Are stronger states more humane? A re-evaluation of ‘exemplary villains’

Abstract

In general, the human rights scholarship finds stronger states to be more humane. In particular, Englehart’s recent book and article assert that ‘petty despots’, rather than ‘exemplary villains’, are the real culprits behind the abuse of human rights. However, some exemplary villains are truly exemplary; providing them with greater powers will only intensify human rights abuses. Indeed, the idea that even dictatorships would better protect rights when stronger contradicts some of the most well-known cases of mass-murder that occurred in powerful authoritarian states. In an attempt to resolve this puzzle, this article argues that, while state capacity does matter, it matters differently for dictatorships and democracies. In both regimes, there are certain types of human rights violations that, from the government’s perspective, happen unwittingly. These unauthorized human rights violations are perpetrated by petty despots trying to benefit from the state’s principal–agent problems. Yet, unlike a democracy, a dictatorship can authorize the violation of certain types of human rights to accomplish its objectives. Whereas state capacity can reduce ‘unauthorized’ types of abuses, it would not necessarily diminish such abuses that are ‘authorized’ by the state. The net effects of state capacity, therefore, will not be as positive for autocracies as they are for democracies. To validate this argument, this article conducts 18 ordered logistic regressions with a time-series cross-sectional dataset that encompasses no less than 142 countries from the period of 1981 to 2002. The empirical analysis provides solid support for the theorized relationship among human rights, state capacity and regime type. As predicted, state capacity improves human rights in democracies but not necessarily in autocracies.

Keywords

autocracy democracy human rights state capacity

Introduction

In general, the human rights scholarship finds stronger states to be more humane (Butler, Gluch & Mitchell, 2007; Acemoglu, Ticchi & Vindigni, 2010; Fearon & Laitin, 2003; Sullivan, 2012; Englehart, 2009; Englehart, 2017). In particular, Englehart’s (2017, 2009) recent book and article assert that ‘petty despots’ (low-ranking government agents), rather than ‘exemplary villains’ (dictators and their close allies), are the real culprits behind the abuse of rights. However, some exemplary villains are truly exemplary; providing them with greater powers will only intensify human rights abuses. Indeed, the idea that even dictatorships would better protect rights when stronger contradicts some of the most well-known cases of mass-murder that occurred in powerful authoritarian states, which begs the question: are stronger states invariably more humane?

For democracies, the answer is rather simple: ceteris paribus, a stronger state would protect rights more effectively than a weaker state, as its policing efforts can prevent state agents from committing unauthorized acts of abuse. In non-democracies, however, the problem is more complex. More power to the state could, on one hand, protect its citizens from any unauthorized abuse perpetrated by ‘petty despots’; it could, on the other hand, also increase a state’s ability to purposefully oppress certain areas of rights by authorizing violations. As a result of these complications, increased state capacity may not necessarily improve human rights when a regime is non-democratic. Put differently, the literature’s apparent discrepancy with common knowledge may have arisen from the fact that existing studies do not test for any interactions between state capacity and regime type. While state capacity does matter, it may matter differently for dictatorships and democracies.

This article will carefully assess the hypothesis that state capacity may not yield as positive an effect on human rights in autocracies as it does in democracies. To validate this argument, this article will conduct 18 ordered logistic regressions with a time-series cross-sectional dataset that encompasses no less than 142 countries from the period of 1981 to 2002. The empirical analysis provides solid support for the theorized relationship among human rights, state capacity and regime type. As predicted, state capacity improves human rights in democracies but not necessarily so in autocracies.

In short, ‘exemplary villains’ deserve another close look. Depending on the circumstances, empowering dictators may not be the answer, because ‘petty despots’ are not always the main cause of human rights abuse.

Literature review

Why do states infringe human rights? Initial investigations into this topic found that a state’s tendency to violate rights is largely shaped by economic, political, and legal conditions. In more recent times, however, scholars started to wonder whether some states are actually incapable of protecting human rights even if they intended to. Enforcing human rights standards requires concentrated law enforcement efforts that are unhindered by information asymmetries; but research finds that weak states have been simply unable to achieve those ends without the proper means. This article seeks to take one step further in this direction, exploring when state capacity can improve rights and when it cannot. Specifically, it aims to investigate whether state capacity enhances human rights prospects regardless of regime type. Although some scholars have already noted the possibility that the effects of state capacity on human rights are contingent on the type of regime, empirical evidence in this area has been far from satisfactory.

Earlier studies of human rights tended to focus on three structural variables: economic development, regime type, and human rights legislation. For many scholars, the economic well-being of a nation was considered necessary for the pursuit of human rights. McNamara (1966), for instance, argued that poor countries are more prone to human rights violations, since the violation of rights is an innately ‘backward’ behaviour. Huntington (1968) suggested that the relationship between development and human rights is a little more complex than that. In his view, it was the modernization of states that put human rights under risk, rather than poverty per se. Thus, the very poor and the very rich nations are the least susceptible to violations of rights, whereas those in the middle – struggling to modernize – are more vulnerable. In an altogether different approach, some scholars found the variation to arise from external economic relationships among countries. For example, Chomsky & Herman (1979) argued that economic ties between rich and poor countries may end up creating more violence in poor nations, as the less well-off seek to formulate conditions receptive to foreign investment.

Another often explored structural element is the political conditions of a nation. In particular, regime type was considered a central variable. In autocracies, the rights of citizens are often violated in the interests of the authoritarian leadership (Escriba-Folch, 2013; Davenport, 1999, 2004; Poe, Tate & Keith, 1999). Different types of non-democracies – military, totalitarian, and authoritarian – have all been found to be prone to human rights violations (Mitchell 2004: 15). Democratic governments, to the contrary, are more accountable to their citizens; and the costs of violating human rights are greater than they are in autocracies. Accordingly, scholars have found democracies to be more protective of human rights than dictatorships. For instance, Poe & Tate (1994) and Poe, Tate & Keith (1999) point to a positive, linear relationship between human rights and democracy, whereas Davenport (1999, 2004, 2007) and Davenport & Armstrong (2004) find the level of repression to be significantly lower in democracies than in autocracies. Englehart’s (2009, 2017) quantitative studies also find that the Polity IV score has positive, linear effects on human rights, specifically political and civil rights.

A few scholars have suggested more complex relationships between regime type and human rights. First, it has been argued that specific dictatorships benefit from violating particular types of human rights. As an example, Escriba-Folch (2013) argues that there exist two different types of repression – restrictions on civil rights and violation of physical rights – and that the effectiveness of each depends on the type of threat a dictatorship faces (Escriba-Folch, 2013). Secondly, echoing Huntington’s argument that ‘modernization’, rather than poverty itself, is the cause of human rights violations, Fein (1995) finds that more murder occurs ‘in the middle’, when a country is in between a democracy and an autocracy. Similarly, Regan & Henderson (2002) argue that more human rights violations occur under hybrid or mixed regimes, while Davenport & Armstrong (2004) and Bueno de Mesquita et al. (2005) suggest that democracy could have non-linear effects on human rights. These scholars explore the possibility of there being a ‘tipping point’, or a minimum level of democracy from which human rights are well kept.

Finally, a group of scholars have emphasized the role of legal devices in the pursuit of human rights. Henkin (1990), as an example, claims that the concept of human rights has become ‘universal as well as international’ and argues that the domain of rights no longer belongs to nation states alone. Relatedly, Bobbio (1996) asserts that the primary challenge in the international human rights regime is ‘legal’. As with Henkin, Bobbio (1996: 12) is convinced that human rights has now achieved ‘universality’. These studies emphasize the role of international organizations, as their central concern lies with the enforcement of international human rights laws.

These earlier approaches suffer from a crucial drawback: no matter how prosperous and democratic a country is, or how well it abides by international legal norms, a state might be simply incapable of enforcing human rights. There is a significant variation in the ability of states to implement domestic and international regulations. From failed states like Somalia to strong, bureaucratic states like the Republic of Korea, states can be placed along a wide spectrum of capabilities. And, as Englehart & Miller (2014) demonstrate through their analysis of women’s rights, enforcing drastic cultural and social change has not been easy even for powerful states. Without the right amount of state capacity then, protecting human rights can be a greater challenge. Variations in state capacity could lead to meaningful differences among states in their protection of human rights, since legal and political commitments cannot be taken seriously in states with low capabilities (Englehart, 2017: 16).

Indeed, a quick glance at the history of human rights records will show that the state needs more than just economic prosperity, democracy and regulations to protect rights. No matter how economically developed or democratic a country is, or how well international and domestic laws articulate provisions for protecting human rights, citizens could still remain prone to violence if the state cannot enforce rules. Despite global economic growth, the spread of democracy and the large amount of human rights legislation created since the end of World War II, neither of the two main human rights indices (PTS scores and the CIRI physical integrity index) improved over time (Englehart, 2017). While Fariss (2014) argues that this apparent stagnation in human rights progress is actually due to a ‘systematic change’ in the way human rights indicators are measured, a large body of cross-national studies maintain that the spread of international human rights treaties and legislation alone has not been effective at improving human rights worldwide (Keith, 1999; Hathaway, 2002; Neumeyer, 2005; Hafner-Burton & Tsutsui, 2007; Englehart, 2009; Hill, 2010). Taken together with the contrasting positive findings of qualitative researchers (e.g. Lutz & Sikkink, 2000; Brysk, 2005; Risse, Ropp & Sikkink, 1999), these studies indicate that the structural variables explored by some of the earlier studies in the literature may improve human rights in some areas but not necessarily in others. They demonstrate that ‘human rights successes are not inevitable; some countries are better candidates than others’ (Hafner-Burton & Ron, 2009: 368).

More recent studies have, therefore, included state capacity in their analyses. Where state capacity was included as a variable, scholars have found it to affect human rights in a positive direction. In their cross-national study of sexual violence, Butler, Gluch & Mitchell (2007) find that sexual violence was lower when agents of the state were subject to tighter control. While not addressing state capacity per se, the authors shed light on the importance of a state’s control over its agents, which is a crucial component of state capacity. Analysing civil conflicts, Acemoglu, Ticchi & Vindigni (2010), Fearon & Laitin (2003) and Sullivan (2012) suggest that human rights violations persist when state institutions are not strong enough to end conflict. In the absence of strong, centralized states, property rights are not properly enforced, tax revenues are not sufficiently collected and public goods are not well provisioned. In such circumstances, state and non-state actors alike have incentives to use violence to secure property, tax and public goods. Finally, time-series cross-sectional analyses from Englehart (2009, 2017) also suggest that state capacity can reduce the likelihood of human rights violations. The author argues that, since most of the violations are caused by ‘petty despots’, a more capable state can better protect rights by circumscribing these everyday extortions.

It must be recognized, however, that state capacity can be a double-edged sword. As Foweraker & Landman (2000: 3) mention in their analysis, we must remember that ‘citizenship both depends upon and is threatened by the state, because what the state gives it can also take away’. As a matter of fact, 71% of all mass killings in the last 200 years were carried out by governments of three arguably very capable states: Nazi Germany, Stalin’s USSR and China (Easterly, Gatti & Kurlat, 2006). So, when are capable states protective of human rights, and when are they destructive? This article will argue that the effect of state capacity on human rights depends on the type of regime: whereas more capable democracies become better at protecting human rights, this is not necessarily true for more capable autocracies. Although some scholars may have already noted this possibility, to my knowledge there has not been any attempt to empirically assess this claim on a global scale. In other words, while the human rights implications of democracy and state capacity have each been studied in the past, there has been a lack of attention to the possibility that state capacity and regime type could jointly affect human rights.

In sum, while the literature has come a long way, not enough research has been done to reveal how regime type could affect the effect of state capacity on human rights. Considering the high number of human rights abuses that occurred in arguably some of the most powerful autocracies in history, there is reason to suspect that an interaction between state capacity and regime type will beget interesting and meaningful results.

Exemplary villains and petty despots: Authorized vs. unauthorized rights abuse

Englehart (2017) makes a good example of the state of the literature. The author argues that, controlling for other structural factors like economic development and the level of democracy, state capacity can positively affect human rights, because many human rights violations are committed by ‘petty despots’ rather than ‘exemplary villains’. ‘Much of the human rights abuse’, Englehart argues, ‘is committed not by exemplary villains but by petty despots, including grasping landlords, corrupt minor officials, local party bosses, gangs of armed thugs and bigoted individuals’ (Englehart, 2017: 3). A stronger modern state, with tighter control over its bureaucracy and a monopolization of violence, can thus significantly improve human rights by policing these unauthorized acts of abuse. Consequently, Englehart (2017) places great emphasis on a state’s ability to police its agents and enforce its laws. A state that fails to effectively keep petty despots in check ends up having high levels of abuse.

However, Englehart (2017) depicts only part of the picture. Human rights violations can be either ‘authorized’ or ‘unauthorized’ by the state. Englehart (2017) mainly deals with unauthorized human rights violations that are perpetrated by ‘petty despots’: low-level

Table I.

Authorized versus unauthorized violations

	Authorized violations	Unauthorized violations
Root cause	‘Exemplary villains’ (principals)	‘Petty despots’ (agents)
Principal–agent interests	Aligned	Not necessarily aligned
Examples	Machiavellian motives Grand Inquisition	Looting Demanding bribes

government bureaucrats, law-enforcement officers and military personnel.¹ States that can enforce rules effectively will be better at preventing these misbehaviours. Yet a key distinction between a dictatorship and a democracy is that, unlike a democracy, a dictatorship also deliberately violates certain types of human rights. As opposed to unauthorized human rights violations, these authorized human rights violations occur because the government – or an ‘exemplary villain’ – wishes to do so. Mitchell (2004) presents two reasons why a government might intentionally seek to violate human rights. First, a dictator may have Machiavellian motives. To mitigate threats against his tenure, a dictator could employ necessary forces against his foes and their (potential) supporters. Second, a dictator of the ‘Grand Inquisitor’ type may utilize force to pursue his ideological agenda (Mitchell, 2004). For example, it is well documented that North Korean women have been forced to abort children of mixed race, to ensure that the regime could maintain what is believed to be Korea’s ethnic purity (UN Human Rights Council, 2014; Myers, 2010). Such a type of rights violations should be considered ‘a policy’ (Mitchell, 2004: 48). It is an agenda that is dexterously carried out by a regime in order to achieve a goal considered crucial by its rulers, and state capacity will not necessarily moderate it (Table I).

What follows is that, while state capacity does matter, it matters differently for dictatorships and democracies. The two categories of rights abuse (authorized and unauthorized) are affected differently by changes in state capacity, and this discrepancy means that the net effect of state capacity varies by regime type. While state capacity decreases unauthorized human rights violations in

Table II.

Human rights violations (by type) in high/low capacity democracies and dictatorships.

	Democracy	Dictatorship
Low capacity	Authorized violations: infrequent Unauthorized violations: frequent	Authorized violations: frequent Unauthorized violations: frequent
High capacity	Authorized violations: infrequent Unauthorized violations: infrequent	Authorized violations: more frequent Unauthorized violations: infrequent

both dictatorships and democracies alike, it does not necessarily reduce authorized human rights violations that are pursued by autocracies. Unauthorized violations become infrequent as a state becomes more powerful, because the government can resolve principal–agent problems by effectively gathering information about any unwanted wrongdoings and punish them accordingly. This aspect of state capacity’s effect on human rights has been well documented by the existing literature. The effect on authorized human rights violations, on the other hand, has not been sufficiently explored. The argument presented in this article is that a more capable authoritarian regime will more adroitly carry out authorized violations. Thus, resolving a dictator’s agency loss may exacerbate some areas of human rights even while improving rights in others. To bring us back to the example of North Korea, a stronger North Korean regime may become able to more efficiently detect the so-called ‘impurities’ created by mixed-ethnicity children and murder these unborn babies without failure. The existence of such purposeful human rights violations recall the words of Rummel (1997: 9), who writes that ‘power kills; absolute power kills absolutely’.

Put together then, the effects of state capacity for human rights would be contingent on the type of regime. For democracies, human rights are better in all respects when a state is more capable. For dictatorships, however, the effect of state capacity is rather ambiguous: it may improve rights that had unintentionally been violated under a weaker state, but enhanced state capacity will not necessarily improve rights that a state deliberately wishes to infringe (Table II).

Formally, the aforementioned relationship among state capacity, democracy and human rights can be described as follows. Where R stands for the number of human rights violations, n denotes the ‘natural level’ of human rights violations, $d$ shows whether a regime is autocratic ( $d = 0$ ) or democratic ( $d = 1)$ ,² c represents civil society’s positive influences on the protection of rights, s denotes state capacity ( $s \in [0, 1]$ ), v_u stands for unauthorized violations of human rights, and v_a stands for authorized violations of human rights, the number of human rights violations R can be expressed as a simple linear function of $n, d, c, s, v_{u}, v_{a}$ :

R = n - d c + (1 - s) v_{u} + (1 - d) v_{a} s

Then, if $d = 1,$

R = n - c + (1 - s) v_{u}

To the contrary, when $d = 0,$

R = n + (1 - s) v_{u} + v_{a} s

\Leftrightarrow R = n + v_{u} + s (v_{a} - v_{u})

The models demonstrate how regime type determines the relation of state capacity to human rights. In democracies ( $d = 1$ ), the relationship between state capacity and human rights violations is strictly negative (since $- v_{u} \leq 0)$ . On the other hand, the association is rather complex in autocracies. Firstly, if the dictator is a ‘Grand Inquisitor’ type ( $v_{a} > v_{u}$ ), infringing rights for the sake of imposing his ideals, improved state capacity only exacerbates human rights conditions (since $v_{a} - v_{u} > 0$ ). Second, if the autocratic leader has more Machiavellian inclinations, he will seek to minimize both authorized and unauthorized violations, since ‘cruelty is not pursued for any other goals or for its own sake, and it must be handled carefully as it carries potential costs in political support’ (Mitchell, 2004: 29). In this case, the effect of state capacity is ambiguous and context-dependent. In times when the Machiavellian dictator finds violence crucial for his own survival, v_a may exceed v_u, but even that would depend on that particular nation’s leniency towards unauthorized human rights violations, since some nations may simply have naturally higher levels of v_u. Figure 1 visually illustrates the implications of the formal model.

Empirically, even a cursory glance at the human rights data indicates clear contrasts between dictatorships and democracies. To measure ‘state terror’, or ‘violations of

Figure 1.

A formal model of state capacity, regime type and human rights using Mitchell’s (2004) categorization of dictatorships

Figure 2.

PTS (Amnesty International) and PTS (US State Department) yearly trajectories

physical or personal integrity rights carried out by a state (or its agents)’, Gibney & Dalton (1996) created the Political Terror Scale (PTS). As it is the most commonly used measure of human rights, Gibney’s most updated PTS scores will be mainly used for creating the summary statistics (Wood & Gibney 2010: 369). Because the dependent variable of interest is the level of respect for human rights, rather than the number of human rights violations, the PTS scores have all been inverted in the following discussions. In addition to the PTS, two indices from Cingranelli, Richards & Clay’s (2014a,b) CIRI index will be analysed. Another widely used indicator of human rights, the CIRI has the upside of having coded different types of human rights violations separately. Using the CIRI will thus make it possible to compare the patterns of authorized versus unauthorized human rights abuses.

Overall, PTS scores exhibit some interesting patterns. First, the time trends differ by regime type (Figure 2). Gibney created two PTS scores using different sources: Amnesty International and the US State Department. Figure 1 shows that, for both of these, the time trends are not the same in democracies and in dictatorships. To give an example, according to both indices, human rights from 2011 onwards improved in democracies but deteriorated in dictatorships. These diverging trends add weight to the suspicion that state capacity could have different implications for human rights depending on the type of regime. Furthermore, as summarized in Figure 3, when PTS scores are juxtaposed against state capacity, the relationship differs depending on the type of regime. Among democracies, state capacity correlates positively with the inverted PTS. Among non-democracies, however, the relationship is more complex. The most capable states are not necessarily the most respectful of citizens’ rights.

While every measurement of rights violations used in this article is a mixture of both authorized and unauthorized rights, we could expect some indices to vastly reflect authorized violations and others to mostly represent unauthorized ones. If so, human rights indices can be further disaggregated into rights most likely to be

Figure 3.

State capacity and physical integrity violations by regime type

violated by the state and rights less likely to be violated by the state, in order to make some initial assessments about this article’s proposed theoretical argument.³ For one, political imprisonment is a rights abuse that is likely to have been authorized by the state. To imprison a large group of citizens for political reasons requires the combined efforts of a state’s judiciary and its prison system; it is thus unlikely to be the product of some rogue agents. By comparison, women’s social rights (such as freedom from female genital mutilation) are more often than not violated by rogue agents, because most of the time the state legally bars its agents from doing so: according to the CIRI data, women’s social rights are legally protected in approximately 83% of country-years in the sample.

What do these indices tell us about the theory at hand? Authorized rights violations tend to show different patterns by regime type. As seen in Figure 4, no political imprisonment (2) is associated with higher state capacity in democracies but not in autocracies. To the contrary, unauthorized rights abuses tend to drop where state capacity is higher, regardless of the type of regime. Figure 5 illustrates how the social rights of women are correlated with higher state capacity in both democratic and authoritarian regimes.⁴ In both cases, the average level of state capacity is higher in countries with greater social rights for women.

As such, the descriptive statistics substantiate the theoretical conjecture that stronger authoritarian states can curb unauthorized violations of human rights but not necessarily reduce infringements that are authorized. If, as Englehart (2017) argues, a capable authoritarian state has better chances of protecting human rights than a failed state, it is unlikely for such variations to arise from the type of regime or from the kind of human rights. There is good reason to suspect that the relationship between state capacity and human rights is more complex.

Re-evaluating exemplary villains

The literature in its present condition is thus missing a critical distinction. It has not tested the possibility that state capacity can have different effects on different types of rights in different types of regimes. To fill this gap, this article will empirically evaluate the effects of state capacity on diverse types of rights by the type of regime. It is expected that state capacity will reduce both authorized and unauthorized human rights violations in a democracy but (if at all) only partially improve human rights in a non-democracy. In an effort to validate this argument, this article will conduct a series of regressions on a global time-series cross-sectional dataset that encompasses no less than 142 countries from the period of 1981 to 2002. Ordered logistical regressions will be employed, because the dependent variables are categorical; Huber-White standard errors will be used to mitigate for the fact that the dataset was not drawn from a random sample; finally, year dummies will be included in the regressions to account for time invariant and almost time invariant factors, but these will not be displayed in the tables. The findings provide strong support for the hypothesized

Figure 4.

State capacity and authorized violations by regime type

Figure 5.

State capacity and unauthorized human rights violations by regime type

relationship among human rights, state capacity and regime type.

Dependent variable: Human rights

The dependent variable will be measured with two of the most widely used datasets in the field – PTS and CIRI. Gibney’s most updated PTS scores will be used as measurements of overall, aggregate levels of human rights violence in a given country-year (Gibney & Dalton, 1996). The PTS will be used here instead of CIRI, because Cingranelli, Richards & Clay’s (2014a,b) dataset uses a potentially problematic technique to aggregate different rights into one index: by summing up the frequency of different categories of rights abuse, the dataset’s aggregation method leads the user to assume that ‘an act of torture is equivalent to a disappearance or that an extrajudicial killing is equivalent to an instance of arbitrary imprisonment’ (Wood & Gibney, 2010: 377). Moreover, as the most widely used dataset in the human rights scholarship, PTS scores are known for their consistency. Having had Mark Gibney as its principal coder since 1984, the coding process reduces the chance of new subjective biases being introduced over time (Wood & Gibney, 2010: 372). Gibney and Dalton created two types of PTS scores based on two different sources: the US State Department’s annual reports on human rights and Amnesty International’s annual reports. As the two measurements will have different biases, both will be used in the analyses. For ease of comparison with other human rights indices, the PTS scores here are inverted, so that high scores represent improved human rights conditions.

To investigate the impact of the independent variable on specific types of human rights, empirical discussions will also employ the CIRI dataset (Cingranelli, Richards & Clay, 2014a,b). Like PTS, CIRI measures ‘state terror’, and, like PTS, it is coded using the US State Department’s annual reports on human rights and Amnesty International’s annual reports. What makes the CIRI dataset stand out is that it allows us to investigate specific types of rights. CIRI separately codes a number of different human rights violations: disappearances, extrajudicial killings, political prisoners, torture, freedom of movement, workers’ rights, and women’s economic, social and political rights. For the purposes of this article, the analyses will specifically deal with CIRI data on extrajudicial killings, political prisoners, torture, workers’ rights, women’s economic rights, women’s social rights and women’s political rights. For all CIRI indices, higher scores indicate better human rights conditions.

Thus, in total, the empirical analysis will use eight different measures of human rights as its dependent variables: PTS, extrajudicial killings, political imprisonment, torture, workers’ rights, women’s economic rights, women’s social rights and women’s political rights.

Independent variable: State capacity

As with many political concepts, state capacity remains a fuzzy and complex idea that is hard to operationalize. To allow for maximum comparability with preceding works in the literature, this article will use a measure that is widely used by scholars of state capacity: tax as a proportion of gross domestic product (Fauvelle-Aymar, 1999; Lieberman, 2002).

In the state formation literature, the collection of tax is considered a major task of the state. During the process of state-building, a state’s ability to raise a strong military (and thereby protect its territory) is highly dependent on its capacity to collect tax revenue (Hintze, 1975; Tilly, 1975, 1992, 2003; Brewer, 1989). Moreover, as opposed to other methods of raising revenue (such as the sale of natural resources), the collection of tax requires extensive use of a state’s bureaucratic apparatus, since states with less agency loss and greater transparency will be able to collect more revenue through tax. As such, the state’s ability to collect tax is a core measure of its capability to maintain itself.

Because this article’s argument directly addresses that of Englehart (2009, 2017), it will be using the same measure of tax as a proportion of gross domestic product as Englehart uses. It will use the best estimates of tax as a proportion of gross domestic product drawn from the World Bank’s World Development Indicators, the African Development Indicators, IMF Country Reports and the Asian Development Bank.

Interaction variable: Autocracy

Whereas regime type has been included in many previous quantitative studies, to the author’s knowledge it has never been used as an interaction variable. To put it differently, existing works have not fully explored the possibility that state capacity may affect human rights differently depending on whether a country is democratic or not. To facilitate comparison with preceding works in the literature – Englehart (2009, 2017) in particular – this article will create a dichotomous ‘autocracy’ variable from the combined Polity IV score. The variable is created as a dichotomous (0,1) variable so that it can be interacted with the continuous ‘state capacity’ variable.

As this article is transforming the continuous Polity IV scale into a dichotomous variable, the precise cutoff point from which a country is classified as an autocracy may significantly affect the results of the statistical analyses. While no method may completely overcome this issue (as indeed the authors of the Polity IV scale never intended it to be used as a dichotomous measure), this article will use two different definitions of autocracy to mitigate the problem. In order to increase the number of autocratic countries in the sample, one set of analyses uses an autocracy variable that classifies countries as autocracies if their Polity IV scores fall below 1. However, since an autocracy as defined by Polity IV is, strictly speaking, a country whose Polity IV score falls below –5, another set of analyses will use an autocracy variable that classifies a country-year as ‘autocratic’ if the Polity IV score is less than –5.

By interacting state capacity with these autocracy variables, the regressions will reveal whether state capacity affects human rights indices differently in autocracies.

Control variables

A few control variables will be included so that the empirical analyses reflect important developments in the human rights literature. All statistical models will control for civil and international conflict, level of economic development and relations with the international human rights regime. In addition, the models will also include the lagged dependent variable to deal with serial autocorrelation and the (logged) size of a country’s population to adjust for the event-based nature of Amnesty International and State Department reports.

Conflict (both domestic and international) has been found to significantly affect human rights abuse (Tilly, 2003; Poe & Tate, 1994; Poe, Tate & Keith, 1999). Not only can the combatants commit abuse against civilians, but even non-combatant state agents may take advantage of the anarchy and abuse members of the public for their own private benefits. This article will use Lacina & Gleditsch’s (2005) dataset on battle-deaths to account for this variance rather than use dummies for the presence of war, since the ‘number of deaths’ data could also cover variations in the intensity of conflict.

The level of economic development (measured here as the logged GDP per capita in constant 2000 US dollars) has also been found to strongly affect human rights in a positive direction. Economic development decreases economic grievances, which is often a cause of petty crimes. Economic development also allows the government to prevent political challengers from arising, thereby diminishing an incumbent’s need to resort to violence.

In addition, as scholars of the legalist tradition argue, the level of human rights in a country is (at least partially) affected by that country’s exposure to international human rights laws. Following Englehart (2009), the models will include two variables to account for this variation. First, they will control for engagement with the international community using Pevehouse, Nordstrom & Warnke’s (2003) intergovernmental organizations (IGOs) data. Countries are given different scores according to their level of engagement with IGOs: observers are given a score of 1, associates are given a score of 2 and full members are given a score of 3. The scores are then added up for each country-year and divided by that year’s global average to create an ‘IGO involvement’ variable. Second, a dummy variable on whether a country, in a given year, had ratified the first optional protocol of the International Covenant on Civil and Political Rights (ICCPR) will be included in each model. The ICCPR’s first optional protocol allows individuals from each signatory country to file complaints to the UN Human Rights Council when their rights have been violated. Signing the protocol thus strongly binds a country to international human rights standards.

Finally, the models will include lagged dependent variables to control for serial autocorrelation and a ‘logged population size’ variable to adjust for attributes of each dependent variable. Previous human rights records can affect the human rights situation of today, and adding a lagged dependent variable controls for that inertial effect. Meanwhile, population size affects the number of human rights abuses recorded by both the US State Department and Amnesty International.

Empirical analysis

Eighteen statistical models have been developed to empirically evaluate this article’s argument. The statistical models lend firm support to the hypothesized relationships among human rights, state capacity and regime type.

For this article’s theoretical argument to stand, the effect of state capacity on the overall level of human rights should be positive for democracies but not necessarily so for autocracies. In addition, the proposed theory will gain further empirical ground if state capacity’s relationships to different types of specific human rights violations – extrajudicial killing, political imprisonment, torture, workers’ rights, women’s economic rights, women’s social rights and women’s political rights – also exhibit disparities between democracies and non-democracies. Therefore, Models 1, 2, 10 and 11 have been designed to evaluate the effects of state capacity on overall levels of human rights violations, while the other models have been formulated to measure the impact of state capacity on specific types of human rights. The difference between Models 1 to 9 (Table III) and Models 10 to 18 (Table IV) is that an ‘autocracy’ in Table III is defined as a country whose Polity IV score is below or equal to 0, whereas an ‘autocracy’ in Table IV is more narrowly defined as an ‘institutionalized autocracy’, whose Polity IV score is under or equal to –6 (Center for Systemic Peace, 2018).

Overall levels of human rights

Empirical analyses regarding overall levels of human rights (inverted PTS scores) reveal that state capacity’s positive effects are rather limited in authoritarian regimes. In all models, the interaction terms between state capacity and autocracy have negative coefficients, even though the coefficients on state capacity are positive and consistently significant at the 5% level.⁵ Put together, the results imply that the positive effects of state capacity on human rights are reduced when a regime is autocratic. As a matter of fact, in both Models 1 and 10, state capacity makes PTS (US State Department) scores even worse in autocracies. In Model 1, the coefficient on state capacity is 3.98 for countries with Polity IV scores above 0 but −0.44 for those with Polity IV scores below or equal to 0. When we apply a stricter definition of autocracy (Polity IV < −5), the negative effects of authoritarianism on state capacity’s influence on human rights become more noticeable. According to Model 10, the coefficient on state capacity is 4.17 for countries with Polity IV scores that are higher than –6 but –0.47 for countries with Polity IV scores under or equal to –6. In both models, the interaction terms are significant at the 1% level. As seen in Models 2 and 11, the effects are more modest and less significant when PTS (Amnesty International) is used instead as the dependent variable. Still, the directions of the coefficients are in line with those of Models 1 and 9; and the interaction term in Model 11

Table III.

Results of logistic regressions (Autocracy = 1 if Polity < 1)

	Overall levels		Specific types of rights violations
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
	PTS (SD)	PTS (AI)	Extrajudicial killing	Political prison	Torture	Workers’ rights	Women’s economic	Women’s social	Women’s political
State capacity	3.98**	3.41*	3.85*	3.84**	3.76**	3.28*	6.05**	4.06**	3.35*
	(1.43)	(1.68)	(1.55)	(1.48)	(1.17)	(1.36)	(1.58)	(0.96)	(1.53)
Autocracy (Polity < 1)	0.05	0.29	0.64*	–0.37	0.49^†	–0.46^†	–0.27	0.16	–0.20
	(0.26)	(0.27)	(0.28)	(0.26)	(0.27)	(0.26)	(0.42)	(0.29)	(0.35)
State capacity*Autocracy	–4.42** (1.45)	–2.95^† (1.66)	–2.84^† (1.63)	–3.48* (1.63)	–4.52** (1.41)	–3.03^† (1.59)	–1.36 (1.81)	–4.30** (1.41)	–0.44 (1.94)
Civil war deaths	–0.00^†	–0.00*	–0.00*	–0.00^†	–0.00	–0.00	–0.00*	–0.00	0.00
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Int’l war deaths	–0.00^†	–0.00**	–0.00^†	–0.00	–0.00	–0.00^†	–0.00*	0.00	–0.00
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
log (GDP/capita)	0.17**	0.18**	0.25**	0.12^†	0.20**	0.01	0.44**	0.12^†	0.03
	(0.05)	(0.06)	(0.07)	(0.06)	(0.06)	(0.06)	(0.10)	(0.07)	(0.07)
log (Population)	–0.29**	–0.15*	–0.22**	–0.30**	–0.31**	–0.15**	–0.17^†	–0.08	0.01
	(0.05)	(0.06)	(0.06)	(0.07)	(0.06)	(0.05)	(0.09)	(0.05)	(0.06)
IGO involvement/global	0.56*	0.16	0.07	0.41	0.42	0.99**	0.96^†	0.67*	0.65^†
	(0.25)	(0.27)	(0.33)	(0.30)	(0.30)	(0.27)	(0.49)	(0.28)	(0.33)
ICCPR membership	–0.19	–0.05	0.03	0.13	–0.08	0.16	0.29	0.32*	0.45*
	(0.14)	(0.13)	(0.16)	(0.14)	(0.14)	(0.14)	(0.23)	(0.14)	(0.17)
Pseudo R²	0.48	0.37	0.35	0.40	0.32	0.39	0.22	0.59	0.70
Log Pseudolikelihood	–2,035.63	–1,912.36	–1,413.46	–1,421.85	–1,572.55	–1,437.82	–1,879.17	–1,017.28	–570.10
Observations/countries	2,748/161	2,069/157	2,192/147	2,189/147	2,170/145	2,172/145	2,212/147	2,102/145	2,159/145

Robust standard errors in parentheses. **p < 0.01, *p < 0.05, ^†p < 0.1.

Table IV.

Results of logistic regressions (Autocracy = 1 if Polity < –5)

	Overall levels		Specific types of rights violations
	(10)	(11)	(12)	(13)	(14)	(15)	(16)	(17)	(18)
	PTS (SD)	PTS (AI)	Extrajudicial killing	Political prison	Torture	Workers’ rights	Women’s economic	Women’s social	Women’s political
State capacity	4.17**	2.97*	3.44**	2.65*	3.48**	3.06**	6.62**	3.61**	2.52^†
	(1.25)	(1.35)	(1.32)	(1.07)	(1.04)	(1.15)	(1.46)	(0.90)	(1.48)
Autocracy (Polity < –5)	0.40^†	0.43^†	0.59*	–0.64**	0.53*	–0.64*	–0.05	0.24	–0.62
	(0.22)	(0.23)	(0.25)	(0.22)	(0.25)	(0.28)	(0.44)	(0.31)	(0.39)
State capacity*Autocracy	–4.64**	–2.24	–2.61^†	–1.13	–4.63**	–3.13*	–1.87	–3.59*	1.34
	(1.30)	(1.40)	(1.44)	(1.24)	(1.27)	(1.58)	(2.01)	(1.66)	(2.19)
Civil war deaths	–0.00*	–0.00**	–0.00**	–0.00*	–0.00^†	–0.00*	–0.00	–0.00	0.00
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Int’l war deaths	–0.00^†	–0.00*	–0.00^†	–0.00	–0.00	–0.00*	–0.00*	0.00	–0.00
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
log (GDP/capita)	0.20**	0.20**	0.24**	0.16*	0.19**	0.01	0.47**	0.13*	0.04
	(0.05)	(0.06)	(0.07)	(0.06)	(0.06)	(0.06)	(0.11)	(0.06)	(0.07)
log (Population)	–0.22**	–0.11*	–0.22**	–0.26**	–0.29**	–0.12*	–0.16	–0.04	–0.00
	(0.05)	(0.06)	(0.06)	(0.07)	(0.06)	(0.05)	(0.11)	(0.05)	(0.06)
IGO involvement/global	0.53*	0.19	0.11	0.39	0.44	0.97**	0.89^†	0.66*	0.70*
	(0.26)	(0.26)	(0.32)	(0.31)	(0.30)	(0.27)	(0.53)	(0.28)	(0.34)
ICCPR membership	–0.10	0.04	0.02	0.16	–0.08	0.17	0.47^†	0.36*	0.42*
	(0.14)	(0.13)	(0.16)	(0.14)	(0.14)	(0.13)	(0.25)	(0.15)	(0.17)
Pseudo R²	0.46	0.36	0.34	0.38	0.31	0.38	0.23	0.59	0.70
Log pseudolikelihood	–1,968.06	–1,878.77	–1,406.37	–1,429.35	–1,555.99	–1,423.44	–1,510.24	–988.91	–568.88
Observations/countries	2,522/145	2,017/145	2,140/144	2,137/144	2,118/142	2,120/142	2,096/142	2,050/142	2,159/145

Robust standard errors in parentheses. **p < 0.01, *p < 0.05, ^†p < 0.1.

Figure 6.

Marginal effects of state capacity on PTS scores

just narrowly misses the 10% significance threshold (p = 0.108).

As the coefficients of ordered logits are not readily interpretable, marginal effects will be used here to explain the impacts of state capacity in interaction with regime type. According to Model 1, a standard deviation (0.10) around the mean value of state capacity (0.19) increases the probability that a country’s inverted PTS score is highest (best human rights record) by 0.07 in a democracy but reduces it by 0.01 in an autocracy. According to Model 2, a standard deviation around the mean value of state capacity increases the probability of receiving the highest PTS score by 0.04 in democracies but only by 0.005 in non-democracies. Figure 6 provides visual summaries of these marginal effects across a wider range of state capacity. It is worth noting how state capacity’s marginal effects compare with those of other variables explored in the literature. In Model 1, for instance, a standard deviation (1.56) around the mean value of log (GDP/capita) (7.42) results in a 0.05 increase in the probability that a country has the highest human rights score (PTS), which is comparable to state capacity’s effect in democracies.

Specific types of rights abuse

Since the CIRI dataset separately codes specific types of rights abuse, it is possible to take a closer look at how state capacity comes into play in different areas of human rights. Interestingly, models whose dependent variables are specific types of human rights display patterns comparable to those of the above models that analyse overall levels of human rights. Among regimes that are democratic, state capacity has consistently positive effects across different types of rights. By contrast, in non-democratic regimes, the effects of state capacity are rather ambiguous. Put together, these results add further support to the theory at hand. Across diverse types of human rights, state capacity has positive effects in democracies but not necessarily in autocracies.

The coefficients on state capacity are positive when the regime is democratic but, in general, either negative (Models 5, 8, 14 and 15) or near zero (Models 3, 4, 6, 12, 17) in non-democracies. In terms of marginal effects, a standard deviation (0.10) around the mean value of state capacity (0.19) increases the probability of a country-year having no extrajudicial killings by 0.09 in democracies but only by 0.04 in non-democracies (Model 3). Similarly, a standard deviation around the mean of state capacity increases the probability of no political imprisonment by 0.09 in democracies but only by 0.01 in non-democracies (Model 4). The disparity between democracies and non-democracies is greatest for the dependent variable for torture. The marginal effects of a standard deviation around the mean of state capacity is 0.07 for democracies (in both Models 5 and 14) but −0.01 (Model 5) and –0.02 (Model 14) for non-democracies. Figure 7 presents the marginal effects for all the specific human rights variables analysed, across a wider range of state capacity.

There are five exceptions to the pattern: Models 7, 9, 13, 16 and 18. In each of these models, the coefficient on state capacity is positive for democracies and non-democracies alike. While it is not within the scope of this article to investigate exactly why these models differ from the others, it is worth pointing out how these results do not necessarily contradict the theory under scrutiny.

According to Models 7, 9, 16 and 18, state capacity improves women’s economic and political rights in non-democracies as well as in democracies. One reason for this variation could be that dictatorships sometimes seek to obtain legitimacy by promoting women’s rights. For

Figure 7.

Marginal effects of state capacity on specific types of human rights

example, there is a general tendency for left-wing dictatorships to purport gender equality, while right-wing authoritarians are more inclined to maintain traditional, patriarchal social systems. Without taking these policy preferences into account, the interaction effects of authoritarianism will remain fuzzy. Relatedly, another reason might be that women’s rights are more likely to be violated by the unauthorized actions of state agents compared to other rights (such as political imprisonment or torture) that, to be infringed, require the joint efforts of various government bureaucracies. If women’s economic rights improve in more capable autocracies as well as in more capable democracies, it may simply imply that the violations of women’s economic rights are more unauthorized than authorized. To see whether this is indeed the case, we should disaggregate CIRI’s original coding into authorized versus unauthorized instances of abuse against women’s economic rights. However, for the purposes of this article, it suffices to demonstrate that these findings do not present a challenge to the present theory unless augmented with further proofs to that end.

Perhaps more troubling for the theory might be the results of Model 13. In previous discussions, political imprisonment epitomized human rights violations that are authorized by the principal, and the theory expects state capacity to only exacerbate this when the regime is non-democratic. Yet, according to Model 13, state capacity makes ‘no political imprisonment’ more likely for both democracies and non-democracies, potentially challenging the theory’s premise that authorized violations will increase in more capable authoritarian states.

Then again, the seeming anomaly does not necessarily challenge the theory at hand. Political imprisonment is not the only type of authorized violation in a dictator’s toolbox. As Mitchell (2004) poignantly argued, dictators can use various methods of violence depending on their needs, meaning that ‘political imprisonment’ may not be the dictator’s preferred mode of authorized violence in every situation. More stable, better institutionalized autocracies may prefer more nuanced, less noticeable forms of control (e.g. censoring the press or social media) than visible rights violations like the sustainment of massive prison camps. Furthermore, since higher capacity states are likely to have more effective judicial systems, dictators will find it difficult to legally imprison someone for political reasons. Indeed, summary statistics in Figure 4 display an inverted-U relationship between political imprisonment and state capacity. Thus, Model 13 should be interpreted with reference to Model 4, which is identical with Model 13 in all respects except its definition of autocracy: whereas Model 4’s definition includes a wider range of non-democracies, Model 13’s includes only ‘institutionalized autocracies’ (Center for Systemic Peace, 2018). If, in fact, more institutionalized autocracies tend to prefer other modes of political control to ‘political imprisonment’, state capacity would, as Model 13 indicates, reduce the likelihood of political imprisonment among these autocracies.

Conclusion

Are stronger states more humane?

The answer really depends on context. First, the relationship between rights and state capacity is contingent on the type of right under discussion. Authorized rights violations are not mitigated by enhanced state capacity. Unauthorized rights abuses, on the other hand, decrease when the state can keep its agents in check. Second, human rights violations vary by the type of regime. Rights abuses in democracies are mostly those of an unauthorized nature. By contrast, autocratic governments are not restrained by such limitations. Consequently, unauthorized violations can be found in any political regime, but authorized human rights violations are characteristic of autocracies.

As a result of these two variations, state capacity is expected to improve overall levels of human rights in democracies, but in non-democracies the effect will depend on the relative amount of authorized violations vis-à-vis unauthorized ones. To assess this hypothesis, this article conducted ordered logistical regressions on a cross-sectional dataset with nine different human rights indices: PTS (US State Department), PTS (Amnesty International), extrajudicial killing, political prison, torture, workers’ rights, women’s economic rights, women’s social rights and women’s political rights. Empirical analysis provided solid support to the theoretical claims. Overall, the 18 statistical models showed how state capacity positively influences human rights in democracies but not necessarily so in non-democracies. In the models, state capacity affected not only overall indicators of human rights but also five different types of specific human rights indices in the predicted directions, adding strong support to the theory.

There have been five exceptions to this pattern, but none of these critically undermined the logic of the theory. Due to limits of the data, it was not possible to gain a deeper understanding of the three anomalies: as the original dataset did not record the nature (authorized or unauthorized) of each occurrence of abuse, it was impossible to investigate the reasons why state capacity positively affected women’s economic rights in democracies and non-democracies alike. Even so, the empirical analyses at the very least highlighted some important complexities behind a relationship that has not been sufficiently explored. To properly investigate these intricacies, the scholarship might need to compile more data on unauthorized violations of rights. Once datasets like Cornett, Haschke & Gibney’s (2016) Societal Violence Scale (SVS) are in full bloom, future researchers might be able to draw a more accurate picture about the variation among different rights, different regimes, and different levels of state capacity.

Footnotes

Replication data

The dataset, codebook, and do-files for the empirical analysis in this article can be found at . All analyses were conducted using STATA 15 SE.

Acknowledgements

I would like to thank my supervisors, Nancy Bermeo and Ben Ansell, for their support and advice, as well as Andrew Eggers for his comments. I also thank my colleagues and cadets at the Korea Military Academy for allowing me the time and academic resources that were put into this project. But most importantly, I thank my family. In particular, I thank my parents, HyeSun and KyeongSeon, and my loving wife, Mi-Jin. I could not have done this without them.

ORCID iD

Seung Hoon Chae

Notes

References

Acemoglu

Darren

Ticchi

Davide

Vindigni

Andrea

(2010) Persistence of civil wars. Journal of the European Economic Association 8(2–3): 664–676.

Bobbio

Norberto

(1996) The Age of Rights. Translated by Allan

Cameron

. Cambridge, MA: Polity.

Brewer

John

(1989) The Sinews of Power: War, Money, and the English State, 1688–1783. New York: Knopf.

Brysk

Alison

(2005) Human Rights and Private Wrongs: Constructing Global Civil Society. New York: Routledge.

Bueno de Mesquita

Bruce

Downs

George W

Smith

Alastair

Cherif

Feryal M

(2005) Thinking inside the box: A closer look at democracy and human rights. International Studies Quarterly 49(1): 439–457.

Butler

Christopher K

Gluch

Tali

Mitchell

Neil J

(2007) Security forces and sexual violence: A cross-national analysis of a principal–agent argument. Journal of Peace Research 44(6): 669–687.

Center for Systemic Peace (2018) About Polity (http://www.systemicpeace.org/polityproject.html, accessed 15 July 2019).

Chomsky

Noam

Herman

Edward S

(1979) The Political Economy of Human Rights. Boston, MA: South End.

Cingranelli

David L

Richards

David L

Clay

Chad K

(2014a) The CIRI Human Rights Dataset, version 2014.04.14 (http://www.humanrightsdata.com).

10.

Cingranelli

David L

Richards

David L

Clay

Chad K

(2014b) The Cingranelli-Richards (CIRI) Human Rights Data Project Coding Manual, version 5.20.14 (http://www.humanrightsdata.com/p/data-documentation.html).

11.

Cornett

Linda

Haschke

Peter

Gibney

Mark

(2016) Introduction to the Societal Violence Scale: Physical integrity rights violations and nonstate actors. Human Rights Quarterly 38(4): 1102–1108.

12.

Davenport

Christian

(1999) Human rights and the democratic proposition. Journal of Conflict Resolution 43(1): 92–116.

13.

Davenport

Christian

(2004) Human rights and the promise of democratic pacification. International Studies Quarterly 48(3): 539–560.

14.

Davenport

Christian

(2007) State repression and the tyrannical peace. Journal of Peace Research 44(4): 485–504.

15.

Davenport

Christian

David A Armstrong

(2004) Democracy and the violation of human rights: A statistical analysis from 1976 to 1996. American Journal of Political Science 48(3): 538–554.

16.

Easterly

Williams

Gatti

Roberta

Kurlat

Sergio

(2006) Development, democracy, and mass killings. Journal of Economic Growth 11(2): 129–156.

17.

Englehart

Neil A

(2009) State capacity, state failure, and human rights. Journal of Peace Research 46(2): 163–180.

18.

Englehart

Neil A

(2017) Sovereignty, State Failure and Human Rights: Petty Despots and Exemplary Villains. London: Routledge.

19.

Englehart

Neil A

Miller

Melissa K

(2014) The CEDAW effect: International law’s impact on women’s rights. Journal of Human Rights 13(1): 22–47.

20.

Escribá-Folch

Abel

(2013) Repression, political threats, and survival under autocracy. International Political Science Review 34(5): 543–560.

21.

Fariss

Christopher J

(2014) Respect for human rights has improved over time: Modeling the changing standard of accountability. American Political Science Review 108(2): 297–318.

22.

Fauvelle-Aymar

Christine

(1999) Political and tax capacity in developing countries. Kyklos 52(3): 391–414.

23.

Fearon

James D

Laitin

David D

(2003) Ethnicity, insurgency, and civil war. American Political Science Review 97(1): 75–90.

24.

Fein

Helen

(1995) More murder in the middle: Life integrity violations and democracy in the world. Human Rights Quarterly 17(1): 170–191.

25.

Gibney

Mark

Dalton

Matthew

(1996) The Political Terror Scale. Policy Studies and Developing Nations 4(1): 73–84.

26.

Hafner-Burton

Emilie M

Ron

James

(2009) Seeing double: Human rights impact through qualitative and quantitative eyes. World Politics 61(2): 360–401.

27.

Hafner-Burton

Emilie M

Tsutsui

Kiyoteru

(2007) Justice lost! The failure of international human rights law to matter where needed most. Journal of Peace Research 44(4): 407–425.

28.

Hathaway

Oona A

(2002) Do human rights treaties make a difference? Yale Law Journal 111(8): 1935–2042.

29.

Henkin

Louis

(1990) The Age of Rights. New York: Columbia University Press.

30.

Hill

Daniel W

Jr (2010) Estimating the effects of human rights treaties on state behavior. Journal of Politics 72(4): 1161–1174.

31.

Hintze

Otto

(1975) The Historical Essays of Otto Hintze. New York: Oxford University Press.

32.

Huntington

Samuel

(1968) Political Order in Changing Societies. New Haven, CT: Yale University Press.

33.

Keith

Linda C

(1999) The United Nations International Covenant on Civil and Political Rights: Does it make a difference in human rights behavior? Journal of Peace Research 36(1): 95–118.

34.

Lacina

Bethany

Gleditsch

Nils Petter

(2005) Monitoring trends in global combat: A new dataset of battle deaths. European Journal of Population 21(2): 145–166.

35.

Lieberman

Evan S

(2002) Taxation data as indicators of state–society relations: Possibilities and pitfalls in cross-national research. Studies in Comparative International Development 36(4): 89–115.

36.

Lutz

Ellen L

Sikkink

Kathryn

(2000) International human rights law and practice in Latin America. International Organization 54(3): 633–659.

37.

McNamara

Robert S

(1966) Speech given in Montreal, Quebec, 18 May 1966. New York Times 19 May 1966, as cited in Huntington (1968) Political Order in Changing Societies. New Haven, CT: Yale University Press.

38.

Mitchell

Neil J

(2004) Agents of Atrocity: Leaders, Followers, and the Violation of Human Rights in Civil War. New York: Palgrave Macmillan.

39.

Myers

Brian R

(2010) The Cleanest Race: How North Koreans See Themselves – and Why It Matters. New York: Melville.

40.

Neumeyer

Eric

(2005) Asylum recognition rates in Western Europe: Their determinants, variation, and lack of convergence. Journal of Conflict Resolution 49(1): 43–66.

41.

Pevehouse

Jon C

Nordstrom

Tim

Warnke

Kevin

(2003) Intergovernmental Organizations, 1815–2000: A new Correlates of War data set (http://www.correlatesofwar.org).

42.

Poe

Steven C

Tate

Neal C

(1994) Repression of human rights to personal integrity in the 1980s: A global analysis. American Political Science Review 88(4): 853–872.

43.

Poe

Steven C

Tate

Neal C

Keith

Linda C

(1999) Repression of human rights to personal integrity revisited: A global, cross-national study covering the years 1976–1993. International Studies Quarterly 43(2): 291–313.

44.

Regan

Patrick M

Henderson

Errol A

(2002) Democracy, threats and political repression in developing countries: Are democracies internally less violent? Third World Quarterly 23(1): 119–136.

45.

Risse

Thomas

Ropp

Steve C

Sikkink

Kathryn

(1999) The Power of Human Rights: International Norms and Domestic Change. New York: Cambridge University Press.

46.

Rummel

Rudolph J

(1997) Power Kills. New Brunswick, NJ: Transaction.

47.

Sullivan

Christopher M

(2012) Blood in the village: A local-level investigation of state massacres. Conflict Management and Peace Science 29(4): 373–396.

48.

Tilly

Charles

, ed. (1975) The Formation of National States in Western Europe. Princeton, NJ: Princeton University Press.

49.

Tilly

Charles

(1992) Coercion, Capital, and European States. Cambridge, MA: Blackwell.

50.

Tilly

Charles

(2003) The Logic of Collective Violence. New York: Cambridge University Press.

51.

UN Human Rights Council (2014) Report of the Commission of Inquiry on Human Rights in the Democratic People’s Republic of Korea. 7 February 2014, A/HRC/25/63.

52.

Wood

Reed M

Gibney

Mark

(2010) The Political Terror Scale (PTS): A re-introduction and a comparison to CIRI. Human Rights Quarterly 32(2): 367–400.