The Failure of “Yugoslavia’s Last Chance”: Ante Marković and his Reformists in the 1990 Elections

Abstract

The last Prime Minister of Yugoslavia Ante Marković was considered by many within the country and in the international community to be Yugoslavia’s last chance for a peaceful transition toward democracy and capitalism. In spite of his popularity, the Reformist party he created failed decisively in the first democratic elections of 1990. We expose the reasons for this failure by analyzing electoral, economic, and sociodemographic data on the level of more than two hundred Yugoslav municipalities where the Reformists put forward their candidates. Our analysis shows that the party’s failure had little to do with the voters’ exposure to the effects of the free market reforms undertaken by Marković’s federal government during this period. Instead, the Reformists’ results were largely determined by the communities’ ethnic makeup and interethnic balance. The Reformists suffered at the hands of a strong negative campaign by the Serbian regime of Slobodan Milošević, and they were squeezed out by the ethnically based parties that benefited from voters behaving strategically in the electoral marketplace dominated by questions of nationalism. The analysis presented here offers important lessons for our understanding of Yugoslavia’s breakup, post-communist transitions in general, and electoral politics in societies on the verge of ethnic conflict.

On July 29, 1990, at the commemoration of the World War II Kozara Battle in northwestern Bosnia-Herzegovina, Yugoslavia’s Prime Minister Ante Marković launched his political party – the Alliance of Reformist Forces of Yugoslavia (Savez reformskih snaga Jugoslavije, SRSJ). He also announced the new party would contest the first democratic elections in the federation’s four republics: Bosnia-Herzegovina, Macedonia, Montenegro, and Serbia (elections in Croatia and Slovenia took place earlier that spring). At that time, Yugoslavia was already on the verge of descending into violent collapse after a decade-long economic and political crisis. Marković implored the assembled mass of nearly a hundred thousand people on the Kozara Mountain not to succumb to the national divisions that were defining the politics of the time: “You do not want conflicts anymore. You do not want blood, particularly not fratricidal blood. You want to send a message that you want to live together” (Grubić and Preradović 1990a). Although he also praised the economic reforms of his government and expressed support for pluralism, democracy, and the market economy, his comments on the need for Yugoslav unity, multi-ethnic tolerance, and respect for the country’s Titoist heritage drew the biggest applause (Filipović 2021). According to local media coverage, the crowd saw Ante Marković as the potential savior of Yugoslavia and the country’s last chance to come back from the brink of dissolution and violence (Čolak 1990a; Grubić and Preradović 1990b). Many international observers, including Western policymakers in highest places, agreed with the crowd on the Kozara Mountain that July. They also believed Marković and his platform of economic reforms offered the only peaceful path out of the crisis for the troubled federation (Zimmermann 1996).

After three months of a rollercoaster campaign, however, voters rejected Marković’s vision. His party’s centrist and liberal platform may have appealed to people in the abstract, but when it came time to cast ballots, voters overwhelmingly supported parties and politicians who ultimately led the country toward violence and breakup. Although Marković and his party have received attention in the literature on Yugoslavia’s breakup (e.g. Andjelic 2003; Gagnon 2004; Glaurdić 2011), and their platform and organizational efforts have recently been thoroughly analyzed (Filipović 2021; Sasso 2020), we have remarkably little firm evidence of the reasons behind the Reformists’ electoral failure to capitalize on the popularity of the federal prime minister and his reforms and to become political actors of any real influence on the evolution of the Yugoslav crisis. Studies focusing on the first democratic elections in Yugoslavia, for example, deal with the Reformists only superficially and instead focus largely either on the communists or their nationalist challengers (e.g. Arnautović 1996; Goati 1998; Herceg and Tomić 1998; Aziri 2013; Stojanović 2014).

Why did the electorate in Yugoslavia’s four republics reject the SRSJ? We believe it is possible to come closer to answering that question by explaining where the electorate rejected the SRSJ. To that end, we construct a dataset consisting of electoral results and a host of economic and sociodemographic data on the level of more than two hundred Yugoslav municipalities where the SRSJ put forward its candidates. Our analysis convincingly shows that the Reformists’ (lack of) success had little to do with the economic conditions or the exposure of the workforce to the destabilizing effects of the free-market reforms undertaken by the Marković government. Instead, the Reformists’ results were decisively determined by the communities’ ethnic makeup and interethnic balance. The Reformists did well in ethnically diverse communities, but not in those marked by polarization between ethnic groups. They also underperformed in communities with greater proportions of Serb voters. Simply put, the SRSJ suffered at the hands of a strong negative campaign by the Serbian regime of Slobodan Milošević and his allies, and it was squeezed out by the ethnically based parties that benefited from voters behaving strategically in the electoral marketplace dominated by questions of ethnicity and nationalism.

The analysis presented in this article fills an important gap in our understanding of the advent of democracy in former Yugoslavia, as well as of the voters’ role in the elections that precipitated the breakup of the country. It also, however, offers an important contribution to the study of democratic transitions in postcommunist Europe in general. In spite of their monumental importance in the history of European democracy, data-driven studies of the foundational democratic elections in Eastern Europe – and particularly in former Yugoslavia – have been rare (Heyns and Bialecki 1991; Kopstein and Richter 1992; Pacek 1994; Kapidžić 2014). Equally important, our study also offers insights into the electoral politics in societies on the verge of violent ethnic conflict. It helps explain where and why political forces trying to overcome ethnic divisions in such societies struggle to make headway with the electorate.

We organize our argument as follows. First, we describe and explain the context leading to the formation of the SRSJ. We then situate Marković, his party, and their electoral performance into the broader literature on the democratic transitions from communism and outline a number of empirical expectations stemming from that literature. After the description of the collected data and method used, we then present the results of our analysis and, finally, conclude with a discussion of its importance for our understanding of Yugoslavia’s breakup and ethnic politics in general.

Ante Marković and the Alliance of Reformists in the 1990 Elections

The last prime minister of the Socialist Federative Republic of Yugoslavia, Ante Marković, took office in March 1989, at the high point of Yugoslavia’s economic crisis. Yugoslavia struggled with foreign debt, falling incomes, rising unemployment, and spiraling inflation throughout that whole decade. Marković was selected to the post of federal prime minister because of his technocratic credentials as a former manager of a large industrial enterprise and as a reformist high functionary in Croatia. The most important difference between Marković and his predecessors was that he was clear about the Yugoslav brand of socialism being systemically unviable. Marković openly admired the free market and believed it had a civilizing role (Marković 1990, 14–22). True to his reputation, the federal government under his leadership initiated a broad program of reforms with the liberalization of prices and imports, more flexible labor legislation, easing of restrictions on private enterprises, plans for the privatization of state firms, and ultimately full convertibility of the Yugoslav dinar through its pegging to the German mark in order to rein in rampant inflation.

Although his reforms clearly proposed a path toward some form of capitalism and were seen by many left-wing commentators as shock therapy, Marković’s labeling of his program as “new socialism” could not be considered as false advertising. His reform package was in most important areas gradualist and his embrace of “privatization from below” that was to turn workers into shareholders was considered as fanciful by his Western neoliberal advisers like Jeffrey Sachs (Sasso 2020, 29). In spite of these critiques from both the left and the right, Marković and his reforms enjoyed public (if not financially tangible) support from important international players such as the IMF, the US administration of George H.W. Bush, and a number of Western leaders (Zimmerman 1996). This support continued even well after it was clear Yugoslavia was violently falling apart in the second half of 1991. Crucially, Marković and his reforms were also, at least initially, hugely popular with the Yugoslav electorate in all republics and provinces – though with some notable regional variation. According to surveys conducted by local media, support for Marković in the spring of 1990 was as high as 79% on the Yugoslav level. It ranged from 59% in Slovenia to 93% in Bosnia-Herzegovina (Woodward 1995, 129).

The essential problem for Marković was the fact that Yugoslavia’s troubles were not only economic, but also political. Throughout the 1980s, the Yugoslav federation was in a serious constitutional crisis that struck at the core of its very existence. Yugoslavia always meant different things to its various constitutive nations and their republics, but by 1989 when Marković came to office, it was clear that any coherent political decision-making was no longer possible as the ruling League of Communists (Savez komunista Jugoslavije, SKJ) was fatally divided between the supporters and the opponents of the Serbian regime of Slobodan Milošević. Milošević saw Yugoslavia’s problems as rooted in its (con)federal structure and he aimed to recentralize the country under his and Serbia’s leadership. Milošević’s opponents, most prominently in Slovenia, balked at his platform that was marrying centralism with Serbian nationalism and instead proposed greater decentralization and liberalization as a solution to the country’s ills. Yugoslav politics in the late 1980s collapsed into a spiral of competing nationalisms that, to varying degrees, were employed by the nominally communist/socialist leaderships of all republics (Haug 2012; Ramet 2006). This clash led to the disintegration of the SKJ at its 14^th Congress held in January 1990, just weeks after Marković launched the main planks of his reform program. Marković famously declared at the congress that Yugoslavia would continue to exist even without the ruling communist party (Sasso 2020, 37). While many at the time saw his statement as wishful thinking, it actually fit perfectly with his worldview and political ambitions. Although a member of the League of Communists for most of his adult life (even actively fighting in the partisan resistance during World War II), by the late 1980s, Marković believed that the state should limit its influence in the economy and he accepted the end of the party’s monopoly on power (Marković 1990, 45–46). The end of the SKJ, therefore, did not necessarily derail his plans because he believed he could offer a pan-Yugoslav political platform that could step into that vacuum of federal political leadership.

Unfortunately for Marković, the elections in Slovenia and Croatia that were scheduled for April 1990 simply came too soon. Even if his political party were ready at the time, it is questionable whether Marković would have chosen to run in Yugoslavia’s two western republics. His popularity in Slovenia was limited, with many on both sides of the political spectrum in this republic criticizing his reforms as centralist. In Croatia, on the other hand, he supported the reformed communists/socialists under Ivica Račan anyway, since they returned the favor and supported his program at the federal level. The reformed communists/socialists in Slovenia and Croatia, however, lost the elections that spring. Parties that grew out of the socialist system received 37% of the votes and 39% of the seats in the Slovenian elections held under proportional representation rules, and 36% of the votes and 25% of the seats in the Croatian elections held under majoritarian rules. The fate of the reformed League of Communists in Croatia (Marković’s party base) in some ways presaged the destiny of the federal prime minister, as the party’s inclusive platform did well in Croatia’s ethnically mixed areas but was defeated handily in mono-ethnic communities.¹

An additional complicating factor for Marković that spring was the fact that, under Yugoslavia’s system of rotation, the President of the Presidency of Yugoslavia, Slovenian Janez Drnovšek, was replaced in May 1990 by the Serb Borisav Jović, who was known as a dogmatic communist close to Milošević and hostile toward Marković (Jović 1995, 195–196). With Milošević dominating Serbia, its two autonomous provinces Vojvodina and Kosovo, as well as Montenegro, and the electoral results in the two remaining republics – Bosnia-Herzegovina and Macedonia – being highly uncertain, Marković realized that he was quickly losing his foothold. He decided to try to capitalize on the popularity he had earned through his economic reforms, and to use the fact that he was the only politician who stood at least some chance among all nationalities in the four republics projected to have their first democratic elections that fall: Bosnia-Herzegovina, Macedonia, Montenegro, and Serbia. He believed his allies could become important power brokers in at least some of these republics and eventually get the most seats in the elections to the federal assembly that were projected to take place subsequent to the republican elections.

After several unsuccessful attempts to complete the constitutional reforms on the federal level that he and his government saw as necessary to establish a functioning and democratic federation, Marković finally officially launched his party, the Alliance of Reformist Forces of Yugoslavia at the aforementioned mass rally on the Kozara Mountain. Thereafter, the SRSJ initiative committees were organized throughout Yugoslavia, with the focus on the four republics where elections were still to take place. A number of small left-wing and pro-Yugoslav parties joined the SRSJ. Within weeks, the number of SRSJ supporters was estimated at two million, with about half a million supposedly registered in the republic seen as key for Marković’s platform: Bosnia-Herzegovina (Verovska 1990a).

What is important to note here is that these initiative committees of the SRSJ were usually organized by reform-oriented former members of the SKJ, leading in many instances to tensions between the SRSJ and the still-ruling party. This was particularly the case in Bosnia-Herzegovina, where personal animosities among the leaders of the League of Communists, the Democratic Socialist Alliance (another spin-off from the ruling regime), and the Alliance of Reformist Forces divided the leftist non-nationalist bloc and had dire consequences on the campaign and the electoral results (Pejanović 2004). In Serbia, the SRSJ used an organizational base which consisted of groups of pro-Yugoslav intellectuals such as the Association for the Yugoslav Democratic Initiative (Udruženje za jugoslovensku demokratsku inicijativu, UJDI), with a number of smaller parties either officially joining the SRSJ (like the Social Democratic Alliance of Serbia, Liberal Forum, People’s Peasant Party, and the Party of Yugoslavs) or maintaining close relations with it (like the League of Social Democrats of Vojvodina and the Democratic Forum of Vojvodina) (Borba 1990b, 1990c). A similar pattern could be observed in Montenegro, where the leaders of the youth section of the League of Communists created their own party and turned it into an organizational base for the SRSJ, together with the Liberal Alliance and the Party of Socialists (Vuković 1990a). Local League of Communists organizations from Bar, Cetinje, and Bijelo Polje also joined the SRSJ, as did the Party for National Equality which was supported by Muslims and Albanians in Montenegro, making the SRSJ the key challenger of the League of Communists of Montenegro – a crucial ally of the Milošević regime in Belgrade (Mićunović 1990a; Vojičić 1990a; Vuković 1990b). Finally, the organization of the SRSJ during that summer was rather successful in Macedonia, though it was largely limited to the republic’s ethnically Macedonian electorate. As in Bosnia-Herzegovina, no collaboration with the ruling League of Communists could be established, partly due to the battle of vanities between the leaders of the SRSJ and the League of Communists (Đurić 1990c; Vlahović 1990).

In spite of these local idiosyncrasies in the way the Reformists mobilized and organized during the summer and fall of 1990, all four republican organizations of the party subscribed to the nearly identical progressive, liberal, and reform-oriented program. In one illustrative example from the local press, the SRSJ was described as a party providing an alternative to both the forces of “the politics of equal stomachs” (i.e. dogmatic communists) and the forces of “the politics of empty stomachs” (i.e. nationalists) (Mrđen 1990). The key point emphasized by Marković was Yugoslavia’s fast transition to a free-market economy and the privatization of socially owned enterprises. Marković believed that pluralism in politics could not be as substantive without the accompanying changes in the economy. He saw himself as an economic liberal and kept reiterating the advantages of a free and united Yugoslav market that could be the foundation for a preserved Yugoslav state. He and his party openly rejected even engaging in the federation vs. confederation debate that dominated Yugoslav politics at the time – likely to their detriment (Sasso 2020). Here it is important to note, however, that for Marković it was crucial to secure more authority for the central government (in other words, for himself) in any new political system that would be built in Yugoslavia (Inicijativni odbor SRSJ 1990; Ninčić 1990a; Verovska 1990b).

This, of course, does not mean that the four SRSJ organizations did not address issues that were specific to the political competitions in their republics, though the questions of ethnic identity and nationalism dominated all four electoral campaigns. In Bosnia-Herzegovina, the SRSJ claimed it “presented a minimum of civilizational, cultural and ethnic values common to all ethnic groups” (Mićunović 1990b; Oslobođenje 1990). This minimum of values, according to the SRSJ, was jeopardized by the nationalist parties. The leader of the Bosnian SRSJ and the rector of the University of Sarajevo, Nenad Kecmanović, described the SRSJ as a multi-national, not anti-national party, which could “stitch the torn up tissue” of Bosnia-Herzegovina’s society (Jergović 1990). The SRSJ especially criticized the new constitution of Bosnia-Herzegovina drafted by the communists because it created a “national democracy.” The SRSJ was also against the electoral system the communists chose for the elections to the republic’s presidency because it solidified the ethnically determined structure of this body. Kecmanović explained that national identity was just one of possible individual identities, and that it would be problematic to limit individuals’ identities to the framework of ethnic groups. He also publicly expressed firm belief in his party’s victory in the upcoming elections (Rakočević-Novaković 1990a). According to party insiders, these public statements closely corresponded with the beliefs of Ante Marković who was “one hundred percent convinced he would win the elections [in Bosnia-Herzegovina]” (Tasić 1994, 209).

In Serbia, the SRSJ positioned itself between what it saw as two dominant collective ideologies: nationalism, most notably presented by Vuk Drašković and his Serb Renewal Movement (Srpski pokret obnove, SPO), and bolshevism, presented by Slobodan Milošević and his League of Communists that transformed itself into the Socialist Party of Serbia (Socijalistička partija Srbije, SPS). The SRSJ embraced the tradition of liberal thought in Serbia, from the 19^th century through the dissident Milovan Đilas in the 1950s and the reform-oriented communists in the 1960s (Večernji list 1990). The main points of the program were multi-nationalism, market economy, democratization, and the promotion of human rights. The slogans read, “End the insanity of ethnic conflicts,” and “Enemies are not necessary to be successful” (Čolak 1990b). As in Bosnia-Herzegovina, the Serbian SRSJ criticized the new republican constitution enacted in September 1990 by the Milošević regime, but here because Serbia effectively circumscribed the power of the federal government. The SRSJ leader Nenad Dimitrijević stressed that there were no solutions for Serbs and Serbia outside of Yugoslavia – a functioning and functional Yugoslavia deserved its chance (Cerović 1990).

In Montenegro, the key issue was national identity and it came to the fore after the Liberal Alliance, which promoted the uniqueness of the Montenegrin nation and the statehood of Montenegro, joined the SRSJ. Pro-Montenegrin politicians saw the SRSJ as protection against the platform of subservience to Serbia supported by the Montenegrin League of Communists (Čakić 1990). As in other republics, the SRSJ promoted pluralism and democracy, market economy, and better relations with ethnic minorities. The Montenegrin SRSJ believed that Yugoslavia should be preserved as a federation, with equality of all nations and republics (Borba 1990d).

The Macedonian SRSJ presented the transition from socialism as necessary, bringing “sweat and tears,” but not blood. The party marketed itself as a link among the republics and a reform force that represented the interests of Yugoslavia as a whole. As in Montenegro, the Macedonian Reformists presented themselves not only as politicians guaranteeing “freedom and rights to all citizens, autonomy to companies, and sovereignty to republics,” but also as guarantors of the recognition of the Macedonian nation. Critically, the party made few real attempts to mobilize Macedonia’s Albanians (Đurić 1990a, 1990b).

Considering its platform, it is not surprising that the SRSJ was the perfect target for criticism from both sides of the political spectrum. This was particularly the case in Serbia and Montenegro, where the communists and the nationalists were united in their criticism of the SRSJ, especially on two key issues: its Yugoslavism and its support for the free market. In Serbia, the SRSJ was accused of being too soft on Croatian, Slovenian, and Albanian separatism, as well as on supposed Islamic fundamentalism in Bosnia-Herzegovina. The leader of the SPO Vuk Drašković claimed that the SRSJ was an “anti-Serb conspiracy on the Vatican-Teheran line” (Radovanović 1990b). Serb nationalists saw the Yugoslavism of the SRSJ as an anti-Serb policy, called the supporters of the SRSJ “salon intellectuals and children from wealthy families,” and gave Marković the message to “go to Zagreb” (Ćosić 1990; Janojlić 1990; Sekulić 1990). When it comes to economic reforms, both the Milošević socialists and the Drašković nationalists claimed that Marković was an agent of foreign corporations who wanted to subjugate the Serbs. They argued that Marković promised to sell all Yugoslav companies to foreigners, and that meant mass unemployment (Radovanović 1990a).

Similar criticisms were heard in Montenegro. Both the League of Communists and its fake opposition in the nationalist People’s Party accused the SRSJ of harming the unity of Montenegro and of Serbophobia. The communist leader Milo Đukanović even personally wrote a newspaper article in which he attacked Marković and the SRSJ for being socially unconscious and not caring for the workers and the “ordinary people” (Đukanović 1990; Vojičić 1990c). This line of criticism was also popular among the leaders of the nationalist Serb Democratic Party (Srpska demokratska stranka, SDS) in Bosnia-Herzegovina who accused the SRSJ of not planning to protect the Serbs in Croatia and of being financed by the Croats (Vujović 1990). On the other hand, criticism from the Muslim Party of Democratic Action (Stranka demokratske akcije, SDA) and the Croatian Democratic Union (Hrvatska demokratska zajednica, HDZ) in Bosnia-Herzegovina was far more muted (Ninčić 1990b).

All of this meant that the campaign in all four republics was heated. Reformist activists were physically intimidated and their rallies were attacked by opponents in Serbia and Montenegro (Borba 1990a; Brailo 1990). The SRSJ in these two republics was also subjected to a constant barrage of negative propaganda from state owned media – something they unsuccessfully tried to counter by forming their own media sources (Bećirović 1990; Rakočević-Novaković 1990b). Due to the pressures of the negative campaign and their own disorganization and incompetence, the Reformists even failed to put forward candidates for most seats contested in Serbia proper. Their campaign was far better run in Macedonia and particularly Bosnia-Herzegovina, where the public spaces were inundated with billboards portraying Marković with the slogan “Come with us, this is the time of change.” They also engaged celebrities like famous football players, actors, and singers of all nationalities, who implored the electorate to remain united (Vojičić 1990b).

In the end, these efforts seemed not to matter much. The Reformists performed far below their expectations and the level of popularity of Marković and his reforms from that spring in all four republics, winning 16% of the votes and 14% of the seats in Macedonia, 9% of the votes and seats in Bosnia-Herzegovina, 15% of the votes and 14% of the seats in Montenegro, and 4% of the votes and just 1% of the seats in Serbia. They failed to gain any meaningful importance – let alone veto power they hoped for – in all four republics. Figure 1 shows the geographic distribution of support for the Reformists. As can be seen, the SRSJ did best in the areas of Macedonia inhabited by ethnic Macedonians, in the coastal region of Montenegro, some urban enclaves in Bosnia-Herzegovina, and in parts of Vojvodina. In Bosnia-Herzegovina, they and the remainder of the leftist spectrum were thoroughly defeated by the nationalist parties SDA, SDS, and the HDZ. In Macedonia, the election led to a highly fragmented parliament, but the Reformists managed to get only fourth place. The Reformist list in Montenegro did manage to become the main opposition force in parliament, but the communists could govern virtually unobstructed with a two-thirds majority they won. And in Serbia, the Reformists became little more than a footnote in the story of the 1990 elections. In the decades following the breakup of Yugoslavia, all of these parties were eventually marginalized. The only exception has been arguably the Alliance of Independent Social Democrats (Savez nezavisnih socijaldemokrata, SNSD) of Milorad Dodik, which has its roots in the SRSJ and which has dominated the politics of the Serb entity of Bosnia-Herzegovina since 2006.

Figure 1. Electoral Performance of SRSJ Candidates in the 1990 Elections

Deriving Hypotheses in the Context of 1990 Yugoslavia

Data-driven studies of the foundational democratic elections in Eastern Europe using real electoral returns have been rare and their findings have largely been context specific. Analyzing electoral results on the level of Poland’s 49 regions, for example, Heyns and Bialecki (1991) found that the pattern of votes for Solidarność in the partially free 1989 elections implied its electorate dominantly consisted of rural voters, rather than those from the industrial urban centers where the movement started. Kopstein and Richter’s (1992) study of the electoral districts in East Germany, on the other hand, arguably found the exact opposite. They concluded that the Christian democratic right built its campaign success in the 1990 elections exactly in areas with high proportions of industrial workers. Moving away from these structuralist arguments, Pacek’s (1994) study of early democratic elections in Bulgaria, Czech Republic, Poland, and Slovakia found that the pattern of aggregate interregional data suggested voters punished incumbents responsible for poorly performing economic reforms and the rise in unemployment.

Although many experts during the early stages of transition thought of East European electorates as terra obscura, findings of political-economic sophistication by East European voters and of their differences rooted in economic structures should not have been particularly surprising. In spite of efforts at eliminating class inequalities, East European countries under communist rule were often marked by highly politicized class differences and the resulting conflicts over economic policy (Whitefield 2002). Moreover, the early economic transition from socialism to capitalism became such a defining period – in the words of Powers and Cox (1997) a “centrally constitutive event” – for the societies of the region that it should have been almost self-evident that political behavior would be fundamentally structured by the experiences and interpretations of the reforms and their economic performance. This argument was most notably proposed by Herbert Kitschelt (1992) in the early days of the transition. Kitschelt held that post-communist party competition was likely to be centered on the issue of differential opportunities provided by the economic transformation. Crucially, unlike in Western Europe, in Eastern Europe the libertarian/cosmopolitan wing of the party spectrum was to be pro-market and thus supported by those socioeconomic strata that stood most to gain from transition.

The natural corollary of Kitschelt’s proposition was that the electoral connection between voters and parties in early post-communism was to be centered on the economic performance of reforms. There is some evidence from this period that seems to support this proposition. In a comprehensive study of voting in Eastern Europe during the first decade of transition, Joshua Tucker (2006) put forward a “transitional identity model” where parties associated with economic reforms were electorally rewarded when the economy was doing well and punished when it was doing poorly regardless of whether they were in power (Tucker 2006). Others, however, found support for the traditional incumbency hypothesis under which parties in power were held responsible for economic performance, regardless of their ideological standpoint vis-à-vis the reforms (Duch 1995; Przeworski 1996; Markowski and Toka 1998).

Deriving hypotheses related to the case of the Alliance of Reformist Forces of Yugoslavia from this literature on the political economy of voter choice in early postcommunist transition is challenging for two reasons. First, from a comparative perspective, it is difficult to find political parties in Eastern Europe during this period that would be similar to the SRSJ in both ideology and political pedigree. Liberal forces in Poland, Czechoslovakia, and Hungary were led by former dissidents, or at least people distanced from the regime. In the case of the SRSJ, all its prominent members built their careers in the SKJ. According to Wolfgang Merkel’s acclaimed typology of transition leaders, had Marković and the SRSJ succeeded, this would have been an example of “change led by elites of the old regime” (Merkel, 2010). Marković’s attempt could perhaps be compared (if we ignore the type of autocratic regimes) to the case of the last prime minister of Francoist Spain, Adolfo Suarez, who founded his own party, won the first democratic elections, and led Spain through the transition (Linz and Stepan 1996). In other words, Marković and his party belonged at the same time to that pro-reform libertarian/cosmopolitan wing of the party spectrum identified by Kitschelt, as well as to the ruling elite. The second reason why deriving hypotheses related to the SRSJ is difficult is that it is also unclear whether the Reformists should be considered as incumbents. This is something Marković himself was very much aware of: he complained to the media that his party was described as a “government party,” but was being “treated as an opposition, and has all the problems of the opposition” (Sasso 2020, 42).

These difficulties notwithstanding, we believe it would be appropriate to propose the following two sets of empirical expectations about the possible sources of electoral support for the SRSJ related to economic reforms. First, although they clearly campaigned against the ruling party, the Reformists were inextricably connected to the economic reforms of the federal government. That leads us to propose that the vote for Reformist candidates should be positively related to better economic performance overall and particularly during the electoral year that was also the year of the economic reforms. If the leftist critics of the supposed “shock therapy” of the Marković government are correct, then the Reformists should have done poorly in less developed areas, as well as in areas that suffered during the time of the reforms. Second, although the SRSJ had its roots in the League of Communists and was firmly committed to Yugoslavia’s Titoist past, it could best be described as a liberal or libertarian/cosmopolitan party firmly committed to economic transition toward capitalism. As such, its success should have been closely related to the workforce structure of the electorate. The Reformists should have done better in communities with workers employed in sectors that stood most to gain from economic liberalization like tourism or trade. Conversely, they also should have done worse in communities with workers employed in sectors more threatened by economic liberalization. This arguably included workers in Yugoslavia’s inefficient industries, but particularly concerned the workers employed in the vast public bureaucracy.

Obviously, party competition in early postcommunist elections may have been strongly influenced by the experiences of economic reforms, but this was hardly the only relevant area of political contestation. In an early review of electoral competition in postcommunist Europe, Evans and Whitefield (1993) suggested that the bases of party competition in this region largely revolved around socioeconomics, ethnicity, and parties’ valence. The nascent party scene in former Yugoslavia may have been in flux, but party ideologies and policy positions were more or less clearly defined. In the case of the SRSJ, it was apparent that its liberal platform represented an attempt to go beyond national identification, to protect Yugoslav/ist heritage, and to transform the country into a constitutionally and economically functioning federation.

This is why we believe that what can help us in hypothesizing about the sociodemographic sources of support for the SRSJ is the sizeable literature on the sources of ethnic identification as Yugoslavs during this period. More than 5% of Yugoslavia’s citizens identified themselves as Yugoslavs in the 1981 census – most notably in parts of Croatia, Bosnia-Herzegovina, and Vojvodina. Aggregate-level data research has demonstrated that Yugoslav identity was more prominent in areas of interethnic contact and higher education (Burg and Berbaum 1989; Kukić 2019), whereas survey-based research has found higher probability of Yugoslav self-identification among younger persons from nationally mixed parentage living in urban communities (Sekulić, Massey, and Hodson 1994). This line of research resonates very well with the long line of studies investigating the relationship between interethnic contact, ethnic prejudice, conflict, and the relative size and distribution of ethnic groups (Montalvo and Reynal-Querol 2005; Esteban and Ray 2011; Esteban, Mayoral, and Ray, 2012). It has to be noted that there is considerable variation in findings in this literature. The general thrust is that interethnic contact and ethnic diversity can help reduce negative ethnic sentiment, but can also lead to conflict under certain circumstances. On the other hand, conflict is much more likely in communities with high levels of ethnic polarization, in other words, where usually two ethnic groups are nearly equal in size and competing for limited public resources. This line of reasoning has been applied in one of the rare data-driven studies of 1990 elections in the Yugoslav republics, where Kapidžić (2014) found that cross-ethnic and non-ethnic voting on the communal level in the 1990 elections to the presidency of Bosnia-Herzegovina was positively related to ethnic diversity and negatively related to ethnic polarization.

Based on these lessons, we believe our third set of empirical expectations should revolve around sociodemographic factors and particularly ethnicity. In line with research on the determinants of Yugoslav identity (which is also in line with the common notions of the foundations of support for liberal parties in general), we believe the level of support for the Reformists should be higher in communities that are younger, more educated, and more urban. Here we are also particularly keen to establish the relationship between the level of support for the Reformists and the communities’ heritage of World War II resistance. The SRSJ clearly wanted to defend Yugoslavia’s Titoist legacy. However, it also presented a challenge to some of the main traits of Titoism: self-managing socialism and the ruling communist party. This is why we are agnostic when it comes to the question of whether the SRSJ should have been more supported in areas that disproportionately bore the brunt of partisan resistance in World War II. We are not agnostic, however, when it comes to ethnicity and ethnic balance. In agreement with research on the foundations of Yugoslav identity and research on ethnic diversity, polarization, and prejudice, we propose the level of support for the SRSJ should have a positive relationship with ethnic diversity and a negative relationship with ethnic polarization. Considering the vicious anti-Marković campaign by the Serbian regime of Slobodan Milošević, we also believe that the level of support for the SRSJ should have a negative relationship with the proportion of the local electorate that identified itself as Serb.

Data and Method

We base our analysis on a comprehensive set of data collected on the level of municipalities in the four republics where the SRSJ ran in the 1990 elections. In spite of shortcomings of aggregate level analyses, we focus on real electoral returns because survey data from that period was notoriously unreliable. Indeed, misleading survey results arguably led to catastrophic miscalculations by a number of electoral actors at the time, particularly in Bosnia-Herzegovina (Andjelic 2003). The dependent variable of our interest is the ln-transformed vote share for the SRSJ candidates on the municipal level. We take the natural log because the distribution of the vote share is highly skewed. Not doing so would cause two problems when using OLS regression. The first is that the residuals would likely have a non-normal distribution, which would create difficulties in the calculation of the p-values, especially in small samples. The second is that the dependent variable, as a proportion, is bounded between zero and one, while OLS regression assumes a continuous outcome variable. This could lead to predictions that exceed the possible. As a robustness check, however, we reran the analyses with a fractional logit model, which confirmed the results reported below.

Our sample includes only the 206 municipalities (out of a total of 322) where the SRSJ actually nominated candidates.² We expose the determinants of voter support for the SRSJ in a string of OLS models, where the base model only includes the number of candidates or lists running in the municipality. The substantive models reflect our three sets of empirical expectations and in turn focus on the state of the local economy, workforce composition, and demographics. All models include dummies for the republics to control for the variations inherent to the four electoral competitions.³ In spite of differences in their electoral landscapes, we believe we are justified in pooling the electoral results in all four republics due to the temporal closeness of the elections and the fact that the four SRSJ organizations clearly exhibited electoral platform coherence.

We model the state of the local economy with ten variables, focusing in turn on the size of the private sector, unemployment, worker productivity, salaries, state sector investments, and spending on social services. Our variable Private Sector represents the proportion of the municipal workforce employed in the private sector in 1990 – the year of reforms of the Marković government that were meant to open up the economy and make private initiative and enterprise easier. Unemployment represents the rate of unemployment captured by the 1991 census. We opt for this figure, even though it was tallied 4–5 months after the elections, because of the same statistical standard employed across the four republics. The variables National Income, Personal Income, Investments, and Social Spending represent the monthly per capita (in the cases of National Income and Personal Income per worker) values for 1990 that have been deflated to 1980 levels and ln-transformed. They are complemented with the four variables representing the change in these values in 1990 compared to the average values during the preceding four-year (1986–1989) period. In essence, we are trying to capture not only the comparative strength of the local economy in the electoral year, but also its change during the first full year or the Marković measures compared to the preceding political cycle covering the period following the last non-democratic elections and the 13^th Congress of the SKJ that were held in 1986.

When it comes to the composition of the municipal workforce, we use the fourteen workforce groups tallied by Yugoslavia’s Federal Bureau of Statistics in 1989 (last year for which reliable data is available) to create a more parsimonious categorization into six coherent groups: Industry (official category of industry and mining), Agriculture (agriculture and fisheries, forestry, water management), Construction (construction, transportation and communication, housing), Trade (trade, hospitality and tourism, crafts, financial services), Social Services (health and social services, education and culture), and Government (socio-political communities and organizations).⁴ Following the prevailing norms in the field, we use Industry as the reference category in our models, but we also calculate the Cox direction effects, also known as a response trace (Cornell 2002), and present the results in an online appendix. Specifically, the models presented below report the effect of an increase in a workforce sector at the expense of the size of the workforce sector Industry. The coefficients reported in the appendix, on the other hand, show the effect of a sector’s increase while spreading the offsetting decrease in the other sectors across all of them instead of concentrating it in one sector. These results are substantively the same to the ones in our more traditional conceptualization presented in the workforce model, but they lend themselves to easier interpretation.

Finally, we model the socio-demographic makeup of the municipalities’ populations with a battery of eight variables focusing on ethnicity and ethnic balance, age and economic activity, education, the urban/rural divide, and the heritage of World War II resistance. All variables are derived from the 1981 census, except for the variable World War II which captures the number of resistance fighters who died in World War II per 1000 municipal inhabitants in the census of war victims Yugoslavia conducted in 1964 (Savezni zavod za statistiku 1966). In our view, these figures are a good proxy for local communities’ partisan heritage and a variable of interest in light of the SRSJ’s platform departures from socialist economic policy, but firm commitment to the ideology of “brotherhood and unity” established during World War II resistance. Here we particularly wish to highlight the three variables that focus on ethnicity and ethnic balance: Serbs, EFI, and EPI. We included in our models the proportion of Serbs in the municipal population because of the clearly negative stance of the Milošević regime and its allies against Marković and the SRSJ. We thus wished to test whether the electoral (lack of) success of the SRSJ in a given community was related to the prevalence of Serbs in the electorate. EFI and EPI, on the other hand, represent the ethnic fractionalization index and the ethnic polarization index. As previously noted, both have been extensively used in the literatures on ethnic conflict and elections (Montalvo and Reynal-Querol 2005; Esteban and Ray 2011; Esteban, Mayoral, and Ray 2012; Kapidžić 2014). Ethnic fractionalization index represents the probability that two randomly chosen individuals from the municipal population belong to different ethnic groups. It can be considered as a good measure of ethnic diversity. The ethnic polarization index, on the other hand, represents the distance of the distribution of ethnic groups in the municipality from bipolar distribution, which represents the highest level of polarization (Montalvo and Reynal-Querol 2005). It can be considered as a good measure of ethnic competition. The descriptives of all variables used in our models are presented in Table 1.

Table 1. Descriptive Statistics

Results

We present the results of our analysis in Table 2, with the first column showing the basic model that only includes the number of candidates or lists running in a given municipality. The three subsequent columns focus in turn on our three sets of empirical expectations (economy, workforce, and sociodemographics), and the last column shows the results of the full model including all three sets of variables. When it comes to our propositions regarding the possible impact of the local economy on the electorate’s level of support for the SRSJ, we suggested that it would be challenging to consider Ante Marković and his party as the incumbents in the classical sense of that word. Marković and his associates operated at times in collaboration and at times in conflict with the ruling Leagues of Communists. In many ways, they could be considered as the liberal splinter party from the League(s) of Communists. The SRSJ certainly ran its electoral campaign on the economic record of the federal government, but the economic conditions in the local communities depended on the complex web of public policy players in Yugoslavia’s devolved system of government.

Table 2. Determinants of Vote for SRSJ

Notes: OLS regression; DV = ln(SRSJ); *p<0.05; **p<0.01; ***p<0.001

In spite of those complicated circumstances, however, we expected the Reformists to do better in communities with more propulsive and productive economies that also experienced growth (or at least did not experience as much of a contraction) during the year of the federal government’s measures. As the results of our Model 1 show, that expectation has been at best only partly confirmed. None of the four variables capturing change in economic circumstances during 1990 was statistically significant, suggesting that the electorate’s support for the SRSJ had little to do with the performance of the economic program of the Marković government. This finding runs counter to those which found electoral results in early postcommunist elections to have been driven by economic performance (Pacek 1994; Tucker 2006) and incumbency (Duch 1995; Przeworski 1996; Markowski and Toka 1998). The Reformists, however, did do better in municipalities that were more economically productive (i.e. with higher National Income per worker) and that had higher levels of Social Spending per capita. The communities that disproportionately supported the Reformists were those with more propulsive and profitable enterprises and with electorates whose education, health, and social wellbeing was invested in by the government. However, in spite of the statistical significance of these two variables, it has to be noted here that Model 1 has rather low explanatory power, putting in doubt any interpretation that the electoral results of the SRSJ depended on economic factors.

The results of Model 2 further support such a conclusion, as it is clear that the level of support for the Reformists had little to do with the structure of the local workforce. We suggested that the vote for the SRSJ should be positively related to the proportion of the workforce employed in categories most dependent on the opening and liberalization of the economy and negatively related to the proportion of the workforce employed in categories most wedded to the socialist system. In making those propositions, we were guided by the early structuralist studies of the first postcommunist elections in Eastern Europe (Heyns and Bialecki 1991; Kopstein and Richter 1992). Model 2 does show that vote for the SRSJ is indeed positively related with the proportion of the workforce employed in trade, hospitality and tourism, crafts, and financial services – arguably the sectors with most to gain from the liberalization of the economy and the measures of the Marković government. Model 2 also shows that the level of support for the SRSJ is negatively related with the proportion of the workforce employed in government bureaucracy – a segment of the workforce likely still committed to the League of Communists, rather than any of its liberal splinters. The low level of statistical significance (0.05) of these two variables and the overall poor explanatory power of the model (virtually no improvement on the base model), however, suggest to us that the Reformists’ success (or lack thereof) had little to do with the calculations of the local workforce about their prospects under the system of economic reforms instituted by the Marković government. Structuralist explanations of electoral results in the first democratic elections in countries like Poland or East Germany find little confirmation in the electoral performance of the SRSJ.

Model 3, capturing the impact of ethnicity and ethnic balance, age and activity, education, the urban/rural divide, and the heritage of World War II resistance on the level of support for the SRSJ, however, clearly presents a strong improvement on the base model. In our third set of empirical expectations, we first suggested that the liberal platform of the SRSJ should have appealed to electorates that were younger and more active, better educated, and more urban. Our predictions have been partially confirmed, with the variables Activity and Illiteracy strongly statistically significant on the 0.001 level and in the expected direction. As expected, the Reformists did better in communities that were more active and educationally advanced. The generational and the urban/rural divide, at least on the aggregate level captured by our analysis, do not seem to have had much of an impact. What also did not have much of an impact was the communities’ heritage of World War II resistance. Considering the SRSJ’s platform departures from socialist economic policy, but firm commitment to the ideology of “brotherhood and unity” established during World War II resistance, we remained agnostic in our expectations regarding the possible effect of this variable on the level of support for Reformist candidates. It seems to have made little difference. Communities with a history of stronger participation in World War II resistance were no more and no less likely to support Reformist candidates than the communities without strong partisan heritage.

What definitely made a difference, however, were ethnicity and interethnic balance. In our discussion of empirical expectations related to ethnicity, we posited that it was likely the Reformists suffered electorally in communities with greater proportions of Serb voters due to the strong anti-Marković campaign by the Milošević regime. Following the insights of the literature on ethnic conflict, interethnic balance, and electoral competition, we also suggested it was likely that the SRSJ performance was positively related to ethnic diversity, but negatively related to ethnic polarization. The Reformists’ liberal platform, we believed, was more likely to be successful in ethnically diverse or fragmented communities where past research has shown there was greater commitment to a common Yugoslav identity and political future (Burg and Berbaum 1989; Kukić 2019). Conversely, we believed their platform was less likely to be successful in ethnically polarized communities where interethnic competition for limited political power and common goods would rather lead voters to support parties with nationally exclusivist platforms. As the results of Model 3 show, those expectations were proven correct. The level of support for the SRSJ is negatively related with the proportion of Serbs in the electorate and ethnic polarization, and positively related with ethnic fractionalization. Our findings here closely mirror those of Kapidžić (2014), who focused on the effects of ethnic diversity and ethnic polarization in the 1990 elections to the presidency of Bosnia-Herzegovina. The Reformists did better in ethnically diverse, but not ethnically polarized, communities with limited numbers of Serb voters. That was likely the case because the platform of the SRSJ highlighting the need for interethnic cooperation and a common Yugoslav future struck a chord with the political views and conceptions of identity prevalent in those communities even prior to the advent of democracy. Since substantive interpretation of findings presented in Table 2 is difficult due to the ln-transformation of the dependent variable, we present the findings related to ethnicity and interethnic balance in graphic form in Figure 2, with the three lines showing the predicted levels of support for SRSJ candidates for the different values of the variables Serbs, EFI, and EPI (with values limited to those actually present in our sample).

Figure 2. Predicted Vote for the SRSJ for Different Values of Serbs, EFI, and EPI

What is particularly important to note here is that all our findings from Model 3 survive virtually unchanged in the composite Model 4, with the few previously significant economic variables in Models 1 and 2 losing their significance. In fact, the composite Model 4 does not do better than Model 3 at explaining the variance in the dependent variable, suggesting that the essence of the explanation for the (lack of) electoral success of the Reformists lay exactly in the set of sociodemographic variables introduced in Model 3. The candidates of the SRSJ presented a liberal Yugoslavist platform that resonated in communities with more diverse (but not polarized), educated, and active electorates that did not heed the nationalist call of the Milošević regime. The problem was that there simply were not enough of such communities in these four republics for the Reformists to do better.

Conclusions

Yugoslavia collapsed into open warfare in the summer of 1991. Ante Marković tried desperately to rally Western support for his economic reforms and to persuade Slovenia and Croatia not to pursue independence. In the end, however, he failed on both counts. The West first offered and then rapidly withdrew its financial support in May 1991 (Glaurdić 2011, 165–166). And the Slovenes and the Croats rejected Marković’s pleas and opted for independence a month later. Marković clung on to his office throughout the worst of the war in Croatia that summer and autumn in a hopeless attempt of doing something, whatever that something may have been. Even though his last year in office brought him little more than ignominy in virtually all corners, Marković’s attempt to chart a peaceful path out of Yugoslavia’s crisis by implementing democratic and free market reforms in the troubled federation are to this day remembered fondly by many citizens of Yugoslavia’s successor states. This “Antestalgia” is not really nostalgia for the actual successes of Marković’s reforms because arguably there were none. It is rather a longing for what may have been, for a peaceful and prosperous future that seemed possible at the time, but was instead overtaken by violence and disaster of the Yugoslav wars (Sasso 2020).

Our analysis helps us get closer to answering why Marković and his liberal platform were rejected by the overwhelming proportion of the electorate in Yugoslavia’s four republics. By studying the determinants of the geographic pattern of support for the SRSJ in the 1990 elections, we are able to shed light on the reasons behind its failure to generate sufficient electoral support to effect any real influence on the country’s precipitous collapse. In our view, three findings from our analysis need to be particularly highlighted. First, the Reformists’ failure to capitalize on the popularity of the federal prime minister and his reforms had little to do with economics. Many leftist academic commentators and contemporaneous critics of Ante Marković saw his economic reforms as neo-liberal “shock therapy” that supposedly scared segments of the population vulnerable to distress brought about by the free market. We found little evidence for this proposition. Reformists’ community-level results had little relation with the structure and economic performance of the local workforce.

Second, Reformists’ lack of success was closely related to the proportion of the local electorate that was ethnically Serb. The SRSJ and Ante Marković were subjected to a vicious attack campaign by the regime of Slobodan Milošević and his acolytes. They were publicly identified by Milošević loyalists as an “ordinary imposter and enemy of the Serb nation” (Jović 1995, 173). This campaign seems to have been very successful.

Third, and in our view most interesting, the Reformists’ performance was closely related to local communities’ interethnic balance. Reformists did well in ethnically fractionalized/diverse communities and they did poorly in communities that were ethnically polarized between two ethnic groups. Just as was the case with budding Yugoslav identity during the 1970s and 1980s, interethnic contact seems to have fostered support for the Reformists, but not in communities where limited public resources were competed over by two closely balanced ethnic groups. This is an important contribution to the literature investigating the relationship between interethnic contact, ethnic prejudice, conflict, and the relative size and distribution of ethnic groups (Montalvo and Reynal-Querol 2005; Esteban and Ray 2011; Esteban, Mayoral, and Ray 2012) because it extends the findings of that literature into the realm of pre-conflict electoral politics.

These findings are also an important contribution to the study of Yugoslavia’s crisis, democratization, and breakup. We are cognizant of all limitations presented by aggregate-level studies. Nevertheless, we believe our analysis offers essential insights into the reasons for the failure of the non-nationalist forces at the ballot box. We also believe these insights are portable beyond the borders and the historical context of 1990 Yugoslavia. They are certainly instructive for our understanding of electoral politics in post-conflict Southeast Europe, not least in contemporary Bosnia-Herzegovina. Ante Marković and the SRSJ may appear to have been unique, but their electoral experience epitomizes similar experiences by parties and politicians running on anti-nationalist platforms in societies riven by ethnonational divisions and conflict. However appealing voters may find those platforms in the abstract, when presented by the starkness of their electoral choice in an environment of competing nationalisms, they often vote strategically and lend their support to parties promoting interethnic competition and not collaboration. It is arguably exactly this sort of behavior, that may be even considered rational in the short term, that has led many multi-ethnic societies into long-term spirals of conflict that have proven to be extremely difficult to overturn.