Abstract
Drawing on insights from political sociology and political economy, this article examines the socioeconomic consequences of entrenched ethnopolitical clientelism in postsocialist Montenegro. As a country that experienced its first electoral transfer of power in 2020, Montenegro's political transition was marked by state capture by the Democratic Party of Socialists (DPS), a dominant party that structured the national economy to serve its political interests and support its clients. Throughout its three decades of dominance, the DPS not only developed effective clientelist mechanisms to secure popular support but also employed populist rhetoric to justify these and other undemocratic practices as necessary defenses against the “ethnonational other”, portrayed as intent on undermining Montenegro's statehood. This rhetoric deepened ethnonational antagonisms and strategically inhibited alliance-building across ethnopolitical cleavages that could have threatened DPS rule. Using multivariate analysis on a unique longitudinal dataset spanning 2007 to 2020 (N = 33,889), this study provides empirical evidence that DPS supporters were more likely to secure employment (mass clientelism) and advance to higher income categories (elite clientelism). The findings also indicate that Serbs faced discrimination not due to their ethnicity per se, but because they were the only ethnonational group overwhelmingly and consistently opposed to the DPS regime and its policies following Montenegro’s independence in 2006.
Keywords
Introduction
Montenegro offers a unique case study for examining the socioeconomic consequences of client politics during three decades of state capture. Its distinctiveness lies in the fact that it is the only European country that had not experienced a change of government at the ballot box in more than a century: since the introduction of parliamentary elections in 1906, the government changed through voting for the first time in August 2020. This article covers the postsocialist period during which the dominant party was the Democratic Party of Socialists (DPS). As the direct successor to the League of Communists of Montenegro (LCM), the DPS can be said to have remained in power continuously from 1945 to 2020, despite internal re-compositions of party elites in 1989 and 1997. Consequently, after the introduction of multiparty system in 1990, the DPS retained complete control over state organs and assets in Montenegro, effectively acting as the owner of the country's infrastructure and resources. This dominance allowed the party to organize the national economy around its political needs and the interests of its clientele (Lazić, 2018; Milovac, 2016; Uzelac, 2003). During its three decades of absolute rule, the DPS not only created a highly efficient patronage network to maintain popular support (Džankić, 2018; Keil, 2018; Laštro et al., 2023), but also developed effective populist rhetoric to justify its clientelistic and other undemocratic practices as necessary defenses against the “ethnonational other,” who allegedly sought to destroy the country's statehood and its strategic political directions (Baća, 2024a; Džankić and Keil, 2017; Komar and Živković, 2016). 1 While the identity of this “ethnonational other” changed over time, the form and dynamics of its material discrimination and discursive stigmatization remained relatively consistent. 2
Without delving into the nature and internal mechanisms of ethnopolitical clientelism, our analysis begins with the well-established fact that client politics was pivotal in shaping state–society relations even before Montenegro became an independent country (Darmanović, 2003; Popović, 2002; Sekelj, 2000). Therefore, the article focuses on socioeconomic effects of clientelism in the context of ethnonational polarization. Using a unique longitudinal dataset that covers public opinion from 2007 to 2020, we address how, and to what extent, the ethnopolitical stigmatization produced by the DPS—namely, portraying and treating Montenegrin Serbs as “enemies of the state” in the post-2006 period—has translated into socioeconomic discrimination against this etnonational group. 3 Contrary to conventional wisdom, we demonstrate that Serbs were not discriminated against due to their ethnicity, but rather because they were the only ethnic-cum-political group that has been overwhelmingly and persistently opposed to the DPS regime and its policies since 2006. 4 Simply put, the dominant ethnopolitical antagonism in Montenegro since the state referendum campaign—namely, between Montenegrins and Serbs, who together constitute the Slavic-Orthodox demographic group—differs from the local ethnonational tensions of the 1990s, or elsewhere in the post-Yugoslav region, as it runs through families. For example, it is quite common for siblings to identify differently in terms of their ethnonational belonging, sometimes even from their (grand)parents, or for an individual to self-identify completely differently in two consecutive population censuses, influenced by their political stance at the time (Baća, 2018; Bešić, 2001; Brković, 2013; Džankić, 2014; Forbess, 2013; Jenne and Bieber, 2014; Kubo, 2007). It is thus unsurprising that research consistently demonstrates the absence of ethnic distance between Montenegrins and Serbs: while both groups consider intermarriage entirely normal, ethnicity becomes problematic only when politicized in relation to issues of state power—specifically, for a member of the other ethnicity to hold a political position in the country (Bešić, 2019). Therefore, rather than referring to ethnonational cleavage, we characterize this antagonism as ethnopolitical.
In this article, we engage in an interdisciplinary dialogue between political sociology and political economy to examine, first, the ways in which state-sponsored populism deepens and widens ethnopolitical cleavages between previously fluid—and often mutually inclusive—ethnonational identities, and second, the extent to which clientelist mechanisms structuring state–society relations discriminate against an entire ethnonational group due to its party preferences. The article is organized as follows: we begin by situating our hypotheses within the sociopolitical and socioeconomic context of postsocialist Montenegro; next, we provide a descriptive overview of the dataset covering independent Montenegro from 2007 to 2020, which we then analyze using multivariate statistics; finally, we discuss our findings and reflect on the broader implications of this study.
Transition without transformation: The Montenegrin experience
While the relationship between the party and the state underwent a fundamental transformation during the postsocialist transition in Central and Eastern Europe (Ganev, 2007), Montenegro experienced what Burawoy (2001) calls “transition without transformation.” Despite three decades of formal postsocialist transition, the ruling party's evolution into a vote-buying “political machine” resulted in state capture (Laštro et al., 2023; Morrison, 2018a; Uzelac, 2003)—a feature that underpins clientelism in many (semi)peripheral states globally (Gans-Morse et al., 2014; Stokes, 2005). Under dominant party rule, Montenegro did progress from authoritarianism to a semi–consolidated democracy (Darmanović, 2003; Strmiska, 2000) and shifted from a planned to a market economy (Đurić, 2003; Uvalić, 1997). However, it also experienced backsliding into a competitive authoritarian regime marked by harsh neoliberal policies during the final decade of DPS rule (Bieber, 2020; Lazić, 2018). Similar to the rest of the postsocialist region, the processes of democratization and privatization were not mutually constitutive and reinforcing (Ost, 2005; Przeworski, 1991). Instead, they were resolved through a “low-level equilibrium” where imperfect democracy coexisted with imperfect markets (Bohle and Greskovits, 2007; Gilley, 2010). This transitional context allowed political elites to redistribute wealth into their own hands through corruption, cronyism, and privatization (Nee, 2000; Rona-Tas, 1994), leading to the emergence of a “transitional bourgeoisie” formed by exploiting inherited state resources for private gain and political connections (Eyal, Szelenyi and Townsley, 1998; Ost, 2000). Such processes rendered the formerly socialist working and middle classes the primary casualties of political transition (Birdsall and Nellis, 2003; Hellman, 1998). Alongside the rise of authoritarian forms of neoliberal policy during the postsocialist transformation, the hollowing out of socioeconomic rights paved the way for the erosion of already deficient democracies, enabling more reactionary forms of politics to fill the ideological void (Bruff, 2014; Ost, 2005). As we demonstrate, this type of politics manifested as state-sponsored populism in Montenegro: the DPS employed ethnopolitical rhetoric to legitimize its clientelistic practices, framing them as essential for defending the state against certain ethnonational groups allegedly intent on “destroying the state,” thereby exploiting societal polarization and marginalizing these collectives (Baća, 2024a, 2024b; Džankić and Keil, 2017; Laštro et al., 2023). For this reason, we refer to it as ethnopolitical clientelism.
In studies on client politics in Montenegro, evidence has primarily been inferred either from available data (Sekelj, 2000) or through selected cases of electoral violence (Mochťak, 2015). During the final years of the DPS regime, scandals involving the party's top officials revealed a well-developed strategy of clientelism, encapsulated by the motto “One person employed, four votes secured” (Jedan zaposleni—to su četiri glasa). This reasoning suggests that providing employment for one individual within a family would lead to votes for the DPS from the entire family (Milovac, 2016; Paleviq, 2020). In this section, we provide historical background for our analysis of the DPS rule, demonstrating that clientelism was a systemic feature of the DPS regime during both the democratization and backsliding phases. Our analysis also reveals that the state-sponsored populist production and political delegitimization of alleged “enemies of the state” was not an isolated case of discrimination in the postreferendum period, but rather a cornerstone of the three-decade-long DPS regime. Nevertheless, following the introduction of multiparty elections, this strategic merger of populism and clientelism shifted depending on which ethnopolitical lines of conflict were most salient for the regime's strategic objectives during different stages of transition and state configurations: the Socialist Federal Republic of Yugoslavia (SFRY) (1990–1992), the Federal Republic of Yugoslavia (FRY) (1992–2003), the confederation of Serbia and Montenegro (2003–2006), and the independent state of Montenegro (from 2006 onwards). While the stigmatization of Montenegrin Serbs through state-sponsored populism in Montenegro has been empirically examined in detail (Baća, 2018; Džankić and Keil, 2017), its implications for employment discrimination in both the public and private sectors remain largely anecdotal. 5 This article fills this gap by situating client politics within the context of ethnonational polarization, focusing on the socioeconomic consequences of entrenched ethnopolitical clientelism.
Postsocialist transition and its socioeconomic discontents
The LCM changed its name to the DPS in 1990, retaining complete control over the state apparatus and key resources inherited from the LCM (Morrison, 2009; Uzelac, 2003). Montenegro's postsocialist transition under this regime is typically divided into two periods: the “flawed/negative transition” and “genuine/positive transition” (Darmanović, 2003; Popović, 2002). Despite the introduction of multiparty system, the period from 1989 to 1997 is characterized by authoritarianism, state takeover of worker-owned property, involvement in Yugoslav wars, and international isolation. In contrast, the post-1997 period is considered a shift toward electoral democracy, the externally supported creation of the civic sector, and the neoliberal restructuring of the economy, all occurring within the context of Montenegro's reemergence on the global stage (Baća, 2017; Baća and Morrison, 2022; Darmanović, 2003; Kilibarda, 2016). 6 The DPS successfully (mis)used the first phase of economic restructuring—when “social ownership” was eliminated and previously worker-controlled property was distributed between state funds and the employees of the respective “social enterprises”—to effectively centralize power through control of all sectors of the economy (Đurić, 2003; Uvalić, 1997). Amid the complete collapse of industry during the 1990s, which left the previously organized and politically potent workforce decimated through mass layoffs and administrative leaves, the government also took it upon itself to support these workers based on their “political suitability” by providing a percentage of lost wages and some social benefits (Đurić, 2003; Lazić and Sekelj, 1997). As this situation was further exacerbated by the rigorous sanctions imposed by the United Nations, patronage emerged as an effective method for the DPS to maintain social peace, influence election outcomes, and exercise control over all aspects of social life.
During the 1990s, there were approximately “300,000 recipients of state funds, or almost the same number as registered voters,” providing the regime with an unprecedented pool of votes (Sekelj, 2000: 62). The decrease in the number of workers in industry was accompanied by a gradual increase in public administration jobs—by the end of the decade, one-third of the workforce was employed in this sector—which was tightly controlled by the party structures (Đurić, 2003). In a context of declining socioeconomic security and widespread impoverishment, fueled by economic reforms under challenging external circumstances—namely, international isolation and regional wars—access to any form of welfare, however limited, became essential for most households, fostering dependence on meager government handouts. Moreover, with approximately one-third of the population living below the poverty line in the 1990s, the deteriorating standard of living not only forced people to rely on clientelist networks for survival but also drove widespread participation in the gray economy and black market (Đurić 1999, 2003; Lazić and Sekelj, 1997; Uvalić, 1997). Since the government itself resorted to smuggling cigarettes and oil due to international isolation, its own illegal activities compelled it to turn a blind eye to the expanding informal economy in order to maintain social peace and garner support, further solidifying its patronage ties with society (Baća, 2024a; Bulatović, 2004; Gallagher, 2003; Morrison, 2009). Ultimately, these socioeconomic relations enabled the DPS to reinforce state capture by combining well-established clientelist networks for political support with the oppression of critics and suppression of rivals (Baća, 2024b, Darmanović, 2003; Mochťak, 2015; Sekelj, 2000), thereby ensuring electoral victories.
When the system began to partially democratize in 1997, the clientelist party machinery was already well-established and fully operational; however, some of its mechanisms were reconfigured to reflect changes in the nexus of polity–economy–society. As the DPS-led government distanced itself from Slobodan Milošević's regime in Belgrade during this period, substantial financial support from the United States and the European Union effectively kept the Montenegrin economy afloat (Đurić, 2003; Gallagher, 2003; Morrison, 2009). 7 Due to a shared adversary, the international community deliberately overlooked undemocratic practices in Montenegro, which continued to persist after Milošević's demise in 2000 (Baća, 2024b; Baća and Morrison, 2022; Morrison, 2018a). 8 With a state monopoly over privatization, the party nurtured a loyal economic elite that “[ensured] that the process developed to their own advantage” (Lazić, 2018: 145; see also Džankić, 2018; Uzelac, 2003). The privatization and distribution of state assets, along with capital accumulation among DPS clientelist networks, became increasingly pronounced and accelerated, particularly with the launch of mass voucher privatization in 2001. This process shifted ownership from state funds to private investment funds controlled by key regime cronies (Kilibarda, 2013, 2016). Simultaneously, the DPS-controlled government continued implementing austerity measures, cutting benefits, freezing public sector wages, rolling back employment protections, and maintaining one of the lowest minimum wages in the region, along with one of the most regressive tax systems (Đurić, 2003; Kilibarda, 2013, 2016; Koman et al., 2015). This socioeconomic configuration made people increasingly dependent on party-state structures and discretionary handouts in the form of benefits, one-off cash payments, and public sector jobs.
With Montenegro's establishment as an independent country in 2006, the DPS assumed control over nearly every aspect of the state, leading to an “aggressive and rapid privatization process” which, coupled with kickbacks tied to the influx of foreign direct investment, “created a new, and often brash, nouveau riche at the expense of many ordinary citizens” (Morrison, 2018b: 143). Despite the privatization strategy being characterized by predatory enclosures and asset stripping—where new owners frequently used bankruptcy as a tool to displace the social costs and risks of privatization onto the public (Kilibarda, 2016; Koman et al., 2015) – the emerging oligarchy was effective not only in prioritizing personal interests over public ones but also in portraying their own enrichment as beneficial to the common good (Laštro et al., 2023; Morrison, 2009, 2018a). By manipulating the political economy of transition, this practice yielded quick profits at the expense of production, making both public and private sectors dependent on DPS-led privatization. Furthermore, the DPS used these accumulated financial resources to strategically cultivate loyal cultural, social, and intellectual elites by advancing their careers across private business, state administration, academia, media, sports, and arts and cultural sectors (Baća, 2024b; Bešić, 2024; Morrison, 2018b). In other words, the revenue generated from privatization was not merely allocated to secure social peace through clientelism; rather, it served as an investment in new elites intended to legitimize the DPS as the sole viable option for the country's democratic path, pro-Western course, and especially for state- and nation-building (Baća, 2024a; Caspersen, 2003; Lazić, 2018). In effect, the regime was well-positioned to present its continued hold on power as the national interest par excellence.
In summary, during the three-decade rule of the DPS, clientelism served not only to maintain voter patronage but also to establish a new entrepreneurial class (Džankić, 2018; Keil, 2018). By the time Montenegro became an independent state in 2006, these elites had emerged as the “winners of the transition,” while the middle class—and workers in particular—of the former SFRY came to be known as the “losers of the transition” (Baća, 2017: 38). However, research in the postsocialist region demonstrates that while client politics primarily target the poorest and most vulnerable socioeconomic groups—who rely on this mechanism to address basic existential needs—it is also upheld by patronage benefiting the wealthiest groups, for whom clientelism serves as a means of capital accumulation (Ost, 2000; Stokes, 2005). As demonstrated in our analysis, these “winners” and “losers” eventually formed two distinct client types—elite and mass—during state capture in Montenegro. We use this analytical distinction to better illustrate the socioeconomic consequences of ethnopolitical clientelism. In the case of mass clientelism, primarily targeting transitional workers and middle class, those who voted for the DPS were more likely to gain employment. In contrast, elite clientelism, directed at a select few close to the DPS, enabled them to enter a high-income category. However, it is crucial to note that, due to the constraints of a shrinking postsocialist economy, it was not feasible for the DPS to extend patronage to the entire voting body—not even during the economic boom following state independence. As a result, the party could only sustain a necessary client base to secure electoral victories. To ideologically reinforce these clientelistic practices, the DPS employed right-wing populism to justify the exclusion of Serbs—portrayed as “traitors of the country” and “enemies of the state”—from access to state resources, prioritizing the appointment of “national loyalists” within the state bureaucracy and thereby rendering clientelism ethnopolitical.
Ethnopolitical polarization and its socioeconomic consequences
In a manner similar to other postsocialist countries in Central and Eastern Europe, the 1990s in Montenegro were characterized by the ethnification of politics, a process through which ethnicity replaces other interpretative frames and becomes the primary driver of political mobilization (Offe, 1997: 51). During the first phase of the political transition, under the strong influence of Milošević's regime in Serbia, the DPS-sponsored ethnonationalist revival in Montenegro “expressed itself rather as Serbian nationalism than as a distinct Montenegrin nationalism” (Bieber, 2003: 12). Consequently, marginalization and oppression were primarily directed at Albanians, Bosniaks, Croats, and even Montenegrins who opposed the regime's policies, branding them as “enemies of the state.” However, once the second phase began, initiated by the internal party split over continued support to Milošević, the DPS's reformist wing—losing support among its traditionally conservative and pro-Serb base—was compelled to form a broad alliance with the progressive opposition. It included Milošević-opposing pro-Serb parties, civic associations, independent intellectuals, and, most importantly, ethnic minorities who made up about 20% of the electorate (Baća, 2017; Darmanović, 2003). By distancing itself from Serb nationalism as a strategy to mobilize support among other ethnonational groups, this DPS-led multiethnic coalition began to rally around Montenegrin national identity, (re)articulating it as civic and inclusive, and hence not based on any particular ethnic composition or religious belonging (Džankić, 2014; Jenne and Bieber, 2014). Conversely, the Milošević-backed opposition, now consisting of the DPS's conservative (pro-Serb) faction and right-wing pro-Serb parties, gradually shifted toward an exclusive Serb nationalism (Bieber, 2003; Popović, 2002). While this polarization was not ethnonational in nature and remained essentially ideologico-political—dividing conservatives from liberals or, more precisely, traditionalists from reformists—the new ruling coalition by the end of the decade began to portray the opposition as an “anti-Montenegro force” (Strmiska, 2000), setting the stage for new ethnopolitical exclusion.
After the fall of Milošević's regime in 2000, tensions between the governments in Podgorica and Belgrade eased, but the population remained divided, and Montenegrin “political elites had interests in the continued separation of the two republics” (Caspersen, 2003: 113). Eventually, progressive pro-Serb parties exited the government in protest, leaving the DPS and its pro-independence allies from civic and minority parties to lead the country. This reconfiguration of the political field positioned Serbs as the sole opposition to the regime's new strategic goal—Montenegro's independence—rendering them “political Serbs” (Baća, 2024b; Bešić, 2024). The previously ideologico-political antagonism was ultimately recalibrated into a relatively new ethnopolitical division within the dominant Slavic-Orthodox population: “Montenegrin identity” became synonymous with support for state independence, while “Serb identity” signified backing for a state union with Serbia (Bešić, 2001; Caspersen, 2003; Kubo, 2007). It is important to note that there is an overwhelming scholarly consensus that, prior to the 2000s, the ethnonational identity categories of “Montenegrin” and “Serb” were not mutually exclusive; rather, they were often interchangeable due to significant overlap (Džankić, 2014; Forbess, 2013; Gallagher, 2003; Jenne and Bieber, 2014). As a result, it was common for members of the same family to identify as either Montenegrin, Serb, or both simultaneously (Baća, 2018; Bešić, 2001; Brković, 2013; Kubo, 2007). However, from this point onward, even families began to fracture along ethnopolitical cleavages.
This transformation of ethnonational identification within the Slavic-Orthodox population, and its evolution into ethnopolitical polarization, was made possible due to its “situational” nature, with its fluid content changing over time as a result of “shifts in the salient political cleavages at the local, national, and international levels” (Jenne and Bieber, 2014: 439). Unlike the high ethnic distance between the Slavic-Orthodox population and other ethnonational groups during the 1990s—which was easily manipulated by state power for political purposes (Bešić, 2001; Bieber, 2003)—the ethnopolitical antagonism within the dominant demographic group had to be engineered “from above” because the ethnic distance between Montenegrins and Serbs was, and still is, virtually non-existent. The only contentious issue between Serbs and Montenegrins revolves around the other group “holding a leadership position in Montenegro,” with no statistically significant interethnic differences in other aspects (Bešić, 2019). Since Serbs were the only ethnonational group opposing independence and remaining staunchly against the DPS, the “civic composition” of the newly established polity was defined by the exclusion of Serbs, following a unique formula of Montenegrins plus ethnic minorities minus “political Serbs” (Baća, 2018: 128–131). As this antagonism permeated everyday life throughout much of the DPS's rule (Banović, 2016; Brković, 2013; Forbess, 2013), it consistently fueled a political narrative in which “pro-independence Montenegrins became linked with the value-neutral civic patriotism, while pro-union Serbs became associated with the ever-threatening ethnic nationalism” (Baća, 2024a: 6). The result was an ethnopolitical divide between “loyal Montenegrins” and “disloyal Serbs.”
This process of politicization of ethnicity—whereby minor sociocultural differences within the Slavic-Orthodox population were strategically transformed into ethnopolitical antagonisms (Baća, 2018, 2024a, 2024b)—proved highly advantageous for the DPS in the post-2006 period. Just as it had labeled pro-independence Montenegrins, Albanians, Croats, and Bosniaks as “enemies of the state” during the SFRY and FRY, the DPS regime now portrayed Serbs as an existential threat to an independent Montenegro. The regime exploited the Serbs’ strong anti-DPS stance, fueling “the ‘fear of Serbs who will take away the not easily achieved Montenegrin statehood’ [which became] the underlying theme of every political campaign in the country” (Komar and Živković, 2016: 790). It thus effectively excluded an entire ethnonational group from the political sphere, branding it as a “threat to public order” (Džankić and Keil, 2017). After 2006, the so-called “unionist bloc” fragmented, yet it remained firmly opposed to the DPS. Since it was not feasible for the DPS regime to extend patronage to the entire electorate, its ethnopopulism justified the exclusion of “disloyal Serbs” from public administration, creating a discriminatory employment system that prioritized “national loyalists.” 9 According to official statistics from the DPS-controlled government in 2019, while Montenegrins made up 45% of the population, they accounted for 82.11% of those employed in state institutions. In contrast, only 7.3% of Serbs held such roles, despite comprising 28.7% of the population (Ministry for Human and Minority Rights, 2019: 33). Similarly, our dataset shows a decline in the percentage of Serbs in public administration, dropping from 28.5% in 2006 to 14.9% in 2020, while the proportion of Montenegrins increased from 57.0% in 2007 to 64.0% in 2020.
In summary, the DPS utilized ethnopolitical populism to portray its leadership as the epitome and indispensable condition of Montenegrin independence, exacerbating existing divisions between two ethnonational groups whose alliance on a shared political platform during elections could mathematically pose the only serious threat to its rule (Baća, 2018; Bešić, 2024; Džankić and Keil, 2017). 10 In the regime's final years, this political antagonism was strategically intensified through a geopolitical lens, with Montenegrins and other minorities seen as naturally inclined toward the West, while Serbs were perceived as aligned with Russia. This division made pro- and anti-NATO stances significant markers of ethnonational identity, further intensifying ethnopolitical polarization (Banović, 2016; Bešić and Spasojević, 2018). Therefore, this article posits that the socioeconomic discrimination against Serbs during the DPS rule was a consequence of their political marginalization: they were not stigmatized for being Serbs but because they constituted an ethnonational group overwhelmingly opposed to the DPS—that is, due to their party preferences. Accordingly, we propose two assertions: the first concerns mass clientelism, suggesting that DPS voters were more likely than “political Serbs” to secure employment; the second concerns elite clientelism, proposing that those closer to the DPS, as opposed to “political Serbs,” were more likely to enter higher-income categories.
Independent Montenegro in numbers (2007–2020)
The dataset used in this article was provided by the Center for Democracy and Human Rights (CEDEM, Centar za demokratiju i ljudska prava), one of Montenegro's most prominent nongovernmental organizations, which has systematically surveyed public opinion in the country from 2007 to 2020 at relatively regular intervals. 11 Covering the period from the independence referendum to the establishment of the first post-DPS government, the dataset encompasses the entire span of DPS rule in independent Montenegro. The data file includes 34 waves of data collection, totaling 33,889 observations, as presented in Table 1. Each wave followed consistent procedures, utilizing multistage random sampling with post-stratification sampling weights based on gender, age, and ethnicity.
Research waves with corresponding number of observations.
In addition to employing demographic categories defined by the Statistical Office of Montenegro—MONTSTAT, we constructed the main dependent variable (DV) based on a widely accepted categorization of political parties, commonly used by scholars, analysts, and the general public in Montenegro: the DPS, pro-Serb, civic, and minority parties, as well as abstainers (Bešić, 2024; Bešić and Spasojević, 2018). The label “minority parties” refers to parties representing ethnonational minorities such as Albanians, Bosniaks, and Croats, while “pro-Serb parties” includes all parties arising from the “unionist bloc” post-2006, which claim to represent Serb interests in Montenegro. The percentage of votes for each party category across research waves is illustrated in the line chart (see Figure 1). While identification with civic parties increased and support for pro-Serb parties declined over time, the primary competition during the 2007–2020 period remained between the DPS and pro-Serb parties, as key representatives of the two main ethnopolitically divided groups.
Party support from 2007 to 2020.
Our first independent variable (IV) is respondents’ employment status, classified into four categories: public sector employment, private sector employment, self-employment, and unemployment. The DPS treated the public sector at both national and municipal levels as its “own property,” making it a primary job source for party clients (Džankić, 2018; Keil, 2018; Milovac, 2016). This strategy facilitated the redistribution of public wealth, including the allocation of national resources for emerging businesses, to the DPS cronies as transitional capitalists. The regime thus became not only the “owner” of the public sector but also, in effect, the private sector that grew during the postsocialist transition (Kilibarda, 2016; Lazić, 2018; Uzelac, 2003). Consequently, we presume that the DPS voters were more likely to be employed in both sectors than supporters of other parties. For self-employment, we do not expect a significant difference between the DPS voters and others, as self-employment could either fall within extended patronage networks or serve as a survival strategy for those opposed to the ruling party. Our second IV is income. Reported income is measured on a detailed 16-point interval scale, ranging from “no income” to “more than €1500” per month. We reorganize this scale according to the 33% criterion, dividing the sample into 3 categories: low income (up to €300), medium income (€300 to €600), and high income (more than €600). We assume that, due to the DPS's exclusive access to state resources—including employment benefits and wealth redistribution to party clients—its voters would report significantly higher incomes than supporters of other parties. The distribution of votes for each party across income and employment categories is presented in Table 2.
Party votes by income and employment sector.
DPS: Democratic Party of Socialists.
We introduced age, gender, and education as control variables, along with context-specific variables such as ethnic and regional belonging. Since employment and income are closely correlated with education level, we used a five-level scale based on national education standards to address this correlation: primary education (or less), secondary vocational education, secondary academic education, tertiary vocational education, and tertiary academic education (including advanced degrees). Ethnicity, as a key structural component of ethnopolitical cleavages, is the primary predictor of party affiliation (Bešić, 2019, 2024). Finally, regional belonging captures variations in socioeconomic development, traditionally categorized on a nominal scale into the north, central, and south (coastal) regions (Bešić and Spasojević, 2018). The distribution of votes across each category of control variables is presented in Table 3.
Party votes by demographic control variables.
DPS: Democratic Party of Socialists.
Given the nature of state–society relations in Montenegro, our hypotheses test the impact of clientelist categories as predictors of party voting, shaped by longstanding ethnopolitical divides. First, we hypothesize that individuals employed in the public sector—and to some extent, those in the private sector—were more likely to support the ruling DPS than other parties (H1). Our second hypothesis proposes that individuals with higher incomes were more likely to vote for the DPS than for other parties (H2). Additionally, since the main party division lies between the DPS and pro-Serb parties, we anticipate that employment in DPS-dominated public or private sectors, along with higher income levels, would reduce the likelihood of voting for pro-Serb parties.
Analyzing the socioeconomic consequences of ethnopolitical clientelism
To test these two hypotheses, we conducted binomial regression analysis to calculate fixed-effect estimates for the IVs, accounting for the nested structure of observations within each research wave. The DVs are organized as five dummy variables for the four party categories: voting for each party was coded as 1, while not voting for that party was coded as 0, with abstainers coded as a separate dummy variable. Accordingly, we created variables for the DPS, pro-Serb, civic, and minority parties. Each regression model presents logit coefficients to evaluate the log odds of the predictors’ influence on voting for each party category. The IVs are also organized into a set of dummy variables, each coded as 1 for yes and 0 for no. To examine the effect of employment status on party voting, we categorize respondents into four dummy variables: public sector employee, private sector employee, self-employed, and unemployed, with unemployment as the reference category. We excluded students, homemakers, pensioners, those not seeking employment, and those who declined to answer. For income, we categorize respondents into three dummy variables: high income, medium income, and low income, with low income as the reference category. Respondents who did not disclose their income or abstained from voting were excluded. Control variables are organized in a standard format: gender was coded as 1 for male and 0 for female. Age is used as an interval scale. Education is divided into five dummy variables: primary education (or less), secondary vocational education, secondary academic education, tertiary vocational education, and tertiary academic education, with primary education as the reference category. Those who did not declare their education were excluded. For ethnicity, we use four dummy variables: Montenegrin, Serb, Bosniak, and Albanian, coding each ethnicity as 1 and others as 0, with Serbs as the reference category. Lastly, we use three dummy variables for regional belonging: north, central, and south, with the south region as the reference category.
By clustering each case within the referent research wave, we deliberately applied fixed-effect estimates to integrate error at both individual and aggregate levels in each regression coefficient. Our research question does not seek to estimate the effect of time on a variable; rather, it aims to assess the impact of the IV on the DV regardless of the time span. This approach is based on the fact that, over three decades, the DPS continuously exerted tight control over the state apparatus, filling public institutions with “national loyalists.” According to our data, these clients were not Serbs, as they predominantly voted for pro-Serb parties during this period. We argue that “being employed” presupposes a hiring process shaped by ethnopolitical clientelism: for an individual to “be employed,” they must first “become employed” through a process that had unfolded over three decades of state capture.
In Table 4, we test H1, which posits that public and private sector employees were more likely to vote for the DPS than for other parties during the period under examination. We present five models, each estimating the probability of the IVs in predicting voting behavior across each party category. Overall, lower education levels correlated with a higher likelihood of voting for the DPS, while individuals with higher education were more likely to favor civic parties. Respondents with less education were also less likely to vote for pro-Serb parties, although there were no significant differences in the probability of voting for pro-Serb parties between the lowest and highest education categories. Logit coefficients for minority parties indicate that only university-educated individuals were significantly more likely to vote for these parties. Regional affiliation also impacted voting patterns: residing in the northern region, compared to the southern region, was associated with a higher likelihood of voting for the DPS, pro-Serb, and minority parties, while it was negatively associated with voting for civic parties. Notable differences were also evident between the central and southern regions, where voters in the central region were less likely to vote for civic parties and more likely to support pro-Serb parties. Finally, ethnicity emerged as a strong predictor of party preference: Montenegrins tended to vote for the DPS and civic parties; Bosniaks and Albanians, in comparison to Serbs, showed a stronger inclination to vote for minority parties as well as the DPS and civic parties; and Serbs almost exclusively supported pro-Serb parties.
Predictors of party voting: binomial fixed effects estimates.
Source: CEDEM *** p < .01 ** p < .05 * p < .1.
(Note: The dependent dummy variable represents party affiliation. The reference category for education is “primary education (or less)”; for ethnicity, it is “Serb”; for region, it is “south”; and for the main predictor, employment, it is “unemployed.”)
CEDEM: Centar za demokratiju i ljudska prava; DPS: Democratic Party of Socialists.
The main findings in the first set of multivariate models highlight employment status as a significant predictor of voting behavior. First, according to the log odds, being employed in the public sector, as opposed to being unemployed, increased the likelihood of voting for the ruling DPS by 2.7 times. Additionally, public sector employees showed the lowest rate of abstention: Model 5 indicates a 95.2% lower probability of abstaining among public sector employees compared to the unemployed, suggesting that public sector employees faced stronger incentives or pressures to vote, likely due to their greater visibility within party structures. In other words, public sector employees were more likely to vote, and when they did, they were nearly twice as likely to support the DPS. Second, the logit coefficients for pro-Serb and civic parties were negative among public sector employees, showing a lower probability of voting for these parties. Specifically, Exp B reveals that public sector employees were 42% less likely than the unemployed to vote for pro-Serb parties and 22.1% less likely to vote for civic parties. Finally, no significant difference was observed between public sector employees and the unemployed in the likelihood of voting for minority parties. Overall, these results indicate that clientelism predominantly excluded individuals who were staunchly opposed to the DPS while it held power—namely, Serbs. Importantly, each model controls for ethnicity, so the key finding is not that being Serb placed individuals in a disadvantaged employment position, but rather that voting for pro-Serb parties was the significant factor.
A similar trend is observed with smaller logit coefficient values for the private sector employees. First, compared to the unemployed, they were more likely to vote for the DPS and less likely to vote for other parties. Specifically, Exp B shows that private sector employment increased the probability of voting for the DPS by 58.1%. Second, those employed in this sector were 13.3% less likely than the unemployed to vote for pro-Serb parties. Third, there was no statistically significant relationship between private sector employment and voting for civic or minority parties. This finding suggests that clientelistic patterns in the public sector also extended into the private sector, albeit less effectively. Entrepreneurs within the DPS patronage networks appear to have employed clientelist or coercive mechanisms to ensure that employees did not abstain and turned out to vote; when they did vote, they tended to support the DPS. Finally, self-employed voters were less likely to abstain. Exp B shows that self-employment reduced the likelihood of abstaining by 30.6%. The data also indicates that being self-employed, rather than unemployed, decreased the chances of voting for a pro-Serb party by 27.3% and increased the probability of voting for the DPS by 47.1%. Again, self-employment was not a significant predictor of voting for civic or minority parties.
In summary, all forms of employment, compared to unemployment, decreased the likelihood of abstention and increased the probability of voting for the DPS. This effect was strongest among public sector employees, somewhat weaker in the private sector, and weakest for the self-employed. This finding suggests that clientelist mechanisms were most effective in the public sector, where the ruling party exercised direct influence and control over individuals. Notably, when abstention was used as a DV, additional significant findings emerged. The data demonstrates that Serbs, compared to other ethnonational groups, were more likely to abstain. Membership in any other ethnonational group reduced the probability of abstention (e.g. Albanians were the least likely to abstain, followed by Bosniaks, who were less likely to abstain than Montenegrins). Essentially, during the DPS tenure, Serbs were less inclined to participate in general elections; however, when they did vote, they tended to cast their ballots against the DPS. It remains unclear whether this pattern stemmed from a lack of belief in electoral change due to the popular narrative of rigged elections (Komar and Živković, 2016), or if the DPS employed tactics to dissuade or limit Serbs from voting, such as alleged ID-buying schemes (Milovac, 2016).
In Table 5, we test H2, which posits that individuals with higher incomes were more likely to vote for the DPS than for other party categories during the examined period. We present four additional models, each calculating the relationship between the IVs and voting choice for each party category. These models utilize a dummy variable for party category with two dummy income variables as predictors, along with the previously defined controls. According to Exp B, individuals in the high-income category, compared to those in the low-income category, were 71.8% more likely to vote for the DPS. Additionally, being in the high-income category showed a negative log odds probability of voting for any other party category. Specifically, individuals in the high-income category were 48.9% less likely to vote for pro-Serb parties than those in the low-income category. However, there was no significant difference between high- and low-income groups regarding the probability of voting for civic and minority parties. Furthermore, individuals in the medium-income category, as opposed to low-income individuals, had a 40.6% higher log odds probability of voting for the DPS. Conversely, being in the medium-income category reduced the chances of voting for pro-Serb parties by 22.5%. Based on this data analysis, we conclude that higher income levels correlate with an increased likelihood of voting for the DPS and a decreased likelihood of voting for pro-Serb parties.
Predictors of party voting: binomial fixed effect estimates.
Source: CEDEM *** p < .01 ** p < .05 * p < .1.
(Note: The dependent dummy variable represents party affiliation. The reference category for education is “primary education (or less)”; for ethnicity, it is “Serb”; for region, it is “south”; and for the main predictor, income, it is “low income.”)
CEDEM: Centar za demokratiju i ljudska prava; DPS: Democratic Party of Socialists.
To further test H2, we reversed the DV and IV in a multivariate model and calculated the Exp B percentages, as shown in Figure 2. We assessed the probability of voting for each party by comparing employed individuals across employment categories to the unemployed, and by comparing those in the high-income category to those in the medium- and low-income categories. As before, students, homemakers, pensioners, those not seeking employment, and those who declined to answer were excluded from the analysis. Each log odds calculation also controlled for gender, age, education, ethnicity, and region. The data indicates that DPS voters were 89.2% more likely to be employed in any sector than the unemployed. In contrast, pro-Serb party voters were 22.1% less likely to be employed. Employment status showed no significant effect for voters of civic and minority parties. Similar patterns emerged when party choice was used as a predictor of income level. The log odds for DPS voters suggest they were 38.5% more likely to be in the high-income category compared to those in the medium- or low-income categories. As expected, pro-Serb party voters, who had higher unemployment rates, were 30.1% less likely to be high-income earners than medium- or low-income earners. Civic party voters exhibited an 11% higher likelihood of being employed compared to being unemployed, while employment status for minority party voters did not reach statistical significance. In conclusion, these findings suggest that employment patterns align closely with the ethnopolitical cleavage between DPS supporters and pro-Serb party voters.
Log odds of employment and high-income categories. (Note: The log odds for the category “employed” among civic and minority parties are not significant. Additionally, the log odds for the category “high income” for minority parties are also not significant.)
Finally, socioeconomic discrimination against Serbs can be observed more directly during the period under examination. In Table 6, we use employment and income as DVs and ethnicity as an IV to demonstrate these direct effects. We conducted a multinomial regression analysis to calculate the probability of being in the high- and medium-income categories compared to the low-income category. The main predictor is Serb ethnicity, coded as 1, while all other ethnicities are coded as 0. The model controls for a full set of demographic factors, including employment sector. Our analysis demonstrates that being Serb, as opposed to belonging to any other ethnonational group, reduced the probability of being in the high-income category by 42.6%, while the probability of being in the medium-income category was 24.9% lower. This finding suggests that discrimination against Serbs during DPS rule also manifested in income levels. Similarly, Table 7 presents a multinomial regression analysis with employment status as the DV, using unemployment as the reference category. The results indicate that being Serb, rather than belonging to any other ethnonational group, was associated with a 40% lower probability of being employed in the public sector compared to being unemployed. The likelihood of Serbs being self-employed or employed in the private sector is similar to that of the unemployed. In other words, this model reveals that employment discrimination against Serbs was specific to the public sector during the period under study.
Predictors of income category: multinomial fixed effect estimates.
Source: CEDEM *** p < .01 ** p < .05 * p < .1.
(Note: The dependent dummy variable represents income level, with “low income” as the reference category. The reference category for education is “primary education (or less)”; for employment, it is “unemployed”; and for region, it is “south.”)
CEDEM: Centar za demokratiju i ljudska prava.
Predictors of employment category: multinomial fixed effect estimates.
Source: CEDEM *** p < .01 ** p < .05 * p < .1.
(Note: The dependent dummy variable represents employment sector, with “unemployed” as the reference category. The reference category for education is “primary education (or less)”, and for region, it is “south.”)
CEDEM: Centar za demokratiju i ljudska prava.
In summary, our findings demonstrate that while in power, the DPS maintained broad support across all demographic groups, highlighting the extent of state capture and the reach of its clientelist networks across education, public services, and the private sector, as well as among diverse income, age, and gender categories. The only notable exception—apart from a minor divergence among low-income voters—was the voting behavior of pro-Serb party supporters, with a clear majority opposing the DPS, effectively positioning the DPS as a catch-all party for all groups except Serbs. In other words, Serbs faced discrimination in both employment opportunities and income levels, as our analysis clearly demonstrates that those employed and with higher incomes predominantly voted for the DPS. These findings suggest that ethnopolitical group loyalties served not only as determinants of partisanship but also as outcomes of systemic discrimination against an entire ethnonational group due to its anti-DPS party preferences. Beyond revealing the socioeconomic impacts of ethnopolitical clientelism during three decades of state capture, these results suggest avenues for future research. Such studies could investigate questions that our survey data cannot address, such as the specific mechanisms through which this clientelism operated in Montenegro.
Concluding discussion
Clientelism is a prominent feature in postsocialist countries of Central and Eastern Europe—particularly in Southeastern Europe—yet few parties have managed to consolidate and sustain such a system as effectively or for so long as the DPS in Montenegro. Despite a formal political transition, substantive democratic transformation did not occur during its three-decade rule. Certain structures dictating state–society relations were retained from the socialist system, including deep state networks and the “leading role” of the party. Popular support from a relatively small electorate in a country of approximately 620,000—with strong kinship ties within a compact territory—made clientelism both effective and efficient during the postsocialist transition. This system was further reinforced by ethnopolitical antagonisms, which provided ideological justification for discriminating against specific ethnonational groups labeled as “enemies of the state” during the DPS's tenure. As our findings illustrate, the 2006 statehood referendum and its aftermath marked a turning point, driving a political realignment within the dominant Slavic-Orthodox population and creating an ethnopolitical divide within it, between Montenegrins and Serbs. This antagonism introduced new ethnopolitical patterns in party support—a cleavage that the DPS strategically exploited by politically delegitimizing Serbs as “disloyal citizens” whose ultimate goal was to “undermine the independent state.”
As the DPS structured the national economy to align with its political objectives and the interests of its clientele, redistributing power, privileges, and material assets over three decades, its clientelist system—alongside broader structural challenges such as poverty, polarization, and austerity measures—impeded the development of alternative market mechanisms for addressing unemployment. Consequently, securing jobs outside the DPS's patronage networks became extremely difficult in both the public and private sectors. Given the fluid nature of ethnonational identity among the Slavic-Orthodox population, where self-identification as either Serb or Montenegrin was often politically driven, a dominant party with a strong ethnopolitical agenda could, under such discriminatory conditions, shape ethnonational self-identification across significant segments of the population. Therefore, building on the well-documented premise that clientelism was deeply entrenched in Montenegro long before 2006, this article shifts focus from investigating the mechanisms of ethnopolitical clientelism to examining its socioeconomic consequences. It does so within a context where the titular nationality (Montenegrins), despite lacking a demographic majority (45%), maintained a super-majority presence in state institutions (82.11%) during the period under study. Contrary to conventional wisdom, our analysis reveals that Serbs were not discriminated against due to their ethnicity/nationality alone but rather because they were the only ethnonational group overwhelmingly and consistently opposed to the DPS regime and its policies since 2006. For this reason, the political stigmatization of Serbs as “enemies of the state” due to their party preferences easily translated into socioeconomic discrimination within both public and private sectors.
Comparative experiences from the postsocialist region suggest that when power shifts, a newly elected party aiming to maintain dominance through clientelism cannot simply replace all individuals loyal to the previous regime. Instead, it often compensates by integrating new supporters to bolster its partisan base, resulting in an “extended public sector” that increases budget expenditures (Pavlović and Bešić, 2019). This article, therefore, shifts focus from foundational aspects of clientelism—such as welfare provision or public sector employment—to an examination of how ethnopolitical clientelism was employed in a context of widespread poverty and labor market precarity, where limited socioeconomic benefits became “valuable” resources distributed by a party-controlled state over three decades. Furthermore, it is crucial to acknowledge that cycles of hiring and firing associated with regime changes offer limited insight into the enduring impacts of single-party clientelism. In Montenegro, however, the ruling party effectively “[became] the state” for thirty years, serving as a “mechanism through which members [advanced] their political and economic interests” (Morrison, 2018a: 181, 167). Therefore, by examining the continuity of clientelism within the framework of multiparty politics in a single country, our findings have broader implications for understanding state capture in the postsocialist region. Specifically, they demonstrate how clientelism can persist as a systemic feature under a dominant party, enduring through phases of formal democratization and subsequent backsliding, with the flexibility to adapt its (ethno)ideological content over time. This adaptability enables the ruling party to maintain a political advantage by continously redrawing ethnopolitical boundaries during state capture, determining which “ethnonational other” is denied access to socioeconomic benefits.
Footnotes
Acknowledgement
The dataset was provided by the Center for Democracy and Human Rights (CEDEM). Both authors contributed equally to the writing of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project has received funding from the European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie grant agreement no. 897686, as well as from the national scientific research project grant Beyond Ethnification: Postnationalist Perspectives of Montenegrin Identity, funded by the Ministry of Education, Science, and Innovation of Montenegro.
