the case for the demonstrations in Kosovo, which were triggered both by a Serb-Albanian conflict at a pilgrimage site, and by control-related conflicts over a mine complex (Hajdari 2015). Nevertheless, ex-Yugoslavian states have all witnessed the recent emergence of societal movements focusing on economic issues and inequality more in particular (Krašovec 2013;Music 2013).

Responses to institutional classification
When elaborating on the reduction of ethnic 'labels' in the 2002 Federation census, Markowitz came up with the notion of 'off-census identity', an internally defined citizenship in which the size of ethnic groups is not seen as a direct reflection of the way people identify themselves outside the census (Markowitz 2010, p. 85). In Yugoslav times, declarations of ethnic belonging were facilitated by a 'pan-ethnic, citizenship-based category' called the Yugoslav-category (Markowitz 2010, p. 79). This category disappeared with the disintegration of Yugoslavia, and was not replaced by something similar. As Ana, a Bosniak-born young woman in her early 20s preferring to identify as a Bosnian, stated: Before the war I did not know what it meant being a Croat, Bosniak or Serb. Because now you have to declare yourself as a Bosniak, as a Croat or as a Serb, which automatically means you are not a Bosnian and Hercegovinian. It is like you have a stamp on your head. 22 Markowitz describes how census respondents suddenly 'rejected mixed or amorphous answers to questions about nationality', and asked 'what are you really?' to place people in one of the four categories (Markowitz 2010, p. 85). The informants need in this sense an overarching, and ultimately more neutral, category in the structure of the current census. They all would like to identify themselves as Bosnians but are forced into one of the categories from the B-C-S structure (Bosniak-Croat-Serb). The option of the 'Other' category is said to have some disadvantages, which will be described in greater detail below. Izmar, a father of two children in an interethnic marriage, stated: 'For 30 years I was a Yugoslav, then suddenly I needed to be somebody else, and then God knows, maybe there is another war coming, what am I going to be after that? This is a confusing thing for all of us'. 23 Beyond the census, Bosnians have to declare their ethnic belonging when applying for, as the informants put it, 'almost everything'-for instance, a university scholarship. They mentioned that they can fill the forms in with 'Bosnian' but then they will be put in the 'others' box. This makes them feel like they do not exist in their own country. Sara stated: The only problem for me here is that I cannot declare myself as anything. Because I have never lived in Croatia, I have never lived in Serbia. I do not feel like any of those … you know. And I am not Muslim, so I cannot declare myself as Bosniak. … I felt really strongly about it, and I wanted to be part of a group, you know. But after a while I simply gave up and said I do not care, you are just being ridiculous, and I do not want to be part of it. 24 Four of the interviewed informants mentioned that they would choose one of the categories if they felt it to be necessary. They argued that it felt appropriate because they accepted that religion is linked to the ethnic category, such as Samim, or because it suits their father's background, such as Edin, or, as Damir did, because their name is identified with that specific group. Damir said: 'I want to be something. It is funny, I am not Serb, not Croat, not Bosniak but I am also not the Other'. 25  is answered by others who decide on the origin of their names. Eight out of 12 informants mentioned in the interviews that they would either way declare themselves as 'Bosnian' and did not intend to adapt to the imposed ethnic structure. Based on both the interviews and past research, 26 it would seem that, especially in Sarajevo, people would like to profile themselves as a Bosnian as an overarching category.
The application procedure for government jobs is one of the situations in which the interviewees needed to declare their ethnic identity. Government institutions hire people according to a government regulated ethnic quota. Samim, a Sarajevo born and raised man in his late 20s, explained: A friend of mine told me a few months ago 'OK you can apply. There is an open position in the ministry of Foreign Affairs, and you can apply because you are a Croat. They are just looking to fulfil this quota'. And I refused that, because I have other qualifications besides that one. 27 Most of the informants were unwilling to identify themselves as belonging to a category they did not consider themselves a part of. In Markowitz's words (2010, p. 10): ' [O]nly Bosnia[k]s, Croats, and Serbs are eligible to hold major government offices, leaving little if any room for mixed-ethnics, minorities, or those who consider themselves Bosnians with no ethnic affiliation'. The existing possibility for the informants was to identify as part of one of the ethnic groups to which their parents belong(ed). Vedran, whose mother is a Croat by heritage, and his father a Serb, says: Sometimes it puts me in a bad position, sometimes in a good position. But for example for this getting a job, in the institutions of the state, OK if you need a Serb I will be a Serb, if you need a Croat I will be a Croat. Whatever you need, just give me a job. 28 This quote clearly shows that being a child from an ethnically 'mixed' marriage can also be an advantage while job-seeking, as it ultimately offers a flexibility that people with a more fixed ethnic identity simply do not have access to. Markowitz (2010) argues that people did not only use the reduced amount of ethnic labels pragmatically when necessary, as many also started to identify, more than ever before, as belonging to one of the particular ethnic categories. The external (institutional) categorisation, in other words, appeared to have also influenced self-categorisation. Markowitz describes how Bosniaks by birth who, in Yugoslav times, used to see themselves as Yugoslavs are now seeing themselves as Bosniaks. Hayden argues that 'while it may be that categories do not reflect the presumed multiplicity and fluidity of identities, the peoples of Bosnia seem to … declare themselves, vote, and fight as if the categories were real' (Hayden 2007, p. 107). Children from interethnic marriages expressed unwillingness towards declaring their ethnic identity. Gagnon mentions, 'Willingly or not, Sarajevans often find that they are first and foremost ethnically marked and marking actors, who are classified, and classify themselves, their neighbours, co-workers, and friends according to essentialistic national identities and loyalties' (Gagnon 2004, p. xviii).

Responses to social classification
Marking (as well as being marked) is indeed an intriguing phenomenon, since one cannot quickly determine-based only on appearance or speech-who belongs to which ethnic group when walking on the streets of Sarajevo. The difficulty of being categorised by others and, in turn, categorising others does not relate to the few people who wear religious symbols or objects-such as a headscarf- (Holbrooke 1998, p. 23;Mesaric 2013) 29 as, in Bosnia & Hercegovina, religion and ethnic identity are often linked (Hayden 2002, p. 157). The Serbian and Croatian languages (before the war simply seen as one Serbo-Croatian language) only differ in the case of a rather limited number of words. Differences in the spoken languages can be identified per district, such as the Sarajevan dialect. Language per se ought not to be regarded as a predominant feature in distinguishing between the various ethnic groups. 30 The absence of clear markers underpins the importance of categorisation at state level. Boris, a PhD student in psychology from Sarajevo, stated: 'Because we do not have these strictly divided groups by colour of skin or whatever, you have to invent the differences. You cannot identify whether a person belongs to one or the other group by simply looking at him or her'. 31 Although people do not recognise each other's ethnic identity on the basis of physical features (Armakolas 2007, p. 92), the informants mentioned that, since the war, all Bosnians have been able to identify one another's ethnic identity based on their name or the names of their parents. According to Donia and Fine, children of interethnic marriages were given foreign or generic names, because their parents 'thought of themselves as Yugoslavs or Bosnians without ethnic allegiances' (Donia & Fine 1994, p. 186). 32 Vedran stated: People are now so obsessed with this ethnic story, recognising who is who, that … when you are reading something and you read the name, you already have the prejudice 'he is a Muslim, he is Croat', you know. You already have prejudices about that man by his name, which is out of [its] mind. 33 The informants demonstrate striking abilities in assessing ethnic identity based on someone's name. This was so natural to them that they thought it was humorous that we were unable to associate ethnic identities with names. They were in this sense happy with their neutral names, as this made others unable to classify them with any degree of immediacy. Examples of neutral names are the fictional names assigned to the informants mentioned in this article. Vedran illustrated the situation through a short anecdote that clearly shows how people with a critical stance regarding ethno-politics do also (unconsciously)  However, the number of women wearing headscarves is increasing (Mesaric 2013). It is one visible sign of the wider re-emergence of religion, such as, for instance, the re-emergence of Muslim dervish cults (Henig 2014). 30 This is not to say that language is no marker at all. During the war, for instance, some scholars that were contracted by the opposing regimes worked on pointing out the difference between the languages, and even (re)invented words to differentiate the languages. In addition, Serbian tourists in Croatia state that they prefer to talk in English, in order not to give away their Serbian identity (de Koning 2011). 31 Interview with Boris, PhD student, Sarajevo, 24 February 2011. 32 See also Bringa (1995, p. 153 36 Other informants stated that people in interethnic marriages strive for the exact opposite, to make their child ethnically unrecognisable by giving him/her a neutral or international name.

Broader strategies: exit, reframing, (partial) separation
Whereas the previous section looked at more immediate/direct responses to ethnic classification, this section discusses in turn the broader strategies employed by our informants in the context of ethnic classification. Within the discussion of these strategies, discursive as well as socio-economic aspects have come to the fore with greater frequency. We have distinguished two opposing strategies: active resistance and reframing on the one hand, and exit (leaving the country) on the other. A third strategy, positioning life within the international, interethnic spaces of Sarajevo, could to some extent represent a middle option between these two strategies.

Exit: leaving the country
In one of the interviews, Sara said: 'When I think about my future, it is somewhere else '. 37 According to the 'Human Development Report BiH Youth 2000' (UNDP-IBHI 2000), 62% of the Bosnian youth would leave the country if they had the opportunity. 38 Many of the informants pointed out that they would like to stay and help the country improve. With no economic improvement, however, most of them would leave. Sara stated: 'I am really, really frustrated that it has been so much time, and things are not really getting better'. 39 Everyone agrees that Sarajevo is a great place to live, and they all would love to stay. However, Bosnia's political situation and its gloomy economic conditions make people think about leaving. As Nina argued: I would really love to live here, and I feel a lot of potential. I love the culture, and I really love being part of this. But if you have a bunch of people who do not want it, and want to destroy it, it is hard to fight it. It is hard to prove [to] them [that] this country is great as it is, we just need to appreciate it more. 40 The frustration of the informants can be heard clearly in these quotations, as they are struggling to contribute to the change of Bosnia's current system without ultimately giving up their values.

Reframing and resistance: examples and limitations
Most informants have made a bottom-up contribution to Bosnia's process of change. As Sara observed: 'Politics is not how things are actually done. All change starts from the lowest level'. 41 In some (rare) instances, informants have actively resisted ethnic categorisation in practice, as demonstrated by the earlier case of a young Sarajevan who refused a job because ethnic identity was one of the selection criteria. It is probable that the ones willing to make such a choice are confident enough to do so as they can access Sarajevo's international space, where NGOs and international firms do not select based on ethnicity only. However, resistance to ethnic classifications is mostly discursive and does often take the form of reframing. The result of the informants' desire for change is their against-the-odds efforts for the recognition of a fourth ethnic category: the Bosnians.
Inter alia, a common Bosnian identity exists in relation to food, music, art, jokes, coffee, football, and dissatisfaction vis-à-vis the international community and the OHR (Hayden 2007, p. 127). An all-inclusive, pan-ethnic Bosnian identity is frequently expressed in daily life, for instance through anecdotes and songs (Woodward 1995, p. 36;Denich 2000, p. 40;Hayden 2002, p. 122). Also, 'most inhabitants of Bosnia-Herzegovina accept a national identity, [although] the majority still prefers to identify themselves primarily as Croat, Serb or Bosniak' (UNDP 2007, p. 3). We will not go into further detail with regard to this latter aspect. In this section, we will instead elaborate upon the recognition of 'Bosnian' as a separate identity label within the country's policies. The discourse establishing and recognising an (all-)Bosnian category resists the division into the three ethnic categories. However, some variants of this discourse are clearly more inclusive than others.
Emir's statement below reflects a very inclusive understanding of the 'Bosnian' category. In an interview, he emphasised that: '… being Bosnian, for people who declare themselves like that, is not just nationality, it is almost like [an] idea; [an] idea of unifying things here'. 42 Edin's following statements are also a plea for the Bosnian category; however, this informant had more difficulties in setting aside the exclusionist thinking that is part of the prevalent ethnic categorisation in society: Like three herds, that is how I see the situation in Bosnia: three herds of sheep. Everyone wants to … if I am Muslim I want to belong to this herd of sheep, if I am Catholic this herd of sheep, or if I am Orthodox this herd of sheep. I am not like a sheep, so I do not like to go in these herds of sheep. … It is like religious herds of sheep. So I am not religious, so I am for Bosnia. I do not care; I will be the fourth herd of sheep. We do not care what nationality you are. You are human, you are good, and will be this fourth herd, and we will be together. 43 In the first part of the quotation, Edin clearly distances himself from the group thinking, here aptly described with a herd metaphor. He does not want to be 'a sheep'. He also refers to being human, which is of course the most inclusive label possible. However, he cannot escape from the group discourse when he speaks about a new Bosnian identity, as reflected by his return to the herd metaphor. Instead of creating an overarching identity, he proposes another, fourth herd. When he says 'we will be together', he only refers to the people within the fourth herd. 41 Interview with Sara, software developer, Sarajevo, 16 April 2011. 42 Interview with Emir, journalist, Sarajevo, 16 March 2011 Interview with Edin, student, Sarajevo, 1 April 2011.
Thus, even young Sarajevans who strive for a more open and flexible framework regarding ethnicity find it difficult to identify Bosnian as a fully inclusive identity. The 'Bosnian' category is not 'empty' as in the case of the 'Moldovan' category in Transnistria (Dembinska & Iglesias 2013). As mentioned earlier, there are quite a few cultural patterns that people would classify as Bosnian. However, Bosnian as a category certainly feels lonely and insecure, even for most of the young interethnic Sarajevans.

Centring life in separate spaces: multi-ethnic and international spaces
A common reaction towards ethnic categorisation is presented by the (deliberate or unconscious) avoidance of places that are ethnically dominated, to find refuge in specific, and often removed, social spheres, namely local non-ethnic groups and arenas or, more importantly, international settings, including EU institutions, international NGOs, and international firms. The orientation towards international structures can be discursive (given EU membership as a potential solution to ethnic divisions) or enacted through daily action (through, for instance, concentrating one's social or working life on international organisations). Many informants argued, in line with Spaskovska, that EU and NATO memberships might 'provide an impetus for partial healing of the inner ethno-cultural fractures' (Spaskovska 2012, p. 392). 44 As previously mentioned, Sarajevo occupies a particular place in the country. In the postwar years, the international community has been very much present in the city's daily life. Sarajevo hosts the Office of the High Representative (which embodies Bosnia's political power), the European Union Police Mission, several UN branches, and numerous NGOs, international businesses, and organisations. As the capital and the most internationally oriented place in Bosnia & Hercegovina, the situation of Sarajevo's interethnic children is different from elsewhere in the country. Barbara Matejčić, a Croatian journalist, mentioned that people generally say that in Sarajevo all places are multi-ethnic and tolerant. 45 Although this might be seen as an overstatement, it is certainly true that there are more multi-ethnic open spaces in Sarajevo than elsewhere in the country.
Most informants interviewed for this article work in NGOs (often having strong international linkages) or in the commercial sector, particularly within larger international companies. Markowitz describes two of her informants as people who 'defy otherness by making it irrelevant to their goals and to their lives' by studying abroad and working for international firms (Markowitz 2010, p. 100). 46 This questions whether the ethnic tolerance nurtured by young Sarajevans in the international space of NGOs and multinational firms does ultimately constitute a place for change within society. Or, in turn, are the informants constituting what Sampson (2002) called a 'project elite'-both part of local society and not part of it-and is the shift to the international space more akin to fleeing rather than transforming Bosnian society?
Despite the importance ethnic identity continues to hold in Bosnian society, some people have continued to downplay its centrality, by focusing on 'not caring' about ethnic issues (Zahra 2010). We would, however, argue that the distancing from broader local society comes not as a result of indifference but as a rather unintended effect of the search for like-minded 44 The discourse of some other informants, as well as that of Armakolas (2007, p. 85), does not depict foreign (EU) involvement as the only solution, and draws on domestic, historical examples, mentioning a 'new Tito' as a solution for the current situation. 45 Personal communication with Barbara Matejčić, journalist, 21 March 2011. 46 See also, Kolind (2007, p. 138). people, or through employment in the international space. A few non-ethnic communities emphasised this ethnically 'neutral' aspect: a case in point is offered by Pontamina, a multireligious choir that consists of people from many different religious, ethnic, and cultural backgrounds and tries to bridge religious and cultural differences in Bosnia & Hercegovina.
Building bridges is Pontamina's aim: this generally corresponds to the informants' desire for an overarching Bosnian identity that transcends ethnic boundaries. Dominant discourses within the international space have little connection with the views of the broader population regarding ethnic relations and Bosnian society. Hayden (2007) points out that the Western understandings of Bosnian history focused too emphatically on multiculturalism and plurality, whereas the Bosnians themselves held no such views. 47 In this regard, Vetta describes the Serbian political campaign for the 2008 elections, observing that Serbs were 'radically polarised between two axes of identity: a modern and progressive European identity … and a traditionalist, nationalist identity' (Vetta 2011, p. 38). To a very similar extent, Armakolas highlights the completion of 'two types of officially sanctioned nostalgias', namely the discourse of the international community and the (formerly powerful) exclusivist Serb discourse (Armakolas 2007, p. 86). Bougarel et al. end their introductory chapter on Bosnia with a statement that relates well to these two nostalgias: '… two separate spaces exist: one for those "stuck" in/with the nation-state form, and one for those with the "ability" to rise above and beyond it' (Bougarel et al. 2007b, p. 272). However, we would tend to note that those in the latter are also somewhat stuck in this very cosmopolitan and open space, which is separated from society at large along lines of class, language, and education.
A clear indication of the limited influence that this multi-ethnic, and hence cosmopolitan, space exerted on the wider local society is represented by the stance of local politicians towards it. Mujkić argues that ethno-politicians will 'tolerate and even welcome' the NGO sector, 'as the only stage or "outlet" where a few angry citizens, most of them liberallyinclined intellectuals, can express their "frustration" without hindrance' (Mujkić 2008, p. 190). Along very similar lines, Sarajlić observed: 'The feeble existence of liberal conceptions of citizenship, sustained mainly by small academic circles in the universities of Sarajevo and Tuzla, is still unable to reframe Bosnian politics and civic life in categories other than ethnic' (Sarajlić 2012, p. 376).
Relating to this description of the different responses to ethnic classification, Table 2 provides an overview of how Bosnians have addressed this issue, by including a short description of both the practical implementation of the limitation associated with these responses.

Understanding strategies and their limitations: beyond the supposed irrationality of ethnic nationalism
A large portion of Bosnia's inhabitants continues to vote for nationalist parties featuring a clear ethnic identity. Support for nationalist parties (and, more broadly, nationalist sentiments) is reproduced in the Western press as a reflection of irrational, emotional behaviour induced by hatred, fear, or frustration. 48 The cosmopolitan segment of the population in Sarajevo seems to generally conform to this conceptualisation. We argue, in line with Kalb (2011), that such an interpretation ultimately ignores the individual economic rationality of this choice, which is mirrored in both working class environments and across Bosnia's middle class.
Due to the discrediting of socialism, no leftwing parties are now championing the interests of the working class across Central and Eastern Europe: most parties, including those inspired by the labour movement, have adopted a pro-European, and implicitly neoliberal stance, marginalising the relevance that the interests of the working class held in their agenda (Ost 2005;Kalb 2011). 49 Throughout the institutionalisation of neoliberal policies in post-war Bosnia, while international agencies and media focused on ethnic issues, the local economy all but collapsed, with industrial production, employment rates, and the social welfare system experiencing significant declines (Donais 2003;Pugh 2006). Interestingly, Bosnia's painfully Not contributing to substantial change, no effect beyond the small multi-ethnic and/or international spaces neoliberal policies were articulated through a discourse on 'the return to normalcy' (IMF 2010, p. 11). 50 Despite a generalised wish for change, the unwillingness to enter politics, join organisations in need of change agents, or make specific choices (such as not accepting jobs that are provided based on ethnicity) led to a deadlock, the essence of which has been captured by Mujkić (2008) through the 'ethnic prisoner dilemma'. Bosnia & Hercegovina's economy, in this sense, might be said to have suffered from a polarised political system that gives too much weight to the matter of ethnicity in its decision-making processes. As Damir illustrated: I am really like in a trap, either to leave the country or not. I want to give myself here, try to put my efforts to change something. At least in a micro-world. But I do not know if it is worth it. … when you talk individually, everyone needs a change, but since probably we are so separated. In a different kind of way, we cannot put ourselves together and change something. But individually there is always [a] need to change. But you cannot gather 1,000 people in one place from all over Bosnia to change something, because people are really divided, have different information. It is really chaos in that sense. 51 Damir's statements reflect Mujkić's ethnic prisoner's dilemma concept, which was developed to describe cases when two (or more) parties operate in settings without effectively communicating, due to distrust or other factors. In these cases, 'by rationally pursuing [their] own interest, [they] both end up worse off than if [they] had acted differently' (Mujkić 2008, pp. 185-91). Mujkić adjusted this concept to the Bosnia & Hercegovina context: individuals expect others to behave primarily as members of an ethnic group, and therefore they themselves act in the interests of their own ethnic group while knowing that this would lead to suboptimal results for the wider society (Mujkić 2008, p. 193). Multiple rationales are at play here. First: if I vote for the multi-ethnic party but the people from the other ethnic groups still vote for their political party, they will be in power, and my interests will not be represented. Or, alternatively: if I do not support the nationalists, I will not have access to the resources they control. In this context, it is important to note that, during post-war privatisations, telecommunications and energy were 'divided along ethno-party lines to provide major sources of revenue for the nationalist parties and their parallel structures' (Pugh 2006, p. 451).
Without inter-group communication or the certainty that the establishment of a multicultural society does indeed represent a shared objective, the stimulation of change might be a difficult, if not altogether impossible, endeavour. Social trust is 'virtually non-existent' in Bosnia, according to a UNDP survey (2007, p. 2). This explains why our informants' tactics are often ambivalent and, to date, bore little fruit, despite their claimed adherence to an ideal multiculturalism: they are stuck in an 'ethnic prisoner's dilemma'. 50 Due to its rapid socio-economic decline, Bosnia's working class is often compared to Africa's (see Vetta (2011) on Serbia). The emergence of cronyism and oligarchic domination led some researchers to outline a Russian trajectory for Bosnia's development (Donais 2003). In response to an ongoing social crisis, international agencies acknowledged the importance of national welfare policies that Bosnia implemented since the early 2000s. Funds allocation by those agencies is still limited. The World Bank welfare budget represents only about a third of the sum committed to managing Bosnia's privatisation process (Pugh 2006, p. 455). 51 Interview with Damir, student, Sarajevo, 25 April 2011. This interview, as well as the others, took place before the 2014 riots. While thousands of people from different ethnic groups did come together for the protests, most observers concluded that this wave of demonstrations did not result in systemic change (see, for example, Dzidic 2015).

Conclusions
The introduction questioned the accuracy of Sarajevo as a 'multi-multi city', arguing that tolerance, in Bosnia & Hercegovina's capital, could be regarded as an old image and a public relations tool targeting the outside world. This article did not address this specific question through an exclusionary approach. Here, we instead attempted to describe how young adults relate to this dichotomy and, most crucially, the way it influences their lives. Neither end of the dichotomy reflects in this sense Bosnia's reality. Sarajevo, we contended, ought not to be described as the ideal 'multi-multi city' (Markowitz 2010), where ethnic and religious groups live peacefully side-by-side. Post-war Sarajevo is a much more ethnically homogenous place, due to large shifts in the population's composition. Informants mentioned that many progressive Sarajevans left the country during the war, while many 'villagers', who fled their houses during the war, eventually moved to Sarajevo. Many Bosnian Serbs relocated to the Republika Srpska (Sell 1999), whether to the eastern part of Sarajevo (called Srpsko Sarajevo-Serbian Sarajevo) or further away. Interestingly, most of our informants rarely visited this part of the town: they did not even regard Srpsko Sarajevo as being part of the city. These considerations suggested that Markowitz (2010), in terming Sarajevo a 'multimulti city', might have failed to acknowledge adequately that her informants (like ours) are mainly highly educated, middle class persons, who are not representative of ethnic stances and interactions within the wider Sarajevo, and therefore form a quite separate 'project elite'.
It would be, however, inappropriate to state that Sarajevo's different ethnic and religious groups live separately. Sarajevo is still more mixed than other areas of Bosnia & Hercegovina, as it provides a number of public spaces where ethnic intermingling continues to occur on a regular basis. This is in stark contrast with, for instance, Mostar, where one multi-ethnic café seems to be a rare place where people of various ethnic backgrounds meet openly and mingle. We therefore argued that multi-ethnic or 'multi-multi' spaces do indeed exist: this conclusion, together with the mix of religious symbols, shapes the conceptualisation of Sarajevo as a multicultural context. Open spaces might be visibly multi-ethnic; we however argued that the separation between the different ethnic groups has occurred on more abstract levels, such as name-giving and ideals. Moreover, access to some multi-ethnic spaces continues also to be limited by class divides.
The other question posed in the introduction in relation to young Sarajevans that come from intermarriages asked: do they, despite their stated ideal of ethno-tolerance, keep the system in place, or are they drivers of change?
The various strategies adopted by Sarajevan youngsters to cope with ethno-politics reveal their struggle with inflexible and static government procedures and categories. They experience a clash between internal and external definitions of ethnic identity. They are committed to a tolerant Sarajevo, and their stated ideal is a multi-ethnic existence. That being said, what do these critical, internationally oriented young adults concretely do to improve the current situation? Choosing 'other' over the three recognised categories when dealing with institutional categorisation could be a form of protest. However, most informants continue to choose one of the three recognised categories. Their strategies mainly consist of distancing themselves from nationalistic tendencies, instead of actively going against them. Informants hardly discuss problems or ideas regarding ethnicity with their friends and acquaintances. Moreover, when we take into account broader strategies (exit, resistance and reframing, as well as life (re-)centring in separate multi-ethnic spheres), it is clear that these strategies are unsatisfactory as they have so far failed to instigate structural change. A rare and recent exception is represented by the case of the young Sarajevan couple that in january 2015-for the first time in Bosnian history-managed to register their child as a 'Bosnian' after a legal fight (jukic 2015). While it is too early to predict the effect of this legal precedent, it surely will not remain an isolated case: two weeks after the ruling, five more requests had been filed to register children as Bosnians (jukic 2015).
However, the hurdles obstructing a move towards less intense ethnic categorisation are not just of a legal or institutional character. One such obstacle-apparent even in the case of our informants, who reportedly want to do away with ethnic categorisation-is represented by the pervasive ethnic factor, which so influences the population that even persons like our informants subconsciously adhere to it. This was evidenced, for instance, by the person who caught himself counting ethnic names after a movie.
The 'ethnic prisoner's dilemma' is another factor limiting the youngsters' strategies (Mujkić 2008). Unless these youth are fully secure in their international career and live in Bosnia's interethnic space, implementing concrete actions against ethnic categorisation might be a detrimental construct. In an economy characterised by low social (and interethnic) trust, and dominated by ethno-political groups, such choices might loosen their networks and reduce their economic possibilities. Thus, openly distancing themselves from ethnic categorisation is not in the immediate self-interest of Sarajevo's interethnic youth.
Finally, in line with Sampson's concept of the 'project elite', we argued that these young adults centred their life on a separate sphere made up of like-minded, progressive people. This makes influencing the other spheres a very difficult process. Identifying a way out of this stalemate is not straightforward, and goes beyond the scope of this article. However, to understand such a possible trajectory, it is important to pay attention not only to ethnic distinctions, but also to class-based divisions-such as those between the more cosmopolitan, somewhat dis-embedded 'project elite' and the more locally embedded majority, or between the 'old' Sarajevans and the often less educated newcomers from the villages. 52 Such distinctions, both discursive and socio-economic, seem to act as overlooked hindrances to transforming ideals of multi-ethnic coexistence into daily realities.