Voting and Human Rights in Democratic Societies

Abstract

The majority of research on human rights focuses on the consequences of regime-type for human rights violations, and overwhelming evidence suggests that democracies are less likely to violate human rights of their citizens as compared to non-democracies. However, a regime-type perspective is unable to account for disparities in human rights violations within democratic and non-democratic regimes. This paper disaggregates regime-type and analyzes the relationship between citizens’ participation and human rights violations. I argue that a participative citizenry, as captured by high voter turnout, is indicative of an active and vigilant populace who are more likely to hold governments accountable and ensure better human rights protections. The paper tests the relationship between human rights and voter turnout among 89 democratic countries from 1976 to 2008. The findings demonstrate that a participative citizenry enhances governmental respect for human rights.

There is near consensus in existing political science research that democracies are better protectors of human rights as compared to non-democratic regimes (see Davenport 1995, 1999; Mitchell and McCormick 1988; Poe and Neal Tate 1994; Poe et al. 1999; Zanger 2000, among others). However, we know relatively less about the differences in human rights protections within democratic and non-democratic regimes.

In order to better understand the relationship between political regimes and human rights, some outstanding questions need to be addressed further, such as what is it about a democracy that matters? Which democratic attributes play an important role that enable democratic representatives to be more respectful of human rights as compared to other political regimes? More specifically, which factors help account for variations in human rights within democratic regimes?

These questions are imperative since democracy is a multidimensional concept with several attributes (Lasswell 1950; Dahl 1971; Linz and Stepan 1996; Hartlyn and Valenzuela 1994; Jaggers and Gurr 1995; Gasiorowski 1990, 1996; Arat 1991; Gastil 1984, 1988, 1989). Thus, we need to identify and assess the effect of individual democratic attributes to get a deeper understanding of regime characteristics that induce better human rights protection. Bueno de Mesquita et al. (2005, 442) emphasize the significance of this approach by stating “…we must disaggregate democracy into its constituent parts so we can parse out the separate effects of different dimensions of democracy”. Moreover, without disaggregating democratic regimes, we run the risk of adopting a broad conception of democracy that may include respect for human rights such that assessing the possibility of a causal relationship between democracy and human rights violations will be relatively difficult due to a tautological relationship (Poe and Neal Tate 1994). This becomes especially problematic if aggregate measures of democracy such as Polity IV are used, which include repression as one its components.

This calls for a nuanced approach to better understand the political dynamics between political regimes and human rights. Recent research sheds light on variations in human rights violations within political regimes and identifies regime attributes that help explain human rights violations across both democracies and non-democracies (Davenport 1997, 1998; Richards and Gelleny 2007; Bueno de Mesquita et al. 2005; Whitten-Woodring 2009; Conrad and Moore 2010, among others).

This paper adopts a similar approach by disaggregating democratic regimes and analyzing the effect of one of the primary attributes of democracy—citizens’ participation—on human rights violations. Citizens’ participation is one of the hallmarks of a democratic society. While all democracies permit citizens’ participation, citizens in some democracies exhibit higher levels of participation as compared to others. I argue that greater citizens’ participation, as captured by high voter turnout, may engender better human rights protection by government officials. A participative citizenry signals the existence of voters who observe the performance of government officials and use the ballot box to convey their preferences to the government. This brings about a more responsive government that is held accountable by the populace.

The paper tests the relationship between citizens’ participation and human rights among 89 democratic countries from 1976 to 2008. The findings demonstrate that a participative citizenry, specifically voter turnout, plays an important role in improving human rights protection of citizens. This paper contributes to the existing literature by shedding light on the variation in human rights violations within democratic regimes and highlighting the role of citizen’s participation. More generally, the paper provides greater insight into democratic politics.

Existing Literature: Democracy and Human Rights

There is a vast consensus among scholars that democratic regimes can rein in their leaders and ensure better protection of human rights of citizens. This literature on regime-type and human rights can be organized under two broad strands of research, the role of citizens in a democracy and the institutional constraints on leaders in democratic regimes.

The first highlights the significance of citizens in democratic societies who can hold political representatives accountable for their actions, thereby making political representatives less willing to use violence as they can be removed from office (Davenport 1999; Poe and Neal Tate 1994). Moreover, citizens in democracies have access to alternative channels to express disapproval towards the government, which reduces reliance of citizens on violent methods against political leaders (Regan and Henderson 2002). Governments thereby have no incentive to resort to violence too since they do not view the opposition as a threat to their rule. An alternative perspective focusing on the role of opposition postulates an inverted u-shaped relationship between political regimes and human rights violations (Fein 1995; Regan and Henderson 2002). Democratic leaders are not threatened by existing opposition while non-democratic leaders subjugate citizens to prevent any kind of threat but semi-democratic leaders are susceptible to higher levels of threats by the opposition, which results in semi-democratic governments resorting to greater violence to suppress any kind of dissent (Regan and Henderson 2002). Note, however, that even though this literature discusses the role of citizens in a democracy more generally, the studies do not specifically propose a theory of citizens’ participation or empirically demonstrate how citizens can ensure responsive behavior from elected officials.

The second strand of research highlights the institutional constraints in democratic regimes that serve as an alternative check on democratic leaders. Coercive agents within democracies have relatively less power, thus accounting for lower levels of violence (Davenport 1999). This is possibly because of the prevalence of division of power among democracies that makes it difficult for any one actor to monopolize means of repression and commit human rights violations (Poe et al. 1999). Additionally, the emphasis on compromise and bargaining within democracies reduce the likelihood of using coercion to resolve conflicts within this regime (Henderson 1991). Approaching the relationship from a different perspective, some scholars propose a threshold effect where democracies are associated with lower levels of human rights violations but this effect is only observed after a certain level of democracy has been achieved (Davenport and Armstrong 2004; Bueno de Mesquita et al. 2005). In summation, the underlying idea conveyed in the majority of above-discussed research is that democracies generally have a better human rights record as compared to non-democratic regimes.

The difference between democratic and non-democratic regimes is evident from Table 1 that displays the percentage of democratic and non-democratic countries across different levels of human rights scores in 2008.¹ Human rights data come from the Political Terror Scale (PTS) and the scale ranges from 1 to 5 where higher numbers indicate greater human rights violations (Gibney et al. 2013).² The table highlights two significant patterns. First, a higher percentage of democracies perform well in human rights protections as compared to non-democracies. Among countries that received a score of 1 (better protection of human rights) in 2008, approximately 83% were democracies while 16% of the countries were non-democracies. Similarly, among countries that received a score of 5 (greater human rights violations) in 2008, approximately 80% were non-democratic countries as compared to 20% democracies. This is consistent with the literature discussed above that emphasizes the role of democratic regimes in protecting human rights. The second pattern that emerges is the variation in human rights violations within democratic regimes. For instance, approximately 37% of countries with a score of 3 and 51% of countries with a score of 4 on the PTS scale were democracies in 2008. This suggests that democracies do not display a homogenous performance in protecting human rights.

Table 1 Variation in human rights violations within democracies and non-democracies in 2008

The presence of such variations necessitates the need to explore the determinants of human rights violations within regimes. Moreover, as mentioned above, political regimes embody several attributes, making it difficult for us to isolate the effect of individual regime characteristics if the discourse primarily focuses on regime-type. Thus, it becomes imperative to analyze within-regime dynamics to gain insight into specific regime features that affect governmental protection of human rights. Indeed, to rectify this shortcoming, some of the recent research focuses on specific regime attributes. These include political competition and participation (Bueno de Mesquita et al. 2005; Davenport 2007), presence of elections (Davenport 1998; Richards and Gelleny 2007; Conrad and Moore 2010), and media freedom (Whitten-Woodring 2009). These studies identify different components of regime attributes that lead to better human rights protection, thereby providing a deeper understanding of the relationship between the two. This paper adopts a similar approach by disaggregating democratic regimes and assessing the effect of citizens’ participation on human rights violations among democracies.

Citizens’ Participation and Human Rights

Democracy “implies responsiveness by governing elites to the needs and preferences of the citizenry” (Verba 1996, (1) where citizens play an important role in facilitating such behavior from elected officials. Citizens’ participation is one of the core attributes of democracy and has been aptly referred to as “the heart of democracy” (Verba et al. 1995, p. 1). It becomes critical that for democracies to work effectively, citizens are vigilant, play a proactive role in politics, and are thereby able to hold political representatives accountable. Passive or inactive citizens, on the other hand, are unable to ensure responsive behavior from elected officials. Thus, the onus of ensuring good governance lies as much on the citizens as it does on political elites. Citizens’ participation is a broad concept and citizens, especially in democracies, have the opportunity of resorting to several forms of participation to influence governmental policy. These include voting in elections, participation in protests, attending public meetings, contributing resources towards campaigns, and writing letters to elected officials, among others.

One way to conceptualize citizens’ participation is through voting behavior. Voting is an established and a frequently used mechanism within democracies through which citizens voice their opinions to the government. The significance of voting is expressed by scholars such as Riker (1965, p. 25) who states, “the essential democratic institution is the ballot box and all that goes with it”. Along the same vein, emphasizing the role of elections Huntington (1991, p. 7) argues, “a twentieth-century political system is democratic to the extent that its most powerful collective decision-makers are selected through fair, honest, and periodic elections in which candidates freely compete for votes and in which virtually all the adult population is eligible to vote.”

One of the advantages of the voting mechanism is that it creates an opportunity for citizens to participate in and influence politics, thereby making it relatively easier for citizens to use this mechanism to signal their preferences to the government relative to other forms of participation.³ Voting is an established institutional process among democracies that citizens can use to influence democratic politics by rewarding good governance and penalizing poor performance in office. Given that democracies are generally more transparent and have greater press freedom, citizens have access to relevant information that helps them gain insight into policy positions of competitors and performance of incumbents in office. Much like other forms of collective participation, citizens view voting as a product of cost–benefit calculus (Riker and Ordeshook 1968; Sanders 1980) such that people are more likely to vote when the benefits of voting are greater than the costs. Herein lies the paradox of voting because in spite of the high costs involved, citizens often vote in large numbers (Downs 1957). While the causes of this paradox are tangential to this paper, of significance are the consequences of voting.

The importance of citizens’ participation on human rights performance has been emphasized in existing research. For instance, Davenport (2007) distinguishes between the voice and veto mechanisms. Voice refers to participation and competition and relates to the ability of citizens as well as competitors to hold representatives accountable while veto refers to constraints on political representatives from using coercion. This paper builds on Davenport’s voice mechanism by emphasizing the role of participation but differs from the study in one significant way. While Davenport conflates the concepts of participation and representation in his empirical analysis, this paper primarily focuses on citizens’ participation, both theoretically and empirically. Other studies that focus on participation primarily assess the role of elections (Richards 1999; Davenport 1998; Richards and Gelleny 2007). While elections play a very important function in a democracy, they do not shed light on variations in participation. This study makes an attempt to do so by specifically assessing the role of voter turnout on human rights performance in democracies.

High voter turnout may signal to the elected officials the presence of two distinct characteristics about the populace, which may engender greater accountability from the officials and ensure better protection of human rights. First, while the collective nature of voting may lead to free-rider problems among citizens (Olson 1965), a higher level of turnout may suggest that citizens are able to overcome these collective action problems to an extent thereby signaling to the elected officials the presence of citizens who actively resort to the voting mechanism to hold them accountable.⁴ Moreover, scholars such as Inkeles (1969) and Verba et al. (1995) point out that high turnout is suggestive of an active populace where people who turn out to vote are likely to be active in others aspects of democratic governance as well. Thus, high turnout connotes the presence of an engaged citizenry who use elections to signal to the political representatives that the populace closely observes their performance and behavior in office. While voting may not enable citizens to convey specific information about their interests and opinions to the government, it does inform elected officials about the general level of satisfaction or dissatisfaction about their performance (Burstein 1999). Since citizens determine the political future of elected officials at the ballot box, the latter are less likely to violate human rights because the people can use the electoral mechanism to remove violators from office.⁵ Second, high turnout also suggests that citizens actively resort to non-violent and legitimate means for conveying their preferences to the government without causing any sort of disruption within society. This may further prevent the government from resorting to repression and violation of human rights as a form of governance.⁶

It is important to note that the mere presence of an electoral mechanism does not mean that citizens automatically vote during elections. Not all citizens may use this opportunity to express dissent against the government. One can visualize a procedural democracy where citizens do not actively participate in governance. A plausible explanation could be that citizens are unable to overcome collective action problems because of the associated costs of participation. Alternatively, citizens may be generally disinterested in politics, suggesting the presence of a weak or disengaged citizenry. Regardless of the reason, this may lead to lower levels of voter turnout where in spite of the presence of an institutional mechanism of voting, citizens do not use the election mechanism to hold representatives accountable. In this scenario, political representatives have few incentives to watch their behavior and be mindful of their performance in office and will consequently be more likely to violate human rights of their citizens.

Alternatively, it is also plausible that citizens are generally active but may not vote during elections, as was the case in the USA in the 1960s and 1970s (see Powell 1986). This may engender high levels of activism but low voter turnout and could be the result of institutional barriers to voting such as registration laws as well as the nature of party and electoral systems.⁷ Scenarios where people are participative in politics but are unable to vote are still likely to signal the presence of alert masses to their representatives. Moreover, activism in a democratic society where there is free flow of information can help in educating other citizens (who may be more likely to vote during elections) about their elected officials’ performance in office. However, it is difficult to ascertain the relative costs of different types of participation and its consequences on human rights performance across democratic countries.⁸ Since general trends across countries suggest that those who are interested in politics are participative in politics as well (Inkeles 1969), including voting during elections, I assume that political activism among citizens is likely to manifest itself in higher levels of voter turnout. This leads to the primary hypothesis, H1: A high level of voter turnout is associated with better human rights protection.

While this paper assesses the effect of voter turnout on human rights violations, it is plausible that the extent of human rights protection or violation has consequences for the nature of democratic politics as well. In the context of this paper, human rights violations may adversely influence citizens’ participation. Voting during elections is one way through which citizens express their satisfaction or dissatisfaction towards elected officials on a range of policy positions. A high level of human rights violations is indicative of a repressive society, which, more generally, could prevent people from expressing dissent. More specifically, human rights violations may reduce voter turnout. Such violations could be carried out with the intention of preventing people from going to the polls and voting for the opposition or just expressing their dissatisfaction with the current incumbents in office. People may not vote if there is a possibility of intimidation during elections. Low levels or non-existence of human rights violations, on the other hand, create a permissive environment such that citizens can participate in democratic politics without fear of intimidation. This is likely to espouse a participative society more generally and may lead to higher levels of voter turnout as well. Thus, human rights violations and voter turnout may influence each other. While the paper primarily explores the effect of voter turnout on human rights, it does address the plausibility of a simultaneous relationship between the two below.

Data and Methods

This paper assesses the effect of citizens’ participation on human rights violations by governments on a global sample of 89 democratic countries from 1976 to 2008.⁹ The sample is restricted to democracies so as to assess the effect of voter turnout when elections are free and fair, as is the case with democracies. A dichotomous measure of democracy is used to classify countries and the data come from Cheibub et al. (2010). The “democracy” variable classifies a country as a democracy if there is an elected legislature and a chief executive and the opposition has the opportunity to win office. This is a minimalistic conceptualization of democracy, which is important for the purpose of this paper as a substantive conceptualization may lead to a tautological relationship between the theoretical variables of interest.

The dependent variable is human rights violations. There are two prominent human rights data sources—the PTS and the Cingranelli-Richards (CIRI) Human Rights Database. PTS primarily refers to the use of violence by the state against its citizens and provides three indicators based on Amnesty International, U.S. State Department, and Human Rights Watch reports. PTS scores based on reports from the U.S. State Department are used for primary analyses while scores from Amnesty International reports are used for robustness tests.¹⁰ The scale ranges from 1 to 5 where higher numbers indicate greater human rights violations. The data are available from Gibney et al. (2013). The CIRI data provide a cumulative physical integrity index ranging from 0 to 8 where higher values indicate better protection of human rights and the data are available from Cingranelli et al. (2013). The index adds four different types of human rights violations, namely, torture, extrajudicial killings, political imprisonment, and disappearance. This is used as a robustness test and the index is reversed such that it is similar to the PTS scale where higher numbers refer to greater human rights violations. The primary independent variable is citizens’ participation, which is measured with voter turnout (as a percentage of registered voters) for legislative elections.¹¹ Voter turnout is alternatively measured as a percentage of the voting age population, which is used as a robustness check. Voter turnout data are available from Banks and Wilson (2013).

A number of control variables are included in the models to account for alternative determinants of human rights violations. The level of democracy is measured with polity2, which ranges from −10 to +10 where higher values indicate a higher level of democracy (Marshall et al. 2013). As most of the existing research discussed above suggests, higher levels of democracy are associated with better protection of human rights. Income is measured with GDP per capita. States with higher income are expected to have greater respect for human rights because stronger economies are less likely to face domestic unrest and thus the government is unlikely to resort to human rights violations (Poe et al. 1999). The size of population is a standard control and the expectation is that states with a larger population are more likely to witness human rights violations due to limited availability of resources (Henderson 1993) or due to the presence of additional opportunities to violate human rights (Henderson 1993; Poe and Neal Tate 1994). The data for income and population come from the World Bank (World Bank 2013).

Involvement in an international or civil war may also increase human rights violations by governments (Poe and Neal Tate 1994) and the data for both variables come from Major Episodes of Political Violence (MEPV) (Marshall 2013). International war is measured as the sum total of variables “international violence” and “international warfare” where each variable ranges from 0 to 10 such that the international war variable ranges from 0 to 20. The international war variable is coded into a dichotomous variable in this paper where values greater than 0 are coded as 1 suggesting the presence of conflict while 0 indicates its absence. Civil war is measured as the sum total of four other variables: “civil violence,” “civil war,” “ethnic violence,” and “ethnic war” and each of these range from 0 to 10 such that the civil war variable ranges from 0 to 40 where higher values indicate higher levels of civil conflict. The civil war variable is coded into a dichotomous variable as well where values greater than 0 are coded as 1 suggesting the presence of civil conflict while 0 indicates its absence. Finally, consistent with most studies on human rights, I include a lagged version of the dependent variable, as a government’s past behavior with regard to human rights may be indicative of future human rights violations as well (Poe and Neal Tate 1994; Davenport and Armstrong 2004; Whitten-Woodring 2009, among others). Additionally, a lagged dependent variable also accounts for alternative factors that are not captured in the models but may influence human rights. All independent variables are lagged by one time period.¹²

While the paper primarily assesses the effect of voter turnout on human rights, it is plausible that there is a simultaneous relationship between the two. In order to determine if that is the case, I conducted a Hausman specification test (Gujarati 2003, pp. 754–756). However, the results did not reveal the presence of simultaneity between the two.¹³ In the absence of simultaneity, the relationship between human rights violations and voter turnout is estimated with an ordinary least squares (OLS) model. Given the nature of the dependent variable, an alternative way to estimate the relationship between the two is to use an ordered logit or probit. An OLS technique makes it easier to interpret the effect of the independent variables on the dependent variable and has been used in comparable human rights studies (such as Whitten-Woodring 2009). However, results are presented with an ordered logit model as well as a robustness test.

Empirical Analyses

Table 2 shows the effect of voter turnout on human rights violations where columns 1, 2, and 3 use different measures of human rights. Table 2 and columns 1 and 2 include the PTS scores based on state department and amnesty international reports, respectively, while column 3 includes the physical integrity index. Higher values on the dependent variables indicate higher levels of violations.

Table 2 Effect of voter turnout on human rights violations (with alternative dependent variables)

Voter turnout is negatively correlated with human rights violations in models 1 through 3 where a lower level of voter turnout is associated with greater human rights violations or a higher voter turnout is correlated with better human rights. This finding is consistent with hypothesis 1. Higher voter turnout suggests the presence of vigilant and proactive populace who has been able to overcome collective action problems and use the electoral mechanism to hold elected officials accountable. Since elections can be used to award and penalize political representatives, elected officials may be especially wary of violating human rights of their citizens when turnout is high. Moreover, since voting is a legitimate means of citizens’ participation, political representatives may also find it relatively difficult to justify resorting to human rights violations when citizens resort to an institutionally accepted mechanism such as voting to signal their preferences to those in office. Overall, the findings demonstrate the significance of voting and the role citizens can play in ensuring that political representatives respect citizens’ human rights.

Figure 1 provides greater insight into the relationship between the theoretical variables of interest. Following Williams and Whittens (2012), Fig. 1 captures the long-term relationship between voter turnout and human rights violations through dynamic simulations. There are two primary advantages to using this approach. First, it enables us to assess the effect of the independent variable of interest over time. Second, we can identify if different values of the independent variable has a statistically distinguishable effect on the dependent variable over time. The figure presents the predicted human rights violation scores over 9 years (with 95% confidence intervals) under three different scenarios, when voter turnout is at its minimum, mean, and maximum levels, while holding the continuous control variables at their means and the dummy variables at 0.¹⁴ The figure demonstrates that in the first year a country with the highest level of voter turnout has a level of human rights violation that is statistically indistinguishable from medium levels of voter turnout. However, it is distinguishable from lowest levels of voter turnout. In the second year, a country with highest levels of voter turnout has a level of human rights violation that is statistically significantly lower than mean and lowest levels of voter turnout and this effect prevails over time. In substantive terms, by year 6, there is approximately a 1-point difference between states with lowest versus highest levels of voter turnout (with human rights violations point estimates at 2.6 and 1.6 for lowest versus highest levels of voter turnout, respectively). The human rights variable is coded on a scale 5-point scale so a 1-point change equals a 20% difference over time. This suggests that substantively the effect of voter turnout is noticeable in the long run.

Fig. 1

Effect of voter turnout on human rights violations

Figure 2 provides further insight by simulating the relationship between the two in Benin, one of the countries in the sample where voter turnout varies from 70% to 55.9%. Figure 2 presents the simulated human rights violation scores over 9 years (with 95% confidence intervals) in Benin as voter turnout reduces from the maximum to the minimum level within the country while holding the continuous control variables at their means values in the country and the dummy variables at their mode values in the country, which is 0. Here again we see that the effect of a reduction in voter turnout from its maximum to minimum value is associated with a 1-point difference (with human rights violations point estimate increasing from 2.6 to 3.6) in human rights violations by year 8. This reiterates that substantively the effect of voter turnout is evident over time.

Fig. 2

Effect of voter turnout on human rights violations in Benin

The control variables perform as expected in most models. Consistent with existing literature, both democracy and income are negatively correlated with human rights violations such that higher levels of democracy and greater income are associated with better protection of human rights. There is near consensus among scholars that democracies perform better than non-democracies in human rights and the findings presented here reiterate this. States with greater income also perform better than states with lower levels of income. Higher income may suggest presence of fewer grievances among the populace, which is turn reduces domestic turmoil and political representatives may not feel the need to resort to violence to maintain order and control within societies. The civil conflict variable is positively associated with human rights violations, which provides support to the idea that the presence of a civil conflict may increase human rights encroachments by governments. However, the interstate conflict variable does not reach statistical levels of significance. Size of population is positively correlated with human rights violations. States with a larger population are associated with greater violation of human rights plausibly because it reduces the pool of available resources for the people and enhances the possibility of domestic unrest, which consequently increases the likelihood that governments may violate human rights. Lastly, the lagged dependent variables are significant in all three models, suggesting that the governments’ treatment of its citizens in the past is indicative of how the latter will be treated in the future.

Robustness Tests

A number of additional tests were carried out to analyze the robustness of the baseline findings. Table 3 presents the first series of robustness tests. The baseline findings as discussed above include a lagged dependent variable, consistent with most studies on human rights. However, a lagged DV runs the threat of reducing the significance of the primary independent variable of interest (Achen 2000). Model 1 presents findings with AR(1) to address serial correlation instead of a lagged DV. Model 2 presents findings with country fixed effects. While this paper assesses the relationship between the theoretical variables of interest across countries, a fixed effects estimation technique enables us to see if change in voter turnout within a country has an impact on human rights violations. Model 3 includes year fixed effects to ensure that the effect of voter turnout on human rights violations is not being driven due to a particular year. Lastly, given that the dependent variable is ordinal in nature, model 4 estimates the effect between human rights violations and voter turnout with an ordered logit model. All four models in Table 3 reiterate the primary baseline finding where high levels of voter turnout continue to be associated with lower levels of human rights violations.

Table 3 Effect of voter turnout on human rights violations (with alternative model specifications)

Table 4 presents the next set of robustness tests. The baseline models measure the primary independent variable with voter turnout (as a percentage of registered voters) and the data come from Banks and Wilson (2013). Models 1 and 2 in Table 4 present findings with voter turnout (as a percentage of registered voters) and voter turnout (as a percentage of the voting age population) respectively and the data for both variables come from an alternative data source, IDEA (International Institute of Democracy and Electoral Assistance 2013). Substantive findings remain the same regardless of how voter turnout is measured or the data source used. Table 5 assesses the relationship between the two among OECD countries (model 1) and non-OECD countries (model 2). High voter turnout is once again associated with lower human rights violations among both developed and developing countries.

Table 4 Effect of voter turnout on human rights violations (with alternative measures of voter turnout)

Table 5 Effect of voter turnout on human rights violations (among OECD and non-OECD countries)

Table 6 presents findings where the sample is restricted to countries that have compulsory voting (model 1) and countries that do not have compulsory voting (model 2). There are 21 countries that have compulsory voting in the sample as compared to 69 countries that do not have compulsory voting. Voter turnout remains significant in both models. This suggests that the role of voter turnout is important not only in societies where voting is compulsory but also where voting is voluntary. While baseline models present findings with the standard controls that have been used in the majority of studies on human rights, recent research has identified other regime attributes that play an important role in improving human rights performance. Table 6, model 3, includes additional controls to assess if the effect of voter turnout remains significant after accounting for these alternative determinants of human rights. These include political competition (Bueno de Mesquita et al. 2005; Davenport 2007), legislative and presidential elections (Davenport 1998; Richards and Gelleny 2007; Conrad and Moore 2010), and freedom of the press (Whitten-Woodring 2009; Conrad and Moore 2010). Political competition is measured with the “parcomp” variable, which ranges from 0 to 5 where higher numbers are indicative of higher levels of competition and the data come from the Polity IV dataset (Marshall, Gurr, and Jaggers 2013).¹⁵ Dummy variables are included for years that have legislative and executive elections and the data come from the Database of Political Institutions (Beck et al. 2001). Freedom of the press is measured with the “speech” variable, which ranges from 0 to 2 where higher numbers indicate greater media freedom and the data are available from Cingranelli et al. (2013). Among the additional controls, political competition is negative and significant indicating that a higher level of competition is associated with lower levels of human rights violations. The other controls are in the expected direction but are statistically insignificant. The theoretical variable of interest, voter turnout, continues to be statistically significant. Lastly, Table 7 presents findings with 2-, 3-, and 4-year lags to further address the issue of reverse causality (models 1, 2, and 3, respectively). Voter turnout continues to have a significant on human rights, further reiterating its significance in improving human rights performance among democratic regimes.

Table 6 Effect of voter turnout on human rights violations (alternative robustness tests)

Table 7 Effect of voter turnout on human rights violations (extended voter turnout lags)

Concluding Thoughts and Policy Implications

The bulk of the existing research on the determinants of human rights primarily analyzes the effect of political regimes on human rights violations and overwhelming evidence suggests that democratic regimes are better protectors of human rights as compared to non-democratic regimes. This paper is an attempt to bring greater clarity to the relationship between political regimes and human rights violations by disaggregating regimes and analyzing the effect of citizens’ participation on the state’s use of repression. In doing so, it reiterates the significance of democratic politics and provides further insight to explain disparities in human rights protection among democratic regimes.

The findings demonstrate the significance of citizens’ participation, particularly voter turnout, in enhancing governmental respect for human rights. The policy implications of the findings suggest that the populace can play an important role in facilitating better protection of human rights. State repression is prevalent in countries that hold elections and permit citizens’ participation albeit in varying degrees. One of the ways citizens can ensure that governments observe human rights is by voting during elections and holding representatives accountable. Failure to do so may adversely affect well-being of the masses. Thus, the responsibility of protecting human rights lies on the government but also on the masses that need to use the electoral mechanism, if they have the opportunity to do so.

There are other avenues of future research that can help shed greater light on the relationship between political regimes and human rights. This is a one of the first attempts that analyzes the role societal actors can play in ensuring better protection of human rights and future research can further explore the relationship between the two. Since citizens in a democracy have access to alternative mechanisms to participate in democratic governance, it would be useful to analyze the relative effectiveness of alternative types of citizens’ participation in facilitating better human rights protections by governments. Furthermore, analyzing the relationship between other regime attributes and human rights can also further our understanding of the link between the two.

Appendix

Table 8 Descriptive statistics

Table 9 Correlation matrix

Table 10 Hausman endogeneity test

Table 11 Findings from 2SLS estimation

Table 12 Variations in voter turnout