STUDY UNEQUAL DEMOCRACIES WHO DOES (NOT) VOTE IN SERBIA Vujo Ili ć and Dušan Spasojevi ć June 2025 Low and unequal voter turnout harms democracy. It may be important to critically evaluate official voter turnout data, which often undervalues election participation due to administrative and migration-related factors. This is not an issue only in Serbia but also the broader region. Voter turnout in Serbia is closer to European averages than often assumed, potentially reaching three-quarters of present voters. Inequalities between voters and abstainers persist, with older, more educated, higher-income, and politically and civically engaged individuals more likely to vote. The government should revise the voter registry and open a public debate on effective diaspora voting rights. Parties should focus on strategic voter mobilisation, reaching out to youth and marginalised groups. Civil society organisations should promote civic engagement as a way to stimulate voting.

UNEQUAL DEMOCRACIES WHO DOES (NOT) VOTE IN SERBIA?

Content  DO WE REALLY KNOW ABOUT VOTER TURNOUT? 2 2 REGISTERED VOTERS AND VOTES CAST 3  REGISTRY AND VOTING-AGE POPULATION 4  VOTER TURNOUT 6  AND REPORTED TURNOUT – A REGIONAL PERSPECTIVE 9  POPULATION AND ACTUAL TURNOUT 11  TURNOUT SOCIAL STRUCTURE IN COMPARATIVE PERSPECTIVE 12  OF 2023 VOTING IN SERBIA 15 9 CONCLUSION AND RECOMMENDATIONS 16 References  17

UNEQUAL DEMOCRACIES – WHO DOES(NOT) VOTE IN SERBIA? 1 WHAT DO WE REALLY KNOW ABOUT VOTER TURNOUT? Voting is the most common form of political partici­pation in democracies. However, voter turnout remains one of the most puzzling aspects of the democratic process. Numerous scholars, from Downs(1957) to Franklin(2004), have attempted to explain why people vote, which factors influence voter turnout, and how context matters, often arriving at somewhat inconclusive answers(Stockemer, 2016). Nonetheless, there is general agreement that low turnout exacerbates inequalities and is detrimental to democracy (Lijphart, 1997). When examining voter turnout in Serbia, an additional layer of complexity arises. By European standards, voter turnout in Serbia is presumed to be low. However, due to administrative peculiarities in maintaining voter registries and longterm migration trends, official turnout figures are lower than the actual turnout. A closer analysis reveals a different perspective: voter turnout patterns in Serbia are more aligned with European averages than previously assumed. Moreover, similar discrepancies exist in many Central and Southeastern European countries. Unlocking this perspective is crucial for understanding voter turnout in Serbia and contextualising voter behaviour across the broader region. This study addresses the issue of voter turnout in Serbia, compares it with European and regional turnout rates, and explores the key factors contributing to undervalued turnout figures. Using multiple data sources, it also examines the social structure of turnout. The findings indicate that Serbia is not significantly different from other European countries regarding the social patterns of voter participation. This insight is both encouraging and discouraging. On the positive side, actual turnout rates in Serbia may be much higher than official figures and comparable to those in established democracies. On the negative side, the social disparities between voters and non-voters in Serbia mirror the patterns of inequality observed across the continent. The study employs several related concepts throughout the analysis – official, actual and reported turnout. Official turnout is calculated as the number of votes cast(numerator) divided by the number of registered voters(denominator). Actual turnout uses the same numerator but divides it by the census-based voting-age population instead of registered voters. Finally, reported turnout reflects the proportion of survey respondents who claim they voted in an election. 2

Registered voters and votes cast 2 REGISTERED VOTERS AND VOTES CAST Over time, participation in elections in Serbia has followed a declining trend, similar to patterns observed in many other European countries(Figure 1). The highest number of votes cast was recorded during the first parliamentary elections in 1990, with 5 million. This figure fell to a historic low of 3.2 million in 2020. However, in the last two decades, the number of cast votes has remained relatively stable, oscillating around an average of 3.8 million. Notably, the number of votes peaked at 4.1 million in the 2008 parliamentary elections, driven by high political polarisation and extensive voter mobilisation. Conversely, the lowest number of votes in 2020 was influenced by an electoral boycott and fears related to the COVID-19 pandemic. The number of registered voters in Serbia tells a different story. During the 1990s, approximately 7 million voters were registered for parliamentary elections. This figure was politically contentious, as the government was accused of manipulating the voter registry, particularly in Kosovo. Since the 2000 elections, the voter registry includes individuals with permanent residence in Kosovo, but their numbers have been stable and more realistic. Even then, the voter registry remained controversial for other reasons. Since 2000, the number of registered voters has shown an unusual, curvilinear pattern. It increased steadily from 6.5 million in the December 2000 elections to 6.77 million in 2012. Afterwards, it began to decline, returning to approximately 6.5 million by 2023, a difference of only 8,000 registered voters compared to nearly a quarter of a century earlier. This trend has raised questions about the accuracy of the voter registry. If the registry is expected to reflect the actual number of voters in Serbia, how can these changes be explained? Did the number of voters genuinely increase for over a decade and then stabilise? Given the widespread depopulation trend in Serbia and the region, how could we explain this? To answer these questions, we must examine Serbia’s population data. Figure 1 Registered voters and votes cast in parliamentary elections, 1990–2023(millions) 8 7 6 5 4 3 2 1 0 1990 1992 1993 1997 2000 2003 2007 Registered voters 2008 2012 2014 Votes cast 2016 2020 2022 2023 3

UNEQUAL DEMOCRACIES – WHO DOES(NOT) VOTE IN SERBIA? 3 VOTER REGISTRY AND VOTING-AGE POPULATION Serbia faces a significant gap between the number of registered voters and the voting-age population. Every Serbian citizen aged 18 or older with legal capacity is eligible to vote. Ideally, the number of registered voters should closely align with the voting-age population as determined by the national population census. 1 Furthermore, like many countries in the region, Serbia is experiencing a shrinking population. As a result, while the voting-age population has been steadily declining, the number of registered voters has continued to rise(Figure 2). By 2023, there was a striking discrepancy of one million between the number of registered voters and the official estimate of the voting-age population. Examining the nature of Serbia’s voter registry is important to understand this discrepancy. Unlike countries where citizens 1 The legal capacity is deprived through a court decision, and even though there is no precise data on how many Serbian citizens have lost their voting rights, the assumption is that it is in the tens of ­thousands(RZSZ, 2023). are required to register for elections actively, Serbia employs a passive voter registration system. The Ministry of Public Administration and Local Self-Government oversees the Unified Voter Register, a permanent database based on the civil registry maintained by the Ministry of Interior. In Serbia, inclusion in the voter registry requires a permanent address (OSCE-ODIHR, 2023). However, enforcement of residency registration laws is lax. Many citizens do not reside at the addresses where they are officially registered. This includes individuals who have changed residences within or between municipalities in Serbia, those who emigrated but retained permanent addresses in Serbia, and Serbian citizens living in neighbouring countries, primarily Bosnia and Herzegovina, who maintain Serbian addresses for various reasons but do not reside there. The one-million-person discrepancy between the voter registry and the voting-age population is likely attributable to Serbian citizens who no longer live in the country and Figure 2 Registered voters and voting-age population, 2000–2023(millions) 7.0 6.5 6.0 5.5 5.0 2000 2003 2007 2008 2012 Registered voters 2014 2016 2020 Voting-age population 4 2022 2023

were not counted in the census. No official estimate exists of Serbian citizens who have migrated permanently, resided abroad, or died outside the country. However, the issue has recently become a subject of heated political debate. The Statistical Office of Serbia representatives have estimated that approximately one million citizens live abroad, while the adult population within Serbia is around 5.1 million(Tanjug, 2022). This context provides a new perspective on voter turnout in Serbia. Voter registry and voting-age population 5

UNEQUAL DEMOCRACIES – WHO DOES(NOT) VOTE IN SERBIA? 4 OFFICIAL VOTER TURNOUT In the first two decades of multiparty elections, the official voter turnout in Serbia experienced significant fluctuations (Figure 3). The official turnout was highest in the 1990 parliamentary elections, the first multiparty elections, at 71 %. However, it steadily declined, reaching 57 % in the 1997 elections. The 2000 elections require a contextual interpretation due to changes in the voter registry following the Kosovo War. Although fewer people voted in 2000 than in 1997, the official turnout appeared higher because its denominator, registered voters, had decreased. In the early 2000s, as the number of votes cast gradually increased, official turnout peaked at 61 % in the 2008 elections. However, by this time, voter registry inflation had already begun, leading to official turnout figures that were lower than they seemed when adjusted for actual population trends. The steady rise in cast votes and official turnout reversed after the 2012 elections. These elections marked a significant political shift, with the Serbian Progressive Party becoming the dominant ruling party. Voter numbers declined, culminating in a historic low during the 2020 elections. This election saw official turnout fall to just below 50 % of registered voters for the first time. However, considering Serbia’s negative demographic trends and the continued inflation of the voter registry, this decline was not as sharp as it appeared. Figure 3 Official voter turnout in parliamentary elections 1990–2023(in%) 100 75 71 70 61 59 61 61 58 57 58 56 53 59 59 49 50 25 0 1990 1992 1993 1997 2000 2003 2007 2008 2012 2014 2016 2020 2022 2023 6

Official voter turnout Figure 4 Official voter turnout in selected European countries(latest comparative data) Belgium(2022) Sweden(2021) Denmark(2018) Iceland(2020) Netherlands(2022) North Macedonia(2022) Norway(2020) Montenegro(2022) Germany(2022) Austria(2022) France(2021) Italy(2021) Spain(2022) Finland(2022) United Kingdom(2022) Slovakia(2022) Greece(2018) Estonia(2022) Ireland(2022) Poland(2022) Hungary(2022) Czechia(2020) Latvia(2021) Slovenia(2021) Bulgaria(2022) Serbia(2020) Portugal(2021) Lithuania(2022) Croatia(2020) Albania(2020) Switzerland(2022) Romania(2019) 0 88.4 87.2 85.9 81.2 78.7 78.7 78.2 76.7 76.6 75.6 74.6 72.9 71.8 68.7 67.6 65.8 63.9 63.7 Average 65.6 62.8 61.7 61.7 60.8 54.6 52.6 49.1 48.9 48.6 47.8 46.9 46.8 45.1 37.8 20  50 75 100 7

UNEQUAL DEMOCRACIES – WHO DOES(NOT) VOTE IN SERBIA? Following the 2020 elections, the number of voters in elections began to rise again. A modest decline in registered voters, attributed to small-scale administrative and political adjustments, contributed to official turnout figures returning to levels seen in the early 2000s. However, when using the voting-age population as a denominator, the actual turnout in the 2023 elections was higher. This increase is unsurprising given the heightened political polarisation in Serbia and the efforts of both the ruling party and the opposition to mobilise voters, especially after a period when the opposition relied heavily on extra-institutional strategies. From a comparative perspective, Serbia’s official turnout remains very low(Figure 4). In the most recent parliamentary elections analysed in the comparative dataset(see Elsässer et al. 2022), the official turnout was 49 %, placing Serbia among the European countries with the lowest turnout. Other countries in this group include Romania(38 %), Albania(47 %), Croatia(47 %), and Bulgaria(49 %). Additionally, countries with traditionally low turnout, such as Switzerland (45 %), Lithuania(48 %), and Portugal(49 %), feature similar figures. Serbia’s turnout is well below the European average of 66 % for parliamentary elections across 31 countries. 8

Official voter turnout 5 OFFICIAL AND REPORTED TURNOUT – A REGIONAL PERSPECTIVE Low official voter turnout exhibits regional patterns that warrant closer attention(DeSilver, 2022). Some countries with notably high or low turnout in Europe are wellknown cases in the literature. For instance, Belgium’s high turnout is attributable to mandatory voting(Miller and Dassonneville, 2016). Switzerland’s low turnout stems from its unique federal system and the strong emphasis on direct democracy(Blais, 2014). However, a distinct pattern of low official turnout is concentrated in Southeastern Europe, parts of Central Europe, and the Baltics. This raises questions: Is Serbia’s case related to low turnout in other Southeastern European countries? One factor is the gap between official and reported turnout in Southeastern Europe. Reported turnout refers to the percentage of surveyed individuals claiming they voted, typically higher than the official turnout(Figure 5). This phenomenon, observed across most European countries, can partially be attributed to survey self-selection bias, as politically more active citizens are also more likely Figure 5 Official and reported turnout in European countries 100 AL RS LV RO 75 CH BG SI HR PT LT DE IS AT NO ES PL GR FI NL ME IT IE EE GB MK HU SK FR CZ SE DK BE Reported Turnout 50 25 25 50 75 100 Official Turnout 9

UNEQUAL DEMOCRACIES – WHO DOES(NOT) VOTE IN SERBIA? Table 1 Voter turnout and population figures in Southeastern Europe Country Group Voter turnout(%) Official Reported Diff. Romania 38 74 36 Albania 47 83 36 Serbia Croatia 49 75 26 I 47 70 23 Greece 64 87 23 Bulgaria 49 71 22 Slovenia 53 70 17 II Hungary 62 75 13 Montenegro 77 89 13 III N. Macedonia 79 80 1 Number of voters(millions) Registered voters Voting-age population Diff. 18.96 17.66 1.30 3.59 2.43 1.16 6.50 5.76 0.74 3.70 3.51 0.19 9.92 8.79 1.12 6.86 5.65 1.21 1.70 1.73-0.04 8.22 8.00 0.21 0.54 0.47 0.07 1.81 1.71 0.10 Migrant stock Number (mil.) Population (%) 3.99 21 1.25 44 1.00 15 1.04 26 1.09 10 1.68 24 0.16 8 0.71 7 0.13 21 0.69 33 Source: Voter turnout – latest data, Elsässer et al. 2022. Number of voters – IDEA, latest figures. Migrant stock, UN, 2020. to respond in a survey. However, Southeastern Europe stands out with its low official turnout and significant discrepancies between official and reported figures. Countries such as Romania, Albania, Serbia, and Croatia exemplify this trend. In these cases, low official turnout combined with large gaps between official and reported figures suggests that inflated voter registries, caused by administrative practices and voter migration, play a significant role, alongside the over-reporting seen in surveys globally. Focusing on Southeastern Europe reveals commonalities that enable meaningful comparisons. Many countries in this region have developed similar passive voter registration systems, inherited under-resourced state administrations, and experienced significant migration towards Western Europe over recent decades. These shared characteristics shed light on their official turnout and population dynamics. Analysing the differences between official and reported turnout in Southeastern Europe reveals three distinct groups of countries(Table 1). The first group includes Romania, Albania, Serbia, Croatia, Greece, and Bulgaria, countries with low official turnout and the largest gaps between official and reported turnout. The second group comprises Slovenia and Hungary, where the gaps are similar to those in other European countries. Finally, Montenegro and North Macedonia comprise the third group, where official turnout is higher, and the gaps are relatively small, even by European standards. 10

Voting-age population and actual turnout 6 VOTING-AGE POPULATION AND ACTUAL TURNOUT In the following step of the comparative analysis, we focus on the difference between the numbers of registered voters and the voting-age population. Registered voters exceed the voting-age population in all Southeastern European countries except Slovenia. 2 This pattern is rare in the rest of Europe, where only four additional countries share this trait: Portugal, Lithuania, Latvia, and Finland. The size of this discrepancy is largest in the first group, while it is smaller in the second and third groups. Consequently, voter registry inflation disproportionately affects the calculation of official turnout in the first group(Table 1). Migration patterns and administrative decisions significantly influence voter registry management. Countries in the first group, with high turnout over-reporting and inflated voter registries, also experience the highest migration rates. 3 Albania is a stark example: of its 3.5 million registered voters, over one million likely reside abroad, heavily distorting official turnout rates. Serbia faces a similar issue, with approximately one million citizens effectively unable to vote in parliamentary elections. In contrast, the second group, with smaller migrant populations, tends to maintain more realistic voter registries, resulting in less severe over-reporting. Understanding regional migration and exploring different methods of calculating turnout leads to the central question: what is Serbia’s actual turnout? In the 2023 elections, the official turnout was 59 %, based on a voter registry inflated by migration. When adjusted for the official voting-age population, the turnout rises to 70 %. Using the Statistical Office of Serbia’s estimate of 5.1 million voters present in the country, the turnout reaches 75 %, much closer to the reported survey turnout. This suggests Serbia’s turnout is closer to European averages, potentially reflecting broader regional trends. But what do we know about the structure of this turnout? The third group, North Macedonia and Montenegro, also has large migrant populations. However, both countries have implemented active measures to mitigate the impact of inflated voter registries. In North Macedonia, voter registry revisions were part of a political agreement between the government and the opposition. Meanwhile, Montenegro is the only country in the region with highly restrictive citizenship laws that prevent dual citizenship. These findings demonstrate that migration and administrative decisions directly influence the size of voter registries and, consequently, official turnout figures. 2 Latest available election information collected by the International Institute for Democracy and Electoral Assistance(IDEA) Voter Turnout Database, https://www.idea.int/data-tools/data/voter-turnout-­ database. 3 Latest 2020 data on the numbers of the size of migrant stock in UN member states, and its percentage of total population from the United Nations Population Division International Migrant Stock 2020 database, https://www.un.org/development/desa/pd/content/international-migrant-stock. 11

UNEQUAL DEMOCRACIES – WHO DOES(NOT) VOTE IN SERBIA? 7 SERBIAN TURNOUT SOCIAL STRUCTURE IN COMPARATIVE PERSPECTIVE Comparative survey data provides insights into the socio-demographic turnout structure in Serbia, examining factors such as age, gender, education, and class and comparing them with trends across other European countries. at 69 %, 79 %, 84 %, and 85 % for ascending age categories. However, a notable gap is emerging between Serbia and other European countries regarding turnout among middle-aged voters. In Serbia, as in other European countries, voter turnout increases with age. The youngest age group(under 30) consistently reports the lowest turnout, with the middle-aged group(30 to 59) also experiencing a decline over the last decade(Figure 6). In contrast, older voters demonstrate consistently high turnout rates, maintaining this trend. Serbia aligns with the broader European pattern, where turnout increases across age groups. In the last parliamentary election across 31 European countries, turnout by age group was recorded The gender structure of turnout in Serbia is generally balanced, resembling trends observed in other European countries. While there was little difference in reported turnout between men and women, patterns shifted between 2012 and 2020(Figure 7). In 2012, men voted slightly higher than women, but by 2020, women marginally outvoted men. Women’s turnout rate surpassed men’s by only two percentage points, situating Serbia within a group of countries where women vote slightly more than men. However, such Figure 6 Age and reported turnout in Serbia 2012–2020 90 80 70 60 50 Under 30 30–44 2012 2016 12 45–59 2020 Over 60

Figure 7 Gender and reported turnout in Serbia 2012–2020 90 Serbian turnout social structure in comparative perspective 80 70 60 Male 2012 Figure 8 Education and reported turnout in Serbia 2012–2016 90 2016 2020 Female 80 70 60 Low Medium 2012 2016 13 High

Figure 9 Class and reported turnout in Serbia 2016 90 UNEQUAL DEMOCRACIES – WHO DOES(NOT) VOTE IN SERBIA? 80 70 60 Routine jobs Semi-skilled jobs Small business owners 2016 Lower grade service class Upper service class small differences may not be statistically significant, as different surveys could yield varying results. Turnout increases with education in Serbia, as seen in most European countries. The lowest turnout is among those with the least education, followed by middle-educated groups, with the highest turnout among those with higher education (Figure 8). This pattern mirrors European averages. In the most recent parliamentary elections across 31 European countries, turnout rose with educational attainment: 69 %, 77 %, and 85 % for low, medium, and high education levels, respectively. Notably, Serbia exhibits smaller differences in turnout between educational groups compared to some countries like Poland, where these disparities are more pronounced. Serbia’s pattern is closer to that of countries like Sweden, where educational differences in turnout are minimal. workers, it is 73 %, rising to 78 % among semi-skilled workers, 82 % among small business owners, 85 % among the lower-grade service class, and 89 % among the upper service class. However, this comparison should be interpreted cautiously, as data for Serbia is limited to a single point of reference in 2016. Building on this comparative data, the final part of this analysis examines the latest 2023 elections. It uses novel data that provides deeper insights into the variables influencing voter turnout. Unlike many European countries, where turnout closely correlates with class, Serbia’s class-based turnout structure in this survey is relatively undifferentiated. The lowest turnout in Serbia was observed among those in routine jobs(74 %), while the highest was among those in the lower-grade service class(82 %). Other categories, including the upper service class(78 %), small business owners(79 %), and semiskilled workers(80 %), fall between these two(Figure 9). The lack of significant variation in turnout by class in Serbia contrasts with European patterns. For example, in 27 European countries, turnout increases linearly. Among routine 14

Determinants of 2023 voting in Serbia 8 DETERMINANTS OF 2023 VOTING IN SERBIA Previous studies indicate that the determinants of voter turnout in Serbia align closely with trends established in the broader scientific literature. 4 According to a review by Todosijev ić (2020), the likelihood of voting in Serbia decreases among lower socio-economic groups, while satisfaction with democracy correlates with a higher probability of voting. Party differentiation variables have the most significant impact on voter behaviour. In a comprehensive review of 30 years of literature, Todosijevi ć also found that education is generally a positive predictor of turnout, household income is typically not significant, no notable gender differences in electoral participation, and political interest and sympathy toward parties positively influence turnout. – Living Standard: Individuals with the highest living standards were 1.667 times more likely to vote than those with the lowest. – Party Proximity: Those aligned with the ruling parties or the opposition were 1.78 times more likely to vote than those not politically aligned. – Civic Engagement: Each additional civic or political activity(e.g., volunteering, protests, and petitions) increased the odds of voting by 1.247 times. – Education did not significantly influence the likelihood of voting in this analysis. This study extends previous findings by analysing voting behaviour in the 2023 elections. We use survey data collected by the Serbian organisation Crta in February 2024. The survey, which included questions on electoral participation and respondents’ socio-demographic characteristics, offers a nuanced and updated understanding of turnout in Serbia. The survey shows that 84 % of respondents reported voting in the December 2023 elections. This figure includes expected over-reporting. Based on adjustments for the voting-age population, the effective turnout likely approached 75 %, but still significantly higher than the official turnout rate of 58 %. These findings confirm some trends while offering new insights. Age remains a consistent predictor of turnout, with older individuals more likely to vote. Unlike most previous studies, women were found to be more likely than men to abstain from voting in 2023. Turnout also increased with higher living standards, which deviates from earlier findings but aligns with broader European patterns. This finding may explain why education did not emerge as a significant predictor in this model. As expected, political alignment and engagement in civic activities strongly correlated with turnout, reinforcing prior research conclusions. The main findings about the determinants of voting can be summarised using odds ratios. 5 These ratios measure the association between a predictor variable and the likelihood of voting, representing the factor by which voting odds change for a one-unit increase in the predictor variable. The key findings are as follows: 5 – Age: The odds of voting increased by 1.03 times for each additional year of age. – Gender: Men were 1.545 times more likely to vote than women. 4 Several works have studied determinants of electoral turnout in ­Serbia, among them the most informative overviews were Lutovac (2007), Todosije vić and Pavlović(2020), and Todosijević (2020). 15 To analyze voting or abstaining in elections, a binary variable, we employed logistic regression, a statistical method designed to predict one of two possible outcomes. The model included several key variables: respondents’ age, gender, level of education, living standard, proximity to political parties, and level of political or civic activism. The sample size consisted of 1,205 respondents(from a total of 1,251), with 1,039 reporting that they voted and 153 indicating they abstained. The model demonstrated strong predictive accuracy, correctly classifying 87.2 % of cases. It was statistically significant. χ 2 (8)= 75.7, p<.001, and exhibited a good fit, as confirmed by the Hosmer-Lemeshow test(significance value above 0.05). However, the model’s explanatory power was limited, with Nagelkerke R 2 = 0.115, indicating that it accounts for a modest portion of the variation in voting behavior.

UNEQUAL DEMOCRACIES – WHO DOES(NOT) VOTE IN SERBIA? 9 CONCLUSION AND RECOMMENDATIONS Election participation in Serbia is not as low as it appears when considering inflated voter registries caused by decades of outward migration and under-resourced election administration. Adjusted for these factors, actual turnout rates in Serbia and other Southeastern European countries may be between 70 and 75 %, approaching European averages. Contrary to common assumptions, no substantial reservoir of abstainers can be easily mobilised to participate in elections. In many ways, participation patterns in Serbia are similar to those in Europe. The likelihood of voting increases with age, education, class, political alignment, and civic engagement. These findings indicate that turnout follows similar patterns of inequality as elsewhere. However, it also shows that a complex set of factors influences voting, suggesting multiple pathways for increasing electoral participation. Recommendations for the government: 1. Revise the voter registry  authorities should prioritise revising the voter registry, as international and domestic election observers recommend. A more accurate registry would improve official turnout figures and increase public trust in the electoral process. 5. Engage Youth  people represent a key demographic for voter mobilisation. Innovative, inclusive programs and initiatives should be prioritised to increase youth participation in elections. 6. Reach Marginalised Groups  strategies should be developed to engage marginalised populations, women, and individuals with lower education levels or living standards. Targeted messaging and outreach activities can encourage their participation. Recommendations for civil society: 7. Promote Civic and Political Engagement  methods of increasing voter participation should be employed. Involvement in civic and political activities, such as community volunteering, signing petitions, or participating in protests, has increased the likelihood of voting. 2. Leverage regional expertise  could benefit from the experiences of neighbouring countries that have addressed similar issues with inflated voter registries. Regional cooperation and knowledge sharing could facilitate effective solutions.  in public debate on diaspora voting rights Before revising the voter registry, authorities should initiate an open and inclusive public debate on the voting rights of the diaspora. It is critical to address the issue of citizens having the legal right to vote without effective access to the process. Recommendations for political parties: 4. Focus on strategic voter mobilisation  parties often aim to mobilise abstainers, assuming their numbers are large. However, they should instead focus on engaging segments of the population with higher abstention rates but potential sympathy for the parties. 16

REFERENCES Blais, André(2014).“Why is turnout so low in Switzerland? Comparing the attitudes of Swiss and German citizens towards electoral democracy.” Swiss Political Science Review 20(4): 520–528. DeSilver, Drew(2022)“Turnout in U.S. has soared in recent elections but by some measures still trails that of many other countries”. Pew ­Research Center, 1.11.2022. Downs, Anthony(1957). An Economic Theory of Democracy. New York: Harper& Brothers Publishers. Elsässer, Lea, Armin Schäfer,& Jonas Wenker(2022). Unequal demo­cracies: Who does(not) vote? Voter turnout trends in the OSCE ­region since 1970. Vienna: Friedrich Ebert Stiftung. Franklin, Mark(2004). Voter Turnout and the Dynamics of Electoral Competition in Established Democracies since 1945. Cambridge: ­Cambridge University Press. Lijphart, Arent(1997).“Unequal Participation: Democracy’s Unresolved Dilemma.” The American Political Science Review, 91(1), 1–14. Lutovac, Zoran(2007). Birači i apstinenti u Srbiji. Beograd: Friedrich Ebert Stiftung, Fakultet političkih nauka, Institut društvenih nauka. Miller, Peter,& Ruth Dassonneville(2016).“High turnout in the Low Countries: Partisan effects of the abolition of compulsory voting in the Netherlands.” Electoral Studies 44(2016), 132–143. OSCE-ODIHR(2023).“Early Parliamentary Elections 17 December 2023, ODIHR Election Observation Mission Final Report”. Warsaw, 28.2.2024. RZSZ(2023).“Punoletni u sistemu socijalne zaštite u 2022. godini”. ­Beograd: Republi č ki zavod za socijalnu zaštitu. Stockemer, Daniel(2017).“What affects voter turnout? A review article / meta-analysis of aggregate research.” Government and Opposition 52(4): 698-722. Tanjug(2022).“RZS: Broj stanovnika i broj bira ča u biračk om spisku su suštinski razli č iti podaci”. Tanjug, 28.12.2022. Todosijevi ć , Bojan(2020). Društvene i ideološke osnove izborne izlaznosti – Srbija 1990–2020. U: Jovano vić, Milan,& Dušan Vučićević (ur.), Kako, koga i zašto smo birali? Izbori u Srbiji 1990−2020. godine. ­Beograd: Institut za političke studije, Službeni glasnik, 203−248. Todosijevi ć, Bojan,& Pavlovi ć , Zoran(2020). Pred glasa č kom kutijom: Politi č ka psihologija izbornog ponašanja u Srbiji. Beograd: Institut društvenih nauka. References 17

IMPRINT ABOUT THE AUTHORS IMPRINT Vujo Ili ć is a research fellow at the Institute for Philosophy and Social Theory, University of Belgrade. His research interests include democratic processes and participation, as well as political conflicts and violence. Most recently he co-edited Participatory Democratic Innovations in Southeast Europe, published by Routledge. Dušan Spasojevi ć is an associate professor at the Faculty of Political Sciences, University of Belgrade. His main fields of interest are political parties, civil society, and the post-communist democratization process. Spasojevi ć is editor-in-chief of the academic journal Political Perspectives and a member of the CRTA steering board. Publisher: FES Regional Office for International Cooperation Democracy of the Future Reichsratsstr. 13/5 A-1010 Vienna Responsibility for content: Johanna Lutz| Director, Democracy of the Future Phone:+43 1 890 3811 301 X:@FES_Democracy democracy.fes.de Contact/ Orders: democracy.vienna@fes.de Design: pertext, Berlin| www.pertext.de The views expressed in this publication are not necessarily those of the Friedrich-Ebert-Stiftung(FES) or of the organization for which the author works. Commercial use of media published by the FES is not permitted without the written consent of the FES. Publications by the FES may not be used for electioneering purposes. ISBN 978-3-98628-724-5 © 2025 ABOUT UNEQUAL DEMOCRACIES Unequal Democracies is a project by FES Democracy of the Future. The main goal is to promote comparative understanding of why inequality in voting, political representation and other democratic processes hurt our democracies. In the series Who does(not) have a seat in Parliament? we analyse the social representation of European parliaments. In the series Who does(not) vote? we investigate election turnout levels across the parameters gender, age, social class and education in European democracies. Both series contain comparative studies and selective country reports. The comparative studies lay out general trends while the country reports provide country-specific analyses about the state of particular national contexts with the aim to develop and discuss political recommendations fordecision-makers. More information at: https://democracy.fes.de/topics/inequality-democracy U

WHO DOES(NOT) VOTE IN SERBIA? Low and unequal voter turnout harms democracy. It may be important to critically evaluate official voter turnout data, which often undervalues election participation due to administrative and migration-related factors. This is not an issue only in Serbia but also the broader region. Voter turnout in Serbia is closer to European averages than often assumed, potentially reaching three-quarters of present voters. Inequalities between voters and abstainers persist, with older, more educated, higher-income, and politically and civically engaged individuals more likely to vote. The government should revise the voter registry and open a public debate on effective diaspora voting rights. Parties should focus on strategic voter mobilisation, reaching out to youth and marginalised groups. Civil society organisations should promote civic engagement as a way to stimulate voting.