Duty to Vote and Political Support in Asia

Elections are the cornerstone of representative democracy. Through their votes, citizens select their representatives in the political system. This, in principle, guarantees the legitimacy of democratic regimes, which could be threatened if many people decided not to participate in the electoral process. Electoral participation is affected by a wide array of factors, not the least of which is the belief that voting is not only a right but also a moral duty. In fact, whether one does or does not feel that she has a duty to cast a ballot is probably the most powerful predictor of the decision to vote or abstain (Blais, 2000; Campbell et al., 1960; Clarke, Sanders, Stewart, & Whiteley, 2004).

Surprisingly, the sources of this strong determinant of turnout have been under researched. Some scholars seem to believe that all citizens feel dutiful (Gerber, Green, & Larimer, 2008), but it makes more sense to assume that not everyone subscribes to the same extent to this public norm. In fact, it is likely that some completely disagree with the view that citizens have a moral obligation to vote in elections and that others are ambivalent, as not all citizens share the same views about citizenship and what is expected from the good citizen (Dalton, 2007). The question thus becomes: What are the psychological orientations that predispose citizens to believe that they do (or do not) have a duty to vote?

The moral, communitarian, and political nature of the duty to vote, as well as its main referent—elections—have led several authors to establish links between this belief and attitudes toward democracy (Downs, 1957) and one’s country (Raney & Berdhal, 2009; Usher, 2011). We hereby posit that these attitudes are in fact antecedents of the duty to vote. More precisely, we hypothesize that the belief that voting is a duty will be stronger among supporters of democracy and/or of their political community. We also expect the duty to vote to be stronger in fully free democracies and ethnically homogeneous countries. To the extent that free elections are associated with the ideal of representative democracy, it should be easier for citizens living in countries where elections are not free to conclude that they do not have a moral obligation to vote. Likewise, citizens in heterogeneous countries are harder to integrate into the larger national community, which will weaken their commitment to collective norms (Cardenas & Carpenter, 2002; Messick & Mackie, 1989; Putnam, 2001).

We also expect these contextual factors to interact with individual attitudes. On the one hand, support for democracy will be more strongly associated with the moral obligation to vote in the presence of more democratic institutions. Under such a context, individual beliefs, and institutional settings are aligned and consistent. Furthermore, support for democracy in full democracies may entail more demanding views of the citizen, whereas democracy supporters in dictatorships are bound to be more concerned with the conquest of basic civil rights and liberties. In turn, the effects of supporting (or not) one’s political community will be amplified in heterogeneous countries, which have a harder time generating and enforcing common moral norms. Those who support their political community even when it is ethnically and culturally divided will internalize and support the social norms that keep them together, especially those regarding the political institutions that unify the country. Conversely, those not feeling positive about their country will turn to their own ethnic community as a reference group when that country is not homogeneous, and will not be prone to adhere to “national” social norms such as the duty to vote.

To test these expectations, we turn to the most diverse world area in terms of political regimes and ethnicity: Asia. The goal is to examine a diverse sample of individuals and countries for which data were available to guarantee maximum variability of the phenomenon under study and the plausible explanatory factors. This, of course, has implications for comparative studies of political culture. A research strand suggests the incompatibility between “Asian values” and some aspects of the civic political culture (Huntington, 1993; Lee Kuan, 1992, 1998), arguing that Confucianism weakens individual values by enforcing order and respect for hierarchy. A contrary school of thought states that Western values have put too much emphasis on individualism, which is not necessarily positive for democracy (Chan, 1995; Robison, 1996). These critical voices have also questioned whether Confucianism implies blind allegiance to the state, referring to Confucian “modern” values such as public-spiritedness and duty consciousness (Harrison & Huntington, 2000, p. 264). Finally, these authors have also highlighted the richness and complexity of the Asian continent when it comes to values and political cultures, noting that the Confucian tradition runs strong only in East Asia (Sen, 1999). This research aims to shed some light on one aspect of the Asian political culture through an examination of the duty to vote. In the process, we are careful not to confine ourselves to countries located in East Asia by including South, Southeast, West, and Central Asian countries. We wish to determine whether the duty to vote is widespread in Asia, whether it is stronger in some countries than in others, and whether it is connected to political support and contextual-level variables.

We test the effects of these individual and contextual factors—and their interactions—by means of two Asian Barometer Surveys (ABS) conducted in 2004 and 2005, gathering data in 21 Asian countries. We first describe the data and then run a series of multilevel estimations that account for the structure of the data. We confirm the effects of the individual and contextual factors. Both support for political community and support for democracy help explain the individual levels of sense of duty, but the first factor seems to be more important. As for the contextual factors, the degree of democracy indeed affects the strength of the duty to vote, but ethnic fractionalization does not seem to play any role. Finally, cross-level interactions are significant: The impact of support for democracy is stronger in more democratic countries, and the effect of support for political community is more pronounced in ethnically fractionalized countries.

Support for Democracy and Political Community as Core Beliefs Shaping the Duty to Vote

Voting out of duty reflects the moral conviction that voting is the right thing to do and that abstaining is bad. In other words, it is a belief that reflects adherence to a social norm about what the good citizen should do (Blais, 2000; Coleman, 1990; Knack & Kropf, 1998; Mueller, 1989; Uhlaner, 1986; Zuckerman & Kotler-Berkowitz, 1998). In the words of Dalton, civic duty entails support for the “traditional” responsibilities of a “good” citizen, which positively relates to a high propensity to vote (Dalton, 2008). This relates to the liberal-communitarian debate on the implications of citizenship. While the first emphasize the rights associated with citizenship, the communitarian approach posits that citizenship requires the fulfillment of a series of responsibilities toward the community (Delanty, 2000; Faulks, 2000). One of these moral obligations would be voting.

But where does the social norm about a “duty to vote” stem from? Although the bases of the beliefs about voting and elections are probably related to some personality traits (Blais and Labbé-St\. Vincent, 2011) and parental practices (Plutzer, 2002; Verba et al., 1995), this study focuses on two basic attitudes that should contribute to the internalization of the belief that citizens should fulfill their duty by voting in elections: Support for democracy and support for the political community.

Indeed, Downs (1957) argues that from a strictly utilitarian perspective, the rational citizen will be inclined to abstain, because of the tiny probability that her vote will decide the outcome of the election, and then suggests that this rational citizen may conclude that she should vote to maintain democracy. According to such a perspective, the act of voting is an expression of one’s commitment to the democratic system. Support for the democratic ideal would thus nourish the belief that voting is a duty. Similarly, several authors have suggested that civic duty is grounded in how one feels about her country. Raney and Berdhal (2009) found a relationship between identification with the Canadian territorial communities (Canada, provinces, and localities) and the duty to vote (2009). But this connection with identity should hold more generally. Attachment to the nation remains a powerful sentiment (Anderson, 1983; Billig, 1995), and such attachment may well elicit a sense of duty to contribute to the welfare of the collectivity (Walzer, 1990). As a result, voting would be a form of compliance with the state, whereby citizens ritualize their fulfillment of moral responsibilities towards the political community (Wuthnow, 1982), making the act of voting a “patriotic gesture” (Usher, 2011).

The first set of arguments for postulating a causal link between attitudes about democracy, feelings about one’s country, and the duty to vote are grounded in theories of political support. Easton’s definition of support states that A supports B when A “orients himself favorably toward B. B may be a person or a group; it may be a goal, idea, or institution” (Easton, 1965, p. 159). Easton distinguished between specific and diffuse support. One feature of diffuse political support is “commitment,” which can be regarded as a reservoir of goodwill toward existing institutions, one of them being the use of elections to select the decision-makers. Another dimension is legitimacy, which refers to the extent to which the political system conforms to an individual’s sense of what is right and good, which involves a moral aspect similar to the one present in duty. In the same spirit, Dennis relates civic duty to diffuse support for the regime, defining duty as the citizen’s feeling of obligation “to contribute his own resources of time and effort even when particular elections are anticipated to be unfavorable or trivial to his own interests” (1970, p. 63). Furthermore, research on political socialization suggests that political support develops early in life (Dennis, 1973; Ehman, 1980; Van Deth, Abendschön & Vollmar, 2011), which makes it a plausible antecedent of the duty to vote.

As for the objects that arouse political support, Fuchs (1993) distinguishes between support toward one’s political community, the political regime, and political authorities. Klingemann (1999) adds support for democracy as a form of government and support for the performance of democratic regimes. Norris (1999) differentiates support for the political community, the regime principles, the performance of the regime, democratic institutions, and, finally, candidates and political elites. Dalton (2004) outlines five dimensions of political support, oriented respectively toward the community, democracy, norms, institutions, and authorities. The most common manifestations of political support are, therefore, support toward the democratic regime and support for the political community, from which we can infer that they are the two main objects of political support and the foundations of more specific beliefs.

Works on political support are often concerned with its consequences on democratic stability and individual political behavior (Dalton, 2004; Muller, 1977; Muller, Jukam, & Seligson, 1982). In that vein, some research shows a relationship between declining support for democracy and declining turnout (Seligson 2002). Similarly, citizens who identify with their political community register and vote more than those who do not (Booth & Seligson, 2009). We would suggest that these relationships between political support and several aspects of electoral participation are in fact mediated by the duty to vote.

Further arguments in favor of the idea that attitudes about democracy and the political community shape sense of civic duty can be found in theories about core values. The structuring role of core values (or beliefs) has been documented in classical studies (Bem, 1970; Feldman, 1988). According to this literature, some values are widely shared and endorsed in societies as cornerstones of their political culture, and they are reinforced by political institutions and politicians’ discourse and rhetoric (Devine, 1972; McClosky & Zaller, 1984). While “values” are “core normative predispositions or abstract visions of the ‘good’” (Marietta & Barker, 2007, p. 50), “core political values” are typically inferred from the presence of a consensus about how government and society should function. According to Schwartz, Caprara and Vecchione (2010), some of the core values revolve around equality and civil liberties, the two most important democratic principles. These authors also mention “blind patriotism,” the “unquestioning attachment to and intolerance of criticism of one’s country” (p. 424). We hypothesize that attitudes toward democracy and one’s political community, construed either as aspects of political support or as core values, shape the belief that voting is a moral obligation, in the sense that those more supportive of democracy and of their political community develop a stronger sense of duty to vote. Therefore, our specific hypotheses on the effects of these two attitudes on the duty to vote are the following:

H1. Stronger support for democracy leads to higher levels of duty to vote.

H2. Those who are more supportive of their political community exhibit higher levels of the duty to vote.

Sense of civic duty is presumably also influenced by contextual-level variables, as the rich literature on the contextual factors that affect turnout suggests, including the socioeconomic context, rules and institutions, as well as specific characteristics of a particular election (Blais & Achen, 2010; Geys, 2006). As duty is an antecedent of turnout, some of those features should shape and foster in the first place the social norm that voting is good and desirable, ultimately affecting turnout rates.

The first of these contextual aspects to be considered is the level of democracy itself. Previous research has found that attitudes beneficial for democracy, such as political efficacy, are positively affected by the degree of democracy (Dalton, 2004). It makes sense, therefore, to expect that sense of civic duty will be higher in more democratic countries. Indeed, democracies are more successful in creating the appropriate context for interpersonal trust, trust in democratic institutions and other aspects of civic culture (Almond, 1980; Muller & Seligson, 1994). Moreover, democracies make great efforts to convey social norms consistent with the democratic principle, such as the necessity of voting in elections (Wright, 1976).

Finally, as the duty to vote is an expression of a social norm, it should be stronger where there is widespread identification with the community (Knack, 1992; Raney & Berdahl, 2009). Campbell’s work (2006) on the causes of participation highlights the crucial role of the sense of duty for turnout. He also finds that politically homogeneous communities are more successful in conveying civic norms such as sense of duty. Members of homogenous communities are more likely to agree on what is right and wrong and to recognize the legitimacy of those who enforce social norms. Other works find similar evidence on the success of homogeneous communities in fostering cooperation and the punishment of those who do not abide by the group rules (Cardenas & Carpenter, 2002; Putnam, 2001). From another point of view, duties toward the state (i.e., paying taxes, voting in elections) will more likely appear where the state represents the individual’s national identity, which enforces thinking in terms of political duties (Hur, 2015). Therefore:

H3. More democratic countries foster higher levels of duty to vote.

H4. More diverse countries are characterized by lower levels of duty to vote.

Last, we expect attitudes toward democracy and the political community to operate differently as a function of the aforementioned contextual factors. For a start, we expect attitudes toward democracy to have a greater impact on duty where democracy is stronger. Consider a democratic supporter in a nondemocratic regime. Her ideal of democracy will be more in line with pillar democratic values, such as freedom of speech and equality, than with civic duties. The perceived urgency of gaining or regaining rights and liberties is bound to trump all other considerations. In countries where there are no elections or these are not democratic, those longing for a democratic regime should consider voting as a right, and not as a duty, either because it is not possible to cast a vote or because it may contribute to legitimating a nondemocratic regime.

Second, we expect support for one’s political community to have a stronger effect on duty where the political community is divided. Indeed, there is a lively debate on the effects of diversity on civic outcomes. Putnam’s (2007) conclusion is illustrative of the state of the debate. While in the short term ethnic diversity may endanger cooperation and lowers altruism, societies that have successfully overcome fragmentation have generated more encompassing identities and new forms of social solidarity. These comprehensive identities should in turn positively impact civic outcomes. As a consequence, being proud or not of one’s country should make a bigger difference in those countries where the collective identity is more ambiguous, that is, in more diverse settings. From another perspective, take an individual who is proud of being a national of her country (the most usual indicator of support for one’s political community) in a divided context. This citizen supports the political community derived from the institutional setting and transcends her immediate ethnic group of reference. This entails being proud of not only her country but also of the institutions that keep the country together, including elections and the act of voting. Therefore:

H5. Individual support for democracy is more strongly related to the duty to vote where institutional norms point in the same direction (more democratic countries).

H6. Individual support for the political community is more strongly related to the duty to vote in divided communities.

The next section addresses the measures of the duty to vote and support for democracy and the political community, the data, and the statistical tests used to test our hypotheses.

Research Design

We use Waves 2 and 3 of the 2004 and 2005 ABS. Both surveys were merged to get information about a total of 28 countries, from which we keep 24 in which the question about the duty to vote was asked. The wording of this question was: “I am going to read out some statements about society and politics. Please indicate how much you agree or disagree with each statement: Citizens have a duty to vote in elections.” Respondents had to choose between the following options: Strongly agree, agree, neither agree nor disagree, disagree, strongly disagree. These options were coded so as to range between 0 (strongly disagree) to 1 (strongly agree).

As for the measure of support for democracy, the question stated “I'm going to describe various types of political systems. Please indicate for each system whether you think it would be very good, fairly good or bad for this country: A democratic political system.” The options were also given values from 0 to 1, from minimum to maximum support for democracy.¹ Finally, support for the political community was measured by means of the classical question used by most scholars concerned with this dimension of political culture (Klingemann, 1999): “How proud are you of being a national of your country?” The response options included: Very proud, somewhat proud, not really proud, and not proud at all. We take national pride to be an affective indicator of support for the political community.²

We have included a number of relevant controls in our analyses. Sex, age, and education are expected to be related to duty; females, the elderly and the more educated tend to be more dutiful (Blais, 2000). We have also included an indicator of personal wealth using individual income. As the question tapping wealth was different in every country, we have created a four-category variable, which ranges from 0 to 1. The minimum value (0) gathers all those who have an income value inferior to 1 SD under the mean for her country. The next category gathers individuals with an income value between the mean and 1 SD under the average income in her country. The third category is for those with an income value higher than the average mean in their country but <1 SD higher. Finally, the maximum value (1) is for individuals with an income that is higher than 1 SD above the average value for the country.

Moreover, we have included additional variables tapping the moral and communitarian dispositions of the individual as well as other regime perceptions. Given that the duty to vote is a social norm and that adherence to norms is more likely among those who identify with their community (Knack 1992; Raney & Berdahl, 2009), we have included a variable tapping this communitarian identification. The ABS asked respondents “of the following lifestyle aspects of life circumstances, please select five that are important to you,” one of them being “contributing to your local community or to society.” We consider that selecting this item among the five most important aspects signals a community-oriented individual who is more likely to embrace citizenship norms such as the one we are concerned with. Also, as the language of duties is germane to religious thinking (Macaluso & Wanat 1979), and both religiosity and duty belong to the moral domain and the sub-domain of authority (Graham et al., 2011). Hence, it makes sense that religious individuals will be more prone to think about voting in terms of duty. For this reason, we have included an indicator of religiosity that takes the value 1 if the individual considers herself a religious person, and zero if otherwise.³ We have finally included two attitudes that are related to perceptions of the political regime. The first is an indicator of trust in parliament.⁴ The second is a variable tapping support for authoritarian regimes. Along with support for democracy, a battery of questions asked people what they thought of a military government and governance by a powerful leader. We have built a “support for authoritarianism” indicator that takes the value of 1 if the individual was warm (fairly good) or clearly supportive (very good) with respect to any of these two alternatives to democracy.⁵ The variable takes the value of 0 only if the individual considered both alternatives “bad,” which approximately one third of our sample did.

All the independent variables measured at the individual level (including socio-demographic controls such as age, gender, and education) have been rescaled to range from 0 to 1. This entails that the coefficients in the multivariate estimations should be read as the effect of going from its minimum to its maximum value.

Contextual data regarding democracy and diversity have been obtained from the Quality of Government database (Dahlberg, Holmberg, Rothstein, Hartmann, & Svensson, 2015), version of January 2015. More precisely, the collapsed version of the Freedom House democratization index (recoded as 0 = not free, 0.5 = partly free, 1 = free) and the Alesina, Devleeschauwer, Easterly, Kurlat, & Wacziarg (2003) ethnic fractionalization index have been used to measure the two contextual factors. After merging the two ABS with these variables from the Quality of Government database, there are 24 countries for which we have information on their citizens’ duty to vote, and 21 countries for which we have complete information at the individual and the aggregate level, our final sample.

The landscape is indeed a diverse one. We have seven ex-communist countries, from which five emerged from the breakup of the Soviet Union in 1991 (Uzbekistan, Kyrgyzstan, Turkmenistan, Tajikistan, Kazakhstan). There are several ex-British colonies (Singapore, India, Pakistan, Sri Lanka, Bangladesh, Bhutan, Malaysia), one ex-US and Spanish colony (Philippines) and a former Dutch colony (Indonesia) along with democracies that have never been under colonial control (Japan, Korea, Thailand). There are three monarchies (Cambodia, Nepal, and Bhutan) and one military regime (Pakistan). Furthermore, most important for our purposes, there is a lot of variation with regard the degree of democracy.

As Morlino (2011) points out, Asia challenges the Western liberal democratic model with some examples of what he calls hybrid regimes, like Singapore and Malaysia in our sample. According to this scholar, another main characteristic of the region is its ability to quickly evolve. Except from Japan’s and Philippines’ relatively early experiences with democracy, most of the democratic countries we consider are young regimes. Moreover, relatively recent events such as the coups in Thailand or the coup attempts in Nepal, Philippines, and Bangladesh have put democracy in a delicate situation (Morlino, Dressel, & Pelizzo, 2011).

As for multiculturalism, South Korea is a good example of a cohesive, homogeneous nation-state, while in Afghanistan, Pakistan, or Indonesia, national identity is challenged by a significant portion of citizens. In some of the more diverse countries in our sample (Thailand, Pakistan, and Sri Lanka), the state has recently used repression, violence, and the suppression of human rights to handle problems with ethnic or religious minorities. The region has also witnessed several anti-Western political doctrines that considered democratic reforms un-Asian, such as the five “shared values” in Singapore (Quah, 1990) or the official state ideology in Indonesia, both emphasizing the importance of family, putting the interests of society before the individual, and imposing consensus and a “virulent organic statism” (Robison, 1996, p. 316).

The diversity of the sample as well as the nature of the hypotheses concerning both individual-level and country-level variables suggest a multilevel approach. When handling pooled cross-sectional surveys, the structure of the errors may reproduce similarities within countries, underestimating the standard errors of the macro-level variables. This can only be fixed by clustering errors by countries or applying multilevel analysis. Thus, hierarchical linear modelling is recommended, given the measure of the dependent variable. This allows us to introduce country-level variables as moderators of individual-level factors (H5 and H6).

Findings

Figure 1 shows the mean value of duty in each of the 21 countries. These values range from 0.99 (Bangladesh) to 0.67 (Turkmenistan), which can be read in terms of the percentage of citizens believing that voting is a duty (from 99% to 67%). South Asian countries score higher, while East Asian countries (Japan, Korea, and Mongolia) come second in line. Southeast and Central (former Soviet Socialist Republics) Asian countries tend to score lower. If we look at the extreme cases, it is not surprising that Turkmenistan comes last. According to the 2013 Amnesty International report, Turkmenistan does not feature independent media, not a single independent human rights organization, does not recognize the right to dissent, and there is virtually no political opposition. Yet a national election was held in 2004—featuring a single party—amidst a climate of fear caused by the widespread use of torture against the enemies of the state (Amnesty International, 2013). On the top of it, Freedom House always includes Turkmenistan among the “Worst of the Worst”—the 10 countries with lowest ratings for both political rights and civil liberties.⁶ Another usual member of the “Worst of the Worst” is Uzbekistan, the antepenultimate country in our list of nations ranked from more to less dutiful. This suggests that the degree of democratic development is probably related to the strength of civic duty. Yet among the top three more dutiful countries we find a mix of partly free (Bangladesh) and free (Thailand, India) countries.

Despite these variations, all the countries exhibit relatively high averages of duty. It is difficult to compare Asian averages with other countries, but one of the last British Election Studies included an almost identical formulation of the duty question in 2010.⁷ If we recode the answers so as to range between 0 and 1 as we did for the Asian countries, the mean value is 0.81; indicating that all the Asian countries included in our sample with a British colonial past have a stronger sense of duty to vote than their former metropolis.

Table 1 shows the mean values of the individual-level variables tapping political support along with the values of the aggregate-level independent variables. The higher averages of support for democracy are found in Thailand, Malaysia, and Sri Lanka, while Turkmenistan, Tajikistan, and Uzbekistan come last. It is in Thailand, India, Philippines, Indonesia, and Bangladesh that citizens are most proud of their country, while Korea, Uzbekistan, and Japan are the countries with lower levels of support for the political community. As for aggregate variables, not free countries constitute a plurality (9 of 21), followed by free (6) and partly free countries (6). Ethnic fractionalization ranges from inexistent in Korea to very high in Afghanistan (0.77) and Indonesia (0.74). Again, the general portrait is one of diversity, which suits our purposes.

Our principal goal is to examine the sources of duty. But before proceeding to that analysis, it is useful to show that belief in the duty to vote has real consequences, that it affects behavior. The ABS included a question about voting frequency. More specifically, it asked “how often” the respondent voted in the national elections (always, most of the time, sometimes, rarely, never). The variable is pretty skewed, as 60% of the sample declares that they “always” vote in elections. We have recoded the variable, so it takes the value of 1 if the individual “always” votes in national elections and 0 otherwise. Next, we have performed a logistic estimation with “always voting” as the dummy dependent variable and duty, age, gender, education, and income as the independent variables.⁸ That logistic estimation indicates that duty has a significant and strong independent impact on the propensity to vote. Figure 2 displays the predicted probability of always voting as a function of sense of duty, controlling for socio-demographic characteristics.

When an individual has no sense of duty, her probability of always showing up at the poll station is 8%. If she has very high sense of duty, the probability increases to 64%. This strong relationship with turnout highlights the importance of understanding what attitudes shape the belief that citizens have a moral duty to vote in elections.

To unravel the relationship between political support and the duty to vote, we have estimated five multilevel linear models where the duty to vote is the dependent variable. The results are presented in Table 2. We start with a null model considering the 21 countries for which we have data on the dependent and all the independent variables. This model serves as a baseline to contrast the fit of subsequent models. We see that, according to the intraclass correlation, 11% of the variance of the duty to vote is owing to phenomena related to the second-level, that is, the characteristics of the countries.⁹ This confirms the suitability of a multilevel approach, as a widely accepted methodological rule of thumb is performing multilevel analyses only when the variation due to second-order factors exceeds 5% (Hox, 2010). All subsequent estimations consider that the data are structured in a hierarchical manner, ensuring that the effect of all individual and contextual variables will not be overstated as a consequence of the similarities of individuals within a country.

The next model includes all the individual-level predictors expected to exert an impact on the duty to vote. The descriptive for these variables are included in the Appendix. We see that gender does not have a significant effect on duty, but that age and education exhibit a positive, significant coefficient. Therefore, older and more educated citizens are more dutiful. Income, in turn, does not exert any significant effect. The two variables related to moral and communitarian attitudes (religiosity and willingness to contribute to the community) do have a positive significant effect on the duty to vote. As for the variables related to regime perceptions, we find mixed effects. The indicator of trust in parliament does have a pretty strong, positive significant effect. The more an individual trusts the parliament of her country, the more she will tend to think that it is her duty to vote in elections. But the variable tapping support for authoritarian regimes does not have a significant effect on duty.

What is more important for the purpose of our research, the two indicators tapping the two likely bases of the duty to vote (support for democracy and for the country) have positive, significant coefficients. As the two variables are coded so to range from 0 to 1, both coefficients can be read as the effect of going from its minimum value to its maximum value, and therefore can be compared. We see that the effect of being proud of one’s country is twice as big as that of being a supporter of democracy. This supports our first and second hypotheses, and suggests that sense of duty to vote is more strongly rooted in orientations toward one’s political community than in beliefs about democracy. We also see that the intraclass correlation decreases to 9.5%, which indicates that the inclusion of our individual-level variables explains about 2% of the initial contextual variation. We can also compare the two pseudo R-squared measures, which suggest that we have managed to explain about 5% of the individual-level variation and about 20% of the variation due to context.

The third model only considers contextual variables. The intraclass correlation slightly improves as compared with the previous model, and so does the second-level R-squared. The variable measuring democratic freedom indicates that the stronger a democracy is, the more dutiful its citizens are.¹⁰ Finally, ethnic fractionalization does not seem to play any role.¹¹

The fourth model includes both individual-level and country-level variables. The slopes of both support for political community and support for democracy have been freed to test the suitability of cross-level interactions with the country-level variables. The variance components of the slopes are significant (not shown), suggesting that the effects of these two aspects of political support do vary across countries. As for the effects of individual and contextual variables, the patterns are similar to those observed in Models 2 and 3, which confirms the robustness of our results. If any, the effect of religiosity, willingness to contribute to the individual’s community, and trust in parliament slightly decrease when considering contextual variables, suggesting that their effect probably varies across countries. This model is clearly better than the previous one, as both R-squares indicate a model improvement.

The last column specifies interactions between an individual-level variable and a contextual factor, as stated in the theoretical expectations. More specifically, the model considers the interaction between support for political community and ethnic fractionalization, on the one hand, and support for democracy and the degree of democracy on the other. According to our pseudo R-squared measures, this last model manages to explain about 6% of the variation in the belief that voting is a duty at the individual level, and about 32% of the phenomenon at the aggregate level. This means that the model is not better than the previous one, a conclusion also confirmed by the Akaike Information Criterion and Bayesian Information Criterion (BIC) values. These measures are often used to compare nonnested models, as they simultaneously consider error and parsimony (Singer and Willett, 2003). Their lower (hence better) values favor in this case simpler models; particularly the one featuring only contextual variables. Yet the purpose of the model displayed in the fifth column is to test the interplay between the two aspects of political support, democracy status, and ethnic diversity.

In this respect, the first interaction coefficient is positive and significant at p < .05. We posited that support for the political community should be more strongly related to the duty to vote in divided communities. This hypothesis is confirmed by the positive, significant coefficient of the interactive variable. The gap in civic duty between those who are and who are not proud is higher in more fragmented countries. When a country is divided, the authorities probably try to enforce public norms to keep the unity of the country, stressing common aspects, such as the rules and institutions that sustain the political system. As a result, being proud of a country that overcomes its inner divisions fosters support for the idea that good citizens should vote.

As for the second interaction, its effect is about half as strong as the one detected for the previous interaction. Yet, it is definitely positive and significant at p < . 05. Hence, we can conclude that support for democracy matters more in more democratic countries. Supporters of democracy are less prone to believe that they have a moral obligation to vote in a nondemocratic country, probably because they have in mind more urgent needs such as freedom of speech and human rights, and also because they understand that the act of voting does not have the same implications in their country than in a full democracy, even if they support the idea of democracy as a standard of progress.

Figures 3 and 4 display the effect of support for political community and for democracy, conditional on the ethnic fractionalization and democratic status of the country. They represent the interaction effects included in the last model of Table 2. The effect of political support increases with higher values of the two contextual variables. Although there are fewer observations in higher values of the contextual variables—hence the larger confidence intervals—the effect of political support in this range is significantly higher than in lower values of the contextual variables. While those who are not proud of their country have pretty much the same predicted level of duty than those who are extremely proud in homogeneous countries, the difference becomes much larger when we move to more divided countries (Figure 3). Similarly, the difference between strong and weak supporters of democracy is small among not free countries, but this gap increases in free countries, where being a democracy supporter is more strongly related to the duty to vote (Figure 4).

Conclusion

Although the belief that voting is a duty is widely considered as the most important predictor of individuals’ decision to vote, we know little about the attitudinal and moral bases of this belief, as well as about the institutional and contextual factors that sustain or weaken it. Given that the belief that voting is a duty entails feeling a moral obligation to perform a concrete act (casting a vote in an election), we assumed that its emergence would depend on the previous acquisition of two more abstract attitudes: Support for democracy and support for one’s political community. As theories on political support and on core political values suggest, these two attitudes are crucial for the development of political culture and are probably acquired earlier in life than the duty to vote.

We also posited that two contextual factors, the degree of democracy and ethnic fractionalization, would affect the emergence of duty. Finally, we expected these two factors to moderate the effect of the two aforementioned aspects of political support. To test these hypotheses, we used two ABS and the Quality of Government database, which allowed us to estimate multilevel models for the belief that voting is a duty. Our estimations confirmed the relationship between both aspects of political support and the duty to vote. What is more important, we found support for one’s political community to matter more than support for democracy. This has implications for the scholarly debate on citizenship norms as well as for politicians and practitioners.

Indeed, previous work on the mapping of the moral domain has highlighted the importance of in-group loyalty (Graham et al., 2011), which seems to be a crucial condition for the duty to vote to emerge. According to our findings, the belief that one has a moral obligation to vote stems more from attachment to one’s country than from the belief that democracy is a good thing, even controlling for individual’s religiosity, communitarian orientations, support for authoritarian regimes, or trust in parliament.

This leads us to the second implication, in this case for the design of more effective turnout campaigns or civic education contents. Stressing the importance of voting as a way to show love or pride for one’s country may have a greater effect on the development of the duty to vote than reminding people about the benefits of democracy. Perhaps the latter benefits are too abstract, whereas citizens have a clear idea about what patriotic pride is and means. Our data confirm previous findings about the low national pride of Japanese, who, like Germans, had to reinvent the bases and meaning of their national identification after their defeat in World War II (Klingemann, 1999). Japan is also one of the countries with lower levels of duty to vote and comes last in line if we only consider fully accomplished democracies. A possible conclusion is that Japan and other countries in a similar situation should reflect about the construction of what Klingemann (op cit. p. 15) calls “salutary patriotism” if they wish to foster sense of civic duty and turnout.

Our models including contextual factors yielded evidence supporting our third hypothesis: The duty to vote is stronger in more accomplished democracies. This should be the case because in such context the social norms are consistent with the concrete institutions that are in place. Nevertheless, we did not find support for our fourth hypothesis, as the levels of the duty to vote seem unaffected by a country’s degree of ethnic fractionalization. This means that the appearance of a social norm regarding the moral obligation to vote is unrelated to the presence of ethnic or linguistic divides.

We finally examined cross-level interactions between these two contextual factors and the two aspects of political support. We posited that the degree of democracy amplifies the effects of support for democracy on duty. The results confirmed our expectation. Indeed, support for democracy has stronger effects on the belief that voting is a duty in more democratic countries. The positive interaction may point to overlapping, consistent messages, both from individuals and institutions, which result in higher levels of duty to vote. It may also be that in dictatorships, democracy supporters are more focused on basic rights and liberties than on duties toward their fellow countrymen and countrywomen. Finally, it might be that citizens will only develop beliefs about their duties as citizens once the political system guarantees them political rights. Of course this also has implications for civic education programs. Before requiring duties from their citizenry, governments must assure that they have done everything to protect citizens’ rights.

The last model in our analyses confirmed our sixth hypothesis. The more divided a country is, the stronger the impact of support for one’s political community on the duty to vote. This result is in line with Putnam’s (2007) previous finding on the relationship between ethnic diversity, national identity, and civic attitudes. Indeed, societies that have overcome their divides can turn diversity into strength, building more encompassing national identities, which bolster the civic duty to vote. This suggests that those who are proud of their diverse country adhere to the symbols, rules, and institutions that keep the country together. The result also suggests that, in fractionalized contexts, the authorities concerned about turnout rates should stress in their civic education curricula all the aspects related to the construction of the country that overcome the ethnic divide, with an emphasis on electoral institutions.

Previous works on Asia’s political culture have highlighted its specificities, such as an inverse relationship between satisfaction with democracy and the actual degree of democracy, or the fact that Asians do not value individual autonomy or freedom as much as Confucianism traditional values (Morlino et al., 2011). Our work nuances these findings in several ways. First, the belief that voting is a civic duty, unlike conceptions of individual autonomy, freedom, or satisfaction with democracy, is stronger in more democratic countries. Second, our results suggest that Asian traditional values—revolving around order, community, hierarchy, tradition, and family—are no obstacle for the development of some attitudes consonant with democracy, such as the duty to vote. Our findings also suggest that the “Asian Values” theory is not suitable for analyzing the phenomenon of the duty to vote. Indeed, duty to vote—a valuable belief for the legitimacy and survival of democracies—is quite high in Asia, and not particularly higher or lower among the Sinic (East Asia) countries.

The precise role of such values, and most especially religion, in the development of the duty to vote should be addressed by further research. Indeed, “duty” is a crucial value in several Asian religions, from the dharma, which prompts Hindus to face up their responsibilities (Mittal & Thursby, 2006) and drives Sikhs for the path of righteousness, to the many duties and moral recommendations of Buddhism (O’Flaherty & Derrett, 1978). Nevertheless, a more diverse sample and an exhaustive questionnaire would be needed to identify the effect of these religious and cultural values. Similarly, cultural “tightness” might partly explain why some Asian countries adhere more to the duty to vote than others. According to this theory, tight cultures have lots of strong norms and low tolerance for deviant behavior, while loose cultures have few and weak social norms and high tolerance of deviant behavior (Gelfand et al., 2011; Pelto, 1968). Tight cultures tend to emerge in more populated societies, and in territories threatened by a hostile environment, leading to fewer political rights and liberties. This theory identifies India as a tighter culture than Japan, and our data confirm that India is a more “dutiful” country than Japan. Further research must pay attention to other religious, cultural, and moral aspects of Asian societies and their interplay with the development of the duty to vote and other citizenship norms. We hope to have opened an avenue of study that other researchers will be able to explore more deeply, with more and better data.

Funding

This work was supported by the Social Sciences and Humanities Research Council of Canada under the Insight Grant program (435-2014-0077).

Footnotes

Appendix

References