STUDY UNEQUAL DEMOCRACIES WHO DOES (NOT) VOTE IN CROATIA? Fewer than one in two Croatians go vote in national elections. Croatia has the forth lowest turnout in Europe. This reports investigates who are the Croatian non-voters and what may be done to reach them. In line with general European trends, the Croatian young, less educated, working class are likely to abstain from voting. Croatian non-voters have low political interest but at the same time – in contrast to other countries – are not disproportionally dissatisfied with democracy. Bartul VuksanĆ usa and Michael Jennewein March 2024 Changing Croatian non-voter behaviour requires tackling persistent structural problems. As one incremental measure we recommend discussing to lower the voting age to 16. For parties, we advocate taking lessons from election campaigns across Europe that specifically targeted non-voters.

UNEQUAL DEMOCRACIES WHO DOES (NOT) VOTE IN CROATIA?

Content INTRODUCTION 2 THE DECLINE IN DEMOCRATIC PARTICIPATION AND SATISFACTION? 3 WHO DOES NOT 4 The first group: The young  4 The second group: The less educated  4 The third group: The working class  4 HIGH ABSTENTION RATES ARE A PROBLEM FOR 6 LOOKING AT 7 Croatian turnout in comparison  7 Young people in the spotlight: An in-depth analysis  9 Who are the Croatian non-voters?  9 CONCLUSIONS AND 13 Voting at 16  13 Potentials for campaigns targeting non-voters  13 Literature  15

FES UNEQUAL DEMOCRACIES – WHO DOES(NOT) VOTE IN CROATIA? INTRODUCTION One central element of democracy is the promise to give everyone an equal say about how to our societies are run. In democracies, elections are the main means of finding out what people have to say. In representative democracies, we elect representatives that represent the diversity of all members of society and their equally different interests. These representatives have the task of forming majorities on issues and how to resolve them – through deliberation and compromise. They support and scrutinize governments to implement these majority decisions. Governments thus derive their legitimacy from winning the hearts and minds of a majority of their citizens. This all sounds beautiful and plausible. But: What if fewer and fewer people take part in elections? If fewer and fewer people exercise their right to decide who should lead the country? What happens to our democracies and the functioning of this representative system if one in every two or even more people abstain from voting? In this report – as part of the FES Unequal Democracies series – we try to shed some light on these questions by looking specifically at Croatia. Croatia is an interesting case because it has one of the highest voting abstention rates of European democracies. More than half of the Croatian electorate does not vote: in the last parliamentary election in 2020 the turnout was only 46.4%. This is a staggering figure, especially when comparing it to other countries. Only three European countries had a lower voter turnout in their last election. Voter turnout is therefore an issue that should be of particular concern to Croatian citizens and decision-makers. The aim of this report is to present evidence on how voter turnout has developed during Croatia’s democratic period, who Croatian non-voters are, and what could be done to improve the situation. In particular, the report contains the following: First, we present comparative data on voter turnout in general and Croatia in particular, based on the Unequal Democracies Comparative Data Set developed for this series(see Wenker 2024). Second, we introduce a novel data set consisting of Croatian survey data since 2000, which allows us to map who the Croatian non-voters are. Finally, we formulate general recommendations about how to meet the challenge of convincing non-voters to vote. 2

The decline in democratic ­participation and satisfaction THE DECLINE IN DEMOCRATIC ­PARTICIPATION AND SATISFACTION Numerous studies have shown that across Western democracies election turnout has been declining steadily over the past decades(e. g. Blais et al. 2004; Ferrini 2013). This is a general trend that applies to all Western democracies. Election turnout is lower everywhere than it was decades ago. A few years ago, observers might have concluded that this could simply be a sign of overall satisfaction with the way societies are developing. Today, we have all the evidence to believe that it is more likely to be a sign of deeply rooted dissatisfaction with the democratic system in general and with political parties, including their politicians and their associated institutions, in particular. Within the European Union, less than half of citizens(47%) are very or fairly satisfied with the way their democracies work in practice. The majority are not very or not at all satisfied. In Croatia, this figure is even lower. Only 29% of Croatians are satisfied with the way their democracy works(Eurobarometer 2023). Figure 1 The steady decline of voter turnout since 1970 First democratic elections in former Eastern bloc 20 First democratic elections in South-Western Europe 10 Below average turnout Above average turnout 0 −10 −20 1970 1980 1990 2000 2010 2020 The points represent individual elections – in each case deviations are shown in comparison with average turnout in the respective country across the whole period. Red points thus correspond to elections with higher turnout than average, blue lower than average. Overall, it clearly shows a constant decline in election turnout since 1970. 3

FES UNEQUAL DEMOCRACIES – WHO DOES(NOT) VOTE IN CROATIA? WHO DOES NOT VOTE? But who are the people who do not vote? Is abstention evenly spread across all social groups? The comparative reports of the Unequal Democracies series covering election turnout data from 30 Western democracies going back until the 1970s have looked at this question(Elsässer et al. 2022; Wenker 2024). And they find that there are three particular groups that drive abstention across Western democracies: a) The young b) The less educated c) The working class people do naturally vote with less knowledge than that, such as out of habit, and for simpler reasons, like for the party they always voted for. At the same time, as a recent qualitative study from Germany has shown, motivations for non-voting can range from forgetting to vote, being simply indifferent, associating stress with voting, to actually being angry with the political class. All these motivations are less likely to occur the higher educated someone is(Hagemayr et al. 2023). THE FIRST GROUP: THE YOUNG Young people turning their backs on democratic participation is not a new finding, as numerous studies have shown (Foa et al. 2020; Franc et al. 2018). Crucially, however, it is not just that young people do not vote, they are also much more detached from other forms of participation in the democratic process. They are less likely to join political parties, they rarely run for office and even more rarely hold office. And most importantly, they feel that their interests are not adequately represented in politics(Bastedo 2015) and are more likely to feel that democracy does not“deliver for them“(Open Society Barometer 2023; Harring et al. 2023). This is not to say that many young people actually show a high level of political interest, but only an outspoken minority of young people are politically active – with large regional disparities. For example, climate activism is driven by young people, but limited to Western European countries such as Germany, and much less so in Eastern or South-Eastern European countries. THE SECOND GROUP: THE LESS EDUCATED The second general predictor of voter turnout is education. The more educated someone is, the more likely they are to vote. This is in line with other research(Early et al. 2023; Scervini& Segatti 2012). One of the key reasons for this finding is the role that knowledge plays in the decision to vote or not to vote. This includes knowing how to actually vote such as where to go or how to postal vote,, what parties or candidates to vote for, reading and understanding their manifestos as well as following their campaigns. Many THE THIRD GROUP: THE WORKING CLASS The third group that disproportionately does not(any longer) participate in elections is the working class. The working class in this context is defined as unskilled or semiskilled workers such as construction workers, bus drivers, shop assistants, and to be distinguished from lower and higher service workers whose jobs require some vocational training and/or university education. This working class effect of driving abstention is even more pronounced when combined with particular age groups. Young people from the working class sector are even less likely to vote than their older counterparts( Figure 2). There are further possible reasons that drive abstention among the less educated and the working class. Firstly, there is the role of knowledge, as mentioned above. In addition, the decision to vote is also a function of someone’s perception of whether their vote actually matters. What does this mean? All voters make the implicit calculation of whether it is worth the effort to vote. This calculation depends on their belief that the representatives they are voting for will actually keep their promises and be able to implement them. In this respect, recent research has shown that democratic decision-making tends to be biased towards the preferences of the better educated and the upper social classes. These studies, which measure the responsiveness of the democratic system to specific preferences, have shown that whenever the preferences of the upper and lower halves of society diverge, many Western democracies(Elsässer et al. 2017; Gilens 2012) have adopted policies favoured by the rich. In Germany, for example, social reforms such as restrictive unemployment benefits or raising the retirement age have been adopted despite an majority of the working class opposing these policies. At the same time, policies have 4

Figure 2 The class effect across age groups 100% 90% Voter turnout 80% 70% 60% 50% Below 30 30–44 45–60 been implemented that mainly benefit the upper classes, such as the reduction of capital taxes or the introduction of tax deductions for stock market losses. Furthermore, social norms may play an important role. As Schäfer& Roßteutscher(2015) have shown, the decision to go vote is also a function of what other people in one’s social circle are doing. For instance, the norm for people with a university degree is that they naturally go to vote because it is something their peers are doing. Deviating from this norm actually requires a justification to their peers. Conversely, in groups and areas where voting is not widespread, the burden of justifying why one deviates from the norm falls on the person who goes to vote when most of their peers do not. One might say: Interesting results, but is this really a problem? All these people have the right to vote. If they choose not to exercise it, so what? Is that problem for democracy? Over 60 Who does not vote? Higher-grade service class Lower-grade service class Small business-owners Skilled workers Unskilled and semi-skilled workers 5

FES UNEQUAL DEMOCRACIES – WHO DOES(NOT) VOTE IN CROATIA? HIGH ABSTENTION RATES ARE A PROBLEM FOR DEMOCRACY There are at least four reasons for why one should be concerned about high abstention rates. Firstly, if for instance, everyone participates equally in an election except a particular group their particular interests are likely to not be adequately represented and accounted for in parliament. This collective abstention becomes specifically a problem if those groups at the same time are disproportionally affected by social and economic grievances. If we look at our three groups from above one may see the relevance. The young are disproportionately affected by issues that concern their(long-term) future like climate change and their general personal prospects, for instance, whether they can afford starting a family. Both the less educated and the working class are disproportionately affected by the ongoing economic and social transformation such as job loss or dealing with high inflation given their limited financial resources. If all those groups do not participate enough – their interests in the matters that affect them are unlikely to be at high priorities of their elected governments. And finally, persistent low election turnout may influence behaviour by parties. If many constituents do not vote anyway, other political engagement such as joining political parties is usually low as well. And understaffed parties and districts are less likely to spur voter mobilisation in the future. Parties will make the calculation if the effort to reach non-voters is worthwhile given their resource constraint. If this constraint becomes persistent one ends up at a continuing cycle of low turnout and mobilisation. As a result, entire populations might distance themselves from their political class and processes which hurts democratic resilience overall(Roßteutscher& Schäfer 2016). Secondly, one main purpose of elections is to evaluate the performance of a government. If only certain groups systematically abstain from this evaluation governments are no longer accountable for the effects of their politics that particularly affect these constituents. If these groups have much less influence in the political deliberation of decisions in actual political processes they further distance themselves from democracy and loose faith in its promise to deliver for them. As a result, they are much less represented in politics, see for instance the very low numbers of young or working-class politicians in parliament and government(Elsässer et al. 2022). At the same time, institutions such as unions or lobby groups that would represent and advocate for their interests are likely to be relatively weaker than their upper class counterparts(see e.g. Schnetzer 2021). Thirdly, if many people abstain the potential to influence the outcome of an election by controlling who participates increases. While research is inconclusive on whether turnout changes election results(Rosema 2007; Schäfer 2012) it is likely that each social group may have particular preferences often at odds with others. So, if one group does not participate some preferences will be underrepresented in democratic decision making. The working class, for instance, is known to be more supportive of redistributive policies than the upper service class. 6

Looking at Croatia LOOKING AT CROATIA But what is the situation in Croatia? In this report we present two sets of data that allow us to shed some light on voter turnout in Croatia. First, we look at where Croatia stands in relation to other European democracies using the UD Comparative Data Set. Secondly, we present an original analysis based on an exclusive primary data set that allows us to delve deeper into who the non-voters in Croatia are. in 2020, turnout was 46.4%. There are some blips, such as in 2000 and in 2015, but overall the trend is clear. In this respect, Croatia is fully in line with the general trend and – when looking at the turnout figures in general – is among the lowest in the EU and within democratic Europe. There are only three European democracies with lower voter turnout in recent elections: Romania(2020: 31.9%), Bulgaria (2023: 40.6%) and Albania(2011: 46.3%). CROATIAN TURNOUT IN COMPARISON Like everywhere else, voter turnout in Croatia has been declining over time( Figure 3). In the most recent election When it comes to who took part in the election, we also find a similar pattern to the general trend. Looking at the most recent election in 2020, we find the effects that were already introduced above: the age effect, the education effect, and the class effect. Figure 3 The steady decline of turnout in Croatian national elections 100% 85% 80% 76% 60% 69% 40% 77% 61% 62% 60% 54% ? 53% 476% ? 20% 0% 1990 1995 2000 2005 2010 2015 2020 2025 7

FES UNEQUAL DEMOCRACIES – WHO DOES(NOT) VOTE IN CROATIA? Figure 4 The social inequality of voting in Croatia 100 80 60 40 20 0 2007 2011 100 80 60 40 20 0 2007 2011 100 80 60 40 20 0 2007 2011 100 80 60 40 20 0 2007 2011 2016 2020 2016 2020 2016 2020 2016 2020 8 Over 60 45-60 30-44 Under 30 Higher-grade service class Lower grade service class Small business owners Semi-skilled workers Unskilled workers Men Women High level of education Medium level of education Low level of education

Looking at Croatia YOUNG PEOPLE IN THE SPOTLIGHT: AN IN-DEPTH ANALYSIS With this picture in mind, we conduct a more in-depth country analysis of non-voters in Croatia. We choose to focus on young people in particular because previous studies(Lamza-Posavec, 2004; Henjak, 2017; Raos, 2020; VuksanĆ usa and Šalaj, 2024; Henjak and Č ular, forthcoming; Širini ć and Dolenec, forthcoming) have shown that young people are indeed notorious abstainers(Smets and van der Ham, 2013: 348). However, it is still unclear whether this trend is the result of age effects, generational socialisation effects or period effects. with reference to the period of their political socialization (Grasso, 2016). We have primarily relied on the generational scheme proposed by Grasso(2014) and updated by Mitteregger(2024), with slight modifications due to the limited time frame of the data used. We have therefore classified respondents into the following groups: World War II generation(born between 1903 and 1945, covering both preand post-war generations), 60s and 70s generation(born between 1946 and 1957), 80s generation(born between 1958 and 1968), 90s generation(born between 1969 and 1981), and millennials(born after 1982). Finally, period effects were operationalised by categorising each survey year as a separate category. Firstly, the effects of age on political behaviour and attitudes can manifest as individual age effects(effects of one's position in the life cycle), as well as cohort or generational effects(effects of common socialisation experiences among individuals) and period effects(effects of large-scale events or the current context)(Neundorf and Niemi, 2014). If we apply the age-cohort-period framework to non-voting, this would imply that young people's abstention may be the result of a lack of voting experience(as they have not yet had the chance to internalise the habit of voting)(s. Franklin, 2004) or their increased focus on tasks and processes related to the characteristic life transition at the time(Dassonneville, 2017: 139–140). An alternative explanation is couched in a generational hypothesis, which postulates that younger cohorts may vote less because, compared to older cohorts, they are characterised by socialisation in a significantly different political-historical and value context. From the perspective of the latter, younger generations are more inclined to post-materialism, in which politics and the social norm of voting are not relatively highly valued(Norris and Inglehart, 2019; Kostelka and Blais, 2021). From a socialisation imprinting perspective, the participation repertoires of younger generations may reflect the depoliticisation of the context in which they come of age, leading to lower turnout(see Grasso et al., 2019). Finally, studies also need to consider period effects, as influential political and other events can mark specific time intervals that may influence political attitudes and levels of electoral participation. Depending on the nature and intensity of these events, they can significantly shape people's political behaviour(s. Lisi et al., 2021). At the same time, it is essential to note that age, cohort, and period are perfectly correlated(once we know two elements, we automatically know the third). Therefore, restrictions on their variance must be applied to entangle their separate effects. Accordingly, this report treats age, cohort, and period as categorical variables(see van der Brug and Rekker, 2021) to avoid their inherent collinearity. Age is thus defined by life cycle categories, where respondents are classified into four categories: adolescents(18–21 years old), early adults(22–29 years old), middle adults(30–64 years old), or late adults(65+ years old)(see Lichtin, van der Brug and Rekker, 2023). In addition, cohorts are defined as political generations, where individuals are categorised In addition to the age-period-cohort backbone of the report, the analysis includes other variables that speak to the non-voting framework outlined in the first part. Specifically, to account for the socioeconomic status of respondents, we have included indicators of education, employment, and wealth(measured as the number of assets an individual owns). Our model also includes measures of diffuse support for democracy(Norris, 1999), satisfaction with democracy (Grönlund and Setälä, 2007), and political interest(Prior, 2010). Demographic controls for settlement size and gender are also included as predictors. Overall, the model operates mainly within the widely used framework of the resource model of political participation(Smets and van Ham, 2013), which emphasises the importance of education, skills, time, experience, and socio-economic conditions for showing up on election day(see Verba and Nie, 1972). The analysis is based on data from several Croatian election studies(CroNES, s., Bovan, Širini ć , and Raos, forthcoming), which is managed by the Faculty of Political Science at the University of Zagreb. This report uses seven nationally representative pre- and post-election surveys conducted between 2003 and 2020, with almost 6,000 respondents, covering a period of gradual decline in voter turnout in Croatia. The dependent variable is binary and indicates the self-reported vote in the last previously held election (0-voter, 1-non-voter). 1 WHO ARE THE CROATIAN NON-VOTERS? First, the results of the binary logistic regression show that gender has no significant effect on voter turnout. Although more tests need to be conducted in this regard(for example, on different types of elections, cf. Kostelka, Blais and Gindengil, 2019), our results are consistent with the finding that the gender gap in electoral participation is gradually disappearing(Smets and van Ham, 2013). Similarly, although European rural and urban residents may differ in their level of political efficacy(García del Horno, Rico and Hernández, 2023), they do not differ when it comes to voting in Croatia. 1 This measure is not an unproblematic one, as individuals tend to overestimate their electoral participation in surveys(see Karp and Brockington, 2005). The same holds for the dataset employed in this study(s. Širini ć and Dolenec, forthcoming). 9

FES UNEQUAL DEMOCRACIES – WHO DOES(NOT) VOTE IN CROATIA? Figure 5 Coefficient plot: Binary logistic regression model of non-voting in Croatia(2003-2020 period). Demography Female Education| ref: Low education Middle High Employment| ref: Employed Retired Unemployed Homemaker Student Wealth| ref: 0 assets 1 asset 2 assets 3+ assets Adolescence Early adulthood Late adulthood Generations| ref: WW2 generation 60s&70s gen 80s gen 90s gen Millennials Interest& Democracy Interest in politics Support for democracy Satisfaction with democracy –1 –.5 0.5 1 1.5 Model: N=5,924, R2=0.121, X2=858.08, p=0.0000. Note: Any effect whose confidence intervals(95%) do not overlap with the 0 line on the x-axis is statistically significant. Although period was included as a factor in the model, it was not included in the graph because the first part of the report already covered voter turnout over the years. At the same time, however, there are some differences in voter turnout across respondents’ education and employment status. That is, as extensively documented elsewhere (see Smets and van Ham, 2013), our results show that education has a positive effect on voter turnout. However, as expected, the differences between the less educated and the respondents with a high-school education are less pronounced(p=0.07) than the differences between the ones with a high-school education and the highly educated(individuals with a BA, MA, and PhD)(p=0.000). Furthermore, compared to those who are employed, unemployed people are more likely to abstain from voting(p=0.001), although with the operationalisation used here we were not able to take into account the socio-economic position of an individual in more detail(see Henjak and VuksanĆ usa, 2019). Moreover, respondents with other employment statuses do not differ significantly from those who are employed. However, retired people are slightly more likely to vote, although this coefficient is not statistically significant at the conventional level(p=0.07). To provide a more concrete explanation of the results, we have included graphs showing the predicted probabilities of non-voting for different education and employment groups. In Figure 6, the left panel shows that individuals with higher education have a 24% probability of not voting, while for those with less education, this probability increases by almost 10%. At the same time, the right-hand panel shows that unemployed people have a 35% probability of not voting, which is 6% higher than for employed respondents. Taken together, these results show that the decline in voter turnout has obvious education and employment contours, as those with fewer resources are the least likely to vote in Croatia. Therefore, we found support for the argument that participation is unequal across social groups in Croatia(see Bovens and Wille, 2017: 70–71; Schäfer and Streeck, 2013: 13–15). This is further supported by the finding that individuals with more assets(2 or 3+) tend to vote more than those with no assets at their disposal(see Figure 5)(Nadeau, Lewis-Beck and Foucault, 2019). We now turn to age-related differences in non-voting. The results in Figure 5 illustrate how both life cycle and generational theories apply in this respect in Croatia. Compared to respondents in middle adulthood(aged 30–64), both adolescents(aged 18–21) and early adults(aged 22–29) tend to participate in elections at much lower rates. This difference is most pronounced among adolescents, an issue we discuss further in the final part of this report. As shown in Figure 7 (right-hand panel), adolescents have a probability of not vot10

Looking at Croatia Figure 6 Less educated and unemployed least likely to vote in Croatia Predictive probability of non-voting by education and employment groups. .5 .5 Probability of non-voting Probability of non-voting .4 .4 .3 .3 .2 Low Middle High Education Note: Confidence intervals are estimated at 95%. .2 loy l e o d yed p p m -em Se E lf Retired loyed p aker em dent Stu Une H m om Employment Figure 7 Adolescents vote the least Predictive probability of not voting by cohort and life phases. .6 Probability of non-voting .5 .4 .3 .2 gen W2 W s gen 60s&70 s gen 80 s gen 90 nials illen M Generations Note: Confidence intervals are estimated at 95%. Probability of non-voting .6 .5 .4 .3 .2 scence dole A ulthood ly ad Ear ulthood le ad idd M dult. Late a Life-cycle ing of almost 50%. This probability drops to 25% and 27% for respondents in middle and late adulthood, respectively. In short and robust terms, the probability of voting in Croatia is almost twice as high for those over 30 as for those under 21. If the habitual theory of voting holds true(Dinas, 2012, also see Blais and Daoust, 2020: 71–90) and adolescents continue to have low turnout rates as they age, a further decline in voter turnout can be expected. 2 2 Still, salience and type of elections should be considered in this regard(Franklin and Hobolt, 2011; Dinas et al., 2024). The differences between cohorts are less pronounced. Nevertheless, the 90s generations(p=0.007) and millennials (p=0.022) are less likely to vote than the WW2 generation(s)(as this category includes both pre- and post-WW2 generations). Looking at the graph of predicted probabilities( Figure 7, left-hand panel), it is noticeable that the 90s generation and the millennials vote less than the 60s/70s and 80s generations. The two youngest generations are estimated to have a 32% probability of not voting, while the same probability drops to 25% and below for older generations. Regarding these generational differences, one might conclude that younger generations in Croatia are 11

FES UNEQUAL DEMOCRACIES – WHO DOES(NOT) VOTE IN CROATIA? Figure 8 Voting increases with age Predicted probabilities of not voting by year of birth .5 Probability of non-voting .4 .3 .2 .1 1900 1910 1920 1930 1940 1950 1960 1970 Note: Confidence intervals are estimated at 95%. Model: N=5,924, R2=0.119, X2=842.93, p=0.0000.. 1980 1990 2000 Birthyear more similar to Western Europeans than their post-communist counterparts(Linek and Petrúšek, 2016). Additionally, the generational magnitudes of non-voting presented here are most comparable to the differences found between the employed and the unemployed. In other words, generational differences between younger and older generations in Croatia explain non-voting as much as unemployment, at least when controlling for basic socio-demographics. We conclude the identification of age-related differences by showing smoothed birth year effects on the probability of not voting( Figure 8)(see Grasso et al., 2019). This was done by adding the birth year variable instead of the political generation one in the model. The results provide further evidence that cohorts born in or after the 1980s (i.e., Millennials, Gen Z) differ from cohorts born before the 1960s(i.e., Baby Boomers, Silent Generation, and Greatest Generations)(Norris and Inglehart, 2019). 3 Interestingly, and in line with the findings of Inglehart and Norris, the opposite trend was found in levels of authoritarianism among birth cohorts in Croatia(Raos and Zakošek, forthcoming). Finally, we turn to the remaining indicators: political interest, satisfaction with democracy, and diffuse support for democracy. The results in Figure 5 again confirm the well-established finding that strong political interest is the most robust suppressor of the probability of not voting. At the same time, support for democracy does not seem to affect voting probability. This shows that voters and non-voters do not differ in their level of normative support for democracy. This contrasts with findings from Germany about non-voters being less democratic(Koch, Meléndez and Rovira Kaltwasser, 2021). Given that satisfaction with democracy reduces the probability of not voting in our model(p=0.012), in Croatia it may be better to think of non-voters as dissatisfied democrats rather than non-democrats. 3 However, this is a relatively simple linear model. Models that offer more flexibility(i.e., allowing for non-linearity, such as GAMs, ­s. Grasso, 2014) might produce more of a U-shaped graph, with oldest groups being somewhat more inclined to non-voting(s. Širini ć and Dolenec, forthcoming). 12

Conclusions and Recommendations CONCLUSIONS AND RECOMMENDATIONS Before formulating specific recommendations on how to counter abstention, let us recapitulate our findings: – Voter turnout in Croatia is one of the lowest in Europe. – Abstention is higher among young people, as the generations born in the 1980s and later are generally less likely to vote. Moreover, abstention among 18–21 year olds is exceptionally high. – Education and unemployment play a role: the less educated and the unemployed are more likely to abstain. Wealth also matters, as people with more assets are also more likely to vote. – Voting in Croatia is also related to people’s interest in politics. Nevertheless, non-voters are not found to be significantly less supportive of democracy in general. They are rather dissatisfied democrats than authoritarians. The challenge of decreasing abstention is twofold. On the one hand, when abstention is as entrenched as it is in Croatia, potential remedies need to address the structural issues, such as the deep institutional distrust in the political system that underlies abstention(Bovan and Baketa 2022). Any serious recommendation in this regard must consider this more complicated problem that cannot be solved with one measure. However, as dealing with institutional trust, political efficiency, political interest and socio-economic structure is well beyond the scope of this report, we turn to solutions that can have a relatively incremental effect. Therefore, the first realm of recommendations elaborates on concrete policy solutions, while the second focuses on campaigns and their reach toward non-voters. VOTING AT 16 There are several institutional mechanisms that can increase youth voter turnout. One of them is compulsory voting (Wattenberg, 2015), but this suggestion has not gained prominence in the public debate in Croatia, even though its effects could be positive by bringing party programmes closer to the preferences of the median voter and reducing polarisation(Oprea, Martin and Brennan 2024). However, the idea that has slowly found its way into academic discourse in Croatia is that of lowering the age threshold for voting(see Šalaj 2024). It is argued that if voting is habitual, in terms of higher voter turnout, it might be better to lower the voting threshold to 16 years of age, as individuals at that age are less exposed to transitional processes and searches than they are at 18 years(see Franklin, 2004; Franklin, 2020; Eichhorn and Bergh, 2021). Therefore, it is suggested that at 16, situational conditions that may discourage participation are less likely, resulting in young people voting in their first election more often and possibly making it a future habit. Although the jury is still out on the effects of lowering the voting age(Bergh and Eichhorn, 2020: 3–7; Rosenqvist, 2020), studies point out that there is not much evidence that it has been harmful where it has been implemented(Eichhorn and Bergh 2020: 238; see also Eichhorn 2018; Wagner, Johan and Kritzinger 2012; Aichholzer and Kritzinger, 2020; Franklin, 2020). Although evidence in this regard is scarce, 15–19 year old adolescents and high school students seem to be more in favour of further lowering the voting age in Croatia compared to those over 20(Ilišin 2017: 230–231). Recently, the first political impulses in this direction have started, mainly, but not exclusively, by the progressive part of the opposition(Toma 2023). POTENTIALS FOR CAMPAIGNS TARGETING NON-VOTERS When it comes to the actual electoral potential of non-voters, most mainstream party strategists follow the logic that winning campaigns is a function of attracting voters from other parties, but rarely how to reach new voters. This is understandable to some extent, for two reasons: Firstly, taking a voter away from a rival is not just a gain of one vote, it is doubly efficient because the rival also loses one vote. Adding a new previous non-voter, on the other hand, is only a net gain of plus 1 in the total number of votes against the competitors. And secondly, targeting non-voters requires parties and candidates to go out of their way, such as tailoring messages and literally going to places to talk to voters where no one has gone before. This, at least, is one finding when we look at partially successful campaigns that have targeted non-voters in the recent past. For example, in the 2017 UK election, Labour surprised the pollsters with a total of 40% of the vote, just behind the Conservatives. The 9.6% increase in voter support was partly due to reaching out to non-voters, research suggests(Dorey 2017). The campaign specifically targeted voters with traditionally low turnout, particularly in university towns and cities with large numbers of young voters, as well as grassroots and 13

FES UNEQUAL DEMOCRACIES – WHO DOES(NOT) VOTE IN CROATIA? civil society groups. One element of relatively successful non-voter campaigns therefore seems to be that atypical campaigns go out their way to reach non-voters. Contrary to popular intuition, populist parties, both left and right, are not necessarily better equipped to reach out to non-voters(Wenker 2024). The most likely reason for this is that populist voters actually have similarly high expectations of democracy’s performance as mainstream voters compared to non-voters(Koch et al. 2023). Our finding that Croatian non-voters are not specifically dissatisfied with democracy is in fact contrary to comparable research from other countries(Koch et al. 2023). But this is good news, as the usual argument that non-voters are notoriously hard to reach for parties is therefore less valid for Croatia. Another recent case to look at in this regard is the 2023 Polish general election. It was accompanied by the highest voter turnout in Poland’s democratic period: 74.5%. A decisive factor was targeted campaigns involving civil society organisations, but also social media activists and others to reach young people(Brändle and Szelewa-Kropiwnicka 2024). While these campaigns may not be easy to replicate in other countries, they demonstrate the importance of building alliances with civil society organisations when running a non-voter campaign. The unique advantage of a non-voter campaign that specifically targets districts or groups with traditionally high abstention rates is that you have the element of being the only offer to these voters. This is the lesson behind surprise campaigns such as the 2023 Salzburg regional election in Austria, where a fringe party surprised pollsters by winning 12% of the vote – many of whom were former voters, because they explicitly campaigned in districts with previously low turnout. Their voter testimonials along the lines of:‘This is the first time a politician is talking to us’(Jennewein 2023). In a nutshell: Non-voter campaigns are hard to pull off. It requires candidates who are willing to put themselves out there, who have a unique profile that resonates with people. They also need a lot of help and cooperation from actors outside the immediate party sphere. But: Non-voter campaigns have the potential to win votes. In the case of Croatia, this potential also exists, as we found that Croatian non-voters, despite being less informed and scoring lower on political efficacy scales(Henjak 2017; Henjak and Č ular forthcoming), are no less supportive of democracy than regular voters. This means that they have not yet given up on democracy, despite being alienated from political representation. At the same time, we know that non-voters in Croatia tend to have pronounced populist attitudes(Raos 2020; VuksanĆ usa and Šalaj 2024), suggesting that they may indeed be populist democrats(Zaslove and Meijers 2023). 14

Literature LITERATURE Aichholzer, J.,& Kritzinger, S.(2020). Voting at 16 in Practice: A Review of the Austrian Case. In J. Eichhorn& J. Bergh(Eds.), Lowering the Voting Age to 16(pp. 81–101). Cham: Palgrave Macmillan. Bastedo, H.(2015). Not‘one of us’: Understanding how non-engaged youth feel about politics and political leadership. Journal of Youth Studies, 18(5), 649–665. Bergh, J.,& Eichhorn, J.(2020). Introduction. In J. Eichhorn& J. Bergh (Eds.), Lowering the Voting Age to 16(pp. 1–12). Cham: Palgrave Macmillan. Blais, A.,& Daoust, J-F.(2020). The Motivation to Vote: Explaining Electoral Participation. Vancouver, Toronto: UBC Press. Bödeker, S.(2012). Soziale Ungleichheit und politische Partizipation in Deutschland. Grenzen politischer Gleichheit in der Bürgergesellschaft. Bovan, K.,& Baketa, N.(2022). Stabilnost i/ili promjene? Povjerenje u institucije u Hrvatskoj od 1999. do 2020. Revija za sociologiju, 52(1), 31–60. Bovan, K., Širini ć , D.,& Raos, V.(Eds.).(forthcoming). Hrvatski bira č i: 30 godina politi č kog ponašanja i mišljenja. Zagreb: Fakultet politi č kih znanosti Sveu č ilišta u Zagrebu. Brändle, M.,& Szelewa-Kropiwnicka, B.(Eds.).(2024). The defeat of populists in Poland: Behind the scenes of the 2023 election campaign. Warszawa: Friedrich-Ebert-Stiftung, Representative Office in Poland. Dassonneville, R.(2017). Age and Voting. In K. Arzheimer, J. Evans,& M.S. Lewis-Beck(Eds.), The SAGE Handbook of Electoral Behavior(2nd edition)(pp. 137–158). London: SAGE Publications. Dinas, E. (2012). The Formation of Voting Habits, Journal of Elections, Public Opinion& Parties, 22(4)m 431–456. Dinas, E., Valentim, V., Broberg, N.,& Franklin M. N. (2024). Early voting experiences and habit formation. Political Science Research and Methods, 12(1), 195–206. Dorey, P.(2017). Jeremy Corbyn confounds his critics: explaining the Labour party’s remarkable resurgence in the 2017 election. British Politics, 12, 308–334. Early, A. S., Smith, E. L.,& Neupert, S. D.(2023). Age, education, and political involvement differences in daily election-related stress. Current Psychology, 42(25), 21341–21350. Eichhorn, J. (2018). Votes At 16: New Insights from Scotland on Enfranchisement. Pariliamentary Affairs, 71(2), 365-391. Eichhorn, J.& Bergh, J. (2021). Lowering the Voting Age to 16 in Practice: Processes and Outcomes compared. Parliamentary Affairs, 74(3), 507–521. Elsässer, L., Hense, S.,& Schäfer, A.(2017).“Dem Deutschen Volke“? Die ungleiche Responsivität des Bundestags. Zeitschrift für Politikwissen­ schaft, 27(2), 161–180. Elsässer, L., Schäfer, A.& Wenker J.(2022). Ungleiche Demokratien. Wer geht(nicht) wählen?. Friedrich-Ebert-Stiftung – Demokratie der Zukunft. Wien. Eurobarometer(2023). European Opinion Poll Survey – December 23. https://europa.eu/eurobarometer/surveys/detail/2966 Foa, R., Klassen, A., Wenger, D., Rand, A.,& Slade, M.(2020). Youth and satisfaction with democracy: Reversing the democratic disconnect? Eurobarometer 2023 December. ECFR Article. https://ecfr.eu/publication/ a-crisis-of-ones-own-the-politics-of-trauma-in-europes-election-year/ Franc, R., Perasovi ć , B.,& Mustapi ć , M.(2018). Youth, history and a crisis of democracy? Perspectives from Croatia. Understanding Youth Participation Across Europe: From Survey to Ethnography, 227–263. Franklin, M.(2004). Voter Turnout and the Dynamics of Electoral Competition in Established Democracies since 1945. Cambridge: Cambridge University Press. Franklin, M.,& Hobolt, S. B.(2011). The legacy of lethargy: how elections to the European parliament depress turnout. Electoral Studies, 30(1), 67–76. Franklin, M. N.(2020). Consequences of Lowering the Voting Age to 16: Lessons from Comparative Research. In J. Eichhorn& J. Bergh(Eds.), Lowering the Voting Age to 16(pp. 13–41). Cham: Palgrave Macmillan. García del Horno, R., Rico, G.,& Hernández, E.(2023). Do they feel like they don’t matter? The rural-urban divide in external political efficacy. West European Politics, online first. Gilens, M. (2012). Affluence and influence: Economic inequality and political power in America. Princeton University Press. Grasso, M. T.(2014). Age, period and cohort analysis in a comparative context: Political generations and political participation repertoires in Western Europe. Electoral Studies, 33(1), 63–76. Grasso, M. T., Farrall, S., Gray, E., Hay, C.,& Jennings, W.(2019). Thatcher’s Children, Blair’s Babies, Political Socialization and Trickle-down Value Change: An Age, Period and Cohort Analysis. British Journal of Political Science, 49(1), 17–36. Grönlund, K.,& Setälä, M.(2007). Political Trust, Satisfaction and Voter Turnout. Comparative European Politics, 5(4), 400–422. Hadjar, A.,& Beck, M.(2010). Who does not participate in elections in Europe and why is this? A multilevel analysis of social mechanisms behind non-voting. European Societies, 12(4), 521–542. Harring et al.(2023). Youth Study: Personal optimism, national pessimism, trust in Europe. Friedrich-Ebert-Stiftung. Henjak, A.(2017). Lojalnost, glas ili izlazak: izborna participacija i potpora novim strankama u Hrvatskoj. Anali Hrvatskog politološkog društva, 14(1), 79–103. Henjak, A.,& Vuksan-Ćusa, B. (2019). Interesi ili nešto drugo? Ekonomski stavovi i njihova utemeljenost u društvenoj strukturi u Hrvatskoj. Revija za sociologiju, 49(1), 37-60. Henjak, A.,& Č ular, G.(forthcoming). Stari rascjepi i nove stranke: analiza bira č kog ponašanja na izborima 2020. u Hrvatskoj. Politi č ka misao. Ilišin, V.(2017). Mladi i politika: trendovi(dis)kontinuiteta. In V. Ilišin& V. Spaji ć Vrkaš(Eds.), Generacija osuje ć enih. Mladi u Hrvatskoj na po č etku 21. stolje ć a(pp. 230–231). Zagreb: Institut za društvena istraživanja u Zagrebu. Jennewein, M.(2023). Amid nothing going right, Austrians lean far-left. IPS Journal. Karp, J. A.,& Brockington, D.(2005). Social Desirability and Response Validity: A Comparative Analysis of Overreporting Voter Turnout in Five Countries. The Journal of Politics, 67(3), 825–840. Koch, C. M., Meléndez, C.,& Rovira Kaltwasser, C.(2023). Mainstream voters, non-voters and populist voters: what sets them apart?. Political Studies, 71(3), 893–913. Kostelka, F., Blais, A.& Gidengil, E. (2019). Has the gender gap in voter turnout really disappeared? West European Politics, 42(3), 437–463. Kostelka, F.,& Blais A.(2021). The Generational and Institutional Sources of the Global Decline in Voter Turnout. World Politics, 73(4), 629–667. Lamza Posavec, V.(2004). Apstinencija na izborima za Hrvatski sabor 2003. godine: tko i zašto nije glasovao?. Društvena istraživanja, 71(3), 383–404. 15

FES UNEQUAL DEMOCRACIES – WHO DOES(NOT) VOTE IN CROATIA? Lichtin, F., van der Brug, W.,& Rekker, R.(2023). Generational replacement and Green party support in Western Europe. Electoral Studies, 83(1), online first. Linek, L.,& Petrúšek, I.(2016). What’s past is prologue, or is it? Generational effects on voter turnout in post-communist countries, 1990–2013. Electoral Studies, 42, 78–90. Lisi, M., Quaranta, M., Real-Dato, J.,& Tsatsanis, E.(2021). Assessing the Impact of Age, Cohort and Period Effects on Partisanship and Support for Mainstream Parties: Evidence from Southern Europe. South European Society and Politics, 26(2), 239–270. McClean, C. T.,& Ono, Y.(2022). Too Young to Run? Voter Evaluations of the Age of Candidates. Working paper. Mitteregger, R.(2024). Socialized with“old cleavages” or“new dimensions”: An Age-Period-Cohort analysis on electoral support in Western European multiparty systems(1949–2021). Electoral Studies, 87, 102744. Nadeau, R., Lewis-Beck, M.S.,& Foucault, M.(2019). Wealth and Voter Turnout: Investigating Twenty-Eight Democracies. Polity, 51(2), 261–287. Neundorf, A.,& Niemi, R.G.(2014). Beyond political socialization: New approaches to age, period, cohort analysis. Electoral Studies, 33, 1–6. Norris, P.(1999). Introduction: The Growth of Critical Citizens? In P. Norris(Ed.), Critical Citizens. Global Support for Democratic Government(pp. 1–27). Oxford: Oxford University Press. Norris, P.,& Inglehart, R.(2019). Cultural Backlash.Trump, Brexit, and Authoritarian Populism. Cambridge: Cambridge University Press. Open Society Barometer(2023). Can Democracy Deliver? https:// www.opensocietyfoundations.org/focus/open-society-barometer Oprea, A., Martin, L.,& Brennan. G.H.(2023). Moving toward the Median: Compulsory Voting and Political Polarization. American Political Science Review, online first, pp. 1–15. Prior, M.(2010). You’ve Either Got It or You Don’t? The Stability of Political Interest over the Life Cycle. The Journal of Politics, 72(3), 747–766. Raos, V.(2020). Struktura rascjepa i parlamentarni izbori u Hrvatskoj 2020. u doba pandemije. Anali Hrvatskog politološkog društva, 17(1), 7–30. Raos, V.& Zakošek, N. (forthcoming). Autoritarnost. In: Bovan, K., Širinić, D and Raos, V.(eds.) Hrvatski birači: 30 godina političkog ponašanja i mišljenja. Zagreb: Fakultet političkih znanosti Sveučilišta u Zagrebu. Rosenqvist, O.(2020). Rising to the Occasion? Youth Political Knowledge and the Voting Age. British Journal of Political Science, 50(2), 781–792. Schäfer, Armin (2012):»Beeinflusst die sinkende Wahlbeteiligung das Wahlergebnis? Eine Analyse kleinräumiger Wahldaten in deutschen Großstädten«, in: Politische Vierteljahresschrift 53(2), S. 240–264. Schäfer, A.,& Roßteutscher, S.(2015). Räumliche Unterschiede der Wahlbeteiligung bei der Bundestagswahl 2013: Die soziale Topografie der Nichtwahl. Die Bundestagswahl 2013: Analysen der Wahl-, Parteien-, Kommunikations-und Regierungsforschung, 99–118. Schäfer, A.,& Streeck, W.(2013). Introduction: Politics in the Age of Austerity. In A. Schäfer& W. Streeck(Eds.), Politics in the Age of Austerity(pp. 1–25). Cambridge: Polity Press. Scervini, F.,& Segatti, P.(2012). Education, inequality and electoral participation. Research in Social Stratification and Mobility, 30(4), 403–413. Schnetzer, M.(2021). Wie Ungleichheit der Demokratie schadet. Wirtschaft und Gesellschaft, 47(3), 297–307. Smets, K.,& van Ham, C.(2013). The embarrassment of riches? A meta-analysis of individual-level research on voter turnout. Electoral Studies, 32(2), 344–359. Smets, K.,& Neundorf, A.(2014). The hierarchies of age-periodcohort research: Political context and the development of generational turnout patterns. Electoral Studies, 33, 41–51. Šalaj, B.(2023). Prema inkluzivnoj demokraciji: spuštanje dobne granice za politi č ko sudjelovanje na 16 godina. Zagreb: GONG. Širini ć , D.,& Dolenec, D.(forthcoming). Izborna participacija. In K. Bovan, D. Širini ć ,& V. Raos(Eds.), Hrvatski bira č i: 30 godina politi č kog ponašanja i mišljenja. Zagreb: Fakultet politi č kih znanosti Sveu č ilišta u Zagrebu. Toma, I.(2023). OKRUGLI STOL. GONG:‘Za glasanje na izborima dobnu granicu treba spustiti s 18 na 16 godina‘. Jutarnji list, September 27, 2023. Van der Brug, W.,& Rekker, R.(2021). Dealignment, realignment and generational differences in The Netherlands. West European Politics, 44(4), 776–801. Verba, S.,& Nie, N.H.(1972). Participation in America: Political Democracy and Social Equality. New York, Evanston, San Francisco, London: Harper& Row. VuksanĆ usa, B.,& Šalaj, B.(2024). Storm in a Teacup: Populist Attitudes and Voting for Populists in Croatia. Anali Hrvatskog politološkog društva, 21(1), 1–28. Wagner, M., Johann, D.,& Kritzinger, S.(2012). Voting at 16: Turnout and the quality of vote choice. Electoral Studies, 31(2), 372–383. Wattenberg, M. P. (2015). Is Voting for Young People? Completely Updated Through the 2018 Election(5th edition). New York, London: Routledge. Wenker, J.(2024). Who does(not) vote?. Friedrich-Ebert-Stiftung – Democracy of the Future. 16

Imprint ABOUT THE AUTHORS IMPRINT Bartul VuksanĆ usa is a PhD Candidate in the Political Science at the Universitat Autònoma de Barcelona(UAB) and Assistant and Lecturer at Faculty of Political Science at University of Zagreb. He holds a BA from University of Zagreb and MA’s from University of Zagreb and Central European University in Political Science. His research deals with political attitudes and voting behavior. Michael Jennewein is a Researcher for Democracy& Economics at Friedrich-Ebert-Stiftung Democracy of the Future in Vienna mainly working on inequality& democracy and right-wing populism. He holds a BA in Philosophy& Economics from University of Bayreuth) and a MA in Technology Governance& Digital Transformation from Tallinn University of Technology. 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 978-3-98628-545-6 © 2024 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 and other democratic processes hurts our democracies. In country reports we provide country-specific analyses about the state of particular national contexts with the aim to develop and discuss political recommendations for decision-makers. More information at: https://democracy.fes.de/topics/inequality-democracy 17

WHO DOES(NOT) VOTE IN CROATIA? Why should Croatians care Who are the Croatian non-­ What should be done? about turnout? voters? Fewer than one in two Croatians go vote in national elections – the forth lowest turnout in Europe. Persistent voting abstention undermines Croatian democratic reslience as more and more citizens become estranged from political processes. In line with general European trends, the Croatian young, less educated and working class are likely to abstain from voting. Croatian non-voters have low political interest but at the same time – in contrast to other countries – are not disproportionally dissatisfied with democracy. Changing Croatian non-voter behaviour requires tackling persistent structural problems. As one incremental measure we recommend discussing to lower the voting age to 16. For parties, we advocate taking lessons from election campaigns across Europe that specifically targeted non-voters.